Remote sensing application in biomass crop production systems in Oklahoma

This study was conducted to evaluate the combined effects of nitrogen and cropping systems on biomass yield and quality and to describe the spatial variation of biomass yield, soil carbon and nitrogen within a switchgrass field. Field plots at Stillwater and Woodward in Oklahoma consisting of five nitrogen treatments and three cropping systems were used for the nitrogen x cropping system study and an 8 ha switchgrass field at Chickasha, Oklahoma was used to describe the spatial variability at fine (2.5 m sampling distance) and coarse scale (10 m sampling distance). Remote sensing technique was used to monitor biomass yield and quality to better understand N requirement and usage for production. Semivariogram were used to evaluate spatial variability of the soil parameters and biomass yield. The results of this study showed that maximum yield was produced at both locations with less than 84 kg N ha 1 and high biomass sorghum has potential to produce biomass yield > 20 Mg ha 1 under normal conditions in Oklahoma. The study results also showed that perennial grass systems are more reliable sources of biomass yield, especially under adverse climatic conditions of Oklahoma. Final biomass yield of high biomass sorghum could be predicted using both broadband (aerial photograph) and narrowband (GreenSeeker) normalized difference vegetation index (NDVI) from July to close to harvest, while biomass yield in the perennial grass was best predicted during June to July. Comparing simple ratios and best narrowband indices with partial least square regression (PLSR) models suggested that while PLSR calibration models produced significantly lower error and higher r2 for predicting biomass yield and N concentration within a growing season, the simple ratios and best narrowband indices were more stable and reliable when used for prediction across growing seasons. Spatial pattern in switchgrass field was described using both ground and aerial imagery. The NDVI computed from aerial imagery provided good precision at the fine scale in describing the spatial distribution of switchgrass yield. Remote sensing application in biomass production systems can greatly improve prediction models for predicting biomass yield and quality in feedstock materials with use of optimal hyperspectral narrowband.