Return to List of Research Summaries
TITLE:
Mesoscale Atmospheric Modeling - Regional Surface Fluxes
INVESTIGATORS:
Roni Avissar (Rutgers), Jim Toth (UA), Susan Moran and Dave Goodrich (ARS)
- Preliminary Coupling of the Regional Atmospheric Modeling System (RAMS) with Landsat Data
Jim Toth (ARS), Jim Shuttleworth (UA), Ghani Chehbouni (ORSTOM)
- Continuing Development of Algorithms for Modeling Surface Fluxes and Assimilating Satellite Data
OBJECTIVES:
1. To assimilate satellite-sensed surface thermal data into the atmospheric model so as to improve the model's partitioning of surface energy fluxes.
2. To make necessary adjustments in the assimilation process for the effects of sub-grid heterogeneity.
3. As integral parts of objectives 1. and 2., insert or assimilate other satellite-derived data (e.g., NDVI).
4. To verify the model results.
APPROACH:
The existing surface scheme in the RAMS model uses two patches (or two components: vegetation/shaded soil and bare soil) for each grid square. In order to assimilate satellite-observed surface temperatures, the model's surface temperatures must be determined as a weighted average of the bare soil and vegetation temperatures for each grid square. In this calculation the fractional coverage of vegetation is critical. Moisture availability (surface soil moisture/stomatal conductance) and albedos are also significant. The leaf area of the vegetated patch is relatively unimportant. For the purpose of partitioning the surface energy fluxes the leaf area is also relatively unimportant when compared to the fractional coverage of vegetation and the moisture availability. Accordingly, preliminary efforts have focused on using the satellite vegetation index to adjust the model's fractional coverage of vegetation and on using the satellite thermal data to adjust the moisture availability. The model is run with 15, 4 and 1-km grid spacings. A large-eddy simulation will be attempted for Lewis Springs.
DATA COLLECTION FROM VARIOUS MODELS:
1. Variable: Large-scale atmospheric state (wind, temperature, moisture). Resolution: Horizontal 30 km, vertical approx. 160 m near ground. Method: National Weather Service Eta model. Location: Approx. 400 km radius of San Pedro. Frequency: Archived at ARS continuously since January, analyses at 03Z and 15Z, 6-hr forecasts at 09Z and 21Z.
2. Variable: Precipitable water (vertically integrated water vapor). Resolution: 30 km. Method: same as 1. Location: All US. Frequency: Archived continuously since mid April, same times as 1.
3. Variable: Mesoscale model output (wind, temperature, moisture, precipitation, surface fluxes, soil variables, vegetation temperature). Resolution: horizontal 15 km with nested 4 km and 1 km grids, vertical 120 m near ground. Method: RAMS model, initial and boundary conditions interpolated from the Eta model. Location: The 1-km nested grids cover the Huachuca Mountains, Lewis Springs and Walnut Gulch; the 4-km grid covers the entire San Pedro basin. Frequency: Synoptic periods, significant precipitation events, and approx. 60% of the remaining dry periods; for all grids a subset of model variables is saved at 3-hr intervals; for selected locations all soil and surface variables, including component surface fluxes, are saved at 15-min intervals.
HELPFUL DATA FROM OTHER INVESTIGATORS
1. Areal average surface fluxes.
2. Mean atmospheric and soil observations.
3. Representative observations of vegetation cover.
4. Surface radiometer temperatures, with location.
5. Any available fluctuation atmospheric variables (approx. 2-5/min) from Lewis Springs.
Return to List of Research Summaries
homepage