Enhancing Land Surface Modeling: Upgrades for NOAA Unified Forecast System Global Forecast Applications

NOAA is collaborating with the US weather and climate science community to develop the next generation fully coupled Unified Forecast System (UFS) for research and operational applications. In response to the biases identified in earlier global model prototypes of the UFS , especially 2-meter temperature cold biases, we have implemented several land surface upgrades to the Noah-MP land surface model for the upcoming prototypes following extensive testing and evaluation. The key modifications include: removing the vegetation type-dependency from the diagnostic option (opt_diag); using radiative temperature as the skin temperature; adopting a single loop alternative approach to address flux inconsistencies below and above the canopy; improving thermal roughness treatment on the bare soil portion of the grid; adjusting momentum roughness length; interpolating temperature between skin temperature and the first soil temperature to set the initial snow temperature based on snow depth; using a constant snow thermal conductivity to reduce snow insulation effects; fixing radiation parameter bugs for grassland/cropland; and adjusting albedo parameters for each soil color category.

A hierarchical model development approach was used to identify processes or code elements to address the biases. Some of these upgrades were initially tested via offline land-only runs and compared with observations to ensure accuracy. These land-only simulations are performed for 10 years to test the impact of proposed changes on the model climatology of surface fluxes and soil/snow states. The final evaluations were conducted using a subset of the full prototype framework with 11 summer cases and 10 winter cases to assess the effectiveness of each modification. The results demonstrated significant improvements in mitigating cold biases and enhancing the accuracy of land surface modeling.