Raingage Network Requirements from a Simulated Convective Complex Weather Modification Experiment

Abstract

A convective complex weather modification experiment was simulated using Monte Carlo techniques. The purpose was to estimate the optimum raingage density for evaluation of a possible future experiment. The data base consisted of radar volume scans made within 150 km of Miles City, Montana, during May?July 1977. A total of 103 convective complexes were identified and tracked from radar data. Raingage networks of various densities were simulated under the lowest-tilt radar scans to estimate total rainfall accumulation from each complex. Randomly chosen rainfall amounts were increased by given percentages to simulate assumed seeding treatments. A Monte Carlo scheme yielded estimates of the number of experimental units required for various combinations of α- and ?-probability levels, treatment effects (percentage of increases) and raingage densities. Applying these results to the numbers of operationally available convective complexes as a function of area gave estimates of the optimal spacing and seasons required to detect a treatment. The results suggest that an unacceptably long field experiment would be necessary to detect treatment effects of 50% or less without some stratification of precipitation data.