Hyperspectral remote sensing of cover crop traits for decision support in precision nitrogen management

Spaceborne hyperspectral sensors offer opportunities to quantify crop traits and inform variable rate nitrogen management, yet their operational use remains limited. We evaluated PRISMA and EMIT imagery across four winter cover crop seasons in Maryland, U.S., to predict traits required by the open source Cover Crop Nitrogen Calculator (CC-NCALC) decision support tool. Traits included aboveground biomass, nitrogen concentration (N%), non-structural carbohydrates (NSC%), holocellulose (HCL%), and lignin (LIG%). Using these traits as inputs, the CC-NCALC estimates nitrogen credits from cover crops based on decomposition dynamics informed by trait data, soil, and weather conditions. Nitrogen credits are the estimated amount of plant available nitrogen that a cover crop contributes to the subsequently planted cash crop. Trait models were developed using partial least squares regression with several cross validation strategies. Biomass ( R 2 = 0.70–0.87; RMSE = 661–951 kg ha −1 ), NSC% ( R 2 = 0.69–0.72; RMSE = 4.53–7.63%), and HCL% ( R 2 = 0.65–0.70; RMSE = 5.04–7.50%) predictions were strong. But HCL% and NSC% predictions degraded under cross-scene and cross-species validation ( R 2 = 0.47–0.54; RMSE = 6.06–9.18%). N% ( R 2 = 0.15–0.79; RMSE = 0.46–0.74%) and LIG% ( R 2 = 0.02–0.35; RMSE = 1.22–1.83%) predictions were more variable; predictions using EMIT were more accurate than predictions using PRISMA, but both approaches demonstrated the need for further increases in accuracies of retrieval approaches for these two traits. Spring trait maps from April 2024 were ingested into the CC-NCALC to estimate nitrogen credits, which ranged from −14.58 to 45.65 kg N ha −1 , with higher values associated with greater legume biomass. Inter-scene comparisons demonstrate EMIT's increased spectral fidelity and trait prediction accuracy compared to PRISMA. Results demonstrate that integrating spaceborne hyperspectral imagery with decision support systems can generate site-specific nitrogen credit maps, advancing hyperspectral remote sensing toward operational nutrient management. Future missions will be critical to help scale these approaches for routine on-farm use.

Authors

Institutions

Publication Details

Journal
Remote Sensing of Environment
Published
2026-11-13
DOI
https://doi.org/10.1016/j.rse.2026.115759
Primary Topic
Remote Sensing in Agriculture
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Hyperspectral remote sensing of cover crop traits for decision support in precision nitrogen management

Brian T. Lamb, Wells Dean Hively, Jyoti S. Jennewein, Steven Brian Mirsky et al.
2 links
Remote Sensing of Environment
Remote Sensing in Agriculture
article

Hyperspectral remote sensing of cover crop traits for decision support in precision nitrogen management

Brian T. Lamb, Wells Dean Hively, Jyoti S. Jennewein, Steven Brian Mirsky, S. Chris Reberg-Horton
article en

Abstract

Spaceborne hyperspectral sensors offer opportunities to quantify crop traits and inform variable rate nitrogen management, yet their operational use remains limited. We evaluated PRISMA and EMIT imagery across four winter cover crop seasons in Maryland, U.S., to predict traits required by the open source Cover Crop Nitrogen Calculator (CC-NCALC) decision support tool. Traits included aboveground biomass, nitrogen concentration (N%), non-structural carbohydrates (NSC%), holocellulose (HCL%), and lignin (LIG%). Using these traits as inputs, the CC-NCALC estimates nitrogen credits from cover crops based on decomposition dynamics informed by trait data, soil, and weather conditions. Nitrogen credits are the estimated amount of plant available nitrogen that a cover crop contributes to the subsequently planted cash crop. Trait models were developed using partial least squares regression with several cross validation strategies. Biomass ( R 2 = 0.70–0.87; RMSE = 661–951 kg ha −1 ), NSC% ( R 2 = 0.69–0.72; RMSE = 4.53–7.63%), and HCL% ( R 2 = 0.65–0.70; RMSE = 5.04–7.50%) predictions were strong. But HCL% and NSC% predictions degraded under cross-scene and cross-species validation ( R 2 = 0.47–0.54; RMSE = 6.06–9.18%). N% ( R 2 = 0.15–0.79; RMSE = 0.46–0.74%) and LIG% ( R 2 = 0.02–0.35; RMSE = 1.22–1.83%) predictions were more variable; predictions using EMIT were more accurate than predictions using PRISMA, but both approaches demonstrated the need for further increases in accuracies of retrieval approaches for these two traits. Spring trait maps from April 2024 were ingested into the CC-NCALC to estimate nitrogen credits, which ranged from −14.58 to 45.65 kg N ha −1 , with higher values associated with greater legume biomass. Inter-scene comparisons demonstrate EMIT's increased spectral fidelity and trait prediction accuracy compared to PRISMA. Results demonstrate that integrating spaceborne hyperspectral imagery with decision support systems can generate site-specific nitrogen credit maps, advancing hyperspectral remote sensing toward operational nutrient management. Future missions will be critical to help scale these approaches for routine on-farm use.

Remote Sensing of EnvironmentVol. 349
Agricultural Research Service (US), United States Department of Agriculture (US), North Carolina State University (US), Beltsville Agricultural Research Center (US), Lower Mississippi-Gulf Water Science Center (US)
National Aeronautics and Space Administration, National Institute of Food and Agriculture
Zero hunger
Remote Sensing in Agriculture
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon