Exploring the Impact of Process Scaling in Forest Dynamics Models: Tree-Level vs. Stand-Level Definitions

The structure and scale at which ecological processes are defined in forest models have direct implications for model behavior, interpretation, and computational efficiency. Most forest dynamics models adopt a fixed spatial perspective for key processes—such as growth, mortality, and regeneration—typically at the tree or stand level. Rarely do models include flexible, parallel formulations of these processes at multiple scales. This Master’s thesis investigates the consequences of defining forest processes at different spatial resolutions within the same model framework. The focus will be on comparing simulations where growth, mortality, and regeneration are modeled at the individual tree level versus more aggregated stand-level formulations. The forest dynamics model to be used is already developed and includes tree-level process definitions. The student will design and implement alternative stand-level formulations of these same processes within the existing model structure. After validation, the model will be used to run comparative simulations across a consistent landscape setup. The goal is to understand how different levels of process aggregation influence projections of forest dynamics (e.g., biomass accumulation, species composition, stand structure) and to assess the computational gains from coarser formulations, especially when scaling up to large forested areas. This thesis addresses both methodological and practical challenges in forest modeling, offering insights into the trade-offs between model resolution and computational efficiency, an increasingly important topic as forest models are applied to larger regions and more complex management questions. **Wanted** We are looking for a motivated Master’s student with a strong interest in forest modeling, process design, and spatial thinking. This project provides an opportunity to explore the structural foundations of simulation models and test how spatial assumptions influence forest projections. You will work with an established model framework and extend it to include parallel stand-level process definitions. This requires analytical thinking, basic coding skills, and the ability to interpret ecological dynamics from simulation results. **We are looking for someone who is:** • Curious about the mechanics and structure of forest models • Interested in spatial scaling and model simplification • Comfortable working independently with code (model is in R, C++, or Python) • Motivated to explore and compare simulation results across scenarios • Able to synthesize model output into ecological interpretations The project is computational in nature but includes strong links to ecological processes and forest management implications. **You will get to** • Work with an advanced forest simulation model and extend its structure • Gain insights into how process definitions affect model results • Compare tree- and stand-level modeling approaches in a controlled setup • Learn about trade-offs between model detail and scalability • Contribute to a methodological study relevant for large-scale forest applications • Potentially co-author a publication based on the thesis findings

The structure and scale at which ecological processes are defined in forest models have direct implications for model behavior, interpretation, and computational efficiency. Most forest dynamics models adopt a fixed spatial perspective for key processes—such as growth, mortality, and regeneration—typically at the tree or stand level. Rarely do models include flexible, parallel formulations of these processes at multiple scales.
This Master’s thesis investigates the consequences of defining forest processes at different spatial resolutions within the same model framework. The focus will be on comparing simulations where growth, mortality, and regeneration are modeled at the individual tree level versus more aggregated stand-level formulations.
The forest dynamics model to be used is already developed and includes tree-level process definitions. The student will design and implement alternative stand-level formulations of these same processes within the existing model structure. After validation, the model will be used to run comparative simulations across a consistent landscape setup.
The goal is to understand how different levels of process aggregation influence projections of forest dynamics (e.g., biomass accumulation, species composition, stand structure) and to assess the computational gains from coarser formulations, especially when scaling up to large forested areas.
This thesis addresses both methodological and practical challenges in forest modeling, offering insights into the trade-offs between model resolution and computational efficiency, an increasingly important topic as forest models are applied to larger regions and more complex management questions.

**Wanted**

We are looking for a motivated Master’s student with a strong interest in forest modeling, process design, and spatial thinking. This project provides an opportunity to explore the structural foundations of simulation models and test how spatial assumptions influence forest projections.
You will work with an established model framework and extend it to include parallel stand-level process definitions. This requires analytical thinking, basic coding skills, and the ability to interpret ecological dynamics from simulation results.

**We are looking for someone who is:**

• Curious about the mechanics and structure of forest models

• Interested in spatial scaling and model simplification

• Comfortable working independently with code (model is in R, C++, or Python)

• Motivated to explore and compare simulation results across scenarios

• Able to synthesize model output into ecological interpretations

The project is computational in nature but includes strong links to ecological processes and forest management implications.

**You will get to**

• Work with an advanced forest simulation model and extend its structure

• Gain insights into how process definitions affect model results

• Compare tree- and stand-level modeling approaches in a controlled setup

• Learn about trade-offs between model detail and scalability

• Contribute to a methodological study relevant for large-scale forest applications

• Potentially co-author a publication based on the thesis findings

• Develop and implement alternative stand-level formulations for key ecological processes (growth, mortality, regeneration) in an existing forest dynamics model. • Run comparative simulations at plot and landscape scales using both tree- and stand-level versions. • Analyze the impact of spatial process resolution on model outputs such as biomass, species composition, and stand structure. • Assess differences in runtime and computational resource demands.

• Develop and implement alternative stand-level formulations for key ecological processes (growth, mortality, regeneration) in an existing forest dynamics model.

• Run comparative simulations at plot and landscape scales using both tree- and stand-level versions.

• Analyze the impact of spatial process resolution on model outputs such as biomass, species composition, and stand structure.

• Assess differences in runtime and computational resource demands.

Supervisors: Dr. Olalla Díaz-Yáñez, Prof. Dr. Verena Griess If the idea of participating in cutting-edge research with practical relevance excites you, please contact olalla.diaz(at)usys.ethz.ch.

Supervisors: Dr. Olalla Díaz-Yáñez, Prof. Dr. Verena Griess
If the idea of participating in cutting-edge research with practical relevance excites you, please contact olalla.diaz(at)usys.ethz.ch.