| Daily weather data (total solar radiation, rainfall, maximum and minimum temperatures, wind run, and atmospheric humidity) are needed for a wide range of applications including those related to agriculture and hydrology. The evaluation of long-term effects of given management strategies requires long records which are seldom available in many locations. Thus, methods to generate daily weather data with similar statistical characteristics as actual data are desirable.; One objective of this study was to modify and evaluate a weather generator, CLIMGEN, and extend its capabilities to include estimating solar radiation and vapor pressure deficit. CLIMGEN requires daily weather data for input, but can also work from monthly summaries and estimate parameters, if missing, from regression equations. It was found that generated weather data compared well with actual data at different locations worldwide. For scenarios where data was assumed limited and input parameters were estimated from regression equations, results were less satisfactory but adequate for most applications.; The second objective was to evaluate the adequacy of generated data in estimating evapotranspiration using both the Penman-Monteith and Priestley-Taylor models. Daily maximum and minimum temperature were the only data assumed to be available, and daily wind speed data was assumed equal to the monthly mean. ET estimates from generated data closely approximated those from actual data, with slightly better agreement for the Priestley-Taylor than the Penman-Monteith model.; A third objective was to compare evapotranspiration determined using the Priestley-Taylor ET model, adjusted for humidity effects using generated and actual data, against the Penman-Monteith model. Estimates obtained at sites with locally-calibrated parameters and actual humidity data closely approximated those from the Penman-Monteith model. The performance of the Priestley-Taylor model using generated data was less satisfactory but acceptable for many applications.; Overall, weather data generation appears feasible and reasonably reliable procedure, useful for long-term analysis of weather-driven processes at locations with limited weather data. |
