Tree communities, airborne remote sensing and ecosystem function: new connections through a traits framework applied to a tropical elevation gradient

Lead Research Organisation: University of Oxford
Department Name: Geography - SoGE


What is the relationship between the composition of an ecological community and its ecosystem function? How do changes in community composition affect carbon and nutrient cycling? How does a shift in ecosystem productivity (e.g. through fertilization) feed through to changes in diversity? These are perhaps the most important questions in ecology day, in the context of direct human pressure on ecosystems and indirect pressure through global atmospheric change. Here we propose to collect the data and develop and evaluate a framework to advance these ideas, in the context of tree community composition of tropical forests.

We take advantage of three powerful tools that our team of investigators and project partners have developed: (i) an elevation transect of study sites in the Andes-Amazon where tree community composition and dynamics have been described in detail; (ii) airborne hyperspectral and lidar data that have recently been collected over this same transect, that enable determination of forest structure and chemistry in unprecedented detail, and (iii) a theoretical framework that utilises plant traits to propose a mechanistic approach to scale from community composition to ecosystem function.

We will add to these datasets by:
1. Conducting an extensive leaf and wood traits collection campaign for seven sites along this transect, and
2. Collecting data on nitrogen and phosphorus cycling.
Then we will develop a 3D model of the forest canopy of each plot (based on forest tree census and lidar data) to:
3. Explore the relationship between leaf traits and tree level characteristics (gross primary production, wood production, above-ground net primary production and nutrient cycling)
4. Scale from individual trees to the whole plot ecosystem characteristics (productivity, wood production, nutrient cycling)

Having developed this detailed framework for relating individual tree properties to plot-level function, we will try to simplify the system to see if ecosystem level properties can be derived from an understanding of the mean value and distribution of traits in a community. Finally, we will explore how well tree-level characteristics can be described by airborne hyperspectra and lidar, and thus explore whether it is possible to describe landscape level ecosystem functioning at the scale of thousands of hectares.

We have assembled a team of leading UK and USA researchers, and have an opportunity to make major advances and novel contributions to these important questions. Ultimately, we seek to acquire a mechanistic understanding of the relationship between forest community assembly and ecosystem level processes. Achievement of this goal would represent a major advance in ecology, in developing a both a theoretical and empirical toolkit with which to reach this goal.

Planned Impact

Scientific community engaged in understanding tropical forest ecology and function, and its sensitivity to environmental change

The scientific beneficiaries have been outlined in the academic beneficiaries section. Many of the needs of this community will be met through conventional routes (paper publication, conference presentation, guest seminars). All investigators and project partners have very strong publication records, and this will continue. In addition, the PI has a strong philosophical commitment to free, fairly rapid and unrestricted availability of our field data for use by the wider community, and have set up a portal where, for example, our recent Amazonian and African data are freely available (

Peruvian Junior Researchers

This project will continue its record of capacity building in Peru by training and further developing 10 Peruvian students and field researchers in data collection and analysis, with the aim of ultimately moving some of them onto PhD and MSc programmes, and publishing aspects of their data as first author -papers. In addition, we seek to train a wider body of students in both countries in the concepts and elements of the research, and encourage and assist local students to attach dissertation projects to the core research. To this end we will hold a wider participation and training workshop in Cusco at the start of the field campaign, and a data analysis workshop later in the field campaign. The workshop is explicitly budgeted for, and responsibility for organising and leading this workshop will be written into the PDRA contracts.

Public and schools seeking to understand and be motivated by modern tropical forest science

Using Pathways to Impact funding on a current NERC grant (focussed on African forests), we have just developed an innovative Geoweb Portal that brings together data, imagery and stories from our global network of ecosystem ecology sites (40+ sites, in Peru, Brazil, Bolivia, Chile, UK, Ghana, Gabon and Ethiopia). The site is launched in Dec 2011 and will be fully populated by mid-2012 ( Innovative aspects of the site include a "plots social network" where students in different regions (e.g. Peru and Ghana) can communicate with each other about techniques, science and questions, and a "public questions forum" where members of the public can pose questions that are answered by researchers and students. The whole Portal is designed in a form where it is easy for researchers and students to upload information, text and imagery.
We plan to incorporate stories and information from the current work within the framework of this Geoweb Portal, both from the field campaigns and from the attractive airborne 3D imagery. A key challenge can be to find time to make continually update and maintain the information on this portal. To incentivise this involvement, we have found that paying a fee of £10/hour for substantive contributions and really motivate and raise the quality of both UK and Peruvian contributions.

Peruvian students and public seeking greater general understanding of forests and climate

We will write some simple introductory material on tropical forests, carbon cycle and climate change (at a more basic level than the concepts researched in this proposal) in the context of our Andes-Amazon transect, that will be translated into Spanish and can be used by secondary schools, university students and the wider public. This material will be distributed to some schools and partner institutes, and made available for free download from the web. A 1-page flyer will also highlight the existence of this resource and be distributed in the Cusco region. A small amount of time to refine/distribute these materials will be written into the PDRA contract.
Description Using a network of sites across an elevation gradient from the Andes to the Amazon, this project aims to understand how variation in tree species composition, the characteristics of those species (functional traits) and forest structure shape variation in ecosystem properties such as primary productivity (the rate of carbon uptake from the atmosphere). Collectively, the studies show correlations in many functional traits and biogeochemical cycling along the gradient, but little variation in overall productivity. The shift in functional traits is driven by species turnover, and the associated shift in leaf traits and ecophysiology serves to reduce the overall sensitivity of productivity to the elevation gradient. These changes in the functional properties of canopies are detectable by airborne hyperspectroscopy at spatial and seasonal scales, suggesting the potential for airborne remote sensing to quantify and map variation ecosystem productivity and function. The results emerging from this environmental gradient represent a landmark study that demonstrates at multiple scales the interactions between plant traits, ecosystem composition and ecosystem function.
Exploitation Route We have shown that plant functional traits are closely related to forest carbon cycling, and that these traits are detectable by remote sensing. Other researchers may use our work as a foundation to predict current carbon cycle functioning across regions, landscapes, and perhaps globally. Furthermore, our results into the mechanisms underlying forest carbon cycle functions, and the remote detection of those mechanisms, may contribute to the ability of global land surface models to anticipate forest responses to global climate change.
Sectors Environment,Government, Democracy and Justice
Description The work under this award is still under way, but is showing exciting new ways to map forest canopy and function through remote sensing. We have presented this work to several NGOs, and to the Peruvian Ministry of the Environment.
First Year Of Impact 2014
Sector Environment
Impact Types Societal
Description ERC Advanced Investigator Grant
Amount € 2,500,000 (EUR)
Organisation European Research Council (ERC) 
Sector Public
Country European Union (EU)
Start 05/2013 
End 05/2018
Title GEMTraits Database and R package 
Description This database contains the plant functional trait data collected as a result of this NERC grant. It is implemented in a POSTGRESQL database, and runs on a server at Oxford University IT services. In addition, we have developed an R package that enables collaborating researchers to easily access and analyze the data coming from this project. 
Type Of Material Database/Collection of data 
Year Produced 2015 
Provided To Others? Yes  
Impact Many of the publications listed as a result of this grant have used this database for their analyses. 
Description Collaboration with Greg Asner, Carnegie Institution for Science 
Organisation Carnegie Institution for Science
Department Department of Global Ecology
Country United States of America 
Sector Academic/University 
PI Contribution We collected plant traits along an elevation gradient from the Andes to the Amazon, and coupled those with our long term research on forest carbon cycles.
Collaborator Contribution The Carnegie Airborne Observatory flew over the elevation gradient in 2013 and collected high-resolution hyperspectral imaging spectroscopy and LiDAR data. The Carnegie Spectranomics program also collected leaves and measured a diverse suite of leaf chemical traits from the same trees the Oxford team was studying.
Impact Publications resulting from this collaboration have been listed in the publication section.
Start Year 2013
Description Prof. Brian Enquist 
Organisation University of Arizona
Department Ecology and Evolutionary Biology
Country United States of America 
Sector Academic/University 
PI Contribution We collected plant traits along an elevation gradient from the Andes to the Amazon, and coupled those with our long term research on forest carbon cycles.
Collaborator Contribution The Enquist Lab at the University of Arizona analyzed individual leaf chemcal and morphological traits on the same trees being measured by Oxford.
Impact Publications resulting from this collaboration have been listed in the publication section. Prof. Enquist is spending his sabbatical with Prof. Malhi at the University of Oxford, funded by the Leverhume Trust.
Start Year 2013
Description Prof. Sandra Diaz 
Organisation University of Cordoba
Department Multidisciplinary Institute of Plant Biology (IMBIV)
Country Argentina, Argentine Republic 
Sector Academic/University 
PI Contribution We collected plant traits along an elevation gradient from the Andes to the Amazon, and coupled those with our long term research on forest carbon cycles.
Collaborator Contribution The Diaz lab collected data on herbivory on the same trees measured by Oxford in the elevation transect.
Impact Publications are listed in the publications section.
Start Year 2013