Max Planck Gesellschaft

Research group: Empirical Inference of the Earth System

Mission | Team | Projects | Publications


Revealing insights from observations by means of data-driven research is a key element in Earth system sciences. Long-term observations of multiple Earth system properties encode our knowledge on how land-surface processes respond to climatic variability and interact with biodiversity. We develop methods to extract the valuable information in these data in order to confront them with models, and gain new insights. We aim at a more profound understanding of changing land ecosystems and their responses to and interactions with climate anomalies.

We are an interdisciplinary team of scientists contributing expertise in "Biology", "Ecology", "Geography" (or should we say "Earth System Science"?), "Computer Science" and "Physics".

Currently our group is following four core topics (a detailed description will follow soon):

  • The emerging Earth System Data Cube
  • Extremes and anomalies in the terrestrial biosphere
  • Biogeography and ecosystem functioning
  • Data analytic methods
  • How to integrate observational evidence in models

Typically, we work at the global scale. But currently we also work on regional land-atmosphere interactions in Colombia.

From left to right: Xuanlong, Daniel, Paul, Milan, Fabian, Julia, Torben, Guido, Christopher, Talie & Rocio

News and events

January 2018: Our latest synthesis paper reveals the impacts of climate extremes across global FLUXNET sites

Jannis von Buttlar just got the key paper of his PhD thesis accepted! Congratulations. In the paper we investigate extreme climatic events. We focus on the impacts of droughts and heat stress on ecosystem-atmosphere CO2 fluxes as measured by eddy-covariance tower measurements across the globe. Our results indicate that short-term heat extremes increased respiration more strongly than they down-regulated GPP, resulting in a moderate reduction of the ecosystem’s carbon sink potential. In the absence of heat stress, droughts tended to have smaller and similarly dampening effects on both GPP and Reco, and, hence, often resulted in neutral NEP responses. The combination of drought and heat typically led to a strong decrease in GPP, whereas heat and drought impacts on respiration partially offset each other. Taken together, compound heat and drought events led to the strongest C sink reduction compared to any single-factor extreme. A key insight of this paper, however, is that duration matters most: for heat stress during droughts, the magnitude of impacts systematically increased with duration, whereas under heat stress without drought, the response of Reco over time turned from an initial increase to a down-regulation after about two weeks. For more information, please refer to the paper (the pre-version is online; the final version will appear online soon).

von Buttlar, J., Zscheischler, J., Rammig, A., Sippel, S., Reichstein, M., Knohl, A., Jung, M., Menzer, O., Arain, M.A., Buchmann, N., Cescatti, A., Gianelle, A., Kieley, G., Law, B.E., Magliulo, V., Margolis, H., McCaughey, H., Merbold, L., Migliavacca, M., Montagnani, L., Oechel, W., Pavelka, M., Peichl, M., Rambal, S., Raschi, A., Scott, R.L., van Gorsel, E., Varlagin, A., Wohlfahrt, G. and Mahecha, M.D. Impacts of droughts and extreme temperature events on gross primary production and ecosystem respiration: a systematic assessment across ecosystems and climate zones. Biogeosciences - Discussions,

Dezember 2017: We have contributed to the BAMS special report explaining Extreme Events in 2016

The American Meteorological Society has just published the annual special report Explaining Extreme Events of 2016 on how man-made climate change could have affected individual extreme events in the past year. This time our group and other researchers at the MPI contributed a chapter to this report. We examined the impacts of the exceptionally warm winter 2015/16 in the Iberian Peninsula, followed by a very wet spring. Using an extensive ensemble of climate-ecosystem model simulations in tandem with eddy covariance data at the Las Majadas del Tietar measurement site, we showed that this anomalous pattern induced an exceptionally high ecosystem gross primary production. We found that these increased uptake rates are also linked to rising CO2 concentrations in the atmosphere, which act as a fertilizer for vegetation. Given the increased probability for warm winter conditions, this pattern is very relevant to understand future impacts of climate extremes. However, one has to consider that there is no evidence for increases in spring rainfall, which was essential to explain the exceptional ecosystem response. This study is among the first papers that directly seek to link anthropogenic changes in climate extremes to ecosystem impacts. The AMS special issue was discussed in many media, e.g. here, here, and here.

The chapter is Sippel, S. , El-Madany, T. , Migliavacca, M. , Mahecha, M.D. , Carrara, A., Flach, M., Kaminski, T., Otto, F.E.L., Thonicke, K., Vossbeck, M. and Reichstein, M. Warm winter, wet spring, and an extreme response in ecosystem functioning on the Iberian Peninsula. Ch 16. In Explaining Extreme Events of 2016 from a Climate Perspective. Special Supplement to the Bulletin of the American Meteorological Society, 98, 80-85, link.

November 2017: We host the BACI progress meeting in Jena!

BACI is the H2020 project supporting efforts towards generating "Essential Ecosystem Variables" (EEVs: exploring the intersection of Essential Climate Variables and Essential Biodiversity Variables) by developing novel downstream data products that are not directly observable from space. We also aim at detecting hotspots of change of the terrestrial biosphere and attribution to fundamental environmental and societal transformations - towards a biosphere atmosphere change index. This week will be a week of mutual updates on recent progress, brainstorming on new ideas, and planning for the last phase of this European effort. Interested? The latest agenda is here.

September 2017: Miguel invited to join the iLEAPS Scientific Steering Committee

iLEAPS – the Integrated Land Ecosystem Atmosphere Processess Study - is one of the global research projects in FutureEarth. In a nutshell, this global network of scientists works towards a deeper understanding of the land-atmosphere interface. iLEAPS is a communication hub to jointly understand the interactions of biological, chemical and physical processes that transform and transport energy and matter through the land-atmosphere interface. The iLEAPS Scientific Steering Committee (SSC) members serve for a period of three years and guide the overall direction and development of the project. For more information click here.

August 2017: Guido and Sebastian teach talented next generation Earth system scientist

The "Deutsche Schülerakademie" offers a two weeks course for bright and academically motivated kids. Guido and Sebastian planned a course for the next generation of Earth system scientist. The very young scientsits learned how to use R, and how to do first analysis with data and models to investigate our Earth system. Please check out this link for more information.

August 2017: Novel concepts, fields and methods: the role of mapping

Miguel presents ideas on new methods at the Resilience2017 sesstion on Novel concepts, fields and methods: the role of mapping. A brief blog on a related session can be found here  here. We got a lot of positive feedback from many coworkers all over the world and are decided to move the Earth System Data Cube and in particular the analytic toolkit forward. We believe that we can make a difference in resilience research as well!

August 2017: Milan published a new paper on multivariate anomaly detection in Earth observations

In our latest paper by Milan Flach et al. we provide an overview and comparison of multivariate anomaly detection algorithms. The problem we address is that Earth observations are not only high dimensional, but also (cross and auto)correlated, and varying at multiple time scales. This paper identifies three multivariate anomaly detection algorithms that clearly outperform others, especially when combined with suitable feature extraction methods. Based on these methods we can now screen global EOs for compund events that are not trivially detectable by univariate anomaly detection methods.

Flach, M., Fabain Gans, Brenning, A., Denzler, J., Reichstein, Rodner, E., Bathiany, S., Bodesheim, P., Guanche, Y., Sippel, S., & Mahecha, M.D. Multivariate Anomaly Detection for Earth Observations: A Comparison of Algorithms and Feature Extraction Techniques. Earth System Dynamics, 8, 677-696, doi:10.5194/esd-8-677-2017.

August 2017: Sebastian receives the Bernd Render Prize for young geoscientists by the DFG! Congrats!

Sebastian Sippel is awarded the Bernd Rendel Prize 2017 for junior geoscientists. He started his doctoral thesis 2014 in our group and first investigated the question how climate variability has been changing over the past decades. Sebastian could show in two papers that changes in climate variability are oftern overestimated and offered an analytical correction. The overarching question of the thesis, however, was how climate extremes influence the carbon exchange between the biosphere and the atmosphere. For these studies he used both observations as well as different modeling apporoaches. He convinced the jury by not shying away from questioning established approaches. Sebastian Sippel received the good news the day of his defense. The award ceremony will take place on September 25, 2017 during the general meeting of the German Geological Society in Bremen.

Summer 2017: Our group has new members

We are very happy that our group has new members and guests!

May 2017: Sebastian defended his PhD thesis @ ETH, congratulation!

We are very happy that Sebastian finished his PhD thesis on Climate extremes and their impact on ecosystem–atmosphere interactions in a very impressive defense at the ETH Zürich. Congratulations from the entiere team - we are proud of you. From left to right: Martin Heimann, Niki Gruber, Sebastian Sippel, Sonia Seneviratne, Miguel Mahecha.

April 2017: Workshop on Using the Earth System Data Cube: Status and Future

In these days we are about to finalize the CAB-LAB project (more known as the “Earth System Data Cube”) and will have a small Final Workshop that will be held at ESRIN, Frascati (Italy) on 12th April, 2017.

The aims of this Workshop are

  • To present what has been achieved so far regarding building the Earth System Data Cube (from an infrastructural point of view),
  • To present the scientific avenues we have been exploring,
  • To discuss how the community evaluates the potential of the project for the future,
  • To discuss new use-cases that can be addressed in future work.

More details about the meeting will be added here soon

March 2017 Talie presents her final study from the PhD thesis in the German TV

Talie Musavi presented her latest study published in Nature Ecology and Evolution in the 3sat-nano: The German title is Alte Wälder atmen besser wee done! Want to see it? Follow this link. Another contribution in the MDR is unfortunatly not available online any more.

Jan 2017: No evidence for increasing precipitation extremes and annual totals in the world's dry regions!

Recently we tried to publish a comment on a Nature Climate Change paper that received a lot of attention, but uses very problematic statistical approaches. However, we were not allowed to publish our critique in the very journal despite of reviewers acknowledging that there were issues in the paper. But, as we consider the topic a very important one it needs to be based on robust analytics. Consequently, we submitted our comment to another journal: Hydrology and Earth System Sciences. Here, again, the reviews were very positive (all can be seen in the discussion forum) and now the final paper is accepted and was published these days:

Sippel, S,, Jakob Zscheischler, Heimann, H., Lange, H., Mahecha, M.D., van Oldenborgh, G.J., Otto, F.E.L. & Reichstein, M. Have precipitation extremes and annual totals been increasing in the world's dry regions over the last 60 years? Hydrolgy and Earth System Sciences, 21, 441-458, doi:10.5194/hess-2016-452.

Fig (right): Trends and 30-year means of spatially aggregated heavy precipitation (Rx1d) in dry and transitional regions combined (Greve et al., 2014). Grey lines indicate our estimation using an aridity-based dryness definition and correcting for statistical artefacts, black lines are estimated equally, but after removing cases of incomplete temporal coverage. All p-values are given for two-sided (one-sided) Mann-Kendall trend tests.

Jan 2017: Older and more diverse forests have a stable photosynthetic capacity

In tandem vegetation and soils serve as global "bioreactors" that constantly take‐up and release CO2. The intensity of this "natural pump" partly shapes the atmospheric concentrations of CO2. However, this is a fragile process i.e. climate extremes and variability can affect it substantially. Consequently, many scientists are seeking to understand which factors are stabilizing the key ecosystem processes such as its capacity to take up CO2 over the years. Writing in the newly launched journal Nature Ecology and Evolution, Talie from our team and coworkers across the world investigate the capacity of forests to dampen the effect of climate variability on the natural CO2 pump. Two factors seem to clearly control this "buffer": Forest age and its species diversity. This finding can be read as a scientific call to preserving old forests and their diversity order to stabilize their functionality! Congrats to this nice success... For more information please go to: Go to the publishers page!

Jan 2017: EGU session submission - Information extraction from satellite Earth observations using data-driven methods

The unprecedented volumes of satellite Earth observation data gathered today allow for thorough investigation of 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. 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 encourage authors to submit presentations on: machine learning applied to geosciences and remote sensing, data-driven methods to analyse spatiotemporal dynamics and causal relationships in Earth observations, enlightening opinions about interface between mathematics and climate science. Submit an abstract!

November 2016: Project workshop in Colombia

A 50 year-old armed conflict in Colombia has influenced natural landscapes and their conservation status both positively and negatively. It remains unclear whether the ongoing conflict resolution would promote drastic changes in land-use promoting either deforestation or conservation. Recent and projected changes in climate constitute a further threat for the conservation of the fragile ecosystems in the country through potential increases in the recurrence rates and intensities of droughts, floods, fires, among other phenomena. This situation makes it crucial to develop an ecological observatory system (EOS) for assessing changes in biodiversity and ecosystem functioning in Colombia. A first meeting developed a discussion forum with diverse experts to explore the possibility of implementing an ecological observatory system for Colombia. Various activities are planned now and will be realized in 2017 and beyond with key scientific units in the country. We are heavily involved since the early planning stage and committed to further develop this activity. For more info: Funded by the BMBF and ESA-iLEAPS.

September 2016: MSCJ Autumn School co-organized by the group

Mahecha, M.D. organized the scientific agenda and logistics of the Michal Stifle Center for Data Driven and Simulation Science entitled Dynamics of natural (eco)systems: theory and applications. This was a mind-blowing activity with amazing speakers! Many thanks to all participants. Course materials are avaiable on demand only.

The latest agenda is linked here.

February 2016: Workshops at the on Extreme Events and Environments from Climate to Society (E3S)

Active and past group members are actively involved in the project Extreme Events and Environments from climate to Society (E3S). In particular, we are organizing two workshops:

January 2016: Jakob receives the Köppen award for his excellent PhD that he realized in our group! Congrats - we are proud!

We are proud of Jakob Zscheischler (now at the ETH Zürich), who receives the Wladimir-Peter-Köppen-Preize awarded by the Cluster of Excellence "Integrated Climate System Analysis and Prediction" (CliSAP) at the University of Hamburg. The thesis was realized in our group and submitted to the Swiss Federal Institute of Technology Zurich (ETH) in 2014. Please find more information here and here or simply see his key papers listed below. One of the key findings of the thesis was that few extremes in gross primary production can largely explain the inter annual variability of this very process. See for instance this paper Zscheischler, J., Mahecha, M.D., von Buttlar, J., Harmeling, S., Jung, M., Rammig, A., Randerson, J., Schölkopf, B., Seneviratne, S.I., Tomelleri, E., Zaehle, S. and Reichstein, M., (2014) Few extreme events dominate global interannual variability in gross primary production. Environmental Research Letters 9 035001.

Fig. to the right: Award ceremony at the MPI for Meteorology, Hamburg.


Phone: +49.3641.57 - extension | E-mail: e-mail - at -

Name Position E-mail Phone Room
Miguel D. Mahecha Group leader mmahecha ...6265 C3.011
Paul Bodesheim PostDoc pbodes ...6231 C2.010
Milan Flach PhD Student mflach ...6288 C3.013
Fabian Gans Geoscientific programmer fgans ...6289 C3.009b
Julia Joswig PhD Student jjoswig ...6211 C1.014
Guido Kraemer PhD Student gkraemer ... C1.020
Christopher Krich PhD Student ckrich ...6240 C1.004
Xuanlong Ma PostDoc ** xma ...6278 C3.002b
Talie Musavi PhD student * tmusavi ...6296 C3.013
Daniel Pabón-Moreno PhD Student ** dpabon ... C1.020
Christian Requena Mesa PhD Student *** crequ ...6242 C1.022
Torben Schick Student helper tschick ...6296 C3.013
Rocio Perez BACI project manager rperez ...6220 C2.012b
* co-oriented by BGI & Jens Kattge
** co-oriented by Mirco Migliavacca
*** co-oriented by us in tandem with Markus Reichstein, Joachim Denzler and the DLR Inst. for Data Science

Former Team Members and Guests

Sebastian Sippel PostDoc Now at the Norwegian Institute of Bioeconomy Research
Franziska Schrodt PostDoc Now at the University of Brighton
Iulia Ilie PhD Student iilie
Jakob Zscheischler PhD Student, later PostDoc Now at the ETH Zürich
Reik Donner PostDoc Now at PIK
Xiuchen Wu PostDoc Now at the Beijing Normal University
Chao Ting Chang Visitor (2012 & 2013) CREAF, Universitat Autonoma de Barcelona
Lina Fürst Master Student Now at the University of Bayreuth
Jannis von Buttlar PhD Student, guest

Our current key projects

BACI: "Detecting changes in essential ecosystem and biodiversity properties - towards a Biosphere Atmosphere Change Index" project is funded by the European Union's Horizon 2020 research and innovation programme (2015-2019); Group members involved: Milan Flach (PhD student) & Paul Bodesheim (PostDoc), Miguel Mahecha (Scientific coordination of the consortium).

The Earth System Data Cube project (formally CAB-LAB) funded by the European Space Agency (ESA) (2015-2017); Group members involved: Fabian Gans (scientific programming) & Paul Bodesheim (PostDoc), Miguel Mahecha (Scientific coordination of the consortium).

The TRUSTEE project is Mari Sklodowska-Curie Innovative Training Network (2017-2020); Group members involved: Daniel Pabon? (PhS student) & Mirco Migliavacca (supervison, teaching), Miguel Mahecha (supervison, teaching).

Key publications by research area

Note, these are only the key papers, please follow this link for a complete list of publications produced since 2013 by this research group. Also note that here we only show the papers where a group member is first or senior author.

Extremes & Anomalies in the terrestrial biosphere

Biogeography and ecosystem functioning

  • Musavi, T.,  Migliavacca, M., Reichstein, M. Kattge, J., Wirth, C., Black, T.A., Janssens, I., Knohl, A., Loustau, D., Roupsard, O., Varlagin, A., Rambal, S., Cescatti, A., Gianelle, D., Kondo, H., Tamrakar, R. & Mahecha, M.D. (2017) Stand age and species richness dampen interannual variation of ecosystem-level photosynthetic capacity. Nature Ecology & Evolution, 1, 0048, doi:10.1038/s41559-016-0048.
  • Musavi, T., Migliavacca, M., van de Weg, M.J., Kattge, J., Wohlfahrt, G., van Bodegom, P.M., Reichstein, M., Bahn, M., Carrara, A. F. Domingues, T.F., Gavazzi, M., Pianelle, D., Gimeno, C., Granier, A., Gruening, C., Havránková, K., Herbst, M., Hrynkiw, C., Kalahari, A., Kaminski, T. Klump, K., Kolari, P., Longdog, B., Miners, S., Montanan, L., Moors, E., Ocher, W.C., Reich, P.B., Rohatyn, S., Rossi, A., Rosenberg, E., Varlagin, A. Wilkinson, M., Wirth, C. & Mahecha, M.D. (2016) Potential and limitations of inferring ecosystem photosynthetic capacity from leaf functional traits. Ecology and Evolution, 6, 7352–7366, doi:10.1002/ece3.2479.
  • Musavi, T., Mahecha, M.D., Migliavacca, M., Reichstein, M., van de Weg, M.J., van Bodegom, P.M., Bahn, M., Wirth, C., Reich, P.B., Schrodt, F., & Kattge, J. (2015) The imprint of plants on ecosystem functioning: A data-driven approach. Int. J. Appl. Earth Observ. Geoinf. 43, 119-131, doi: 10.1016/j.jag.2015.05.009.
  • Reichstein, M., Bahn. M., Mahecha, M.D., Kattge, J. & Baldocchi, D. D. (2014) Linking plant and ecosystem functional biogeography. PNAS, 111, 13697-13702, doi:10.1073/pnas.1216065111.
  • Mahecha, M.D. and Schmidtlein, S. (2008) Revealing biogeographical patterns by nonlinear ordinations and derived anisotropic spatial filters. Global Ecology and Biogeography, 17, 284–296, doi:10.1111/j.1466-8238.2007.00368.x.

On observations and models

Data analytic methods

An analysis of 68 of our abstracts using a manually standardized list of keywords. We visualized these data useing t-SNE considering co-ocurrence statistics. Key-word size and colour intensity corresponds to abundance; red-orange colours are key-words that describe our domain knowledge, blue-green colours our methods.


We try to publish our code in a transparent way. Most tools can be accessed from the BGI-JENA organization on GitHub. Currently you find the following contributions from our group:

  • CABLAB.jl - Julia package for Data Cube analysis Julia
  • MultivariateAnomalies.jl - Julia package for Multivariate anomaly detection Julia
  • RSCAPE - R package for base respiration and Q10 estimation R
  • DRR - R package for Dimensionality Reduction via Regression R
  • coRanking - R package for co-Ranking matrix R
  • dimRed - R package Framework for dimensionality Reduction R


Missing: Iulia, Rocio, Guido

Directions | Disclaimer | Data Protection | Contact | Internal | Webmail | Local weather | PRINT | © 2011-2018 Max Planck Institute for Biogeochemistry