WMOD00523 - Laserfiche WebLink

Home Browse Search

midway from the HAS to t~e TAR on the west edge ofthe plateau top, GTR near the TAR, and GSO I km due south (crosswind) of GTR. These three are referred to as the "core" gauges, located within 2 km or less of one another. The downwind gauge GDN had a P-value of 0.09, somewhat suggestive of seeding effectiveness, but not meeting the 0,05 criterion, These test results. from the entire EU population, provide a strong statistical suggestion that propane seeding increased the "seasonal" 3.5 month experimental period snowfall on the windward slope and upwind portion of the plateau top. The "sc~eding signal" is less obvious at the gauge furthest downwind which had a lower likelihood of routinely being under the seeding plume because of variations in wind direction. Table I. Summary for all available 98 EUs with no partitioning: 47 pairs, 50 seeded and 48 nonseeded EUs. Pooled Re- Target - randomized I-tailed 95% Target GNO Sample I-tailed Confidence Gauge Correlation MDR P-value Interval GSC 0.85 1.22 0.06 1.02 - OCJ GTR 0.88 1.22 0.04 1.04 - OCJ GSO 0.89 1.23 0,03 1.06 - OCJ GDN 0.84 1.21 0.09 0.98 - OCJ Whereas a significance test can imply there is a seeding effect, confidence intervals can quantify the possible range of effect which seeding may cause. Rerandomization, the repetitive simulation of the experiment using the same data but with different random seeding decisions, was required for the derivation of P-values and confidence limits shown in Table I. The same set of 1000 rerandomized MDRs used to derive the P-values Was applied to estimation of confidence limits. For example, a test of significance using a = 0.05 might estimate from the MDR that treatment provided an additional 10% SWE. The upper limit of the confidence interval for the I-tailed application, which assumes seeding can increase but not decrease snowfall, cannot be specified. Hence, upper limits in the tables are given by the "00" symbol meaning infinity. The lower limits for the core gauges in Table 1 range from 1.02 to 1.06 corresponding to minimum precipitation increases of2, 4 and 6%, respectively. The lower limit for gauge GDN is 0.98 meaning that it is possible that decreased precipitation could have resulted there, but by no more than 2%. That gauge also has the highest P-value(lowest significance). All the confidence interval values are valid at the 95% level. It is reiterated that these are minimum possible values and actual increases may be substantially higher. The sample MDRs range between 1.21 and 1.23 implying about 22% more precipitation for the seeded EUs than for the nonseeded control EUs for the entire available sample population of98 EUs with no partitioning. Partitioning Applying statistical tests to EU groups which meet one or more meteorological criteria can give insight to the credence of physical hypotheses. This is termed "partitioning" or "stratification." The most obvious and physically reasonable partitioning of the available EUs was by HAS wind direction. No real seeding effect can exist if the wind did not transport the seeding plume over the target gauges. Considerable prior plume tracing with tracer gas and AgI was conducted with the same HAS seeding site, primary intended target, TAR, and its associated gauge, GTR. These earlier data sets were analyzed in detail for this experiment. They showed that plumes released at the HAS when winds there were from about 170 to 270 deg routinely reached the GTR/T AR area with typical 2 km across-the-wind plume widths by that downwind distance. Plumes released when HAS winds had any northerly component did not target that area but were transported further south. The obvious wind partition was for HAS median wind directions of 270 d€:g true or less, as the lowest EU value was 169 deg, This wind direction partition provided a total of 32 pairs, and a total of35 seeded plus 34 nonseeded or 69 EUs. The target-control correlations of Table: 2 are almost as high as those of Table l. Strongly suggestive P-values of 0.03 and 0,05 are shown in Table 2 for gauges GSO and GTR, respectively. Downwind gauge GDN had a respectable 0.06 and 7 ._J~""L.