<![CDATA[Salinity is the most important parameter for controlling the biological components of ecosystems, seas, and estuaries, which also control the components that make up an ecosystem. Conventional water quality monitoring is considered inaccurate and inefficient in terms of energy and time. Therefore, research is needed to predict sea surface salinity as a type of water quality monitoring using remote sensing reflectance or Remote Sensing Reflectance (RRS) from Landsat imagery. The Landsat image data used is level 2 Surface Reflectance (SR), which is ready to use without additional processing by the user, whereas previous research required corrections to the image data to obtain Surface Reflectance image data. This study aims to determine the performance of the prediction model produced by using five combinations of Landsat image bands. The data used are Landsat 8 OLI image data (recording date 05 August 2024) downloaded from the USGS website and in situ salinity data in the Bangka Strait sea (09 March 2025), as many as 5 samples that can be used. The obtained data were processed using multiple linear regression analysis with Rrs as the independent variable and in situ salinity as the dependent variable. The salinity prediction model consisted of five band combinations. The analysis produced R² values for each model combination of 0.8166287408, 0.935603228, 0.820745745, 0.869209652, and 0.574027060. The RMSE validity tests for combination 1, combination 2, combination 3, combination 4, and combination 5 were 2.41327, 1.43012, 2.38602, 2.03811, and 3.67817. Then for the NMAE value, namely 10.10152205%, 5.32713015%, 9.58011308%, 8.8868031%, and 14.51012574%. The combination rankings that have the best prediction performance are combination 2, combination 4, combination 3, combination 1, and combination 5. So the best model in predicting seawater salinity is the combination of the 2 prediction models, with its constituent band components being band 1, band 2, and band 4.]]>
Copyrights © 2025
