Research

What we do

A seasonal shift in δD observed over North West Africa by the TROPOMI instrument reveals distinct evaporation and mixing regimes.

Stable Water Vapour Isotopologues

Isotopologues serve as natural fingerprints that record the history of water evaporation, condensation, and transport.
They reveal exactly where a body of moisture originated and the conditions of that source region.
The ratio of heavy-to-light isotopes changes during phase transitions, exposing hidden thermodynamic processes.
We develop Level 2 satellite products from sensors operating in the Shortwave Infrared (SWIR), such as GOSAT and TROPOMI.
These products can provide a global benchmark to test General Circulation Models, aid in our understand tropical convection and cloud formation, and provde a constraint for models simulating precipitation to reduce climate prediction uncertainty

Emulating Radiative Transfer

Radiative Transfer Models (RTMs) are key tools that allow us to exploit indirect satellite measurements of the Earth's system.
While there is a large family of models which are generally faster at lower spectral resolution, higher-resolution line-by-line (LBL) models incur significant computational overhead.
We are developing a range of LBL RTM emulators that can reproduce the more computationally expensive calculations with high accuracy using modern machine learning approaches.

Example output from initial OptiRad model, which emulates LibRadtran at 0.001 nm at visible wavelengths over 6000 times faster than the original model.

Validation of retrieved clear-sky surface pressure from 12,000 global OCO-2 scenes obsevred over different land surfaces using our Hybrid OE scheme.

Hybrid Optimal Estimation

The statistical method of Optimal Estimation, or OE, is commonly used to infer geophysical quantities from satellite observations of the Earth system.
While OE, as a framework for satellite retrievals, has many benefits, in an era when satellite data is only getting more complex, it can struggle to produce results within the short latency required by many users.
Therefore, we are building Hybrid OE frameworks that can utilise machine learning models to speed up complex and slower components, whilst retaining the benefits of OE.

Energy Balance Climate Models

These are simplified, low-tier complexity diagnostic-prognostic climate models suitable for rapid climate process calculations and climate simulations on a personal computer or laptop
Reduction and linearization of sets of partial differential equations into a set of physical parametrisations defined by an array of variables
This group of models suitable for understanding the impacts of instantaneous radiative forcing and the energy distribution across various model interfaces, climate sensitivity experiments, understanding the energy budget of the system and drivers leading to the imbalance and hydrological cycle intensification
We are developing models that leverage the vast quantity of environmental intelligence from satellites, ground-based measurements, and modern reanalysis for a better understanding of how the hydrological cycle will respond to climate change.

nEBM temperature anomaly of the sea surface layer against the selected CMIP6 ocean surface temperature output and their differences relative to 1961-1990 reference period.

Mean evaporative source of precipitation over Indonesia (hatched region) for December to February and June to August 1986-2014 using the scaled-flux tracers in the Met Office Unified Model.

Atmospheric Moisture Transport

The transport of water vapour through the atmosphere is a key component of the hydrological cycle. It provides information about the dynamics within the atmosphere.
We are using non-isotopic water tracers in the Met Office Unified Model to identify changes in the atmospheric moisture transport.
This information can be used to understand changes in precipitation and evaporation patterns, helping to identify areas potentially affected by future climate hazards such as droughts and flooding.

Who we work with

Page updated

Google Sites

Report abuse