Wheat Yield Forecasting Based on Landsat NDVI and SAVI Time Series

Abstract

<jats:p>Due to increase demand of food grain in the world, assessment of yield before actual production is important in making policies and decisions in agricultural production system. For a large area, forecast models developed from vegetation indices derived from remote sensing satellite data possesses the potential to give quantitative and timely information on crops over large areas. Different vegetation indices are being made used for this purpose, however, their efficiency in estimating crop yield is needed to be certainly tested. In this study, wheat yield forecast was derived by regressing ground truthing yield data against time series of spatial vegetation indices for the 2013 to 2019 growing seasons. These spatial vegetation indices derived from Landsat 8 image data: Normalized Difference Vegetation Index (NDVI) and Soil Adjusted Vegetation Index (SAVI) were compared to evaluate the most appropriate index that performs better in forecasting wheat production at Karcag, Kunhegyes and Ecsegfalva settlements in J&amp;aacute;sz-Nagykun-Szolnok county, in the Northern Great Plain region of central Hungary. The best time for making wheat yield prediction with Landsat 8- SAVI and NDVI was found to be the beginning of ripening period (160th day of the year) with higher correlation between the vegetation indices and the wheat yield. The validation results revealed that the model from SAVI provides more consistent and accurate forecasts yield compared to NDVI. The SAVI model forecast yield for the validation years, 2018 and 2019 were within 6.00% and 4.41% of the final reported values while that of NDVI model were within 8.31% and 6.27%. Nash-Sutcliffe efficiency index is positive with E1= 0.99 for the model from SAVI and for NDVI, E1=0.57, which connote that the forecasting method developed and evaluated performs acceptable forecast efficiency.</jats:p>

Cereal yield forecasting in semi-arid region of Algeria using MODIS-NDVI

“…Interestingly, the strength of the relationship was similar or even stronger when compared to the findings of our study, with R²= 0.64 for barley, and R²= 0.643 for wheat. In a study on predicting the grain yields of wheat Adeniyi et al (2020), proves that the use of Normalized Difference of Vegetation Index (NDVI) derived from Landsat 8 time series data, from 2013 to 2019 growing seasons, are effective in predicting winter wheat yield in Jász-Nagykun-Szolnok county (Northern Great Plain region of central Hungary). The highest determination coefficient (R²=of 0.569) was found on the 160 th day, which is lower than the value obtained in the current study (R² = 0.643).…”

Sentinel‐2 accurately estimated wheat yield in a semi-arid region compared with Landsat 8

“…The notes from remote sensing shed light on some of the changes that relate to the earth, soil, plant cover and the difference between plant health (Kamilaris et al, 2017b). There has been an increase in the importance of paying attention to and using the data collected from satellites to monitor and predict the yield of cereal crops this is caused by their ability to produce data by location coverage, real time coverage and objective d at product growth (Adeniyi et al, 2020).…”

“…Various plant diversity indicators have been developed from visible satellite sensors that can provide a lot of information about plant health and biodiversity (Xu et al, 2011). There is a relationship between vegetation indicators (NDVI) and seasonal initial yield, because of this relationship vegetation density can be used as an indirect measure of initial grain yield through clear formation activity of agricultural plants during a certain period before harvest (Adeniyi et al, 2020). The main aim of this study is to predict wheat and barley yield and investigate the relationship between remote sensing derived vegetation indices to experiment with as many vegetation indices as possible with the aim of improving grain yield prediction.…”