LB 42-357 - Determining paramo peatland distribution in the Colombian Andes using multidate, multisensor remote sensing data.

Páramo peatlands are important carbon and nutrient sinks, water regulators, and provide quality habitat. The páramo peatlands store more carbon per hectare below ground as peat soil, than all the above-ground biomass of tropical forests. However, these are highly disturbed ecosystems due to a combination of increasing temperatures in the tropics, and activities such as agriculture, grazing and drainage, causing peatlands to quickly lose soil carbon to decomposition, and contributing to atmospheric greenhouse-gas concentrations. Approximately 43% of the Andean páramos are found in Colombia, however peatland coverage and their carbon stocks have not been estimated. An urgent need exists for local authorities to develop peatland planning and management strategies with local communities to undertake climate change mitigation policies for the country. Our main goal is to create a national peatland map for Colombian Páramos. To achieve this, I am utilizing multi-date, multi-sensor radar and optical imagery, along with ground-truth data, as inputs for a machine learning algorithm called random forest. As part of our preliminary results for the mapping efforts I will be presenting results for the first mapped region in the department of Nariño, Colombia. Our model had 82% overall accuracy. We found that páramo peatlands in Nariño account for 10.5% of the total páramo complexes in the area. Pasture peatlands, which represent disturbed peatland, represent one fourth of the total peatland areas. More exploration of these areas will help concentrate efforts for restoring and managing affected peatlands. In conclusion, understanding peatland distribution will not only be useful for quantifying páramo peatlands carbon stocks but also for prioritizing areas for peatland conservation and restoration.