Evaluation of Landsat 8 Spectral Data Capabilities and Laboratory Spectroradiometer for predicting corn yield (Case study: Moghan Agro-Industry)

Introduction: Monitoring the growth stages and yield of crops in agricultural areas is essential for food security and farmers' income forecasts. Progress in remote sensing has greatly contributed to the process of monitoring of various developmental stages of agricultural crops and the evaluation of their yield (Anastasiou et al., 2018; Shi & Mo, 2011). Remote sensing (RS) and global positioning systems (GPS) can be used to evaluate the changes in crop dynamics, including its yield and spatial diversity (Dadhwall & Ray, 2000). Spectral vegetation indices (SVIs) are a combination of the spectral absorption and spatial distribution of plants in different electromagnetic spectral range and are used to measure the characteristics of a product. SVI provides a simple method for measuring spectral responses of plants throughout the season, which uses fundamental differences between soil and plants, and often as a kind of relationship between the energy of electromagnetism reflected in red and infrared wavelengths the red near (NIR) is expressed. Green healthy plants exhibit relatively low reflections in the visible range of the electromagnetic spectrum (high absorption of light for photosynthesis); however, its reflection is usually high in the near infrared region (Al-gaadi et al., 2016). Therefore, in this study, the remote sensing method spectrometric data were used to predict the yield of corn in the Moghan plant in northern Ardebil province.
Materials and Methods: Landsat-8 satellite images were prepared during four growth stages of corn, and simultaneously at the dates when the satellite images of the study area were taken, spectroscopy of the plant samples was performed using the Field Espect-3 spectrometer. In this study, 30 corn fields were selected in the Moghan Plain to estimate the yield of corn. First, using the GPS device, the position of the farms was determined. Then, at different growth stages e.g. four-leaf stage, growth differentiation stage, flower emergence stage and physiological maturity, soil and plant samples were prepared according to standard methods and then the specimens were measured. Vegetation indices of NDVI, SAVI, MNDVI and OSAVI were calculated based on satellite data and laboratory spectrometers.
Results and Discussion: The results in both cases- the use of Landsat 8 satellite images and laboratory spectrometer at flower emergence stage- showed that the correlations for coefficient of determination for leaf area index and yield were from 54% to 72%, which were more robust as compared to other growth stages. On the other hand, the evaluation of the indices obtained from the spectrophotometric results and the use of spectral data and the comparison between the two showed that the correlation for coefficient of determination for NDVI and SAVI was 70%, which were determined as effective indices for estimating leaf area index and yield using satellite imagery. While MNDVI and OSAVI indices were 72% and 69%, respectively, they were found to be the most suitable indicators for estimating leaf area index and yield based on the results of laboratory spectroscopy. Therefore, using satellite images, these indices are more likely to be present, while for MNDVI and OSAVI indices, particular wavelengths are studied, and given the fact that the laboratory spectrometer shows the slightest variations in each wavelength, these indicators can also be considered as robust.
Conclusion: The results of the study showed that among the growth stages of the plant, flower emergence stage was the best for predicting the yield of the crop, and on the other hand, NDVI and SAVI indices resulting from Landsat 8's satellite imagery were found to be the most robust in predicting crop yield and MNDVI and OSAVI indices were found to be the best predictors of crop yield based on the results of spectral laboratory data.