ANALYSIS OF SEAGRASS ABOVEGROUND CARBON STOCK DYNAMICS IN PARI ISLAND, 2021-2023, USING SENTINEL-2

Seagrasses are marine plants that efficiently store carbon. Understanding their role in climate change requires information on seagrass area and carbon content, which is currently lacking in Indonesia. The objectives of this study are to (1) develop a mapping model of seagrass aboveground carbon stock (AGC) dynamics based on percents of seagrass cover, (2) map AGC dynamics using multitemporal Sentinel-2 imagery and (3) analyze patterns and factors affecting AGC dynamics. This study used two regression models, random forest regression (RFR) and stepwise regression (SWR). The RFR regression model produced a more accurate and consistent AGC map with R2=0.21 (RSME=5.04gC/m2) for the Ea class and R2=0.24 (RSME=1.99gC/m2) for the ThCr class. Meanwhile, SWR produced an accurate AGC map for the EaTh class with R2=0.15 (RSME=2.90gC/m2). Both models were applied to Sentinel-2 images for 15 months, from April 2021 to December 2023. The highest AGC for the RFR model was shown in October 2021 with 0.104 tons of carbon and for the SWR model in December 2023 with 0.105 tons from a total seagrass cover area of 1.15km2. Biophysical variables like rainfall can affect AGC dynamics. As rainfall increases, the AGC estimate tends to increase. © 2025, Romanian Inventors Forum All Rights Reserved