A New Hail Suppression Project Using Aircraft

"NON-REVIEWED" ,’-~RIL 2000 [(RAUSSET A NEWHAIL SUPPRESSIONPROJECT USING AIRCRAFTSEEDING IN ARGENTINA

¯ Terry W. Krauss* RoelofT. Bruintjes HugoMartinez WeatherModification Inc. National Centerfor Atmos- UniversidadNacional de Fargo, North Dakota pheric Research Cuyo - CONICET Boulder, Colorado Mendoza,Argentina

INTRODUCTION OPERATIONALSEEDING PROGRAMDETAILS The Province of Mendozain western Argentina (32 S., 68 deg W.) is one of the most hail prone The current operational seeding period is from areas of the world. Lossesto agriculture dueto October15 th to April 1~t. Themain radar is cen- hail historically are >10%of total annualagricul- trally located at Cruz Negra(33.447 deg south, tural production. The region is knownas the 68.967 deg west) near Tunuyan.The project area "Landof Wine"and the principal crops are grapes covers the three agricultural oasesof Mendoza- plus a large variety of fruit andvegetables (e.g. SanMartin in the north, Tunuyanin the center, apples, cherries, peaches,apricots, plums,mel- and SanRafael in the south. A mapof Mendoza ons, garlic, tomatoes). Theprovince comprises province and the target oasesis shownin Fig- diverse territory of mountains,upland areas, and ure 1. The programconsists of 3 Piper Cheyenne plains. The western border of Mendozais occu- II seedingaircraft, 1 LearJet research-seederair- pied by the mainAndean cordillera, with a number craft, two C-band weather radars with TITAN of peaks above 6400m.Climatic conditions are processingand display software, two Vaisala ra- ¯semi-arid, andthe annualrainfall for the Capital diosondesystems, plus all the necessarystaff. City Mendozais 200 mm. TwoCheyenne II are stationed at SanRafael, the

Hail suppression programsus- ing Russian rocket technology have been conductedin the region since 1985with someinter- ruptions (Makitov 1999). A new airborne hail suppressionproj-. ect has nowbeen implemented ..... by Weather~.Modification ..Inc.....-; (WMI)of Fargo,i...North~Dakota;.... for the Ministry of Economy, Governmentof " Mendoza, gentina. A one-year program was conducted from Dec. 4th 1998 to April 15th, 1999 to. demonstratethe capabilities of airborne hail suppressiontech- niques and methodologies. In October1999, in responseto an international request for bids,. WMIwas awarded a 5 yr con- tract to continuethe operational" seeding programand implement a technologytransfer and re-. searchprogram. Figure. 1: Mapof MendozaProvince, Argentina showing the three areas of operations:, North (Mendozaand San Martin), Central (Tunuyan), and South *Correspondingauthor. (San Rafael). Email: [email protected] Me~o;taTake-Offand Lar~gs (4-~e¢-98 to 28-Mar.99- 206Ffghts) the hail growth process. Someof this complexity is reviewedin the paper of Foote (1985) that at- temptsto classify a broad spectrumof .storm types ,o....o ..... " i ._.!_J_ [ ..i_!.__.!... ’ . - ...... according to both dy- namical and microphysical processesthought to ~ To ¯ - - L .~ ...... ~__..1...._..[: _2...... ~___-:.--.~-_~ _L be critical to hail production. Thegrowth to large hail is primarily hypothe- sized to occur along the edgesof the main storm updraft wherethe merg- ~meof Day(UTC) ing feederclouds interact with the main storm updraft (WMO,1995). Figure 2: The frequency of occurrence and cumulative distributions of air- The cloud seeding hy- craft take-off and landing times for all flights as a function on time. Local pothesis in Mendozais .time is 3 hrs behind universal (UTC) time. Approximate summer sunset based on the concept of and sunrise times are indicated. "beneficial competition". Beneficial competition assumesa deficiency of other CheyenneII and the Lear Jet are stationed natural ice nuclei in the environmentand that the injection of silver iodide (Agl) will result in the at Mendoza. The Cruz Negra radar operates continuously. Radarimages are madeavailable productionof a significant number.of"artificial" ice nuclei. Thenatural andartificial ice crystals "com- on the VVWVVinternet at 15 min intervals. All of the project’s radar data, meteorologicaldata, air- pete" for the available super-cooledliquid cloud craft data, and reports are recorded onto CD- water within the storm. Hence,the hailstonesthat ROMfor the Ministry of the Economy.The audit- are formedwithin the seededcloud volumeswill ing of the operationalprogram is being conducted be smaller and produce less damageif they shouldsurvive the fall to the surface.If su~cient by the NationalUniversity of Cuyoin Mendoza. nuclei are introducedinto the newgrowth region of the storm, then the hailstones will be small CONCEPTUALHAIL MODEL enoughto melt completely before reaching the ground. Cloud seeding operations alter the mi- crophysics of thetreated clouds, assumingthat The cloud seeding is based on the conceptual the presentprecipitation processis inefficient due modelof Alberta hailstorms which evolved from the experimentsand studies of Chisholm(1970), to a deficiency of natural ice nuclei. Cloudseed- ing doesnot attemptto competedirectly with the Chisholmand Renick (1972), Bargeand Bergwall energyand dynamics of the storm. Anyalteration (1976), Kraussand Marwitz (1984), and English of the storm dynamicsoccurs as a consequence (1986). Direct observational evidencefrom the of the increasedice crystal concentrationand ini- instrumentedaircraft penetrationsof Coloradoand Alberta stormsin the 1970’sand early 1980’sindi- tiation of riming andprecipitation sizedice parti- cates that hail embryosgrow within the time cles earlier in the cloud’slifetime. Thesilver-iodide reagentsinitiate a condensation-freezingprocess evolving "main"updraft of single cell stormsand within the updrafts of developing"feeder clouds" and produce enhancedconcentrations of ice crystals that competefor the available, super- or cumulustowers that flank mature"multi-cell" and "super-cell" storms (see e.g. Foote, 1984; cooledliquid water in a storm andthereby reduce the growth of large damaginghail. The seeding Krauss and Marwitz, 1984). The computationof also initiates the precipitationprocess earlier in a hail growth trajectories within the context of cloud so that cloud hydrometeorsmay proceed measuredstorm wind fields provided a powerful newtool for integratingcertain parts of hail growth via an ice-phase(graupel) to rain mechanismin- stead of continuingto growto damaginghail. theories, andillustrated a striking complexityin "NON-REVIEWED" 75 APRm2000 CLOUDSEEDING METHODOLOGY" -6°C,-8°C, -10°C, and-12°C.The flares used in Mendozaproduce between1012 and 1013 ice nu- Stormcells (defined by radar) with maximumre- clei per gramof pyrotechnic between-6°C and - flectivity >35dBZ within the cloudlayer abovethe 10°C. The newgeneration ICE pyrotechnic pro- -5oC level, located within the project areas or duces>101~ ice nuclei per gramat -4 C. Ratesof within a 20 mintravel time "buffer zone"upwind of. ice crystal formationin the CSUisothermal cloud the project areas, are seedingcandidates if they chamberwere extremely fast (63%active within are deemedto be a potential hail threat. Radar one to two minutes, and 90%active within four meteorologists are responsible for makingthe minutes) and the apparentice formation mecha- "seed" decision anddirecting the cloud seeding nismwas identified as condensation-freezingnu- missions, incorporating the observationsof the cleation (Demott,1999). Significant ice formation pilots into their decisions.Patrol flights are often activity was measuredas warmas -4°C. Little launchedbefore clouds within the target area dependenceof yield on cloud liquid water content meetthe radar reflectivity seedingcriteria. These or clouddroplet distribution characteristicsis ex- patrol flights are meantto providea quicker re- pectedbased on the current tests. High yield and sponseto developingcells. In general,a patrol is fast acting agentsare importantfor hail suppres- launchedin the eventof visual reports of vigorous sion since the time-windowof opportunityfor suc- towering cumulusclouds or whenradar cell tops cessful intervention of the hail growthprocess is exceed20 kft height over the higher terrain along often less than 10 minutes. the western border on days whenthe forecast calls for thunderstormswith hail potential. 1998-99 FLIGHTSTATISTICS During the demonstrationproject of 1998-99,two Cloud top seeding can be conducted between hundredand ten (210) aircraft flights (515 flight -8°C and -15°C. The 20 g pencil flares fall ap- hours) were conducted on 64 days betweenthe proximately4000 ft (approximately10°C) during period Dec. 4th, 1998and March28th, 1999. Of their 35-40 s burn time. The seeding aircraft these, 161flights (437hrs) wereseeding flights, penetratethe edges of single convective cells and41 flights (66 hrs) werefor weathersurveil- meetingthe seedcriteria. For multi-cell storms,or lanceonly. Therewere 111 night time flights (286 storms with feeder clouds, the seedingaircraft hrs) and99 day-timeflights (229hrs). Thedistri- penetrate the tops of the developing cumulus bution of take-off andlanding times accordingto towers on the upshearsides of convectivecells, the time of day is shownin Figure2. as they growup throughthe -10°Caltitude. SEEDINGMATERIALS Cloud seeding was conducted on 52 days. The seedingconsisted of 18,480 droppable/ejectable ¯ Silver-iodide is dispensedusing droppable and/or flares (369,600gm seeding material), 2,015 burn- end-burning pyrotechnics. Acetonegenerators in-place flares (302,250gm seeding material), and were also used during the 1998-99 campaign, 642litres of acetonesolution (11,096gms of sil- however, they have nowbeen abandoneddue to ver-iodide). The total massof seeding material the penalty they_ pose. on aircraft performance. released was682,946 gmso WMInow uses silver,iodide formulation flares.. manufacturedby--Ice Crystal Engineering(ICE) Operationswere conductedin the north oasis on Davenport, ND. The ejectable flares contain 20 32 days, dispensing 268.5 kg of seed material. gmof seedingmaterial andburn for approximately There were 10 days with reported hail damagein 35 sec and fall approximately4000 ft. The end- the north. Operationswere conductedin the cen- burning flares (BIP) contain 150 gmof seeding tral oasis on 24 days, dispensing126.9 kg of seed material and burn for approximately5 rain..The material. There were 9 days with reported hail ice nucleating effectiveness of the flares manu- damagein the central oasis. Operations were factured by ICE, have been tested at the Cloud conductedin the south oasis on 34 days, dis- Simulation and Aerosol Laboratory at Colorado pensing287.4 kg of seed material. Therewere 16 State University (Demott, 1999). The primary dayswith reportedhail damagein the south. productof the laboratory characterizationis the ¯ "effectiveness plot" for the ice nucleant which. ATMOSPHERICCONDITIONS gives the numberof ice crystals formedper gram A summaryof ~omeof the most important atmos- of nucleantas a function of a rangeof cloudtem- pheric parameters measuredby the radiosonde peratures. Specified temperaturesfor testing in balloon launcheddaily at 16:00 UTC(13:00 local the isothermal cloud chamberare typically -4°C, time) at the CruzNegra radar site is given in Ta- 76 -JOURNALOF \VE...\THER NIODIFIG,,-\TION ble 1. The period for this summaryis from Dec. The average -10°C height (used for a common 6t" , 1998to March31 st, 1999inclusive. Thedata seedingaltitude) wasin the range18 - 19 kft. The are summarizedfor all days, days with thunder- mid-level windswere typically southwesterly(from storms, and days on which damaginghailstorms 250-260 deg true) and the average storm cell occurred. speed was 12 - 14 kts. A commonobservation from the soundingswas the dramatic increase in Eachof the 22 days with hail damagewas fore- the wind speedabove 20 kft, which corresponds cast as a hail day by the project meteorologists to the height of the Andes Mountains. The using the index called the CDC(Convective Day mountainshave a dramaticeffect on storm forma- Category) as defined by Strong (1979). Storm tion that is not fully understood.The mountains cloud base heights averagedbetween 8-10 kft are believed to causea cappinginversion which and the averagetemperature at cloud base was helps to suppressthe prematurerelease of con- 8°C to 12°C. Pilots usually reported that the vectiveinstability, but also are thoughtto provide feeder cloud baseswere between10-12 kft. Se- a lifting mechanismfor triggering thunderstorms vere hail storms tended to have the lower and after sunset, possibly throughthe formation of a warmercloud bases. The average surface dew low-level-jet streamand Bernoulli type effect. point on dayswith hail was11.46°C and the pres- Moreresearch is required into the effects of the enceof high surface humidity wasa goodpredic- mountainson the hailstorms in Mendoza, tor of severestorms.

Table 1: Summaryof important atmosphericparameters measured by the radiosondeballoon launcheddaily at 16:00 UTC13:00 local time) at the CruzNegra radar All Days ThunderstormDays Hail DamageDays Parameter Units. Average StdDev Average StdDev :Average StdDev FCST CDC None -0.20 1.75 0.73 1.53 1.71 0.72 Precip. Water In 0.72 0.19 0.79 0.20 0.90 0.19 0 deght. Kft 14.22 1.55 14.36 1.41 13.90 1.65 -5 deght. kft 16.51 1.39 16.54 1.34 16.13 1,55 -10 deght. Kft ¯ 18.77 1.33 18.71 1.31 18.50 1,49 CloudBase Ht. Kft 9.01 2.63 9.13 2.49 8.06 1.92 Cloud Base Deg C 8.70 4.55 9.83 5.07 12.28 3.85 Temp. Max. Storm Top Kft 31.63 9.30 ¯ 35.76 7.35 39.29 3.47 Temp.max. Deg C 30.47 4.74 32.21 3.93 32.36 3.87 DewPoint Deg C 8.30 3.94 9.42 4.27 11.46 2.60 ConvectiveTemp. Deg C 31.73 7.01 32.71 5.10 30.63 4,56 Total Totals Index None 44.65 7.95 48.16 6.02 52.14 3.85 Lifted Index None 0.49 4.00 -1.56 3.40 -3.64 2,13 StormDirection Degtrue 252.58 64.00 258.62 58.80 278.35 47.21 Storm Speed knots 15.40 8.65 13.66 6.06 12.82 4.64 Cell Direction Degtrue 250.33 64.03 255.36 63.63 262.53 71.68 Cell Speed knots 15.60 8.88 14.00 6.51 ;12.76 5.01 700 mbwnd. Dir Degtrue 192.38 132.52 196.45 133.40 172.41 137.20 700 mb wnd. Spd knots 8.80 5.28 9.68 5.86 10.41 6.49 500mb. Wnd.Dir. Degtrue 260.56 45.21 268.15 38.41 274.29 38.50 500 mb. Wnd. knots 25.21 12.57 24.98 13.23 23.47 11.34 SPd. 300rob, Wnd. Dir. Degtrue 260.71 50.83 265.35 41.31 273.24 34.77 300rob. Wnd. knots 48.37 21.33 48.61 21.56 50.88 25.30 Spd. Outlook CDC none -0.13 1.52 0.44 1.18 0.52 1.17 ,,NON-REVIEWED" 77 LEAR JET RESEARCH-SEEDER 1000 feet below the growing cloud .top. There was no natural ice detected so the cloud was The Lear Jet research-seederaircraft conducted seededduring the penetration¯ with 7 ejectable five research experimentsand four seedingmis- flares (ICE, 20 gmAgl type). Thecloud Continued sions during Februaryand March1999 to char- to growso the LearJet ¯climbedto -14°Cfor the acterize the microphysical conditions of the second¯penetration at 21:13:59and a spike in the clouds, to confirm the microphysicalseeding ef- ice concentration from the 2DCprobe was ob- fects, andto demonstrateits superior ability to served inside the cloud. The King liquid water seed clouds. More research flights have been content and the 2DCice concentration for cloud conductedin 2000. TheLear Jet is equippedwith pass number1 and 2 are shownin Figure 3. The 408 flares and has the performance to move first passhad a peakliquid water content of 2.2 quickly from storm to storm and seed many. g/m3 and no ice. The secondcloud penetration stormsduring a single mission. Thecloud physics hada peakliquid water contentof 1.9 g/m3 and a instrumentation consisted of a PMS-FSSPand maximumone secondice concentration of 902/L.. OAP-2DCand was able to measure the atmos- The 2Dimages show that the high ice concentra- pheric conditions, c~oudvariables such as the tion seeding signature consisted of mostly 100 cloud droplet and ice crystal concentrationsand microndiameter thin plate-like crystals. This is sizes, andmicrophysical responses to seeding. the fastest responseto silver-iodide seedingthat the authors have ever measuredwithin a cloud Liquid water contents > 1 g/m3 have beenmeas- andverify the. fast ice nucleationrates measured ured at temperaturesas cold as -38°C. The liquid in the CSUcloud chamber.The older formulation water contents decreaserapidly at temperatures TB1silver-iodide flares typically took approxi- colder than -38°Cas an apparentresult of homo- mately 6 min before the ice signature was de- geneousfreezing of the water droplets. Thecold tected using 2DCprobes ¯ (Krauss et al., 1987). temperaturesby the Lear Jet have confirmed an Thenew formulation ICE probes, therefore, repre- abundantsupply of super-cooledwater for hail sent a significant decreasein responsetime in the stone growththrough a very large vertical extent life-cycle of precipitation developmentfor conti- as a typical occurrencein Mendozastorms. nental clouds.

Droplet concentrations as measuredwith an FSSP-100vary between 300 and 600 cm3 just above cloud base indicating a continental air- mass. Cloud bases are generally between+8°C and+12°C. The limited observationsto date indicate that the coalescenceprocess is not active in convective clouds in the area. This wasfurther supported b~/ radar measurmentsthat indicate that the first echousually develops¯at temperatures colder than -10°C. The cloud droplet mean- volume-diameter at -10°C is approximately 15 urn, and at-20°C approximately 25 um. These are characteristic of continentalclouds, however, from these limited observations,it is not possible to determinethe presenceor absenceof low con- ¯ Figure 3: The King liquid water content and the 2DCice centrations of large drops (e.g. drizzle drops of concentration for cloud pass number1 (right) at 21:11:27, 100um diameter) and their effect on the growthof and cloud pass number2 (left)at 21:13:59. hail in theseclouds. The third cloud penetration was madeat the - 15°Clevel at 21:17:50 and a peak liquid water AGI SEEDINGCASE STUDY content of 1.5 g/m~ and peakice concentrationof 788 /L were measured. The King liquid water On 2-March-2000, a seeding experiment was content and the 2DCice concentration for cloud conducted on an isolated cumulus-congestus pass number2 and 3 are shownin Figure 4. The cloud. The convective cloud base on this after- 2Dice crystal" imagesfor a portion of penetration noon was at 9300 ft MSLwith a temperature of 3 are also shown.The ice crystal seedingsigna- 8°C.. Thefirst penetrationwas made at 21..:11:27 ture wassplitting the updraft coreof the cloudand UTC,at a temperatureof -10 C, approximately depleting the liquid water. In contrast, the first 78 ,JouRN,",J.. 01;" ~,’VIL~THER,~IODIFIGATION natural ice is most often detected at the edgesof the cloud.

The 4th cloud penetration was madeat 21:17:50 at the -14°C level and recorded a peak liquid water of 1 g/m3 and a peak ice concentration of 162/L. The 5th cloud penetration was madeat 21:20:37 at the -16°C level and recorded a peak liquid water of 1.5 g/m3 in a newlydeveloping tu_rret, but a liquid water concentration of 0.6 g/m~ in the older, seeded portion of the cloud. The peak ice concentration was 82/L. in the seeded portion but .essentially zero in the newly growingportion. The 6th cloud penetration was madeat 21:22:55 at the -14°C level and recorded a peak liquid water of only 0.2 g/m3 and a peak ice concentration of 84 /L. The King liquid water content and the 2DCice concentration for cloud pass number 5 and 6 are shown in Figure 5. A radar echo of 20 dBZ had formedat this time.

Theevolution with time of the ice size spectra for. the cloud penetrations number2 to 6 (after seeding) is shown in Figure 6. The high concentrations of ice and growth to large sizes are clearly shown. The legend indicates the average ice concentration across the ice plume during the penetration..This single cloud seeding experiment confirms the effectiveness of the ICE flares and supports the beneficial competition and cloud liquid water depletion steps in the conceptual hail suppression model.

CROP DAMAGE EVALUATION

Crop damagestatistics obtained from the Ministry of Economy(Direccion de Prevencion de Contin- Figure4: TheKing liquid watercontent and the 2DCice ¯ gencias) indicate that the 1998-99season had the concentrationfor cloud pass number2 and3. The2D ice lowest hail damagein the last 10 years. The crystal imagesfor a portion of penetration 3 are also ¯ shown. The distance between vertical lines for 2DC damagedarea for 1998-99 was 41%of the previ- probeshadow images is 800 microns. ous 9 year average. Furthermore, only 33%of the year’s damageoccurred after the seeding project begin, since 67%of the damagesoccurred before Dec. 4t", 1998. A graph of: the area of hail damage(equivalent to 100%) for the three oases since 1989-90 is shown in Figure 7. The season was not unusual in climatic terms with respect to precipitation and temperature, however, it maybe that the 1998-99 season may have had below "NON-REVIEWED" 79 normal storm intensities and frequency of severe room for improvement. It is therefore important hailstorms. Preliminary indications for the 1999- that studies to investigate the effectiveness of hail 2000 season indicate manymore severe storms, suppression and the economical impacts continue therefore, the area of hail damagewill certainly be in Mendoza.In considering the outline Of a possi- greater for the second seeded season. The air- ble future research program two primary objec- borne seeding effectiveness can not be evaluated tives have beenidentified; using these measures in only one or two years, therefore, no .conclusions regarding the effec- ¯ Objective A: To learn more about the nature tiveness of the seedingare madeat this time. of hailstorms, both natural and seededin order to provide: 1. Evidence for the effectiveness of hail suppression seeding 2. Evidence for the economic benefit of hail suppression(cost benefit analysis) 001 3. A basis for improving the effectiveness of hail suppression

0.001 ¯ Objective B: To educate and train a new c](XX]l generation of scientists in Argentina to carry on this researchin the future.

As part of the current five year program, four professors from the Universidad de Cuyoin Men- doza have been integrated into the program and two students are currently working With WMIstaff Figure 6: The evolution with time of the ice size spectra in the program.In addition, eight Argentine Pilots for the cloud penetrations number2 to 6 (after seeding). are flying as co-pilots on all the seeding,missions.

DfRECCIONDEpREVENCION DECON11NGENCIAS SUPEFiCIE~AFECTADA$ AL100% DE DANO CONCLUSIONS AND RECOMMENDATIONS

In conclusion, airborne seeding technology and hail suppression seeding concepts and method- I " ’’ " ’ . " ’ ologies were successfully demonstrated during the summerof 1998-99. Considerable knowledge and experience were obtained regarding the hailstorms of Mendoza. In addition, the people of Mendozahave learned a lot about the capabilities and methodologiesof aircraft seeding operations. Preliminary results indicate that the 1998-99hail damageto crops in the province of Mendozawas the lowest during the last ten years, however, the second season is experiencing above average Figure I: Area of hail damage (equivalent to 100% dam- hailstorm frequency and damage,therefore, it is age) for the three oases since 1989-90. The damage area muchtoo early to draw any conclusions regarding before seeding and after seeding began on Dec. 4th, 1998 is shown separately for the past season. the effectivenss of the airborne seeding program at this time. A research and evaluation program must now be implemented to determine the im- PHYSICAL STUDIES pact of the seeding on the frequency and occur- renceof hail. Physical studies formeda small but important part of the first two years activities and will hopefully The Lear Jet has already conducted several re- continue in the future. Although hail suppression search missions within high altitude cloud regions has shown signs of success in the past (WMO that have not been possible in the past. Some statement on Weather Modification; AMSState- seeding signatures were detected, however, more ment on Weather Modification), the technology is microphysical measurements are required to ¯ as yet not scientifically established and there is document the concentrations of super-cooled 8O ¯ JOURNAL OF WEATHERNIODIHGATION ~V’OLU~IE(3 water, especially at temperatures colder than -20 English,M., 1986:The testing of hail sul~pressionhy- C. Muchmore data is required to characterize the pothesesby lhe Alberta Hail ProiecLPreprints, initiation and evolution of the natural ice concen- 10th Conf. WeatherModification, Arlington, trations and the effect of seeding on the super- Amer.Meteor. Soc., 72-76. cooled liquid water of the storms. Furthermore, Foote, G.B., 1984:The study of hail growthutilizing more research is recommendedto determine the observedstorm conditions. J. Climate. Appl. Meteor.,23,84-101. variety of effects that the mountainshave on the Foote, G. B., 1985:Aspects of cumulonimbusclassifi- formation, structure, and characteristics of hail- cation relevant to the hail problem.J. Rech. storms. A research and technology-transfer pro- Atmos.,19, gram is being implemented to address these is- Krauss,T.W., and J.D. Marwitz, .1984: Precipitation sues during the next four years. processeswithin an Alberta supercetl hailstorm. J. Atmos.Sci., 41, 1025-1034. Acknowledgements: The authors wish to thank Krauss,T.W., R.T. IBruintjes, d. Verlinde andA. Kahn, the Ministry of Economy,Government of Mendoza 1987.Microphysical and radar observationsof seededand non-seededcontinental cumulus for their support of the operational programand clouds. J. ClimateAppl. Meteor.Vo1.26, 585- scientific studies. 606. Makitov,V., 1999:Organization and main results of the REFERENCES hail suppressionprogram in the norther areaof the province of Mendoza, Argentina. J. Barge, B.L., and F. Bergwall, 1976:Fine scale struc- WeatherModification, 31, 76-86. ture of convectivestorms associated with hail Strong, G. S., 1979:A convectiveforecast index as an production. Rep. 76-2 (2 Vols.), Atmos.Sci. aid in hail suppressionevaluation. Proceed- Div., A~bertaResearch Council, Edmonton. ings, 7th Conf. Inadvertent and Planned ¯ Chisholm,A. J., 1970: Alberta hailstorms: A radar WeatherModification, Amer. Meteor. Soc., study and model. Ph.D. Thesis, McGill Uni- Banff, 2pp. versity, Montreal,287pp. World Meteorological Organization (WMO),1995: Chishotm,A. J., andJ. H. Renick,1972: The kinematics WMOMeeting of Experts to Reviewthe Pres- of multicell and supercellAlberta hailstorms, ent Status of Hail Suppression,Golden Gate Alberta Hail Studies, 1972.Research Council National Park, SouthAfrica, 6-10Nov. "1995, of AlbertaRep. 72-2, 24-31. Prof. R. List, Chairman,WMO Geneva. DeMott,P.J., 1999:Report to WeatherModification Inc. onTests of the ice nucleatingability of aero- sols producedby ICE pyrotechnics. Dept. At- mos.Sci., ColoradoState Univ., Report, Fort Collins.