Imprints of Soil Moisture Memory on Hydrological Predictability in SubX Subseasonal Forecasts
Abstract
Analyses of both observational and simulated soil moisture data identify 1-2 months as a typical time scale for the memory associated with root zone soil moisture anomalies. Thus, if the land surface at some location in a subseasonal-to-seasonal (S2S) forecast system is initialized with a specific soil moisture anomaly, some aspect of that anomaly will still typically be present, say, 3-4 weeks into the forecast and can thereby influence evapotranspiration, surface runoff, and other hydrological processes at such leads. This contribution of soil moisture initialization to subseasonal hydrological predictability, however, varies significantly in space and time. Through in depth analysis of the subseasonal forecasts produced for the SubX project by a state-of-the-art S2S forecast system (the NASA GMAO S2S forecast system), and working from ideas provided by Seneviratne and Koster (2011), we demonstrate that the global distribution of the predictability of soil moisture is strongly tied to the ratio Var(W)/[Var(W)+Var(P)], where Var(W) is the variance (across years) of the total water in the soil column at the start of the forecast and Var(P) is the variance of the precipitation volume falling in the first 15 days of the forecast, i.e., prior to the subseasonal averaging period. This ratio is indeed found to explain most of the variance in soil moisture predictability across the globe in any given month as well as the seasonality in soil moisture predictability at most locations. The SubX data also indicate that evapotranspiration (ET) predictability (and, by extension, air temperature predictability) is strongly connected to this ratio as modified by independently established connections between soil moisture and ET. Finally, the initial soil moisture state for a forecast is shown to have a curious impact on predictability estimates: drier initial soils promote a higher predictability for soil moisture, whereas wetter initial soils promote a higher predictability for ET (and, again by extension, for air temperature).
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMA226.0004K
- Keywords:
-
- 3322 Land/atmosphere interactions;
- ATMOSPHERIC PROCESSES;
- 3339 Ocean/atmosphere interactions;
- ATMOSPHERIC PROCESSES;
- 3362 Stratosphere/troposphere interactions;
- ATMOSPHERIC PROCESSES;
- 3373 Tropical dynamics;
- ATMOSPHERIC PROCESSES