UCLA

Sea surface temperature (SST) variability has been shown to have predictive value for land precipitation, although SSTs are unable to fully predict intraseasonal to interannual hydrologic extremes. The possible remote effects of large-scale land surface temperature (LST) and subsurface temperature (SUBT) anomalies in geographical areas upstream and closer to the areas of drought/flood have largely been ignored. Here evidence from climate observations and model simulations addresses these effects. Evaluation of observational data using Maximum Covariance Analysis identifies significant correlations between springtime 2-m air temperature (T2 m) cold (warm) anomalies in both the western U.S. and the Tibetan Plateau and downstream drought (flood) events in late spring/summer. To support these observational findings, climate models are used in sensitivity studies, in which initial LST/SUBT anomaly is imposed to produce observed T2 m anomaly, to demonstrate a causal relationship for two important cases: between spring warm T2 m/LST/SUBT anomalies in western U.S. and the extraordinary 2015 flood in Southern Great Plains and adjacent regions and between spring cold T2 m/LST/SUBT anomalies in the Tibetan Plateau and the severe 2003 drought south of the Yangtze River region. The LST/SUBT downstream effects in North America are associated with a large-scale atmospheric stationary wave extending eastward from the LST/SUBT anomaly region. The effects of SST in these cases are also tested and compared with the LST/SUBT effects. These results suggest that consideration of LST/SUBT anomalies has the potential to add value to intraseasonal prediction of dry and wet conditions, especially extreme drought/flood events. The results suggest the importance of developing land data and models capable of preserving observed soil memory.