ABSTRACT

Soil moisture has two properties that give it a special role
in the climate system: (i) its variations lead to variations in
evaporation, which in turn can induce variations in (for example)
precipitation and air temperature, and (ii) soil moisture anomalies have
an intrinsic "memory" spanning weeks to months. Given these properties,
initializing an Earth system model with realistic estimates of soil
moisture may yield some skill in precipitation and air temperature
forecasts one month out. This talk examines this potential, addressing
in addition the form of the soil moisture - evaporation relationship --
a nonlinear functional relationship that imposes strict limits on where
such skill can be realized and indeed imposes unique fingerprints of
soil moisture variability on the meteorological record.