Water Stress Prediction Model in Rainfed Sugarcane Fields Using Temporal Landsat Data Based on Random Forest Regressor.
Saved in:
| Title: | Water Stress Prediction Model in Rainfed Sugarcane Fields Using Temporal Landsat Data Based on Random Forest Regressor. |
|---|---|
| Authors: | Suharso, Aries1 aries.suharso@unsika.ac.id, Herdiyeni, Yeni2 yeni.herdiyeni@apps.ipb.ac.id, Tarigan, Suria Darma3 sdtarigan@apps.ipb.ac.id, Arkeman, Yandra4 yandra@apps.ipb.ac.id |
| Source: | IAENG International Journal of Computer Science. May2025, Vol. 52 Issue 5, p1393-1406. 14p. |
| Subjects: | Temporal databases, Machine learning, Sugar plantations, Landsat satellites, Random forest algorithms |
| Abstract: | Sugarcane drought (water stress) at the critical germination and tillering phases results in a decrease in milled sugarcane production every year. Using temporal data from Landsat 8 satellite images and local meteorological data, this study attempts to assess sugarcane water stress using upper canopy reflectance (LST). The difficulty in obtaining when calculating canopy temperature is quite complicated, we propose a sugarcane water stress indices prediction model based on the Random Forest Regressor machine learning algorithm with other multivariate vegetative feature data NDVI, NDWI, NDDI, OSAVI, LSWI combined with local daily climate data in the form of air temperature, humidity, rainfall, sunshine hours and wind speed. This data is relatively easier to obtain compared to LST. The research study was carried out at the Indonesian Sugar Research Center's sugarcane plantation in Djengkol Kediri, East Java, Indonesia. Observations were focused on the sugarcane plantation area that was suspected of experiencing the most severe decline in milled sugarcane yields (G33). Initial analysis showed an increase in water stress phenology during the germination - tillering phase between June - October in 2021 and 2022. Through cross-correlation tests, and time lag effect tests, we compiled the dataset. The prediction performance results of our proposed Random Forest Regressor Model achieved the best performance with the vegetation dataset without LST achieving an accuracy of X2 = 91.08% and MAPE = 8.93%. These results emphasize that the multi-feature method, excluding LST, was effective in forecasting variations in the sugarcane water stress index. Consequently, this approach is anticipated to mitigate potential losses in future sugarcane milling productivity. [ABSTRACT FROM AUTHOR] |
| Copyright of IAENG International Journal of Computer Science is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
Be the first to leave a comment!
