برآورد شاخص سطح برگ گیاه ذرت با استفاده از تصاویر ماهواره‌ای در پلتفرم گوگل ارث انجین و مقایسه آن با خروجی مدل WOFOST

ABSTRACT

Introduction: Corn, as one of the key agricultural products and a fundamental pillar of food security worldwide, is cultivated in all parts of the world due to its high resistance and adaptability to various climatic conditions, and has long been of interest to farmers due to its high production potential and diverse applications. The leaf area index (LAI) is a key parameter for assessing plant growth. Therefore, accurate measurement and continuous monitoring of LAI are essential for optimal management of corn fields and accurate crop yield prediction. The leaf area index is a key tool for assessing and monitoring vegetation cover changes. The main objective of the present article is to estimation the leaf area index of corn using satellite images in the Google Earth Engine platform and compare it with the output of the WOFOST model, which is innovative in two aspect among research in Iran: one is the use of the WOFOST model, and the other is the use of the capabilities of the Google Earth Engine platform in estimating LAI values and comparing the values with each other.

Material and Methods: This study used Landsat 9 images from the 2023–2024 statistical period within the Google Earth Engine platform. The corn growing period in the Kalibar region of East Azerbaijan province, which spanned from April 14 to September 14, was determined using SMADA software. Additionally, we employed the WOFOST model to examine and compare the leaf area index (LAI) values of corn crops, a key crop for food security. For this purpose, NDVI, SAVI, and LAI indices were calculated. Additionally, r^2, RMSE, and MSE were used to verify the results.

Results and Discussion: First, NDVI, SAVI, and LAI indices were calculated, with the lowest and highest NDVI on 2024/08/25 and 2023/04/24 being -0.228 and 0.691, and the lowest and highest SAVI on 2024/08/25 and 2023/07/22 being -0.342 and 0.937, respectively. The lowest LAI index recorded was zero, while the highest was 5.968, observed on July 22, 2023. The results showed that the RMSE and MSE values of the LAI index based on the WOFOST model were below 0.5 and were equal to 0.376 and 0.334, respectively. Also, the coefficient of determination (r^2) between the WOFOST model and satellite images is 0.857, and the highest LAI starts from day 185 of crop growth and continues until day 225. Additionally, the highest coefficient of determination (r^2) between LAI and NDVI is related to 2024/09/10 with 0.961, and the lowest is to 2024/06/06 with 0.795. The overall correlation value between the LAI index and NDVI is 0.937.

Conclusion: The findings of this study demonstrate that integrating remote sensing data with crop growth simulation models such as WOFOST can be a powerful tool for monitoring and assessing vegetation dynamics, especially the Leaf Area Index (LAI). Spatiotemporal analyses of the SAVI and NDVI indices revealed that the southern and southwestern regions of the study area exhibited the highest values of these indices due to dense vegetation cover. Conversely, barren lands caused the lowest values in the north and east. A strong positive correlation between SAVI and NDVI (with a coefficient of determination of 0.857) confirms that these indices can be used complementarily to evaluate vegetation health and density. Furthermore, the high agreement between LAI values derived from satellite imagery and WOFOST model predictions (with low RMSE and MSE values of 0.376 and 0.334, respectively) underscores the model’s accuracy in simulating plant growth parameters. These findings suggest that calibrating dynamic crop growth models with satellite data can be a practical solution for rapid monitoring and yield prediction in large-scale agricultural applications. Keywords: Leaf Area Index (LAI), NDVI, SAVI, Google Earth Engine, Landsat, WOFOST.