Motivation:
Soil moisture, snow, and land surface temperature all strongly impact the water and energy fluxes from the land surface to the atmosphere, and hence land-atmosphere coupling and atmospheric predictability. NASA EOS products provide global observations of surface soil moisture, snow water equivalent, snow cover, and land surface temperature based on data from AMSR-E and MODIS. EOS products alone, however, are not sufficient for forecast initialization because of gaps in spatial and temporal coverage and because key model variables cannot be observed from space (e.g. root zone soil moisture). Similarly, NASA EOS land surface products alone are not sufficient as background inputs for other EOS retrieval algorithms (e.g. CERES products). If used in a land data assimilation system, NASA EOS land surface products have tremendous potential for improvements in the characterization of the land surface water, energy, and radiation budget, the prediction of weather and short-term climate, and the background fields for other EOS algorithms.
Hypotheses:
- Off-line, multi-variate assimilation of EOS land products into the NASA Catchment land surface model will improve the initialization of operational NASA seasonal forecasts, and thus the forecasts themselves
- Coupled, multi-variate assimilation of EOS land products into the NASA GEOS-5 land-atmosphere model will provide more consistent estimates of land and atmospheric conditions, improve the initialization of NASA weather and sub-seasonal forecasts, and thus improve the forecasts themselves.
Prior accomplishments:
The proposed work builds on previously funded research under two NASA-EOS grants, the first for seasonal forecast initialization from EOS soil moisture assimilation products (PI: Reichle) and the second for impact studies of EOS land surface temperature products (PI: Bosilovich). Our team of investigators comes with the experience needed to ensure rapid progress on the ambitious work plan. Specific areas of expertise include advanced land data assimilation systems, remote sensing, bias correction and quality control of satellite data (Reichle), assimilation of land surface temperature, the global water and energy cycle, reanalysis products (Bosilovich), cold land processes, snow water equivalent retrieval (Tedesco), land modeling, and seasonal predictability (Koster).
Proposed work:
In the first project phase, we will jointly assimilate AMSR-E soil moisture and snow water equivalent retrievals and MODIS land surface temperature and snow cover data into the uncoupled NASA Catchment land surface model using a distributed ensemble Kalman filter (EnKF) fitted with well-tested modules for bias correction and advanced on-line quality control. The EnKF will produce enhanced estimates of global land surface conditions that will be used for initialization of NASA seasonal forecasts. A retrospective data set based on off-line assimilation of EOS land surface products will be generated and disseminated. The second phase will focus on assimilation of EOS land products into the coupled land-atmosphere model. Coupled assimilation has the potential to produce enhanced land and atmospheric conditions that are fully consistent with each other - a powerful advantage for initializing weather and sub-seasonal forecasts and obtaining superior real-time estimates of land surface conditions for NASA instrument teams (e.g. CERES). We will conduct a highly quantitative analysis of land surface fields obtained through assimilation, quantifying their uncertainty and evaluating their relationship to raw satellite retrievals. The incremental impact of EOS land products on estimates of land surface conditions and forecast skill will be assessed relative to what can be achieved without these data. Through the proposed integration of NASA EOS data we will gain a deeper understanding of the quality and predictive potential of NASA's key land surface satellite products.
