European Geosciences Union General Assembly 2018

IE4.5/AS5.14/BG1.22/CL5.26/EMRP4.35/ESSI2.12/GD10.7/GI1.7 Information extraction from satellite observations using data-driven methods (co-organized)
(Co-)Convener: Miguel Mahecha
The unprecedented volumes of satellite Earth observation data gathered today allow for thorough investigation of the Earth's climate system and its interactions with the biosphere. Several international research initiatives and scientific projects are focused on the application of mathematical and statistical methods to extract insights about the functioning of the hydrosphere, biosphere and atmosphere, from this emerging Earth information data-cube. However, this data-cube is highly dimensional, thus innovative data mining and big data tools are required, and machine-learning methods – such as neural networks, tree ensembles, random forests or Gaussian processes, among others – can offer new means to extract valuable information in a rigorous manner.
As we progress through an exponential increase in satellite data availability, this session aims to bring researchers together to discuss the current state in big data and machine learning applications to Earth sciences and remote sensing. We aim to both (a) discuss current efforts, and (b) identify common challenges for the future.

(Co-)Convener: Markus Reichstein; Dorothea Frank
This session explores the linkages between climate extremes, biosphere and societal dynamics. I.e. emphasis is laid on 1) what impacts are caused by climate extremes on various aspects of the biosphere (incl. e.g. productivity, biogeochemical cycling, biodiversity) and society (e.g. political, economic aspects), and 2) which feedbacks exist amplifying or moderating the intensity, duration or extent of climate extremes. Empirical, theoretical and modelling studies from local to global scale are all highly welcome.

(Co-)Convener: Markus Reichstein,Sönke Zaehle
Plant traits extend the range of earth observations to the level of individual organisms, providing a link to ecosystem function and modeling in the context of rapid global changes. However, overcoming the differences in temporal and spatial scales between plant trait data and biogeochemical cycles remains a challenge. This session will address the role of plant species, biodiversity and adaptation / acclimation in the biogeochemical cycles of water, carbon, nitrogen and phosphorus. We welcome conceptual, observational, experimental and modeling approaches, and studies from the local to the global scale, including e.g. remote sensing observations.

(Co-)Convener: Markus Reichstein, Mirco Migliavacca
During the last decade, technological developments in field spectroscopy, isotope flux measurements and quantum cascade lasers have enabled alternative approaches for constraining ecosystem-scale photosynthesis and respiration.
In this session we aim at reviewing recent progress made with novel approaches of constraining ecosystem gross photosynthesis and respiration and at discussing their weaknesses and future steps required to reduce the uncertainty of present-day estimates. To this end we are seeking contributions that use emerging constrains to improve the ability to quantify respiration and photosynthesis processes at scales from leaf to ecosystem and global. Particularly welcome are studies reporting advancements and new developments in CO2 flux partitioning from eddy covariance data, the use of carbonyl sulfide, sun-induced fluorescence and stable isotopes approaches. Modelling studies which enhance our fundamental understanding of ecosystem-atmosphere CO2 exchange at global scale or make use of these emerging new constraints in data assimilation schemes are also welcome.

(Co-)Convener: Sönke Zaehle
Human activities are altering a range of environmental conditions, including atmospheric CO2 concentration, temperature, precipitation and nutrient availability. Quantifying and predicting the combined effect of these changes on biogeochemical fluxes is challenging because carbon, nutrient, and water cycles are intricately linked in terrestrial ecosystems. Terrestrial ecosystem models are used for this purpose. These are routinely tested and calibrated against data from ecosystem flux measurements, remote sensing, atmospheric inversions and ecosystem inventories. While these constrain the current mean state of the terrestrial biosphere, they provide limited information on the sensitivity of ecophysiological, biogeochemical, and hydrological processes to environmental changes.
This session focuses on how ecosystem processes respond to changes in CO2 concentration, temperature, water and nutrient availability. It aims at fostering the interaction between experimental and modeling communities by advancing the use of experimental data for model evaluation and calibration. Contributions include both experimental and observational studies, as well as modelling exercises spanning a range of scales and conditions: soil microbial activity, plant ecophysiology, nutrient cycling, and ecosystem level dynamics.

Venue: European Geosciences Union General Assembly 2018 | Vienna | 8 – 13 April 2018
Co-convener:Nuno Carvalhais
Monitoring and modeling of vegetation and ecosystem dynamics is fundamental in diagnosing and forecasting Earth system states and feedbacks. However, the underlying ecosystem processes are still relatively poorly described by Earth system models. Confronting terrestrial biogeochemical models at multiple temporal and spatial scales with an ever-increasing amount and diversity of Earth observation data is therefore needed.
In this session, we present the most recent advances in:
(1) the production of global land surface biophysical and biochemical variables from satellite and in-situ observations;
(2) assessment of plausibility, validation and intercomparisons of these products;
(3) their use in studying global ecosystem dynamics related to, e.g., climate variability and change;
(4) developments of terrestrial biogeochemical models to allow for the integration of new observational datasets
(5) benchmarking and improvement of these models through statistical and model-data integration techniques.