PNNL

Abstract

Sequestration by injection and deep geologic storage is a promising mitigation tool for reducing atmospheric carbon releases. Cost-effective and reliable containment verification methods are necessary to minimize local environmental impacts. One potential monitoring tool is spatiotemporal mapping of vegetation through remote sensing. For the detection and assessment of future impacts, we define a baseline crop response from 26 years (1986-2011) of Landsat data for 400 km2 surrounding a potential injection site near Jacksonville, Illinois. Crop condition was quantified through the normalized difference vegetation index (NDVI) calculated near peak response. The normal score transform (y_NDVI) was applied to each field to accentuate spatial patterns and normalize differences due to planting times. We tested crop type and soil moisture (Palmer Crop Moisture Index (CMI)) as explanatory variables. Spatial patterns in y_NDVI were similar between corn and soybeans—the two major crops. Linear regression between y_NDVI and the cumulative CMI (CCMI) for the growing season showed complex interactions between crop condition, topography, soils, and annual moisture. Wet toposequence positions (depressions) were negatively correlated to CCMI and dry positions (crests) positively correlated. However, only 40% of the landscape showed a statistically significant linear relationship. Accordingly, we defined a univariate tolerance interval to identify fields with non-typical y_NDVI responses. The technique was tested on an independent 2013 image and 63 fields with potentially anomalous responses were identified out of a possible 1,483. In addition, the correlation between y_NDVI and CCMI has broad application for long term crop performance monitoring and edaphological soil mapping.