Species-based Mapping of Carbon Stocks in Salt Marsh: Tianjin Coastal Zone As a Case Study

Binyue Kang,Xiaoqin Chen,Zhibo Du,Weiqing Meng,Hongyuan Li
DOI: https://doi.org/10.34133/ehs.0052
2023-01-01
Ecosystem Health and Sustainability
Abstract:Because of geographical position and high carbon storage potential, coastal salt marshes are recognized as an essential component of blue carbon and play an indispensable role in regulating climate and reaching carbon neutrality targets. Nonetheless, accurately mapping salt marsh carbon stock on a regional scale remains challenging. The framework of mapping salt marsh carbon stock was developed by using machine learning (temporal-phenological-spatial) models, vegetation index aboveground biomass inversion models, and above/belowground biomass allometric models. Here, we employed Sentinel-2 time series images based on Google Earth Engine in combination with field survey data to produce a 10-m map of salt marsh carbon stocks in the Tianjin coastal zone (TCZ). The total and average carbon stocks of TCZ salt marsh vegetation in 2020 were approximately 6.24 x 103 Mg C and 45.02 Mg C/ha, respectively. In terms of vegetative species, the carbon stock was ranked by Spartina alterniflora (2.89 x 103 Mg C) > Phragmites australis (1.74 x 103 Mg C) > Suaeda salsa (1.61 x 103 Mg C). The carbon density of 3 representative salt marsh species sampled in Tianjin were calculated: S. alterniflora (18.63 Mg/ha) > P. australis (6.49 Mg/ ha) > S. salsa (1.40 Mg/ha). The random forest algorithm shows the best performance in classifying, with an overall accuracy of 87.21%. This work created the replicable and generic technical framework for mapping carbon stocks in salt marshes, which supports blue carbon accounting and provides case support for "nature-based solutions."
What problem does this paper attempt to address?