Simulating spatial distribution of coastal soil carbon content using a comprehensive land surface factor system based...

Area of Science:

• Environmental Science
• Remote Sensing
• Soil Science

Background:

• Surface soil carbon content (SCC) in coastal areas is influenced by complex factors, making spatial distribution analysis vital for carbon storage assessment and understanding variation drivers.
• Remote sensing data offers ecological insights valuable for establishing comprehensive land surface factor systems (CLSFS).

Purpose of the Study:

• To map the spatial distribution of surface soil carbon content (0-30cm) in the Yellow River Delta.
• To evaluate the accuracy and applicability of different simulation algorithms (SFR, MFR, PLSR, BPNN) using a CLSFS.
• To identify key factors influencing SCC spatial distribution in coastal environments.

Main Methods:

• A comprehensive land surface factor system (CLSFS) was developed, integrating spectral information, ecological indices, spatial location, and land cover from remote sensing data.
• Four simulation algorithms, including single-factor regression (SFR), multiple-factor regression (MFR), partial least squares regression (PLSR), and back propagation neural network (BPNN), were employed for SCC mapping.
• A 10-fold cross-validation approach was utilized to assess the uncertainty and accuracy of the simulation algorithms.

Main Results:

• The back propagation neural network (BPNN) model demonstrated the highest accuracy in SCC mapping, with a mean root mean squared error of 2.78g/kg, outperforming SFR, MFR, and PLSR.
• The study area exhibited distinct spatial heterogeneity in SCC, with generally lower content alongshore (excluding estuaries) and higher content in the western regions.
• Farmland and wetland vegetation showed higher mean SCCs (18.31g/kg and 17.98g/kg, respectively) compared to water areas, salterns, and bare land.

Conclusions:

• The comprehensive land surface factor system (CLSFS) is effective for simulating surface soil carbon content spatial distribution in coastal areas.
• Land-sea interaction and human activities are identified as joint drivers influencing SCC spatial distribution.
• The BPNN model offers a highly accurate method for SCC mapping, contributing to better carbon management strategies in coastal zones.