Using a 20 year historical sample of climatological data, significant to evaporation, a stochastic model was developed using a first order markov process. The model generated new 20 year sets of climatological data for deterministically estimating the potential evaporation from an oil shale retort water ree surface, using data collected at Laramie, Wyoming, and the Kohler-Nordenson-Fox equation which was shown to give a good estimate of the evaporation of retort waters. The model used climatological data from the area of Laramie, Wyoming, which was then used to generate 1000 new 20 year sets of data. Mass balance techniques were then used to evaluate the new data sets and determine the maximum pond depth and the allowable input in volume per unit area for each of the 20 year sets of data. The maximum pond depth for all sets and the lowest total allowable input were considered to be the two key design paramters in the design of evaporation ponds for oil shale retort waters.