Evaluation of Multi-Model Simulated Products in NLDAS-2

Currently, NLDAS is a quasi-operational system that supports U.S. operational drought monitoring and seasonal hydraulic prediction, in particular for the National Integrated Information System including U.S. Drought Monitor (USDM) and NCEP Climate Prediction Center Monthly Drought Briefing. Detailed information about NLDAS can be found at NOAA (http://www.emc.ncep.noaa.gov/mmb/nldas) and NASA (http://ldas.gsfc.nasa.gov/nldas/) websites. NLDAS consists of four stages. The first stage is to establish NLDAS infrastructure (2000-2005) and test performance of the models, the second stage is to upgrade models and surface forcing to generate long-term NLDAS products (2006-2009), the third stage is to evaluate quality and reliability of these long-term products using as many as available in situ observations and satellite-retrieved data (2009-2011), and the fourth stage is to apply these products to U.S drought analysis and monitor (2012-beyond).

This presentation gives an overview of our overall evaluation results performed during the last three years including evaluation tools, in-situ observations and satellite-retrieved data used in NLDAS-2, and performance of different models. The evaluated products include streamflow/total runoff, evapotranspiration, sensible and latent heat flux, ground heat flux, soil moisture, soil temperature, and land surface skin temperature. These evaluations cover different spatial scales vary from basin to continental scale and time scales varying from hourly to yearly time scale. After we summarize our evaluation results, we also show some preliminary results from recent efforts to further improve individual models and suggest some possible directions to improve different NLDAS-2 land surface models in future.

NLDAS has become mature enough for NCEP operational implementation (planned for the near future). At the same time, we recognize that the current NLDAS is not an “actual” land data assimilation system because remotely-sensed estimates of land-surface states such as soil moisture and snowpack, and in-situ observations such as streamflow and soil moisture, are not yet assimilated into current version of NLDAS. The NCEP/EMC NLDAS team is collaborating with the NASA Goddard Hydrological Sciences Laboratory to add their Land Information System (LIS) to the current NLDAS system which would allow assimilation of remotely-sensed data and in-situ observations, e.g. via an ensemble Kalman filter approach.