Deep Tech and Environment Conservation:
How Remote Sensing and Satellite Imagery
Can Help Us Save the Planet
Climate change is the biggest threat facing Earth today! Our planet, home to a growing population of over 7.8 billion people and countless species of flora and fauna, has changed significantly over the last few decades. These widespread environmental changes have prompted governments and public bodies to increase investment in nature conservation in an attempt to arrest and possibly reverse the effects of rampant climate change. Research indicates that around USD 50 billion flows into Nature Conservation projects every year. While the Fourth Industrial Revolution contributed to a major part of the instability associated with climate change, the advancement of remote sensing and data collected through satellite imagery offer solutions to identify problems and help policy makers to do root-cause-analysis and make decisions.
Remote sensing involves understanding changes in broad landmasses through the detection and monitoring of the physical or chemical characteristics of an area by measuring from a distance the radiation reflected or emitted by the examined area. Remote sensing is used by governments and environmentalists across the globe to get an accurate understanding of the earth while making policy decisions, especially from the perspective of ongoing rapid climate change. Remote sensing techniques are especially useful in maintaining an inventory of soil moisture or water bodies, identifying crop types and health, and estimating crop yield.
There is now a growing call for organisations to balance profit and social impact. The applications of these technologies are relevant to several industries, but nowhere has it generated the kind of impact as it has in the field of sustainable agriculture and forestry. How we produce our food has a big impact on Earth’s resources. Data gathered from satellites and ground sensors help us in understanding our planet better and, as a result, it helps in supporting our farmers to better plan their planting, nurturing, harvesting, and selling their produce.
There are a plethora of satellite images available in the past decade that are at various levels of spatial and temporal resolution, with limited coverage. However, in the year 2017, the private Earth Imaging company, Planet Labs, deployed a flock of satellites on an Indian Space Research Organisation (ISRO) rocket. They help scan the Earth daily, at high spatial resolution (3-5m), with wider geographical coverage. They provide satellite-imagery-based information for applications that include, among others, tracking crop yields and health, managing agriculture and natural resources for civil governments, emergency and disaster management, and also helping detect deforestation, illegal mining, or other changes in the ecological landscape for conservation and sustainability. We also recently started our partnership with Planet Labs for piloting some of their data for our needs.
Presently, the biggest impact is in the area of conservation. Satellite data has proven beneficial in watershed management and is used in India for basins of Krishna, Yamuna, and Tapi river. A national-level project, Integrated Mission for Sustainable Development (IMSD), undertaken by the Department of Space, covered an area of about 84 million hectares spread over 175 districts in India. In selected watersheds under the project, the implementation of rainwater harvesting demonstrated a number of benefits, one among them being the increase in the agricultural development of once-barren regions in the area.
CropIn has combined satellite imagery with machine learning (ML) algorithms in order to strategize and implement a large-scale river conservation project in Central India. River conservation projects are of paramount importance in agriculture-based economies like India, where farmers are moving away from their dependence on uncertain weather conditions and infrequent rains. Rivers are the lifeline for these farmers and are a primary source of water for irrigation. However, river conservation projects have to be well planned and timed. These can also become increasingly expensive. It is at this point that technology plays an indispensable role, as it offers a system that can monitor activity along the river basin, encourages farmers to continue adopting sustainable practices, and gives confidence to policymakers to continue efforts.
Investigating and comparing the past historical data with the current data coming from the satellite imagery of these locations allowed us to investigate the impact of tree crop plantation along the river basin. The plan took into account the river basin boundary along with the plantation plan to monitor changes in water capacity, tree density, drought, and precipitation over a four-year period. Measuring water capacity alone is an inaccurate analysis of impact since this might be due to short changes in the weather conditions or might be due to the intrinsic characteristic of that particular geographical zone. The algorithm monitored changes in each of these elements over the period of time to determine the impact of tree plantation on agricultural crops in the area under study.
Remote sensing also helped in providing a faster and more accurate measure to evaluate the impact of these conservation projects. Traditionally, it would take, probably, several decades to ascertain the impact.
There are several indices and derivatives that can be obtained from the optical imagery other than the popular NDVI(normalised difference vegetation index) for vegetation health. CropIn used derived features from the satellite data, ground sensor data, weather data, and custom-built an ML model on top of these to continually monitor over time and analyse the tree density change over the four years along the river basin over the 15 districts.
The tree density change is estimated using a satellite-imagery-based aggregation model whose results are then correlated with the plantation statistics accrued during the plantation drive to estimate regions where the plantations survived and grew over the years. In these identified regions, further analysis is carried out to find out the factors that majorly contributed to surface water change. Surface water change in itself is quantified using the historical satellite data available from the region. It involved detecting the river extent, followed by measuring the surface area change using satellite derived indices. An ML model is built to assess the change in the surface water area and water retention due to the following contributing factors: a) change in tree cover, b) precipitation change, and c) drought conditions. High-resolution imagery of the area under study further validated these results. We can also determine the increase or decrease in the cultivated land over time, giving an indication of the correlation between tree density increase, water levels, and spur in agriculture activity in the area. There are some open challenges with change detection of the water levels, the agriculture activity, and tree density. One important issue is the effect of confounding variables linked to seasonal, annual, and long-term precipitation changes in the region. The effects of these confounding variables have to be removed during the analysis to measure the real impact.
Another case study is the performance evaluation of the Mahi Right Bank Canal project in Gujarat. This evaluation, in turn, helped in crop inventory, generation of vegetation spectral index profiles, and estimation of crop evapotranspiration.
The continual detection of changes in the region, thanks to the above methodologies, can reduce the time taken to calculate the ROI and impact of conservation projects. We believe that the insights and metrics can also provide government and policy bodies in prioritising and implementing projects with greater efficiency.
Reach out to us to find out more about how CropIn is using technology to help governments and development agencies fight climate change with smarter and more sustainable agriculture.
Watch this video to learn about Land Use Analysis powered by SmartRisk®.
 CPIC releases new blueprints to boost investment in nature conservation. Global Environment Facility, 2020.
 The Fourth Industrial Revolution. Deloitte, 2020
 Disaster Management. Geospatial World, 2010
 Ray, Shibendu & Dadhwal, Vinay & Navalgund, Ranganath (2002). Performance evaluation of an irrigation command area using remote sensing: A case study of Mahi command, Gujarat, India. Agricultural Water Management. 56. 81-91.