Page 233 - India Insurance Report 2023- BIMTECH

India Insurance Report - Series II 221

Sentinels have modernised Europe’s agricultural data collection and assessment systems, reducing
field inspections and enhancing the efficiency of payments to farmers who suffered crop losses. Some
examples are:

(i) The agricultural monitoring system in Poland is based on the machine learning techniques applied
to Sentinel data to identify and monitor crops.

(ii) COVID-19 Earth Observation Dashboard was developed by the National Aeronautics and Space
Administration (NASA), European Space Agency (ESA), and Japanese Aerospace Exploration Agency
(JAXA) to monitor the impacts of COVID-19 on the agricultural situation worldwide.

(iii) The wealth of data from multiple satellites is being used to track crop planting and harvest patterns,
and crop health progression now influences the export policies of agricultural commodities by
many developed economies.

3. Indian Perspective –Innovative and Sustainable Solutions Available for India

i. Using Hyper-Local Weather Data : Uptake of crop insurance in India would significantly benefit
from weather data availability at the scale of individual villages (2x2 km). This new data can replace
the current data available at a 25x25km area covering multiple numbers of villages and thus can
remove the significant amount of basis risk involved therein, leading to customer dissatisfaction.

Sophisticated machine learning algorithms are now capable of assimilating data from the available
ground weather stations, atmospheric reanalysis, and polar and geostationary satellites, producing not
just rainfall and temperature but also other variables such as humidity, solar radiation and wind speed.

These new data sets properly sanitized are already available for a long period spanning more than 20
years and offer exciting opportunities for creating new parametric insurance products that target
specific crops and specified perils over much smaller unit areas, say a village.

ii. Availability of Scientifically Modelled Historical Crop Yield Data :The lack of reliable crop yield data
has been one of the biggest limitations for expanding and sustaining the crop insurance program in
India. Estimates from manual Crop Cutting Experiments are gradually being found to be unreliable
and do not capture the spatial variations necessary for setting proper premium rates at smaller
insurance units, say a village. The advent of remote sensing coupled with ecosystem modelling now
allows us to estimate crop yields at spatial scales as fine as 10m for all major crops.

The Radiation-Use-Efficiency yield models integrate weekly satellite data and weather data to estimate
gross primary production, which is then converted to crop yield. A limited number of protocol-
based CCE results help in calibrating the relationship between gross primary production and crop
yields. Using the RUE models, historical crop yield data have been generated at the scale of 2x2 km
(say at the village level) for all major crops for the last 23 years from 2001-2022. These crop yield
data offer exciting possibilities for actuarial analysis and help in setting proper premium rates.

iii. Availability of Geostationary Satellite Data : Persistent cloudiness during the kharif season reduces
the opportunities for observing landscapes from orbiting optical satellites. While orbiting radar