Leveraging machine learning and open accessed remote sensing data for precise rainfall forecasting
Rainfall forecasts are essential for human activities enabling communities to anticipate any impacts. Rainfall events correlate with other natural and hydro-meteorological phenomena, which can be used in modeling and prediction. This study used daily CHIRPS for the Gajahwong watershed in Yogyakarta,...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Komunitas Ilmuwan dan Profesional Muslim Indonesia
2025-07-01
|
| Series: | Communications in Science and Technology |
| Subjects: | |
| Online Access: | https://cst.kipmi.or.id/journal/article/view/1638 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Rainfall forecasts are essential for human activities enabling communities to anticipate any impacts. Rainfall events correlate with other natural and hydro-meteorological phenomena, which can be used in modeling and prediction. This study used daily CHIRPS for the Gajahwong watershed in Yogyakarta, Indonesia as the precipitation data. It also used Sea Surface Temperature, Land Surface Temperature (Day and Night), Minimum and Maximum Temperatures, Solar Radiation, Wind Speed (U and V components), Cloud Pressure (Top and Base), and Cloud Height (Top and Base) as the parameters. Further, data processing was performed by means of the Google Earth Engine (GEE) platform. Machine learning methods, including Support Vector Regression, Gradient Boosting Regression, Random Forest, and Deep Neural Networks, were applied. The correlation analysis revealed that only the Wind Speed V-component showed significant correlation with rainfall, other seven parameters showed moderate and four showed weak ones. Meanwhile, accuracy assessments indicated that Support Vector Regression had the most accurate predictions accompanied by Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Squared Error (MSE), R2, and Coefficient Correlation (CC) at 1.366, 0.947, 1.866, 0.948 and 0.982 respectively. This study demonstrated that utilizing openly accessible atmospheric datasets processed through the GEE could yield reliable rainfall predictions, facilitating informed decisions on a wide scale. The methodology is adaptable and can be reproduced for any comparable research or operational purposes. |
|---|---|
| ISSN: | 2502-9258 2502-9266 |