LB 15-137 - Linking gross ecosystem production with remote sensing in spring wheat-corn-soybean rotations

Thursday, August 10, 2023

5:00 PM – 6:30 PM PDT

Location: ESA Exhibit Hall

Presenting Author(s)

NS

Nicanor Z. Saliendra

Ecologist
USDA ARS Northern Great Plains Research Laboratory
Mandan, North Dakota, United States

Coauthor(s)

ML

Mark Liebig

USDA ARS Northern Great Plains Research Laboratory, North Dakota, United States
CW

Craig Whippo

USDA ARS Northern Great Plains Research Laboratory, United States
DA

David Archer

USDA ARS Northern Great Plains Research Laboratory, United States

Background/Questions/MethodsAs participants in the Long-Term Agroecosystem Research (LTAR) – Croplands Common Experiment, we compared business-as-usual (BAU) and aspirational (ASP) practices in two fields ( >20 ha) near Mandan, North Dakota, USA. The ASP practice included cover crops, which were presumed to alter carbon fluxes relative to the BAU practice without cover crops. Currently, our understanding of carbon flux dynamics in semiarid cropping systems is limited, as is the capacity of remote sensing tools to quantify such dynamics in a region known for highly variable growing conditions. Therefore, our objective was to develop empirical relationships with ground-based GEP (gross ecosystem production) as a function of satellite-based NDVI (normalized difference vegetation index) in spring wheat, corn, and soybean grown with or without cover crops under no-till, rainfed practices in a semiarid environment. We quantified GEP in both BAU and ASP fields using eddy covariance methods from 2016 to 2022 in conjunction with remote sensing acquired from satellite. We calculated the NDVI, surface temperature, and normalized vegetation water index (NDWI) using Analysis Ready Data from reflectance images of Landsat. Results/ConclusionsWe found a strong, positive linear relationship (R2 = 0.85, P < 0.0001, n = 170) between GEP and NDVI across ASP and BAU fields when 85 non-cloudy, clear Landsat scenes were available through seven growing seasons. However, detailed analyses of GEP-NDVI seasonal dynamics for each field revealed the relationship to be linear from start-of-season to peak-growth and non-linear (quadratic) from peak-growth to end-of-season. Thus, the estimation of ground-GEP with satellite-NDVI as its estimator must involve both linear and non-linear empirical models for the entire growing season. Additionally, we found canopy surface temperature from late-July through September was 1.5 °C cooler and NDWI was 60 % higher in the ASP than BAU field during a drought year in 2020 when both GEP and NDVI were higher in corn inter-seeded with cover crops than corn without cover crops. Results from this study emphasize the importance of remote sensing in the modeling and upscaling of carbon fluxes from the field to the landscape scale in semiarid cropping systems.