Modelling forest carbon stocks on the Canary Islands

Area of Science:

• Forestry
• Ecology
• Remote Sensing

Background:

• Island-wide forest carbon mapping is limited.
• Forest carbon data is vital for management, climate mitigation, and conservation.
• This study addresses the need for high-resolution island forest carbon data.

Purpose of the Study:

• To create the first high-resolution (50m) spatial map of forest carbon for the Canary Archipelago.
• To model total, aboveground, and belowground carbon density for 18 forest types.
• To identify key drivers of carbon stock variation.

Main Methods:

• Combined field data (Spanish National Forest Inventory plots) with airborne laser scanning and Sentinel-2 satellite data.
• Utilized fine-scale interpolated climatic variables.
• Employed machine learning models, including Boosted Regression Tree and Random Forest.

Main Results:

• The Canary Islands' forests store an estimated 10.26 Tg of carbon.
• Canarian pine forests hold the largest pool (57%), while humid laurel forests exhibit exceptional densities (413.2 ± 149.5 Mg C ha⁻¹).
• Structural attributes and moisture availability were primary drivers of carbon stock variation.

Conclusions:

• Developed a high-resolution, accurate carbon map for complex oceanic islands using integrated data and machine learning.
• Highlighted exceptionally high carbon densities in Canary Islands' forests, especially humid laurel forests.
• Provided a baseline for conservation, forest management, restoration, and climate policy.