A Comparison Study of Surface Water Cycle in Land Data Assimilation Systems Using Different Noah Land Model and Different Treatments for Surface Runoff and Ground Water

Skillful forecast at sub-seasonal to seasonal time scales rely heavily on well-simulated atmospheric response to slowly varying boundary forcings from both ocean and land surface. The critical importance has been recognized. Ocean indices have been used in a variety of climate predictions. In contrast, land anomalies, especially soil moisture and associated surface runoff parameterization and ground water treatment in land models, have been notably difficult to demonstrate. The surface physical runoff parameterization has direct impact on how much water can penetrate to the soil as soil water, and ground water treatment involves water table dynamics and deep soil moisture, which are important to soil moisture memory, which has been identified as one source of predictability.

The Noah Land Surface Model (LSM) is the land component of NCEP Coupled Forecast System (CFS) used for seasonal predictions. The Noah LSM treats the surface runoff as a sum of surface overflow resulting from infiltration excess and field capacity, and the subsurface runoff as free drainage where the water table depth is assumed to be well within the soil layers. To overcome this shortcoming, a new Noah-based land model with Multiple Parameterization options (Noah-MP) was developed over the past. The Noah-MP allows for options for different physical parameterizations, especially options for surface runoff and ground water treatment. In this study, surface water cycle components from Noah MP offline simulations with different choices for runoff and ground water parametrizations covering a 31-year period are compared to each other. The comparisons are also carried out with those obtained from the Global Land Data Assimilation System (GLDAS) in NCEP CFS Reanalysis and Reforecast (CFSRR), where the traditional Noah LSM is used. The results will be discussed.