Validation of the MODIS albedo product and improving the snow albedo retrieval with additional AMSR-E data in Qinghai-Tibet Plateau

Abstract

Land surface albedo is crucial for land surface radiation and energy budgets. In this study we compared the MODIS 16-day albedo product (MCD43A3) with field-measured data in Qinghai-Tibet Plateau. The validation data were used from 4 automatic weather stations(AWS) locations, spanning the year 2002-2008. Results indicate that MCD43A3 albedo product in snow-free seasons is in good agreement with ground-based observations, with a bias of ±0.02-0.05. But in snow season of Qinghai-Tibet Plateau, the MCD43A3 albedo product reaches a high bias. One of the possible reasons is that the amount of bidirectional reflectance observations may not be sufficient for getting the high quality surface albedo retrieval because of cloudy weather during the snowing days. Another reason may be that the heterogeneity of snow surface and complexity of snow grain. It is well known that the snow albedo is influenced by many parameters. However, the accumulated daily maximum temperature is shown to be a good predictor of the snow albedo. And also the snow albedo may effected by snow depth and snow water equivalent. In this paper, we improved a snow albedo retrieval model through daily maximum temperature from AWS and SWE from AMSR-E which can provide time series observations during snowing and snowmelt period. Also the AMSR-E SWE product has a coarse-resolution (25km) and has some uncertainties, the results show better correlation with the field-measured snow surface albedo. The 16-day average value of this algorithm performs well when there is snow in spring.