Enter Zoom Webinar


Towards evolvable physics-based plants and landscape processes in terrestrial biosphere models

The terrestrial biosphere exerts disproportionate influence on Earth's climate, making improvements in its representation key to reducing climate uncertainty. After 50 years of development, land surface models contain detailed processes of energy fluxes, photosynthesis, hydrology, C-N-P cycles, and land-use within coarse non-interacting grid cells. Remaining discrepancies in fidelity to observed carbon and water cycles appear primarily related to deficiencies in the representation of forests and human activity. These include the omission of spatial processes of disturbance, migration, adaptation, and management. Also missing is the generative process of life, evolution, which gives rise to life history strategies, trophic-metabolic networks, leaf economics, local adaptation (i.e., optimality, acclimation), and plant behaviour. Despite improvements in representing vegetation demography by utilizing emergent properties of allometric scaling, canopy geometric realism remains low. This may bias carbon and water cycles per radiative transfer and coupled processes of photosynthesis, regeneration, evapotranspiration, heterotrophic respiration, and disturbance.

We believe that physics-based botanical models, forest landscape models, and terrestrial biosphere models may soon merge into new multi-scale models. While low-dimensional representations of forests are often used to improve computational efficiency and cope with a dearth of 4-D forest observatories, deep learning may be combined with new autonomous scanning systems - proximal and/or remote - such as our proposed global tower-based '5DNet' to infer evolvable 4-D physics-based models. This includes learning multi-generation tree models with 4-D traits from image and/or laser scanning time-series. To date, 4-D ontogeny has been inferred from individual scans of mature trees, multi-plant phenological events have been tracked in real-time, and the self-similar and -organizing nature of plants has been used to efficiently compress tree models down to their generating parameters. Achieving leaf-to-global scaling may require co-processor acceleration and fusing deep learning with 3-D radiative transfer modeling to infer global surface properties. An additional focus on evolution and human activity comes as 21st century land surface models mature into general simulations of life on Earth.

This Union Symposium presents exciting work toward achieving this moonshot in Earth observation and systems modeling.

Convener: Adam EricksonECSECS | Co-conveners: Rico Fischer, Sujay Kumar, Annikki Mäkelä, Nikolay Strigul

Activation of the text chat sets a cookie in your browser that is automatically deleted at the end of the conference.

A chat user is typing ...
SHIFT+ENTER for line break
We are sorry but we encountered a problem while running the chat US4 . Please reload this browser window. In case this message is shown again after reloading, please contact us at: egu21@copernicus.org. We are sorry for this inconvenience.

Mon, 19 Apr, 15:00–17:00

Chairperson: Adam Erickson


Eetu Puttonen et al.

Light detection and ranging (lidar) has become an essential tool in mapping and change detection in different environments over the last 20 years. Laser scanners capture point clouds to create accurate digital snapshots of their surroundings. These snapshots tell about the structural information in the scene and can be readily returned to again and again to detect and measure any changes with multi-temporal measurements. However, multitemporal measurements cannot typically resolve the change events nor can they resolve more high frequency dynamics that happen on daily or weekly basis in the scene. Also, lidar systems operate still mainly with single wavelength limiting their usability in classification tasks. First multi- and hyperspectral systems have been already demonstrated, but have yet to break through in wider usage. Finnish Geospatial Research Institute (FGI) has been prototyping with different 3D measurement systems for the last 10 years to improve multitemporal mapping (4D) solutions. The prototypes include both hyperspectral and long-term multi- and hypertemporal lidar systems, and their combinations in static and mobile configurations. FGI started early on to experiment with hyperspectral laser sources (2007) and successfully demonstrated the first hyperspectral laser scanner prototype in 2012. The system was later used in detecting intraday vegetation dynamics in 2015. Multitemporal multispectral ALS measurements have been conducted since 2015 in Evo and in Espoolahti. The first long-term multitemporal studies with FGI mapping platforms were started with ALS to monitor changes in forests (1998) and built environment (2001) and with mobile laser scanning in studying the erosion of an arctic river basin (2008) annually.  Multitemporal ALS studies with vegetation started in 1998 in Kalkkinen and in 2007 in Evo followed with bi-temporal studies with TLS. Test Site Evo has been acquired with ALS. In 2020, Evo test site was granted Academy of Finland Research Infrastructure (RI) status. The RI will collect a 30-year-long time series with annual measurements using various laser scanning sensors for investigating single tree growth processes, forest dynamics, understanding cyclic forest while having variation at diurnal and annual scales and forest monitoring technologies. Vegetation dynamics monitoring was extended in 2020, when FGI started set up a permanent TLS measurement station in a boreal forest. The TLS station accurately detects structural changes of hundreds of tree crowns around it. The experiment aims to detect the changes of phenological state the trees and further link them with the environmental parameter variation. 4D measurements have successfully demonstrated their potential in extending the information available from laser scanning systems. To improve the usage of these novel information, automated pre-filtering of the vast data amounts already at sensor level will be imperative. Different lidar platforms can operate throughout the spatial scale from millimeter precision all way to national coverage. Thus, development of new scalable lidar RIs open new possibilities to complement already existing infrastructures.

How to cite: Puttonen, E., Hyyppä, J., Litkey, P., Batista Campos, M., Hyyti, H., Wang, Y., Kukko, A., Kaartinen, H., Matikainen, L., Vastaranta, M., Junttila, S., Vaaja, M., and Alho, P.: State-of-the-art in 4D measurements, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16530, https://doi.org/10.5194/egusphere-egu21-16530, 2021.

Werner Rammer and Rupert Seidl

In times of rapid global change, the ability to faithfully predict the development of vegetation on larger scales is of key relevance to society. However, ecosystem models that incorporate enough process understanding for being applicable under future and non-analog conditions are often restricted to finer spatial scales due to data and computational constraints. Recent breakthroughs in machine learning, particularly in the field of deep learning, allow bridging this scale mismatch by providing new means for analyzing data, e.g., in remote sensing, but also new modelling approaches. We here present a novel approach for Scaling Vegetation Dynamics (SVD) which uses a deep neural network for predicting large-scale vegetation development. In a first step, the network learns its representation of vegetation dynamics as a function of current vegetation state and environmental drivers from process-based models and empirical data. The trained model is then used within of a dynamic simulation on large spatial scales. In this contribution we introduce the conceptual approach of SVD and show results for example applications in Europe and the US. More broadly we discuss aspects of applying deep learning in the context of ecological modeling.

How to cite: Rammer, W. and Seidl, R.: New deep-learning based approaches for forest modeling beyond landscape scale, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16528, https://doi.org/10.5194/egusphere-egu21-16528, 2021.

Renato Braghiere

Addressing the impact of 3D vegetation structure on shortwave radiation transfer in Earth System Models (ESMs) is important for accurate weather forecasting, carbon budget estimates, and climate predictions. While leaf-level photosynthesis is well characterized and understood, estimates of global level carbon assimilation in the literature range from 110 to 175 PgC.yr-1. I will explore how neglecting canopy structure leads to significant uncertainties in shortwave radiation partitioning, as well as second order derived canopy properties, such as leaf area index (LAI). I will also cover how modeled carbon assimilation of the terrestrial biosphere is impacted when a satellite derived clumping index is incorporated into the UKESM. Finally, I will touch on how the clumping index might be integrated into hyperspectral ESMs to explore the theoretical relationship between canopy structure and photosynthesis.

How to cite: Braghiere, R.: Better representing vegetation canopy structure in Earth System Models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16531, https://doi.org/10.5194/egusphere-egu21-16531, 2021.

Mike Harfoot et al.

Ecosystems are facing unprecedented pressures as a result of human activities. At the same time, ecology as a discipline is increasingly demanding more mechanistic understanding of what causes observed ecological patterns, in part for the development of the science but also to help mitigate impacts. Here, I will present the Madingley Model (www.madingleymodel.org), a General Ecosystem Model that aims to provide a mechanistic understanding of how ecosystems, on land and in the seas, are structured and how they function, and for how anthropogenic changes might alter that structure and function. I will discuss the model’s current capabilities, how it is being used, and highlight some necessary and exciting future directions for development.

How to cite: Harfoot, M., Tittensor, D., Hoeks, S., Krause, J., Arneth, A., Doughty, C., and Abraham, A.: General Ecosystem Models, moving towards modelling responses and effects of whole ecosystems, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16569, https://doi.org/10.5194/egusphere-egu21-16569, 2021.

Miłosz Makowski et al.

Over the last years, the role of forests in climate change has received increased attention. This is due to the observation that not only the atmosphere has a principal impact on vegetation growth but also that vegetation is contributing to local variations of weather resulting in diverse microclimates. The interconnection of plant ecosystems and weather is described and studied as ecoclimates. In this work we simulate ecoclimates by modeling (1) the local climate-response of individual plants in large-scale ecosystems, (2) the vegetation impact on the atmosphere, and (3) the soil hydrology. We employ interactive state-of-the-art methods for simulating ecosystem growth and weather dynamics to enable a realistic animation of vegetation growth, plant competition and cooperation mediated by light and soil water, as well as cloud transitions. Our plant ecosystem model simulates the growth of individual trees with branch-level geometry. We couple an ecosystem with a weather model to locally sample weather variations over time, which enables us to simulate the long-term climate-response of individual tree models. Simultaneously, the composition of an ecosystem affects the development of weather: individual trees vary in how they release vapor or transfer heat to the air. Our framework allows us to interactively explore the growth- and climate-response of individual trees and of an ecosystem as a whole.

How to cite: Makowski, M., Hädrich, T., Michels, D., Pirk, S., and Palubicki, W.: Interactive Simulation of 3-D Ecoclimates, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16600, https://doi.org/10.5194/egusphere-egu21-16600, 2021.

A chat user is typing ...