Laserfiche WebLink
<br />Current methods of collecting and updating information on land-use <br />practices frequently are costly and time consuming. This is especially true <br />for a large region like the Yampa River basin where land use is diverse and <br />often subject to change both at regular and irregular intervals, A dynamic <br />system such as this requires regular updates of information if the <br />information is to be a realistic tool for planning and subsequent development <br />and allocations of regional natural resources, <br /> <br />Landsat images may be a source of data to assist in determining land use <br />and monitoring its changes. Landsat provides repetitive coverage of large <br />areas which may be suitable for determining land-use categories on a regional <br />scale and updating the information at regular intervals, A number of studies <br />have been conducted that use Landsat data for land-use classification. These <br />studies have resulted in varying degrees of success. However, few studies <br />applying Landsat data to land-use classification have been conducted in <br />regions that are as complex and diverse both culturally and physically as the <br />Yampa River basin, The Yampa River basin is representative of several <br />regions in the Rocky Mountains which are undergoing rapid development of <br />energy resources. The basin's complexity, diversity, and potential for rapid <br />development make it a challenging area for testing the feasibility of using <br />Landsat data for land-use classification, <br /> <br />Both photographic-interpretation and digital-image processing techniques <br />were used for land-use classification. Photographic-interpretation tech- <br />niques were applied to band-7 black-and-white photographs of Landsat images <br />for February 25 and 26, 1975, and to a color-composite photograph of the <br />Landsat image for August 24, 1975. Image scales of 1:1,000,000 were used in <br />this analysis. Computer-compatible digital tapes of the Landsat image for <br />August 24, 1975, were used for digital-image processing and analY5is, <br /> <br />Digital-image processing and analysis techniques were emphasized in <br />land-use applications primarily because of their abil ity to classify large <br />areas much more rapidly than photographic-interpretation methods, Therefore, <br />photographic-interpretation techniques were used only for identification of <br />land-use features that would be difficult to identify by spectral char- <br />acteristics (digital analysis) alone and which would be relatively easy to <br />identify by incorporation of spatial characteristics, such as linearity. <br /> <br />The land-use classification system suggested by Anderson, Hardy, Roach, <br />and Witmer (1976) was used as a guide in the definition of possible land-use <br />categories (table 2), Also, results from the Landsat classification were <br />compared with table 2 to determine the level of classification that could be <br />obtained, for various categories, using digital-image processing, <br /> <br />Land-use categories and representative areas (training sets) for each <br />category were defined using a combination of black-and-white 1 :24,000-scale <br />aerial photographs and available ground information. The black-and-white <br />aerial photographs obtained in 1973 were the most recent available coverage <br />of the area at the time of the analysis. This proved to be somewhat of a <br />disadvantage in selection of land-use categories and representative training- <br />set areas on the 1975 Landsat imagery. However, sufficient sites representing <br /> <br />8 <br />