Explore IWMI’s journey using Earth Engine for climate crisis and food security in South Asia
By Giriraj Amarnath, Principal Researcher, Disaster Risk Management and Climate Resilience, and Surajit Ghosh, Regional Researcher, Water Risk and Data Sciences Specialist, International Water Management Institute
When a flood is approaching or a drought is on the horizon, decision makers in planning, water, agriculture, and disaster management departments need to closely monitor the situation and quickly act. Satellite imaging technology can shed light on the impact of flooding and drought, but access to the technology has been limited to people with powerful computers and high-speed internet access, given the size of satellite image files.
How can decision makers on the ground get fast access to satellite images when they really need them, even if they don’t have the latest and greatest computers to store and process them? In our work using Earth Engine at the International Water Management Institute (IWMI), through the Climate Change, Agriculture and Food Security (CCAFS), and the Water Land and Ecosystems research programs of CGIAR and Indian Council of Agricultural Research (ICAR), we’re realizing just how versatile Earth Engine can be for water management decision makers around the world.
We combine remote sensing, ground observations, and socioeconomic data to create knowledge products for applications in agriculture, water scarcity, and security. Among our early offerings are a crop harvest- and weather-monitoring tool, a drought-monitoring package, a flood-mapping product, and a global map of surface water bodies. Although these products have only recently been ported into Earth Engine, IWMI bases them on years of work developing similar offline products.
Underpinning food security during the pandemic
We developed the crop harvest- and weather-monitoring tool to help authorities manage food production during the pandemic. As Covid-19 began to spread, state governments in India needed to find ways to minimize the movement of people without interrupting food supplies.
By combining indices with weather maps, planners were able to clearly identify crops that were in imminent need of harvesting ahead of storms, and to organize farmworker logistics accordingly.
“We used the satellite images from Earth Engine to identify crop-growing areas that were ready for harvest, and to assist agencies and farmers in making timely decisions around sowing the rabi [winter] crop during the pandemic,” says Dr. P. S. B. Anand, Principal Scientist at the Indian Institute of Water Management of the Indian Council of Agricultural Research. “Our aim was to delay some crops arriving at the market, in order to reduce the throngs selling produce at any one time.”
Facilitating drought-planning at the regional level
In the past, our monitoring capabilities used multiple workstations for image processing and product development, which delayed the distribution of knowledge products. We solved this problem with Earth Engine. Our algorithm accesses both historical and near real-time multisource satellite data and advanced computation capabilities to produce drought severity maps in just a few seconds, and takes less work by users.
The new drought-monitoring product is an Earth Engine version of IWMI’s existing South Asia Drought Monitoring System, for which the Institute won the Geospatial World Excellence Award 2020. Through this system, authorities responsible for managing drought in Afghanistan, Bangladesh, Bhutan, Maldives, Nepal, India, Pakistan, and Sri Lanka can access 500 meter-resolution drought severity maps created using satellite data, and a 15-day drought forecast.
IWMI produces more than 30 drought indices by accessing Earth Engine and third-party data sources (for example, subseasonal forecast data from IITM ERPAS and ESA ASCAT) for drought monitoring and early warning, to power tools such as the Afghanistan Drought Early Warning Decision Support (AF-DEWS) supported by World Bank. The new technology has made the product widely accessible, including to users at the state level in India who are employing it to monitor and manage drought in their own regions. These frameworks can be scalable to any geography given the global access of Earth Engine data sources.
“The maps provide us with near real-time information on drought severity for rapid decision making,” says Dr. K. V. Rao, Principal Scientist at the Central Research Institute for Dryland Agriculture (CRIDA), also part of ICAR. “They’re utilized to monitor drought areas, issue advisories, and guide implementation of contingency measures.”
Using space technology to enhance flood management
The flood-mapping product draws on research conducted during development of IWMI’s existing Index-Based Flood Insurance for which the Institute won the 2020 GEO Sustainable Development Goals (SDG) Award. The product combines a hydrological model and 10 meter-resolution satellite images from ESA’s Sentinel-1 mission. Rainfall data for a particular catchment is added to the model, which shows how runoff will travel and collect.
Satellite images are used to verify the depth and duration of any flood, and identify farmers eligible for compensation. Insurance is just one application for the flood-mapping product, however. Organizations in India, Bangladesh, and Sri Lanka participated in capacity-building exercises on how to use IWMI’s Earth Engine offerings, and are now finding ways to customize the flood-mapping tool for their own use.
Hydrometeorological agencies in South Asia, in particular, plan to use IWMI’s Flood Insurance solution to identify where weather stations should be established. Department staff can query the flood-mapping product and see areas where rainfall regularly exceeds 250mm, and then examine locations of existing weather stations. If no station exists in an area of high rainfall, agency staff can plan to install one. Once the station is in position, it will send out a warning when precipitation reaches a particular level, so the team can prepare for flooding downstream of the site.
In the southern Indian city of Chennai, authorities are already using our surface waterbody and emergency response map to support flood management. Here, rapid development has caused wetlands to shrink, making managing floodwaters difficult. When staff examined surface waterbody maps for different years, the results clearly showed that natural water storage had declined over time. This information is now helping planners find a solution. One possibility is constructing small upstream reservoirs to store water in the way that natural ponds once did.
Going forward, the growth in cloud-based computing and open-access earth observation data will bring opportunities to develop new environmental applications. These can help to underpin collaboration between the public and private sectors — for example, through private-sector insurance companies working with meteorological departments and agricultural institutes to develop innovative crop insurance products.
Importantly, these low-cost technologies will directly benefit vulnerable smallholder farmers and communities by helping to mitigate and build resilience to the impacts of climate change.
