Cooperative Institute for Research in Environmental Sciences

Snow and Ice Enhancements to the Land-Surface model used in WRF-based Rapid Refresh

Tatiana G. Smirnova (1,2), John M. Brown (2), Stan Benjamin (2)

(1) CIRES, (2) NOAA ESRL

The Land Surface Model (LSM) described in this presentation was originally developed as part of the Rapid Update Cycle (RUC) model development effort at what is now the Earth System Research Laboratory (ESRL). The goal of this work was to provide more accurate lower boundary conditions for the Rapid Update Cycle (RUC). This LSM (henceforth, RUC LSM) has been operational as part of the RUC since 1998. The RUC LSM describes complicated atmosphere/land surface interactions in a simplified fashion so as to avoid excessive sensitivity in the RUC system to multiple poorly defined surface parameters. The simple parameterizations in RUC LSM have proved to be physically robust and capable of realistically reproducing evolution of soil moisture, soil temperature and snow in the cycled model, where surface fields are initialized from the previous 1-h model forecast. Our RUC LSM development has adapted some ideas from existing land surface schemes, but also conceived of some unique features which in our opinion can give more accurate representation of physical processes.

Monitoring of model performance over the years has motivated several modifications to the RUC LSM leading to better performance. Beginning in 2002, the RUC LSM became available to the WRF community as part of the yearly WRF code releases, and is implemented in the WRF-based Rapid Refresh (RAP), which will replace RUC at NCEP in March of 2012. With the much larger RAP domain covering the whole North American continent as well as large ocean areas in high latitudes, the RUC LSM needed further validation and development for application in the tundra permafrost regions and over sea ice, including heat conduction in the sea ice, and snow accumulation and ablation on top of the sea ice. The modifications to the RUC LSM were extensively tested off-line in one-dimensional framework using data from SnowMIP2 (Snow Model Intercomparison Project for forest snow processes) experiment, then were monitored and verified in the coupled developmental version of RAP, and finally were implemented in the operational RAP at NCEP.
These recent modifications to ice and snow treatment in RUC LSM and results from one-dimensional and coupled RAP validations are described in this poster.