Ads. Space Res. Vol. 9. No. 1, pp. (t)207—(1)215, 1989 0273—1177/89 $0.00 + .0 Printed in Great Britain. All rights reserved, Copyright © 1989 COSPAR ESTIMATION OF PROPERTIES OF ALPINE SNOW FROM LANDSAT THEMATIC MAPPER Jeff Dozier Center for Remote Sensing and Environmental Optics, University of California, Santa Barbara, CA 93106 U.S.A. and Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109 U.S.A. ABSTBACT Estimation of snow characteristics from satellite remote sensing data requires that we distinguish snow from other sur- face cover and from clouds, compensate for the effects of the atmosphere and rugged terrain, and interpolate snow albedo over the entire solar spectrum from measurements at a few wavelengths. We also need to account for topo- graphic effects without requiring that satellite data be precisely registered to digital elevation data, because the poor quality of most digital elevation data introduces considerable noise into calculations of slope and azimuth. From simula- tion of a range of snow types and various atmospheric profiles, over possible illumination conditions, we can develop typ- ical spectral signatures above the atmosphere over mountainous terrain. Landsat Thematic Mapper data of the south- ern Sierra Nevada are analyzed to distinguish several classes of snow from other surface covers. Snow can be reliably mapped at all sun angles encountered in the mid-latitudes, and large surface grain sizes can be distinguished from areas where the grain size is finer at the snow surface. Because of saturation in TM band 1, estimation of the degree of con- tamination by absorbing aerosols is not feasible. INTRODUCTION Snow, glaciers, ice in lakes and rivers, and ice in the ground comprise the frozen part of the land portion of the hydro- logic cycle. Water in these frozen states accounts for more than 80 per cent of the total fresh water on Earth and is the largest contributor to runoff in rivers and ground water over major portions of the middle and high latitudes. Snow and ice also play important interactive roles in the Earth’s radiation balance, because snow has a higher albedo than any other natural surface. Over 30 per cent of the Earth’s land surface is seasonally covered by snow, and loper cent is per- manently covered by glaciers. Snow cover represents a changing atmospheric output resulting from variability in the Earth’s climate, and it is also an important changing boundary condition in climate models. Thus it is important that we monitor the temporal and spatial variability of the snow cover over land areas, from the scale of small drainage basins to continents. Over the last 10-20 years satellite remote sensing has opened the possibility of data acquisition at regular intervals, and operational as well as research-oriented satellites have provided information on snow cover. The Landsat systems, in particular, have provided multi-wavelength visible and near-infrared data for hydrological and glaciological research at the scale of drainage basins. The most important information that can be derived about the snow cover from measure- ments from satellite of the reflected solar radiation are maps of snow-covered area and rates of snow-cover depletion and estimates of spectral albedo throughout the solar spectrum. Satellite remote sensing data data provide information on the spatial distribution of parameters of hydrologic impor- tance, snow-covered area, surface albedo, and snow water equivalence. For the seasonal snow cover, remote sensing has been used to improve the monitoring of existing conditions and has been incorporated into several runoff forecasting and management systems. The alpine snow cover and alpine glaciers in the mid-latitudes are important to our understand- ing of global and regional climates, and to our use of water resources /1,2/. Much of the uncertainty and sensitivity in the global hydrologic cycle lies in these reservoirs of frozen water, and their recent melting appears to account for much of the recent rise in sea level /3/. The most common use of remote sensing in snow studies is to monitor snow-covered area /4,5/. These efforts have been carried one step further by including satellite-derived measurements of snow covered area as an index in a snowmelt runoff model /6/. The next step is to use satellite radiometric data to measure or estimate snow water equivalence and snow surface properties that are necessary for the calculation of the surface energy balance /7/. Identification of snow during daylight hours is straightforward during clear weather, because of high albedo of snow in the visible wavelengths /8,9/. Automatic discrimination between snow and clouds is possible with a wavelength band between 1.55—1.75 jim, where snow is dark but clouds are bright /10,11,12/. The last decade has been marked by advances in our understanding of the radiative properties of snow. The important conclusions from the considerable work in the optical wavelengths, where all radiation of importance to the energy bal- ance is contained, is that the characteristics of snow reflectance can be modeled by the radiative transfer equation /13/. In the visible wavelengths ice is highly transparent, so the albedo of snow is sensitive to small amounts of absorbing impurities /141. In the near-infrared wavelengths ice is more absorptive, so the albedo depends primarily on grain size /8,9/. While grain size and the amount of absorbing impurities can be roughly estimated with present-day satellite sen- sors sensors /12,15/, the advent of high-spectral resolution data in the 1990’s should enable us to measure snow reflectance and estimate surface snow properties more accurately. In alpine areas the problems of spatial resolution are severe. Measurement of snow and ice properties by remote sensing requires satellite data on similar spatial scale to that of the topographic relief, i.e. a few tens of meters. Such data are available, either from the Landsat Thematic Mapper (TM) or from the French Systeme Probatoire d’Observation de Is Terre (SPOT). The TM has the better spectral coverage, whereas SPOT has a finer spatial resolution and pointing capa- bility. In either case, the extremely large data volume make the analysis difficult. How we use such data to measure key areas of the alpine snow and ice cover is a challenging issue. (1)207