Navigation auf


URPP Global Change and Biodiversity

Project 3: Integration of cross-scale effects minimizing feedback omission: a latitudinal gradient approach

Project Team

Michael Schaepman
Felix Morsdorf
Fabian Schneider
Carla Guillén Escribà
Chengxiu Li

Research aims - This project deals with the problem of scale between drivers of biodiversity change. The project will assess regional-scale genetic and functional diversity and associated ecosystem functions, and up-scale this information across the latitudinal range of test sites.

Assessing cross-scale effects by extending scales – Cross-scale effects are particularly important when assessing biodiversity at large, regional and small scales together. However, conflicting or even disagreeing scale effects may occur when using observational data for which temporal, spatial or even spectral scales cannot be aligned. Multi-temporal observational data are used to derive changes in plant functional traits at large spatial resolution for large scale applications. On the other hand, regional-scale data are used to assess floristic diversity, canopy biochemistry and structure, as well as species- and ecosystem-level productivity. Both approaches have been proven to deliver relevant results for the assessment of biodiversity from space, but are prone to feedback omissions, when scale mismatches occur.

The regional-scale approach will link leaf optical trait models and spectral databases with genetic diversity. Selected sites along the latitudinal gradient will be monitored using a high resolution imaging spectrometer (Airborne Prism Experiment (APEX) and full waveform laser instruments. Two model species important for the supply of ecosystem services in temperate environments will be investigated (common beech [Fagus sylvatica]and Norway spruce [Picea abies]). Monitoring and modelling of genetic and phylogenetic diversity will be achieved by analysis of leaves along phenological cycles and subsequent spectral, chemical and genetic analysis. Combined high-spectral-resolution remote sensing, integrated with spatial-genetic variation models will support large-scale inferences about changes at the genetic level that are difficult to obtain directly. Genetic diversity will be surveyed, in collaboration with project 5, first with microsatellite markers. The genome size of Fagus sylvatica is relatively small (543 Mb), and some genomic information for the family Fagaceae is publicly available. If appropriate, genome-wide polymorphisms can be surveyed using next-generation sequencing (NGS). Population structure and demography will be estimated, and association mapping will identify genes responsible for chemical and spectral diversity. Genome-wide expression analysis using NGS and microarrays will reveal environmental triggers of phenological events. Integration of scale effects will use coupled radiative transfer models (Schaepman et al., 2009), allowing attributing cause–effect relationships of cross-scale effects.

The large-scale approach will use globally continuous multi-temporal data and analyse them for vegetation productivity hotspots across all test sites of this project. Spatio-temporal modelling will allow detecting greening and browning trends as well as changes of phenological cycle length between 1983 and 2011. Observations of terrestrial net primary production are available from time series. Hotspot analysis will be validated using temporally discontinuous high-resolution data. This will facilitate regionally validated global maps of ecosystem service delivery, indicating regions of most rapid productivity changes along the latitudinal range of test sites.

Converging scales of Earth observations, models and field experiments – Much of global change is driven by feedback mechanisms at spatial and temporal scales smaller than those currently incorporated in global change models. This severely limits ability to predict, mitigate and adapt to environmental change at local and regional scales. Modelling, experiments, and data are gradually converging across spatial and temporal scales where processes at their typical length scale can be compared. 3D Vegetation Laboratory (3DVegLab) is a project to fully characterize a temperate mixed forest with the dominant species Fagus sylvatica and Picea abies using volumetric and geometric 3D-radiative transfer models and full-waveform laser scanning. The core site at Laegeren is 300 x 300 m and allows for incorporation of relevant plant functional traits into parameterized 3D models of vegetation structure. For temporal surface aspects at this site, FLUXNET, AERONET and PhenoCam infrastructure are used. Estimates of net primary productivity (NPP) from remote sensing and FLUXNET data will be cross-calibrated and validated in combination with dynamic vegetation models (e.g., LPJ-GUESS) and footprint models. The 3DVegLab design allows for explicitly research of scaling issues. For example, we can test the explanatory power of field- and remotely-measured variables to predict ecosystem processes (Morsdorf et al., 2009), and the potential continuous scale-dependence of biodiversity-ecosystem processes relationships.

Coupling physically based soil–vegetation–atmosphere transport models – The use of empirically based indices has prevented remote sensing from establishing objective predictions of global change. The use of physically based soil–vegetation–atmosphere-transport models allows inference of estimates of canopy characteristics from remote sensing data (Schaepman et al. 2009) and estimates of carbon fluxes and pools as well as biodiversity in space and time. Cutting edge systematic approaches using spectral fingerprinting of dominant species allow spectral databases to serve as modelling baseline for inversion- and radiance-based approaches. Important plant functional traits can be predicted using these approaches, from regional to large scales. Novel approaches to remotely measure physiological processes can be combined with photosynthesis models to estimate plant functional traits specifically related to productivity. These results can be aggregated to produce abundance distributions of plant functional types (PFTs). It is the goal of this approach to produce continuous fields of abundance distributions of PFTs at regional to global scale. Validation is performed using extended ground networks, such as the sites of the Swiss Biodiversity Monitoring programme.

Expected contributions to research theme – Increasingly, Earth observation data and products are used to assess biodiversity from space. Since large-scale spectral, spatial and temporal high-resolution instruments have become available, significant advances have been made towards monitoring of biodiversity from space. Sound forecasting of global change and biodiversity will not only rely on the above methods and research, but will increasingly include legacy data and field observations. This will render ours a policy relevant approach with substantial contributions to the Biodiversity Observation Network (BON) of the Group on Earth Observation (GEO).

Weiterführende Informationen

Biodiversity and Earth observation

GEO Biodiversity Observation Network GEO BON | Airborne Prism Experiment (APEX)