On the Utility of HYSPLIT Trajectories Driven by Operational NWP Analyses and Forecasts for Evaluating and Forecasting Cloud Seeding Plume Pathways and Targeting Effectiveness

Trace chemical analysis, while a useful technique, has some inherent limitations in its utility, either alone or in conjunction with high-resolution precipitation gauge analysis, as a means of evaluating the potential effectiveness of a cloud seeding program. In addition to saltation and snow drift effects, as well as concerns about the representativeness of the snow sampling conducted at individual sites with respect to a larger target area, settling, melting and refreezing processes all complicate the interpretation of a trace chemical analysis of a snow column.

For the past five seasons of the Wyoming Weather Modification Pilot Project (WWMPP), the Desert Research Institute conducted sampling and trace chemical analysis of snow in the Medicine Bow, Sierra Madre and Wind River mountain ranges in Wyoming, as well as for a control site in Colorado's Park Range. The goal of the sampling in the Medicine Bow and the Sierra Madre ranges was to provide physic-chemical evidence that would specifically support the WWMPP Randomized Statistical Experiment (RSE) as well as the companion AgI Seeding Cloud Impact Investigation Study (ASCII) that was a two-year study led by the University of Wyoming. Results, presented in a companion paper, indicate that in some instances the various processes described in the preceding paragraph may be in play. However, there is also the possibility that wind and stability conditions in the vicinity of the cloud seeding generators and the target areas are such that seeding material was not transported into the clouds that provided the precipitation impacting the target areas. In other words, targeting was not optimal.

One means of addressing this latter concern is to perform trajectory and dispersion analyses driven by model analyses for which one has significant confidence in their fidelity. While the use of high-fidelity, high-resolution models is optimal, not all cloud seeding programs will have the ability to run or access to runs of such high-resolution models. Thus, it is worth examining whether more widely available models and analysis approaches have utility for diagnosing transport of seeding material, whether in a post-event retrospective analysis or a pre-event forecast.

In this presentation we will discuss recent work in this regard utilizing the NOAA HYSPLIT trajectory/dispersion modeling system driven by 13 km NAM (WRF-NMM) analyses for specific Randomized Statistical Experiment (RSE) cases during the last five years of the WWMPP. From those case studies, we will make inferences as to the degree to which specific environmental conditions may have led to suboptimal targeting, and to the degree for which model/HYSPLIT runs prior to the event might have been useful to refine the targeting of the seeding operations.