The Comparison of Summer-seasonal Surface Water Cycles in the NCEP Coupled Forecast System (CFS) Using Noah and Noah Multiple Parameterization (Noah-MP) Land Surface Models

Surface water processes are integrated parts of the global water cycle. On an annual basis, the partitioning of precipitation between evaporation and runoff at the surface has important implications for climate change analyses and climate predictions as it directly impacts on seasonal variations in stream flow and soil moisture, whose anomaly is considered as one of the important sources for climate forecasts. The representation of these processes at climate time scales heavily relies on the parameterization schemes used in climate models.

The Noah land surface model (LSM) is the model used in the NCEP's Climate Forecast System (CFS) for climate predictions. It originates from the Oregon State University LSM, but it is equipped with many physical improvements and updates added by researchers. The main features relevant to this study is that the evaporation control in Noah LSM is based on the Penman-Monteith equation with a combined surface layer of vegetation canopy and ground, which generally leads to higher canopy resistance. The Noah LSM is configured with four shallow soil layers with a fixed depth of 2 meters and free drainage at the bottom soil layer, where the ground water effects are neglected. To overcome the limitations of the Noah LSM, the Noah model with Multiple Parameterization (Noah-MP) options was developed by the great Noah community. The community Noah-MP LSM contains a separate vegetation canopy defined by physical and radiometric properties. In addition, an unconfined aquifer is attached to the bottom of the soil to allow the water table move freely up and down. These new parameterizations are are expected to have a great impact on surface water climate and subsequent climate prediction skills. To examine the impact from using the new Noah-MP options on CFS water cycle and prediction skills, summer-seasonal T126 CFS reforecast experiments are carried out using different parameterization options for selected years. Preliminary results will be presented and brief discussions will be followed.