ESRL/PSD Seminar Series

Abstract

From an atmospheric modelers' point of view, the land/water surface may be treated as passive bottom boundary. The conventional land/water surface model parameterizes feedbacks (heat, momentum, and water exchange) between the bottom surface and the atmosphere, given meteorological data and surface properties and conditions. A modern land/water surface model, on the other hand, can simulate surface conditions and feedbacks at the same time, and therefore allows air-land/water interaction. Take Simple Biosphere Model (SiB) as an example. SiB can be driven by precipitation, wind, longwave and shortwave radiative flux, air temperature, and air humidity. The land/water surface conditions, such as soil moisture, canopy temperature, etc., then evolve with time, which are used to calculate feedbacks to the atmosphere. A unique feature of SiB is that it also produces CO2 flux in addition to the conventional feedbacks. The CO2 flux, CO2 concentrations, and other conventional feedbacks in coupled model simulations can be evaluated against observations to confirm model validity. However, from stomatal conductance on a leaf to the simulation of a "CO2 front" in the conterminous U.S., several assumptions exist in the upscaling process. This incurs difficulty in model diagnosis for CO2 budget calculation, while similar difficulty may apply to other physical processes.

Similar to the modern land/water surface model, an Ocean General Circulation Model (OGCM) can be driven by meteorological data to provide bottom boundary conditions as well as feedbacks to the atmosphere. A tropical cyclone case in the Bay of Bengal will be discussed for its impact on surface heat flux and downward ocean heat pumping in the Hybrid Coordinate Ocean Model (HYCOM) simulations. The importance of reconstruction of wind forcing that drives HYCOM will be emphasized. Different from the land/water surface model, any OGCM contains a dynamic core and mixing scheme due to its fluid nature. The point observations in the ocean thus have larger effective range. As a result, the ocean model evaluations can rely on the check of both the feedbacks to the atmosphere and the ocean prognostic variables. The model diagnosis, however, must consider various dynamic and physical processes simultaneously.