Effects of Cloud Seeding in West Texas: Additional Results and New Insights

Abstract

Additional results and new insights have been obtained regarding theeffect of randomized dynamic seeding of supercooled convective clouds inwest Texas. These have resulted in a revised conceptual model that aidsin understanding the new results and in designing future physicallybased experimentation and cloud modeling studies.Attention is focused initially on further evidence for seeding effectson convective cells, which are the treatment units for the experiments.A total of 183 cells {[}93 seeded (S) and 90 nonseeded (NS)] have beenidentified and their properties computed through analysis ofthree-dimensional, volume-scan, C-band radar data using cell-trackingsoftware. The results indicate that AgI seeding increased the maximumheights by 7%, the areas by 43%, the durations by 36%, and the rainvolumes by 130%. Cell merger occurred nearly twice as often in theAgI-treated cases. The rainfall and merger results are significant atbetter than the 5% level using rerandomization procedures.Additionally, it was found that AgI-treated cells produced more rainfallthan untreated cells of the same height.Time-height reflectivity composite cross sections for the S and NS cellsrevealed stronger reflectivities aloft immediately after seeding for theAgI-treated cells. This region of enhanced S reflectivities expanded anddescended with time, reaching the earth's surface by 40 min afterinitial seeding.The next step focused on the areas surrounding the treated cells. A newand improved `'focused area'' approach, involving calculations for radiiof 7, 10, 15, 25, 50, and 100 km around each treatment position wasdeveloped and applied. The rainfalls from the seeded cells exceeded therainfalls from the nonseeded cells at radii less than 10 km early in thetreatment period, and this apparent effect spread to larger radii withtime. These results are consistent with a positive effect of AgItreatment on rainfall that begins on the cell scale, where the seedingtakes place, and spreads outward with time.Analysis of the 34 (17 S and 17 NS) small mesoscale convective clusters(i.e., the experimental units) obtained to date produced ratios of S toNS rainfalls of 1.37, 1.27, 1.37, 1.26, and 1.27 for the time periods0-30, 0-60, 0-90, 0-120, and 0-150 min, respectively, after initialseeding. None of the results has strong P-value support. The ratios arelarger and more significant when the five experimental units that failedto meet the qualification criteria are eliminated from the sample.The revised dynamic seeding conceptual model gives much more attentionto microphysical processes than before. How dynamic seeding likelyinduces greater cell rainfall without an appreciable increase in maximumcell height in some cases is addressed by the model. The model is shownto be useful also in identifying which cloud processes should be thefocus of recommended future physically based experimentation and cloudmodeling.