Regular monitoring and mapping of the Snow Cover
Area (SCA) is important to manage the natural resources and to assess the
impact of climate change on SCA. But, over inaccessible Western Himalayas, the
estimation of the SCA is one of the challenging tasks due to its complex and
rugged topography. SCA was mapped so far with optical sensors. Generally, the
optical sensor data is affected by the presence of the cloud cover and is more
sensitive towards interference from environmental effects. Alternatively,
several developments were made by various authors to map the SCA using active
or passive microwave data all over the globe. It is proven that
scatterometer data has the potential to retrieve the snow cover information.
But the mapping of SCA using active microwave satellite data is still in its
initial phase due to lower accuracy as compared to optical sensors data.
Due to the recent advancements in classification
algorithms such as Sub-Pixel Classification (SPC) and Super-Resolution Mapping
(SRM), there is a clear requirement of extensive exploration of different
classification algorithms for SCA estimation especially over complex undulating
Western Himalayas using SCATSAT‒1 data. The integration and evaluation of
advanced classification algorithms with scatterometer data are some of the
primary issues in the remote sensing field. Therefore, with an investigation on
such critical issues, it may be possible to extend the applicability of
SCATSAT‒1 data in different applications. It is also expected that such the analysis will improve the performance of a scatterometer in the precise mapping
of SCA over rugged terrain surface.
Methodology to generate Snow Cover Maps
This study is capable to evaluate the Indian Satellite SCATSAT-1 data for snow cover mapping. Three classifiers KMC, SVM, and LSM were used for the generation of binary snow cover maps from SCATSAT-1 backscattered values. For better accuracy of the maps super-resolution mapping technique in conjunction with NDSI data has also been used. The results of classifiers and SRM have been compared with NDSI images generated from Landsat‒8 and MODIS data on clear sky days, respectively. This means that LSM-SRM on SCATSAT-1 is a good candidate for classifying the SCMs at Ku-band but it requires an additional dataset. Further, such outcomes will be beneficial in the study of the cryosphere, especially on cloudy days. LSM-SRM results show that regular snow cover mapping for the generation of a time series maps in high‒altitude Himalayas and other regions of extreme climate sensitivity, where cloud cover plays an important role during the winter season, maybe possible by fusion of scatterometer data with optical sensor data. Currently, the application range of SCATSAT-1 may be limited due to some difficulties such as a) incidence angle variations affect the backscatter values, and b) difficult to identify pure endmember properties due to low spatial resolution. Further studies to overcome the limitations in examining SCA from scatterometer data may be useful for trend analysis of early snowmelts and seasonal snow cover variations.
Sood et al. (2020) address the potential
of SCATSAT‒1 data for snow cover mapping over a
part of Western Indian Himalayas. Three classifiers KMC,
SVM, and LSM were used for the generation of binary snow cover maps from
SCATSAT-1 backscattered values. For better accuracy of the maps
super-resolution mapping technique in conjunction with NDSI data has also been
used.
Reference: Sood V., Gusain H.S., Gupta S., Singh S., Kaur S. 2020. Evaluation of SCATSAT-1 data for snow cover area mapping over a part of Western Himalayas, Advances in Space Research, Elsevier.
Link for full study: https://doi.org/10.1016/j.asr.2020.08.017
