Community Research and Development Information Service - CORDIS

H2020

BACI Report Summary

Project ID: 640176
Funded under: H2020-EU.2.1.6.

Periodic Reporting for period 1 - BACI (Detecting changes in essential ecosystem and biodiversity properties – towards a Biosphere Atmosphere Change Index: BACI)

Reporting period: 2015-04-01 to 2016-03-31

Summary of the context and overall objectives of the project

Space data offer multiple opportunities for monitoring ecosystems and their transformations, e.g., in response to human interventions or climate extremes. Today, these data are available in unprecedented spatial, spectral, and temporal resolutions. This development is complemented with the increasing availability of a wide range of ground data on diverse aspects of ecosystem functioning and ecosystem structure, and other parameters relevant to fully describe the functional biogeography of ecosystems. In this broad context, the BACI project aims to tap into the yet-to-be realized potential of existing and scheduled space-borne Earth observation (EOs) in conjunction with ground data to derive new essential ecosystem variables and detect changes in ecosystem functioning. The project is coarsely separated into seven highly interwoven scientific work packages (complemented by one management and one dissemination work package):
Firstly, BACI is exploiting existing data archives and new observations as they become available. Of particular importance is the equitable capitalization of optical and radar space data, which has been rarely realized in this scientific arena. While the optical data are often interpreted in terms of ecosystem functioning, radar data mostly reveal ecosystem structure – thus providing a highly integrative view. A key aspect for these endeavors is to provide a generic, scalable framework for combining data from multiple streams. The result is a system ‘state vector’ representing the state of a point/region on the land surface at a given time with defined uncertainty. The output will be a description of surface state with constrained uncertainty that can be ingested directly into various types of further analysis, without requiring conversion to higher-level model-derived products both regionally, with ‘cut-outs’ and options for near-real time assessments. BACI is presently developing the methodology for creating a state vector from both archived and newly upcoming data.

Secondly, we need to address the diversity of ground data to later enable the BACI team to translate space data into new data products. Hence, BACI is dedicating resources to set up a ground data database, which includes harmonized and standardized data from different sources. In addition to their obvious uses, we are also working to provide realistic assessments of uncertain for all the variables included in the database. Five main data types were identified as major components of the ground database, i.e. plant traits and phenology; ecosystem parameters derived at Fluxnet sites; synthesis datasets of biodiversity in plants and birds; synthesis dataset of tree ring records; and finally LiDAR derived vegetation structure and biomass data, including inventory information. As each dataset type is generated independently and according to different protocols, the challenge now being undertaken is achieving the integration, harmonization, and uncertainty assessment for each of the data types.

Third, while the ‘state vector’ of space data allows for accurate detection of changes, it does not necessarily provide data sets in readily interpretable units. For instance, land-atmosphere fluxes of carbon dioxide or phenologies of certain traits are well constrained by space data, but not directly derivable. Ground data offer these “direct” types of observations is the relevant units, and we therefore need to integrate space and ground observations to derive novel data streams that are not directly observable at global scales. BACI focuses on deriving what we call “Essential Ecosystem Variables” i.e. variables that are essential to monitor the fundamental interactions and feedbacks in the Earth System between the biosphere and atmosphere. The specific objective is to derive novel downstream data products by integrating Earth observations and in-situ data with machine learning methods. One part of this work package aims at resolving high frequency patterns of biosphere-atmosphere fluxes.

Fourth, the most exploratory part of the BACI project is dedicated to developing a generic index of change. This index should be nontrivial, i.e. able to detect a variety of different types of anomalies and extremes. This novelty index (BACIndex) ideally allows e.g. for detecting sudden events and abnormal changes in the multivariate EO data streams. In the context of the project, specific emphasis will be on detecting abrupt changes that are relevant to essential ecosystem variables, i.e. properties relevant to the functioning of terrestrial ecosystems, biosphere-atmosphere exchanges of matter and energy, and biodiversity related properties. So far, our main focus has been the comparison and development of methods and techniques able to automatically detect multivariate abnormal events. Numerous existing methods together with newly developed methods coming from the BACI consortium have been intensively tested and their results compared.

Fifth, BACI is well aware that novel data products need to be evaluated against independent data. Likewise, change detections exercises are ideally underpinned by additional lines of evidence. Hence, we selected regional study areas, as well as fast track sites for which regional evaluations are foreseen. Various regional validation sites cover a broad latitudinal transect of diverse areas in Europe and Africa, and a large range of types and drivers (human induced and climate). New, unexpected hot spots are constantly added during the lifetime of the project. For instance, as a response to the recent drought event in Ethiopia 2015/2016, the district of South Wollo, Ethiopia was added to our regional efforts. Research on the development of an accurate validation framework has started where we focus on the combined assessment of space, time, and thematic accuracy.

Sixth, BACI seeks interactions with the land use community and practitioners to prioritize research directions within our project in order to warrant optimal use of upcoming BACI (and non-BACI) products. In this context we are applying various user consultation strategies to identify knowledge gaps. Later we expect to bring integrate the produced data in a common framework with socio-ecological analyses.

Seventh, we are exploring novel avenues to enable the exploitation of space data for biodiversity studies. Amongst other studies, we are specifically working towards linking raw reflectance measures to plant biodiversity. We hope to achieve a new level of accuracy in biodiversity predictions compared to the current state-of-the-art (largely depending on climate data only and therefore lacking spatial accuracy).

In summary, the BACI project is a cascading workflow ranging from the assembly of various space and ground data, to deriving novel data products essential to monitor ecosystem functioning, to detecting system changes. Based on these products, we perform regional assessments and explore implications for socio-ecological as well as biodiversity research. The project sees itself as bridging element between the technical aspects required in the H2020 “Industrial Leadership” pillar to the “Societal Challenges”. We work in close collaboration with other FP7 and H2020 projects on data valuation as well as on co-interpretation and advancing the concepts of “Essential Variables” for the study of ecosystems in a changing environment.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

WP1: The project management team aimed to foster scientific collaboration and e.g. supported the partners in organizing the first set of workshops on e.g. use consultations. Additionally, a concept paper was submitted with the help of all partners on the conceptual basis of the BACI work and how this links to the concepts of “essential variables”. From a management perspective, the workflow was implemented i.e. all project management mechanisms are in place including an external advisory board that helps identifying critical issues early on. Overall, the project is structured like a working cascade; WP2 and WP3 are the “data packages” that carry the burden of kicking off the integration of work.

WP2/WP3: Both WPs have kicked-off their conceptual work and advanced substantially. In particular, the “surface vector” (WP2 )is defined based on existing and forthcoming Earth observations. Fist studies that link WP2 and WP3 activities have been published. Likewise, the ground database has been initiated and many input data streams are already in place to be ingested further.

WP4: This WP has the task to capitalize on both WP2 and WWP3 data to derived new data products. In fact, new sampling strategies for efficient “upscaling” activities were implemented such as an algorithm to derive high-frequency CO2 and energy fluxes at sub-daily resolution. These technical advances will allow soon to derive first large scale data products to be released.

WP5: The developments on the “biosphere atmosphere change index” itself is not yet implemented based on real data, but quite advanced from a methodological point of view. A wide range of existing methods for detecting multivariate abnormal events comparison were implemented in an efficient computing language and tested.
WP5 also developed a new algorithm for extreme events detection. All methods have been tested for their suitability to detect changes in real world data e.g. as provided by the Surface Vector.

WP6: The regional validation has not yet concretely started in terms of testing products and improving them. However, selecting focus areas and conceptually preparing product evaluation strategies made major advances. The role of this WP will be increasingly important in the next phase.

WP7: Organized a workshop on “Identification of promising applications for EO for land system science and sustainability science based on a user consultation workshop”. The results were reported to the consortium, led to controversial discussions, and certainly will help the consortium tailoring data products to the needs of a wider audience.

WP8: Has worked on two aspects: Firstly, WP8 prepared the inter-linkage of several community metrics with several indices describing the spatial variability of the primary productivity and at various scales (250m, 2500m and 25km). Pilot analyses using bird distribution data were performed and now need to be extended. Secondly, also WP8 has carried out a user consultation meeting with a broad community of researchers and practitioners. Here we have identified three main priority areas: improving projections of species distribution under future climate and land use scenarios, predicting changes in community composition, and monitoring protected areas (such as NATURA 2000).

WP9: Implemented several dissemination activities, such as the website, posters or presentations at conferences.

Overall, all tasks falling in the reporting period started and all deliverables were provided. We also achieved all milestones.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

We worked on the development of the concept of “Essential Ecosystem Variables” to bring the idea of “essential variables” closer to the community working on Earth system models we show how fundamental interactions among Essential Climate Variables and Essential Biodiversity Variables govern the states, processes, and functions of ecosystems. We advocate an emerging "Essential Ecosystem Variables" (EEVs) framework and its use for future monitoring activities. If this synthesis activity is successful, BACI could have a major societal impact.

In technical terms, we developed and Earth Observation assimilation approach for synthesizing multiple sources of optical and microwave data in to surface state vector (SV). The prototype of this method is very generic and beyond the state of the art. We expect that this idea would be very relevant to exploit a wide range of space data beyond BACI.

We also implemented machine learning models for high-frequency CO2 and energy flux upscaling by fusing EO and in-situ data. We anticipate that this data product will substantially push the research field and provide an unprecedented reference data set for next generation of global Earth System Models. This is of high relevance for societal challenges such as climate change assessments via coupled carbon cycle climate models.

Also the exploration of satellite data for advancing the fields of biodiversity research is highly relevant to e.g. GEO BON. We anticipate that we are able to empower biodiversity research beyond current state-of-the-art approaches in terms of predictive accuracy and spatial detail.

Related information

Record Number: 190140 / Last updated on: 2016-11-08
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top