WMOD00473 - Laserfiche WebLink

Home Browse Search

2000 JOURNAL OF CLIMATE AND APPLIED METEOROLOGY VOLUME 22 TABLE 4. Ridge temperature partitions: Wilcoxon (W) and MRPP probabilities for each winter tested separately. Experimental days Zone I Zone 2 Zone 3 Range (oC) All sins W MRPP W MRPP W MRPP Part A 1970-71 data ~ -7.0 61 27/34 0.10 0.05 0.47 0.46 0.20 0.09 ~ -9.0 46 18/28 0.06 0.05 0.31 0.46 0.16 0.12 ~ -11.0 30 12/18 0.03 0.02 0.05 0.06 0.004 0.003 ~ -13.0 25 9/16 0.04 0.04 0.03 0.05 0.004 0.003 Part B 1971-72 data ~ -7.0 70 38/32 0.21 0.29 0.33 0.46 0.35 0.40 ~ -9.0 54 26/28 0.07 0.12 0.17 0.28 0.20 0.26 ~ -11.0 38 19/19 0.08 O.ll 0.28 0.34 0.24 0.34 ~ -13.0 24 13/11 0.04 0.05 0.05 0.08 0.16 0.22 probability values are not as low as those of the former winter, the important point is that similar results were achieved in each of two separate winters. The 100 days with mean Ridge temperatures of -90C and below yielded low probability values in Zone 1, the intended target, while the partition of -130C and below yielded low values by both tests in all three zones. These two subpopulations have consequently been chosen for more detailed illustration of results. Table 5 shows various statistics for the seeded and nonseeded populations for the three zones. It can be seen that correlation coefficients are highest between the control gages and those of Zone 1, ranging from 0.75-0.85. Precipitation amounts are highest, and days with zero precipitation least frequent, at the con- trol and Zone 1 gages-as would be expected with higher elevation sites. The frequency of precipitation, mean and median precipitation amounts and corre- lation coefficients generally decrease with the lower elevation sites further downwind in Zones 2 and 3. Therefore, the results there should be interpreted with more caution. The mean and median precipitation amounts at the control gages are seen to be less for the seeded days than for the nonseeded. This difference is in the di- rection of the type I statistical error suggested in pre- vious analysis of the BRE as discussed in Section 1. In contrast, mean and median precipitation amounts are similar for seeded and nonseeded days in Zone 1. It is interesting to examine the spatial variation of the differences associated with seeding. This has been. done by calculation of the Wilcoxon probability and mean double ratio for each gage and plotting of the results in Fig. 3. It should be expected that calculation of these statistics for individual gages will yield more "noise" in the results than use of the mean of all gages in a zone as done previously. Local topography and forest cover would be expected to increase the variance in single site measurements. . Figure 3 shows the spatial pattern for the 100 days with Ridge temperatures ~ -9.0oC. It indicates that the lowest probability values and highest double ratios tend to be in the northern portion of Zone 1, the intended target. However, three additional gages have probabilities less than 0.05 along the southern edge of Zones 1 and 2 and one gage has a value of 0.05 far downwind in the foothills of the next mountain range. The highest probabilities and lowest double ratios tend to be in the central and northern portions of Zones 2 and 3, or east of the low probability gages in the north- ern portion of Zone 1. Thus, there does seem to be a general pattern in probability values with either high or low numbers usually grouped together. It would be disquieting to observe alternating high and low values in close proximity in a random fashion. The AgI plume tracing results and generator cali- brations previously cited suggest that mean ice particle concentrations should have ranged from about ten to . several hundred per liter in the AgI plumes for the partitions ~ -90C. Concentrations at warmer tem- peratures would be expected to be much lower. Thi~ offers a'plausible phy~ical explanation for the statistical suggestions. b. Partitioning by 700 mb temperature As discussed in Section 4, the 700 mb temperature (hereafter noted by T7(0) is believed to approximate the AgI plume-top temperature above the Main Ridge. The mean height of the 700 mb surface was 2950 m, so T700 was closely assoCiated with the Main Ridge temperature measured at 2595 m. The reasoning for using T700 in addition to Ridge temperature to partition the experimental days is as follows. The Ridge tem- peratures are averages for the entire 24 h day. They indicate nothing regarding the presence of clouds suit~ able for snowfall formation. Often, however, storms lasted for only some fraction of the day. Therefore, I'. '1 "