NONLINEAR DYNAMICS OF SOIL-MOISTURE AT CLIMATE SCALES .1. STOCHASTIC-ANALYSIS

Hydrologic and climatic time series are often characterized by smaller fluctuations around persistent anomalies. Occasional transitions occur between these prolonged runs of above or below average conditions. A statistical-dynamical model is developed to represent the surface hydrology of large continental regions. The model has physically based components, and it includes a parameterization for the local recycling of precipitation. It is shown that stochastic fluctuations in this equation may give rise to separate and distinct statistical modes with noise-induced transitions between these modes. In an example situation the characteristic long-term probability distribution of soil moisture states is shown to be bimodal. The implications of this result are important for the interpreting and the statistical modeling of hydrologic records.