Large-scale distribution of functional diversity using trait-based approaches and remote sensing observations

Team members

Michael Schaepman

Felix Morsdorf

Summary

Mapping biodiversity from space requires up-scaling from local to regional (i.e. as established in project 3 in the first phase of the URPP). Scaling from regional to large-scale and/or global coverage requires careful assessment of separating interannual and intra-annual effects, in addition to ‘mixed pixel’ hypotheses (eg species are not pixels). Upscaled combinations of morphological and physiological traits, based on EBVs (essential biodiversity variables) will be transferred to satellite based LiDAR and spectral observations. The project will use regional scale data gathered in Switzerland, together with detailed biodiversity and trait-based inventory data and test and implement methods for two spaceborne sensors, namely Sentinel-2 and GEDI. Key contribution is the assessment of functional diversity measures at extended scales using morphological and physiological traits. Productivity-diversity relationships will be able to be assessed at large scales, across ecosystems.

Research

We will use the combination of morphological and physiological traits established in Schneider et al (2017) and extend those to large-scale observations. We will use regional scale earth observation data gathered on Laegern and Borneo sites to upscale our methods to regional and global scale. On this scale, multi-spectral images of Sentinel-2 will be used to retrieve physiological traits from inverse modeling in addition to modelled GEDI waveforms for morphological traits. GEDI waveforms will be modelled using airborne laser scanning data already available, this approach is well established and this way we eliminate any risk associated with a potentially delayed launch of the system.

We plan to proceed as follows:

1.         Review of current prioritized essential biodiversity variables (EBVs) for their suitability regarding their derivation at larger scale using the observational approaches mentioned above. Of special relevance is the added complementarity by fusion of various approaches, i.e. the inclusion of a LiDAR instrument, which is currently not foreseen for most EBVs. Test of EBV derivation on the core forest sites of the URPP (Laegern and Borneo).

2.         Development and implementation of unified models for the selected best performing EBVs based on step 1. Inverse and forward assessment on how well a limited set of traits can sufficiently describe large scale changes of functional diversity. Validation will be performed using URPP GCB obtained in-situ data, modeling as well as using existing trait databases and other datasets.

3.         Studying the differences originating from exhaustive (EO) and point-based (observational networks) sampling and trying to link changes in the EBVs with those of drivers of global change across the sites under consideration.