Peatlands are a unique wetland class that sequesters soil carbon for millennia. Peatland carbon stocks are known to be some of the largest in the world. However, little is known about the distribution of peatlands in tropical areas, especially small mountain peatlands found in the tropics due to poor precision of current maps. The climate mitigation potential of peatlands relies on the protection of threatened areas or the restoration of degraded sites. Identification of the regional potential of climate change mitigation requires knowledge of peatland distribution, regional patterns of disturbance and land use change, and patterns of GHG emissions and absorptions. Remote sensing of peatlands in tropical mountains have many challenges: incorrect returns from radar on rough landscapes, permanent cloud cover, seasonality, differences between flooding and soil saturation from high precipitation, confusion between land cover types, and lack of relevant ground truthing points. Here, we focused on understanding the distribution and diversity of peat accumulating ecosystems on a subregion of the central Andes of Colombia using a remote sensing approach. Our analysis strategy included an intensive field campaign with over 500 field points and a multiplatform-multisensor remote sensing mapping strategy. Ground truthed points were randomly assigned and data on land cover, land use, and soil carbon down to 40 cm was collected. Remote sensing analysis was based on a combination of C and L radar bands (Alos Palsar 1 and 2, Sentinel 1), optical imagery (Landsat and Sentinel 2), and high-resolution topography. We discriminated the different land cover classes using a random forest algorithm. We identified 11 land cover classes with four different peatland classes: graminoid, shrub, cushion plants, and pasture peatlands. The overall precision was above 85% with the lowest precision differentiating between pasture peatlands and pastures on mineral soils. The region’s 429,000 ha were dominated by forest (40%) and shrub vegetation (28%). Peatlands covered nearly 10% of the area whereas disturbed areas covered nearly 12% of the region. Peat soils under intensive pasture represented 0.1%. Carbon content was similar in natural peatlands. Our results prove the accuracy of the mapping technique used and highlights the extension and relevance of these ecosystems in the Central Cordillera in Colombia. These results will be of paramount importance for future studies of climate change mitigation and a tool to help developing countries in achieving their targeted GHG emission reductions.


Peatlands, Colombia, Disturbance, NDC, Tropical Andes, climate mitigation

Juan Benavides, Erik Lilleskov, Angela Lafuente, Nicolás Patrick, Laura Bourgeau-Chavez, Michael Battaglia, Rodney Chimner

Presentation within symposium:

S-50 Tropical South American Peatlands: distribution, impacts of disturbance and knowledge gaps

Using remote sensing to map degraded mountain peatlands with high climate mitigation potential in the wet Tropical Andes