The initiation of a hydrologic drought may depend on large-scale or teleconnective causes; however, local positive feedbacks in the land-atmosphere system are believed to contribute to the observed persistence and intensification of droughts. In this study a basic linearization technique is combined with a nonlinear stochastic model of land-atmosphere interaction to analyze and, more importantly, quantify feedback mechanisms that arise in the coupled water and energy balances at the land surface. The model describes land-atmosphere interaction by four coupled stochastic ordinary differential equations in soil moisture, soil temperature, mixed-layer humidity, and mixed-layer potential temperature. The solution is a physically consistent joint probability distribution. The steady and perturbation-induced parts of the model equations are decomposed into the dependence of each component physical process upon each model state. Because of the negative correlation between soil moisture and soil temperature, the physical mechanisms that serve to restore each state individually (largely soil moisture control of evaporation and temperature dependence of saturation specific humidity) act as significant anomaly-reinforcing mechanisms for the other state. ¿ American Geophysical Union 1996