Periodic Reporting for period 3 - BACI (Detecting changes in essential ecosystem and biodiversity properties – towards a Biosphere Atmosphere Change Index: BACI)
Período documentado: 2017-10-01 hasta 2019-03-31
The project is divided into seven scientific work packages:
1) BACI explores the potential of integrating existing data archives and new observations as they become available (i.e. from Sentinel 1 & 2). Of particular importance is the integration of optical and radar data. While optical data allow tracking the phenology of the ecosystem and responses to short term anomalies in ecosystem functioning, radar data are expected to reveal structural properties of ecosystem and their changes. BACI aims to provide a generic, scalable framework for combining data from multiple wavelength.
2) BACI works on the integration of a wide range of in-situ data from different sources. As with the EOs, the project puts much emphasis on realistic uncertainty estimates. The main components of the in-situ database are global plant traits at various aggregation level for 18 traits, ecosystem parameters derived from eddy flux measurements across ecosystems, synthesis data sets of plants and bird diversity for Europe; synthesis data sets of annual ring data sets for Europe; and finally, vegetation structure and biomass data derived from LiDAR, including stand information. We also conduct a series of experiments on integrating these data with EOs.
3) The project aims at deriving new downstream data products. For example, we seek to produce global data products on land-atmosphere fluxes of carbon dioxide, water, and energy or ecosystem scale phenologies. As these variables are can only be observed in-situ, we need to integrate EOs and in-situ observations. In BACI we realize this integration with machine learning methods. BACI focuses on the derivation of so-called ""Essential Ecosystem Variables"", i.e. variables that are essential for the monitoring of the fundamental interactions and feedback in the earth system between biosphere and atmosphere.
4) A key element of BACI project is the development of a generic index of change. This index should be able to detect a variety of different types of anomalies and extremes. This novelty index enables the detection of sudden events and abnormal changes in multivariate EO data streams. For now, BACI focuses in particular on the identification of abrupt changes relevant to essential ecosystem variables.
5) To guarantee highest quality standards, BACI performs rigorous quality assessments. Both the generated downstream data products as well as the generic index of change are validated with independent reference data. Different regional validation sites are used for validation efforts.
6) BACI embraces a component that allows us to contextualize detected changes. In particular, we consider socioeconomic changes as key drivers of change in land ecosystems.
7) BACI explores new ways to ingest space data into the study of biodiversity patterns and its changes. For instance, we explore to potential of radar data as predictors for species distribution modelling. The project also explored the value of the change index for monitoring protected areas in Europe.
"
· Generating the concept and software for computing the surface state vector. Providing the first version of the state vector to the partners and gathering their feedback.
· Generating the consistent database of the in-situ observations and providing the output to various scientific initiatives within and beyond the project.
· Generating a series of downstream data products. In particular, BACI produced four global data streams in half-hourly fluxes of CO2, H20, and energy at 0.5 degrees. We also worked in multiple efforts to specialize plant functional traits and interpretation as a function of climate and edaphic factors.
· Generating algorithms for change detection and computing a first index of change, which is currently under evaluation.
· Generating a validation framework that can now be fueled with the outputs of the project, i.e. the index of change and downstream data products.
· Conceptualizing the integration of social change indicators to the project.
· Evaluating the potential of new Earth observation for species distribution modelling, and the assessments of protected areas in Europe. Also assessing beta-diversity patterns with BACI products.
· Establishment of a general data exchange portal.
· Realization of multiple outreach and dissemination activities across scientific projects and considering a wider target group.
Technically, we have developed an earth observation assimilation approach to synthesize multiple sources of optical and microwave data in a single surface state vector (SV). The prototype of this method is very general and goes beyond the state of the art. We expect that this idea would be very relevant to use a wide range of space data beyond BACI,
We also implemented machine-learning models for high-frequency CO2 and energy flow upscaling by merging EO and in situ data. Now that this product is ready, we expect this data product to significantly advance the research field and provide an unprecedented reference data set for the next generation of global Earth system models. This is highly relevant for societal challenges such as the assessment of climate change through coupled climate models of the carbon cycle. But we have also made major advances in terms of upscaling and interpreting plant trait data and interpreting their ecological meaning at the global scale.
We have also explored the potential of EOs for biodiversity monintoring i.e. the use of radar data to describe habitat heterogeneity. In general, we believe that research into satellite data to promote biodiversity research is also of great relevance for GEO BON, for example. We now participate in GEO BON activities, like the data management tasks and have also interacted with regional BONs to explain the concept of data as developed here. We assume that we are in a position to strengthen biodiversity research beyond the current state of the art in terms of prediction accuracy and spatial detailing.