Land Significantly Contributes to East Asian Summer Monsoon Rainfall Predictability
The East Asian summer monsoon (EASM) rainfall provides water for over 20% of the world’s population. Researchers significantly improved simulations and hindcasts of EASM rainfall at an interannual timescale by integrating a new, weakly coupled data assimilation system that uses observational data to constrain simulated soil moisture and temperature into a coupled climate model. This allowed them to capture the notable shift to a "wetter-South-drier-North" rainfall pattern in China in the early 1990s. They attribute the improvements to the strong land-atmosphere coupling over large portions of China, which allows soil moisture and temperature to influence precipitation and monsoon circulation.
Previous strategies that exploit ocean memory for Earth system prediction at an interannual timescale have met with little success improving interannual predictions of EASM rainfall. This study highlights the significant contribution of land to these predictions. Integrating land contributions can increase their accuracy and benefit the substantial populations influenced by the large variability of summer monsoon rain that can lead to floods and droughts. It demonstrates the need to systematically evaluate the contributions of land versus the ocean for interannual climate prediction worldwide.
EASM rainfall exhibits large seasonal-to-interannual variability. Land processes influence surface thermal conditions and can provide an important source of predictability for EASM rainfall. Researchers used a recently developed, weakly coupled data assimilation system in which land observations constrain soil moisture and temperature in a coupled climate model to evaluate the influence of land on EASM predictability for the first time. Numerical experiments performed using the coupled climate model with and without land data assimilation showed that land data assimilation improves the modeling of soil moisture and temperature, which have important effects on the surface water and energy balances. These effects can propagate in a coupled model featuring a weakly coupled data assimilation system to the atmosphere through land-atmosphere interactions, which are generally strong in the monsoon region and subsequently impact the large-scale precipitation pattern. The researchers performed hindcast experiments, a method of validating a mathematical model, that used initial states to predict events that already occurred. They initialized hindcast experiments with the well-balanced states determined by a coupled simulation with the weakly coupled data assimilation system, which showed significantly improved rainfall prediction skill over East China and the Tibetan Plateau.