Crop Yield Estimation Using NDVI: A Comparison of Various NDVI Metrics

Loading...
Thumbnail Image
Date
2021-03
Authors
Wilton, Michael
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The objective of this study is to examine and compare multiple Normalized Difference Vegetation Index (NDVI) metrics for estimating crop yield. There are several terms used to describe NDVI metrics including: process methods, aggregation techniques, phenological indices, and aggregation metrics. The various NDVI metrics included in this study are maximum NDVI (MaxNDVI), integrated-NDVI (INDVI), Minimum NDVI (MinNDVI), relative annual range of NDVI (RREL), days to maximum NDVI (DTM), and days from maximum NDVI (DFM). NDVI data was accessed from the NASA Moderate Resolution Imaging Spectroradiometer (MODIS) over a 13-year period (2006-2018). County level corn yield data was from the United States Department of Agriculture (USDA) National Agriculture Statistics Service (NASS) database. Temperature data was gathered from the Puget Sound Regional Synthesis Model (PRISM). Regression analysis was conducted to examine the performance of various NDVI metrics for estimating crop yield. The results indicate that MaxNDVI is best able to estimate county level corn yields. This research aids in understanding the ability of the various NDVI metrics to estimate crop yield. This information will assist those designing satellite-based crop yield forecasting and index-based crop insurance models.
Description
Keywords
Crop Yield Estimation and Forecasting, Index-based Crop Insurance, Satellite, Remote Sensing, MODIS, Normalized Difference Vegetation Index (NDVI), NDVI Aggregation, NDVI Processing Methods, Maximum NDVI (MaxNDVI), Integrated-NDVI (INDVI)
Citation