Document Type : Original Research Paper
Authors
1 Department of Surveying Engineering, Faculty of Engineering, University of Zanjan, Zanjan, Iran
2 Iranian Space Research Institute,,Tehran,,Iran
3 Iranian Space Research Center, Tehran, Iran
Abstract
Atmospheric water vapor is a critical meteorological parameter influencing climate studies, weather forecasting, and climate change modeling. Precipitable water vapor (PWV) is measured using satellite data, particularly from the MODIS sensor, which provides near-infrared (NIR) and infrared (IR) products. The IR product, widely used due to its day-and-night data availability, faces accuracy challenges under varying atmospheric conditions. This study aims to enhance the accuracy of MODIS IR PWV data using machine learning and assess calibration effects on day and night data. Data from 10 radiosonde stations in Iran (2019–2020) served as the reference. Three datasets—original MODIS, fitted, and modified—were analyzed. A multilayer perceptron (MLP) model was designed for calibration, and its performance was evaluated using root mean square error (RMSE) and correlation coefficient (R). The MLP model significantly improved PWV accuracy. During daytime, RMSE decreased from 3.72 mm to 2.63 mm, and R rose from 0.81 to 0.86. At night, RMSE dropped from 4.9 mm to 3.16 mm, with R improving from 0.76 to 0.78. Overall, RMSE reduced from 4.48 mm to 2.92 mm (fitted) and 3.03 mm (modified), while R increased from 0.77 to 0.87 and 0.85, respectively. The MLP model effectively reduced errors in MODIS PWV data, enhancing reliability for weather forecasting and climate studies. Its innovation lies in calibrating satellite data for day and night conditions.
Main Subjects
)