4:15 PM - 4:30 PM
[AHW22-19] Investigation of Seagrass Meadow Community in Coral-Reef Lagoons by Remote Sensing Technology, Indonesia
Keywords:blue carbon, nutrient cycle, seagrass meadow, mapping
Introduction
Carbon sequestration is crucial for environmental stability, however, estimating its presence, particularly in the form of blue carbon, poses challenges. Seagrass meadow, prevalent in coastal ecosystems, play a significant role in carbon sequestration. However, human activities have led to a decline in seagrass meadow populations. Nonetheless, human activities can also exert positive impacts on seagrass meadow. For instance, a decrease of nutrient like nitrogen and phosphorus in the Seto Inland Sea has a negative impact on cultivated laver. Such nutrient, supplied by human activities, can have positive impacts on coastal seagrass meadow. From the perspective of nutrient cycle, abundant nutrient supply is believed to positively influence seagrass meadow. However, effect of human activities on coastal seagrass meadow is complex. Thus, case studies demonstrating increased seagrass meadow activity due to abundant nutrient supply are scarce. With observational data on seagrass meadow abundance covering broad areas, the relationship between nutrient supply from land and seagrass meadow could be discussed by comparing multiple locations. Considering the labor-intensive of surveys, assessing seagrass meadow abundance over broad areas, remote sensing is the best observation method. This study investigated seagrass meadow distribution in lagoons using remote sensing.
Materials and Methods
The satellite imagery used for supervised classification was Sentinel-2 (ESA) data captured on April 14, 2022. High-resolution images from Google Earth obtained on March 10, 2022, were used to generate 650 points training data. 80% of this training data was used for supervised classification, while 20% was used for validation the classified result. Supervised classification was conducted using the Random Forest method. The classification accuracy was evaluated using a confusion matrix. There are over 100 lagoons within the study area, and polygons were created for all lagoons. The area of seagrass meadow was calculated for each lagoon. Biomass per unit area of seagrass meadow was calculated using biomass data from previous studies on Indonesian lagoons, and the biomass (above and below ground) was derived from the area of seagrass meadow.
Results and Discussion
Table 1 shows the classification accuracy. The overall accuracy was 85%, with a producer's accuracy of 78.8% and a user's accuracy of 83.9% for seagrass meadow classification. Misclassifications were predominantly identified as coral reefs. Figure 1 shows the mapping results of seagrass meadow. Larger islands, particularly those in the south, exhibited larger areas covered by both land and seagrass meadow. The island with the largest area covered by seagrass meadow was Pari Island (0.74 km²). The total area of seagrass meadow within the whole study area was 61.3 km², and the estimated biomass was 1082.7 dry weight tons. Analyzing the relationship between the proportion of seagrass meadow and other classification categories in the top 25% of lagoons by area revealed a positive correlation between the proportion of seagrass meadow and terrestrial areas, with an R² of 0.385 and a p-value of <=0.0005. By comparing multiple lagoons, it was possible to demonstrate the potential for nutrients supplied from terrestrial areas to enhance the area of seagrass meadow.
Acknowledgement
This research was supported by KAKENHI(21KK0192).
Carbon sequestration is crucial for environmental stability, however, estimating its presence, particularly in the form of blue carbon, poses challenges. Seagrass meadow, prevalent in coastal ecosystems, play a significant role in carbon sequestration. However, human activities have led to a decline in seagrass meadow populations. Nonetheless, human activities can also exert positive impacts on seagrass meadow. For instance, a decrease of nutrient like nitrogen and phosphorus in the Seto Inland Sea has a negative impact on cultivated laver. Such nutrient, supplied by human activities, can have positive impacts on coastal seagrass meadow. From the perspective of nutrient cycle, abundant nutrient supply is believed to positively influence seagrass meadow. However, effect of human activities on coastal seagrass meadow is complex. Thus, case studies demonstrating increased seagrass meadow activity due to abundant nutrient supply are scarce. With observational data on seagrass meadow abundance covering broad areas, the relationship between nutrient supply from land and seagrass meadow could be discussed by comparing multiple locations. Considering the labor-intensive of surveys, assessing seagrass meadow abundance over broad areas, remote sensing is the best observation method. This study investigated seagrass meadow distribution in lagoons using remote sensing.
Materials and Methods
The satellite imagery used for supervised classification was Sentinel-2 (ESA) data captured on April 14, 2022. High-resolution images from Google Earth obtained on March 10, 2022, were used to generate 650 points training data. 80% of this training data was used for supervised classification, while 20% was used for validation the classified result. Supervised classification was conducted using the Random Forest method. The classification accuracy was evaluated using a confusion matrix. There are over 100 lagoons within the study area, and polygons were created for all lagoons. The area of seagrass meadow was calculated for each lagoon. Biomass per unit area of seagrass meadow was calculated using biomass data from previous studies on Indonesian lagoons, and the biomass (above and below ground) was derived from the area of seagrass meadow.
Results and Discussion
Table 1 shows the classification accuracy. The overall accuracy was 85%, with a producer's accuracy of 78.8% and a user's accuracy of 83.9% for seagrass meadow classification. Misclassifications were predominantly identified as coral reefs. Figure 1 shows the mapping results of seagrass meadow. Larger islands, particularly those in the south, exhibited larger areas covered by both land and seagrass meadow. The island with the largest area covered by seagrass meadow was Pari Island (0.74 km²). The total area of seagrass meadow within the whole study area was 61.3 km², and the estimated biomass was 1082.7 dry weight tons. Analyzing the relationship between the proportion of seagrass meadow and other classification categories in the top 25% of lagoons by area revealed a positive correlation between the proportion of seagrass meadow and terrestrial areas, with an R² of 0.385 and a p-value of <=0.0005. By comparing multiple lagoons, it was possible to demonstrate the potential for nutrients supplied from terrestrial areas to enhance the area of seagrass meadow.
Acknowledgement
This research was supported by KAKENHI(21KK0192).