Intercomparison of same-day remote sensing data for measuring winter cover crop biophysical traits

Authors: PERABHAKARA, K - University Of Maryland; THIEME, A - University Of Maryland; LAMB, B - Us Geological Survey (USGS); Jennewein, Jyoti; McCarty, Gregory; HIVELY, W - Us Geological Survey (USGS)

Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/28/2024
Publication Date: 4/6/2024
Citation: Perabhakara, K., Thieme, A., Lamb, B., Jennewein, J.S., Mccarty, G.W., Hively, W. 2024. Intercomparison of same-day remote sensing data for measuring winter cover crop biophysical traits. Remote Sensing. 24(7):2339. https://doi.org/10.3390/s24072339.
DOI: https://doi.org/10.3390/s24072339

Interpretive Summary: Winter cover crops have proven to be an effective method to reduce sedimentation and nutrient runoff into waterways, and to increase environmental benefits including improved soil health and water quality and remote sensing has been demonstrably successful as a tool to monitor winter cover crop growth and performance, To assess cover crop vigor (green fractional cover and biomass) we used normalized difference vegetation index (NDVI) which is comprised of a normalized difference of visible red and near infrared reflectance values but there are some measurement uncertainties associated with the needed spectral measurements. This study assessed whether same-day measurements among satellite platforms are sufficiently correlated with each other when adjusted for spatial resolution differences. Satellite imagery was corrected to surface reflectance and was compared to top of atmosphere reflectance across visible and near-infrared bands. The results of this study show that all sensors were highly correlated with each other for both visible and near-infrared bands and that surface reflectance NDVI was highly correlated with percent vegetative groundcover and biomass and aligned well with predictive equations developed with similar datasets. This information will reduce uncertainties in measurement of winter cover crop performance when using satellite spectral data.

Technical Abstract: Accurate estimation of winter cover crop biomass and percent vegetative groundcover is essential for assessment of environmental benefits. Winter cover crops are planted during the fall and achieve higher ground cover and biomass towards the following spring, with winter severity having a significant impact on biomass accumulation throughout the fall-to-spring season. Remote sensing serves an in ideal tool for monitoring cover crop performance, particularly over large spatial extents . In addition to remote sensing imagery, advances have been made in the use of proximal sensors integrated with GPS for on-field measurements. The comparability of such measurements between platforms, as well as between processing levels is important, but such assessments have been limited for cover crop applications. Theoretically, ground-based measurements contain minimal atmospheric interference and should compare accurately with measurements from satellites that have been corrected to surface reflectance. This research effort examines the relationships between SPOT-5, Landsat 7, WorldView-2 satellite imagery with handheld proximal sensors on two days during the 2012-2013 winter cover crops season. Satellite imagery was corrected to surface reflectance and was compared to top of atmosphere reflectance across visible and near-infrared bands, and for the normalized difference vegetation index (NDVI). Surface reflectance products were more correlated with proximal sensors compared to top of atmosphere, with intercepts closer to zero, regression slopes nearer to the 1:1 line, and less variance between measured values for band reflectance and NDVI.Additionally, all instruments and processing levels were compared to percent green vegetative groundcover and in situ biomass measurements. Surface reflectance NDVI was highly correlated with percent vegetative groundcover and biomass and aligned well with predictive equations developed with similar datasets.