The potential predictability of the effects of warm season soil moisture anomalies over the central United States has been investigated in a series of general circulation model (GCM) experiments with the National Center for Atmospheric Research (NCAR) Community Climate Model version 3 Land Surface Model (CCM3/LSM). Three different types of experiments have been developed, all starting in either March (representing precursor conditions) or June (conditions at the onset of the warm season): (1) anomaly runs with large, exaggerated initial soil moisture reductions, aimed at evaluating the physical mechanisms by which soil moisture can affect the atmosphere; (2) predictability runs aimed at evaluating whether typical soil moisture initial anomalies (indicative of year-to-year variability) can have a significant effect, and if so, for how long; and (3) threshold runs aimed at evaluating if a soil moisture anomaly must be of a specific size (i.e., a threshold crossed) before a significant impact on the atmosphere is seen. The anomaly runs show a large, long-lasting response in soil moisture as well as surface temperature, sea level pressure, and precipitation; effects persist for at least a year. The predictability runs, on the other hand, show very little impact of the initial soil moisture anomalies on the subsequent evolution of soil moisture and other atmospheric parameters; internal variability is most important, with the initial state of the atmosphere (representing remote effects such as SST anomalies) playing a more secondary role on seasonal and shorter timescales. The threshold runs, devised to help resolve the dichotomy in anomaly and predictability results, suggest that, at least in CCM3/LSM, the vertical profile of soil moisture is the most important factor and that deep soil zone anomalies exert a more powerful, long-lasting effect than do anomalies in the near-surface soil zone. We therefore suggest that soil moisture feedbacks may be more important in explaining interannual to decadal droughts evident in the historic and recent prehistoric records, but less important on monthly to seasonal timescales.