Winter time orographic cloud seeding effects in WRF simulations Lulin Xue (xuel@ucar.edu), Sarah Tessendorf and Roy Rasmussen -- National Center for Atmospheric Research, Research Applications Laboratory Introduction Introduction Model setup Model setup Results Results Summary Summary References References The goal of this study is to use a numerical model to investigate the feasibility of oro- graphic cloud seeding from existing ground-based generators and aircraft seeding tracks in the Payette, Eastern Idaho, and Western Wyoming regions operated by Idaho Power Inc. Table 1. Case summary The Weather Research and Forecast (WRF) model coupled with an AgI point-source module (Xue et al., 2011) was run at 2 km horizontal resolution using Real Time Four Dimensional Data Assimilation WRF (RTFDDA-WRF) forecast data for 10 seeding cases including both ground-based and airborne cases from the 2010-2011 winter season. The results of four cases (Nov. 27, Dec. 02, Dec. 19 and Feb. 16) will be pre- sented in detail (Table 1). A single 2 km horizontal resolution domain covering the snake basin was used to in- vestigate the seeding effects (Fig. 1). Figure 1. Terrain height of the 2 km resolution domain covering the snake basin. Payette water shed and sub- water sheds over east- ern Idaho region are on- lined in black. Ground generator locations are indicated by solid dots. Black: Payette, White: Northern group of east- ern Idaho region, Blue: Sourthern group of east- ern Idaho region, and Green: Proposed Wyo- ming sites. Red sector is aircraft track A4B and green sector is A2B. To investigate winter time orographic seeding effects, a control simulation and a seeding simulation were performed for each case. Figure 2 shows the comparisons of accumulated ground precipitation for the seeding period between WRF simula- tions and available SNOTEL observations. Figure 2. WRF simulated accumulated precipitation over the seeding periods (color shaded in mm) for Nov. 27 (top left), Dec. 2 (bottom left), Dec. 19 (top right) and Feb. 16 (bottom right) cases and corresponding SNOTEL accumulated precipitation (color filled black circle with the same scale as WRF simulated results). It is noticed that the WRF simulated precipitation compared qualitatively well with the SNOTEL observations. Since no quality control has been done on the hourly SNOTEL data, some mismatches between simulations and observations might due to incor- rect SNOTEL data or inaccurate simulation of the precipitation in WRF. Relatively good representation of ground accumulated precipitation is the first step towards seeding effect assessment. Table 3. Precipitation differences between seeding and control cases (acre feet and percentage increase) It is noticed that seeding effects are mostly located outside of the target areas in Dec. 19 and Feb. 16 cases. This feature is the result of inefficient transport of AgI particles into the cloud over target regions. Figure 6 illustrates a typical surface wind pattern over this area. The surrounding complex terrain over upper snake region deviates the wind field and limits the transport of AgI particles. 1. Xue, L., C. Liu, A. Hashimoto, R. Rasmussen, and D. Breed, 2011: Simulations of seeding effects on winter oro- graphic clouds using a two-moment microphysics scheme with a AgI point source module. Jan. 2011, 18th Confer- ence of Weather Modification, Seattle, WA. 2. Thompson, G., P. R. Field, R. M. Rasmussen, and W. D. Hall, 2008: Explicit Forecasts of Winter Precipitation Using an Improved Bulk Microphysics Scheme. Part II: Implementation of a New Snow Parameterization. Mon. Wea. Rev., 136, 5095–5115. Figure 5. Precipitation difference between seeding case and control case (color shaded in mm) for Nov. 27 (top left), Dec. 2 (bottom left), Dec. 19 (top right) and Feb. 16 (bottom right) cases. Water sheds are onlined in red. Green sector in Dec. 2 plot indicates flight track A4B. Winter time orographic cloud seeding events were simulated by WRF model coupled with an AgI point-source module. The sensitivity results show that seeding effects are positive for all cases simulated here. Seeding effects are propotional to seeding rates. Airborne seeding technique is more effective than ground seeding. 18 km resolution RTFDDA WRF forecast data over the western US (operational fore- casts for Wyoming Weather Modification Pilot Program) was used to drive the WRF model. The model configurations are listed in Table 2. The ground generators have an AgI releasing rate of 20 g/h. Aircraft tracks are all at altitude of 3353 m AMSL. For Dec. 2 case, 3.6 kg AgI were released during 2.4 hours seeding period. Table 2. Model configurations Figure 3. Snapshots of WRF simulated super cooled liquid water path (color shaded in g/m 2 ) for Nov. 27 (top left), Dec. 2 (bottom left), Dec. 19 (top right) and Feb. 16 (bottom right) cases. Water sheds are onlined in red. The snapshots were taken in the middle of the seeding periods. Figure 4. AgI particle number concentration at 3000 m AMSL in logarithmic scale (color shaded) for Nov. 27 (top left), Dec. 2 (bottom left), Dec. 19 (top right) and Feb. 16 (bottom right) cases. The snapshots were taken in the middle of the seed- ing periods. The second step is to have the super cooled liquid water content right in the model. The whole idea of winter time orographic cloud seeding program is to bring the super cooled liquid water in the air down to the ground through nucleating water drops into ice crystals and subsequent growing of these ice crystals by diffusion, riming and aggregation processes. The Thompson micro- physics scheme was proved to have advantages over other schemes in WRF in simulating super cooled liquid water content cor- rectly (Thompson et al., 2008). The super cooled liquid water path in g/m 2 in the middle of each seeding period is illustrated in Fig. 3 for each case. They compared qualitatively well with the radiometer observations (not shown). The third step is to simulate the dispersion and diffusion of the seeding agent (AgI particles in this case) correctly in the model. Although offline dispersion models are good tools to investigate this problem, the physical linkage be- tween the distribution of AgI particles and seeding effect requires an online approach. Figure 4 depicts the AgI number concentrations at 3000 m AMSL in the middle of each seeding period in logarithmic scale. Significant amount of AgI particles were transported to levels where super cooled liquid water resides. Seeding effects are expected for these cases. When all the physical processes related to cloud seeding are simulated by the model, the seeding effect can be assessed. The precipitation difference between seeding simulations and control cases are showed in Fig. 5 in mm. All cases showed precipitation enhancements in the domain, basin and target regions (Payette and eastern Idaho water sheds). Negative differ- ences in Dec. 19 case are attributed to the convective feature of the storm. Figure 6. Surface wind vectors at 00:00 01-14-2011 UTC as simulated by WRF model. The wind in upper snake region is blocked and regulated by the complex terrain and shows a re- turnning feature. Figure 7. New generator locations that were tested for Payette and eastern Idaho region (blue dots). Proposed aircraft tracks are indicated by red sec- tors (A1, A2 and A3). Pay- ette water shed and sub- water sheds over eastern Idaho region are onlined in black. The finding of inefficient transport of AgI particles over upper snake region leads to sensitivity tests of additional generator locations and alternative aircraft tracks (Fig. 7). YSU PBL scheme and different seeding rates (1/2, 2 and 5 times of default seeding rate) were also tested. The results are listed in Table 3. (D), (B) and (T) represent Domain, Basin and Target. * are track A1 results for EIDW1219 case. ** are track A2B re- sults for A4B1202 case. *** are track A1 results for NEID0216 case. For Dec. 19 case, 60% additional AgI amount released from track A2 or A3 generated 35-47% more precipitation over the whole domain compared to seeding effect from ground generators, and 150-200% more precipitation over the basin and target areas. Domain size Resolu�on Ver�cal layers Radia�on PBL Land surface Microphysics 420 x 200 2 km 60 (stretching) CAM MYJ Noah Thompson + AgI Nov. 27 Dec. 2 Dec. 19 Feb. 16 Nov. 27 Dec. 2 Dec. 19 Feb. 16 Case date Range Ground gens Gen Times UTC Aircra� track Aircra� �mes UTC 11/27 Paye�e All 1540-2215 none n/a 12/02 Paye�e None n/a 4B 1502-1729 12/19-12/20 Eastern ID All 2200(19)-0500(20) none n/a WY All 2300(19)-0900(20) none n/a 02/16-02/17 Eastern ID North 2300(16)-0400(17) none n/a Case Base YSU* Extra Gens** S05*** S2*** S5*** PAY1127 (D) 685 (0.25%) 614 (0.22%) 810 (0.28%) 451 (0.16%) 981 (0.34%) 1467 (0.51%) PAY1127 (B) 406 (0.18%) 393 (0.17%) 517 (0.21%) 265 (0.11%) 577 (0.24%) 851 (0.35%) PAY1127 (T) 395 (3.5%) 357 (3.5%) 413 (3.9%) 241 (2.3%) 506 (4.8%) 722 (6.8%) A4B1202 (D) 805 (0.29%) 836 (0.30%) 426 (0.16%) 611 (0.22%) 1002 (0.37%) 1257 (0.46%) A4B1202 (B) 421 (0.23%) 412 (0.22%) 323 (0.18%) 303 (0.16%) 534 (0.29%) 669 (0.36%) A4B1202 (T) 451 (8.1%) 454 (8.1%) 295 (5.3%) 333 (6.0%) 559 (10.1%) 676 (12.2%) EIDW1219 (D) 6330 (0.41%) 4961 (0.32%) 6935 (0.44%) 4335 (0.28%) 9388 (0.60%) 13481 (0.86%) EIDW1219 (B) 1466 (0.15%) 1286 (0.14%) 1609 (0.17%) 1111 (0.12%) 2466 (0.26%) 3397 (0.36%) EIDW1219 (T) 1491 (0.56%) 1223 (0.46%) 1496 (0.56%) 1036 (0.39%) 2399 (0.90%) 3322 (1.3%) NEID0216 (D) 451 (0.08%) 217 (0.04%) 544 (0.10%) 366 (0.06%) 698 (0.12%) 932 (0.17%) NEID0216 (B) 154 (0.05%) 83.6 (0.03%) 227 (0.08%) 210 (0.07%) 428 (0.15%) 540 (0.19%) NEID0216 (T) 164 (0.10%) 92.1 (0.05%) 238 (0.15%) 231 (0.14%) 433 (0.27%) 557 (0.35%) Nov. 27 Dec. 2 Dec. 19 Feb. 16 Nov. 27 Dec. 2 Dec. 19 Feb. 16 A1 A2 A3 View publication stats View publication stats