Storm Simulation

The linkage of the SHM, the MM, the THM, and the WQM is central to the SRBEX effort (Figure 3). It consists of the SHM providing the initial soil water content to MM5, MM5 providing simulated precipitation data over the SRB to the THM, and the THM providing the hydrographs at various points throughout the basin to the WQM, which then uses them to predict the chemical nature and reactivity of the runoff.

In order to perform a first test of the linked-model strategy and to identify bottlenecks in the transfer of data from one model to another, we chose a single storm event. The following summarizes the steps taken.

As a first step, the Penn State Department of Meteorology meteorological records and the USGS water-supply records were searched for storms that would fulfill the following criteria: 1) near, but not overbank full discharge on a regionally extensive set of gages. This gives a good volume of water without the complications of overbank flooding on a regional extensive storm. 2) No snow on the ground or in the storm. We did not want the complication of snow melt in the model. 3) Leaf off conditions. We did not want to have the added complication of dealing with transpiration in the THM at this early stage. This procedure produced 10 possible storms.

As a second step, the investigators needed to determine which of these strong hydrographic responses would be best to model. The working assumption was that the MM would be more effective modeling precipitation from a typical storm. Similarly, it was assumed that the THM could reproduce the hydrographic trace generated by an archetypal storm's precipitation. Thus, of the storms associated with the high-discharge events, which one was most representative?

Synoptic climatology was used to answer this question. In earlier EOS-supported research, Yarnal (1993) quantitatively compared most of the common techniques used to classify atmospheric circulation. Yarnal and Draves (1993) applied this knowledge to determine those storm systems and their tracks related to elevated discharge from one small watershed. These results suggested that it should be possible to determine subjectively the most representative event by studying the weather maps and storm trajectories associated with the strong hydrographic responses at WE-38. The analysis indicated that the 9 April 1980 storm was clearly the best choice, possessing a well-defined, normal evolutionary sequence for wave cyclones.

Although synoptic climatology successfully identified the most representative storm system, this step asked many seemingly simple questions that we could not answer. Does a single basin respond the same way to each storm event, or does its hydrographic trace change with each system? Do all basins respond the same way to a storm event, or does each basin have a unique hydrographic signature? Do basin responses to storm systems change with basin scale? Is there a more rigorous way to select representative events? These and related questions prompted the analysis presented in the "Synoptic Climatology" section (see above).

Once the 9 April 1980 storm event was identified as the best choice for our first linked-model experiment, the third step was to define the MM nested domains for this simulation. A set of three nested domains, centered on the MCW, were created, each having a mesh of 61 by 61 grid points at horizontal grid spacings of 36-, 12-, and 4-km. The remaining MM5 initial and bound ary conditions were then created from observations and a 48-hour simulation was performed. The MM5-simulated precipitation fields were output on an hourly basis. Since the MM5 simulation was performed on the CRAY-YMP, we had to develop software to read the precipitation output and gather the required information from the MM run, transfer it to the ESSC SUN workstation network, and convert it to a GIS-compatible format. Standard GIS tools can then be used to regrid the precipitation fields for use by the THM as a forcing term in the surface hydrology balance (Lakhtakia et al., 1995).

Synthetic Aperature Radar Studies at Mahantango Creek Watershed

Radar backscatter measurements provide an alternative source of information about soil water content and vegetation cover. Interpretation of these measurements is complicated by their sensitivity to terrain and other features of natural and agricultural landscapes. This is especially evident in hilly regions such as the SRB where natural ridge top forests and streambed vegetation are intermixed with contour-plowed fields and pastures.

To evaluate the applicability of radar data under SRB conditions, we are examining the influence of terrain, soil moisture, and vegetation canopy on synthetic aperature radar (SAR) backscatter. The data from the July, 1990 MACHYDRO mission, conducted over the Mahantango Creek Watershed, are being used in conjunction with nearly simultaneously acquired ground data, including soil moisture, land cover, and digital elevation models. The SAR data were collected with the NASA/JPL fully polarized C-, L-, and P-band AIRSAR instrument. This work is being conducted in three parts.

Radiometric influence of terrain on backscatter is examined for all bands and polarizations. Terrain is represented as slope and down-and-across slope curvature as determined from three digital elevation models. Each model has different resolution and x, y, and z accuracy characteristics. The results of this part of the investigation will provide insight concerning the influence of these characteristics on their use for terrain correction procedures.

SAR imaging from two incidence angles is examined as a method for detecting differences in canopy characteristics based on the changes in the two-way trip of the waves through the canopy. Detected canopy properties are compared to a map of land cover to assess the sensitivity of the angle difference method.

Soil moisture is estimated with an inversion procedure using two-frequency backscatter. The estimated soil moisture values are then compared to soil moisture values determined from field survey. The multifrequency approach has a reduced sensitivity to differences in the dielectric properties of soils based on their texture.

Human Dimensions of Environmental Change in the SRB

Research on the human dimensions of environmental change spotlighted the policy relevance of global-change projects and, specifically, SRBEX (Yarnal, 1995a). This work concluded that much basic research on environmental change cannot advance policy directly, but new projects must determine the relevance of their research to decision makers and build policy-relevant products into the research. Similarly, ongoing projects must judge whether it is possible to alter or add to the present science design to make the research policy relevant.

To be pertinent to policy makers and water resource managers, SRBEX must add a significant human component. Specifically, new objectives must: These objectives require the integrated assessment of hydroclimatic, socioeconomic and decision-making systems. However, although this research is critical to the ongoing vitality and relevance of SRBEX (Yarnal, 1992), it is clearly beyond the scope of the present program. Therefore, there has been a major effort in 1995 to procure funding for this important work. So far, three large proposals - involving more than a dozen social scientists from several departments and three colleges at Penn State - have been submitted to conduct socioeconomic and policy research on hydrologic change in the SRB (Fisher and Shortle, 1995; Liverman, 1995; and Yarnal, 1995b). Links to SRBEX are integral components of the proposed research.

Table of Contents Previous Section Next Section Download Report
Graphics and HTML text markup by Phil Kolb
Comments to: R. White, raw@essc.psu.edu