Cloud-maven

After pasting my incredibly LONG review of this article in this space, I realized it was too long for a blog, so will just put a link here where that strange person who would want to read it will find it.  I’m following in the paths of F. A. Gifford and Fred Sanders, of whom it was said their reviews were sometimes LONGER than the manuscripts they were reviewing!

Review of Freud et al 2015 (monograph sized)

Below is my pathetic attempt to put my review into a viable/readable blog that I will be working on to “clean up”, add figures from the original and “other”.  So look for changes as time goes by.  I don’t expect the pdf to change.   The figures are in the pdf above, btw.

A continuing theme of reviews like this is, as in other cases,  I was not asked to review this manuscript in my expertise before publication, even though I was known by the authors as an expert on Israeli clouds, weather, and cloud seeding.  Perhaps the editor of this journal,  who chose the reviewers did not know???   Right now, its “Ready, Fire, Aim” to quote a book title describing the premature publication of this blog.

Oh, that Israel-4 cloud seeding experiment?!    The result of the seven season randomized experiment was published by Benjamini et al. (2023, J. Appl. Meteor.).  The “primary” evaluation found no viable increase in precipitation due to cloud seeding, in agreement with my 2017 review of this cloud seeding potential paper and all my prior work on Israeli clouds (Rangno 1988, Quart J. Roy. Meteor. Soc.) and cloud seeding (Rangno and Hobbs 1995, J. Appl. Meteor.)

Cloud seeding potential was grossly exaggerated by Freud et al. (2015).  I wanted to scream to the Israel National Water Authority, “Don’t do a cloud seeding experiment based on these findings!”  But, it was too late.

==================================================================================================

A comprehensive review of

“Cloud microphysical background for the Israeli-4 cloud seeding experiment”

by Freud, E., H. Koussevitsky,  T. Goren and D. Rosenfeld,

(the article which is reviewed below was published in 2015 by Atmospheric Research)

with a recap of the Hebrew University of Jerusalem’s cloud seeding experiments

as seen by a long-time observer (me)

1. A. L. Rangno[1]

Opening remarks

“What is scientific knowledge? When is it reliable?”, ask Kenneth Foster and Peter Huber (1997), in “Judging Science:  Scientific Knowledge and the Federal Courts”, MIT Press.  “Scientific knowledge” might be considered anything published in a peer-reviewed journal.  Having passed the peer review “filter” gives, or should give, that which appears in a peer-reviewed journal particular reliability on which our scientific knowledge grows.

Much of this review addresses the second question raised by Foster and Huber, the reliability of published science. Implicitly, this review also addresses the quality of the peer-review “filter” in the cloud seeding domain:  “Has it improved since the 1960s through 1980s when hundreds of pages of faulty cloud seeding claims concerning experiments in Colorado and Israel were published, ones that misled our best scientists?”  In this review of a recently published paper, we must answer, “No.”

In this review[2], we examine the reporting of current and past research by researchers at one of the world’s great universities, the Hebrew University of Jerusalem (HUJ). The researchers at the HUJ in the article to be examined are reporting in a domain of science particularly susceptible to controversy, cloud seeding (e.g., Changnon and Lambright 1990); a field of study afflicted by a quasi-religious “confirmation biases”, and polarization that have corrupted it repeatedly.

Moreover, can the contents in articles published in peer-reviewed journals always be trusted in the polarized domain of cloud seeding with so much at stake for experimenters who report on their own work (jobs, prestige, confirmation of a priori beliefs, etc)?  And, it is experimental work that is unlikely to have attempts at independent replication, our best safeguard against faulty claims, due to the high cost of field experiments.  There may be no more vulnerable field to corruptive influences than cloud seeding due to these factors.

Conversely, pressure arises when you can’t find more rain on your cloud seeding experiment’s seeded days; you may be deemed a failure by others who KNOW that seeding worked.   You may lose funding and your job since you found no response.  This is because the value of a carefully conducted cloud seeding experiment with no viable proof of a seeding effect is undervalued.  These factors converge to produce repeated false “happy” results in cloud seeding publications.

The idea of making it rain or snow on command, as it were, has corrupted many a good scientist since the earliest modern days of this field when Schaeffer dropped dry ice in a layer of Altocumulus clouds in 1946)[3].    The controversy largely results from reports of cloud seeding successes (precipitation increases) by the original experimenters whose findings are subsequently overturned by independent scientists “upon closer inspection” in re-analyses.  This cycle of reverses has gone on since Brier and Enger (1952) right through the era of randomized experiments (e.g., Levin et al. 2010)!   Randomization apparently did not remove experimenter bias.   The Israeli experiments have been subject to this rise and fall cycle.

In Freud et al. (2015)—hereafter, F2015, this rise and fall cycle in the Israeli experiments conducted by the HUJ-CSG is not explained as candidly as it should have been for their Atmospheric Research ( AR) readers and funders, thus prompting this review.  This review scrutinizes all aspects of the F2015 paper which apparently led to additional cloud seeding in Israel.   Because  F2015 was also part of a proposal to the Israel National Water Authority (INWA), it is also reviewed with more rigor for that reason.  Proposals are written by “Party A” to extract monies from “Party B”;  it’s a sales pitch.  When this is done, one might find that Party A has painted a rosier scene for possible accomplishments (his “product”) than is viable for Party B.  In droughty times, it’s always the cloud seeding salesman that wins.  And it’s a win situation for funders who want to show constituents that they are doing something about a drought.

The review of F2015 is organized by topics.   It also contains a “corrective history” of cloud seeding in Israel since Israeli 1.   The history is based on the hope that it will be useful to potential future reviewers of  manuscripts in cloud seeding that might emanate from the HUJ’s “cloud seeding group” (hereafter, “HUJ-CSG”;  referring to those HUJ authors over the decades that have authored or co-authored papers on cloud seeding).  This history may also be of interest to young graduate students within the HUJ science department who may not be informed about it.

Much of this critique of F2015 is due to the authors’ incomplete and misleading descriptions of prior cloud seeding results reported by the HUJ in the abstract and “Introduction” segments.  The former chief editor of Science magazine says it all:

“The difficulty is that positive claims are sometimes made against a background of    unrevealed negative results.” 

—————– Donald Kennedy, 2004 Science editorial:  “The Old File Drawer Problem.”

The word “independent” is highlighted in this review due to the remarkable number of times that outside, independent researchers, when examining the findings published by the HUJ-CSG in peer-reviewed journals, beginning with Rangno (1988-hereafter R88), could not substantiate or reversed them.   This may give the HUJ-CSG the unenviable reputation as a frequent producer of published, but “unreliable, irreproducible” results.  The primary reason is that the HUJ-CSG publishes ambiguous results but describes them as though they were “in concrete”, hiding the ambiguities in their findings that others bring out later.  What should be troubling is that there are likely more ambiguities in HUJ-CSG publications hidden in what appear to be solid findings that have not yet been examined by independent researchers.

HUJ are you listening?

“One-sided citing” is often practiced by the HUJ-CSG, and is seen again in F2015.  Those instances are called out in this review.   “One-sided citing” has recently been condemned in the American Meteorological Society book, Eloquent Science, by David Schultz (2009).  Here’s what Schultz had to say about one-sided citing:

“One-sided reviews of the literature that ignore alternative points of view, however, can be easily recognized by the audience, leading to a discrediting of your work as being biased and potentially offending the neglected authors (who might also be your reviewers!).”  (Yep, me!)

Moreover, there is material damage to your fellow scientists when they are not cited when they should be, as by the HUJ-CSG in this and other cloud seeding articles.  The impact in one’s field; promotions, awards, impact and status, is usually determined by an impact metric, such as the number of citations of your work.

Too, there is implicit damage to the reputation of the home department and institution from which one-sided citing emanates.

HUJ, are you listening?   After all, “the Hebrew University of Jerusalem is Israel’s premier academic and research institution,” as the HUJ states on its web page.  Thus, it should be held to a high standard of research reporting.   The HUJ should be appropriately troubled to read what it finds here.

It is strongly recommended that the HUJ-CSG authors read Schultz’ book, and also that a course in scientific writing and ethics be taught at the HUJ.  (Perhaps many research institutions involved in cloud seeding research would benefit from such courses.)

One-sided citing in F2015 also demonstrates a poor AR peer review process prior to publication, perhaps by seeding partisans (“one-sided reviewing”?)  The obvious poor peer review of this article in the manuscript stage is a further reason why I have troubled to spend time doing a “comprehensive” , “no-stone-unturned” review.

———————–Organization of the Review——————

1)  Overall assessment of this paper and its claims about cloud seeding potential in northern Israel.

2) Who should be assessing seeding potential in Israel?  Ans.  Not the HUJ-CSG, as told by their history.

3)  The shifting cloud microstructure reports over the decades from the HUJ-CSG: how did it go so           wrong in the first place?  (Ans.  “We don’t know yet.”)

4)  The wrongful scientific consensus on Israeli cloud seeding:  how did it go so wrong?

Ans.  Inadequate peer-reviews,  insufficient skepticism and too much trust on the part of scientists reviewing manuscripts on cloud seeding.

5)  A mandated list of publically-available data from the F2015 field program described in this paper,       and why (see above remarks).

6)  “Filling in the blanks”:  Correcting F2015’s incomplete descriptions of Israeli 1, 2, and 3, and operational seeding  (in separate, titled “modules.”)

7) If you want to go further:  the remainder of the original article, absent the abstract and Introduction sections follows in black type with “inline” reviewer’s comments following highlighted statements.

8) After the References follows reviewer’s disclaimers, baggage, conflicts of interest, a prior convictions about cloud seeding, and his background that qualifies him for this review, etc., items that should always be mentioned.

Line numbers have been added so that rebuttals or support for this review can be easily accomplished.

I don’t apologize for the length of this review.  After all, reviews are intended to prevent faulty science from reaching the pages of journals, and help authors in explain their work.   It didn’t happen under AR prior to the publication of F2015.

The excessive length?   The HUJ-CSG has “earned it”; it can’t be trusted in the seeding domain under its current leadership IMO.  If this sounds overly provocative, or even outrageous, read on:

“It didn’t come out of a vacuum[4].”

This review is virtually identical to the one I would have done for AR for F2015 and the HUJ-CSG’s accompanying proposal for Israeli-4 to the INWA if I had been asked.  In polarized research environments, reviews of manuscripts by must be especially intense, every sentence checked for accuracy, every claim suspect until clearly proven;  the discovery of omissions of critical data called out and discussed as in this review.  F2015 encapsulates all that is right and wrong with the HUJ-CSG cloud seeding literature.

The decision from here on the manuscript version of this article would have been:

“Accept, but ONLY upon the authors fulfilling the required major revisions, posting mandated data online, and implementing corrections to their partial descriptions of prior work.”

As a proposal to initiate the Israeli-4 randomized experiment?

“Reject.”

There is insufficient evidence to guarantee a successful outcome.  I review F2015 as a proposal that I myself would be asked to pay for (as an Israeli citizen will be doing).

This “review” will contain compliments and condemnations.   Expect to get mad at someone, maybe me.  The language is sometimes blunt, and the question of misconduct is raised[5].

2. Overall assessment of F2015

 

This article, due to its “Jeckyl-Hyde” properties, presents a unique review challenge; some of the best conscientious scientific writing by the HUJ-CSG is in this very article; the kind of writing marked by caveats and qualifications made in a circumspect manner.

On the positive side, too, it was pleasant and unexpected to read in F2015 that the HUJ-CSG has finally agreed with long standing work, first published by R88, that the Israeli clouds are overall unsuitable for seeding with silver iodide due to their propensity to form precipitation/ice readily.  This finding was re-iterated in Rangno and Hobbs (1995-herafter RH95).   Due to an oversight, or small-mindedness, however, F2015 do not cite the original altruistic work.

In contrast to the positive writing mentioned, there are elements of writing in this paper that cast a dark shadow on the authors, their home institutions, and ultimately, the journal that their article appeared in, AR.   Partial, and therefore, misleading descriptions of the prior Israeli experiments are found in several places, along with omissions of important references to work that was critical of HUJ-CSG findings, thus representing a stunningly “gorgeous” exhibition of “one-sided citing” in F2015, as is often practiced by the HUJ-CSG.

Nonetheless, the HUJ-CSG’s idea that inland clouds reforming over the Golan Heights/Mt. Hermon region might have seeding potential has merit and is worthy of further investigations and corroboration and expansion of the results in this paper by independent groups before any seeding takes place.

Viable seeding potential in the Golan region, however,  is not demonstrated in this paper, as will be shown in subsequent commentaries.

Factors to consider in the extreme northern Israel seeding scenario:  Cloud tops in northern Israel are usually significantly colder than those in the central and southern regions of Israel (Gagin and Neumann 1974, hereafter GN74; RH95).

GN74 reported modal radar tops on rain days in the north target of Israeli 2 near -19°C, whereas in the south target of Israeli 2, they were -16°C.  GN74 (written in 1972) in preliminary analyses, used those temperature differences to explain why seeding was seemingly effective in the north target (more activation of AgI), but not in the south target due to less activation of AgI[6].

Today, however, due to R88, Levin 1992, 1994, and Levin et al 1996, hereafter L929496, and in F2015, we know thatboth cloud top temperatures given in GN74 would be associated with high ice particle concentrations as turrets mature; they would be highly unsuitable for seeding, a situation this reviewer feels would extend into the Golan on many, if not most, storm days.

3)  Who should evaluate cloud seeding potential in the Golan?  Not the HUJ-CSG

Perhaps the most surprising and troubling aspect of F2015  is that the assessment of seeding potential in the Golan is being undertaken by HUJ-CSG; “troubling” due to that organization’s prior reporting history in that domain, i.e., its inability to detect (or hid) the nature of Israel’s clouds for decades with so many tools at its disposal, and the inherent conflict of interest such a new study represents.

Philosopher George Santayana said it: “Those who cannot remember the past are condemned to repeat it.”

Moreover, if the present “background” paper by F2015 is the primary reason why a new randomized cloud seeding experiment, Israeli-4,  was undertaken, then yes, the failed past of the HUJ-CSG’s seeding experiments (Israeli 1, 2 and 3) will be repeated providing that Israeli-4 is evaluated by independent, non-HUJ statisticians and scientists divested of cloud seeding interests!

The evaluation of the clouds in this article in AR is not complete enough to provide confidence that seeding is going to increase rain by an economically worthwhile amount in the Golan.   A new randomized experiment, or operational seeding, both appear to this reviewer to be unwise or at least, premature undertakings in spite of Israel’s urgent water needs.  Cloud seeding “salesmen” are drawn to drought like bees to a jar of spilled honey.  Funders, too, are anxious to open their tills to show their constituents that they are doing something about drought.  It’s a win-win situation for both.  Thankfully, the INWA in a burst of courage, went “randomized” and not into direct operational seeding based on the HUJ-CSG claims.

The sad history of the HUJ-CSG reporting in the cloud seeding domain is reprised for the reader or reviewer who is/was likely unaware of it so that the faults in F2015 can be seen in the context of  “damage control.”   These faults also represent “baggage” carried by the HUJ-CSG and their efforts to recoup the HUJ-CSG’s damaged credibility when they re-write and minimize key elements of that seeding history in F2015.

What “baggage”, you ask?

What if the HUJ-CSG knew for decades that the seeding operations started by the INWA after Israeli 2 in 1975 and  costing millions over the following 30 plus years were actually targeting inappropriate, highly efficiently precipitating clouds rolling in off the Med with little or no seeding effect likely?

What alternative would that group have other than posting one unreliable finding after another that seeding had worked anyway on such clouds?

Is it really possible for the HUJ-CSG to acknowledge error and cover up of that cloud knowledge as likely happened and apologize to the INWA and to the people of Israel?    Absent that apology, the HUJ-CSG group should be terminated IMO.

The appearance of F2015 showed that the HUJ-CSG has yet one more chance (after Israeli 1,2, 3, and in their evaluations of operational seeding) to skew findings to impress an apparently still naïve INWA and their countrymen about  seeding potential in the Golan.

This is the perfect example of the “fox guarding the hen house,” analogous to the Phelps-Dodge Mining Corporation being solely responsible for the environmental impact statement of its next mine.  That Phelps-Dodge assessment could be right; but it would never be trusted as such without outside,  independent confirmation.

Why doesn’t the INWA recognize this inherent, and, obvious to all familiar with the history of seeding in Israel, HUJ-CSG conflict of interest?

Why weren’t outside,  independent groups, such as the University of Wyoming, Stratton Park Engineering Company, NCAR, Droplet Measurement Technologies, Tel Aviv University, etc., brought in by the INWA to make this assessment?  The people of Israel and their media should be asking this question.   The HUJ-CSG forfeited its right to do such an assessment decades ago due to their incompetence in assessing their clouds.

Lastly, because of our experiences with the HUJ-CSG and the unreliability factor in their cloud seeding publications, I URGE that  INWA  have outside, independent groups experienced in airborne work collect data and report on the seeding potential of Israeli clouds in conjunction with critical supporting ground measurements of the nature of ice crystals and precipitation at Mt. Hermon as a next step before experimentation begins. (OK, I’m late to the party, but that’s what I would have recommended to the INWA).

 

4. The shifting sands of the HUJ-CSG cloud microstructure reports over the decades and their shifting implications on cloud seeding

The essential question that should be asked by all of us is, “Why did it take so long for HUJ-CSG to discover the true nature of their highly efficiently precipitating clouds, given all the tools they’ve had available to them since the later 1970s? The tools at their disposal included multiple radars, including a vertically-pointed 3-cm one (Gagin 1980), aircraft, IMS rawinsondes, and the skill of the IMS forecasters who were well aware of the efficiency (shallowness) of Israeli clouds when they begin to precipitate[7].  And this information was there in plain sight for all of us (e.g., GN74, Fig 13.4, GN76, Figs. 2 and 3, Druyan and Sant 1978).

The inability of the HUJ-CSG to detect the nature of their clouds can be either seen as an example of astounding incompetence or one of investigable misconduct if those within the HUJ-CSG knew that their clouds were not as they described them in journals so many times; but, rather, hid that knowledge from their peers and the INWA to keep their jobs, their seeding programs, and misbegotten prestige intact.

Donald Kennedy (2003), in another Science editorial on research fraud:

“In the real instances of research misconduct we know about in biology and physics, the motive appears to have been career enhancement, pure and simple. It is, after all, a competitive world, and the incentive to gain reputation can be powerful. But other motives may appear in those social sciences that bear upon major policy issues.”  (Always was and always will be—reviewer’s comment)

 

You decide.

The reason why the Israeli government is now attempting to once again see if the clouds of Israel can be made to produce economically useful amounts of water via seeding is because none of the prior Israeli experiments produced credible rainfall increases due to seeding.  And that due to the unsuitable clouds for seeding.

5) The false scientific consensus on cloud seeding created by the HUJ-  CSG

 That the HUJ-CSG could not discover the true nature of their clouds until recently, and did not report results of experiments in a timely manner[8], also produced an erroneous “consensus” view in the science community that cloud seeding had been “proven”in Israel (e,g, Tukey et al. 1978, Mason 1980, 1982, Kerr 1982, American Meteorological Society 1984, Dennis 1989, World Meteorological Organization 1992, Young 1993).

So what?

A wrongful consensus damages all of science!   This false “consensus” was published in numerous textbooks, and in countless popular articles along the lines of, “Israelis make it rain in the desert.” Who will retroactively correct those statements?

6. Mandated publically-available data requirements from the F2015 cloud sampling program

Due to the past history of much-published, but faulty, cloud seeding research by the HUJ-CSG,  this reviewer would have mandated prior to publication of F2015 and before any cloud seeding experimentation took place, that the HUJ-CSG provide the following critical data online so that the wider community of experts could have evaluated the findings in F2015.  These should be provided at this time:

1) A table of flight data with dates and times of flights, linked to synoptic maps and satellite imagery for those flight days, and the ability to access flight videos from this program.

2)  A table detailing some of the microstructural measurements, following Hobbs and Rangno (1985), Table 1:

1. a) maximum ice particle concentrations found in each sampling zone on each day over widths of                         of 300 m and 1-km,
2. b) cloud top temperatures and heights of sampled clouds
3. c)cloud base temperatures and heights of sampled clouds
4. d)flight level temperature at which sample was obtained,
5. e) height of the sample below cloud top,
6. f) widths[9] near cloud top of the clouds that were sampled,
7. g) average and maximum liquid water content in each “study” cloud,
8. h) sizes of droplets <13 µm diameter and >23 µm diameter within the H-M temperature zone of                        -2.5° to -8° C  clouds,
9. i) Large size tail of the FSSP droplet spectrum (“threshold diameter”, after Hobbs and Rangno 1985)
10. j) average and maximum droplet concentrations in study clouds,
11. k) estimate of stage that the cloud was in when sampled
12. l)results of any ground work in the Golan/Mt. Hermon
13. m) back trajectories for flight sampling times over the Mediterranean Sea, and over the Golan

The mandated data elements for outside researchers to examine may seem like an undue burden to the HUJ-CSG since some of these require care to obtain.  But these requirements have to be put in the perspective of how much the prior erroneous and incomplete reports by the HUJ-CSG cost not only the people of Israel in ineffective cloud seeding but also by neighboring Arab countries that undertook similar ineffective seeding operations, and finally, by the cost of the failed attempted replication of the ersatz Israeli results in Italy (List et al. 1999).

7. a) The Israeli 1 experiment as described by F2015 and the counter evidence to that description; “filling in the blanks”

F2015 Abstract:  “These clouds were seeded….producing statistically-significant positive effects…”

This partial description of results for the first two experiments in the authors’ abstract should have been removed by the prior reviewers.  Right from the start the HUJ-CSG authors began misleading less-informed AR readers, some of whom will likely not go beyond the abstract.

There are MANY reasons why the Israeli 1 rain increase results are suspect, all known by the authors of F2015.  They deflected the full story about those reasons to, “positive effects.”  We reprise those reasons why no one any longer believes that Israeli 1 nor Israeli 2, have credibility as successes in rain production for AR readers who might have taken the faulty abstract prima facie:

Prior to the appearance of F2015 in AR,  INDEPENDENT  assessments (this cannot be overstressed) of  the statistical results of seeding in Israel  found them questionable at best  (RH95, Silverman 2001[10], Levin et al 2010, 2011, the latter in reply to Ben-Zvi et al 2011).   Those independent assessments are strongly supported by clouds that we now know are unsuitable for seeding due to the copious natural formation of ice at modest supercoolings.  Indeed, the Israeli clouds have never been suitable for seeding in spite of the numerous published reports by the HUJ-CSG to the contrary.

Conversely, it was those erroneous ultra “ripe-for-seeding” cloud descriptions by the HUJ-CSG in the 1970s and 1980s that gave the statistical results of the Israeli 1 and 2 experiments credibility; that seeding really had increased rain (e.g. Kerr 1982, Mason 1982, Dennis 1989).

Thus, the F2015 authors cannot reasonably believe that actual rainfall increases occurred in Israeli 1 and 2 although they inform the AR reader of “positive effects” in the abstract.  Seeding clouds whose high natural efficiency to form ice is as high as that anywhere in the world, cannot lead to statistically-significant results in a “static-style” cloud seeding experiments, as the Israeli experiments were.

Gagin (1986):

“While it is important to record the effects of seeding on rainfall at the ground, the statistical evaluation of this parameter alone cannot constitute an acceptable result of a successful seeding effect.”   Q. E. D.    End of story.

Turning our attention specifically to Israeli 1 and why the indications of rain increases are no longer credible:

1) The clouds of Israel have been found largely unsuitable for seeding as the HUJ-CSG themselves finally have discovered and now report in F2015, in agreement with several other studies, initially by R88.  If there are no suitable clouds for seeding, then statistical results cannot be due to seeding effects, but rather must be due to “lucky” random draws, or other mischief such as omitting data, post-facto cherry-picking controls, etc.

It is worthwhile to reprise another GN74 statement concerning statistical results without a cloud foundation, a well-known principle within the weather modification community:

“…no statistical evaluation will be of real value unless these results are substantiated by detailed physical considerations.”   

It’s the clouds that drive the credibility of the statistical results, not the other way around.

2) Evidence for a Type I statistical error in Israeli 1 was reported by the Chief Meteorologist of that experiment, Mr. Karl Rosner,  who pointed out to Wurtele (1971) that the region that exhibited the highest statistical significance on Center seeded days,  the “buffer zone” (BZ) between the two targets, could barely have been seeded (“5-10%” of the time).  The HUJ seeding unit did not give this “red flag” enough attention; that is, they did not attempt to reconcile the wind analysis by Shimbursky (in GN74) with their own Chief Forecaster Rosner’s evaluation.   In contrast to their Chief Meteorologist’s view, they considered the BZ inadvertently seeded (GN74).

Let us also quote Neumann et al. (1967) on the BZ issue:

“An acceptable and unbiased way of omitting most unsuitable days was to restrict the analysis to the 327 rainy days, defined as days with some precipitation in the buffer zone by the Mediterranean, a location where rainfall is unlikely ever to have been affected by seeding in either experimental area.”

 

3) The BZ seeding issue was examined independently and in more detail by Rangno and Hobbs (1995-hereafter RH95).  In RH95 was concluded, based on their wind analysis when rain was falling in Israel (when seeding would have been expected to be taking place) that it would take a very bad pilot to have inadvertently seeded the BZ when instructed not to do so.  This conclusion due to the very narrow low-level wind envelope concurrent with rain in Israel.

The Shimbursky wind analysis in GN74, on the other hand,  had only a once-a-day IMS rawin launches, which may or may not have been associated with clouds and rain, and thus, could not address the direction of winds solely coincident with rain falling in Israel at the time of the launch as did RH95.

4) Additional evidence for a Type 1 error in Israeli 1 was presented in RH95 due to more rain on seeded days in the immediate coastal zone of Israel over which the seeding aircraft virtually flew, a conclusion that was reached earlier by Neumann et al 1967, also due to logistical considerations.  Quoting Neumann et al (1967):

“After some 2 1/2 seasons of operational seeding experience, it was noticed that flying was effectively limited in such a way as to affect only the interior parts of the two areas.”

 

Gabriel (1967:

 It was claimed that the seeding plane generally did not fly outside visual contact with the coastline so that there could have been no seeding effect near the coast; hence a 10 km

wide coastal strip must have been unaffected.

 

Gabriel and Baras (1970)

The second modification was suggested in 1963 in view of the actual flight patterns of the seeding plane. It seemed that seeding could only effectively reach the interior parts of the two areas and that analysis should therefore be restricted to these sub-areas.

Rosenfeld (1997) in a series of grand speculations concerning Israeli 1, offered an alternative to RH95 and those by the experimenters:  he argued that some of the airborne-released AgI dispersed downward from the initial releases in the southwesterly or westerly flow, was eventually into in a thin offshore flowing layer near the ground.  That portion of the plume, Rosenfeld conjectured,  then went  offshore, was ingested by seedable clouds (which we know do not exist; never did) at the right distance upwind for the AgI to rise up into the those offshore clouds, nucleate at appropriate levels (usually above 3.5 km ASL),  grow into precipitation-sized particles that fell out just in time on the Israeli coast, thus “explaining” the indications of a bias in storms (more rain on Center seeded days on the Israeli coast).

Whew.  Irving Langmuir comes to mind….

On Rosenfeld’s behalf, there is an occasional offshore flowing shallow layer during less vigorous synoptic rain situations, and one largely confined to early morning hours.  But, Rosenfeld provided no statistics on how often that shallow offshore flow occurred while seeding was taking place in Israeli 1.

5) Too little seeding (an average of but 4 h per day and only about a kilogram of AgI, Gabriel 1967) was carried out and dispersed in Israeli 1.   RH95 demonstrated that the area claimed to have had increased rainfall (under the seeding line, sidewind, downwind, in the target) was not commensurate with the amount of seeding material released in Israeli 1, a conclusdion reinforced by unsuitable-for-seeding clouds (R88, F2015).

6) The line seeding carried out by a single aircraft flying 75 km up and down the Israeli coastline was deemed ineffective by RH95 in having seeded enough clouds to have produced an effect; more rigorously corroborated in later modeling studies by Levin et al. 1997.[1]

 

7. b)  The Israeli 2 experiment as described in F2015 and the counter evidence to that description; filling in more blanks

In Israeli 2, the HUJ-CSG, realizing the poor seeding strategy it had used in Israeli 1, added a second aircraft, and 42 ground generators in Israeli 2 (NRC-NAS, Panel on Weather Modification, 1975).  In Israeli 1 there had been but a single aircraft, and four ground generators located in the far NE of Israel (GN74).

The authors omit for the AR reader, the fact that the north “positive” effect in Israeli 2 was found to be  the product of an astoundingly one-sided random draw[2] for heavy storms throughout Israel (Gabriel and Rosenfeld 1990[3]), and also one that affected Lebanon, and Jordan (RH95).   The extraordinary random draw saw the home of HUJ-CSG in the south target experience the most rain of all the stations available to RH95 on north target seeded days!

Ignoring the report of extreme rain in the south target on its control days, those when the north target is being seeded,  Rosenfeld and Farbstein (1992) proposed that dust-haze,  when combined with AgI, had caused overseeding of clouds in the south target; that is, rain had actually been decreased while the north target, having fewer “dust-haze” days, experienced rain increases due to seeding.  This assessment gained wide traction for a time (e.g., Simpson 1990, Presidential Address to the Amer. Meteor. Soc., Young 1993), but like so much HUJ-CSG work, it was unreliable.

Let us quote Gabriel and Rosenfeld (1990) on the rain in the south target:

“Otherwise, one would need to explain why there was so much more rain in the south when the north was being seeded; as of now, no explanation is available, especially as the prevailing wind direction is from the southwest[4].”

Ergo, rain in the south was not “decreased” by seeding, but rather, in no way could seeding overcome the astounding random draw for north seeded days[5].  And, as we expect, the finding of extreme rain in the south target on north seeded days is not mentioned by Rosenfeld and Farbstein.  How could they?

The rain throughout Israel on north target seeded days once again points to efficiently precipitating clouds that are doing just fine without seeding with AgI.

Levin et al 2010 examined Israeli 2 in more detail than did RH95.  They attributed the apparent seeding-induced extra rainfall in the interior of the north target as due to a bias in strong synoptic systems on north target seeded days, one that created a misperception of seeding (see also, Levin et al 2011, reply to Ben-Zvi  et al. 2011).

Ironically, the Levin et al. (2010) findings are strengthened in F2015 when they observed that the intrusion of large aerosols inland from the Mediterranean on windy days in the north increased rain efficiency of inland clouds.

The Israeli 2 experiment had several design options (GN74), the first of which was the “crossover” evaluation, the same evaluation mode as was used in Israeli 1.  The true tragedy of Israeli 2, however, was in the omission of numerical results of random seeding in the crossover scheme, and in the south target by the HUJ-CSG after 1974.  It was an omission that kept the scientific community in the dark about how the full Israeli 2 had actually turned out.

While it is true, as GN81 pointed out, that the larger area of seeding done in the south in Israeli 2 reduced the correlation with the north, thus making seeding effects harder to detect in Israeli 2 compared with the Center/North crossover comparisons in Israeli 1.  Nevertheless, a crossover analysis based on the prior Center target gauges could have been presented with whatever caveats the authors wished to add.  The prior Center target gauges were in zones S1 and S3 of Gabriel  and  Rosenfeld (1990).  However, it is clear from the SAR’s of those regions in Gabriel and Rosenfeld (1990) that the crossover result of  Israeli 2 was not going to replicate Israeli 1.  At this point, if GN81 had done what they should have, they would have displayed that null result, and provided some thoughts on why it had happened.

It took 15 years after Israeli 2 ended, and at that,  spurred by a letter writing campaign begun in the winter of 1986 by the Israeli experiments’ own Chief Meteorologist, Mr. Karl Rosner, to “out” both the crossover results (-2%) and the negative SAR (-15 to -20%) for the south target.  And this was ONLY after the lead HUJ-CSG experimenter passed in 1987.

We have to assume that without Mr. Rosner’s public call for the exposition of the crossover and south target results, they would still be hidden from view within the HUJ-CSG.  Why weren’t authors, Gabriel and Rosenfeld, or others at the HUJ, troubled by this omission over the years after Israeli 2 ended in 1975[6]?  (Ans.  We don’t know why.)  Donald Kennedy does:

“The difficulty is that positive claims are sometimes made against a background of    unrevealed negative results.” 

Furthermore, the early reports of a “confirmation” of the results of Israeli 1 due to the partial reporting of an Israeli 2 “success,” limited to the north target in a  target-control evaluation, spurred the decision by the INWA to begin an “operational seeding” program in 1975 (GN76), one that produced no viable results for more than 30 years when evaluated by independent scientists.[7]  And, perhaps, those Israeli 2 partial results also convince the INWA to begin Israeli 3 in the south target of Israeli 2.

Did the INWA know about the south target results of Israeli 2 before they began Israeli 3?  We don’t know what inspired Israeli 3.  INWA, F2015:  please tell us.

Where, too, was the outside cloud seeding community, one that failed to raise post-publication questions about the results of random seeding in the south target of Israeli 2 following GN81? ( I count myself in this oversight…and blame.)   Statistician Jerzy Neyman, who closely monitored cloud seeding publications and had a reputation for commenting on them, would surely have caught the Israeli 2 omissions had he not passed in 1980.   There is a lot of blame to go around.

Not reporting the south target results of Israeli 2, one that superficially suggested decreases in rain on seeded days also suppressed the inevitable questions that would have arisen:  “How could there be a suggestion of decreased rain on seeded days with such ultra “ripe-for-seeding” clouds (with warmer tops than in the north) as had been described by the HUJ-CSG over so many years?

Without doubt, the wheels would have come off the HUJ-CSG’s seeding train with full reporting of Israeli 2 in a timely manner.  And the report in Science that there had only been “one success in 35 years” in cloud seeding (Kerr 1982) would never have occurred.

We reprise the speculation of Gabriel and Rosenfeld (1990) near the end of their statistical analyses, one that was to be confirmed 20 years later by Levin et al. in 2010:

“The most plausible explanation (for the statistical results of Israeli 2) is one of random variation, with the north-seeded days being more rainy inland, especially towards the northern and southern edges of the experimental region.”

 

7  c)  The reported, primary effect of seeding in Israeli 2: duration:  where it stands today (it’s not credible).

Virtually the entire supposed seeding effect claimed in the Israeli 2 experiment was due to a greater duration of rain, not greater intensity  (GN81, Gagin and Gabriel 1987, hereafter, “GG87”).  Seeding with AgI, they reported,  caused cold-topped (-15°C to -21°C),  non-precipitating clouds to precipitate exactly like natural clouds, thus extending the duration of rain on seeded days (north target evaluation only).     The increase in that one temperature range was a whopping 46%, all of it due solely to extended duration caused by seeding!

That duration finding no longer makes sense in the face of highly efficiently precipitating Israeli clouds.   It never really did upon close,  independent inspection.

 

 

7 d) The Israeli 3 experiment:  its delayed reporting, and its significance

The first interim report about the progress of Israeli 3 randomized experiment from the HUJ-CSG was in 1992 (Rosenfeld and Farbstein), 17 years after it began[8].  Decreases in rain were suggested in that experiment year after year.  The final result, reported by Rosenfeld (1998) at conference, was that after 19 winter seasons and nearly 1000 random decisions there was an indication of a decrease in rain of 8% on seeded days (non-statistically significant).

In F2015, the authors, following in the footsteps of GN76, cannot divulge for the AR reader the indication of decreased rain on seeded days in Israeli 3.  Instead they cloak that result, as did GN76 for the Israeli 2 south target, with the correct but carefully chosen word, “inconclusive”.  “Inconclusive” could refer to increases or decreases in rain that are merely not statistically-significant; reporting indications of decreases, even if not statistically significant,  would have raised more interest.

The 8% suggested reduction in rainfall on seeded days after 19 winters and nearly 1000 random decisions were important on two accounts: that 1000 random decisions after 19 winter seasons could lead to a result so far from a null one after so many random draws,  assuming there really was no seeding effect as Rosenfeld (1998) asserted.  IF Rosenfeld and Farbstein (1992) are right about “dust-haze” combined with AgI, then the HUJ-CSG in the conduct of Israeli 2 and 3, combined,  have decreased rain for 25 years in the central and southern parts of Israel.

Israeli 3 also demonstrated that the clouds of Israel are unsuitable for glaciogenic cloud seeding.  There is nothing substantially different between the clouds that affect the central and southern regions of Israel from those that affect the north in terms of microstructural behavior except that the clouds in northern Israel are overall generally colder from top to bottom.  In both regions they glaciate effectively.

And, who would undertake a randomized seeding experiment for 19 years knowing you might have a natural draw so far from zero that can’t even be overcome by an actual 10% seeding-induced increase in rain[9]?   (One wonders about the quality of the random draw again…)

7 e)  Operational seeding:  the descriptions in F2015 and the counter evidence to those descriptions; filling in still more blanks

F2015:  “Subsequently operational seeding in the north of Israel was carried out between 1975 and 2013.”

The above quote by the HUJ-CSG authors’, writing in 2015 in AR, shows that they cannot bring themselves to report that the “operational seeding” program in the north of Israel, as it was originally formulated, was terminated in 2007 (Sharon et al 2008[10]).    We quote Sharon et al. 2008 for the AR reader:

 

“3b. A revision and elaboration of possible future seeding activities

In view of Kessler et al.’s initial finding in Section 2a and the ensuing controversy, a joint forum of the

National Water Authority and other professionals involved, met early in 2007. After 50 years of uninterrupted seeding activities, the forum decided to discontinue the program at the end of that rainy season (April 2007) and instead, consider the initiation of a new updated experiment, Israeli IV.”

Why was the original program terminated by the INWA and its panel[11]?

Answer:  The independent panel of experts (Kessler et al 2002, Kessler et al. 2006, distilled by Sharon et al 2008) could find no additional runoff due to seeding into the target, Lake Kinneret (aka, Sea of Galilee) over the 30 plus years of seeding.  They also found that the small increases (6%) reported by Nirel and Rosenfeld (1994); Rosenfeld and Nirel (1996) could not be substantiated when more data from later years were used.   This null finding for the operational program by Kessler et al. was later corroborated by independent scientists at Tel Aviv University (Levin et al. 2010, 2011, the latter in reply to the comments of Ben-Zvi et al 2011).

At this point, one wonders who the reviewers of an article in the respected journal, AR, were?

How could they let such partial statements and an article absent vital references appear?  One must also assume that the reviewers for AR were woefully ignorant of the history of the Israeli experiments to have allowed the F2015 distortions reach the print stage of a major journal.  Truthfulness and a full exposition of events is not part of the HUJ-CSG’s understanding of how to compose a manuscript for a scientific journal.   Their manuscripts and publications, in essence, need to not just be reviewed, but “policed” for accuracy.

F2015:  “Additional statistical analyses showed that the orographic precipitation responded most sensitively to the seeding experiments (Givati and Rosenfeld 2005).”

F2015 claimed in the above publication, that air pollution had decreased precipitation exactly by the amount that seeding was increasing it in the interior of Israel (thus explaining the lack of operational seeding effects reported by the independent panel).  The finding by Givati and Rosenfeld (2005)

could not be substantiated independently by either (Alpert et al 2008, Halfon et al 2009a[12], b) nor by Levin et al. 2010.   The pollution claim by Givati and Rosenfeld (2005) was found to be ambiguous at best.

 

It is useful to quote Thom (1957) on how Givati and Rosenfeld came up with their “pollution” findings (as was demonstrated by Alpert et al 2008 and Halfon et al 2009):

 

“If one takes the liberty of choosing among minimum distance controls, he can often find any result for seeding that suits his purpose, either positive or negative.”

That the results of operational seeding as reported by the HUJ-CSG could not be validated by independentresearchers on four occasions (Kessler et al 2002, Kessler et al. 2006, Halfon et al 2009, and Levin et al. 2010, 2011 (reply to Ben-Zvi et al) is one of the most important aspects of the seeding history in Israel.

For those who know this story, we again see one-sided citing in this discussion; the HUJ-CSG in F2015 cannot bring themselves to cite those publications, damaging those independent workers, themselves, and whether they realize it or not, their own home institutions.

HUJ, are you listening?  Isn’t it best if you policed the HUJ-CSG publications instead of someone like me doing it after they’re out in print?

8. A critique of the cloud sampling and its effects on the cloud microstructural properties reported: how unbiased are they?

This discussion will make it clear why F2015 must make videos from their research flights publically available; these are crucial to understanding what they did and how they did it.

Median concentrations in “hard”, newly risen turrets over the Mediterranean, as the authors show in Figure 16 from their Mediterranean cloud targets, produces a bias toward low ice particle concentrations and higher liquid water contents compared to values that would be found as those same clouds as they matured minutes later.   The authors know this.  The authors are to be commended for adding the description of the stage that the Mediterranean clouds were in when they were sampled.   This really helps the experienced reader weight those “medians” in the Figure 16.

But did the authors’ sample “hard” but meteorologically inconsequential, slender, “chimney” Cu as well in Figure 16?  These are ones we know produce little ice/precip compared to their wider brethren even if they are at the SAME cloud top temperature (e.g., Schemenauer and Isaac 1984, Rangno and Hobbs 1991).  Cloud top width, in particular, is an extremely important metric and should be listed in an online supplement by F2015.

Some background on the width issue:  the apparent error of penetrating isolated, likely “chimney” Cumulus with “smooth, hard tops” Cumulus clouds instead of the wider complexes that produce appreciable rain in Israel was one that was made in the original cloud reports coming out of the HUJ-CSG in the 1970s and 1980s (e.g., GN74, GN76, Gagin 1975, 1981).

Due to that sampling strategy, it misled those HUJ-CSG scientists on the ice content in their clouds, and ultimately their journal audiences.  In contrast to those early cloud samples, GG87 reported that the average duration of a shower in Israel was 23 minutes.  With just a 10 ms^-1 wind carrying that shower along, that rain would have fallen from a cloud system at least about 13 km long; at least 26 km long if it was carried along in a 20 ms^-1 wind!

So, what DID F2015 sample that was relevant to Israeli shower clouds?  We don’t know.  But its critical that we do know.

Another source of bias toward low ice concentrations can result if the authors, in fact,  did sample wider, high ice-producing complexes, but only penetrated the very newest, “hard” turrets in them (typically on the upwind side) in which the explosion of ice had not yet occurred.

Lastly, another critical metric for understanding the quality of the HUJ-CSG authors’ measurements in Fig. 16 is the height of sampling below cloud top.  Sampling too close to cloud top, say, tens of meters instead of a few hundred meters, also leads to a low detection of precipitation-sized particles (and also why R_e satellite derivations can mislead in newly risen turrets, and in ice-spewing, supercooled water-topped layer clouds common in orographic settings.  This again reinforces the need for videos of the F2015 study.

9. The HUJ-CSG’s continuing trouble with ice

The HUJ-CSG has carried out several airborne sampling programs with modern droplet and hydrometeor probes over the past 25 years, beginning in the middle 1990s (Rosenfeld and Lensky 1998).  Further airborne studies were conducted by Lahav and Rosenfeld (2000), Rosenfeld et al. 2001), and by F2015.   Maximum ice particle concentrations and the degree of ice multiplication were not divulged by the HUJ-CSG in any of these publications!

In F2015, we finally have ice particle concentrations!   However, what we get are unsatisfactory-for-seeding-potential, “median” concentrations obtained  in “hard” turrets for those clouds in the Mediterranean.   This prevents us from understanding the “ice life cycle” of Israeli clouds.

Ice particle concentrations, and their origin,  has been one of the continuing mysteries of cloud microstructure (e.g., Mossop 1970, 1985);  in particular,  a primary enigma is how secondary ice forms in clouds with top  temperatures warmer than about -15°C, as occurs prolifically in Israel.    AR readers are likely aware of the 2017 Amer. Meteor. Soc. Monograph  (Field et al.[13]) that focused on the origin of secondary ice and the puzzle it still represents.

Israel is in a region of confluence of various aerosols that affect clouds.  Reports of the degree of ice multiplication from Israel would have surely helped fill in some microstructural blanks in our knowledge.  What a shame the HUJ-CSG can’t address this question in their cloud sampling programs!

Too, where are the droplet spectra and graupel concentrations in the Hallett and Mossop secondary ice-producing temperature zone of -2.5° to -8°C?  Nowhere to be found.    More missing pieces to the knowledge required to understand the clouds of Israel.    Why does it feel like I am reading a paper written in 1968?  The authors could have done better with the “black glove” technique, or a foil sampler (Koenig 1963).

From the reviewer’s experiences in 1986, ice multiplication in the clouds of Israel is rampant.  But, remarkably, only L929496, has addressed this issue in Israel been addressed over the past 43 years since Gagin (1975). The study of ice multiplication, a staple of airborne cloud microstructural studies over the past 60 years (e.g., Mossop et al 1972).  T he HUJ-CSG seems to have trouble since the early 1980s of honoring this standard.

This discussion above begs the question about why F2015 didn’t target mature and dissipating portions of Mediterranean clouds after high ice particle concentrations had formed?  Its inexplicable.  The statement by F2015 that mature, heavily precipitating clouds were avoided due to aircraft safety considerations is not credible to us researchers in airborne studies.  Heavy icing, graupel, hail are found in the younger turrets, not in the maturing ones where icing is subdued due to the formation of high ice particle concentrations that consume the liquid water.

Neither can the HUJ-CSG’s median ice concentrations be compared to the concentrations of ice in Israeli clouds reported by L929496 who was somehow able to sample the high ice-containing regions of Israeli clouds that F2015 found so dangerous.

Specific examples of HUJ-CSG reporting from airborne work over the years since Gagin 1975:

Rosenfeld and Lensky (1998) flying on shower days in pursuit of a comparison between satellite and aircraft measurements of R_e, wrote that they  did not carry a 2-DC probe for hydrometeors on their aircraft on the two flights with showers (!).

Lahav and Rosenfeld (2000), in a ten flight sampling program having a 2-DC probe, nevertheless refrained from reporting ice or other hydrometeor concentrations, while titling their paper, “Microphysical Characterizations of the Israeli Clouds from Aircraft and Satellites.”    But that’s not what they did, or, at least reported.  They only reported partial results.  Sound familiar?

Rosenfeld et al. 2001 had a 2-DC on their research aircraft, but once again, refrained from reporting  concentrations of hydrometeors in the clouds they sampled.  What they did report was that there was a “large number”, or that they were” increasing” or “decreasing.”  What’s going on here?  How does such reporting like this make it into a prestigious journal like PNAS?

This reviewer’s guess, from his 1986 Israeli field project, is that the HUJ-CSG has found in those several airborne projects mentioned above that they have an “embarrassment of ice particle riches” and wants to keep those concentrations and the degree of ice multiplication away from AR readers and funders who might consider cloud seeding.   Ice multiplication in clouds has always been considered an impediment to successful “static” glaciogenic cloud seeding (e.g., GN74, Dennis 1980).

Is there another motive by the HUJ-CSG for keeping maximum concentrations from us?  Yes, another critical omission of results by the HUJ-CSG.    To repeat:

“The difficulty is that positive claims are sometimes made against a background of       unrevealed negative results.” 

This is why, again, that  independent groups are critical in evaluating seeding potential in Israel!

INWA, are you listening?

Gagin (1975), in support of ripe for cloud seeding clouds, claimed that ice particle concentrations do NOT increase with time.  This finding could have been evaluated, too, at many points in the HUJ-CSG airborne studies.

Even more relevant is that those very same ice-filled clouds over the Mediterranean Sea are going to be swept downwind and into the Golan within ~30-40 min during rainy spells, making it doubly important to have reported those maximum concentrations and followed their evolution downstream in route to the Golan.

In sum, yes, you CAN mislead with an aircraft, even one with a full complement of 21^st century cloud microstructure instruments via omission and sampling biases.

The points raised in the foregoing section concerning the airborne sampling carried out by F2015 emphasizes all too well why in-flight videos must be made publically available as they are at the University of Washington and also for a number of other projects at NCAR.

To summarize the dark history of the HUJ-CSG:   over a period of several decades: it misled their own people and the world repeatedly about their clouds, withheld statistical results for Israeli 2 for 15 years, results that would have raised so many questions, delayed for 17 years the reporting of results from the Israeli 3 randomized experiment that, too, would have raised numerous questions.   And under its new leadership, that following the passing of Professor Gagin still can’t seem to publish reliable results about their cloud seeding work, either.  Neither can they fully report on their clouds such basic information as the degree of ice multiplication that Israeli clouds exhibit.

Furthermore, and this may the most telling of all, every paper the HUJ-CSG has published and that has been examined by outside, independent investigators, has been found to be unreliable.  Will more unreliable HUJ-CSG claims be turned up in their literature?  It seems inevitable.

One can predict confidently that in the future the HUJ-CSG will, without major changes to its leadership,  and without a more skeptical INWA concerning the claims coming out of that group, “repeat history.”

——–End of “Commentary” on the F2015 Abstract and Introduction—

Following the Abstract and Introduction segments, F2015 improves demonstrably; it is written extremely well in places, representing the best in what we think of as scientific writing.  However, there are still a few lapses and reviewer-required “clarifications” that will be addressed as they appear.  The type in black is that from the original article.

The objective of this paper is to present the available knowledge on  the cloud  properties in northern Israel,  which supported the decision to commence with the Israel-4 experiment, as briefly described above. Section 2 describes thetypical synoptic conditions during the rainy  days  and  the dynamics of the  clouds as they  interact with the sea  and  the topography. Section  3 describes the  methodology of the physical experiment,  the  cloud  physics aircraftinstrumentation,  flight  patterns and  methodology of data analysis. The methodology of supporting satellitemicrophysical retrievals is also given  in this section. The results of the measurements with respect to aerosols and theway they determine the microstructure at cloud base are given  in Section  4.1. The subsequent vertical evolution ofcloud microstructure with height above  cloud  base  and  initiation of rain are described in Section 4.2. The mixed-phase processes and availability of super-cooled cloud water are presented in Section 4.3. Finally, a summary ofthe results and a discussion of the suitability of the clouds over  the Golan  Heights to glaciogenic seeding aregiven  in Section 5.

2. Synoptic, dynamic and macro-physical considerations

2.1. Meteorological conditions

The  synoptic systems that are  responsible for  more than

90% of the annual precipitation in northern Israel occur  with cyclones passing through the  north-eastern part of the Mediterranean Sea, these cyclones are  referred to as Cyprian Cyclones  (Goldreich, 2003). A rain  event typicallystarts with the passage of a cold  front  followed by  a  thermal low  that develops in  the  cold  air-mass behind thefront   due  to  the relatively warm sea and the lee effect of the Turkish  mountains to the north. An upper trough withrelatively cold  air aloft is associated with the  cyclone,  which increases the thermal instability and  favors  thunderstorm  formation over  the  sea. First precipitation over  land  typically starts with the arrival of the cold front, as the air ahead of it is characterized by dry and often dusty air from the Sahara desert. Fig. 2 shows the synoptic conditions on a typical rainy day in Israel, at the time of the cold front  arrival.  As the winds veer from southerly towesterly, the low-level air  becomes moister and  the cloud-base elevation lowers to a typical level of 500-1000 m (allabsolute heights are given  above  sea  level).  The thermal instability reaches its maximum in the thermal low afterthe passage of the cold front. The average wind speeds are often greater than 10 m/s and they contribute to theorographic component of the precipitation. When the low-pressure system moves to the east the instability is graduallyreduced. However, due to the long trajectory of the cool low-level air over the warm sea the instability and moisture supply supports continued rainfall for another day or two.

Minor:  The IMS refers to periods of rain as “rainy spells.”   This is because they often consist of several consecutive days with recurring showers.  In Israeli 1, there was a period of 17 consecutive days with rain.  The F2015 highlighted sentence might be revised to reflect the occurrence of “rainy spells.”

2.2. Cloud dynamics

Fig. 3 illustrates the cloud and  precipitation characteristics on a typical rainy  day in the  air-mass behind the coldfront.  It is a west–east cross-section across northern Israel,  and  it is based on our  observations and  impressionsfrom  the physical experiment. The schematic figure  is intended to illustrate the main features of the cloud  and  precipitation processes that often  take place  as the air-mass travels eastwards across  the land  and  over themountains.

Figure 3 and its caption:  “…clouds in a typical rain day on northern Israel, as a west-east cross section from the Mediterranean Sea to the Golan Heights.  Rain originates in convective clouds over the sea as graupel and hail.  The sea spray from the rough sea seeds the clouds hygroscopically.  Orographic cloud over the Galilee are often seeded by the remnants of convection.  The clouds evaporate into the Jordan Valley and reform over the Golan Heights.”

This figure is over-simplified, perhaps out of necessity due to the many scenarios that would have to be presented to reflect the several stages in Israeli storms as troughs come and go as documented in the comprehensive satelllit studies by Rosenfeld in 1980, 1982, by Gagin 1981, R88, RH95).

F2015 show a higher level stratiform region extending from a line of convection at the Israeli coast in Figure 3.  However,  it is but one snapshot of an incoming storm, not a semi-permanent pattern as the AR reader might infer.  There are many other scenarios that could have been drawn, which could have been as online supplements to fill in the picture of Israeli rainy days more. Figure 3 is really much closer to a depiction of the coastal convergence zone that tends to develop in the later nighttime and morning hours during lesser onshore flow periods (e.g., Neumann 1951, Goldreich 2003).

Satellite imagery reveals that during major events (lines of convection organized by upper level troughs) clusters of Cumulonimbus clouds and stratiform rain areas barge into Israel en masse.  While the roots of the of convection might be lessened when they move inland in wintertime, as F2015 note,  these mesoscale/synoptic scale systems march across Israel with much of the deep, precipitating cloud system intact.  They do not separate completely, as a rule, into an upper layer and lower layer as shown in Figure 3.   Too, some of the rain that falls is due to aggregates in quasi-stratiform debris clouds.

The Stratocumulus over the hill regions in Figure 3 might be accurate for that specific moment of an approaching storm.  However, drizzle, and mist-like rain (due to collisions with coalescence) often falls from such hill-topping overcast low clouds topping out,  as shown in Figure 3,  at about 3 km ASL[1].   It should have been pointed out that such low clouds, “auto-precipitate.” They frequently top out at temperatures where seeding them would not be effective (> -5°C, e.g.,  R88).

In later winter and spring, shallower Stratocumulus clouds tend to lift off the lower hills as temperature rises during the day, and by later afternoon can sprout into small Cumulonimbus clouds (glaciated clouds) under cold trough situations.  Also, due to the local strengthening of the onshore flow during the afternoons, incoming convection can weaken and may disappear altogether in the coastal divergence zone responding to that strengthening flow.

The deepest clouds in springtime are frequently over the inland hill regions rather than offshore as evidenced by an increase in lightning frequency inland during the late winter and spring.

Minor:  Figure 3 also indicates that the bases of the clouds reforming over the Golan are virtually as low as those on the hills of Galilee.  This seems implausible under the loss of precip water and downslope motions in prevailing westerly flow.  Bases should be higher over the Golan, and probably closer to the freezing level?  Or?

As mentioned in  the  previous section, the  lower troposphere during a typical rainy  day  is quite moist and  unstable due to the relatively warm sea (sea  surface temperature never drops below 17 °C) and colder air aloft(typically below + 2 °C at 850 hPa during the rainier days).

The cycle of Mediterranean Sea temperatures from fall to late winter is significant.   It starts out at around 22°C and descends gradually to 17° C.  Cloud base temperatures, hence, water content in clouds and the propensity to form warm rain and copious ice, also are impacted, likely further reducing seeding opportunities in the warmer fall/early winter period.

This favors the formation of fairly  deep convective clouds over  the sea.  This  is also  why hailstorms and  lighting activity are  more probable near the coastline and less common further inland (Altaratz et al., 2003; Goldreich, 2003).

Lightning activity is about equal over the Mediterranean Sea and inland areas in November and March (Altaratz et al. 2003).   This equality likely extends into April as well.

When these convective clouds move inland, they become separated  from   their main sensible heat and moisturesource and  weaken quickly.

To repeat, the roots of convection may weaken, but these major complexes usually don’t separate with a clear zone between upper and lower clouds as shown in Figure 3; its just one of many scenarios.  F2015 do not support their Figure 3 scenario with frequencies of occurrence during the winter.  How does it compare with the frequencies of other storm scenarios?

However, the moist air that is pushed inland by the strong westerly winds is forced upwards  by   the  topographyof  the  western  Galilee,   and therefore an orographic component is added to the weakening convective  clouds.  The  annual precipitation amount in  the upper Galilee peaks at 1000 mm, as compared to the ~ 600 mm at the coast.

When the air continues eastwards it descends about 800-1000 m to the Jordan (Hula) valley, the clouds, unlessthey are synoptically forced,  tend to dissipate and  break up. This leaves the valley   with  an   annual  precipitation  amount  of  only

~ 500  mm.  This is the  area  where often  visual  flight rules  can apply  and  below-cloud aerosol measurements can  be  made safely.

The slopes of the Golan  Heights force  the air to rise  again and  to produce new  orographic clouds with smallerdroplets.

Check+

The crest  of the Golan Heights rises  gradually from  a height of 400  m  in  the south to  1200  m  in  the  north. Further north, Mount Hermon rises  up to an elevation of 2800  m. The Golan Heights  are   45–70  km  away   (shortest  distance) from   the coastline and  the  clouds there are  normally less  convective than the clouds overthe Galilee and over the sea, because much of the instability is consumed upwind.

 “…normally less convective… ”  True,  except maybe for November  and spring when convection over inland regions is apparently more evenly distributed if lightning occurrences are an indicator (Altaratz et al 2003).  What are the seeding implications of enhanced convection in the Golan?  Do the authors know?  Likely less potential if deeper clouds in the Golan glaciate at modest supercoolings.

The  terrain over  the Golan  Heights is also  less  complex than over  the Galilee  and therefore the flight throughthe  clouds over the  Golan Heights is often relatively smooth. The top of the ridge is only about 10-15  km  downwindfrom  the foothills where the  clouds form. This leaves the clouds little time to  convert their water into precipitationbefore they start to dissipate and lose their water content back to the atmosphere.

We reprise Figure 1 from F2015:  Isn’t the target too small and close from seeding release points for seeding to result in a fallout of precip on it?  Where does the water go in the lee of the Golan, to Syria or into Lake Kinneret?   (This diagram should have had a kilometer scale.)

GN74 estimated 30 min for precip fallout from line seeding, as do F2015.   In the HUJ-CSG publications we find various distances from the aircraft line seeding at which the fallout of seeding-induced precipitation may have occurred, possibly due to variations average wind velocities used .  These ranged from 25-35 km downwind (Gagin and Neumann 1976;  35-50 km  (Gagin 1981); 25-45 km (Gagin and Neumann 1980) ; “about 40 km” (in GG87) to “25-35 km” in F2015.

So which is it?

 And what wind speeds were used in F2015?  And why?  What climo did you use?  The estimate contained in Gagin 1981 (35-50 km) appears to be too great a distance for a precip fallout for this project to be successful except in deep southwest flows.

What are the flight levels of the line seeding aircraft?  Why is this information not in F2015?   This is critical information since if those seeding lines are too low, the seeding material will not reach the necessary heights for nucleation and precip fallout in time to help runoff.  (Perhaps this information is contained in the full proposal?)

How can airborne or ground releases target the region north of the dashed line in Figure 3, including the wettest location, Mt. Hermon, when seeding plumes have to climb to between 700 and 500 mb levels to nucleate effectively and when the primary wind flow is westerly?  In winds even a little north of westerly, it does not appear that seeding is going to be viable in the wettest portion of the Golan unless you get help from Lebanon.

And isn’t the downwash into the Hula Valley going to be detrimental to seeding the Golan?

See the wind rose for 850 mb for those times that rain is falling in Israel from RH95.  It would appear that  many periods of precip  in the north part of the Golan, anyway, will not be able to be seeded efficiently.

Lastly, seeding can reduce upslope precipitation by reducing the degree of riming on upslope precipitation.  Riming is reduced as new, tiny AgI ice crystals form reduce or deplete the upslope supercooled LWC.  The trajectories of ice crystal/snow precipitation are less steep, raised since riming accelerates the downward fall of snow and that will be lessened in seeding.  Will it be made up by more ice crystals?

Ground measurements of precipitation should have been incorporated into the proposal to the INWA or mentioned in F2015 in supporting their contentions about seeding potential in the Golan.

In case there are synoptically-forced or mature clouds above, which survived the  descent to the Jordan (Hula)valley (Fig. 1), they often  precipitate through the orographic clouds and  seed  them naturally from  the top.

What is the frequency of occurrence of this scenario?

The  annual precipitation amount  over   the ridge   increases from  600 mm  in the south to more than 1200  mm  over Mount Hermon to the north, mainly depending on the surface elevation (Goldreich, 2003).

2.3. Aerosol dynamics

The trajectory of the air  mass  that is associated with the Cyprian   Cyclone  passes various regions on  its approach to Israel. Each of these regions leaves  its signature on the aerosol properties. One of the  common features of the warm sector of the approaching low-pressure system is the reduced visibility due to increasing loads ofdesert dust in the air. The winds with the southerly component get stronger as the cold-front comes closer,  lift the dustparticles from  the  surrounding deserts and keep  them airborne. The low-level convergence raises the dust particleshigher up,  so they can  travel greater distances. The dust concentrations are normally higher in central and southern Israel  than in  the  northern part.  This  is due  to  a number of reasons: 1) Northern Israel is farther away  from  the main dust sources; 2)  The  south-westerly flow  ahead of the  cold  front travels some distance over  the south-eastern Mediterranean before arriving to northern Israel, and is occasionally associated with pre-frontal rains   that can wash out  some   of  the  dust particles; and 3) The Israeli deserts in the south, where sand and dust  storms are  common, can  also  contribute to the aerosol population in central and  southern Israel.  After the passage of the cold-front with the arrival of the cool air, the visibility tends to quickly  improve.

Dust particles can  act as cloud  condensation nuclei (CCN) and  may  also  serve   as  natural ice  nuclei   (IN). The  seeding potential of the  clouds is therefore expected to depend on the presence and  the concentrations of thedust particles in  the boundary layer   and   in  the free  troposphere. This  was  the reason, according to  Rosenfeldand  Farbstein (1992), to why they found a positive seeding effect both in northern and  in central Israel when theystratified the experimental data  based on dusty and non-dusty days. Consequentially, one can assume that therelease of silver iodide may have the intended effect of accelerating the precipitation formation only when and  where natural IN supplies are limited and ice does  not readily form in the super-cooled clouds.

Another common and  important aerosol type that is quite abundant when strong westerly winds are prevailing is thesea salt aerosols. The breaking waves and  rough seas trap bubbles of air in the sea water. As the bubbles float backand  reach the surface of the water, they burst the thin film of the seawater and release small  drops into  theatmosphere — the sea spray.  The largest drops may quickly fall back to the sea, but those that stay airborne just 5

long enough have  the chance to remain in the air for a much longer time period as they evaporate and get smaller and  more concentrated. The vertical mixing in the  boundary layer  together with the convergence associated withthe low- pressure system assist  in raising these sea salt particles to the cloud  base  level.  These  fairly coarse and hygroscopic particles are the first to act as CCN and  make the largest droplets in the cloud, which serve  as embryosfor subsequent raindrops.

Concerning “sea spray” and large droplet formation in Israeli clouds:

The discussion of sea spray and bubbles strongly resembles that of Woodcock (1953).  These researchers stand upon the shoulders of Woodcock but do not cite him.  This again points to weak reviewing of the pre-publication manuscript.  We quote Woodcock (1953) below:

 “It is suggested that bursting air bubbles in “white caps” on the open sea are a major source of the salt nuclei, and that a greater portion of the sea surface may act as a source of these particles during average winds than might be judged from the relatively small area usually covered by white caps”.

 

We note with interest that Woodcock was likely wrong in his initial 1953 finding (e.g., Woodcock et al. 1971).Woodcock’s finding that large particles are more numerous on “average wind days” rather than days associated with a large number of whitecaps, supports the occurrence of warm rain on more days than just those with numerous whitecaps and stronger winds in Israel.

More work is needed on this finding by F2015, preferably, again, by independent groups.

L929496 should be cited here.  In Levin’s studies it was found that large CCN comprised of sulfate-coated desert particles led to large droplets in Israeli clouds starting at cloud base.  He did not report on sea surface conditions, however.

————————

However, rather than strong winds and “sea spray”, the occurrence of large droplets in clouds are likely to be modulated by cloud base temperatures and low or moderate droplet concentrations in clouds over Israel and the Mediterranean, with relatively weak cloud base updrafts in which only the largest CCN are activated.

 In an R88 case study, clusters of Cumulus congestus clouds moving in from the Mediterranean Sea on a nearly calm day and produced light rain showers with cloud tops only near 0°C.  Cloud base temperatures on that day were above the average quoted range of 5° to 9°C, about 11°C.

Cloud base temperatures vary substantially in Israel, not only changing as the air mass trajectory changes, but also due to the warm to cool cycle of the eastern Mediterranean from fall to late winter and spring where the temperature can start at 22°C at the start of the rain season,  and ends up at 17°C in mid-late winter.

In RH95, it was noted that cloud base temperatures in Israel varied from 12°C to 5°C.  There would be approximately 40% more water available for condensation with the highest cloud base temperatures compared to the coolest ones, given the average cloud base altitude of about 800 m above sea level.    We note that the authors are aware of this effect of cloud base temperatures, but they do not present those temperatures.

Herut et al. (2000) analyzed the chemical composition of nearly 600  samples of rain  water collected all aroundIsrael during five rainy  seasons. They found that the  sea-salt fraction of the  rainwater composition is influencedmainly by the distance from the Mediterranean Sea, with a decrease from 73% of sea salt fraction in the coastalsamples in the north to 55% in the samples from  the Golan  Heights. They also  reported that the contribution of non-sea-salt precursors to the salinity of the rainwater was much greater in the south due to higher input of continentalcomponents and  lower annual precipitation there.

3. Methodology of the microphysical measurements

3.1. The research aircraft

A twin-engine turboprop Beechcraft King-Air C90 aircraft was instrumented for the  cloud  physics measurementsduring the rainy seasons (November through April) between 2009 and 2013.  At the start of every  season the airplanewas fitted with the aerosol and cloud-microphysics instrumentation as well as the data acquisition and displayinghardware and software. The entire system was  tested on the ground periodically while the airplane was standing-byready for suitable weather to arrive.

3.2. Instrumentation

For measuring the  concentrations and properties of the aerosols we used  a CPC (condensation particle counter), acloud condensation nuclei (CCN) counter and an aerosol spectrometer.

The CPC (TSI model 3781) is a water-based condensation nuclei counter that measures the total concentrationof particles larger than 6 nm  in diameter, at a 1-Hz  temporal resolution  (Hering  et   al.,  2005).  The   simple design,   fast response and  continuous measurement help  detecting variations  in  aerosol concentrations that could  be  related to the atmospheric thermodynamic structure, pollution sources and/ or aerosol nucleation events. It is aninstrument that is fairly easy to handle and  maintain and  is considered reliable.

The  CCN counter that was  purchased from  Droplet Measurement  Technologies (DMT)  for  the  experiment is  a continuous-flow streamwise thermal-gradient counter (Lance et al., 2006). It measures the concentrations of theparticles that were activated into  small  droplets at  a set super-saturation, as well as the sizes of the activateddroplets. The CCN counter can  measure continuously at  a constant super-saturation or alternatively cycle  thoughuser-defined super-saturations for measuring the CCN spectra. The downside of changing super- saturations is that ittakes a couple of minutes for  the temperatures to stabilize and  the actual super-saturation to settle around the  required super-saturation. This also happens when the sample temperature or the pressure changes due to changes in flight altitude. We were therefore flying most of the time with a constant super-saturation (typically 0.5%), exceptduring the dedicated time for measuring the CCN spectrum at a constant flight level below the cloud base (typicallytwo-thirds of the way from  the  surface to the  cloud  base). The effects  of the  sample pressure on  the  actualsuper-saturations as  well  as  the  temperature changes within the  instrument were accounted for during the  data  analysis and quality control.

During   the rainy   season  of  2009–2010  we   had   a  DMT manufactured

aerosol spectrometer (PCASP-X2) onboard (Tan et al., 2010). As opposed to the CPC and the CCN counter, it does not expose the aerosols to  any  super-saturation, but  actually does  the  opposite; it dries  the  air sample for avoidingaerosol swelling due to absorption of water vapor. The PCASP-X2 measures the diameters and concentrations of aerosolsin the range 100 nm to 10 μm. The instrument was  mounted inside the cabin.  The air intake was  isokinetic only  up  toaerosols of about 2  μm, thus truncating the sampling of much larger aerosol than 2 μm.

            Another  DMT instrument on  the   plane was   the Cloud, Aerosol  and  Precipitation Spectrometer (CAPS)(Baumgardner et al., 2000). It consists of two spectrometers (CAS and CIP) and sensors for measuring thetemperature, relative humidity, static and  dynamic pressures as  well  as  a  hot-wire for  measuring cloud  liquid  water content (LWC). The CAS (Cloud and  Aerosol Spectrometer) measures particles and  droplets at the diameter range of 0.5 to 50 μm. The instrument is mounted on a pylon under the wing  and  measures directly the airstreamfrom  the cloud.  The measured aerosol spectrum is therefore sensitive to the  relative humidity. Accounting for thiseffect  is not  possible without knowing the chemical composition of the  aerosols, so we  mainly used  this  probe as asecond cloud  spectrometer for particles larger than 2 μm, as a backup for the main cloud droplet probe (CDP) and forquality control.

The CDP is a DMT-made cloud droplet spectrometer that measures the concentrations and  sizes of the cloud droplets in the  2–50  μm diameter range (Lance et al., 2010). This range is divided into 30 bins, which are muchnarrower than the bins of the CAS (in the cloud droplet size range). Both probes size each droplet that crosses theirsampling volume, based on the amount of light that is scattered forward when the laser  beam hits the droplet.

The DMT-made Cloud  Imaging Probe  (CIP) (Baumgardner et al.,  2000) provides 2-D  images of  precipitationparticles based on their shading pattern on a 62-element array of photo- diodes. The CIP that has been used hadpointy tips, to minimize error due  to shattering of large  particles, and  a resolution of 15  μm so  the  nominal widthof the array  corresponds to a length of 930 μm. The CIP allows identifying the different habits of the ice particles aswell  as distinguishing them from  rain/ drizzle.  It is not possible to directly derive the mass  of the precipitationparticles when ice is present due to their complex form and sensitivity to their orientation. However, the number concentration of the particles (after software partial removal of splinters from shattered particles) along  with particleimages can be useful  for identifying different microphysical phases in the clouds.

Details of the methodology used to remove “splinters from shattered particles” by F2015 is mandatory due to the HUJ-CSG’s prior excess removal of “splinters” from in-cloud measurements that apparently misled them about ice formation in their clouds (e.g., Gagin 1975).  We also note that F2015 used a 2-DC probe with “pointy-tips” as described by F2015 to minimize artifacts in the first place.

So, to re-iterate something that shouldn’t need to mentioned more than once, “Why can’t F2015 report concentrations of ice particles, beyond the medians Figure 16?”

Other parameters that were being   recorded during the flights were the air temperature, the relative humidity andthe navigation data from the  GPS system.

3.3. Flight patterns and execution

During the four rainy seasons of the physical experiment, 27 research flights  were conducted. Each flight lastedtwo and  a half  to three hours on  average.

How different were the synoptic settings?

A table of dates and times of flights is mandatory as are related synoptic maps (surface and 500 mb maps) preferably those embedded with satellite IR imagery.  IMS, aircraft soundings, and, if available, Beirut rawinsonde profiles, should also be made available.   Radar imagery should also be posted online for these flights.

These will help corroborate the authors’ findings; to insure that cherry-picking of a few particular synoptic regimes hasn’t been done to “improve” the overall apparent magnitude of seeding potential.  (One can observe the lack of trust of the HUJ-CSG on the part of the reviewer here in evaluating seeding potential with the baggage they now carry. Sorry, HUJ.)   If they did select only certain flow regimes, because only a few produce targeting possibilities, they should have stated this.

Fig. 4 displays a typical flight track. The black curved line shows the ground path, while the colored line  isprojected and  colored according to the flight altitude. The numbering relates to the geographical locations that arementioned below.

The take-off was  either from  Sade-Dov  airport in Tel Aviv (denoted as point “1a”  in Fig. 4), or from  Ben-GurionInternational airport which is ~ 20 km to the southeast (point “1b”).  The flight typically started with flying out to theMediterranean Sea at low  level  in order to assess the roughness of the  sea and  to safely measure the aerosolsbelow cloud base (point “2” in Fig. 4). This was followed by profiling the deepest clouds in the region from bottom to top,away from local pollution sources over land, while heading north (to point “3”). The profiling was done either at aconstant climbing rate of ~ 500 ft/min in case the cloud layer was  continuous or  by  flying  horizontally through cloudsand climbing stepwise 500–1000 ft in the cloud-free air, in the case of well-defined convective clouds.  After reaching andsampling the tops of the convective clouds over the sea we were normally able to look  eastwards towards the hilly regions of northern Israel, and  identify new  clouds over  the  Galilee. We would then fly to these clouds and profilethem from their tops to the lowest safety flight level (6000 ft) in a spiral  (point “4”).

We note that the sampling height over the Galilee district is about 3500 feet above the bases of the Mediterranean Sea and coastal clouds.

 Sampling “new” clouds might bias ice concentrations to lower values than actually developed a little later. Perhaps the authors don’t mean “new” in the sense of cloud stage?

The next step  was to descend to below the cloud bases over

the Jordan  valley  between the western and  eastern mountain ranges (i.e.,  between the  Upper   Galilee  mountainsand  the Golan  Heights) in order to measure the aerosols inland.

Heights of cloud bases and temperatures should have been given here; preferably listed in a comprehensive table for all flights and different sampling regions.

This was  either done over  the Hula Valley (point “5a”) or over  the Sea of Galilee (“5b”), depending on the  weatherand/or the  air traffic control directions. The CCN counter was set at this point to cycle through three super-saturations(normally 0.3, 0.6 and 0.9%) for approximately 15 min, while we were flying in circles at a constant altitude, and  tryingto avoid  areas with rain. The third cloud  profile  was  done over  the  Golan  Heights, starting with the cloud basesover the eastern edge of the valley and  above  the  slopes (point “6”).  Due to the proximity of the Syrian  border, therest  of the  climb  over  the crest  was  mostly done either in spiral  ascent or in a number of north–south legs, each  approximately 5 min  long,  perpendicular to the  westerly wind direction, until  reaching the cloud top or the heightwhere the  cloud  was  fully glaciated. At that point we typically started heading back south, unless we had a chance tocomplete another profile   or  measurement that we  were not  able  to  complete earlier. Finally we landed at Herzliyaairport (point “7”).

The maneuvering was reasonable in consideration of the realities of the area.  However, as the authors know, sampling on the upslope side of mountains leads to more LWC and less ice than would be found farther downwind.  Again, is radar coverage of the sampling area on the fight days available?  Did the aircraft have recorded radar imagery?  If so, can it be made available for each zone that sampling took place?  Are there ground hourly precip reports?

3.4. Data analysis  and quality control

The main software onboard the research aircraft for real- time data  acquisition was  PADS (Particle Analysis  and  Display System). PADS has been developed and  is maintained by DMT.  This data  was  subsequently processed byour  own  procedures for merging PADS and non-PADS datasets, extending the analysis from   the   research  aircraftmeasurements,  as  shown  in Section  4.2.

Interesting commentary here on what happens to DMT’s PADS processing package.  What exactly is meant here,  “processed by our own procedures”?  What changes, if any, are made from what PADS puts out?  Does DMT agree with this revision to your software’s output?  These may be harmless, but they should be discussed.

4. Results and discussion

4.1. Aerosols and cloud base microstructure

The research aircraft was not equipped with instruments to study the chemical composition of the cloud  and  rain  water. However,  we  noticed that often  on  windy days,  after flying through a cloud,  there were white streaks ofsalt left on  the windshield of the aircraft after the evaporation of the  cloud water streamers.  Fig. 5  shows what theaircraft windshield looked like after passing through a cloud over the Sea of Galilee on 3 Feb 2010, as an example.This is one of the expressions of high  salinity of  the  cloud/rain water in  Israel,  as  had  been studied by Herut et al. (2000) and  mentioned in Section  2.3.

Nice photo.

Another  and  more quantitative expression of  the  abundance of sea  spray  is achieved by comparing the  aerosol size distributions  (ASD)  that  were measured by  the PCASP-X2 below cloud   bases   (Fig.  6).

Each  curve   shows a  60-second averaged ASD at elevations between 400 and  700 m. The black, blue and greencurves show the ASD measured on 28 Feb 2010, which was  a fairly windy day (mean wind of 10.3 m/s  at Haifa Port during the flight  time). The sea  was  rough and  full with white caps — and hence we would expect an extensivedischarge of sea  spray.

We note again that Woodcock et al (1971), in later studies of the chemical composition of rain, did not find the expected association between salt in rain.  We would like to see independent confirmation, of course.

 It can  be  seen  that these three curves have  the greatest concentration of super-micron aerosols — even  when the  measurement was  made approximately 45 km inland (the green curve) over  the Hula  Valley. These  largestaerosols are normally the first to activate into droplets at cloud base at rather low super-saturations, as they act as giantCCN (GCCN).

There  were probably more GCCN than what is shown in Fig. 6, but the experimental setup and the inlet of thePCASP-X2 caused the truncation of  the  ASD at  aerosol dry  diameters greater than 2  μm,  as  mentioned in Section 3.2.  However, despite the slight undercounting of the super-micron particles, it may still be useful to lookinto the geographical differences in their concentrations.

          Fig. 7 presents the  statistics of the  concentrations of super- micron (diameter N  1 μm)  aerosols in  the  marine boundary layer and at the foothills of the Golan Heights measured during four  different flights.   The dataset comprises the 60-second averaged PCASP-X2  concentrations measured at  a maximum elevation of 1000 m above  the ground. In-cloud and  noisy measurements were filtered out, so each  box in Fig. 7 is based on  5  to  20  one-minute  averaged aerosol distributions.  The green text  above  the boxes  denotes the mean wind speed and direction that were measured at Haifa Port at the  time  of the flights.

One can see that the last flight (12 Apr 2010) stands out in the  sense that this  was the  only flight where the super-micron particle numbers at the foothills of the Golan Heights were not considerably smaller than in  the marineboundary layer.  On the other three flights  the large aerosol concentrations over the sea  were 2–5  times greaterthan ~ 45 km  inland. The smallest difference and  the greatest inland concentrations (excluding the last flight) wereon 28 Feb 2010 — the day with the strongest winds. This implies that relatively little time was  available for the  large  particle to  settle on  their way  inland. The  greatest difference and  lowest inland concentrations was  on  26  Feb2010, with the weakest winds out of the three earlier flights.

Levin et al. 2010 attributed their finding of synoptic bias that explained the north target results of Israeli 2 to to stronger synoptic systems with strong winds that drove the precip max farther inland.  The authors’ finding seems to support the idea that large, Mediterranean Sea-derived aerosols also played a role in creating an ersatz seeding effect in the interior regions of that experiment.

The  wind speed at  the shore is not the only  factor that determines the absolute concentrations of the super-micron particles  and   certainly not  their  rate   of  removal  through cloud  and  precipitation  processes.Furthermore,  there could be additional sources of super-micron particles, other than sea spray,  such  as desert dustand  local  pollution. The PCAPS-X2 does not distinguish between the different particles, this allows only  aqualitative estimation of their relative contribution to the observed values.

The flight notes and pictures from the flight of 12 Apr 2010 indicate hazy conditions over central and northernIsrael, with desert dust being  the  probable main constituent of the super-micron aerosols. That is also the day withthe weakest winds recorded at the shore and  lowest concentrations of large aerosols near the coast, so the seaspray  production may be the lowest on that day, compared to the three others. Low sea spray production and highregional dust loading can  explain the relatively low concentration of large  aerosols over sea, and  the small differences between the marine boundary layer  and  the foothills of Golan Heights, as were observed on that day(Fig. 7).

What were the offshore cloud droplet concentrations for 12 April 2010?

Both desert dust and sea salt have  the  potential of accelerating the  precipitation formation, but  through different microphysical processes. The sea salt is an efficient CCN/GCCN, which can  produce large  cloud  droplets and drizzle particles, and   hence  speed  up  the warm rain   processes.  The  larger droplets are also more likely tofreeze and accelerate the rate of secondary ice splinter production (Mossop and  Hallett, 1974). Desert dust particles,on the other hand, may not be as effective CCN as sea salt, but they tend to be more efficient ice nuclei that mightmake cloud seeding with silver-iodide redundant  on dusty days. This may be something to account for in operational cloud seeding.

The sub-micron particles contribute much more to the aerosol total number concentrations than the super-micron particles. A large  source of small  aerosols is local  pollution, especially when  the  air-mass travels inland over  densely- populated areas,  such  as in central Israel. This is demonstrated by comparing the blue and green curvesin Fig. 6 with the black curve — all from 28 Feb 2010. The first two curves demonstrate the fairly  low  backgroundaerosol concentrations on that day because the green curve,  despite showing an  inland distribution,  is from  the rather sparsely-populated north. The  black curve,  on  the other hand, shows the aerosol size distribution downwindfrom  the heavily-populated Tel Aviv area  where the concentrations of  both   sub-  and  super-micron particles increase. But it is the sub-micron population that dominates the number concentrations, thus being  responsible fordoubling of the PCASP-X2 concentrations from  ~ 350 cm− 3 over  the sea to~700 cm−3  about 10 km downwind ofTel Aviv.

The remaining three curves in Fig. 6 show the  aerosol size distributions on relatively calm days; one with hazyskies — due to dust that was  transported from  SW (12  Apr  2010) with mean winds of 7 m/s  at Haifa Port; and  one  with no haze  (26

Mar 2010) and  mean winds of only 4 m/s.  The ASD of 26 Mar

2010   (in   purple)  shows  the  least   super-micron  particles despite being  sampled only  5

km  inland.  With  weak winds and calm seas the production of sea spray was very limited, but local air pollutionprobably caused the high  concentrations of the sub-micron particles. The  red  ASD in  Fig. 6, which was measureddownwind from  the heavily industrial area  near Haifa, had  the highest total aerosol concentration of all other ASDs. There were about 1000 cm−3 particles in the 0.1 to 2 μm size  range.  These  particles make the largest partof the  CCN population.

Fig. 8  presents a  comparison  of  two CCN spectra from another flight. The blue curve shows the CCN concentrations at three different super-saturations (nominal 0.3, 0.6 and 0.9%), as measured over  the sea  on  2  Jan 2012  at  ~ 600  m.

What were cloud base heights, if any?  What time of day was this measurement made?  Onshore or offshore flow?  The Bet Dagan sounding nearest the time of the flight should be presented so we know this case is not a fair weather day with the aerosols clamped down by stable layers.  Also, over what distances were these data collected?

 The  CCN concentration at 1% super-saturation is estimated at 360 cm− 3, while the total aerosol concentrationmeasured by the CPC was 700–1000 cm-3. The red  curve,  on the other hand, shows the  higher CCNconcentrations ~ 45 km inland. The low-level winds on  that day  had  a  clear  southerly component,  which could have  brought the pollution from  the industrial area  near Haifa to the Hula  Valley. The CPC concentrations below theclouds were 3000–3500 cm-3  while the CCN concentrations at  1% super-saturation were close  to 2000  cm-3. The  actual cloud base super-saturation was probably lower, as the highest cloud droplet concentrations in that area  barely exceeded 1000  cm-3 on that day.

The cloud drop concentration and size distribution at cloud base are determined primarily by the properties of theaerosol population (concentrations,  sizes  and  chemical composition) as well  as the cloud  base  updraft. GreaterCCN concentrations and  stronger updrafts lead  to nucleation of more and  smaller cloud  droplets at  the cloud  base.

     Fig. 9 shows a comparison between the shapes of  the cloud  droplet size  distributions (DSD) slightly above the cloud  base  for a windy, a hazy  and  a calm day — all over the sea. All DSDs have a mode near 10 μm, a totaldroplet concentration of ~ 120  cm-3  and a liquid  water content of ~ 0.04 g/m3. However, the difference that standsout between the DSDs is the “tail” of the largest droplets. While  on the calm  day  of  26  Mar  2010,  the DSD is quite symmetric around the  mode,  it becomes more and  more skewed with increasing concentrations of super-micron aerosol concentrations (see matching colors in Fig. 6). On the windy day of 28 Feb

2010,  the concentration of the ~ 20 μm cloud  droplets is about an order of magnitude greater than on the calm  day,with the concentrations of the hazy day in between.

Would like to see more November – December data, that period when the Mediterranean Sea temperatures are warmer; cloud bases, too!

A comparison of near cloud-base DSDs of more than 20 clouds sampled during 12 different flights  over theMediterranean Sea and the Golan Heights, is presented in Fig. 10.

There is no separation between windy, hazy  or calm  days  because the emphasis here is on the geographicaldifferences. Because  the cloud  droplet initially grow  (in  diameter) rather quickly,  only clouds within a  hundredmeters from  the  cloud  base  were considered —  as  long  as  the mode of the  volume-weighted distribution wasbelow 10 μm and the mean LWC was between 0.01  and  0.1  g m− 3.  These  filters were applied in  order to avoid  including DSDs from  clouds whose bases  could  not  be documented due to safety and air traffic controlrestrictions, as well  as in order to exclude highly  diluted clouds.  In addition, the cloud   base   altitudes had   to be within 300  m  in  both locations on the  same day.  This is to prevent large  variations in cloud  base  temperatures that determine the  amount of condensable water vapor near the  cloud  base.

Superlative considerations, but where are the ACTUAL cloud base temperatures?  Nowhere to be found!  The reader will want to know what they were to compare with “historical” reports of cloud base temperatures.  Also, without care, low LWC’s and small droplets can be found in clouds with bases that are evaporating upward providing a false indication of where an original base was.  Here’s where flight videos would be helpful to outside researchers evaluating these claims.

          The integrated LWC of the mean cloud  base DSD over  the sea and  over  the Golan  Heights, as indicated inFig. 10, is comparable (0.052 vs. 0.045 g m-3, respectively). This facilitates the comparison of the  shapes of the observed distributions  because it suggests that the  differences in the distance of the compared samples from  the  actual cloud  base  and/or different exposures to entrained cloud-free air,  are  probably not the main contributors to  the   non-overlapping DSDs in  Fig. 10. This  leaves  the different mean aerosol properties and  cloud base updrafts as  the probable causes for  the different DSDs. However,  the  greater instability over  the sea  isexpected to produce stronger updrafts that would result in a larger number of CCN to activate there — assuming acommon CCN spectra. This would shift the blue DSD in Fig. 10 to the left, i.e. to smaller sizes.  But the mode of theblue  DSD is actually ~ 4 μm larger than the  red  DSD (9.0  vs. 4.8 μm). So apparently the different mean aerosolproperties and their resulting CCN spectra, as the example in Fig. 8 shows, are  the  main factors in shaping the mean near cloud  base  DSD.

The droplet number concentrations that corresponds to the DSDs in Fig. 10 are 225 and 480 cm-3 over the sea andthe Golan Heights, respectively.

These are very credible concentrations from the author’s experience in Israel, except on some haze and smoke,  afternoon “fair weather” Cu days (Cu mediocris, even Cu congestus) in the Golan  (where the reviewer was briefly in ‘86),  droplet concentrations would likely be higher than 480 cm^-3.

The larger number of activated droplets over the western slopes of the  Golan  Heights results in  the smallerdroplet sizes  near the cloud  bases  there, compared to over the sea.

Furthermore, the  proximity of the  Mediterranean clouds to the  source of the sea  spray  tends to result in a tail  oflarger droplets, as the comparison of the mean near cloud base DSDs in Fig. 10 shows. This is due to the activation ofthe giant CCN there (which are much fewer than the rest  of the CCN). The  90th percentile DSDs in  Fig. 10  (filled   circles   above   the curves) indicate that the cloud base DSD over the Golan Heights can also have a considerabletail of large droplet. This is probably caused by  activated sea  spray,   which has  been transported inland withstrong winds, because the  dust particles, as  discussed above,  are  less efficient as CCN.

Here is another finding about large aerosol particles that supports the Levin et al 2010 reanalysis of Israeli 2 and the finding that strong synoptic situations led to the misperception of a seeding effect in the interior of the north target.

 The largest cloud  drops at the tail of the  DSD have  greater fall speeds than the  smaller cloud droplets,  and  therefore they can  collide  with and  collect the small   droplets and   grow   further.  Those  droplets also  have higher probability of freezing in sub-freezing temperatures and hence help  producing precipitation-sized particles moreeffectively. The clouds over the Golan Heights, however, most often have  a smaller tail and  hence convert theircloud  water into precipitation-sized particles by warm processes less efficiently.

4.2. Vertical evolution  of cloud microstructure

The convection that occurs from synoptically or topographically-induced updrafts raise  the cloud  droplets that nucleate at cloud  base  to higher elevations and  colder temperatures.

This is such a common descriptor combination it’s hardly worth pointing out, but a temperature cannot be warm or cold.  Peter Hobbs:  “A cup of coffee can be warm or cold, but not a temperature.”  A temperature refers to an object that is warm or cold.  It itself, a number, cannot be warm or cold, it tells one aspect of the physical state of an object.

The droplet condensational growth is determined by the number of activated CCN and the height above cloud base(Freud et  al.,  2011). The  rate   of  droplet  coalescence is  determined mainly by droplet size and  is related to the 5th power of cloud drop effective radius (re) (Freud and Rosenfeld, 2012). This rate depends also on cloud  dropletspectrum width and  concentrations. When re exceeds 14 μm rain tends to initiate. This process takes normally a  fewtens   of minutes and  requires that the typical convective cloud  in Israel would exceed a vertical extent of 2 to 3 km(Freud and Rosenfeld, 2012).

This process normally takes “tens of minutes”? It’s not is clear what the authors are referring to.  Is “tens” of minutes from the first visible evidence of a cloud?

If not, F2015 need to walk along the beaches of Israel when the skies are boiling with Cumulus to Cumulonimbus transitions as I did.  RH95 show incontrovertible evidence of the rapid glaciation of Israeli clouds, within a few minutes.

But let us assume a Cumulus cloud forms (becomes visible) at 800 m above the Med.  It contains a 5 m sˉ¹ updraft.  From an 800 m it would reach the freezing (typically about 2500 m above sea level on rainy days), in 340 s, and the 700 mb level (3000 m) in 440 s, where the temperature is typically about -5°C.  To reach 4000 m ASL, about 600 mb, would take 640 s since the cloud first appeared above the Med Sea.  The temperatures are now typically-10°C or lower on Israel rain days at 600 mb (4000 m).

      The 700 to 600 mb (4 km ASL) temperatures are typically in the zone where ice is initiated in Israeli clouds (-5°C to -10°C).   At still lower cloud top temperatures, the ice concentrations would increase (in non-chimney clouds).

Ice would be appearing rapidly in such a cloud in just over ten minutes from its initial appearance as a visible cloud, and certainly, if it reached the -12°C to -15°C, would contain hundreds per liter of ice particles from its first appearance, and in just a few minutes after surpassing the freezing level.   Go to your radars and look at the time to first echo.

   But why do we have to go through this simple example?

Please allow me to take your research plane up next winter (with a 2-DC probe that works).  How about if I bring DMT’s Darryl Baumgardner or Duncan Axisa along to insure accurate 2-DC concentrations from the DMT CAPS probe?  I bet that in the first two hours of flight I can find the highest concentration of ice particles that has ever been REPORTED by the HUJ-CSG in 40 years.  I may not find the highest that they have measured; only the highest that they have REPORTED.  (They seem not to want their funders, nor the rest of the science community, to know just how high ice particle concentrations can be in Israeli clouds.  Is there another reason why they don’t report them in airborne study after airborne study?)

The authors, or at least one of them, is well aware of this fast-glaciating behavior of Israeli clouds and has been for at least 30 years (e.g., Rosenfeld 1997; Rangno and Hobbs, 1997.)

Perhaps the phrase, “tens of minutes” was a careless entry in the original ms?

Fig. 11 shows the evolution of the DSD with increasing height in convective clouds over the Mediterranean Sea on 12Jan 2012. The mean DSD of every  cloud  pass is given  as a single  curve  in the plot. The clouds were repeatedlypenetrated close to their tops as they grew,  until  small  precipitation particles (diameter ~100 μm) were detected bythe  CIP at the altitude of 3000 m, on that day.

 How close to cloud top?  Too close to cloud top misses the formation of precip because they lag the smaller droplets at the tippy-top.

These cloud-top “precipitation embryos” become much larger (N 1 mm) as they fall through the cloud and collect othercloud droplets and precipitation particles. On other days, precipitation embryos were detected in the growingconvective  clouds at other altitudes depending on the  specific  conditions (Freud and  Rosenfeld, 2012). However,most of the time that happened when the volume-weighted mode of the DSD exceeded ~ 20 μm.

For comparing the vertical evolution of the DSD for clouds growing in different aerosol environments, it is easier to represent the entire DSD with a single  number,  such  as  the cloud  drop effective radius. This was  derived from  the binned

“EQUATION”

The values of re, the mode of the volume-weighted DSD and the  droplet mean volume radius (rv) are linearlycorrelated as most DSDs have  a generally similar shape (Freud et al., 2008). Freud  and  Rosenfeld (2012)  showed  that  these  relations  are independent of geographical location. They also showed both theoretically and  empiricallythat the  mode of the  DSD and  rv have  values that indicate the  initiation of rain.  This  is  why vertical profiles ofre, derived from satellite retrievals (Rosenfeld and  Lensky,  1998) can  be  used  in  real  time  to  estimate the droplet sizes at cloud  top, as well as their potential for forming precipitation even  before it is detectable by the  precipitation radars. This concept is already in use  in the  operational cloud seeding program in Israel.  An example of the retrieval of the  cloud  microstructure for 10 Dec 2012,  at 10:56 GMT (12:56 LT),  in the form  of re vs. cloud  top  temperature, is presented in Fig. 12.

This retrieval is based on NPP/VIIRS high-resolution data, which allows effective analysis of differences in the  microstructure of clouds developing in rather small user-selective areas (Rosenfeld et al., 2014).

What is the depth of cloud from which these retrievals were derived?  One meter? 10 m? 100 m? 1-km?  This should be stated here for the AR reader.  Ice will certainly fall out of super cooled liquid-topped layers in the Golan; R_e will mislead in those instances.  Another reason for ground obs in the Golan.

 There are three T–re  profiles for different regions in this  figure; the  Mediterranean Sea (area 3), the Galilee (2) and  the Golan  Heights (1).  The median re  (the bright-green curve  in each sub-plot) in areas 3 and 2 reached 15 μm at a temperature of about -5 °C. In addition, the  T–re  profiles in these two areas indicate that  cloud   glaciation  has   occasionally occurred  at temperatures as  warm as  -10  °C.

Can F2015 supply numbers to quantify the “occasionally” statement?  Also, isn’t this old news, about 30 years old, about clouds glaciating/raining when their tops are >-10°C?  (See R88; not cited here, as one would expect by the HUJ-CSG).  Also, from the IMS, 1986:  “We get good rains from clouds with tops at -10°C.”  This quote demonstrates how the every day weather forecasters within the IMS could have prevented the early cloud misinformation published by the HUJ-CSG, that clouds had to be much colder–topped before they rained.

Glaciation is  indicated by spikes of high  re  and  by  a red  tone of the cloudy pixels.

In contrast, the median re  of the clouds over  the Golan  Heights (area 1) did  not reach the  precipitation thresholdof 14  μm (Rosenfeld and Gutman, 1994) and did not show any indication of glaciation. The cloud  tops thereappear to contain mostly super-cooled drops despite developing above  the -10  °C isotherm.

To emphasize, in more stratiform clouds, the top of ice-producing clouds is often supercooled droplet clouds, as reported on several occasions (e.g., Cunningham 1957, RH85, Rauber and Tokay 1991).  The R_e reported here is likely to be misleading.  Ground confirmation is required for assertions of “non-precipitating clouds” as indicated by an R_e value.

Perhaps, since F2015 do not discuss it, they are not aware of the problem in using R_e when precipitating clouds have liquid tops from which ice emanates?   Can they supply hourly precipitation data, radar imagery or weather reports for these times in the Golan?

While there may be, indeed, supercooled LWC at cloud top, it is from that liquid top that ice crystals are spawned, grow, and breakup increasing ice particle concentrations below top (Hobbs and Rangno 1985).   In turn these consume water vapor via deposition/riming in ice supersaturated conditions.  Is F2015 sure they want to disrupt this system by seeding a liquid topped ice-producing layer?

While  Fig. 12  presents a  snapshot of a  single  case,  as  an example of how the different re profiles look in cloudsdeveloping in different areas,  it is useful  to use  a statistical approach and examine additional cases.  Fig. 13 showshow  the  cloud  droplet effective radius varies with the cloud top temperature. It is based on all high-resolution satelliteretrievals from the rainy season of 2012–2013 that were suitable for obtaining the  T–re  profiles for  clouds over theMediterranean Sea and over the Golan Heights. In total,  seven retrievals from  days  with seeding potential, according  to  the  criteria of  Israel-4,  contained  clouds with varying tops in  both areas.   The  cloudy pixels   were grouped according to  their 11.8  μm brightness temperature to  clusters with the same temperature ± 1 °C, and then differentpercentiles of re were calculated for each group. Fig. 13 displays the first and third quartiles for each  cluster and area  as dashed lines, and the median re in each  area  as a solid  line. Only clusters containing more than 100  data  points are  displayed for more robust statistics.

All the curves in Fig. 13 show an increasing re with colder cloud  tops,  but  the  change tends to be more substantialover the  sea. The median re over  the  sea (blue solid  curve) crosses the precipitation threshold of 15  μm alreadyat − 3 °C, even before the  silver   iodide can  have   any  effect,  compared to − 12 °C over  the Golan  Heights (redsolid  curve).

I applaud the authors for this information, highlighted above.  It gives one hope that we may not travel down the same path that the HUJ-CSG has traveled so many times before.  We rephrase the authors’ somewhat obtuse description more simply for the reader:

The clouds moving into Israel from the Mediterranean Sea are unsuitable for seeding with silver iodide. 

This is a new finding for the HUJ-CSG, and it should have been given more attention in F2015.  R88 should have been cited here, who concluded virtually the same thing 27 years before F2015 were able to discover it.

By not citing those who went before, it makes F2015 look like the work of small minds, not that of disinterested scientists only interested in truth.  Omitting relevant work is devious, harms conscientious workers, and degrades the HUJ.

Is this really how F2015 want to represent their home institutions?  It would seem so.

This means that on average, the clouds over the  Golan  Heights have  to acquire a greater vertical extent than theclouds over  Sea to  start precipitating. The probable cause for that is the greater aerosol and  CCN concentration(e.g.  Fig. 8) and  less  sea  salt particles further inland (Fig. 7). But the  different characteristic dynamics andthermodynamics of the clouds may contribute to the differences as well.

As  in  Fig.  12,  Fig.  13  also   shows  indications  of  early glaciation over the sea at rather warm temperatures ofaround − 10 °C,  where  the  75th  percentile of  re   is  close   to  the saturation value  of 40  μm. Furthermore,  thecurves indicate that the clouds over  the sea  tend to  have  a greater vertical extent and  reach colder temperatures.

The temperature bugaboo strikes again.

This is probably caused by the stronger convection over  the  sea  that encourages the formation of  deep clouds,  while the clouds over  the  Golan Heights normally have a more orographic nature and a layered structure, accordingto our subjective observations.

In November, and in later winter and spring, thunderstorms are as common at inland hill regions as over the Mediterranean (e.g., Altaratz et al. 2003), so the idea of perpetual stratiform clouds over the Golan as suggested by F2015 is flawed and needs to be revised to draw this out.  Too, clouds may not be ripe for seeding in these deeper convective situations that occur in November and later in the spring.  Further independent airborne work is needed.

This is also supported by the differences in the  patterns of rain  durations and  intensities between the coastal and hilly stations in Israel (Goldreich, 2003), where rainfall at the coastal plain  is much more intense and has a shorterduration.

It is important to keep in mind that the colored curves in the subplots of Fig. 12 can be considered as verticalprofiles in the convective clouds because the cloud  base  temperature do not tend to vary much within the sameconfined area. re also shows little variance in a given area and altitude (Freud et al., 2008) and the  cloud  top  re  at  agiven  temperature does  not  vary  much during the lifetime of a convective cloud (Lensky and Rosenfeld, 2006). Fig.13, however, is based on seven different profiles, not necessarily with the  same cloud  base temperature and probably with varying aerosol properties and atmospheric thermodynamics. This is why the curves in Fig. 13 cannot really bereferred to as vertical profiles that represent the  development of individual deep convective clouds.  They  should be treated as  a  general statistical view of the T–re  relations in the indicated areas.

These are excellent qualifiers, as scientific writing should contain when necessary.

        The  large number of flight  hours and  the penetration of clouds at  various heights and  in differentgeographical areas facilitate a similar statistical analysis to what is presented in Fig. 13 for satellite data,  with thein-situ data.

The flight hours are really not so large at <81.

What is far more important, to repeat, is the synoptic setting in which the flights actually took place.  As a reviewer, I would have mandated a supplemental appendix with synoptic maps and satellite imagery for each flight be supplied by the authors prior to publication.

I am sorry to say that the HUJ-CSG long ago forfeited the right to make general statements without comprehensive backup material that independent researchers can investigate for reliability.  Furthermore, and I repeat for emphasis, videos of flights should be made available on request or made

available online.  Most of us “senior researchers” who know well the trail of the prior reports from the HUJ-CSG, understand why more evidence is required from them in their cloud seeding reports than might be the case for another institution.  “Fool me once, shame on you; fool me twice, shame on me.”

The 1 Hz measurements inside the clouds represent a spatial averaging along a ~ 100  m  long  path.  Thesemeasurements from  15  research flights  that were conducted during the physical experiment were groupedaccording to the measured temperature at the flight  elevation. The percentiles of the CDP-derived  re  were calculated  for  each   group  of  measurements  and   no  other filtering or normalization was applied except thegeographical separation between the clouds over the Mediterranean Sea and the Golan Heights. Fig. 14 shows acomparison between the re values of the clouds sampled over  these areas.  Here  too the apparent trend ofincreasing re with colder temperatures is more pronounced over  the sea  than over  the Golan  Heights, as  in Fig.13.

The  two profiles appear to  separate around the temperature of − 5 °C, where the marine cloud seems to maintain its convective profile,  i.e. increasing re  with colder temperatures. Over the  Golan  Heights, however, the  effects  ofelevated cloud layers   with embedded  convection and/or  secondary droplet nucleation, may  explain the offset  ofthe peak  between re and LWC (see Fig. 17).  This  is  supported by  the statistics of the droplet numberconcentrations around those peaks (not shown here). The differential likelihood of freezing of larger cloud droplets,elevated cloud  layers,  cloud-top entrainment and precipitation, may  all contribute to maintaining a low  median re attemperatures below − 10 °C.

References supporting the freezing of larger drops could have included, for example,  Vali 1971, Pitter and Pruppacer (1973), among many others.  FYI:   HR85 used the large size end tail of the cloud droplet spectrum (the so-called, “threshold diameter”) as measured by a FSSP-100,  in newly risen, low ice containing Cumulus turrets as a predictor of later maximum ice particle concentrations.  It worked pretty well except in short-lived narrow “chimney Cu”.

The  percentile-profiles of  the   satellite-derived re   of  the convective clouds over  the sea  exhibit greatervariance than those of the orographic clouds over the Golan Heights (Figs. 12 and  13). This is partly caused by theso-called 3D effects  — the variable illumination of the cloud  tops by the sun  due  to their inhomogeneous physicalstructure (Marshak et al., 2006). The cloud   top   re  of  cloud   surfaces that  incline away   from   the satellite, i.e. ina direction that increases the sun–cloud–satellite angle,  would be  overestimated (even more in  shaded cloud surfaces). The cloud  surfaces that incline towards the  satellite would have  their re underestimated due to theopposite effect, resulting in a falsely wide  re distribution for clouds with a non- horizontal cloud  top surface. This, ofcourse, does  not apply  for in-situ derivation of re, as in Fig. 14.

Outstanding distinterested writing by the authors describing issues, both above and below!  Here the paper has the feel of real science!  I’m excited.

Furthermore,  the  absolute re  values that were calculated from  the  CDP measurements,  are  typically smaller inFig. 14 than the satellite-derived re shown in Fig. 13. Possible  sources for this apparent discrepancy include

1) Our  focus  was  on  documenting the  evolution of the  cloud DSD as it grew  and  on measuring the formation ofprecipitation embryos and  the  initiation of effective precipitation.

What is “effective precipitation”?

In addition, we tried to avoid  flying higher than ~ 5 km or in heavy precipitation from  mature clouds due  to flight safety  and  performance reasons.

F2015 stated that they had targeted growing, “hard-topped” Cumulus, filled with supercooled liquid water that would cause heavy icing on their twin-engine, turboprop Beechcraft King Air C-90 aircraft.

Icing is one of the great dangers of sampling Cumulus congestus clouds at below freezing temperatures.  When Cumulus congestus transition to those modest Cumulonimbus clouds of the Mediterranean loaded with ice crystals, the supercooled liquid water disappears and they are virtually harmless to aircraft as an icing threat.

F2015 seem to be indicating that they are unaware of the reduction of icing hazard when this transition to the modest Cumulonimbus clouds, the vast majority are not strong enought to produce lightning.  We’re not sure what hazard they could possibly be thinking about.

 High ice particle regions of clouds have, as a rule, little icing in them and less turbulence, and MANY in Israel in those low theta-e air masses will be found with tops less than 5 km ASL.  In this airborne researcher’s experience in similar clouds in the Washington State coastal waters, such clouds have never posed a hazard to our aircraft except via icing because of sampling too many new turrets too quickly.   (Icing buildups can be removed by descending to lower, warmer levels.)

The highlighted sentence is incomprehensible as a reason not to have sampled the maturing and dissipating portions Cumulonimbus clouds in the eastern Med.

This is why  our  sampling strategy did not cover the area uniformly but instead favored clouds in their early  growingstages — unlike the satellite retrieval.  The aging  of the  deep convective clouds allows more time for  the  collision and   coalescence process to produce large  droplets and  increases the  chances for  ice formation.Therefore  the underrepresentation of the  mature clouds in  the  in-situ measurements and  in  Fig. 14 reducesthe re values.

Qualifying one’s measurements doesn’t get any better than this!  Thank you, authors! Just for the record, this reviewer believes that GN74—written in 1972) represented one of the best examples of objective writing within the often murky domain of weather modification/cloud seeding, where “confirmation bias”, monetary considerations so abound that they deflect true science into something else.

But, this outstanding description begs the question, “How were L929496 able to sample mature, full of ice Israeli clouds, and F2015 were not?”  Isn’t the HUJ-CSG more experienced in airborne research? To the cynical, it could look like an excuse not to sample high ice concentrations in even modest Cumulonimbus clouds.  Authors:  please explain why L929496 was, and you weren’t able to do this.

 2)  The CDP measures only  droplets with a diameter smaller than 50 μm. If larger drops or even  small  iceparticles are present at the  cloud  top or slightly below it (or  even  as a thin cloud layer above  it) then thesatellite-derived re might be overestimated Point (1) above also helps explaining why the blue profile in Fig. 13  is  deeper than the   red  profile,   but in  Fig. 14  it  is shallower. We basically stopped climbing when cloud waterwas mostly converted to precipitation or glaciated. This happened on average at a  slightly warmer temperature  over  the sea,  and hence the shallower profile  in Fig. 14. The clouds over  the  sea often reached much higher altitudesthan over the Golan Heights (e.g. Fig. 12) due  to the  more unstable conditions.

To repeat, the situation highlighted  is often reversed early winter, and certainly during the spring in major troughs when thunderstorms are as common inland as over the Sea.  This needs to be revised to reflect this reality.

The satellite retrievals  reflect this,  but the   in-situ measurements  do  not because of the sampling strategy

4.3. Formation of hydrometeors

Supercooled cloud  water tends to be quite abundant in northern Israeli rain clouds.

F2015 do not provide specific data on this “tends to be abundant” claim anywhere in this paper, nor do they seem to realize that supercooled water is often short-lived in Israeli clouds due to the rapidity of ice formation.

Most of the rain that falls in northern Israel during the winter initiates as mixed phase hydrometeors in clouds withtops  that are  well  above  the  0 °C level.

Omit, “well”;   “above the 0°C level” is sufficient for the “initiation” of precip.  Furthermore, in low stratiform clouds that are thickening upward as an upper trough approaches, drizzle and mist-like rain from warm rain processes are initiated (as was observed on several occasions during the reviewer’s 1986 experiences in Jerusalem).  One suspects drizzle will also occur in hill locations in the far north of Israel as well.  It seems strange to tell people who live in Jerusalem about this!

This again speaks to the critical importance of having independent ground obs at Mt. Hermon;  maybe Har Kana’an, as well.

This has been documented with the help  of the  CIP that provided 2D images of the  hydrometeors in  its  sampledvolume of cloud/ air.  Although  the  CIP does  not  allow  distinguishing between spherical super-cooled liquid  dropsand slightly irregular frozen drops at diameters smaller than ~ 100  μm — due  to its 15 μm resolution, riming offrozen drops and  graupel particles can be easily   identified.  These  particles  provide  clear   indication  of ongoingmixed-phase processes in the clouds.  Quantifying the ice  content in  the  cloud  with a  2D imaging probe is  highly uncertain, however we were still able to see and  document the increasing concentrations and  sizes  of the iceparticles as the clouds matured and the  water content decreased.

Here’s what a Droplet Measurement Technologies rep has to say about the “highly uncertain” claim in F2015 concerning its CAPS probe:

“I agree with you that the authors could have reported the ice particle concentrations from the CIP if they wanted to. I see CIP images in Fig 15 (of F2015), so the authors should be able to get particle concentrations. The CIP does report particle concentrations and processing software is capable of producing high quality particle concentrations.”

Doesn’t this make the “highly uncertain” claim in F2015 a falsehood?  Its does to this reviewer.  But where were the other reviewers, if any, of this article???

The American Meteorological Society (AMS) has recently issued a Monograph (Field et al 2017) focused on secondary ice particles and the efforts to explain them, authored by our leading scientists.  Most of the observations discussed in that Monograph are due to measurements with a 2-DC probe!

A possible cynical translation of the HUJ-CSJ “highly uncertain” claim about ice content:   “We measured really high ice particle concentrations similar to those of L929496  but we,  F2015, wish not to reveal in an article about cloud seeding potential having such high concentrations of ice as we found.  We therefore have invoked the excuse that the “mature” clouds of Israel were too dangerous for our aircraft to sample where all that ice was.  Its amazing to us that Zev Levin was able to do it!”

2-D artifact reduction:  Since the 1970s, there have been ways in software of reducing 2-DC artifacts caused by, say, shattering on probe tips.  Indeed, in F2015 they inform us that they used 2-DC probes with “pointy tips” to reduce shattering artifacts.  But then they reverse course, and tell us they can’t report what they found with that probe.  This is sad, because it appears to be another omission of important data by the HUJ-CSG.

Maybe we shouldn’t be surprised at such an omission, except that it appears in a peer-reviewed journal.   What ARreviewers, ad nauseum,  allowed an omission of such critical data in an evaluation of cloud seeding potential?

If this type of “editing” of findings in F2015 was what convinced the INWA to proceed with rudimentary operational seeding or “Israeli-4”, then the HUJ-CSG has misled them once again.   It is likely, in such an experiment with an unbiased draw, that no statistically-significant results from seeding will accrue (providing the evaluation is conducted by independent non-HUJ statisticians).

Fig. 15  shows some 2D images of the hydrometeors that were observed at  different altitudes over  the sea  on  12  Jan 2012,  as an example.

The images are quite nice; shows what we can do today.

However, a series of single strips of hydrometeors chosen by F2015 in Figure 15, does not suffice to prove much.  They have omitted (sound familiar?) those images and concentrations of sheaths and needles, those crystal habits that form at temperatures  > -10°C  which would document the prolific operation of a secondary or other ice particle production mechanism that operates in the clouds of Israel.    The aerosol environment of Israel, with its large CCN and often high concentrations of smaller droplets due to pollution in the Hallett-Mossop riming/splintering zone creates the “perfect storm” for ice multiplication in their clouds, as F2015  should know when they cite Mossop (1978).

 Figure 15 is not acceptable to this reviewer without extensive 2-D imagery with “best possible” concentrations available online.

The CIP particle size distributions for the same cloud/area were shown in Fig. 11. According to the combinedCDP/CIP particle size distributions, the mode of the cloud  DSD reached ~ 20  μm at  an  elevation of slightly below3000  m. This is also where the CIP water content of ~ 100  μm particles  became  more  significant. Fig. 15  showsthat these precipitation embryos are  spherical,  i.e. probably still  in  the liquid  phase, although the temperature isclose to − 8 °C.  The LWC at  that elevation reached 1.6  g m-3  and  there was  no indication for ice. Higher  up inthe  cloud  the particles became even  larger and  started having a more irregular shape.  The irregular particles seenat temperatures warmer than − 5 °C are not likely to have frozen at the  observed elevation, but rather have  fallen  from  above   while riming smaller particles. This leads  to another small  mode in  the far  right  of the  CIP size distribution (Fig. 11) rather than a more continuous decline in LWC with increasing particle diameter in the  CIP range.

            Documenting a similar process of cloud droplet growth, the formation of precipitation embryos and theirfreezing over the Golan Heights was much more challenging due to a number of reasons.  It  was  difficult to  follow   the   same cloud   element through

its lifetime due  to the more orographic nature and layered structure of the  clouds.

The authors recover from the omissions discussed previously and write an outstanding couple of qualifying sentences, the type that normally accompany a great science piece!  I’m excited again!

And even  if distinct convective cloud tops were identified, they often quickly drifted eastwards into  Syria, where itwas  not possible flying to. The impression was  that often the clouds there would consist only  of supercooled waterwith no  precipitation, unless having embedded convection or ice particles falling from above.

The above is good, and the use of the word, “impression” makes it even better.  Thank you authors!

Fig.  16  presents the   profile   of  the concentrations of hydrometeors with diameter greater than 100  μm in  clouds over the Sea and the Golan Heights — as measured by the CIP. It is based on the same 15 flights  used  forthe  analysis in Fig. 14 and  both figures have  some common features.  As with the cloud  droplet effective radius,the  concentrations of the large hydrometeors increase with decreasing temperatures  in  the developing convectiveclouds over  the  sea. This is because as the cloud grows and its top becomes colder,  the cloud droplets and  the hydrometeors that do not fall become larger,  and  also because the freezing of the largest droplets, which may  push them over the 100 μm cutoff size used  in Fig. 16, is more likely.

The  statistics of  the hydrometeor concentration over  the Golan  Heights, on the other hand, shows a maximumconcentration between − 5 and  − 10 °C, with median concentrations close  to 10 L-1. The inclusion of cloud  layers with colder bases, smaller droplets and  less  LWC might explain the decreasing hydrometeor concentrations atcolder temperatures over  the Golan Heights. This might not, however, explain the finding that the CIP concentrations inthe − 5 to − 10 °C temperature band are considerably higher over the Golan Heights than over the sea.

This doesn’t make sense.  The reality would be just the opposite if they had been more adventurous with their aircraft sampling in the Med.  The reason for the lower concentrations in the -5° to -10°C temperature range over the Mediterranean where ice multiplication is rampant, compared to those concentrations found in the Golan,  is the sampling by F2015 of young, newly risen,  “hard” turrets before the ice explosion occurs over the Med.

It is worth noting that the temperature of maximum hydrometeor concentration over  the  Golan  Heights coincideswith the activation temperature of the silver-iodide which was released upwind during most of the sampling.

WHY would one sample, at great operating cost, clouds when seeding was underway upwind???  Isn’t that a waste of precious INWA resources?  Was this unintentional?  Are there no data from your flights that were gathered in the Golan without seeding taking place?  Did you measure ice nucleus concentrations, ones that could have alerted you to this problem?  This, unfortunately, raises the issue of incompetence unless there were mitigating circumstances.

It is also close to the temperature of maximum Hallet–Mossop ice multiplication rates, which may be suppressed instrong convection (over the sea) due to relatively short residence time of the graupel in the optimal temperature band for ice splinter production (Hallett and Mossop 1974).

Suppressed H-M process?  This is non-sense.  Gimmee your plane!  These authors need to go outside and watch there incoming Cumulus congestus clouds as they convert to ice (and small Cumulonimbus clouds) within minutes after reaching very much above the -5°C level.  With proper aircraft sampling, the authors would have found a ton of H-M generated (and/or that by other processes) ice crystals (needles and sheaths) in the temperature zone (-3° to -10°C) where H-M-produced ice can reside (as at Mt. Hermon on many occasions).

One wonders why the analysis goes so badly here?  Who was directing the aircraft?  Who wrote this segment?

But this process may  be  enhanced by  the embedded convection that often occurs over  the Golan  Heights.  Thisis because the mixing of the convective cloud  element with the stratiform cloud  with a colder base,  can  result in abroad, or even  bi-modal, droplet size distribution, which is favorable for ice splintering (Mossop, 1978). It is notpossible, however, to quantify the relative contribution of each factor to the observed hydrometeor concentrationswithout a very  detailed case  by case  analysis, because the variation from  cloud  to cloud,  and even  within a cloud  is rather high.

Send me the videos from these flights, your ice and droplet concentrations and spectra, and I promise to help you understand what you did!  Better yet, give me your research aircraft for the month of January and we will find some ice, and some new findings commensurate with those of L929496.

4.4. Availability of supercooled cloud water

The supercooled liquid  water content of the clouds over the Mediterranean Sea and over the Golan Heights, atdifferent temperatures, is presented in Fig. 17. The water content was integrated over the binned CDP data and, ifnecessary, slightly adjusted to match the measured content of the Hot-wire probes. Fig. 17 is based on the samedataset as Figs. 14 and 16. Note that the abscissa in Fig. 17 is in the log-space due  to the large  possible variationin  LWC. At  the initial stages,  as  the clouds develop,  their tops   get cooler   and   the   LWC  values increase bothover  the Mediterranean Sea and  over  the Golan Heights. At − 4 °C, 25% of the  measurements that were taken over the Golan  Heights had  a LWC greater than 0.4  g m− 3, while over  the  Mediterranean Sea that number,  at  thesame temperature, was slightly smaller.

As the authors noted previously, they don’t know what happened to the LWC in the clouds at the upwind side of the Golan because it went over toward Syria.  That should be re-iterated here, lest the AR reader be misled by the F2015 claim that this is unambiguously potential water for seeding purposes.  That potential remains to be seen without more comprehensive measurements, such as ground measurements of ice crystal concentrations, habits, riming characteristics at the top of Mt. Hermon, which is often in the H-M temperature zone during precip events in Israel.

Mt. Hermon crystal data would make a stupendous paper of itself.  But could we trust the HUJ-CSG, with so much seeding baggage, to do it objectively?  I really don’t think so.  What a shame that a “senior” researcher, who is extremely familiar the output of the HUJ-CSG over several decades, feels this way.

Further up the LWC values over the Golan Heights started decreasing, while the LWC over the sea  continued toincrease as  the  cloud   tops developed further.

As mentioned in Sections 3.3 and 4.2, the sampling strategy over  the  Mediterranean Sea and  over  the Golan Heights was often  different because of the different typical cloud  structure/ dynamics as well as flight safetyrestrictions. While over the sea “hard” tops of deep convective clouds were more abundant and favored, over theGolan  Heights the maneuvering was more  restricted and  the  sampling covered the  area  more homogeneously.

Here and below, the authors divulge critical pieces of information about sampling that represents the highest form of objective, scientific writing for their readers!

Preferential sampling of developing cloud  tops  with relatively little history of mixing over  sea,  versus measuring mostly orographic layer  clouds over  the Golan  Heights,  can explain the  greater LWC values over  the sea  and the positive trend all the way up. More than half of the samples over the sea at  a  temperature around − 16  °C had  a  LWC greater than 1 g m− 3.

The quote above about 1 gm m^-3 at -16°C concerns liquid water that would be very short-lived,  indeed, as I believe the authors know, or should know.    Would it last at that level even five minutes in a Langrangian sense?  Give me your research plane and we will see!

The  likely   explanation  for  the LWC “bump” only   over the Golan  Heights is the sampling of cloud  layers  withbases at  colder lower temperatures  and   hence reduced LWC. As Fig. 3 illustrates, the  clouds over  the Golan Heights tend to  have  a structure of layers (on the days that we chose to sample on) and  a more homogeneouscoverage of the area  would include sampling of those layer  clouds.  This explanation is supported by Fig. 14, asnew  high  cloud  layers that are separated from the clouds below, are expected not only to have  low values of LWC,but also small re (Fig. 14). The LWC in those elevated cloud  layers  would increase little with height due to the coldlow temperatures, but re might increase rapidly due to the low concentration of CCN and hence also cloud  droplets, even  in the undiluted parts of the clouds.

This doesn’t make any sense to this reader; “re might increase rapidly..”   Where?

Another possible contributor to the changing trend of LWC over   the   Golan   Heights  at  temperatures  below − 8  °C is the conversion of liquid  cloud  water into  ice  (which is not measured by the CDP or by the Hot-wires). Theconcentrations of the  hydrometeors (Fig. 16) are at maximum at that temperature and  that might not be justcoincidence — as discussed in Section 4.3. In order to determine the causes and the effects  of the observedrelations, as well  as the contributions of the  seeding operations versus the  natural processes, other methods arerequired.

Again, why spend so much on research targeting natural clouds when seeding is going on?  Its crazy.

So, within the scope of this study, it is practically impossible to quantify the relative impacts of the explanationsabove  to the observed different trend of LWC versus the temperature over the sea and over the Golan Heights without analyzing a larger dataset and  a more dedicated planning and  filtering.

Yes! Thank you, authors, for this candid assessment.  I would love to take part in those additional studies!  “Call me!”   Furthermore, having a senior researcher as a kind of “Resident Skeptic” would improve the credibility of future HUJ-CSG publications.

However, despite the reduction in LWC with altitude, and  regardless of its true explanation, there are still  non-negligible amounts of  super-cooled LWC over  the Golan   Heights.

The end result of the supercooled water they observed at the upwind side of the Golan is not known.  Caution is advised in what to make of this statement.

As far as  cloud seeding  is  concerned,  the relevant question is  the precipitation efficiency of  the orographiclayer  clouds.  Obviously, this type of cloud  is inexistent over sea.

Fig. 18  provides the probability distribution and  the cumulative distribution function of the  supercooled water content over  the  (upwind side of the) Golan  Heights at temperatures  colder lower than − 8 °C.

You can’t emphasize the sampling bias contained in Figure 18 enough; otherwise it will likely mislead the INWA or other organizations into funding cloud seeding with no real return, as we saw the HUJ-CSG caused the INWA to do in the past at great expense to the people of Israel.

While some of the water might have been converted to ice due  to the  seeding operations, the  remaining liquid  water might mark the  potential of additional precipitation enhancement.  It shows that 55% of the measurements in theclouds had less than 0.2 g m-3 of super-cooled water, but it means that 45% had more. 20% of the data had  morethan 0.5 and  ~ 4% even more than 1.0 g m-3 of liquid  water with a potential to freeze with silver-iodide.

This sentence is lacking the very nice qualifiers you incorporated in this piece at other points.  Were these questionable portions written by only one of the authors?  One wonders.  But, unless the writing is partitioned in a footnote, all of the writers are responsible for a paper’s content.

Since the flight pattern over the Golan Heights was  more or less fixed, this represents the  true areal  fractional coverage of these clouds during the flights there. Therefore, this represents significant amount of super-cooled cloud water and indicates that ice  nuclei may  still  be  a  limiting factor in  the conversion of LWC into ice and further toprecipitation. So there might be room for extending the seeding efforts that target the clouds over  the  Golan  Heights torealize more of their seeding potential.

5. Summary and conclusions

A key  factor  that determines the seeding potential of the clouds is the quantity of super-cooled water in the clouds,  the time that the water droplets remain in the super-cooled region and  the temperatures at  which supercooled waterpersists.

Add this sentence after, “persists”:  “We were unable to do all the necessary maneuvers to establish these parameters.” ( It doesn’t mean you failed….)

More  super-cooled water, longer time and  colder lower temperatures,  all  lead   to  greater  seeding potential  as  there  is  a possibility to convert more of the super-cooled water into  ice and  precipitation.

After analyzing 27 research flights during four rainy seasons (2009–2013) we  see  that although the  naturalprocesses of converting  cloud   water  into   precipitation  is   often    quite efficient,

Yep.

there  are  still  occasions with good  potential for precipitation augmentation by glaciogenic cloud  seeding.

        Remains to be seen.  “Good” potential was not demonstrated in this paper.  There were too many unanswered questions.  This sentence appears to have been written for funders, not for those of us senior researchers who know the paper trail of the HUJ-CSG literature well.  One wonders how F2015, who wrote so eloquently about the drawbacks of this study, and the questions that remain, could have written this sentence?  Here the questioning and objectivity have disappeared.

One natural seeding process is  the  hygroscopic seeding by  the abundance of sea  salt aerosols in the boundarylayer  close  to the coast due  to the strong winds and  rough seas.  Another natural seeding process is called  the“seeder–feeder” mechanism,   where hydrometeors from   more mature  and   higher clouds fall through the clouds atlower levels and collect or rime their water content.

        In this scenario, there well might be an underlying, shallow orographic Stratocumulus layer that is present before the rainy/icy complex of Cbs arrive.  If precipitation is falling over mesoscale regions, downdrafts can be present that weaken of dissipate the lower Stratocu layer as the complex barges into the hill regions and the Golan.  This is a reviewer speculation; I just don’t think we have enough information to know what exactly happens, how much supercooled water there is when weakening Cb clusters/lines organized by upper troughs arrive in the northern mountains.   Observations at Mt. Hermon, to repeat, are critical in answering this question.

The results of this study have  been already applied for the design of the Israel-4 cloud seeding experiment.

        I find this statement sad, because this study/proposal was not reviewed properly, and those who assembled it have so much cloud seeding baggage, as is demonstrated in this review.  I feel sorry, too, for the Israeli people who paid millions based on prior faulty interpretations of cloud seeding results and seeding potential based on fictitious cloud descriptions.   The HUJ-CSG for too long has prospered on the backs of the Israeli people with their biased, self-serving presentations of clouds and seeding potential.

            Finally, why wasn’t I sent this manuscript in the first place?  One of the authors, DR, knows I am “above ground”, and I do not oppose seeding in proper circumstances.  I do know Israeli clouds pretty well….

This was done by  an  addition of  an  eastern seeding line  for  targeting the clouds developing over  the  Golan  Heights and  by positioning the ground generators accordingly. The main objective of the Israeli rain enhancementprogram is to increase the amount of water that reaches the Sea  of  Galilee,  therefore adding an eastern seedingline  contributes to  fulfilling  this objective in four ways: 1) The Golan Heights contribute most of the water to the Sea  of Galilee;  2)  The  clouds over  the Golan  Heights typically undergo less natural seeding by mature cloudsand sea spray;

There is quite a bit of conjecture in item 2.  This study didn’t quantify how much “less natural seeding by mature clouds” happens in the Golan.  Such a study would take several winters in this reviewer’s opinion, of combined satellite andsurface observations, and many more flights than in this study.   To repeat, such a study shouldn’t be done by the HUJ-CSG for the highest credibility.  They had their chances over so many years to get it right, and couldn’t do it.

Cloud tops are colder in extreme northern Israel and offshore region compared with those areas to the south (GN74; RH95) as noted.  Decreased cloud top temperatures lead to more ice and deeper clouds.  Presumably more Cbs, organized by troughs and by the Cypress low, would move onshore and into the Golan in most rainy day situations.  Whether the underlying orographic forcing would supply enough water for seeding purposes in the context of these mesoscale systems remains unknown.

 3) The clouds over  the Golan Heights occasionally have relatively high  concentrations of supercooled water attemperatures in which silver iodide can serve as efficient ice nuclei;

Add this qualifying sentence:  “Due to flight limitations, it is not known whether this liquid water persists long enough to be a viable seeding target.”

 4) There  is a greater seeding potential for the less convective clouds (King and  Ryan, 1997), such  as thoseclouds over  the Golan Heights.

Who knows?  Don’t count on it, Israel.  Be cautious.

A further indication of citation sloppiness, and a poor manuscript review of this article; “King and Ryan” should be Ryan and King (1997).

This whole article has degraded AR in my opinion.

            The Israel-4 randomized seeding experiment started in November 2013.  The expected duration of theexperiment is about 6 winters, during which the randomized seeding will be done in the  same consistent way, and  further accompanied by the cloud  physics measurements by aircraft and  satellites.

Acknowledgements

This study was funded by  the Israeli  water authority for supporting the rain enhancement project.

Very sad to read that this article, or some other version of it in a proposal, led to INWA to seed operationally or otherwise.     Doesn’t the INWA have any cloud expertise within it, or consultants to rely on?  It doesn’t seem like it.

Nevertheless, the INWA is to be admired and congratulated for having the wisdom to do a randomized experiment even if deemed premature, rather than just initiate another operational seeding project,  while asserting to its public, with little evidence, that rain is being increased.  It doesn’t happen in the US!

REFERENCES

Alpert, P., Halfon, N., and Levin, Z.. 2008:  Does air pollution really suppress precipitation in Israel? J. Appl. Meteor. and Climate, 47, 933-943

Altaratz, O., Levin, Z., Yair, Y., Ziv, B., 2003. Lightning Activity over Land and Sea on the  Eastern Coast of  theMediterranean.  Mon. Weather  Rev.  131, 2060–2070

American Meteorological Society, 1984:  Statement on planned and inadvertent weather modification. Bull. Amer. Meteor. Soc., 65, 1322.

Baumgardner, D., Jonssonc, H., Dawsona, W., O’Connor, D., Newtona, R., 2000.

The cloud, aerosol and precipitation spectrometer: a new instrument for cloud investigations. Atmos. Res.,  59,251–264.

Brier, G. W., and I. Enger, 1952: An analysis of the results of the 1951 cloud seeding operations in central Arizona.  Bull. Amer. Meteor. Soc., 23, 208-210.

Ben-Zvi, A., Rosenfeld, D., Givati, A., 2011 Comment on  “Reassessment of rain experiments and operations in Israel including synoptic considerations” by Zev   Levin,   Noam  Halfon and  Pinhas  Alpert  [Atmos. Res.   97   (2010)513–525]. Atmos. Res. 99, 590–592.

Blyth, A., and Latham, J., 1998:  Comments on cumulus glaciation papers of P. V. Hobbs and A. L. Rangno.  Quart. J. Roy. Meteorol. Soc., 124, 1007-1008.

Cunningham, R. M., 1957:  A discussion of generating cell observations with respect to the existence of freezing or sublimation nuclei.  In Artificial Stimulation of Rain, H. Weickmann, Ed.  Pergamon Press, NY.,  267-270.

Dennis, A. S., 1980:  Weather Modification by Cloud Seeding.  Academic Press, NY, 145.

Dennis, A. S., 1989: Editorial to the A. Gagin Memorial Issue of the J. Appl. Meteor.,  28, 1013.

Druyan, L. M., and Sant, Y., 1978: Objective 12 h forecasts using a single rawinsonde.  Bull. Amer. Meteor. Soc., 59, 1438-1441.

Field, P., Lawson, P., Brown, R. A., Lloyd, G., Westbrook, C., ,e D. Moiseev, D., Miltenberger, A., A. Nenes, A., Blyth, A.,  Choularton, T., Connolly. P., Buehl, J., Crosier, J., Cui, Z., Dearden, C., DeMott, P., Flossmann, A., Heymsfield, A., Huang, Y., Kalesse, H., KANJI, Z. A., Korolev, A., Kirschgaessner, A., Lasher-Trapp, S., T. Leisner, T., McFarquhar, G., , Phillips, V., , J. Stith, J., Sullivan, S., 2017:  Secondary Ice Production: Current State of the Science and Recommendations for the Future.  Amer. Meteorol. Soc.,  Monograph, 58, 7.01 – 7.20. DOI: 10.1175/AMSMONOGRAPHS-D-16-0014.1

Freud, E., Rosenfeld, D., 2012. Linear relation between convective cloud drop number concentration and depth forrain initiation. J. Geophys. Res.  117, D02207. http://dx.doi.org/10.1029/2011JD016457.

Freud,  E., Rosenfeld, D., Andreae,  M.O.,  Costa, A.A., Artaxo, P., 2008.  Robust relations between  CCN  and  the  vertical evolution of  cloud drop  size distribution in deep convective clouds. Atmos. Chem. Phys. 8, 1661–1675.

Freud, E., Rosenfeld, D., Kulkarni, J.R., 2011. Resolving both entrainment-mixing and number of activated CCN in deep convective clouds. Atmos. Chem. Phys. 11, 12887–12900. http://dx.doi.org/10.5194/acp-11-12887-2011.

Freud, E., Koussevitsky, H., Goren, T.,  and Rosenfeld, D., 2015: Cloud microphysical background for the Israeli-4 cloud seeding experiment, Atmos. Res., 158-159, 122-138. doi.org/10.1016/j.atmosres.2015.02.007

Gabriel, K. R., 1967: The Israeli artificial rainfall stimulation experiment: statistical evaluation for the period, 1961–1965.  Proceedings, Fifth Berkeley Symposium on Mathematical Statistics and Probability, Vol. 5, L. M. LeCam and J. Neyman, eds., University of California Press, 91–113

Gabriel, K.R., Rosenfeld, D., 1990. The second Israeli rainfall stimulation experiment:  Analysis of  precipitation onboth targets. J. Appl.  Meteorol. 29, 1055–1067.

Gagin, A., 1975: The ice phase in winter continental cumulus clouds.  J. Atmos. Sci., 32, 1604–1614.

Gagin, A, 1980:  The relationship between the depth of cumuliform clouds and their raindrop characteristics.  J. Res. Atmos., 14, 409-422.

Gagin, A, 1981:  The Israeli rainfall enhancement experiments.  A physical overview.  J. Wea. Modif., 13, 108-120.

Gagin, A, 1986: Evaluation of “static” and “dynamic” seeding concepts through analyses of Israeli II and FACE-2 experiments. In Precipitation Enhancement–A Scientific Challenge, Meteor. Monog., 21, No. 43, Amer. Meteor. Soc., 63-70.

Gagin, A., Neumann, J., 1974. Rain stimulation and cloud physics in Israel. In: Hess, W.N. (Ed.), Weather andClimate Modification. John Wiley and Sons, pp. 454–494

Gagin, A., and J. Neumann, 1976: The second Israeli cloud seeding experiment–the effect of seeding on varying cloud populations. Proc. II WMO Sci. Conf. Weather Modification, Boulder, WMO Geneva, 195-204.

Gagin, A., Neumann, J., 1981. The  second Israeli randomized  cloud seeding experiment: Evaluation of the results.J. Appl.  Meteorol. 20, 1301–1311.

Givati, A., Rosenfeld, D.,  2005. Separation between cloud-seeding and air- pollution effects.  J. Appl.  Meteorol. 44,1298–1314.

Givati, A., Rosenfeld, D., 2009. Comments on “Does air pollution really suppress

precipitation in Israel?”. J. Appl.  Meteorol. Climatol. 48, 1733–1750.

Givati, A., Steinberg, D., Binyamini, Y., Glick, N., Rosenfeld, D., Shamir, U., Ziv, B.,

2013. 2013. The Precipitation Enhancement Project: Israel-4 Experiment (טקיורפ רטמה תרבגה: לארשי יוסינ 4). The WaterAuthority, State of Israel, 55.

Goldreich,  Y.,  2003.   The  climate   of   Israel,  observations,   research  and applications.        Kluwer Academic/Plenum Publ,  New York, NY,  p. 298.

Halfon, N., Z. Levin, P. Alpert, 2009:  Temporal rainfall fluctuations in Israel and their possible link to urban and air pollution effects.  Environ, Res. Lett., 4, 12pp. doi:10.1088/1748-9326/4/2/025001

Hallett, J., and Mossop, S.C., 1974. Production of secondary ice particles during the riming process.  Nature 249, 26–28.

Hering, S.V., Stolzenburg, M.R., Quant, F.R., Oberreit, D.R., Keady, P.B., 2005. A laminar-flow, water-basedcondensation particle counter (WCPC). Aerosol Sci. Technol. 39, 659–672.

Herut, B., Starinsky, A., Katz,  A., Rosenfeld, D., 2000. Relationship between the conditions along a  transition zonebetween large deserts and Mediterranean climate, Israel. Atmos. Environ. 34, 1281–1292.

Hobbs, P. V.,  and Rangno, A. L., 1978: A reanalysis of the Skagit cloud seeding project.  J. Appl. Meteor., 17, 1661–1666.

Hobbs, P. V., and Rangno, A. L., 1985: Ice particle concentrations in clouds. J. Atmos. Sci., 36, 2523-2549.

Hobbs, P. V.,  and Rangno, A. L., 1998:  ‘Reply to “Comments by Alan M. Blyth and John Latham on ‘Cumulus glaciation papers by P. V. Hobbs and A. L. Rangno. Quart. J. Roy. Meteorol. Soc.,         124, 1009-1011.

Hobbs, P. V., Lyons, J. H., Locatelli, J. D., Biswas, K. R., Radke, L. F., Weiss, R. W., Sr., and Rangno, A. L., 1981:  Radar detection of cloud-seeding effects.  Science, 213, 1250-1252.

Kennedy, D., 2003:  Research fraud and public policy.  Science, 300, 393.

Kennedy, D., 2004:  The old file drawer problem.  Science, 305, 451.

Kerr, R. A., 1982: Cloud seeding: one success in 35 years.  Science, 217, 519–522.

Kessler, A., A. Cohen, D. Sharon, 2002:  Interim analysis of the cloud seeding in Northern Israel. A report submitted to the Israel Hydrology Institute and the Israel Water Management of the Ministry of Infrastructure, In Hebrew.

Kessler, A., Cohen, A., Sharon, D., 2006:  Analysis of the cloud seeding in Northern Israel. A report submitted to the Israel Hydrology Institute and the Israel Water Management of the Ministry of Infrastructure, In Hebrew. 117pp.

Lahav, R., Rosenfeld, D., 2000. Microphysics characterization of the Israeli clouds from aircraft and satellites.Proceedings of the 13th International Conference on Clouds and Precipitation, pp. 732–735.

Lance,  S., Nenes, A., Medina, J., Smith, J.N., 2006.  Mapping the operation of the DMT continuous flow CCN counter.  Aerosol Sci. Technol. 40,  242–254.

Lance,  S., Brock,  C.A., Rogers, D., Gordon, J.A., 2010. Water droplet calibration of the Cloud Droplet Probe (CDP) and in-flight performance in liquid, ice and mixed-phase clouds during ARCPAC.  Atmos. Meas. Tech.  3, 1683–1706.

Lensky,  I.M., Rosenfeld, D., 2006.  The  time-space exchangeability of  satellite retrieved relations between cloud toptemperature and particle effective radius. Atmos. Chem. Phys. 6, 2887–2894.

Levi, Y., Rosenfeld, D., 1996. Ice  nuclei,  raiINWAter chemical composition, and static cloud seeding effects in Israel.J. Appl.  Meteorol. 35, 1494–1501.

Levin, Z., 1992: The role of large aerosols in the precipitation of the eastern Mediterranean. Paper presented at the Workshop on Cloud Microphysics and Applications to Global Change, Toronto. (Available from Dept. Atmos. Sci., University of Tel Aviv).

Levin, Z.,  1994:  The effects of aerosol composition on the development of rain in the eastern Mediterranean.  WMO Workshop on Cloud Microstructure and Applications to Global Change, Toronto, Ontario, Canada. World Meteor. Org., 115-120.

Levin, Z., Ganor, E., Gladstein, V., 1996. The effects of desert particles coated with sulfate on rain formation in theeastern Mediterranean. J. Appl.  Meteorol. 35, 1511–1523.

Levin, Z., Kirchak, S. O., and Reisen, T., 1997: Numerical simulation of dispersal of inert seeding material in Israel using a three-dimensional mesoscale model.  J. Appl. Meteor., 36, 474-484.

Levin,  Z., Halfon, N., Alpert, P., 2010. Reassessment of rain enhancement experiments and operations in  Israelincluding synoptic considerations. Atmos. Res. 97, 513–525.

Levin, Z., N. Halfon, and P. Alpert, 2011:  Reply to comments by Ben-Zvi, A., D. Rosenfeld and A. Givati on the paper: Levin, Z., N. Halfon and P. Alpert, “Reassessment of rain experiments and operations in Israel including synoptic considerations.” .  Atmos. Res., 98, 593-596.

List, R., Gabriel, K. R., Silverman, B. A., Levin, Z., Karacostas, T., 1999: The Rain Enhancement Experiment in Puglia, Italy: Statistical Evaluation.  J. Appl. Meteorol., 38, 281-289.

Marshak, A., Platnick, S., Várnai, T., Wen, G., Cahalan, R.F., 2006. Impact of three‐

dimensional radiative effects on satellite retrievals of cloud droplet sizes. J. Geophys. Res. 111, 1984–2012.

Mason, B. J., 1980:  A review of three long-term cloud-seeding experiments.  Meteor. Mag., 109, 335-344.

Mason, B. J.,, 1982:  Personal Reflections on 35 Years of Cloud Seeding.  Contemp. Phys., 23, 311-327.

Mielke, P. W., Jr., 1979: Comment on field experimentation in weather modification. J. Amer. Statist. Assoc., 74, 87-88.

Mossop, S. C., 1970: Concentrations of ice crystals in clouds. Bull. Amer. Meteor. Soc., 51, 474-479.

Mossop, S.C., 1978. The influence of drop size  distribution on the production of secondary ice  particles during graupelgrowth. Q. J. R. Meteorol. Soc. 104, 323–330.

Mossop, S. C., 1985: The origin and concentration of ice crystals in clouds. Bull. Amer. Meteor. Soc., 66, 264-273.

Mossop, S.C., Hallett, J., Hallet, J., and Mossop, S. C., 1974. Ice crystal concentration in cumulus clouds: Influence ofthe drop spectrum. Science 186, 632–634.

National Research Council-National Academy of Sciences, 1973: Weather Modification: Progress and Problems. T. F. Malone, ed., 258 pp. (Available from the National Research Council, Washington, D. C.)

Nirel, R.,, and D. Rosenfeld, 1994: The third Israeli rain enhancement experiment-An intermediate analysis. Proc. Sixth WMO Scientific Conf. on Weather Modification, Paestum, Italy, World Meteor. Org., 569-572.

Neumann, J., 1951:  Land breezes and nocturnal thunderstorms.  J. Meteor., 8, 60-67.

Neumann, J., Gabriel, K. R., and Gagin, A., 1967: Cloud seeding and cloud physics in Israel:  results and problems.  Proc. Intern. Conf. on Water for Peace.  Water for Peace, Vol. 2, 375-388.

Pitter, R. L., and H. R. Pruppacher,1973: A wind tunnel investigation of freezing of small water drops falling at terminal velocity in air. Quart.  J. Roy. Met. Soc., 99, 540-550.

Rangno, A. L., 1979.  A reanalysis of the Wolf Creek Pass cloud seeding experiment.   J. Appl. Meteor., 18, 579–605.

Rangno, A. L., 1988: Rain from clouds with tops warmer than -10 C in Israel. Quart J. Roy. Meteorol. Soc., 114, 495-513.

Rangno, A. L., and Hobbs, P.V.,. 1991: Ice particle concentrations and precipitation development in small polar maritime cumuliform clouds. Quart. J. Roy. Met. Soc., 117, 207-241.

Rangno, A. L., and Hobbs, P.V., 1993: Further analyses of the Climax cloud-seeding experiments.  J. Appl. Meteor., 32, 1837-1847.

Rangno, A.L., Robbs Hobbs, P.V., 1995. A new look at the Israeli cloud seeding experiments. J. Appl. Meteorol. 34,1169–1193.

Rangno, A. L., and Hobbs, P. V., 1997: Comprehensive Reply to Rosenfeld, Cloud and Aerosol Research Group, Department of Atmospheric Sciences, University of Washington, 25pp.  (Available at http://carg.atmos.washington.edu/)

Rauber, R. M., and Tokay, A.,  1991:  An explanation of supercooled water at the top of cold clouds.  J. Atmos. Sci., 48, 1005-1023.

Rosenfeld, D., 1980:  Characteristics of Rain Cloud Systems in Israel Derived from Radar and Satellite Images.  M. S. Thesis, The Hebrew University of Jerusalem, 129pp. (Available from the Department of Meteorology, Hebrew University of Jerusalem, Jerusalem, Israel) in Hebrew, no translation yet.

Rosenfeld, D., 1997: Comment on “A new look at the Israeli Cloud Seeding Experiments”, J. Appl. Meteor., 36, 260-271.

Rosenfeld, D., 1998.  The third Israeli randomized cloud seeding experiment in the south: evaluation of the results and review of all three experiments. Preprints, 14th Conf. on Planned and Inadvertent Wea. Modif., Everett, Amer. Meteor. Soc. 565-568.

Rosenfeld,  D.,  Farbstein,  H.,  1992.  Possible  influence  of   desert  dust  on seedability of clouds in Israel. J. Appl. Meteorol. 31, 722–731.

Rosenfeld, D., Givati, A., 2005. Evidence of orographic precipitation suppression by  air  pollution induced aerosols in the western U.S.  J. Appl.  Meteorol. Climatol. 45, 893–911.

Rosenfeld, D., Gutman, G., 1994. Retrieving microphysical properties near the tops of  potential rain clouds by multispectral  analysis of  AVHRR data. Atmos. Res. 34, 259–283.

Rosenfeld, D., Lensky, I.M., 1998. Satellite-based insights into precipitation formation processes in  continental andmaritime convective clouds. Bull. Am. Meteorol. Soc. 79, 2457–2476.

Rosenfeld, D., Nirel, R., 1996. Seeding effectiveness — the interaction of desert

dust and the southern margins of  rain cloud systems in  Israel. J. Appl. Meteorol. 35, 1502–1510.

Rosenfeld, D., Rudich, Y. and Lahav, R., 2001: Desert dust suppressing precipitation:  a possible desertification feedback loop. PNAS, 98, 5975-5980.   doi/ 10.1073/ pnas.101122798

Rosenfeld, D., Liu, G., Yu, X., Zhu, Y., Dai, J., Xu, X., Yue, Z., 2014. High resolution (375 m) cloud microstructure asseen from the NPP/VIIRS Satellite imager. Atmos.  Chem.  Phys.  14,   2479–2496. http://dx.doi.org/10.5194/acp-14-2479-2014.

Ryan, B.F, King,  W.D., 1997. A Critical Review of the Australian Experience in

Cloud Seeding. Bull. Am. Meteorol. Soc. 78, 239–254.

Schaefer, V. A., 1946: The production of ice crystals in a cloud of supercooled water droplets.  Science, 104, 457-459.

Schemenauer, R. S., and Isaac, G. A., 1984:  The importance of cloud top lifetime in the description of natural cloud characteristics.  J. Climate Appl. Meteor., 23, 267-279.

Schultz, D., 2015.  Eloquent Science.  American Meteorol. Soc., 412pp.

Sharon, D., A. Kessler, A. Cohen, and E. Doveh, 2008:  The history and recent revision of Israel’s cloud seeding program.  Isr. J. Earth Sci., 57, 65-69.

Silverman, B. A., 2001. A critical assessment of glaciogenic seeding of convective clouds for rainfall enhancement. Bull. Am. Meteor. Soc., 82, 903-924.

Tan,  W.,  Yin,  Y., Chen, K., Chen, J., 2010.  An  experimental study of  aerosol particles using PCASP-X2  andWPS™.  Geoscience and Remote  Sensing (IITA-GRS),   2010 Second IITA International  Conference on  vol.  1.  IEEE, pp. 586–589.

Thom, H. C. S., 1957: An evaluation of a series of orographic cloud seeding operations.  Final Report of the Advisory Committee on Weather Control, Vol. II, Government Printing Office, 25-49.

Tukey, . Brillinger, D. R.,  and  Jones, L. V.,  1978b: Report of the Statistical Task Force to the Weather Modification Advisory Board, Vol. II.  U. S. Government Printing Office,  pE-3.

Vali, G., 1971: Quantitative evaluation of experimental results on the heterogeneous freezing nucleation of supercooled liquids. J. Atmos. Sci., 28, 402-409.

Woodcock, A. H., 1953:  Salt nuclei in marine air as a function of altitude and wind force.  J. Meteor., 10, 362-371.

Woodcock, A. H., Duce, R. A., and Moyers, J. L., 1971. Salt and raindrops in Hawaii.  J. Atmos. Sci., 28. 1252-1257.

Woodley, W. L., and D. Rosenfeld, 2002: Comments on “A critical assessment of glaciogenic seeding of convective clouds for rainfall enhancement.” Bull. Amer.Meteor. Soc., 83, 739–740.

World Meteorological Society, 1992: Statement on the Status of Weather Modification.

Wurtele, Z. S., 1971: Analysis of the Israeli cloud seeding experiment by means of concomitant meteorological variables. J. Appl. Meteor., 10, 1185-1192.

Young, K., 1993: Microphysical Processes in Clouds.  Oxford University Press, pp427.

Reviewer “baggage” module: bio, a priori convictions, and a little about the reviewer’s 1986 11-week cloud investigation in Israel

Background:  Chased storms, cloud photographer.

I have worked on both sides of the seeding fence, having participated in operational seeding projects in South Dakota twice, India, Washington State, and in the Sierras of California.  I was a forecaster for the Colorado River Basin Pilot Project (CRBPP), a large, multi-million dollar advanced randomized cloud seeding experiment in Colorado from 1970-1975.  Beginning in 1976,   I worked in airborne studies of clouds and the origin of ice in clouds in the University of Washington’s Cloud and Aerosol Research Group for the 28 years.  I was a “flight scientist” for NCAR during the 2006-07 winter during their Saudi Arabia cloud seeding potential study.

A priori convictions concerning cloud seeding:

Cloud seeding works in limited situations, principally via the seeding of non-precipitating, supercooled clouds (e.g. Hobbs et al 1981, an experiment suggested by the reviewer; “wonderful experimentation”).  Whether there is economically viable return from such seeding, I don’t know.  I begin every talk I give with, “Cloud seeding works!”, and show a couple of examples of effects of seeding (those resembling Schaefer’s dry ice seeding experiments in 1947)

Motivation for this review:

The review of F2015  was motivated by the absence of post-publication comments on F2015 over the past three years.   And it comes from the emotional reaction after I read it recently for the first time: “Someone has to do something about this!”, the very same feeling I had before I went to Israel in 1986 as a skeptic of ripe-for-seeding cloud claims emanating from the HUJ-CSG.  (That same emotional reaction first happened during the CRBPP which led to a “career” of own-time scrutiny of suspect cloud seeding publications (e.g., Rangno 1979).

The 1986 Israel cloud investigation

 At the end of 1985 I resigned my job with the Cloud and Aerosol Research Group over issues of credit, and went to Israel in early January 1986 to investigate their clouds[1].  I was skeptical that those many descriptions of them in the peer-reviewed literature and in conference presentations were correct.   I did not go to Israel without “baggage,” but had done a considerable amount of “homework.”

A note-sized paper asserting that the Israeli clouds were not as they were being described in journal articles and elsewhere, had been rejected in 1983 (B. Silverman, Ed., J. Appl. Meteor.,  personal communication).

That rejection was instrumental in my 1986 trip.

The several reviewers’ negative takes on my 1983 “Comment”, including that of the leader of the HUJ-CSG at that time who lectured me in person at the 1984 Park City weather mod conference about how wrong I was about his clouds,  had no effect whatsoever about what I thought about them.  He had also been a reviewer of that “Comment.”

To repeat, I had done my homework, plotted numerous soundings when rain was falling at launch time or within the hour at Beirut and Bet Dagan, and had scrutinized all of the HUJ-CSG cloud reports in great detail.  Even the daily rainfall in the Israel Meteorological Service (IMS) monthly weather reports indicated something was terribly wrong with clouds that were being described as highly inefficient rain producers.

While in Israel in 1986, that same leader of the HUJ-CSG prevented this writer from visiting the HUJ-CSG-controlled radars to evaluate cloud top heights during rain spells.  I wanted to see what top heights were, and indirectly via IMS rawinsonde soundings, obtain top temperatures.  (The HUJ-CSG leader, “AG,” did allow a brief “show and tell” visit to his Ben Gurion AP radar in February as a storm approached[2] as part of our third and last meeting in February 1986.)

To disallow a bona fide worker in the field of cloud microstructure and weather modification access to data/measurements to test claims about the Israeli clouds was demonstrable scientific misconduct IMO[3].   The clouds of Israel can only be studied in Israel, and as scientists we must be open to cross-checking of our results; having our findings tested.  It’s what we do in science (e.g., see Blyth and Latham’s 1998 criticisms of the glaciation papers of Hobbs and Rangno and our Reply (Hobbs and Rangno 1998) as a great example of open criticism).

But this was not what the leader of the Israeli experiments understood about science.  If he was deluded about his clouds, or was correct about them, I would have been welcomed, I thought.  If he had “contrary knowledge”, I would be blocked.  Q. E. D.

My findings (R88), in lieu of radar data, were based on Israel Meteorological Service’s (IMS) four-times-a-day rawinsondes, and were published in 1988.  They strongly suggested that there was a serious problem with the HUJ-CSG’s ultra “ripe-for-seeding” cloud descriptions.  Rosenfeld and Farbstein (1992) within the HUJ-CSG, belatedly discovered that “dust-haze” when present, apparently produced efficiently raining clouds in Israel, findings that supported R88.   Dust-haze has been around for quite awhile in Israel.

Later independent aircraft reports by Tel Aviv University (Levin 1992, 1994, Levin et al. 1996, hereafter L929496) also supported R88 and further exposed the faults in the HUJ-CSG cloud descriptions.

Today, and at last in F2015, the HUJ-CSG finally acknowledges, implicitly, why the leader of the Israeli experiments refused this visitor access to radars in 1986:   the high precipitating efficiency of Israeli clouds was going to be obvious in radar cloud top imagery.[4]  F2015 offers additional confirmation of those long ago findings in R88.  Lahav and Rosenfeld (2000), and Rosenfeld et al. 2001 had also found unsuitable clouds for seeding, but did not state that as explicitly as did F2015, or only alluded to “some clouds.”

Too, the high precipitation efficiency of Israeli clouds, too, is not limited to “dust-haze”; it never was.  The belief that the Rosenfeld and Farbstein (1992) attribution of “divergent seeding effects” to dust-haze was unreliable was the “acorn” of conviction that led to the “oak” of RH95.[5]

Rather than “dust-haze, that high efficiency is due relatively low droplet concentrations in Mediterranean clouds moving into Israel.    Larger droplets in clouds accelerate the formation of precipitation via several processes.

End of “baggage” module.

[1] Done on “own time, own dime”; not on grant monies.  I went to Israel in 1986 after I became convinced that the cloud reports by the HUJ-CSG were in substantial error and that the people of Israel were likely paying for the seeding of unsuitable clouds: “Someone has to do something about this!”  It was a “do-gooder” thought, and felt I COULD do something about it due to my background in airborne studies in the Cloud and Aerosol Research Group at the U of Washington, as a weather forecaster, as a storm chaser, and as a cloud photographer.  I spent 11 weeks in Israel watching the clouds and storms and working within the IMS.  I lived off my savings ($$$$) the rest of 1986 after I left Israel while preparing the manuscript that became R88.  It was submitted to the QJ in January 1987.  “Quasi-altruistic”?  Yep. I wanted to also demonstrate that I was the BEST at “outing” bogus cloud seeding work that none of my gullible peers suspected, to be completely candid.  Yes, it was a little megalomaniacal…

[2]Marked by a mid-level overcast of Altostratus and some lower Altocumulus with dust.

[3] After being asked to leave the offices of the HUJ-CSJ and not come back at the end of my 2^nd meeting with AG in Janaury, I wrote to several scientists around the world about that event. Those I wrote to were: P. V. Hobbs and Lawrence F.  Radke, University of Washington; Gabor Vali, University of Wyoming; Roscoe Braham, Jr., North Carolina State University; and S. C. Mossop, CSIRO, Australia.

[4] Those within the HUJ-SU had about ten years of viewing and recording storms on their Enterprise 5-cm radar by the time of my  1986 visit.  Was no one “minding the store”?

[5] Also done on “own time, own dime”; not on grant monies.  Crackpot alert?

[1] As experienced in this writer’s 11-week stay in 1986, events in which embedded, much taller clouds sprang forth as the upper trough approached.

[1]Not cited by the HUJ-CSU in this paper. We expect studies that offer counter views to F2015 findings not to be cited by them at this point.  Perhaps they are taught that at the HUJ?  HUJ, are you listening?

[2] The reviewer believes that it is critical that the INWA or other organization certify that the list of random decisions provided prior to each season for Israeli 2 by statistician Ehud Shimbursky are, in fact, the ones that were used in Israeli 2.  This reviewer is doubtful, to add yet another layer of darkness to this analysis.  It’s just too tempting for those in the cloud seeding realm to say that, “Yes, we seeded that heavy rainstorm.”

[3]The authors omit the indication of the remarkable biased random draw of Israeli 2 experiment in which the rain in the south target was “statistically significant” in terms of standard error from average amounts on rain days (Gabriel and Rosenfeld 1990).

[4] Also from Gabriel and Rosenfeld (1990):  “The easiest explanation is to ascribe everything to chance, and accept H00 (the null hypothesis) . But the majority of the analyses run counter to this simplistic summary. Most of the evidence favors a positive effect of seeding in the north, and there is more evidence for a negative effect in the south than for a zero effect.”  This interpretation no longer holds water.

[5] The random draw for Israeli 2 was much different than for Israeli 1.  In Israeli 1 (1961-1965 daily randomization period), there were few consecutive draws of the same sign (13%) of the total of draws (Gabriel 1967, Table II).  In Israeli 2, consecutive draws of the SAME sign predominated (58%).  The list of random days for Israeli 2 was provided to Peter Hobbs in 1983 by A. Gagin.  Further investigation will likely find that this is the root cause of the lopsided draw in Israeli 2.

[6] From the Weather Modification Association’s Code of Ethics:  “Falsification: changing or not reporting appropriate data or results (i.e. the purposeful omission of conflicting data or information with the intent to falsify results; deceptive selective reporting). Isn’t that the purpose of “one-sided” citing as well?   Of note:  The American Meteorological Society does not have a Code of Ethics, but rather “suggestions” or “guidelines” for professional conduct.

[7] See the discussion of operational seeding below for details.

[8] Sound familiar? We repeat the WMA Code point for your consideration:  falsification: changing or not reporting appropriate data or results(i.e. the purposeful omission of conflicting data or information with the intent to falsify results; deceptive selective reporting).

[9] We can only imagine how these results would have been spun around if this had been a suggestion of an 8% increase in rain.  How many statistical tests would be tried?  Would the authors really just say, “inconclusive”?

[10] Sharon et al (2008), and the reports that it was distilled from (Kessler et al. 2002, 2006), are nowhere to be seen in the HUJ-SU authors’ paper.  Should we be surprised at this point?  I don’t think so.  This is not science as we know it.  HUJ, are you listening?

[11] While the original INWA program was terminated, some additional experimental-operational seeding did carry on, beginning with the 2007-08 rain season; it was terminated again in 2013 (according to F2015).  The reviewer does not yet know the details to the “supplemental” seeding following the termination in 2007.

[12] Alpert et al 2008 and Halfon et al 2009, and the exchange between Givati and Rosenfeld, also go uncited, strengthening a pattern of deception by the HUJ-SU; such actions by the authors that can no longer be attributed to oversights, but are meant to keep the ARreaders in the dark about the quality of the HUJ-CSG research.

[13] “Cast of thousands!” (except me..)

[1] Retiree, Cloud and Aerosol Research Group, Atmos. Sci. Dept., University of Washington, Seattle, 1976-2006.

[2] I dedicate this review to Jerzy Neyman, whom I greatly admired for his careful and voluminous criticisms of papers in the cloud seeding arena; to K. Ruben Gabriel, who was able to “stay above the fray” as a careful and objective statistician for the Israeli experiments, and lastly to Karl Rosner, Mekoroth’s Chief Meteorologist for Israeli 1 and 2, for his ability to remain objective within the “halls of seeding.”

[3] The great Nobel Laureate, Irving Langmuir, comes to mind, who once bitten by AgI, believed that any rainfall event could be explained by cloud seeding had there been any, no matter how large or far away from the release the event was.

[4] The above provocative conclusion comes from someone who has followed the HUJ-CSG reports over the past 35 years or so; is it me or them?

 

[5] The reviewer is not and has never been a faculty member at a university and therefore can raise such an issue with impunity unlike those with faculty status due to the “live and let live” de facto rules of the “faculty club.”

[6] This was the last mention of a numerical result of seeding in the south target of Israeli 2.

[7] The several IMS forecasters I spoke with when I worked within the Israel Meteorological Service (IMS) in 1986 were well aware of the “efficiency” that Israeli clouds exhibited.  One stated, “We get good rains out of clouds with tops at -10°C.”

 

[8] It had to be done by an outsider.

[9] Width (and implicitly, cloud top lifetime) has a direct bearing on the production of precipitation (e.g., Schemenaur and Isaac 1984; Rangno and Hobbs 1991.

[10]Whom also goes uncited (strangely) since one of the authors of F2015 (DR) commented on his BAMS cloud seeding review.  Small-mindedness here as well?  Or is it the desire by the authors to hide alternative views from their funders and journal readers?

2.  Durango, Colorado: Where it all began so long ago

How an idealistic view of science decayed into cynicism-followed-by- activism during a large randomized cloud seeding experiment in Colorado, eventually leading to a trip to Israel to evaluate their clouds

This section is kind of a slog about my Colorado experiences….but, I wanted to hit a FEW highlights of what was an epiphany about science for a rather naive person just out of college, me, that occurred in Durango, Colorado. This was my very first job as a weather forecasting meteorologist after graduating from San Jose State College.

(Skip if busy….though if you do, you will miss some personal ridicule, a movie, accolades, a possibly libelous newspaper headline caused by me, and details of a monetary science prize from the United Nation’s World Meteorological Organization that me and Peter Hobbs received for our work in weather modification. Yes, in 2005 I became, “Prize-winning meteorologist”, Art Rangno…

It is sad for me to have to point out something about the above “prize”, however. Like my HS and college baseball career, (all 2nd team this; all 2nd team that), the prize described above was really a consolation one, to insert a truth-in-packaging note. Other workers got lots more than we did that year. On the other hand, 32,000 Chinese weather modification workers got the SAME amount as Peter and I got that year; hah, less than a US dollar each!

OK, back to serious text…

…that Durango job was a dream come true for me, since I only wanted to be a weather forecaster since I was a little kid (even, somehow, forecasted weather for my 5th grade class–had an aneroid barometer in the “cloak room”). And there I was in the beautiful little town of Durango, Colorado, right out of college, forecasting weather for an important scientific experiment! My life could not have been better!

How I got to the point where I would be so skeptical of peer-reviewed cloud seeding literature that I would travel thousands of miles in question of cloud reports from the world’s leading cloud seeding scientist, however, began here during the huge Bureau of Reclamation randomized cloud seeding experiment called the Colorado River Basin Pilot Project (CRBPP). Read on.

3.  The movie explaining the Colorado experiment; a tribute to its size and importance

To depart for a second, it was a project so huge that it had its own movie, the cloud seeding “documercial,” Mountain Skywater, with a soundtrack by a local Durango artist, Clarence “Gatemouth” Brown!

Departing even further from serious text, it is with extreme modesty that I point out that I was the STAR of this 28 minute movie; I never dreamed that I would be a STAR in a movie (!), but there I am, as was declared by the Commissioner of Reclamation in those days, Ellis Armstrong. He attended the 1972 release of the film in Durango and gave me an autographed photo of several of us with him in which he proclaimed on it that I was the STAR. I only speak maybe two sentences in the whole thing! It was a pretty humorous take. I do cite it in my filmography, however.

Watching this movie you will get a sense of that cloud seeding era and how it was thought that a cloud seeding success in this randomized experiment was going to be a slam dunk in the San Juan mountains around Durango. There wasn’t a lot of questioning in those days about the work that this massive project was based on; namely, several stunning randomized experiments conducted and reported by Colorado State University (CSU) scientists in the late 1960s–contracts were being signed in 1968 for the CRBPP work about when the Climax II experiment was only about half completed! (And that, my friends, was a gigantic goof, as you will read.)

Also from the movie you will get a sense of the CRBPP’s scope and how well-planned it was overall. The precip measurements were made by those who didn’t know what the experiment day call was, seeded or not seeded. It doesn’t get better than that, and the BuRec deserves some mighty big accolades for that; trying to do it right. They were so confident, too, that they said that in spite of randomization (in which only half the days are seeded), that the CRBPP would produce an extra 250,000 acre-feet of water from the target watersheds.

Also in “doing it right”, and before the CRBPP began, the BuRec proclaimed in its PR literature that they would hire an independent statistical group to evaluate the results of this mega-experiment. It doesn’t get better than that, either. It was a display of confidence about the outcome of the experiment.

Aside: For the other seeding operators out there whose films you might see, this admonishment: “Randomize, baby, randomize”. Prove your claims the right way. Also, to seeding funders: employ independent panels to evaluate what you’ve been getting from commercial seeding as the Israeli’s bravely did.

————————-

4.  Scientific idealism begins to slip away in Durango

However, during the CRBPP I lived through journal peer-reviewed literature (J. Appl. Meteor.) that many of us knew was bogus but no one challenged. I, too, participated in a “Code of Silence” that kept our outside peers in the dark about important discrepancies that were being discovered in the CSU cloud and cloud seeding hypotheses during the CRBPP. These discrepancies turned out to cause the undoing of an otherwise well-planned experiment by the Bureau of Reclamation’s Atmospheric Water Resources Management Division, as it was called then (just “BuRec” in this piece). “Management” of atmospheric water was a word that also spoke to overconfidence.

At the same time, while in awe of the BuRec’s planning, it was strange to me that the personnel with them were immune from learning from those of us in the field about problems in their interpretations of the CRBPP’s results.

An example: BuRec personnel submitted a paper to a Florida conference in 1974, several years after the CRBPP had started, purporting that “carryover seeding” effects (those days when a control day followed a seeded day) had compromised the CRBPP because heavy snow often fell on that second “control” day. They then assumed that any heavier snow on the 2nd day MUST be due to seeding effects from leftover seeds that didn’t get blown away. They then grouped such carry over days, or portions of such days, into the actual days chosen for seeding and got better suggestions of increased snow due to seeding for the CRBPP overall.

However, no seeding effects were being detected in the first few years on single days that were seeded. Therefore, it was a crazy idea that somehow the seeding agent, silver iodide, turned into super-seeds after we turned off the seeding generators.

Of course, there was a natural explanation for this phenomenon when two days in a row were selected for experimentation.

I wrote a long letter in 1974 explaining why the findings in that BuRec preprint were bogus. When we randomly selected a second day in a row for experimentation, it was because an incoming storm was so large and heavy that it took two days for it to go by, or it was just beginning on the last hours of the first day. Not surprisingly, the heaviest part of the storm was on the second day, and usually early on.

I showed the BuRec data that control days that followed a control day, the second control day also had heavy snow, especially in the early going just like they were inferring was due to inadvertent “carryover” seeding of a control day after a seeded day. You could claim in a similar way from my examples that not seeding on a control day caused heavy snow on a following control day; silly. I had much more argumentation as well.

My explanation fell on deaf ears.

I concluded my commentary to them in 1974 about their ersatz findings with a line they couldn’t refuse to act on: I said they needed a “Resident Skeptic” at their headquarters in Denver.

A couple of weeks later, the CRBPP Project Monitor from the BuRec, Mr. Bill Douglas, presented me in person with a framed, Dr. Archie M. Kahan “Certificate of Honorary Resident Skeptic Award.” The presentation, in which he read the words on the Certificate, got a lot of chuckles from our staff who gathered around to see it. Archie Kahan, whose signature appears in the lower right, was the head of that BuRec cloud seeding division.

5.  My Well-earned Resident Skeptic Award from Dr. Kahan and the BuRec

Here is that “Certificate”, one really meant, I thought anyway, to ridicule someone they didn’t take seriously. Well, there were some at the BuRec, like the late Olin Foehner, who did take me seriously. I was only trying to help, guys…. You’ll have to zoom in to read the text.

Note the upside down Bureau of Reclamation logo in the lower left hand corner. It was to be prophesy for the division that sent me this “award.” Due to various missteps, of which the CRBPP was one, and a wetter period of years in the later 1970s into the 1980s, interest in cloud seeding virtually disappeared and their office was shutdown.

6.  Decay of idealism accelerates in Durango

More disillusionment with the BuRec and journal literature came when their preprint about carry over effects in the CRBPP was published in 1975 in the peer-reviewed, J. Appl. Meteor. There was no mention of the synoptic situation that I had described that compromised their findings. To them, inadvertent contamination of CRBPP days was too good an argument to let go of to help boost the results for a failing 10 million dollar experiment. Nor did I comment on it; I had no experience in journal matters and it never occurred to me to do so.

7.  The choice of the evaluators of the CRBPP

Another decline in confidence about the science of the CRBPP occurred when the BuRec, instead of choosing an independent group to evaluate the CRBPP as they said they would do before the project started, hired a cloud seeding group to evaluate it! While the group they hired went under the company name of Aerometric, Inc., most of the team of evaluators were really from North American Weather Consultants, led by Robert D. Elliott, President of NAWC. NAWC was largely a commercial cloud seeding company with many seeding projects and at one point was seeding commercially so enthusiastically in Utah that it contaminated some control days of the CRBPP! “Aerometric-NAWC” was chosen as the evaluator when it was clear, after just two years of random decisions, that the CRBPP was NOT going to replicate the CSU seeding results.

Perhaps the BuRec needed a friendly bailout, someone to put a happy face on a science disaster. (Footnote: I had worked for NAWC as a summer hire in 1968 and loved it and the great people there. Tor Bergeron stopped by! Still, it wasn’t a good choice by the BuRec to have them evaluate whether cloud seeding worked.)

8.  The informational “black hole” during the CRBPP: important findings came in from the field but never went out to peers

In mid-stream of the CRBPP, the BuRec called a meeting in July 1973 to try to understand what was going wrong with it. Why wasn’t it going to replicate the CSU work? Mainly, it was due to a few critical CSU assumptions that were not supported by data, such as the 500 mb temperature being an index of cloud top temperatures, and therefore, as it had been assumed, a reliable index of seeding potential. After all, the CSU experiment seeding effects were stratified by 500 mb temperatures repeatedly in the published literature; they had no data on actual cloud tops. Neither of those parameters, 500 mb temperatures or cloud top temperatures, are reliable indicators of seeding potential.

Nor were there widespread non-precipitating, reasonably deep clouds ripe for seeding ahead of and behind periods of natural precipitation, clouds that CSU scientists had inferred existed because the claimed increases in snow they reported, were solely due to the greater duration of snowfall on seeded days. Seeding had no effect on natural precipitation they concluded.

No such thick, non-precipitating cloud was found to exist in the CRBPP. This was largely due to the fact that cloud tops during storms were almost always colder than -15°C in storm situations, and usually considerably colder. Those cold tops naturally produced substantial ice concentrations without being seeded. High natural ice concentrations in clouds pretty much decimates seeding potential.

In closing that 1973 meeting, consisting of a who’s who in weather modification from universities and companies around the country, the Chief of the BuRec’s cloud seeding division, Dr. Archie M. Kahan closed it by observing that the CSU physical hypotheses, “were not as strong as we had been led to believe.”

It was an understatement.

But these important findings presented at that BuRec conference remained husbanded with those at that meeting. The “Code of Silence” was in full operation. The discrepancies were not to be “outed” until 1979 in Hobbs and Rangno (J. Appl. Meteor.) and in my reanalysis of the CSU Wolf Creek Pass experiment that same year in that journal. (The former article was originally part of the draft manuscript I brought in to Prof. Hobbs, but he deemed it something that should be reported separately.)

9.  Another pivotal event in 1974

I remember how excited I was, too, when a National Academy of Sciences 1973 report, Climate and Weather Modification; Problems and Progress, came through the Durango office in 1974. The NAS Panel on Weather Modification (Malone et al.) stated that the CSU cloud seeding work had “demonstrated” cloud seeding efficacy on a “deterministic basis”.

What was exciting when I read that NAS report in 1974?

I knew by then that an assessment by our best scientists with the NAS, and that a scientific consensus on the CSU experiments, as we would say today, was wrong! It was interesting to me later that Peter V. Hobbs, for whom I was to work, was a co-author of that optimistic report concerning the CSU experiments.

10.  A final blow to idealism and of the credibility of the cloud seeding literature

The final straw, however, was a much-cited article in 1974 in the J. Appl. Meteor. titled, “The Cloud Seeding Temperature Window.” The two authors had used constant level pressure surfaces to index cloud top temperatures in several seeding projects to come up with a cloud top temperature window of -10° to -25°C for successful cloud seeding. This temperature range was thought to characterize clouds with tops this cold that were deficient in ice particles, but would have supercooled liquid water in them that could be tapped by cloud seeding. It turned out to be a too optimistic a temperature range as later research showed.

Moreover, the lead author of this article had been told by three different people on separate occasions in my presence not to use a constant pressure level as an index of cloud tops in the Rockies. Nature does not constrain cloud tops so that they can be indexed by a constant pressure level temperature in the atmosphere.

The other author of “The Cloud Seeding Temperature Window” was in the midst of evaluating the storm day rawinsondes of the CRBPP; he was the leader of the Aerometric-NAWC evaluations team chosen by the BuRec. He absolutely knew that stratifications by a constant pressure level was not a viable way to index cloud tops from our data. When I asked that 2nd author the next time he came through the Durango office about that article, “How could you write that?” He simply replied, sheepishly it seemed to me, that he had just, “gone along with” the lead author.

So that was it.

I never again trusted the cloud seeding published literature. Cynicism 1, Idealism, nil. It didn’t matter, either, how highly regarded the literature was. It still might be inaccurate, corrupt, I thought. I often wondered, too, why that “Window” article was cited so much. I presumed it must be by readers that did not know much about synoptic weather and cloud top fluctuations.

11.  A regrettable personal media eruption in late 1975 that required an apology in person at CSU

I remained quiet until the CRBPP experiment ended in 1975, which also allowed me to retain my great job in the nice little town of Durango, Colorado–ah, the plight of whistleblowers……

But then I erupted in November 1975 after the CRBPP ended when it was safe and I had no job. Here’s that whistleblowing eruption as seen in the Durango Herald, one that I feel I have to disclose in this “blook” to give an idea of my potential biases:

You will notice that I referred to “Watergate” in the Herald headline. As I left the Durango Herald office with the reporter, Mike McRae, I muttered a mistake. I said, “if what I have begun to work on turns out, it could be the Watergate of meteorology”, meaning it would make a big splash. It was a poor, if current and accessible metaphor, but it implied wrongdoing on the part of CSU scientists. I was away when the article came out and was devastated to see what Mike had written after a careful 1-2 h recorded interview in his office. He had promised to let me examine the article before it came out, but called the evening before I left and said he wasn’t able to do that, adding, “trust me.”

I left the next day for Fresno, California. I got that Durango Herald issue about a week after it came out while I was there working briefly for Tom Henderson, and Atmospherics Inc.

After I returned to Durango from Fresno, I sped off to CSU to apologize in person for my lapse to the leader of the CSU experiments, Professor Lewis O. Grant. I had also submitted a “retraction” to the Herald clarifying what I meant. I did see that reporter Mike in the Durango supermarket, and, after I only shook my head at him, he said, “Never trust a newspaper reporter.”

Q. E. D.

But Mike’s article in which I stated I was going to reanalyze ALL of the CSU prior experiments, as you will read, was to have a profound effect that neither of us could have imagined at the time.

12.  The apology and the after effects of the 1975 Durango Herald article

I was able to meet with Professor Lewis O. Grant, the leader of the CSU experiments in his CSU office as soon as I got there, . I groveled and apologized for my possibly libelous newspaper gaffe. He was real nice about it, actually. And, moreover, even when I said I still questioned his seeding experiments and asked for data, like the list of random decisions, he did not hesitate. He was an idealist; questioning was a part of science and he understood that.

Professor Grant’s attitude was not shared by the leader of the experiments in Israel, I am sad to say as Sir John Mason’s letter illustrated.

I kept Professor Grant apprised of my work from Durango as I went along with it as I said I would. As the Wolf Creek Pass experiment began to fall apart in my reanalysis, he even wrote that I had found something important. He was a true scientist.

I also learned from Professor Grant’s graduate student, Owen Rhea, who had started out as the CRBPP’s lead forecaster in 1970 and, along with Paul Willis, had hired me, that the Durango Herald article got back to the National Science Foundation who asked of CSU, “What’s going on?”

According to Owen, due to that Durango Herald article in which I was claiming that I myself would reanalyze ALL of their work, CSU scientists began reassessing their Climax experiments at that time. Those, too, eventually fell apart “upon further review”; their own. Its always best if you find your own problems and report them first before someone else does.

First, in 1978, the earlier claimed evidence of inadvertent downwind increased snow due to seeding at Climax, was found to be due to a synoptic (weather pattern) bias on seeded days. Gone.

Then, in October 1979, at a joint conference of weather modification and statistics at Banff, Canada, Owen Rhea, Professor Grant’s graduate student, verbally withdrew the claims that seeding had increased snowfall in the Climax experiments. Paul Mielke, Jr., the lead CSU statistician, had already done this in a short commentary in the J. Amer. Stat. Assoc. in March of that year, also noting that the stratifications could not have partitioned seeding potential. Climax I and II, gone.

A lucky draw on seeded days had occurred in both Climax experiments; pretty remarkable, though Climax II was to receive some “help” as it turned out, exposed in later independent reanalysis in 1987 by yours truly, with Hobbs.

At that same conference at Banff in 1979, I presented my now published, “Reanalysis of the Wolf Creek Pass cloud seeding experiment” in the May 1979 issue of the J. App. Meteor.) It, too, like the Climax experiments, was the result of a lucky draw and favorable selection of controls by the experimenters, but ones chosen after the experiments had begun, a no-no for experiments because it opens to door to confirmation bias and cherry-picking.

That was my first presentation at a conference. The year before, I had played “center microphone” for a similar conference in Issaquah, Washington. That is, I ran around with a microphone for attendees that had questions for speakers. I was a real “gopher” just the year before.

All in all, the Banff conference was a devastating one for those involved in cloud seeding at CSU, and for those organizations such as the BuRec that had placed such big bets on the CSU experimenters’ original reports.

13.  Pre-1979 Banff conference palpitations and why; the human part of being a science worker in a conflicted environment

The Banff 1979 program that I was going to present in was published in the Bull. Amer. Soc. in May 1979. I was shocked to see that it indicated that CSU faculty would address my paper before I gave it. Thankfully this did not happen. I was an amateur compared to the faculty at CSU, and I was sure all that time before the October Banff conference after seeing the program in May, that my work would be cut to pieces and I would get up red-faced with nothing to say. I had palpitations that whole summer of this nightmare scene, and even redid my paper. Perhaps I had made egregious errors; I was the one that was biased and couldn’t see it.

The evening before my talk in October, I ran into Professor Grant, and he informed me at that time that they were not going to address my work after all. Whew. I had even considered not going; the fear of humiliation was that bad!

Paul Mielke, Jr., also came by, and he simply said, “We screwed up.” I admired him for that and his courageous 1979 article in the J. Amer. Stat. Assoc. In essence, in that article, he had stated that there was no real basis for the 10 million dollar CRBPP the BuRec had signed up for. Can you imagine? The BuRec REALLY did need a “Resident Skeptic!”

The 1979 Banff talk went fine, even got an accolade and a laugh, and I ended by saying, “Who wouldn’t have believed all this evidence was NOT due to cloud seeding?”, trying to put the best face on the CSU seeding collapse that evening. It was an amazing trifecta of indications that seeding had increased snow that CSU scientists had encountered and embraced, but were now gone.

But that was not to last.

CSU scientists began looking again at their Climax experiments and began publishing claims that they had resuscitated valid increases in snow in those experiments in 1981, though they were smaller ones, stratifying the data again by 500 mb temperatures asserting or implying that they had something to do with cloud tops and cloud seeding potential. It was quite a discouraging blow if you care about science.

Neither I, nor Owen Rhea of CSU, could let such claims go unchallenged and we each reanalyzed the new Climax experiment reports, both of us finding a second time in the following years that those claims of increased snow due to seeding by the experimenters were ersatz. There’s much more on this, but will end this discussion here for some hint of brevity.

And, so, while the story today is centered on my work in Israel, the full ppt “book” has a lot of backfill to my experiences in Durango like the ones above, experiences that caused me to distrust any publication regarding a cloud seeding success without extreme scrutiny, the kind that reviewers of journal manuscripts mostly don’t have the time or inclination for.

14.  1983, a real no-no: a request for an independent panel to investigate the reporting of the Climax I randomized experiment

This was a painful chapter, but in trying to be totally candid, it has to come out. There are likely still those out there that know about it, though, as I wrote in my request for this to the Amer. Meteor. Soc., I hoped it would remain completely behind the scenes. It did not. Prof. Grant himself later told an audience that he was under investigation.

Here’s why: CSU statistician, Prof. Paul Mielke in 1979 J. Amer. Stat. Assoc., while withdrawing the claims that the Climax experiments had increased snowfall, observed that both experiments, Climax I and II, had experienced favorable draws that created the impression that snow had been increased on seeded days. It was a courageous post. Here’s what he wrote:

“Very recently, in connection with design studies for a possible experiment of this type in central and northern Colorado mountains, station-by-station precipitation analyses of the Climax I and II experimental units were made for all available hourly stations in Colorado. The resulting maps of seeded to non-seeded mean precipitation amount ratios and non-parametric teststatistic values plotted over the western half of Colorado indicated (for meteorological partitions such as warm 500 mb temperatures) that the Climax experimental results were part of a region-wide pattern rather than an isolated anomaly produced by seeding. In particular, these recent results cast serious doubts on consistency of apparent effects associated with replicated five-year winter periods of the Climax I and Cllimax II experiments.

Later, however, while looking for something else, I ran into this statement at the very end of the article by Mielke et al. (1970, J. Appl. Meteor.), an article accepted for publication on June 30, 1969:

“In an attempt to better define the area extent of the differences between the seeded days and non-seeded days beyond the boundary of the experimental network, available data from all Weather Bureau stations in Western Colorado are currently being investigated.”

Mid-1969 was time that large contracts were being formulated by the BuRec and signed by contractors involved with the CRBPP. One, at least, had already been signed in 1968, the one with CSU scientists for a CRBPP design document, whose interim document was released in October 1969.

What to do after I ran into what seemed to be a “smoking gun”?

It seemed inappropriate to me to have the CSU scientists answer such a profound question on which millions of dollars might depend on the answer: “What happened to the 1969 study that was “underway”? So, I stewed for quite awhile on this seeming “smoking gun.”

Millions of dollars would have been saved, of course, if the CSU scientists had discovered/reported in 1969 the evidence that Climax I had been compromised by a “lucky draw.” It can be assumed that the BuRec would have backed off their plans for the randomization of the CRBPP; perhaps had gone into a research mode with ground and air measurements, or canceled the project altogether to ruminate on what really happened in Climax I. Note: it was well known at E. G. & G., Inc, that CSU scientists opposed randomization of the CRBPP on the basis that, “it’s already been done” (in their own experiments).

Ultimately, in 1983, following a reaction to the CSU scientists’ responses to my friend, Owen Rhea’s reanalysis of the Climax II experiment, I wrote up my request and sent it in to several organizations including CSU, the AMS and NAS. The AMS didn’t know how to go about this (D. Landrigan, personal communication) and I got no response from the NAS.

There was, however, an internal investigation by CSU faculty panel that found no problems in the reporting of the Climax I experiment. I also received a threat of legal action by then Acting Colorado State University President, Robert Phemister if I persisted in my calls for an investigation of the CSU reporting. I didn’t. I still wish that there had been a wider look besides that by CSU faculty, one of whom was a co-author of a seeding paper.

I really hated to do it, knowing the fallout. But, what would you have done if you found the 1969 Mielke et al. “smoking gun?” I just didn’t think they should answer a question with millions of dollars riding on the answer.

I let this issue go downstream, but you can only imagine how CSU and their sympathizers that found out about my unprecedented action might have felt about me. I had asked for an investigation of the most beloved persons in all of weather modification, Lewis O. Grant and Paul Mielke, Jr., both of whom I actually liked as people!

Peter Hobbs, when he found out, was livid; he was not involved because he was on sabbatical in Germany. No one was involved but me. But, I got a raise the next year, 1984. ??

I presented a paper at the Park City, UT, weather mod conference in 1984 with all those present from CSU who knew what I had done. Gads, how did I make it through that one! The tension was so thick. My paper, one that later became part of an AMS Monograph with the other presentations, was titled (I had been assigned this title), “How good are our conceptual models of orographic cloud seeding?”

—————————–

15.  Tension highlight at Park City

It was during this conference that Prof. A.G. from Israel took me aside to sternly lecture me about how wrong I was about the clouds of Israel (from my 1983 rejected article by the J. Appl. Meteor. that asserted they weren’t being described correctly. It was also at that time that he informed me that he had been a reviewer, one of course, that helped reject that paper. His lecture had no effect whatsoever on what I thought about those clouds, why I hopped a plane to Israel two years later.

————————–

If you have read our papers on the Climax experiments, you will know that there was suggestions of tampering with the key NOAA target gauge precipitation data in Climax II (Rangno and Hobbs, 1987, 1995, J. Appl. Meteor.) The values used by the CSU scientists in their analyses were not the ones that were published by NOAA for that gauge; the values that the experimenters used increased the supposed seeding effect, but modestly (4%). And there were many other discrepancies in the 500 mb temperature assignments of storms from those published by NOAA that also “helped” the Climax II experiment “replicate” Climax I.

In contrast, errors were negligible in Climax I; all the precipitation data were the same as in the NOAA publications, for example. Climax I was gifted by a lucky draw of storms with NW flow on seeded days, the direction from which Climax receives it greatest daily precipitation and the set of control stations chosen by the experimenters halfway through the experiment, the least. Climax II had no such luck. Check it out below:

To my knowledge, the results of the 1969 begun Mielke et al. investigation were not made known to the BuRec until Mielke’s 1979 J. Amer. Stat. Assoc.) comment.

———————————

Why would anyone do call for a behind the scenes investigation that would only have negative fall out for everyone involved? I felt I was representing those people outside the cloud seeding community who really paid for the CRBPP. That, too, was the way I felt about my trip to Israel. OK, I know you’re rolling your eyes now, but it was true, I really did think, “Someone has to do something about this!” If I was arrogant (“confidant” is a better word) it was because I thought I could do something given my particular cloud-centric background. I think a lot of “activists” think this way; that they can do something.

16.  Intermission and time for a “Get a life!” note

Following the above comments, it seems like an appropriate point for a reader to erupt with, “Get a life!” See the note at the very end of the science portions of thes piece if that’s what you might be thinking at this point, which is not an unreasonable thought at all.

I did have an outside life somehow. I was single during most of this time, too. There’s no way you could be married/have a partner, and be doing what I was driven to do. Playing baseball, doing some extracurricular forecasting on the radio and for the Washington Huskies comprised most of that outside life.

OK, enough intermission….

17.  A “fruitful perception”

Not trusting cloud seeding peer-reviewed literature, no matter how highly regarded it was, was a fruitful perception. I think you can see why by now!

Over the following twenty years after Durango I reanalyzed, with Prof. Peter Hobbs as my co-author on all but one article, no less than six peer-reviewed, journal published cloud seeding experiments. Not one was the success the original experimenters claimed it to be! PDFs of these reanalyses, and other commentaries on cloud seeding in the literature can be found here:

https://cloud-maven.com/publications-boring/

Important Footnote: To fill out my CV even further on the above page, I have even included my rejected papers and non-submitted reviews as well to make it look bigger than it really is. Of course, those latter items REALLY don’t count in official CVs except to ME. I am hoping to one day to have, as other scientists do, a subset of my papers published: “The Collected Rejected Papers of Arthur L. Rangno.” The volume would be quite thick.

All those published reanalyses and commentaries, and articles/reviews that weren’t accepted or not even submitted, was a vast amount of material I had created, and they were accomplished on my own initiative, my own time (except one, the Skagit reanalysis, was on Peter Hobbs’ time, but my initiative). That is, I worked on these kinds of things on my weekends, evenings, before work, after work at the office, etc. , on and on over years, probably amounting to thousands of volunteer hours to evaluate and “out” faulty cloud seeding claims and to get my views of the cloud seeding arena into print. I even drafted most of my own figures.

I had no funding, of course, for these, well…”altruistic” efforts, as I thought of them. I just felt I had the skills to expose faulty cloud seeding literature being a forecaster and a “cloud man.” I also felt I had a duty to do it since it was likely that no one else would.

To readers: anybody down here?

18. Peter V. Hobbs and his group’s work in cloud seeding

---

19. Life beyond science volunteering: some anecdotes, some humor…maybe

The almost fanatical activity described above can be also be seen as a “crackpot alert.” But, maybe a good one? Yes, and you might well be thinking, as noted, “get a life!”

Well, I did have some outside activities, like playing baseball in a hot semi-pro league called the Western International League, so there. Eight guys were signed off my team over the several years I played on it; one, Mike Kinunen, was pitching for the Twins the next (1980) summer and the guy that batted 3rd in front of me, made the last out of the 1980 college World Series in Omaha playing for the #5 Hawaii Rainbows (defeated by the Arizona Wildcats!) I was the oldest starting player in that league in those halcyon days of my late 30s. In case you don’t believe me:

In my last playing year, I was the recipient of the Jim Broulette “Mr. Hustle” Award in 1980. No, it wasn’t for being a great player, but rather for being an “inspirational” one, which is not as good as being given an award for being great (I had an off year..). FYI, this what I looked like during the era of ruining cloud seeding papers except I wasn’t wearing a baseball uniform when I was doing that.

In a further nostalgic sports report and waste of your time, after the WIL, I pitched batting practice for the Seattle Mariners, 1981-1983. An anecdote about that:

I showed up for a tryout at a workout they were having on the U of WA Husky baseball field in 1981 after the MLB strike had ended and, after pitching BP there, I got to be one of the regular Mariner BP pitchers in the Kingdome, an unpaid job, btw. You get tickets behind home plate. It was so much fun, but stressful. There was an uneasy quiet if you threw as many as three balls that weren’t smacked.

They released me at the end of 1983 because the “guys” were complaining that my ball had too much movement in BP; I was “cutting the ball”, giving it extra spin (private communication, Steve Gordon, backup catcher, 1983). (Unbelievable).

The Mariners of note in those days were Tom Paciorek, Dave Henderson, Bruce Bochte, Richie Zisk and Gaylord Perry, the latter who said my BP was “horrible” in 1983 after he joined the Mariners– he didn’t hit it so well. Of course, he was a washed up pitching buffoon in those days–what would he know about hitting? (Just kidding, Gaylord.) I did throw harder than normal BP pitchers and off or near the pitching rubber, just like I did for my WIL teammates who loved my BP. They wanted zip on the ball like real pitching and I thought the MLB players would, too. And they did, too, that’s why I got “hired” in the first place.

————————

Forecasting for the Washington Husky baseball and softball teams.

I was also the de facto weather forecaster for Washington Husky baseball and softball teams calling rain delays, tarp placements and removals and such beginning in the mid-90s. I had met the Husky baseball coach during my WIL experiences and began forecasting for softball during the 1996 NCAA regional tournament in Seattle which was impacted by numerous showers and even a thunderstorm.

The weather during these spring sports seasons is occasionally showery in Seattle, lots of Cumulonimbus clouds form on those kinds of days, rather than the easy to predict day-long rains from fronts. Radar was pretty useless in showery situations. Why? Because the lifetime of showers is short, and the Huskies could play in SOME rain, just not too hard. So, an incoming shower had to be evaluated by eyeball to assess whether it was dissipating or not; was it all ice or what, and would it rain hard enough to require a tarp and a rain delay? So that’s how I did it, almost completely by eyeballing showers, their movement and growth pattern and assessing their stages.

When the tarp was on the softball diamond during showery days, it was almost harder to call when it should be removed since it took about 45 min to get the game going again; the players had to warm up, besides taking the tarp off themselves. This meant predicting whether a shower would even form in that 45 min time frame, and if so, would it affect the game? The worst possible scenario was that you said to remove the tarp, everyone warmed up again, the crowd came back into the softball stadium, and then it rained hard right after that. It was a stressful volunteer job. Fortunately, that did not happen. I was lucky.

It sounds disconnected, but this was exactly the kind of skill I took to Israel in 1986.

———————————

Before the Husky forecasting era, I had been a forecaster on two different radio stations in Seattle, KUOW-FM (1987-1992), an NPR affiliate in which I came on during “Weekend Edition”, and on a local rock station, KZAM-FM, M-F, for about six months in 1982. For both stations I was doing very short-term forecasts for Seattle using the time of day, such as “no rain through 11 AM, then rain beginning between 11 AM and 2 PM”, etc. When I started these efforts, Seattle had no dedicated weather radar! Doppler weather radar became available only in 1992. In place of radar, you had to use upwind station reports, satellite imagery, know the “territory”, and eyeball the cloud situation along with knowing what the computer model predictions were, and then evaluate how the cloudscape, obs, and how the model predictions were meshing with what the sky was doing.

Perhaps, for sophomoric entertainment, you would like to hear one for KZAM-FM in 1982. In listening to this (sorry, its not real clear), we have to remember that, as the LA Times wrote in 1981, weather forecasting at that time was an era of “clowns and computers” as they headlined. You were expected to come up with some “schtick” if you were a media weather forecaster. And I was encouraged to do so by KZAM-FM. It got a little wild, as you will hear. To stay with the theme of “sports and weather”, I reprise my “sports-like” 1982 weather forecast on KZAM-FM, one that mentioned Gaylord Perry in context with a low pressure in the Gulf of Alaska with “moisture and rotation on it”, known for cheating by throwing spitballs. And damn him for criticizing my BP! It’s a little muffled, but you’ll get the idea. Remember I was forced to do this by the forecasting motif of the day….

OK, I am having more fun now as I remember those crazy days….I still worked at the U of WA cloud group full time during these efforts, too. Good grief, how did I manage all this?

1. 1. 1. The background for the trip to Israel: The British, among other groups, can’t get in to study the ripe-for-seeding clouds in Israel
      2. The British can’t get in: Sir John Mason’s letter
      3. Why I thought I could do something
      4. Story board concerning a trip to Israel to see their clouds
      5. About the clouds I was supposed to see in Israel
      6. 1983, an early jab at “faulty towers”; a paper that questioned the Israeli experimenters’ cloud reports is submitted and rejected
      7. About getting the 1986 cloud study published
      8. The best example of rapidly glaciating clouds I saw
      9. Why was the 1986 study submitted to a foreign journal?

1.  

1.  The background for going to Israel in 1986: no one could get a research plane in to check out those ripe-for-seeding clouds described by the experimenters

By the early 1980s, the events in the journal literature I had experienced during the 1970s caused me never again to believe in a cloud seeding success prima facie no matter how highly regarded it was by national panels and individual experts, as the Israeli experiments were in those days. In Part I, I discussed what happened to me in Durango, a transformation from a pretty naive, idealistic person “going in”, to one that knew those experts alluded to above could be easily fooled and were more naive than I was starting out!  Just too much resting on the reporting a cloud seeding success; glory, funding, status, confirmation of a priori beliefs, for a “let the chips fall where they may” attitude.

The ripe-for-seeding clouds that I went to see in Israel were ones that the HUJ scientists described repeatedly in journals and in conference presentations. They were the foundation for the belief that seeding them had resulted in statistically-significant increases in rainfall that  reported in two, separate randomized cloud seeding experiments, Israel-1 and Israel-2. Those experimenters’ cloud reports explained to the scientific community WHY cloud seeding had worked in Israel and not elsewhere. These first two Israeli experiments were hailed by the scientific community, and in 1982 after the second experiment were described in Science magazine as the ONLY experiments in 35 years of seeding trials that had proven rain increases had been induced by cloud seeding.  (Yes, there was a dreaded scientific consensus that these experiments had proved cloud seeding.)

At the time I went to Israel in 1986, and some of the reason for going, was that no major outside research institution, curious about those Israeli clouds, had been able to get their research planes in to check them out. At least six attempts had been rebuffed (Prof. Gabor Vali, University of Wyoming, personal communication, 1986). The attached letter below from Sir John Mason, former head of the British Royal Society and author of, “The Physics of Clouds,” wrote to me about his attempt to get the British research aircraft into Israel and coordinate such a mission with the lead Israeli cloud seeding experimenter, Professor A. Gagin (hereafter, Prof. AG) of the HUJ. You will find it illuminating about why outside researchers couldn’t get in.

2.  British unable to get in

Why I thought I could do something

So, in going to Israel in 1986 and by then having ten years of experience under my belt in airborne cloud studies with the University of Washington’s Cloud and Aerosol Research Group (CARG), as a weather forecaster, as a former storm chaser (summer thunderstorms in the deserts of Southern California and Arizona, Hurricane Carla) and importantly, as a cloud photographer, I felt I could fill a vacuum left by those rebuffed airborne research missions. Peter Hobbs, the director of our group, put it this way: “No one’s been able to get a plane in there.” It was a very curious situation in itself.

3, Story board concerning my trip and its results

(Hit the expand button in the lower right hand corner for a full view.)

But Peter Hobbs also chided me before just before I went to Israel about my skepticism of the Israeli cloud reports; that I seemed to be indicating to him that I knew more about the clouds of Israel than those who studied them in their backyard. He also added that he thought I was “arrogant.” Wow. Peter, too, was still mad at me for resigning from his group just before a big project and raising a ruckus about it. Moreover, I had scrutinized the HUJ cloud reports in considerable detail, and had submitted a paper on the problems with them in 1983. I had a solid background for my assessment.

Why I resigned from a job I loved, is another long story (oh, not really; you know, it was the old “authorship/credit issue”). But it’s one that ends happily with a reconciliation a couple of years later, which doesn’t always happen! We both benefitted from that reconciliation. We needed each other.

My trip to Israel was self-funded and self-initiated. It may sound ludicrous, but I also felt that by going to Israel I was going to be able to do what those rebuffed airborne missions could not do; evaluate the clouds of Israel sans aircraft. I had flown in hundreds if not thousands of clouds using high-end instrumentation, and when you’re directing research flights as I did for the U of WA research group on studies of ice-in-clouds, mostly Cumulus and small Cumulonimbus ones, you visually assess clouds before going into them and then fly into them sampling the best parts and then see what your instruments have told you about them (concentrations of drops and ice particles, etc.). You get a real quantitative feel for how much ice they’re going to have in them by their external appearance.

So, by just visually assessing the Israeli clouds and estimating their thicknesses and top heights, I would know from my airborne work and background whether the reports about the ripe-for-seeding clouds were correct. Upon closer inspection, there were several odd aspects in the Israeli experimenters’ cloud reports.

Too, if I was right about the clouds of Israel, that they were starting to rain when they were relatively shallow (highly efficient in forming rain, as we would say), say, topping out at 3-4 km (roughly 10 kft to 14 kft) above sea level, the Israeli people might well be wasting millions of dollars over the years by trying to increase runoff into their primary fresh water source, the Sea of Galilee (aka, Lake Kinneret) by seeding unsuitable clouds. They had started a commercial-style program in 1975 after Israel-2, the second experiment, had been reported as a success in increasing rain.

During the first daylight hours of the first showery day I saw, shallow Cumulonimbus clouds full of ice were rolling in from the Mediterranean to Tel Aviv, ones that had been preceded by true drizzle and thick misty rain in Jerusalem the night before. I KNEW that soon that the cloud reports from the HUJ experimenters were grossly in error. Drizzle, btw, was not supposed to fall from Israeli clouds because they were too polluted. Drizzle, instead, meant they were ripe to produce ice at temperatures only a little below freezing due to having large cloud droplets capable of coalescing into bigger drops.

Of course, there were other experienced research flight scientists in cloud studies out there I am sure that could have done the same thing as I did. But, I was the one that went. (Spent a lotta money doing what I thought was an altruistic act, too.)

4.  About the clouds I was supposed to see in Israel

So, what are clouds that are plump with seeding potential supposed to be like? Just that; fat and pretty tall. The clouds that responded to seeding were reported to be those with radar-measured “modal” tops with heights at levels from balloon soundings that were between -12°C and -21°C with the major rain increases due to seeding in the lower half of that temperature range. These would be clouds rolling in off the Mediterranean that were about 3-5 km thick, topping out around 15,000 to 20, 000 feet or so above sea level. Such clouds were described as having a tough time raining, according to the experimenters at the HUJ. They either barely rained, or not even at all, until they were seeded. The effect of seeding in their statistical analyses of the Israeli experiments was that seeding had increased the duration of rain, not its intensity. This was a finding compatible with how the experimenters seeded and also, without direct evidence, led to the inference of deep clouds that didn’t rain until seeded, surrounded the taller ones that did. The experimenters had used just a little bit of seeding agent (silver iodide) released by a single aircraft flying long lines along the Israel coastline near cloud base in showery weather.

It all made sense. Mostly…unless you really got into the details of their reports, in which the devil resides. And I had done that by 1983. See below for a “detective meteorology” module in which the cloud reports of the HUJ experimenters are closely scrutinized and I concluded something was seriously wrong with those reports.

5.  1983:  A paper questioning the Israeli cloud reports is submitted and rejected: a call to action… eventually

In 1983, after plotting dozens of rawinsonde soundings when rain was falling at or fell within an hour of the launch time at sites at Bet Dagan, Israel, and at Beirut, Lebanon, (see first figure in ppt above) I came to the conclusion that the clouds of the eastern Mediterranean and in Israel were, shockingly, nothing like they were being described as by the HUJ experimenters at conferences and in their peer-reviewed papers. I also looked at their published cloud sampling reports and it was clear to me that the clouds that the experimenters had sampled were not representative of those that caused significant rain in Israel; they were too narrow, did not have enough ice particles in them; they did not sample the wide Cumulonimbus complexes that produce the rain for tens of minutes to more than an hour at a time during Israel’s showery winter weather, often marked by thunderstorms.

I submitted a manuscript in July 1983 to the J. Clim. Appl. Meteor. that questioned the experimenter cloud reports. It indicated that rain frequently fell from clouds with tops >-10°C which according to the experimenters’ reports, was never supposed to happen. It was rejected by three of the four reviewers. (Peter Hobbs, the leader of my group, was on sabbatical in Germany and was not happy I had submitted a journal paper without his purview!)

I was undaunted by the reviewers’ take and the rejection; I was pretty sure my findings were correct, which they were proved to be by aircraft measurements in the early 1990s. (Rejected authors, take heart! You may have something really good.)

The problem for reviewers of that 1983 submission?

How could the HUJ experimenters not know about what I was reporting if it was true?

The many rebuffed outside airborne attempts to study Israeli clouds, such as that by Sir John Mason mentioned above, suggested otherwise. I was to fester over this rejection for the next couple of years before deciding to go to Israel and see those clouds for myself, becoming a “cloud seeding chaser”, maybe the first!

6.  About the publication of the 1986 cloud study

Peter Hobbs called Prof. AG a few months before he passed in 1987 to let him know that my article on the clouds of Israel, derived from my 1986 cloud investigation, was going to be published in the Quarterly Journal of the Royal Meteorological Society. The title? “Rain from clouds with tops warmer than -10°C in Israel,” something that the lead experimenter had maintained for many years never happened. In fact, such rain was quite common, as the Israeli experiments Chief Forecaster, Mr. Karl Rosner, states in a 1986 letter to me (posted below), and as I also saw in 1986 during my investigation. Prof. AG passed three months later. Undoubtedly, the appearance of my paper was going to bring many questions and stress for Prof. AG.

7.  The best example of rapid glaciation of shallow cumuliform clouds that I saw in Israel

Shallow Cumulus congestus transitioning to modest Cumulonimbus clouds rolled in across the coast north of Tel Aviv on January 15, 1986. This day’s scene was especially good because of the lack, mostly, of intervening clouds toward a small line of Cumulus and Cumulonimbus clouds. The first shot was taken at 1556 LST and the second shot just four minutes later, 1600 LST. The rising turret peaking between clouds in the first shot had transitioned to ice in that time, taking its possible load of momentary supercooled water with it! This kind of speed occurred repeatedly on this day, and other days when I was there.

Recall that lead of the Israeli cloud seeding program, Prof. AG had asserted in his papers that ice particle concentrations in Israeli clouds did not increase with time. (!) Author’s comment: Not possible.

I estimated the tops of the clouds at 4 km ASL and the temperature at -14°C +3°C based on rawinsonde data. Cloud bases were a relatively warm 10-11°C. This estimate was later verified by radar by Rosenfeld (1997, J. Appl. Meteor.); our full discussion of these photos, including the error in time by Rosenfeld (1997), is found here along with replies to his other comments:

Copies of these slides, with the times above annotated on them, were sent in 1986 to Dr. Stan Mossop, CSIRO, Australia, Prof. Roscoe R. Braham, Jr., North Carolina State University, and Prof. Gabor Vali, University of Wyoming.

8.  Why was my 1986 Israel cloud study submitted to a foreign journal, the British Quarterly Journal of the Royal Meteorological Society?

Ans.: Neither Professor Peter Hobbs nor myself believed that my 1987 manuscript could be published in journals under the auspices of the American Meteorological Society (AMS). So, we went “foreign.”

I believe that this also relates to the problem I have today with the current “Rise and Fall” manuscript submitted to the BAMS under its current leadership. Perhaps the BAMS editors and its leadership feel they are “protecting” Israel, its science, and the HUJ by rejecting a manuscript about faulty Israeli science with elements of misconduct?

My rejected manuscript in 1983 had already suggested that the AMS audience and its reviewers were not ready to hear what I was going to report, and once again I was going to report that the clouds were markedly different than was being described by the HUJ seeding researchers.

The problem with submitting to the AMS, again? Too many American scientists had heard repeatedly in conference presentations or read in peer-reviewed journals about Israeli clouds plump with seeding potential and low in ice content to low cloud top temperatures (to -21°C) as they were being described by the lead experimenter.

In 1983, and again in the 1987 submitted manuscript, I was reporting that the clouds of Israel had little seeding potential due to how readily they rained naturally when cloud top temperatures were barely cold enough for the seeding agent to even work.

So in 1987 we believed that what I was reporting would not fly in an American journal. The major problem again for AMS journal reviewers would be, as it was in 1983:

How could the HUJ experimenters not know this?

Overseas reviewers, however, such as a Sir B. J. Mason, et al (I don’t know who the reviewers actually were) were likely to be more circumspect, and not at all surprised by mischaracterizations of clouds by members of the cloud seeding community. And they were more circumspect. (As they likely would be with my present manuscript.)

My 1987 submitted manuscript was accepted and published in the January 1988 issue of the Quarterly Journal. My conclusions about the general nature of Israeli clouds have been confirmed on several occasions beginning in the early 1990s in airborne measurements by Tel Aviv scientists and later by others. I had indicated to Prof. AG and several other scientists to whom I wrote to from Israel, that from ground observations the clouds of Israel were producing “50-200 ice particles per liter at cloud top temperatures >-12°C” and that “ice was onsetting in Israeli clouds at top temperatures between -5°C and -8°C.” Of course, these were fantastic statements at the time for those scientists that I wrote to, but they were verified by Tel Aviv scientists whose paper with cloud tops and ice concentrations appeared in 1996 (J. Appl. Meteor., Table 4).

The HUJ researchers, however, could only discern this general characteristic of Israeli clouds in 2015; that precipitation onsets only a little above the freezing level. The Israeli experiments’ Chief Meteorologist, Mr. Karl Rosner, already knew this in 1986 (see his letter), as did the Israel Meteorological Service forecasters I spoke with in 1986. What’s wrong with this picture?

Moreover, as happens in conflicted science environments, the HUJ authors of the 2015 paper could not bring themselves to cite my 1988 paper that had reported 27 years earlier what they were finally discovering about their own clouds in 2015. What does this kind of citing tell you about the science emanating from this group at the HUJ? A lot.

The cause of such high precipitation efficiency, the 2015 HUJ authors asserted, was “sea spray cleansing” of clouds coming across the Mediterranean Sea from Europe. This made them ready to produce precipitation at modest depths with only slightly supercooled cloud tops. The Mediterranean Sea is approximately five million years old.

It was, however, in 1992 that the HUJ seeding researchers first discovered that shallow clouds with slightly supercooled tops do rain in Israel; but only in the specific situation when they were impacted by “dust-haze”, and then mostly on the southern margins of showery days, they reported.

So, why did it take HUJ researchers so long to learn this about their “sea spray cleansed” clouds with all the tools at their disposal? Only the current HUJ seeding leadership can tell us.

9.  About the new Israeli randomized cloud seeding experiment and the airborne study that prompted it

Israel, abandoning any idea that the prior cloud seeding experiments had “proved seeding”, again indicative of a terminus, has started over with a new experiment to see, if in fact, cloud seeding works. It’s called, “Israel-4”, now in its seventhseason. No preliminary results have been reported, which is odd. In contrast, the seemingly successful first two Israeli experiments had many interim reports.

Unfortunately, the funder of this new experiment, the Israeli National Water Authority, hired the HUJ “seeding unit” to evaluate seeding potential in the Golan Heights region in preparation for the start of Israel-4, a mistake akin to having the fox guard the hen house.

I reviewed the published article that came out of that HUJ research in 2015 (Atmos. Res.) that described itself as the background airborne cloud study for the new experiment. After reading it, I was not sure it had even been reviewed!

It clearly exaggerated seeding potential in my view; the 2015 authors could not even disclose ice particle concentrations and the rapidity at which they develop in Israeli clouds, critical information for seeding evaluation purposes, claiming that they could not measure ice particle concentrations due to inadequate equipment (manufactured by Droplet Measurement Technologies, Inc.)

I wondered, too, why I wasn’t selected as a reviewer by Atmos. Res. of that 2015 article? My decision on the manuscript would have even been: “accept, pending MAJOR revisions”! This article had some of the best objective writing by the HUJ’s “seeding unit.” But also had a Jeckyl-Hyde aspect where misleading statements kept popping up and along with over-optimized seeding scenarios.

And to the INWA? I would have implored them:

“Don’t do a cloud seeding experiment based on this paper! Get outside researchers to evaluate seeding potential!”

If Israel-4 fails to produce rain via seeding, the faulty HUJ assessment of seeding potential in the Golan will be the cause; the fox will have guarded the hen house as well as expected. And that faulty paper will be consistent with the work of the HUJ seeding group since the early 1970s, work that consistently exaggerated the seeding potential in Israeli clouds and seeding results.