This project will develop new methods to study the changing temperature of the Earth’s surface, a need recognised to be very important by international space agencies and environmental scientists. A major challenge for Earth Observation (EO) is to understand the relative influence of global and local effects within the Earth system. Increasing human population and activities contributes very short-term local effects that can have significant long-term global effects. Multi-scale measurements are therefore required in order to understand how these changes evolve over time. Thermal variations in particular are important to isolate, as they contribute and invariably drive the overall energy balance. For this we need accurate (< 0.5 K) measurement of Land Surface Temperature (LST) at local (< 100 m) scales (to resolve fields and cities for example) and knowledge of the composition of the Earth’s surface. The provision of accurate high-resolution LST is one of the major challenges of EO science and has application in many areas including urban meteorology, climate change, forestry, agriculture and hydrology. This project will apply new mathematical approaches – optimal estimation and artificial intelligence (AI) to retrieve LST from remote sensing platforms.

Current infrared satellite EO sensors such as SLSTR on Sentinel-3 and VIIRS on JPSS Suomi typically offer LST approaching < 0.5 K in accuracy but their spatial resolutions are of order 1 km. Some higher spatial thermal imaging capability for LST measurement is available through Landsat and ASTER but their limited number of spectral bands (typically two bands at 11um and 12um), limited temporal sampling and lower accuracy (< 2.0 K) restricts scientific advances and uptake of applications from these missions. Indeed, by extending both the spectral coverage and spatial resolution the range of applications can be widened to include, LST, Sea Surface Temperatures (SST) in coastal waters, emissivity, land classification, volcanology, fire radiative power, cloud masking, aerosols, and trace gases. A sophisticated thermal instrument, most likely multispectral, will fly as a next generation Copernicus instrument (nominally called Sentinel 8). This project will provide key inputs to relevant new data techniques and instrumentation.

High-spatial resolution LST image of New York city (from ASTER)


This project will consolidate, review and refine a set of scientific and instrument. This will be achieved by developing a state-of-the-art optimal estimation (OE) methodology for simultaneous retrieval of both LST and Land Surface Emissivity (LSE) tested against AI. Current methods for retrieving both LST and LSE are limited to semi-empirical or use multi-channel separation methods where the temperature and emissivity are estimated separately. The OE and AI approaches will develop a novel algorithms where the LST and LSE are retrieved simultaneously. The approach will be generalised to apply to both existing sensors and new sensors.

The project will carry out testing of the methods on both simulations and real data from hyperspectral aircraft measurements (as well as the new Ecostress instrument to be flown by NASA on the International Space Station or ISS). Once verified, the new scheme will be used to identify the performances, modelling and design of new satellite sensors.

Training and Skills

CENTA students are required to complete 45 days training throughout their PhD including a minimum 10-day placement. In the first year, students will be trained on environmental science, research methods and core skills. Throughout the PhD, training will progress from core skills sets to master classes specific to this project’s themes. Specialist training will include atmospheric radiative transfer theory, non-linear data methods and infrared radiometry; The National Centre for Earth Observation will provide national-level training, meetings and business contacts. Additional space industry skills related to EO data and mission design will be carried out with the CASE partner, Airbus Space and Defence.


Year 1: Basic research skills training; familiarisation with literature, existing LST/LSE datasets and retrieval methods. Development of LST only retrieval.

Year 2: Completion of combined LST & LSE AI retrieval algorithms. Application to existing data from ASTER, MODIS and SLSTR. Work at CASE Partner, Airbus Space and Defence, in Stevenage, UK.

Year 3: Finalise data approaches/instrument design. Application of methods to NERC Owl aircraft and ISS Ecostress data. Further work at Airbus Space and Defence. Writing the thesis will take place during the final year, but papers will be published throughout the project. There will be opportunities to present at international meetings (UK and overseas).

Partners and collaboration (including CASE)

This project will be offered as a CASE award with Airbus Space and Defence. Prof John Remedios is Director of the National Centre for Earth Observation (EO) Observation (NCEO) and leads research on surface temperatures. Dr Gary Corlett is an expert in IR radiometry and surface temperature revivals. Prof Mark Sims is director of the university’s Space Research Centre. Dr Paolo Arrigo is head of Earth Observation and Science at Airbus Space and Defence.

Further Details

Contact John Remedios, University of Leicester, jjr8@le.ac.uk


Gary Corlett, University of Leicester, gkc1@le.ac.uk