Author(s): J. M. Segura; V. Yevjevich
Linked Author(s):
Keywords: No Keywords
Abstract: A stochastic method of modeling the soil moisture profile at n points is presented starting with the precipitation tim e series. This multiv ariate time process is first standardized by using the periodic mean and periodic standard dev iation (for simplicity, that stan dard deviation here is assum ed a constant). So only a random dependent process is m odeled in detail. The developed method is then calibrated on a real case, w ith the observed soil moisture at 14 points alon g a depth of 4.20 m. The modeled and the observ ed soil moisture profiles are relatively very close in this case of using 36 model parameters for four depth points and a total of 648 pieces of data. The model was then used to simulate 300 years of pro files of soil moisture content. By neglect ing the water uptakes by plan t roots and by using the Darcy's equation, the rec harge to groundwater was computed. Th e probab ility distribution of the recharge is the Jshaped gamma distribution with a high skewness. The groundwater recharge resulted as an intermittent, non-linear, non-normal process with few rare large even ts of precipitation making most of the replenish-ment of underground reservoirs. Therefore, t his method consists of stochastic modeling of soil moisture profiles and a determ inistic transform of the soil moisture in to the groundwater recharge.
Year: 1985