On the Transition of Land Modeling and Data Assimilation to Operations

Land is one of the key components of the global and regional forecasting systems for weather, water, and climate prediction. In the past two decades, my group has developed and transitioned new model parameterizations and value-added data to operational centers (NCEP, ECMWF) and community models (CESM, WRF). Here I will overview our recent efforts in transitioning land modeling and data assimilation to operations.

First, we developed a consistent treatment of land surface roughness length in the Noah land model (Zeng et al. 2012). Its implementation in the NCEP operational GFS model (Zheng et al. 2012) substantially reduces the daytime cold bias in forecasted surface skin temperature and 2-m air temperature. The reduced model bias also substantially increases the amount of assimilated satellite data from surface-sensitive infrared and microwave channels.

Second, we have developed the daily 4 km snow water equivalent (SWE) and snow depth data from 1982-2016 over conterminous US based on in situ snowpack measurements and gridded precipitation and 2-m temperature data (Broxton et al. 2016a; Dawson et al. 2017; Zeng et al. 2018). Using this dataset, SWE from the initialization of NCEP forecasting models (GFS and CFS; Dawson et al. 2016) and reanalysis (CFSR; Broxton et al. 2016b) and from the GFS and CFS forecasts (Broxton et al. 2017) was found to be deficient. We also found that SWE initial states affect the subseasonal to seasonal prediction in CFS (Broxton et al. 2017). We are currently working with NCEP scientists to improve the snow modeling and data assimilation for subseasonal to seasonal prediction and drought monitoring, and the new results will also be included in the presentation.