Agumassie Genet, who is one of the MSc students associated with the PhD candidates, successfully defended his MSc thesis entitled “Retrieving Leaf Area Index for Maize and Rice Crop Yield Estimation from Earth Observation Satellites, the Case of Lake Tana Sub Basin”. See the abstract of his thesis below.
Abstract
The monitoring of crop growth conditions and production estimate is important for the economic development of any nation. The conventional methods of crop monitoring are time-consuming, expensive and unable to cover large geographical areas. Hence, remote sensing data integrated with ground measurement has been used as a potentially valuable tool to extract biophysical variables like leaf area index (LAI), biomass and phenology, which are essential inputs in crop monitoring and yield estimation. The objective of this study was to estimate crop yield of maize and rice crops from high resolution leaf area index that retrieved from Sentinel 2 and Landsat 8 OLI imageries, and ground leaf area index measurements in Lake Tana sub basin. A regression equation was implemented to estimate leaf area index from Landsat-8 OLI and Sentinel 2 derived normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI) and the two band enhanced vegetation index (EVI2). The results showed that the LAI derived from Sentinel 2 exhibited strong nonlinear relationships with leaf area index measured on the ground for the two crops with RMSE less than 0.52m2 /m-2 except NDVI for rice (0.78m2 /m-2). Sensitivity analysis between the three vegetation indices and green leaf area index shows NDVI has the highest sensitivity to green leaf area index than others at the early growth stage of both crops. Leaf area index estimated from sentinel 2 using NDVI and EVI2 at the crop reproductive stage showed a strong correlation with rice yield with a coefficient of determination of 0.77 and 0.70, respectively. The overall results provide insights on the feasibility of multispectral data particularly, sentinel-2 with 10meter resolution for estimating high resolution leaf area index in a very fragmented agricultural landscape like Ethiopia. Thus, large scale crop monitoring and yield prediction activities can be done accurately through remote sensing data with the use of ground LAI. However, it has to be tested under different environmental conditions before being applied on a larger scale with limited field data.
Keywords: Crop monitoring, Leaf area index, Crop yield, Vegetation Indices, Multispectral image
