Assessing the Spatio-Temporal Controls of Field-Scale Nitrogen Uptake Using Biophysical Modeling and High Resolution Satellite Imagery.

Crop nitrogen (N) uptake can vary widely at the field scale due to heterogeneous landscapes, where multiple variables including terrain, local water content, soil type, temperature, and microbial activity impact the efficiency with which water and nutrients in the soil contribute to biomass production. Due to this inherent heterogeneity, precision N management requires a greater understanding of the spatio-temporal variability of factors that affect crop N use efficiency (NUE) and water use. Remote sensing of plant greenness (a result of both leaf area and chlorophyll content) using NDVI is the dominant approach for field-scale site-specific management of aboveground N content; however, by validating high temporal (~10 days) and spatial resolution (5 m) resolution satellite imagery (RapidEye) with extensive field campaigns, our results suggest that mapping aboveground nitrogen can be improved using a red-edge spectral band (R² = .72, RMSE = 20.5 kg/ha N). From this, using a N balance approach (accounting only for N applied at planting), we mapped nitrogen uptake (kg/ha) at different growing stages throughout the 2012 and 2013 growing season. The importance of water stress on crop N uptake is identified using spatially distributed (10 m) biophysical models (CropSyst-Microbasin and the Soil Moisture Routing model) parameterized using detailed soil characterization and terrain data.  These models, applied at daily and hourly timesteps, distinguish both regions of high water stress such as clay-rich, eroded summit positions as well as regions where the topographic redistribution maintains adequate supplies of water late into the growing season. This modeling provided the opportunity to determine the extent to which topography, variability in soils, and climate drive both the temporal and spatial distribution of soil water and the extent to which this controls crop nitrogen uptake patterns. In addition to high resolution satellite imagery, we compared results with courser resolution Landsat imagery (30 m), to assess the viability of freely available data for nutrient management in the topographically complex fields of the Palouse region in Northern Idaho and Eastern Washington.