Dynamic Vegetation Modelling for Climate Prediction

Lead Research Organisation: UNIVERSITY OF EXETER
Department Name: Engineering Computer Science and Maths

Abstract

This project will ensure that JULES-ED becomes a central part of the next generation of the Hadley Earth System Model (HadGEM3), so that interactions between the land-surface and climate can be accounted for with unprecedented realism. Firstly this requires confronting the currently uncoupled version of JULES-ED (including SPITFIRE, ECOSSE and FUN) with a wide range of land-surface observations. Comparison to observations will be used to focus further model improvements and to constrain uncertain model parameters (WP1). The improved version of JULES-ED will then be coupled to HadGEM3 and tested in climate-vegetation simulations of the 20th century. These historical HadGEM3-JULES simulations will be evaluated against observational datasets of climate and land surface characteristics, and compared to simulations with the current HadGEM-MOSES/TRIFFID model (WP2). Once JULES-ED has been shown to be performing well, a series of parallel simulations for the 20th century will be carried-out at the Met Office to isolate the impacts of the new components (i.e. nitrogen cycling, wildfires and land-use change) on the evolution of the land carbon sink and vegetation structure and functioning through the 20th century (WP3). Finally, the coupled HadGEM3-JULES model will be integrated (i) over the standard AMIP period (1979-2008) to assess the importance of interactive vegetation for seasonal to decadal predictability; and (ii) over a 21st century concentration scenarios (the RCP4.5 scenario used for the IPCC 5th Assessment Report). The simulated changes in climate, ecosystem services (e.g. provision of primary productivity and fresh water) and land-carbon sink will be compared to existing Hadley Centre runs to isolate the impact of JULES-ED on future projections (WP4).
 
Description Through this grant, we made improvements to the dynamic global vegetation model (DGVM), JULES; and we collected data for evaluating and benchmarking Earth System Models. The UK research community uses JULES to study land surface processes pertinent to the terrestrial carbon cycle - for example plant photosynthesis, respiration, and land use changes due to human interference. DGVMs classify global vegetation into Plant Functional Types (PFTs) and predict the global vegetation distribution. We increased the diversity of vegetation types represented in JULES by expanding the number of PFTs from 5 to 9. The PFTs now represent a broader and more realistic range of plant responses to climate variability, climate change, and increasing atmospheric CO2 levels. This affects the response of the land surface to climate and the carbon cycle/climate feedbacks.

In developing the PFTs, we relied on two primary sets of observations. First, parameters for the new classifications were based on a global database of plant characteristics (the TRY database: TRY-db.org). Second, new versions of the model were optimized and evaluated against observations from towers in the global FLUXNET database (fluxnet.ornl.gov). These towers measure the exchange of latent heat, sensible heat, and CO2 between the land and atmosphere. The CO2 flux can be decomposed into the uptake of CO2 through photosynthesis (or gross primary productivity, GPP) and the release of CO2 through respiration.

Development of the PFTs was based on rigorous comparison between modelled and observed fluxes at the FLUXNET sites. The seasonal cycle of CO2 exchange was improved in JULES and biases were reduced. For example, the model previously had a bias that resulted in too much CO2 being released to the atmosphere during the Northern Hemisphere summer. This made it impossible for fluxes from JULES, fed through a transport model, to match the observed seasonal cycle of atmospheric CO2 in Barrow, Alaska. Now there is increased photosynthesis and reduced respiration during the summer, which combine to improve the net ecosystem exchange of CO2 in the northern high latitudes. In addition, we have included new data sources for benchmarking the model, including a globally gridded dataset of GPP based on the flux tower network and remotely sensed chlorophyll fluorescence, which correlates with GPP.

We optimized the model against observed GPP, respiration, and latent heat flux. Through this exercise we identified the parameters that are most helpful for optimisation, and realistic upper and lower bounds for some parameters available from the TRY database. We also realized that optimising for single sites does not yield a universal set of parameters for the model (in other words - what is optimal for one site may not be optimal for another). This led to a refined optimisation technique using observations from several sites at once, and this work is underway. Another important lesson from the optimisation is that improved parameters alone were not enough to match the observations in the tropical forest sites. At these sites, a combination of deep roots and relaxed soil moisture stress were needed.
Exploitation Route Newly developed version of JULES is being used by the MetOffice for implementation in UKESM the UK earth System model to be used for climate simulations as part of the next IPCC assessment.
Sectors Environment
 
Description As mentioned before, The new version of JULES developed within the project is now being implement in the UKESM for use in future IPCC climate projection.
Sector Energy
Impact Types Policy & public services
 
Title JULES dynamic vegetation model with 9 Plant Functional Types 
Description Dynamic global vegetation models are used to predict the response of vegetation to climate change. They are essential for planning ecosystem management, under- standing carbon cycle-climate feedbacks, and evaluating the potential impacts of climate change on global ecosystems. JULES (the Joint UK Land Environment Simulator) represents terrestrial processes in the UK Hadley Centre family of models and in the first generation UK Earth System Model. Previously, JULES represented five plant functional types (PFTs): broadleaf trees, needle-leaf trees, C3 and C4 grasses, and shrubs. In this new version of JULES, trees and shrubs were split into deciduous and evergreen PFTs to better represent the range of leaf life spans and metabolic capacities that exists in nature. Also, we distinguished between temperate and tropical broadleaf evergreen trees. These two changes result in a new set of nine PFTs: tropical and temperate broadleaf evergreen trees, broadleaf deciduous trees, needle-leaf evergreen and deciduous trees, C3 and C4 grasses, and evergreen and deciduous shrubs. 
Type Of Material Computer model/algorithm 
Year Produced 2016 
Provided To Others? Yes  
Impact Publication of the new model in GMD (Harper et al., Geosci. Model Dev., 9, 2415-2440, 2016) JULES 9PFT version available to the overall JULES community Inclusion of this new version of JULES in the UKESM at the MetOffice for use by the MetOffice and the climate community for CMIP6 supporting the IPCC. 
 
Description Collaboration with the MetOffice on dynamic vegetation model developments in JULES 
Organisation Meteorological Office UK
Country United Kingdom 
Sector Academic/University 
PI Contribution Anna Harper, the PDRA hired on that NERC project continued to work with us at the University of Exeter, and continue the efficient collaboration with the MetOffice. The new version of JULES developed by Anna Harper accounts for 9 Plat Functional Types (as opposed to 5 in the previous version). Anna collaborated with the MetOffice land surface team (Andy Wiltshire) and the UK-Earth System Model (UKESM) team (led by Colin Jones) to ensure that the new version of JULES would be included in the current version of UKESM in order to participate to teh CMIP6 exercise and provide inputs to the IPCC. Anna spent a significant fraction of her time, going to teh MetOffice, integrating and testing the new version of JULES before inclusion in the ESM
Collaborator Contribution The MetOffice was in charge of the coupling of JULES 9PFT within the ESM and benchmarking the new version.
Impact The main output, as described before is a new version of JULEs, with 9 PFTs within the UKESM. This is the most significant JULES development regarding vegetation dynamics over the last 10 years.
Start Year 2011