methods may require LULC information at discrete times to analyze the environment properly. Currently, the USGS has available comprehensive LULC data for the United States. As currently planned, future LULC data collection will be user driven and not collected on a routine repeat basis. In the past, USGS acquisition of LULC data has been slow, and the data for adjacent quadrangles may have been collected years apart and by differing interpreters using different collection techniques. In these circumstances, the resulting data are variable in their spatial accuracy and attribute quality. For these reasons, users of vulnerability assessment methods may need to seek additional sources of LULC data (e.g., interpretation of synoptic satellite data).

The Level II Anderson classification scheme used by the USGS is detailed enough for simple applications and provides the framework for more specific information where additional detail is needed. However, data requirements for more complex or comprehensive methods may quickly outdistance the attribute characteristics of this scheme. For example, Anderson Level II does not distinguish pasture land from cropland or irrigated from nonirrigated lands, yet these factors may affect vulnerability significantly.

The 1987 Agricultural Census and 1987 National Resources Inventory (NRI) are additional sources of land use information for state and national assessments. The data collection techniques and classification schemes differ between the Agricultural Census and the NRI, making comparisons of data difficult. The Agricultural Stabilization and Conservation Service (ASCS) and SCS land use and land cover data collected at the farm and ranch level are valuable because of their level of detail and currency, however, they lack a digital geographic reference.

The utility of digital LULC data for ground water vulnerability studies and other applications could be enhanced by the development of a standard LULC classification scheme to achieve consistency in data at the national scale. The development of such a classification scheme will require cooperation among scientists, other user groups, and image interpretation specialists. The scheme should include hydrologically significant land use classes developed to support vulnerability assessments and other ground water investigations.

In addition, much more thought should be given to mapping land cover as a mutually exclusive data category, with land use attributes added to the cover polygons as appropriate. The existing Anderson classification scheme intermixes land use and land cover categories so that a single land use or single land cover classification is not possible. Ancillary land use and land cover attributes would enhance the utility of the simple classification scheme now in use. Examples might include estimates of surface roughness, amount of impermeable surface, and type of cultural features present.

Since land use and land cover are dynamic attributes (agricultural cropping



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement