Applications of Weather Data, Satellite Data, and Other Geospatial Data for Improving Crop Insurance and Agricultural Risk Management
Loading...
Date
2021-03-31
Authors
Roznik, Mitchell
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This dissertation consists of three essays investigating improvements to crop insurance by applying weather data, satellite data, and other geospatial data. The first essay investigates the use of weather data and satellite data to improve a temperature-based crop insurance policy in Alberta, Canada. The analysis used advanced methods from geostatistics, including Universal Kriging and Generalized Additive Models. Results suggest that the more advanced interpolation methods reduced interpolation error and provided a more useful measure of temperature for insurance policies.
The second essay evaluates a number of traditional crop yield updating methods and also proposes a new updating method that may be used to improve crop insurance. For illustration purposes, farm yield data is used from 1133 canola farms in Alberta, Canada from 2002 to 2017 (16 years) and national yield data from Statistics Canada are used to evaluate the yield trend updating methods. Crop yield updating methods are needed because improvements in technology have led to higher crop yields over time. Lower past yields need to be updated at higher levels to bring them up to the level of today's technology, otherwise the producer may be under insured. The updating approaches investigated are 10-year average (model A), national yield trend (model B), county yield trend (model C), and a new mixed method (model D). The results suggest that all four crop yield updating methods performed similarly. This analysis may be of interest to crop insurance analysts and policy makers in Canada, the United States, and other countries.
The third essay investigates if higher satellite resolution can improve the accuracy of crop yield estimation. In the analysis, the years 2008-2018 of NASA's MODIS satellite image collection over the contiguous United States were examined for four crops: corn, soybeans, spring wheat, and winter wheat. Crop yields were regressed on a vegetation index (NDVI) at three satellite resolution levels (i.e., 1km, 500m, and 250m). Improvements in satellite resolution increased the relative accuracy of the crop yield estimation by improving the quality of the vegetation index (NDVI) measurements. Further improvements in yield estimation accuracy may be expected in the future as satellite resolution improves.
Description
Keywords
Satellite Data, Weather Data, Big Data, Risk Management, Crop Insurance, Geospatial, Agriculture, Kriging, Bayesian Regression, Remote Sensing
Citation
Roznik, Mitchell, C. Brock Porth, Lysa Porth, Milton Boyd, and Katerina Roznik. "Improving agricultural microinsurance by applying universal kriging and generalised additive models for interpolation of mean daily temperature." The Geneva Papers on risk and Insurance-Issues and practice 44, no. 3 (2019): 446-480.