Monitoring Snow Cover in Himachal Pradesh, India using Remote Sensing
- Arpit Shah
- Oct 17, 2021
- 8 min read
Updated: Dec 12, 2025
SECTION HYPERLINKS
BACKGROUND
The Cryosphere comprises everything on Earth’s surface—glaciers, sea ice, polar caps and snow. Although these regions are inhospitable for most forms of economic activity, they are vital to humanity because they regulate global climate systems by reflecting incoming solar radiation. A decline in polar ice is often treated as a leading indicator of widespread Global Warming. For this reason, monitoring change in the Cryosphere is of deep scientific interest within the Earth Observation community.
This post focuses on monitoring Snow Cover in Himachal Pradesh, India, between 2018 and 2021. The state is studded with mountains from its northwest to southeast, and these high altitudes receive significant winter precipitation in the form of snow. Snowmelt is the primary source of water for the major rivers of northern India, including the Beas, Chenab, Ravi, Sutlej and Yamuna. The linkage is direct and critical:
Less Snowfall → Less Snow Cover → Less Snowmelt → Lower River Discharge → Threat to Water Resources and Livelihoods
My motivation for writing on this topic was a report published in August 2021, which stated that Snow Cover in Himachal Pradesh reduced by 18% between October 2020 and May 2021. The study—conducted by HIMCOSTE and SAC Ahmedabad—analysed AWIFS imagery from ISRO's ResourceSat-2 satellite.
For my own study, I processed multispectral imagery from the Ocean and Land Colour Instrument (OLCI) onboard ESA’s Sentinel-3 satellite. I analysed two sets of imagery:
Four images acquired between Winter and Summer of 2021→ Objective: observe the rate of snowmelt.
One image acquired during Summer in each of the years 2018–2021→ Objective: observe year-on-year trends in Snow Cover.
A calendar-view of the imagery acquisition timeline is shown in Figures 1 and 2.
METHODOLOGY
Below is an example of Sentinel-3 OLCI multispectral imagery visualised in RGB mode:
To map Snow Cover, I applied the Normalized Difference Snow Index (NDSI). Traditionally, NDSI relies on Short-Wave Infrared (SWIR) reflectance because clouds and snow differ significantly in this region of the spectrum. However, Sentinel-3 OLCI does not capture SWIR wavelengths. Therefore, I used an adapted NDSI formulation developed by Kokhanovsky et al. (2019), which delineates snow using Near-Infrared reflectance—the highest SWIR-free alternative available from OLCI.
Figure 4 depicts the complete processing chain used to convert raw OLCI imagery into Snow Cover Maps using SNAP software. The adapted NDSI equation is implemented via Band Maths operator within the chain.
Sharing a brief explanation of the key processing steps-
Rayleigh Correction— Sentinel-3’s OLCI instrument captures reflected solar radiation across 21 spectral band in Top of Atmosphere (TOA) mode and functions as a passive sensor, meaning the energy source is the Sun rather than an onboard emitter. You may be familiar with Rayleigh Scattering—the phenomenon where fine atmospheric particles scatter shorter wavelengths of sunlight (especially blue), giving the sky its characteristic color.
To obtain an accurate representation of surface-level reflectance, this scattering effect must be removed. The Rayleigh Correction algorithm converts TOA reflectances into rBRR (Bottom of Rayleigh Reflectances), thereby eliminating distortion introduced by atmospheric scattering.
Subset— Subsetting a satellite image can involve:
a) clipping the geographic extent,
b) removing unused spectral bands, and/or
c) removing additional metadata included with the imagery package.
Subsetting significantly reduces dataset size, which speeds up processing. For this Snow Mapping workflow, I applied only a geographic subset, limiting the imagery to the Himachal Pradesh study area.
Reproject— Reprojection assigns a specific Coordinate Reference System (CRS) to the output. Satellite imagery is initially captured in sensor geometry—essentially an arbitrary XY coordinate space that does not inherently correspond to Earth’s geography.
Map Projection, on the other hand, involves converting information from Earth’s 3-dimensional curved surface onto a 2-dimensional map. Reprojection ensures that each pixel’s location in the dataset corresponds correctly to real-world geographic coordinates, aligning the imagery with an Earth-based reference system
A step-by-step video demonstration of these methods (applied to a different study area) is available in the linked tutorial.
OUTPUTS AND INTERPRETATION
Processing a single Sentinel-3 image with the adapted NDSI yields the Snow Cover map shown below:
Snow maps for February, March, April and June 2021 are displayed in Figure 6 below:
A sharp reduction in Snow Cover is visible by June 2021, as expected with advancing summer. Interestingly, the map for April 2021 shows a temporary increase due to unseasonal snowfall—typically less dense and prone to rapid melting, which can trigger avalanches.
Year-on-year comparisons for a summer day in 2018–2021 are shown in Figure 7 below:
It is evident that June 2021 exhibits significantly lower Snow Cover than May 2019 and May 2020. The lowest output of all four years appears in May 2018, which is consistent with findings from a separate scientific study on Snow Cover during that period.
In Figure 8 below, I have combined all the comparable monthly outputs from 2021 (Figure 6) into a consolidated map:
Similarly, in Figure 9 below, I have combined all the comparable yearly outputs from 2018-21 (Figure 7) into a consolidated map:
While I have interpreted the outputs visually thus far, it is also possible to derive quantitative insights directly within SNAP
Figure 10 depicts the geographic extent of Snow Cover for the four processed outputs from 2021:
To summarize these findings, a ~36% decrease in Snow Cover is observed between February 2021 and June 2021. Interpreting this figure requires caution: February is a winter month, whereas May and June fall in summer. Naturally, a portion of the Snow Cover melts during this period. What constitutes a normal melt percentage, however, is something I am not familiar with.
That said, if you recall the scientific study mentioned earlier, the researchers reported an 18% decrease in Snow Cover between October 2020 and May 2021—an output they classified as unusual and attributed to Climate Change. From this, we can infer that the standard seasonal melt is well below 18%. Hence, the 36% decrease observed in my study is indeed alarming.
This discrepancy also raises a methodological question for the earlier scientific study: An interesting methodological question arises here: why did the researchers choose October 2020 as the baseline month? October marks the end of the monsoon season, when Snow Cover is typically at its lowest. Comparing it to a peak summer month—where Snow Cover is again naturally low due to extensive melt—seems an unusual choice. While I am willing to believe the researchers had a compelling rationale, this example highlights why we should avoid accepting research findings in news portals at face value. Headlines typically emphasize eye-catching numbers, whereas the underlying methodology is rarely explained
It was precisely this uncertainty around “normal” melt rates that prompted me to conduct year-on-year comparisons using imagery acquired on approximately the same summer date each year. These outputs complement the month-on-month findings from 2021 and help validate whether Snow Cover is indeed declining—thus strengthening the Climate Change hypothesis.
Figure 11 depicts the Snow Cover extent for the four outputs from 2018 to 2021:
To summarize these results, a 42% decrease in Snow Cover is observed in the June 2021 output when compared to May 2019, and a 34% decrease when compared to May 2020. This reinforces the pattern seen in the 2021 analysis: the effects of Climate Change are clearly visible.
However, a 25% increase in Snow Cover appears when comparing June 2021 to May 2018. Although this seems counter-intuitive to the Climate Change narrative, one can argue—rightly—that the impacts of Climate Change are not linear or unidirectional. Increased volatility in weather patterns is a hallmark of anthropogenic warming.
Two additional factors likely contributed to this anomaly:
Incomplete spatial coverage in the May 2018 image. As shown in Figure 12, a small part of Himachal Pradesh is missing from the 2018 scene. While this omission would reduce the Snow Cover estimate for that year, the missing area is modest, and the increase would most likely be in the single-digit percentage range
Figure 12: May 2018 Imagery does not completely cover the violet-shaded extent of Himachal Pradesh Record heatwave conditions in May 2018. News reports confirm that the region experienced unusually high temperatures during that period. This would have accelerated snowmelt, leading to the lowest Snow Cover extent among all eight datasets. Indeed, in the remaining seven outputs, Snow Cover consistently exceeds 10,000 sq. km, indicating that 2018 is an outlier rather than evidence against Climate Change.
PARTING THOUGHTS
Himachal Pradesh is not an exception—Climate Change and anthropogenic warming are impacting the Cryosphere globally. In the Sierra Nevada mountain range in the USA, for example, I observed a 90% reduction in Snow Cover between February and May 2021:
Remote Sensing remains indispensable for such Earth Observation workflows. The high temporal frequency, regional coverage, and adequate spatial resolution of modern satellites allow researchers to monitor environmental change continuously and at scale. The insights derived from satellite-based analysis help governments and institutions implement risk-mitigation and adaptation strategies in the face of accelerating climatic shifts.
Feel free to share your thoughts.
ABOUT US - OPERATIONS MAPPING SOLUTIONS FOR ORGANIZATIONS
Intelloc Mapping Services, Kolkata | Mapmyops.com offers a suite of Mapping and Analytics solutions that seamlessly integrate with Operations Planning, Design, and Audit workflows. Our capabilities include — but are not limited to — Drone Services, Location Analytics & GIS Applications, Satellite Imagery Analytics, Supply Chain Network Design, Subsurface Mapping and Wastewater Treatment. Projects are executed pan-India, delivering actionable insights and operational efficiency across sectors.
My firm's services can be split into two categories - Geographic Mapping and Operations Mapping. Our range of offerings are listed in the infographic below-
A majority of our Mapping for Operations-themed workflows (50+) can be accessed from this website's landing page. We respond well to documented queries/requirements. Demonstrations/PoC can be facilitated, on a paid-basis. Looking forward to being of service.
Regards,
