Reconstruction of Snow Water Equivalent and Snow Depth Using Remote Sensing Data

Abstract

Snow water equivalent (SWE) and snow depth are some of the most important quantities in describing the properties of the accumulated snow during winter, which is a source of runoff during spring season. Here, our objective was to reconstruct the spatial dynamics of SWE and snow depth over a study area in eastern parts of the northern Alberta during the period 2007-09. The employed methods consisted of: (i) delineating snow presence from Moderate Resolution Imaging Spectroradiometer (MODIS)-derived normalized difference snow index (NDSI)-images, (ii) calculating heating degree days (HDD) from MODIS-based surface temperature images, (iii) modelling net solar radiation, and (iv) integrating all of the above steps in the frame of a process based snow-melt model and SWE ground data as well. We used ~45% of the ground data (i.e., ~19 data points) in calibrating the values of base temperature and heating degree day coefficient for the model. Then the remaining ~55% of the ground data (i.e., 23 data points) were used in validation. It revealed that the agreement between the model and measured SWE-values were reasonable (i.e., 59%, 72%, and 62% of the time values were within ±20% deviations during 2007, 2008, and 2009 respectively). The root mean square deviation (RMSD) between the measured and modelled SWE-values were also reasonable and found to be ±24.75 mm in 2007, ±25.05 mm in 2008, and ±23.99 mm in 2009. Overall, the SWE-predictions at all of the measurement sites were on an average 7.5% higher in 2007, 10.2% lower in 2008, and 1.9% lower in 2009 than that of ground-based measurements. During the period 2007-2009, we found that the study area-specific average values of SWE and its depth were 177 mm and 694 mm respectively.

Keywords: MODIS, normalized difference snow index, heating degree days, net solar radiation, process based model, snow depth