top of page

Adatos' Peat Depth Model

The recently released Global Peatlands Assessment 2022 highlights the importance of peatlands and the threats to their survival. It also provides peatland maps that can be used as a valuable baseline for peatland conservation decisions.

Ecologically, peat is an extremely important natural resource that contains up to 33% of the world’s soil carbon despite only covering 3-4% of the planet’s surface. It has therefore become a key focus area and tool for mitigating climate change. Peat is also important for the livelihoods of many local communities and provide an irreplaceable habitat for a range of plants and animals. However, peatlands around the world are degrading fast due to land use change and peat extraction for fuel and horticulture. This calls for urgent action to improve sustainable management of these wetland ecosystems.

The advancement of technologies around remote sensing and artificial intelligence (AI) means that project developers are increasingly using digital monitoring, reporting, and verification (DMRV) for climate and sustainability. Specifically, in our experience, DMRV has been sought after for accelerated pre-feasibility studies for potential project sites.

Adatos is a leading AI provider in carbon forestry, mangroves and precision agriculture. Specific to peatlands, Adatos has a fully satellite-based AI model trained on 20 relevant datasets that estimates peat depth. The beauty of using satellite data is that each pixel is rich in spectral information, enabling high resolution, pixel-based mapping. This contrasts with standard peat mapping, which interpolates between widely spaced ground truth points.

Adatos’ peat depth model typically categorizes peat depth into 1 meter classes, as seen in this image. Adatos’ peat depth model has demonstrated cumulative 5% peat depth errors which lends confidence and credibility to the output. Image Source: Adatos

The peat depth model was developed and verified using ground truth and field surveys in a tropical peatland, with input from relevant expert organizations. When tested against in situ measurements, the peat depth model has a 5% cumulative error for belowground carbon stock. This gives us confidence that it is appropriate for similar tropical peatland environments, whether as a guide for field sampling or measure of carbon. Furthermore, with the increase in resolution of satellite imagery, Adatos can produce peat depth maps at 30m resolution across large areas.

Moving forward, Adatos plans to further develop the peat depth model by incorporating more ground truth data from different peatland geographies. This would enable:

  1. Increased applicability and model accuracy across more environments (e.g. temperate peatlands), and

  2. Enhanced model precision for shallow and deep peat ranges.

Contact Adatos now to discuss how you can harness the benefits of DMRV and use Adatos’ peat depth model for your projects.


bottom of page