STOCHASTIC MODELING OF TRANSECT-TO-TRANSECT PROPERTIES OF GREAT-BASINRANGELAND STREAMS

We developed a transect-to-transect stochastic model of stream width a nd pools. The model included Markov chains for transect-to-transect tr ansitions among pools, nonpools, and part pools, a two-parameter beta distribution for pool fraction when the transect is a partial pool, an d a two-parameter gamma distribution to model standardized widths. The model accurately reproduced the mean and standard deviation of both w idth and transect-based measures of pool area. However, the model smoo thed the distributions thereby reducing the skewness of the observed d ata. The physical tendency for streams to have very few transects that are almost! but not totally, all pool or nonpool implies that stochas tic models based on continuous distributions will have more density in the tails than observed on the stream. The model represented odd-shap ed empiric distributions of pool area which indicates it is very flexi ble and can accurately model a variety of streams.