BAYESIAN OPTIMIZATION OF A DYNAMIC VEGETATION MODEL
(2025) In Master’s Theses in Mathematical Sciences MASM02 20251Mathematical Statistics
- Abstract
- In this thesis, we apply Bayesian Optimization to estimate the parameters of the LPJ-GUESS dynamic vegetation model with a focus on Methane (CH4) emissions in the Siikaneva wetland in Finland. Previous research has used other statistical methods, such as Markov Chain Monte Carlo (MCMC), to estimate the LPJ-GUESS model parameters, achieving a root mean square error (RMSE) of 0.023. Our objective is to assess whether Bayesian Optimization using a Gaussian Process can yield more accurate parameter estimates with lower error, using fewer computer resources.
To evaluate this, we first conduct a twin experiment using synthetic data, testing different acquisition, covariance, and loss functions in the Gaussian Process. We found that the Log... (More) - In this thesis, we apply Bayesian Optimization to estimate the parameters of the LPJ-GUESS dynamic vegetation model with a focus on Methane (CH4) emissions in the Siikaneva wetland in Finland. Previous research has used other statistical methods, such as Markov Chain Monte Carlo (MCMC), to estimate the LPJ-GUESS model parameters, achieving a root mean square error (RMSE) of 0.023. Our objective is to assess whether Bayesian Optimization using a Gaussian Process can yield more accurate parameter estimates with lower error, using fewer computer resources.
To evaluate this, we first conduct a twin experiment using synthetic data, testing different acquisition, covariance, and loss functions in the Gaussian Process. We found that the Log Expected Improvement (Log-EI) acquisition function offered a good balance between accuracy and computational cost. The Matérn covariance with ν = 2.5 was found to be suitable for the optimization, and RMSE proved to be more effective than mean absolute error (MAE) as a loss function. We initially assumed a zero-mean Gaussian Process, but then incorporated a mean structure.
When applying the optimization to real observations of CH4 emissions, Bayesian Optimization improved parameter estimates and reduced the RMSE to 0.013 — a better result than previously obtained when using MCMC. These results suggest that Bayesian Optimization is an effective approach for parameter estimation in the LPJ-GUESS model. (Less) - Popular Abstract
- Methane is a powerful greenhouse gas that significantly contributes to climate change, and one of its main sources is wetlands. In this project, we focus on the methane emissions from a Finnish wetland called Siikaneva. We are using a computer model called LPJ-GUESS, which helps us understand how the ecosystem behaves, including how much methane is re leased. However, this model relies on unknown parameters—input values that influence how the model works—for which we do not know the exact values. Previous research used advanced statistical techniques to estimate the parameters, which resulted in a prediction error of 0.023. This project aims to reduce the prediction error.
In our project, we applied a statistical method called Bayesian... (More) - Methane is a powerful greenhouse gas that significantly contributes to climate change, and one of its main sources is wetlands. In this project, we focus on the methane emissions from a Finnish wetland called Siikaneva. We are using a computer model called LPJ-GUESS, which helps us understand how the ecosystem behaves, including how much methane is re leased. However, this model relies on unknown parameters—input values that influence how the model works—for which we do not know the exact values. Previous research used advanced statistical techniques to estimate the parameters, which resulted in a prediction error of 0.023. This project aims to reduce the prediction error.
In our project, we applied a statistical method called Bayesian Optimization to see if we could get more accurate parameter estimates and reduce the prediction error. Bayesian Optimization focuses on finding the minimum value of a function by smartly choosing which points to evaluate. We began with a twin experiment. This means that we created our own synthetic data where we know the exact values of the parameters. This allowed us to test how to best configure our optimization. We examined different options to decide which point to evaluate next and measure the difference between the observations and the predictions.
After optimizing our setup, we applied it to the real methane emission observations from the Siikaneva wetland. The results showed a prediction error of 0.013, which is lower than what was obtained using other statistical methods. We conclude that Bayesian Optimization is an effective method for estimating parameters in the LPJ-GUESS model, especially when studying environmental data like methane emissions from wetlands. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9203793
- author
- De Leon, Eric LU
- supervisor
- organization
- course
- MASM02 20251
- year
- 2025
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Bayesian Optimization, LPJ-GUESS, Acquisition Functions, Methane Emissions, Zero-Mean, Mean Structure, Gaussian Process
- publication/series
- Master’s Theses in Mathematical Sciences
- report number
- LUNFMS-3131-2025
- ISSN
- 1404-6342
- other publication id
- 2025:E61
- language
- English
- id
- 9203793
- date added to LUP
- 2025-06-23 16:19:27
- date last changed
- 2025-06-23 16:19:27
@misc{9203793,
abstract = {{In this thesis, we apply Bayesian Optimization to estimate the parameters of the LPJ-GUESS dynamic vegetation model with a focus on Methane (CH4) emissions in the Siikaneva wetland in Finland. Previous research has used other statistical methods, such as Markov Chain Monte Carlo (MCMC), to estimate the LPJ-GUESS model parameters, achieving a root mean square error (RMSE) of 0.023. Our objective is to assess whether Bayesian Optimization using a Gaussian Process can yield more accurate parameter estimates with lower error, using fewer computer resources.
To evaluate this, we first conduct a twin experiment using synthetic data, testing different acquisition, covariance, and loss functions in the Gaussian Process. We found that the Log Expected Improvement (Log-EI) acquisition function offered a good balance between accuracy and computational cost. The Matérn covariance with ν = 2.5 was found to be suitable for the optimization, and RMSE proved to be more effective than mean absolute error (MAE) as a loss function. We initially assumed a zero-mean Gaussian Process, but then incorporated a mean structure.
When applying the optimization to real observations of CH4 emissions, Bayesian Optimization improved parameter estimates and reduced the RMSE to 0.013 — a better result than previously obtained when using MCMC. These results suggest that Bayesian Optimization is an effective approach for parameter estimation in the LPJ-GUESS model.}},
author = {{De Leon, Eric}},
issn = {{1404-6342}},
language = {{eng}},
note = {{Student Paper}},
series = {{Master’s Theses in Mathematical Sciences}},
title = {{BAYESIAN OPTIMIZATION OF A DYNAMIC VEGETATION MODEL}},
year = {{2025}},
}