Time scales of layered soil moisture memory in the context of land-atmosphere interaction

The time scales of layered soil moisture memory in the Common Land Model ( CLM) coupled with the National Center for Atmospheric Research Community Climate Model, version 3 (NCAR CCM3) have been examined using a 50-yr climate simulation. Such soil moisture memory has been characterized in terms of the spatial, seasonal, and vertical variations of 1-month-lag autocorrelation coefficients and the corresponding e-folding decay time scales. To understand this land memory mechanism, in terms of the variations that occur in the model, a cross-spectral analysis has been applied to the soil moisture profile with precipitation ( P), runoff ( R), evapotranspiration ( ET), transpiration, and the residual of P 2 ET 2 R, respectively, together with an examination of the surface water budget of the annual cycle. These collectively provide physical insights on time scales of layered soil moisture memory in the context of land - atmosphere interaction. The major findings are: 1) soil moisture memory in warm climates can be at least several times longer for drier conditions than when it is sufficiently rainy; and 2) under wet conditions the time scales of soil moisture appear to be controlled by temperature-dependent climatic demand; but for drier conditions they appear to depend largely on increasing time scales for the coupling of soil moisture to ET and especially runoff.