Satellite imagery has been widely promoted as a decision-making tool in agricultural production. Many agricultural software companies, crop consultants, and farm managers have integrated satellite imagery into their data analytics for production monitoring and management. Such data and analytics show promise for supporting field zone mapping, identifying relative variations in crop yield, and monitoring water use. There are tradeoffs between satellite platforms in terms of the scale and usefulness of the data collected depending on the application at hand. There are also cases where satellite imagery can be useful for mapping the impacts of extreme events such as flooding, drought, or derechos.
In June 2021, a 100-year flood event occurred in Southeast Arkansas. When such events happen, it is imperative that impact estimates are conducted as soon as possible to identify how and where to best support producers. The public imagery typically available for agricultural applications comes primarily from the NASA/USGS Landsat or EU Copernicus Sentinel missions. While imagery from these platforms comes at sub-monthly time-steps and relatively high spatial resolutions (30 and 20 meter pixels, respectively), it often isn’t available for several weeks after the images are taken due to processing. However, commercial imagery such as that produced by Planet Labs, offers opportunities for higher spatial resolution of 3-5 meters at near daily time-steps and becomes available in hours to days. When more than 15 inches of rain dropped over Southeast Arkansas, Planet Scope imagery was used to estimate the flood extent over cropland and illustrates the potential for more rapid cropland monitoring applications.
Figure 1 shows the color-infrared (CIR) imagery, flood extent, and heavy flooding (>1 ft) by crop type based on USDA NASS’s cropland data layer (CDL). Planet Scope records four bands of multispectral data for each pixel, including red (R), green (G), blue (B), and near-infrared (NIR). Many producers are familiar with the normalized difference vegetation index (NDVI), which is derived from NIR and R bands and corresponds to crop greenness. NIR reflects strongly when vegetation is present, and absorbs heavily when water is present. This relationship is very useful, as in this case, for mapping flooded fields. Based on the flood extent identified with thresholds for NIR, the CDL, and ground reference information provided by local county extension agents, there were 254,323 soybean, 190,150 rice, 54,817 corn, and 34,864 cotton acres estimated to have heavy flooding during this event in Arkansas. This information was generated within days of the flood event for the seven hardest hit counties. After a few weeks, impact estimates were generated for the broader area including 12 counties. Economic impacts from the event were estimated at approximately $60 million for corn, $6 million for cotton, $68 million for rice, $71 million for soybeans, and $1 million for wheat at an approximate total of $206 million.
Satellite imagery applications for agriculture are most commonly thought of for in-season monitoring and post-season assessment of maximum greenness to map field management zones. However, a lesser known but useful application of satellite imagery is for flood or natural disaster mapping. When flood events, derechos, or other natural disasters occur, estimating the extent of crop damage in a short timeframe is of utmost importance to best support recovery. We often rely on conversations and phone calls with county extension agents, producers, and crop consultants to gain an initial estimate of crop impacts. Satellite imagery can provide an additional tool for making damage estimates in those critical days and weeks following an event, especially when combined with on-the-ground conversations and validation. Planet and other satellite platforms will continue to play a role in improving and supporting agricultural production as wider access becomes available. This case represents one of many opportunities for satellite imagery to increase in its adoption and applications in crop production.
Figure 1: Estimates of Flood Extent and Acreage Impacts in Southeast Arkansas, June 2021.
Davis, Jason, and Aaron M. Shew. “Flood Impact Estimates on Cropland using Satellite Imagery“. Southern Ag Today 2(3.3). January 12, 2022. Permalink
