Impact of Satellite derived Vegetation Condition on Crop Yield Prediction in Eastern Ontario

Abstract:

Crop condition derived from Earth observation satellites has been used as a tool to track the impacts of variable weather and climate for decades on agriculture production, primarily using optical vegetation indices such as the Normalized Difference Vegetation Index (NDVI). The simplicity of this index has permitted its relatively easy integration into national monitoring systems, with observations collected over multiple decades using satellites such as AVHRR and MODIS, which provide frequent temporal coverage at national and global scales. NDVI has been shown to be well correlated with crop yields at the regional level in the Canadian Prairies (Mkhabela et al., 2011) and it has been used to operationally forecast crop yields within the Canadian Crop Yield Forecast (CCYF) system in combination with meteorological indices (Chipanshi et al., 2015). As the ability to model crop yields at finer spatial scales improves, the need for more precise information on vegetation condition becomes more crucial. The NDVI index suffers from numerous deficiencies, including saturation at high canopy biomass, confounding responses to both crop density and chlorophyll content and impacts of cloud cover. The use of indicators that more accurately represent measurable vegetation characteristics can create models that allow better identification of the precise impacts of weather or other factors on plant physiological processes (Liu et al, 2012). The objective of this study is to compare the use of alternative indicators to NDVI in both biostatistical and process based crop models to examine impacts on crop yield modelling at local and regional scales in eastern Ontario corn crops over the period of 2017 to 2019. These are compared to measured crop yields at both local (field) and regional scales.