Title: Estimating Guilan rice production potential in drought conditions through preparing paddy map using remote sensing
Author: Mojtaba Rezaei
Abstract: A large area of Iran’s rice lands is scattered in the coastal areas of the Caspian Sea, especially the Mazandaran, and Gilan provinces, accounting for more than 70% of the country’s rice cultivation area. Monitoring the paddy fields has a special place in management and decision-making. Extracting agricultural census through traditional methods and mapping, despite the high accuracy, is costly and time-consuming. In the vast geographical areas, remote sensing (RS) techniques have a high potential in producing statistical due to their ability to capture images with high spatial resolution and to produce images in different parts of the electromagnetic spectrum. This study aimed to investigate the ability of RS images to estimate the rice cultivation areas in Gilan province using 6 supervised classification methods including maximum likelihood-ML, CART, random forest-RF, SVM, GME, and simultaneous use of RF with NDVI index. The ML method was performed in the ENVI environment using 6 Sentinel 2 image frames in 2018 and the others were done through the GEE environment using the same image. The classification results showed that the RF method with growth indices with a kappa coefficient of 0.94 had the highest accuracy compared to other methods. Estimation of the province’s cultivation area by RF method and NDVI showed that the total land area of the province is about 229 thousand hectares, which is slightly less than the existing statistics of the Jihad Agricultural Organization of the province (230 thousand hectares) and the regional water company (245 thousand hectares). In total, taking into account about 5% of the infrastructure facilities, the total net land of the province is estimated at 217 thousand hectares. Meanwhile, Rasht city with about 59 thousand hectares has the most and Rudbar with 2500 hectares have the lowest area under cultivation.
Keywords: Rice GEE, Area, remote sensing.