The complexity of using a feedback parameter to quantify the soil moisture‐precipitation relationship

The long time scale of soil moisture dynamics makes it a potentially important predictor in seasonal to interannual precipitation predictions. However, where and to what extent soil moisture can influence precipitation is still a topic of debate. A feedback parameter was applied by several studies to quantify both the magnitude and direction of soil moisture's impacts on precipitation based on both observational data and model output. Its effectiveness as an indicator of soil moisture impact on precipitation was questioned in recent studies. Using the NCAR CAM3‐CLM3 as a modeling framework, this study examines whether the feedback parameter can correctly identify regions of strong soil moisture impact on precipitation as compared with other more physically based numerical experimental designs. It is found that the effectiveness of feedback parameter as an indicator for soil moisture impact on precipitation is significantly reduced by the presence of low‐frequency precipitation variability. Such low‐frequency variability (which may be caused by large scale oceanic forcing) propagates to the land surface, causing similar variability in soil moisture, and the presence of the low‐frequency signal in both variables leads to spuriously large feedback parameter values. Our results suggest that the feedback parameter is not a reliable indicator for the impact of soil moisture on precipitation.