Land-atmosphere interactions in numerical weather and climate prediction: Perspectives from coupled land-atmosphere data assimilation

Abstract:

The role of land-atmosphere interactions in weather and climate prediction has been well recognized. However, accurate representation of land-atmosphere interactions in the numerical modeling systems remains a challenging problem. Despite significant development and improvements in land surface models in the past two decades, how to effectively couple the land surface model with an atmospheric model is still under development. The related issues depend not only on the sophisticated land surface model itself but also on the accurate specification of soil states and land surface characteristics (such as land use and land cover, snowpack, etc.) as well as water and energy fluxes in the land-atmosphere interface. Meanwhile, due to deficiencies in representing land-atmosphere interactions in the numerical models, uncertainties in near-surface atmospheric and boundary layer conditions are significant in the atmospheric models. So far, most of the previous studies have emphasized the influence of the land surface model and soil moisture on weather and climate prediction, while there is a lack of attention to the atmospheric boundary layer due to limited observations. The impact of atmospheric states (except for precipitation), especially these near-surface and boundary layer atmospheric conditions on soil states, has not received much attention. Modern remotely sensed observations bring significant opportunities for studying land-atmosphere interactions with data assimilation. However, so far, most of the coupled land-atmosphere data assimilation systems use a weakly coupled method, in which land and atmosphere data assimilations were conducted separately. With the recent development of a strongly coupled land-atmosphere data assimilation, cross-covariances between land and atmospheric variables enable the land and atmospheric variables to be corrected simultaneously, resulting in effective land atmosphere coupling and improved both soil and atmospheric states. In this invited presentation, the author will present recent results of strongly coupled land atmosphere data assimilation with discussion on questions regarding 1) the coupled data assimilation as a tool for studying the land-atmosphere interaction, 2) the influence of coupled data assimilation in improving numerical simulation and forecasts with better representation of soil states, near-surface air conditions, and atmospheric boundary layer structures, and 3) the impact of enhanced soil states, surface water and energy fluxes, and near-surface atmospheric conditions on the accurate weather prediction. The challenges, progress, and direction in studying land atmosphere interactions with numerical modeling and data assimilation will also be discussed.