Detecting peatland drains with Object Based Image Analysis and Geoeye-1 imagery

Area of Science:

• Environmental Science
• Remote Sensing
• Ecology

Background:

• Peatlands are crucial for the global carbon cycle, offering vital ecosystem services like carbon sequestration.
• Drainage disrupts these services, and accurately mapping peatland drains is challenging and costly.
• The spatial extent of drains is often unknown, hindering effective peatland management.

Purpose of the Study:

• To develop and assess an object-based image analysis (OBIA) method for mapping peatland drains.
• To determine the accuracy and cost-effectiveness of OBIA compared to traditional methods.
• To utilize drain maps for assessing carbon flux impacts and management strategies.

Main Methods:

• Object-based image analysis (OBIA) applied to very high-resolution satellite imagery (Geoeye-1).
• Accuracy assessment using error matrix (overall accuracy, kappa statistic) and completeness, correctness, and quality (CCQ).
• Cost comparison with manual digitisation and field surveys; drain maps used for carbon flux estimation.

Main Results:

• OBIA successfully detected nearly 500 km of drains with high accuracy (overall accuracy 94%, kappa 0.66).
• Sub-area accuracy ranged from 95-97%, with CCQ scores of 85%, 85%, and 71%.
• OBIA was at least 55% more cost-effective than field surveys or manual digitisation; mapped drains reduced CO2 flux estimates by 19%.

Conclusions:

• OBIA provides an accurate and cost-effective method for mapping fine-scale peatland drains over large areas.
• Accurate drain mapping is essential for understanding peatland carbon dynamics and informing climate change mitigation strategies.
• This method supports Annex 1 parties in accounting for greenhouse gas emissions and developing effective drain-blocking strategies globally.