Characterising, mapping and modelling peatland carbon stocks across Amazonia

We invite applications for the following project from which the best 2 candidates will be selected for interviews which will take place in Edinburgh between 21-24 February 2017.

Until a decade ago, the quantity of carbon stored in peatlands in the tropics was thought to be small and overwhelmingly concentrated in SE Asia. It is now clear, largely through two previous PhD studentships we have co-supervised, that extensive peatlands exist in the Pastaza-Marañón Basin in western Amazonia (c. 35,000 km2, storing 1-8 PgC; Draper et al. 2014), and in the central Congo Basin (145,500 km2, 6-47 PgC: Dargie et al. Nature 2017). For context, peat in Scotland stores ~1.6 PgC. Tropical peatlands account for a substantial part of global soil carbon storage and are very important in terms of potential regional greenhouse gas fluxes, as land use change can quickly return their carbon stocks to the atmosphere.

Tropical peatlands are mostly forest-covered and remote, making them difficult to survey. We and our students have developed an approach to predicting the distribution of peat based on mapping the wetland vegetation types (e.g. seasonally flooded forest, palm swamp, dwarf pole forest) that overlie peat, using remote sensing products that measure aspects of land cover (Landsat), hydrology (PALSAR radar), and topography (SRTM). Extensive field data have been used to develop and test our predictive models, and establish the peat thickness under different vegetation characteristics.

Key Research Questions:

1. Can we use better field and remote sensing data to reduce uncertainties on carbon storage estimates of peat in the Pastaza-Marañón Basin of Peru (currently at 95% confidence 1-8 PgC)?
2. Can we extend these methods to estimate peat stocks across Amazonia?
3. Can we create useful outputs from these maps and estimates for NGO’s and governments of the region, to assist with policy decisions and conservation planning?

Methodology: Building on our previous work in the Amazon, the student will:

1. Make a step change in our ability to predict carbon distribution. Our modelling provides the best estimates yet of peat carbon storage in western Amazonia but uncertainties around our central estimate remain large (1-8 PgC at 95% confidence). We will reduce these uncertainties by (1) incorporating multi-date radar data into our model, which we believe can resolve differences in seasonality of flooding; (2) collecting field data to validate interpretation of the multi-date radar data; (3) extending our network of ground reference data to poorly-sampled parts of the basin; and (4) testing the meaning of ‘speckle’ in all our vegetation maps, which at face value indicates the presence of very small (<100 m) scattered peatlands across parts of the landscape; we do not know if these are real or not.
2. Extend our modelling in other parts of Amazonia, using existing data and undertaking additional fieldwork (e.g. Yasuni, Ecuador; Madre de Díos, Peru) to test the performance of our model in other regions. This would be a significant step towards mapping peat distribution across the whole of Amazonia.
3. Tailor our modelling of peat distribution to the needs of stakeholders. We have worked with WCS-Perú to integrate our current outputs with existing vegetation maps used for conservation planning. Significant changes in class definitions and resolution were necessary. Similarly, carbon conservation projects, local communities, and global land surface modellers have different requirements. We will seek input from stakeholders from the outset, and ensure that our maps are made freely available in a variety of ways (as raw data layers, in easy-to-use web maps, and as pdf reports in Spanish, Portuguese and English).
The project will involve three fieldwork phases spread over the first two years of the project:
- ~2 months in the Pastaza-Marañón Basin
- ~10 days in Yasuni
- ~10 days in Madre de Diós

Funding for these field trips, as well as lab analyses in the UK (carbon and nitrogen analysis), trips between St Andrews and Edinburgh, and money to attend meetings and conferences (UK and international), is included in the scholarship. In total £9625 of research costs are associated with this grant, along with in kind support from the Universities of St Andrews and Edinburgh for field equipment, knowledge exchange (workshops at the start and end of the project), and travel for PhD supervisors.