Cross-Comparison between MODIS and VIIRS Snow Cover Products for the 2016 Hydrological Year

Shubhechchha Thapa,Andrew G Klein,Parveen K Chhetri

doi:10.3390/cli7040057

Abstract

The VIIRS (Visible Infrared Imaging Radiometer Suite) instrument on board the Suomi-NPP (National Polar-Orbiting Partnership) satellite aims to provide long-term continuity of several environmental data series including snow cover initiated with MODIS (Moderate Resolution Imaging Spectroradiometer). Although it is speculated that MODIS and VIIRS snow cover products may differ because of their differing spatial resolutions and spectral coverage, quantitative comparisons between their snow products are currently limited. Therefore, this study intercompares MODIS and VIIRS snow products for the 2016 Hydrological Year over the Midwestern United States and southern Canada. Two hundred and forty-four swath snow products from MODIS/Aqua (MYD10L2) and the VIIRS EDR (Environmental Data Records) (VSCMO/binary) were intercompared using confusion matrices, comparison maps and false color imagery. Thresholding the MODIS NDSI (Normalized Difference Snow Index) Snow Cover product at a snow cover fraction of 30% generated binary snow maps are most comparable to the NOAA VIIRS binary snow product. Overall agreement between MODIS and VIIRS was found to be approximately 98%. This exceeds the VIIRS accuracy requirements of 90% probability of correct typing. The agreement was highest during the winter but lower during late fall and spring. MODIS and VIIRS often mapped snow/no-snow transition zones as a cloud. The assessment of total snow and cloud pixels and comparison snow maps of MODIS and VIIRS indicate that VIIRS is mapping more snow cover and less cloud cover compared to MODIS. This is evidenced by the average area of snow in MYD10L2 and VSCMO being 5.72% and 11.43%, no-snow 26.65% and 28.67% and cloud 65.02% and 59.91%, respectively. While VIIRS and MODIS have a similar capacity to map snow cover, VIIRS has the potential to map snow cover area more accurately, for the successful development of climate data records.

Highlights

Snow cover extent is one of the most useful indicators of climate conditions [1,2,3]
The optimal fractional snow cover threshold of the Moderate Resolution Imaging Spectroradiometer (MODIS) swath snow map was determined by examining snow cover area at four different FSC thresholds (20%, 30%, 40% and 50%) and comparing the resulting binary snow maps with those of Visible Infrared Imaging Radiometer Suite (VIIRS)
This study suggests that VIIRS is slightly overestimating (6% more) the snow when compared to MODIS at the 30% FSC threshold which appears to produce the most comparable snow cover maps

Summary

Introduction

Snow cover extent is one of the most useful indicators of climate conditions [1,2,3]. Precise monitoring and mapping of snow cover extent are critical for understanding climate change, climate dynamics and hydrological phenomena. The information on spatial extent and distribution of snow cover is an important input factor for climate and hydrological modeling [4,5,6,7], weather forecasting [8], determining Earth radiation budget [9], estimating snowmelt runoff for water resource management and hydropower production [10,11] and for modeling flood hazard potential from snowmelt [12,13]. Satellite remote sensing allows for the detection and mapping of snow cover extent (SCE) at regional and global scales and has a potential for extending local in situ snow measurements [2].

Objectives

Methods

Results

Discussion

Conclusion