Determination of snow depth and water equivalent by remote sensing
Steinhoff, Harold W., author
Barnes, Albert H., author
Environmental Resources Center, publisher
This exploratory study was designed to investigate the possibilities of using inexpensive aerial remote sensing methods to measure the snowpack and its water content. The relation of snow depth and elevation on the same aspect (north or south) was definitely linear but the slopes of the regression lines varied between months and between years. Thus no uniform prediction was possible of snow depth over an entire watershed from a single measurement at one point. Example: The regression equation for March 4, 1972 on the north aspect was YSNOW DEPTH(dm)= -24.5 + (1.10 ± 0.51)XELEV(100m). Snow depth and water equivalent were consistently related to melt date, sometimes quite strongly. The addition of vegetation density to the equation significantly increased the proportion of variability of snow depth which is accounted for. Little strength was added by including the other environmental factors, - aspect, elevation, and slope degree. However, elevation was more strongly related to snow depth early in the spring and aspect was more strongly related later. Prediction of water equivalent was improved by including degree of slope early in the season, and by including aspect and then slope later in the season. An example of multiple regression equation with a multiple R of 0.81 is YSNMAR= -14.0 + 0.08XMELTDA+ 0.095XVEGDEN+ 0.004XELEV - 0.021XFROMN- 0.053XSLOPE where subscripts mean SNow depth (dm) in MARch, MELT DAte (year-day), VEGetation DENsity (100-foot candles), ELEVation (meters), azimuth FROM due North less than 180° (degrees), and degree of SLOPE (Percent). Therefore by measuring the melt date and environmental variables one could predict snow depth and water equivalent, once these equations were established for a given area. Melt date can be measured by observation from two aerial flights at three-day intervals in early spring. The relationship between the elevation of the snow-melt line and time was determined to be linear for each of the three years observed. The average slope of the linear regression was 19.64 meters per day (0.82 meters per hour) during the snow melt period. This rate of recedence was consistent during the three-year period of observation. The variation in date over the three-year period for the snow-melt line to be at a given elevation, was approximately equivalent to the time period for the snow-melt line to raise 600 meters. The photogrammetric determination of snow depths over the area was restricted by the limited number of ground control targets. The measurement and computational procedures for the photographic imagery were adequate for the intended purpose. The determination of the location of additional visible ground control in the available photography would permit more definitive results. The point measurement of the elevation of snow fields can be accomplished if there are sufficient shadows due to vegetation or image texture due to dirt or surface irregularities. Our conclusion is that determination of snow depth and water equivalent by remote sensing from aircraft is possible. We have uncovered the basic principles but further work is needed to develop the method.
Submitted to Office of Water Research and Technology, U.S. Department of the Interior.
Snow surveys -- Remote sensing