Fine Scale Meteorological Simulations of the Houston-Galveston Metropolitan Area with LANDSAT-Derived High-Resolution Land Use and Land Cover Datasets

To characterize the effects of urbanization on the local atmospheric conditions, a fine scale meteorological modeling that has a higher resolution than the current grid size, usually around at 4- or 5-km, would be necessary. Such a model would require an adequate description of the atmospheric dynamic/thermodynamic conditions as an input at around 1-km grid resolution. Meteorological modeling at 1-km resolution has been attempted by several researchers. However, one of the common concerns was the limited resolution and accuracy of the critical input data such as the land use coverage data. For example, the current USGS 25-category land use data linked with the MM5 meteorological modeling system is roughly at 1-km resolution data elements somewhat out-dated (with the reference year 1990). Besides it uses only one urban category that does not distinguish among the built-up urban, residential, planted trees, and road and pavement areas. Our experience of meteorological modeling for the Houston-Galveston Area (HGA) show that the data can describe the air-land surface exchange processes at best up to 4-km MM5 grid resolution and the 1-km grid meteorological simulation would not be possible because of the poor quality of the land use input data. Recently, with the support of the Texas Forest Service (TFS), GEM has generated highly accurate land use (LU) and land cover (LC) datasets separately for the Houston and the surrounding eight county areas using the 30 meter resolution LANDSAT satellite imagery and ancillary datasets of varying spatial resolutions for the reference year 2000.

This paper focuses on the specification of the land surface conditions inside MM5 by replacing the original USGS 25-category data with the new LU/LC data within the Houston eight counties. The objective of this paper is to study the impacts of using the new LU/LC datasets on fine scale meteorological simulation. To fully exploit the advantage of the new LU/LC dataset, a comprehensive land surface model (NOAH LSM) is utilized in order to provide necessary evapotranspiration and moisture diffusion processes, which are not included in the simple slab soil model in MM5.

The usage of new LU/LC datasets in 4-km simulation domain, although represent appreciable differences in the detailed spatial patterns, does not show obvious improvements over the original land use data. However, the utilization of LU/LC datasets at the 1-km domain shows significant difference and better performance than those with the original land use data. With the updated and finer LU/LC data, fine scale meteorological modeling is able to achieved.