The feasibility of satellite remote sensing and spatial interpolation to estimate cover crop biomass and nitrogen uptake in a small watershed

Abstract

The adoption of winter cover crops has been identified as one of the most effective best management practices to reduce nonpoint nitrogen (N) loss via subsurface drainage in a watershed in the midwestern Corn Belt. Understanding the variation of cover crop growth and N cycling is vital for watershed modeling efforts that simulate cover crop adoption. However, there is a dearth of watershed cover crop studies that describe the variation in cover crop growth and N cycling and compare the ability of spatial analytical methodologies to predict cover crop biomass and N uptake within diverse agronomic management practices and heterogeneous soil landscapes. Therefore, the objective of this study is to compare the feasibility of satellite remote sensing and spatial interpolation methods to predict cover crop biomass and N uptake in a small watershed (100 to 10,000 ha). This study was undertaken during 2015 to 2017 in the Lake Bloomington watershed (374 ha) in McLean County, Illinois. Within the small watershed, daikon radish (Raphanus sativus L.)/oats (Avena sativa L., R/O), annual ryegrass (Lolium multiflorum)/daikon radish (A/R), and cereal rye (Secale cereale L.)/daikon radish (C/R), were adopted on 78% of row crop land area for both years. Strong linear relationships were observed between soil adjusted vegetation index (SAVI), enhanced vegetation index (EVI), and triangular vegetation index (TVI) and cover crop biomass, (average R2 of 0.77, 0.78, and 0.76, respectively) and N uptake (average R2 of 0.68 for all three vegetation indices). Cover crop biomass and N uptake estimated by the SAVI method were 86% and 85%, 93% and 94%, and 107% and 98% of the ground observed value during 2016 spring, 2016 fall, and 2017 spring, respectively. Moreover, spatial pattern and different cover crop management fields in the study watershed were also captured by the SAVI method. Ordinary kriging (OK) and inverse distance weighting (IDW) showed similar mean cover crop biomass and N uptake values for fields with cover crops; however, both spatial interpolation methods showed lower prediction R2 values than that of the SAVI method. The results of this study suggest that the combination of spatially accurate satellite imagery and limited ground sampling could be used for repeated small watershed assessment of cover crop growth. Furthermore, this can be used to understand cumulative cover crop impacts on soil and water quality in response to conservation practices and weather within a drainage basin.