Intercomparison of Soil Hydrology Modules in the MIT

The availability of water at or near the surface determines the way incident radiative energy is partitioned at the ground surface. The goal of this thesis is to determine if better hydrological representation in the MIT Integrated Global System Model will improve its climate prediction capability. This thesis compares the performances of the hydrological modules in the MIT Climate Model and the Natural Emissions Model (NEM) with the off-line National Center for Atmospheric Research (NCAR) Land Surface Model (LSM version 1.0) for Ecological, Hydrological, and Atmospheric Studies. The models are forced with the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) data, and outputs are validated using FIFE Intensive Field Campaigns measurements. Validation and analyses include comparisons between daily and diurnal model outputs and FIFE measurements, and evaluations of diurnal root mean square errors (RMSE). All three models simulate FIFE conditions well. The NEM is particularly good at tracing the diurnal trend of most diagnostic parameters; but the large and numerous fluctuations in this model's outputs result in large diurnal RMSEs as well. Many of the errors in this model are due to deficient representation of soil moisture movement in its shallow soil column. The deep lower soil layer in the hydrological module in the Climate Model over-drains the thin upper soil layer; the dryness of the upper layer adversely affects energy partition at the land-atmosphere boundary. The NCAR LSM avoids many of the problems encountered with the other two modules and simulates FIFE conditions best; the doubled computational requirement is its main drawback. Hence, comprehensive hydrological representation in climate models will improve climate prediction capacity by providing consistent and more accurate hydrological inputs to all submodels. Thesis Supervisor: Dara Entekhabi Title: Associate Professor