Quantifying and mapping blue carbon ecosystems in Bahrain waters

Abstract

Seagrass habitats in Bahrain have experienced significant impacts from dredging and reclamation activities associated with coastal development projects initiated since the 2000s. The primary objectives of this study were to quantify and map the distribution of Blue Carbon ecosystems in Bahraini waters, evaluate the degradation and destruction of these ecosystems, and account for losses in sedimentary carbon. Previous classification maps for Bahrain were re-evaluated to establish the total seagrass area for the preceding years. An enhanced pre-processing methodology was employed to generate new classification maps, which resulted in lower accuracies ranging from 41.18% to 61.05%. The effect of turbidity on the ability to accurately map the survey area was investigated using two approaches: (i) remote sensing reflectance employing Sentinel-2 S2A and Landsat 8 data from the Sentinel Application Platform Case 2 Regional Coast Colour Processor (2016-2021) to assess water quality and map turbidity, total suspended matter, and chlorophyll-a; (ii) estimating turbidity through a correlation between in-situ turbidity and total suspended matter, substituting the latter's values from the processor outputs. Mapping and validation of chlorophyll-a and turbidity relied on the relationships between reflectance bands and in-situ observations, resulting in an R2 value of 0.45 (p = 0.00016), a correlation coefficient of 0.67, and a significant linear relationship between the variables. Sampling locations were chosen within densely populated seagrass areas, and seagrass substrate was extracted using a box corer to analyse sediment carbon content corresponding to different covers. Underwater photos were taken, and multiple quadrants were employed to determine the cover for each location. Landsat 8 and Sentinel-2 images were acquired concurrently with field sampling, and image pre-processing was performed. The Depth-Invariant Index (DII) was calculated for all band combinations, and the most suitable DII band combination was used to generate seagrass classification maps. Linear regression models were conducted between the DII and seagrass cover to create estimate maps for seagrass cover within the seagrass classification. By utilising the seagrass cover estimation maps and sediment carbon values from the field survey, we developed below-ground sediment carbon estimation maps for the seagrass area in Bahraini waters. The average sediment carbon content within the seagrass ecosystem, calculated from the 2018 and 2019 surveys, was 2.54 ± 0.19 kg/m2 for Landsat 8 and 2.40 ± 0.31 kg/m2 for Sentinel-2. Estimated sediment carbon for 1985 (1.89 ± 0.50 kg/m2), 2005 (1.73 ± 0.95 kg/m2), and 2015 (1.89 ± 0.73 kg/m2) indicate a significant carbon stock to be considered for the country's carbon budget and IPCC directions. However, the seagrass area has decreased by nearly half from 1985 (1875 km2) to 2019 (656-953 km2), leading to an estimated carbon loss of approximately 1,122,114 ± 940474 -1,968,342 ± 944979 Mg C. The same model was used to estimate the sediment carbon content for the seagrass area within the Arabian Gulf, revealing carbon content associated with sediments in the Gulf ranging from 0.80 to 3.69 kg/m2, with averages of 2.46 ± 1.53 kg/m2 (24.6 ± 15.3 Mg C/ha) for Landsat 8 and 2.69 ± 1.58 kg/m2 (26.9 ± 15.8 Mg C/ha) for Sentinel-2. The seagrass cover of Bahrain decreased over the years, with the change synchronised with carbon stock, particularly since 2001 when dredging and reclamation activities became widespread. The study confirmed the effectiveness of remote sensing sensors, such as Landsat and Sentinel, in predicting seagrass cover percentages, enabling monitoring of seagrass on spatial and temporal scales. Covering the entire territorial waters of Bahrain for further specification input is crucial to detect temporal variations resulting from seasonal changes and potential climate change impacts. This research highlights the importance of incorporating seagrass ecosystems into the country's carbon budget and mitigation strategies while emphasising improved management and conservation measures to prevent further seagrass habitat loss due to coastal development activities. By effectively monitoring and understanding the dynamics of seagrass ecosystems, stakeholders can better address the challenges posed by climate change and anthropogenic activities, ensuring the preservation of these vital blue-carbon ecosystems for future generations.

Date of Award	Dec 2023
Original language	English
Awarding Institution	University of Brighton
Supervisor	Matthew Brolly (Supervisor), Niall Burnside (Supervisor) & Chris Carey (Supervisor)