JULES EMulator of ecosystem services (JEM)

Lead Research Organisation: University of Exeter
Department Name: Mathematics


Ecosystems provide many services, such as timber, food, and water. They are a natural protection against floods, store carbon, and are locations of cultural heritage and recreation. To be sustainable and profitable, future landscape decisions need to consider the various services provided by ecosystems as well as their responses to climate change. The UK's commitment to reducing emissions (to at least 80% below 1990 levels by 2050) means that carbon storage and greenhouse gas emissions from land are central to future land use and management. More ambitious emission reduction targets and their implications are currently being discussed, increasing the need to build land use decisions upon an understanding of the integrated water and biogeochemical cycles.

Process-based land surface models are excellent tools for predicting carbon, water, and surface heat exchanges between the land and atmosphere. The UK's community land surface model, JULES, is used in the Met Office Unified Model (UM) framework for applications ranging from climate change to weather forecasting. JULES predicts productivity of vegetation, carbon accumulation in vegetation and soils, soil water content, runoff, and emissions of greenhouse gases from vegetation and soils - meaning it can be used for evaluating the impacts of land use change and climate change on several ecosystem services. Within the UM, JULES is run at a km-scale resolution for weather forecasting. JULES is also used for evaluating the global carbon cycle: for example, each year JULES is used to calculate the net land sink of carbon and emissions due to land use change in the Global Carbon Project. However, in these simulations the spatial resolution (>50km) is inadequate for addressing landscape-scale issues.

The computational cost of running JULES at km-scale resolution is very high. At a 1.2km resolution, there were nearly 80,000 grid cells. Running on 144 cores, a 10-year simulation took 70 hours. Decisions on land use changes should incorporate longer timescale changes and uncertainty quantification. An emulator of JULES would allow for more robust statistical calculation of changes in carbon stocks and the water cycle and their uncertainty and open the way for better calibration of JULES. During this short project we will design and build emulators of JULES to inform future evaluation of ecosystem services related to the carbon and water cycles.

Planned Impact

The emulator will enable statistical analysis of changes in productivity that would occur due to land use change. This is useful for decision makers involved in land use and land management. To ensure our work produces useful outputs we will engage with Defra, who we see as a key stakeholder in this project. We will also engage with the Project Coordination Team early in the grant to ensure our project integrates with other work in the program and to maximize the opportunity for wider engagement. This is essential as the limited time and resources for this project limit the engagement we can undertake on our own.

This project will serve as an essential building block in a larger project, which will utilize emulators of more ecosystem services. This will be useful within the research community and for those involved in landscape decisions, enabling an improved understanding of impacts of land use and land management decisions on ecosystem services under a changing climate. For example, the UK government is considering input from the Committee on Climate Change on Building a Net Zero economy, which will likely include one or more forms of land management for carbon storage. Understanding how land use change and land management will affect the carbon cycle, as well as quantifying the uncertainty in the impacts, will be essential for planning future land use.
Description Our project focused on building an emulator of a process-based land surface model. Due to the short duration of the project, our objective was to lay the foundation for future work by building an initial Gaussian process emulator to be used for global sensitivity analysis, uncertainty analysis due to uncertainty in model parameters and climate change, and calibration. We demonstrated the capacity for this emulator and discovered some methodologies to reduce the computational cost of building the emulator.
Exploitation Route The example of the emulator built in this project could be applied to any land surface model. We have incorporated the early work in this project into a successful proposal for a UKRI Greenhouse Gas Removal Demonstrator project.
Sectors Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Environment

Description Our early work on building an emulator of a high resolution land surface model for the UK has been incorporated into a Greenhouse Gas Removal demonstrator project. The outcomes of this project will result in a decision support tool for policy makers. The project and its associated outputs will occur over 2021-2026.
First Year Of Impact 2021
Sector Environment,Government, Democracy and Justice
Impact Types Policy & public services

Description Greenhouse Gas Removal Plus (GGR+): Sustainable Treescapes Demonstrator & Decision Tools
Amount £5,621,807 (GBP)
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 04/2021 
End 11/2025
Title Gross Primary Productivity simulations of Great Britain for emulation, 2001-2010 from JULES land surface model 
Description This data is an ensemble of Joint UK Land Environment Simulator (JULES) simulations, ran for a select set of 1kmx1km grid cells in Great Britain, each with a different set of parameter values, from 2001 to 2010. The data includes simulated Gross Primary Productivity (GPP) for 5 different plant functional types on an 8-day average. (Broadleaf trees, Needleleaf trees, C3 grasses, Shrubs, and Cropland) as well as the weighted combined sum. These simulations were obtained to facilitate statistical emulation, which is why a wide range of grid cells and parameter values were used (both also provided in the data set). 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
Impact This forms background research for further use of land surface models and emulators in the BBSRC project GGR+. 
URL https://catalogue.ceh.ac.uk/documents/789bea37-0450-4822-9857-3dc848feb937
Description Forest Research 
Organisation Forest Research
Country United Kingdom 
Sector Public 
PI Contribution Through follow-on work from JEM, I now work with FR in modelling future tree productivity in the UK. We will use the techniques demonstrated in JEM to build emulators of models used by my team and by FR.
Collaborator Contribution FR is the authority on UK forests, and the people I work with are experts in forest dynamics. So far they have provided me with data for calibrating and developing my model. In the future we will further develop our models in tandem.
Impact We are now working together through the BBSRC project: Net Zero Plus.
Start Year 2020
Title Gaussian Process Emulator of carbon and water cycle for the UK 
Description We developed methodology to emulate a land surface model, focusing on carbon and water cycle impacts of land use change. It is in the process of being published at which point the code will be publicly available. 
Type Of Technology Software 
Year Produced 2020 
Impact This method will be used to develop a decision support tool for policy makers, related to land use decisions and the UK's commitment to reach net zero emissions of carbon by 2050.