Skip to main content

Training on Remote Sensing for Ecosystem modElling

Periodic Reporting for period 2 - TRuStEE (Training on Remote Sensing for Ecosystem modElling)

Reporting period: 2018-10-01 to 2020-09-30

TRuStEE (Training on remote sensing for ecosystem modelling) is a pioneering innovative training network designed to provide multidisciplinary training for 12 Early Stage Researchers (ESRs) on the use of the still under-exploited potential of novel remote and proximal sensing techniques to face grand societal challenges as the increasing pressure of environmental change on ecosystem functioning and land-atmosphere interactions. The innovative aspects of the TRuStEE network are hereafter summarized:
i) to shape a new generation of scientists with combined, complementary and interdisciplinary skills in ecosystem modelling, plant physiology, remote sensing technologies and big data analysis;
ii) to train young scientists to exploit the recent technological development in the field of proximal and remote sensing (e.g. drones, fluorescence sensors) and Earth Observation data from new satellite missions (e.g. ESA-FLEX, NASA-GEDI) for the description of the ecosystem functioning and the prediction of ecosystem services;
iii) to improve the understanding and modelling of biosphere-atmosphere interactions and ecosystem services allowing the exploration of inter-linkages, synergies and feedbacks between ecosystem functional properties and plant traits measurable from remote observation.
To address these challenging aims, the TRuStEE network is composed by 9 beneficiaries and 8 partner organisations well known internationally for their interdisciplinary and innovative research and activities. The academic partnership is complemented by the non-academic sector involving companies operating in the project topic. The integration of the interdisciplinary expertise of different consortium members creates a number of synergies within the network that benefits the young scientist’s training.
The TRuStEE network of excellence designed a research programme with 12 multidisciplinary and intersectoral individual research projects that span 4 key areas and the relative work packages (WP).
In WP1, the link between the essential biodiversity variables categories (EVBs), the ecosystem functional properties (EFPs) and the vegetation optical properties was analysed using complete datasets of remote sensing information acquired and processed at different spatial, spectral and temporal resolutions. Complementary, a deep study on innovative instruments and data chain processing was performed. An innovative spectrometer for vegetation optical property acquisition was made operative and deployed on a flux tower. Robust techniques for intelligent data capture from UAVs were also developed to ensure good quality data acquisition. An operational data chain processing for UAV images was finally developed.
In WP2, the link between vegetation fluorescence, hyperspectral optical indices and vegetation functioning was investigated on different terrestrial ecosystems in order to verify the possibility to use optical indices related to photosynthesis in ecosystem models.
A first dataset of coupled active and passive F measurements was obtained using a newly automated platform. The data were collected over a Free-air CO2 enrichment experiment in Germany. Furthermore, the anisotropic properties of F and reflectance at different wavelengths were empirically characterised using high resolution radiance data collected over different canopy types with a mobile goniometer capable of collecting data at different view angles.
Time series of fluorescence and vegetation optical indices obtained with unattended high resolution field spectrometers, were also analysed in order to decouple the low and fast temporal dynamics and to link them to plant traits and ecosystem functional properties.
In WP3, the expertise and new understanding of vegetation systems emerging from WP1 and WP2 were transferred in real case studies in different. EFPs and plant traits (PTs) maps were derived from both ground and remote sensing data in target areas where plant variation was induced by different factors. For example, seasonal changes in canopy evapotranspiration were estimated in a tree-grass ecosystem using an energy balance model. Canopy structure was estimated with Structure from Motion techniques applied to data collected from UAV. Maps of early stress conditions were obtained through spectral indices and fluorescence.
In WP4, a review of the methods and techniques for the upscaling of important variables concerning vegetation and the environment was made. A suite of statistical data-driven and machine learning techniques was developed to up-scale PTs and EFPs from in-situ data at different spatial scales using remote sensing data . An open source package to derive at global scale EFPs using eddy covariance observations was developed using spectral information retrieved from Sentinel-2 with machine learning techniques. The uncertainties of the globally produced maps were also evaluated.
TRUSTEE scientific results were disseminated in scientific open access papers. Reports are available and exploitable on the TRUSTEE website with the link to repository of open source code created for the data processing. The datasets collected remotely with different sensors and platforms were organized in a WebGIS and are now open and accessible to public. In particular, a handbook of exercises linked to this dataset was prepared and published on the project website to enhance the data exploitability. These exercises were organized as a tutorial for undergraduate and master students interested in remote sensing application for ecosystem modelling.
TRuStEE research activity was focused on the methodological and technical constraints that hamper a systematic incorporation of remotely sensed data in ecosystem models, including scalability and multi-source data integration issues.
Main TRuStEE impacts were to enhance the career perspectives and employability of the students through a full training and a research activity on innovative technologies that are highly used nowadays. The project will have a short term impact on the fellow’s careers due to the new skills acquired by their academic/non-academic experience. The training schedule and content were thought to provide both an immediate benefit for the ESRs by strengthening their skills and abilities to successfully complete their PhD projects and a long-term benefit by improving their employability and their entrepreneurship capability. We think the ESRs were exposed to a good scientific environment that did strengthen their theoretical background. The summer school and courses improved their technical skills in operating UAVs and innovative instrumentation and data processing through open source software that can be customized for operational application in many environmental fields.
The new results published in scientific journal and the know-how acquired will have impacts in the remote sensing and earth observation sector. The dissemination activities planned by the consortium did not only target the scientific community but also the society in general. Few outreach activities were organized, targeting undergraduate students and society in general, within the lifetime of the project. We hope we were able to show to new generations, approaching the choice of future studies and jobs, that science is challenging but fun and that new technology such as UAVs and earth observations can help to change the future of the earth towards a more sustainable development.