Plant Optical Types to Predict Ecosystem Impacts of Plant Invasions (INPLANT)

Start-End 01/12/2015 -  30/11/2018
Programme STEREO 3
Contract SR/01/321
Objective Plant species invasions are among the most important threats to the functioning of the earth’s ecosystems. Invasion biologists have mainly focused so far on the effects that invasive plant species have on native populations and communities. The response of ecosystem functioning to invasion has received considerably less attention. Although it has been proposed that invasive species impact can be predicted based on a comparative analysis of their functional traits and these of the native community, this approach has rarely been used and is potentially compounded with problems regarding feasibility and potential for generalization. Here, we will explore the capability of new remote sensing instruments to define optically distinguishable functional types (‘optical types’) as a means to quantify the effects of invasive plants on ecosystem functioning. Current state-of-the-art hyperspectral sensors allow to characterize the vegetation chemistry and canopy water content, as such allowing a discrimination between subtle physiological differences among plant species, and at the same time providing a straightforward link to ecosystem effects of the species. We predict that the optical types to be developed will outperform, or at least complement, the conventional functional trait approach when predicting changes in ecosystem functioning through plant invasion, both regarding accuracy and potential for generalization up to larger spatial scales.

The general objective of this research proposal is thus to develop a novel ‘optical types’-based approach to evaluate and to predict the impact of invasive plant species on ecosystem functioning. This new approach will not be based on functional plant traits but on optical plant traits. The principal idea behind the proposed approach is that invasive plant species often have physical-chemical properties that differ from native species. These physical-chemical properties can be directly linked to ecosystem functioning changes following invasions, and they can be largely quantified based on the optical reflectance spectra of the species, which are detectable due to recent developments in hyperspectral remote sensing technology.

In this proof of concept study we will focus on herbaceous species that grow in open habitats.

Method We will study three model plant species that are invasive in NW Europe, in their most common habitats. For each of the studied systems we will quantify ecosystem functions along an invasion gradient, and retrieve trait data for the native and invasive communities from both existing databases and field measurements. Optical types will be derived based on hyperspectral data from individual plant species, collected using both field spectroradiometers and airborne spectroscopy. We will use multivariate statistical techniques (including Redundancy Analysis (RDA) and Partial Least Squares Regression (PLSR)) as well as process-based leaf-canopy optical models (LCOMs) to identify those optical traits and conventional functional traits reflecting ecosystem function changes following invasions.

This proposal should be considered as a proof-of-concept study, which, if successful, will result in a better understanding of the impact of invasive plant species and deliver a useful tool for performing environmental risk assessment analyses in invaded ecosystems. The presented concept of optical types offers an innovative and potentially transformative approach for linking optical properties (accessible with remote sensing) with ecosystem function. The approach is expected to be complementary with more traditional functional trait-based approaches, on top offering the additional benefit for upscaling to the landscape level. The optical type approach is also very promising in the broader global change field, as it may allow to predict ecosystem functioning response to other global change processes such as climate change or large scale eutrophication.


As already mentioned, from an operational perspective our approach has potential as a Risk Assessment tool. A Risk Assessment of the impact of the invader on its hosting ecosystem is indispensable to provide the critical information to support decision-making on what measures have to be implemented and who should enforce them. Once a potential invader is identified,  its spectral properties can be measured using a field spectroradiometer (i.e. proximal sensing). The invader’s optical type can then rapidly be determined by comparing the spectral properties of the invader to a publically available online library of optical types that remains to be established. Our results will allow to initiate the establishment of such a database that links spectral properties of individual species to potential impacts on ecosystem functions. This database should be considered similar to the traditional trait databases that provide lists of trait values for each species. So, once we have identified the optical type we immediately know the harmful potential of the invader.
Furthermore, the project results will contribute to the set up of a remote sensing supported pipeline for an operational warning system. Yet, as mentioned earlier, this proposal should only be considered as the initial step in which we try to proof the concept. Beyond the scope of this initial work, additional further work will be needed to evaluate the portability of the approach to other invasive plant species, ecosystems and ecosystem functions. Providing generic methods of monitoring ecosystem function over large areas and for a diverse range of species while developing operational early warning systems is beyond the scope of this proof of concept study, but is the long term objective of the recently established Remote Sensing and Terrestrial Ecology group headed by the PI.

Website link
Team Member: FEILHAUER Hannes FAU Erlangen-Nürnberg University - Institut für Geographie
Coordinator: SOMERS Ben KULeuven - Division Forest, Nature and Landscape
Team Member: HONNAY Olivier KULeuven (Katholieke Universiteit Leuven)
Sensors used
Related presentations BEODay 2016 - 13. INPLANT: Plant optical types to predict ecosystem impacts of plant invasions
Related publications
Datasets used