PhD: Linking Earth Observation, Soil Carbon and Hydrological Function Across UK Agricultural Landscapes

University of Leeds

Leeds, UK 🇬🇧

Agricultural soils are critical for food production, carbon storage, biodiversity and water regulation, yet many agricultural landscapes have experienced long-term declines in soil health. Regenerative agriculture has emerged as a promising approach to reverse these trends through practices such as reduced tillage, cover cropping, diversified crop rotations and improved management of soil organic matter. Despite growing interest among policymakers, farmers and environmental organisations, there remains limited evidence at regional and national scales regarding the extent of regenerative agriculture adoption and its measurable impacts on soils.

Recent advances in Earth Observation provide an opportunity to address this challenge. The European Copernicus Sentinel missions now offer frequent, high-resolution observations of agricultural landscapes, while national and continental-scale land cover and crop mapping products are becoming increasingly available. At the same time, advances in machine learning enable the extraction of soil properties from large Earth Observation archives.

This project will build directly upon recent research conducted at the University of Leeds that developed machine learning methods to identify bare soil exposures from Sentinel-2 imagery and reconstruct bare-soil spectra for estimating soil organic carbon. The PhD will extend these approaches by integrating additional Earth Observation datasets, including Sentinel-1 radar observations, Sentinel-2 optical imagery, crop type information, land-cover products and emerging Copernicus agricultural monitoring services.

The overarching aim is to determine whether regenerative agricultural systems can be identified from remote sensing observations and whether these systems exhibit measurable differences in soil condition and landscape function compared with conventional management.

The project will address three core research questions:

  1. Can regenerative and conventionally managed agricultural fields be distinguished using Earth Observation data and machine learning approaches?
  2. Do fields managed under regenerative practices exhibit systematically different soil organic carbon dynamics and bare-soil characteristics compared with conventional systems?
  3. Can Earth Observation reveal differences in hydrological function, specifically soil moisture response, associated with regenerative management?

To answer these questions, the student will combine field-scale and regional datasets describing agricultural management with multi-sensor Earth Observation data available through cloud-computing platforms such as Google Earth Engine, and developing your own code using Machine Learning and explainable AI.

The research will provide some of the first large-scale assessments of whether regenerative agriculture can be monitored routinely from space. The outcomes will contribute to understanding how agricultural management influences soil carbon storage, soil exposure and water regulation, supporting evidence-based policy development related to climate mitigation, environmental land management and sustainable food production.

The successful student will receive interdisciplinary training in Earth Observation, remote sensing, machine learning, geospatial analytics and environmental data science. They will gain extensive experience working with large satellite datasets, develop advanced coding and modelling skills, and engage with stakeholders interested in monitoring sustainable agricultural systems.

55 days remaining

Apply by 17 January, 2027

POSITION TYPE

ORGANIZATION TYPE

EXPERIENCE-LEVEL

DEGREE REQUIRED

Register for World Toilet Summit 2026 - Thailand - November 17-19, 2026