Using Remotely Sensed MODIS Snow Product for the Management of Reservoirs in a Mountainous Canadian Watershed

摘要

The objective of this study is to analyze relationships between the variability of MODIS derived snow covered area (SCA) and flood dynamics for the management of reservoirs. The study was conducted in the Nechako River Basin in British Columbia (Canada). Relationships were analyzed between daily SCA calculated from filtered MODIS images and the fraction of the flood volume that had already entered the reservoir during the snowmelt period. Results indicate a robust (R-2 = 0.87) and significant (p < 0.001) correlation over a 14-years (2000-2013) period. A similar analysis was conducted for each sub-basin, producing an even more robust relationship (R-2 = 0.90) for the Tahtsa Lake sub-basin. Two distinct relationships corresponding to years with either large or small spring runoff were identified by classifying years according to their maximum snow water equivalents (SWE), peak inflows and flood volume, thus improving the correlations. These empirical relationships are a simple forecasting tool that does not require any additional data other than MODIS SCA and SWE measurements. This straightforward approach can be applied to other mountainous watersheds dominated by snow accumulation and melt.