Improved soil moisture detection to understand climate threats to Amazon rainforests

See also: https://eo-cdt.org/projects/improved-soil-moisture-detection-to-understand-climate-threats-to-amazon-rainforests/

Tropical rainforests exchange more carbon dioxide with the atmosphere than any other land biome. This partly reflects their geographic size, but also high metabolic rates driven by warmer temperatures and generally high moisture availability, together with land use change. Changes in the loss or gain of carbon from tropical forest ecosystems can substantially affect the global carbon balance (1), but accelerating climate change threatens to cause these forests to act as much smaller absorbers of carbon dioxide than in previous decades (2).  In Amazonia an increasing frequency of climate extremes, particularly drought and warming, are key drivers of these changes, but the effects remain poorly understood, limiting our ability to predict change and related risk (3). Adequate monitoring of soil moisture availability is a fundamental requirement to understand changes in the carbon balance of tropical rainforest. Satellites offer powerful monitoring capability, but alone they are limited by resolution of the signal in the data. This PhD will connect novel intermediate-scale non-invasive measurements of soil moisture using new ground penetrating radar (GPR) with the analysis of satellite-based detection and direct measurement of soil moisture using soil moisture probes at two field sites in Amazonia.

All three measurements are unified by using the dielectric constant of water to detect change in moisture content in soil, but they differ in their resolution and error. The use of GPR to detect and understand changes in soil moisture will enable us to fill a critical gap between high resolution/small area sensor-scale and low resolution/large area satellite-scale measurements. The outcome will allow large advances in the monitoring of forest responses to drought, leading to completely new insight into ecosystem functioning at the site level, and new capability to develop early warning methods for large-scale detection of regionally-dangerous drought risk.

The PhD student will gain expertise in high-level computational skills including development of multivariate machine learning algorithms, advanced change detection techniques applied to microwave-based satellite data (e.g. Sentinel-1 C-band SAR) and advanced statistical approaches to signal detection. In addition, the student will deliver field datasets from two well-studied locations in Amazonia: one in the east, at the world’s only long-term (> 5yrs) hectare-scale drought experiment in tropical rain forest (4); and a second in the south, where drought and warming threaten Amazonian forests strongly (5). Both sites have exceptional datasets on forest physiology and soil moisture (up to 20 years of soil moisture data). The student will test new GPR-based measurements with data from networks of sensors already installed at each site, and to connect their analysis with satellite data-streams, thus developing a powerful understanding of soil moisture constraints on Amazonia rainforests.

The supervisory team has a sustained track record in understanding drought impacts on tropical rainforests, and in high-level statistical and computational analysis of biological and remote sensing data, publishing at the highest levels (eg, 6,7). The student will collaborate with two leading field Brazilian rainforest research groups in the Brazilian states of Mato Grosso and Pará (eg. 8), and will also collaborate with NASA’s Jet Propulsion Laboratory, being invited to spend a period of training at JPL with Prof S. Saatchi (eg, 9). The student will be based at the within the School of Mathematics and the School of Geosciences at University of Edinburgh, which has an exceptional track record in radar remote sensing and tropical forest ecology; the wider supervisory group (Leeds, Edinburgh and NASA/JPL) will bring additional expertise in both areas and statistical analysis, and substantial extra scientific networks.

SENSE CDT

This PhD is part of the NERC and UK Space Agency funded Centre for Doctoral Training "SENSE": the Centre for Satellite Data in Environmental Science. SENSE will train 50 PhD students to tackle cross-disciplinary environmental problems by applying the latest data science techniques to satellite data. All our students will receive extensive training on satellite data and AI/Machine Learning, as well as attending a field course on drones, and residential courses hosted by ESA (Rome). All students will experience extensive training on professional skills, including spending 3 months on an industry placement. See http://www.eo-cdt.org