Laserfiche WebLink
<br />Landsat data can provide the information necessary to map areal snow <br />extent in large watersheds (Barnes and Bowley, 1974). This information, used <br />in conjunction with traditional ground surveys, may increase the accuracy of <br />forecasting runoff amounts. The increased accuracy of forecast would improve <br />the efficiency of allocating water derived from seasonal snowpack. <br /> <br />The Image-IOO System was selected for determination of areal snow <br />extent. Digital tapes of the February 26, 1975, Landsat image (fig. 14) were <br />used for this analysis. The Image-l00 System was selected for use in this <br />application because it provided the opportunity to map only those areas <br />within the Yampa River basin boundaries and because spectral complexity was <br />not a significant factor in this image, so the faster parallelepiped-decision <br />function was suitable for classification. <br /> <br />In cloud-free areas of the western part of the Yampa River basin, deep <br />snow in open areas has a distinctive spectral signature, with large digital <br />values in all four Landsat bands. It was a simple matter to define a <br />training set for any small area and to subsequently classify all open areas <br />covered by deeper snow. <br /> <br />The Landsat image included many areas with digital values less than <br />those for deeper snow in open areas but much greater than those for other <br />land-cover types. These areas were classified qualitatively as thin or <br />partly shadowed snow, and very thin or shadowed snow. Shadowed snow is most <br />likely the result of various densities of exposed vegetative cover. Ground <br />surveys would be needed to determine whether the snow in these areas was deep <br />enough to be significant for predicting future streamflows. In relatively <br />open areas of thin snow, Landsat data might be used to show the areal extent <br />of snow thickness, as measured at a few points on the ground. In open areas <br />of thicker snow, however, present landsat data provide no information on <br />thickness, water equivalent, or free-water content. <br /> <br />Clouds have approximately the same spectral signature as snow, and <br />shadows obscure land-cover features. Some cloud cover is common in the <br />River basin during the winter and could cauSe problems in delineating <br />extent on some of the Landsat images. <br /> <br />cloud <br />Yampa <br />snow <br /> <br />Concern has been expressed by some scientists about snow areas being <br />obscured by dense conifer forests (Rango, 1975, p. 61, 231). This did not <br />appear to be a major problem in the snow-covered areas of the western part of <br />the Yampa River basin. In this area, there are few stands of dense conifers, <br />and many groups of conifers occupy less area than the 1.1 acres (0.45 ha) of <br />a Landsat pixei. Thus, most of the deep-snow class appeared to be continuous <br />on this Landsat image. However, in mountainous areas of 'the eastern part of <br />the basin, which supply the majority of the snowmelt runoff, conifer stands <br />are dense in many locations and could create problems in digital classifi- <br />cation. <br /> <br />A basin-boundary map was registered and digitized similarly to the <br />approach discussed previously for the August 24, 1975, Landsat-image land-use <br />analysis. However, in this instance, map registration was a time-consuming <br /> <br />32 <br />