Niwot Ridge Long Term Ecological Research site, Boulder, Colorado
View of the Niwot Ridge Long Term Ecological Research (LTER) site near Boulder, Colo. This alpine tundra and subalpine forest site is a member of the LTER network, Critical Zone Observatories, and National Ecological Observatory Network, some of the many research networks that provide data used in Earth system model evaluation and help facilitate collaborations between empiricists and modelers to improve our understanding of Earth’s climate and ecological systems. Credit: William R. Wieder, National Center for Atmospheric Research, Boulder, Colo., and Institute of Arctic and Alpine Research, University of Colorado Boulder

The year 2016 marked a new stage in modern human history: For the first time, the average carbon dioxide (CO2) measurement at Mauna Loa Observatory in Hawaii surpassed 400 parts per million (ppm) for every month of the year. Atmospheric CO2 levels are expected to continue to rise, altering processes in terrestrial environments that have additional feedbacks with Earth’s climate. Accurately representing these land-atmosphere interactions in Earth system models is critical to producing reliable projections but remains a challenge. On 11 August 2017, speakers in the closing Ignite-style session at the Ecological Society of America’s annual meeting presented current and emerging research priorities for Earth system modeling—sharing how empirical and modeling approaches can improve model representations of terrestrial ecosystems. Session presentations and discussions led to two suggestions for achieving this goal.

Suggestion 1: More Data Types, More Details

Earth system models should better quantify and incorporate the impacts of human activity on biophysical (e.g., water and energy) and biogeochemical (e.g., carbon and nitrogen) cycles. Several of the session’s presentations emphasized the importance of improving model depictions of human activities, including fire, agriculture, and logging, and their effects on terrestrial landscapes.

One way to accomplish this goal is to expand the types of data used in model development. For example, model depictions of fire could be improved by incorporating not only quantitative data on human land use, demography, and economics but also qualitative data on human livelihoods that provide further details on human interactions with fire. Developing a database to catalog changes in ecosystem traits by disturbance type—and also disturbance severity and recovery time—could facilitate a new generation of model-data experiments and benchmarks.

Suggestion 2: Better Integration of Empirical and Modeling Approaches

The Earth system modeling community should create and share tools to train Earth system scientists to work at disciplinary interfaces. During the Ignite-style session, attendees expressed interest in understanding how they can apply their field-collected ecological data to Earth system model development.

Training students to use both modeling and empirical approaches will prepare the next generation of scientists to better address challenges in global change science.

Speakers detailed their experiences transitioning from empirical ecology to Earth system modeling and emphasized the importance of collecting empirical data in a format useful for evaluating model structure and parameterizations. A central database that aggregates available training opportunities and workshop materials would improve communication between empirical and modeling communities.

The panel also discussed ways to introduce students early in their education to principles and applications of modeling through short exercises in higher education courses. To increase the effectiveness of these types of active learning exercises, we suggest that their impacts on student learning undergo evaluations that are published in peer-reviewed education journals (e.g., CourseSource) so that the evaluations and classroom exercises can be widely accessible to the global change science community. Training students to use both modeling and empirical approaches will prepare the next generation of scientists to better address challenges in global change science.

With these two approaches, the discipline of global change science can maintain its currency and better address existing and emerging research needs in an ever-changing, 400-plus-ppm CO2 world.

The coauthors thank the Ignite-style session speakers for their presentations, which provided inspiration and content for this report. U.S. Department of Agriculture National Institute of Food and Agriculture project 2015-67003-23485 supported the 11 August 2017 Ignite-style session.

—Susan J. Cheng (email:, Cornell University, Ithaca, N.Y.; Nicholas G. Smith, Texas Tech University, Lubbock; also at Purdue University, West Lafayette, Ind.; and Alison R. Marklein, Lawrence Berkeley National Laboratory, Berkeley, Calif.; also at University of California, Berkeley


Cheng, S. J.,Smith, N. G., and Marklein, A. R. (2018), Modeling global change ecology in a high–carbon dioxide world, Eos, 99, Published on 16 March 2018.

Text © 2018. The authors. CC BY 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.