A more stable and extensive analysis of climate is necessary to simulate long-term impacts associated with climate change. The exponential dispersion model (EDM) family of distributions, a popular choice when characterizing precipitation levels and temperature in different climate environments, is being considered for its applicability to near-surface disposal performance assessments. In this study, the EDM family is examined to determine if there is a preferred distributional form within the family for these parameters using data from two sites whose climate environments are quite different. One site is in a semiarid environment and the other is in a humid environment. In addition, the merits of selecting a different distributional form to represent each calendar month of precipitation and temperature data are explored.

Results show that the gamma distribution was most often determined to be the best fit to recorded precipitation data. When considering temperature, however, the Weibull distribution proved to be a better fit. These results suggest that greater precision may be possible when temperature and precipitation serve as inputs to modeling activities, if these parameters are allowed to be represented by different distributions and derived by calendar month. Ultimately, the approach provides a more far-reaching examination of historical records and provides an increase in confidence, when used in the evaluation of long-term climate impacts associated with near-surface disposal facilities. © 2013 Wiley Periodicals, Inc.