Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Etats Unis → 2015 → Estimating Nitrogen Content of Dryland Wheat Fields Using Landsat Imagery

University of Idaho (2015)

Estimating Nitrogen Content of Dryland Wheat Fields Using Landsat Imagery

Finch, Sam Louis

Titre : Estimating Nitrogen Content of Dryland Wheat Fields Using Landsat Imagery

Auteur : Finch, Sam Louis

Université de soutenance : University of Idaho

Grade : Master of Science (MS) 2015

Inefficient use of synthetic nitrogen fertilizers on wheat fields increases costs of production and can cause numerous undesirable environmental consequences. Specifically, increasing plant nitrogen use efficiency may decrease agricultural sources of nitrous oxide, nitrate leaching, and soil acidification, while maintaining crop yield and quality. A main cause of inefficient nitrogen fertilizer application in industrial agricultural systems is a dearth of information about in-field nitrogen variability in crops. The objective of this study was to determine whether free, publicly available Landsat imagery could be used to predict N content in wheat plants across a spatially heterogeneous landscape. We derived field-scale N content by extrapolating plot-level destructive samples using commercially available Rapid Eye satellite data over five farms located in the Palouse of eastern Washington and northern Idaho, USA, in 2012 and 2013. We then used these results to compare with Landsat-derived Normalized Difference Vegetation Index to predict N content and map spatial variability at peak biomass. Statistical analyses of modeled versus observed N content show that Landsat-derived NDVI can predict N content of wheat crops to within 13 to 28 kg/ha of the N content map predicted using calibrated RapidEye data. We found statistically significant relationships between predicted N content and NDVI, with R2 values ranging between 0.20 to 0.82 depending on the field site. NDVI may be inferior in estimating crop N content relative to other vegetation indices (e.g. the normalized difference red-edge index) not calculable using Landsat data. However, because Landsat NDVI could be used to describe significant amounts of variation in crop N content, these data are likely to be useful for improving N fertilizer management (e.g. through establishing fertilizer zones) at the field scale. When considered in conjunction with Landsat data being publicly available and globally relevant, Landsat-derived NDVI could be a useful and cost-effective tool for improving N management of wheat crops worldwide

Mots clés : Agronomy, Landsat, N content, Nitrogen fertilizer, Environmental science, Remote sensing, Health and environmental sciences, Earth sciences, Biological sciences, Nitrogen, NDVI


Page publiée le 19 novembre 2015, mise à jour le 21 décembre 2017