|Johnson greg l,|
|Ballard edward b,|
Submitted to: Journal of Applied Meteorology
Publication Type: Peer reviewed journal
Publication Acceptance Date: 3/11/1996
Publication Date: N/A
Citation: Interpretive Summary: Where climatic records are either not available or are complete for only short periods of time there is often a need to synthetically generate sequences of weather using computer programs which mimic the true climate of a given location. Two such climate simulation models, commonly called "weather generators," were compared for their performance in matching true, historically observed climate at six widely-dispersed locations in the continental United States. Model performance varied, depending on the climatic variable (precipitation, temperature or solar radiation) and the statistics of interest. In general, the USCLIMATE model was found superior in replicating climatic variability and in matching day-to-day changes in temperature and solar radiation, especially during the spring and fall transition seasons. Both models replicated precipitation amounts and durations quite well.
Technical Abstract: Two stochastic weather simulation models were compared for their performance in replicating observed precipitation, temperature and solar radiation variables at six locations in the United States. Statistical tests of significance were performed on a variety of standard and derived weather variables, including means, standard deviations, extremes and ranges, over monthly and annual time periods. Differences and similarities in the performance of the models were noted. Model selection was determined to be a function of the intended application, with neither model found to perform adequately in all areas of testing.