Land Surface Characteristics Data
Objectives
The various models require several types of land surface characteristics
information, including
In general, these data require two types of preprocessing before they
can be passed to the models:
- Extraction and/or computation of the physcial parameters needed by
the models, and
- Aggregation and reprojection of the data from the scale and map
projection in which they were distributed to the scale and map
projection required by a particular model.
The currently available geographic datasets and any dataset-specific
preprocessing are discussed below. Other hypertext pages provide
additional information on standard tools for
aggregating and reprojecting rasterized or gridded data, processing
and analysis of remotely sensed information, and
the local database
facilities for storing and retrieving the data.
We have acquired USGS digital elevation model (DEM) data at 3-arcsecond
resolution (corresponding to grid spacings of approximately 90 meters
north-south by 70 meters east-west) for the entire Susquehanna River
Basin, and at 30-meter resolution for for about 80% of the 7.5-minute
map quadrangles in the Basin. A 5-meter DEM dataset for the WE38
intensive study area within the Mahantango Creek Watershed, derived from
aerial stereophotography, is also on hand.
The Pennsylvania Geological Survey has made available to us a
provisional vector coverage of geololgical formation surface contacts
for the entire state. For the part of Pennsylvania within the
Susquehanna River Basin (SRB), the formations have also been classified
by general lithology type, using the six categories of
sandstone/quartzite, interbedded sandstone and shale, shale/siltstone,
carbonate, gneiss/schist, and other (basalt, diabase, etc.). Digital
geologic data are not yet available for the parts of the SRB in New York
and Maryland.
The USDA-NRCS (Natural Resource Conservation Service) has recently
released the State Soil Geographic (STATSGO) Data Base containing soil
characteristics for the 48 conterminous United States. Development of
software for extracting the dominant soil texture class within each grid
cell of an arbitrary model domain and generating look-up tables giving
corresponding values of soil physical properties required by the models
(e.g., soil water (SW) content at saturation, minimum soil suction,
saturated hydraulic conductivity, SW at wilting point, etc.) is nearing
completion.
Soil properties derived from the STATSGO database appear to be
significantly more realistic than the estimated values which many models
have used in the past. Some preliminary results by Miller and Lakhtakia
(1994a; 1994b) are
described in the 1994 Annual Report.
For regions within the Chesapeake Bay drainage area, the EPA EMAP
program has published a landcover dataset at 25 meter resolution,
derived mainly from Landsat Thematic Mapper imagery.
Models whose domain extends outside the SRB can use a 1-km gridded
landcover dataset, recently released by the EROS Data Center (EDC),
which covers the 48 conterminous states and was derived by analyzing
satellite imagery (AVHRR) in conjunction with climatological, land
elevation, and ecoregion information. This dataset uses a somewhat
different classification scheme than the older LUDA landuse/landcover
dataset, which was compiled on a set of 1:250,000 scale base maps
covering the entire U.S. from aerial photography acquired, for the most
part, during the 1970s.
The EDC and LUDA datasets both divide landuse/landcover into a larger
number of classes than are physically meaningful for current atmospheric
models; standard GIS software has been used to aggregate the classes to
the smaller number required by the models. As with the soils data,
preliminary results (Lakhtakia and Miller,
1995; see also the 1994 Annual Report) indicate that high
resolution landcover grid derived from the EDC data describes landcover,
and the vegetation properties associated with it, more adequately than
the lower resolution dataset, archived at NCAR, which has generally been
used with these models.
The USGS has published digitized representations of the boundaries of
the major watersheds ("8-digit" watersheds) for the entire U.S., with
accuracies corresponding to a map scale of 1:250,000; these have been
added to our EOS database. Since the gauging stations used for studies
of water flow and quality are not, in general, at the outflow points of
these watersheds, a separate watershed divide coverage has been created
which defines the catchment areas associated with the 61 USGS and PA
Department of Environmental Resources water quality sampling stations
within the PA portion of the SRB (Richards and Kump,
1994b).
A digital dataset defining all streams and bodies of water in PA,
digitized from 1:100,000 topographic maps, is also in the database.
Last change: 12 May 1995,
R. A. White / raw@essc.psu.edu