[Dataset] Optimizing the Temperature Sensitivity of the Isoprene Emission Model MEGAN in Different Ecosystems Using a Metropolis-Hastings Markov Chain Monte Carlo Method
Software and Data Repository for MEGAN Optimization Experiments Using MHMCMC This repository is associated with the following manuscript, which has been submitted to JGR Biogeosciences: Christian Alexander DiMaria, Dylan B. A. Jones, Valerio Ferracci, et al. Optimizing the Temperature Sensitivity of...
| Autores: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Tipo de recurso: | conjunto de datos |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositorio: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/390262 |
| Acceso en línea: | http://hdl.handle.net/10261/390262 https://digital.csic.es/handle/10261/390244 |
| Access Level: | acceso abierto |
| Palabra clave: | Optimization Ecosystem Isoprene Model Monte Carlo Observations http://metadata.un.org/sdg/3 http://metadata.un.org/sdg/9 Ensure healthy lives and promote well-being for all at all ages Build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation |
| Sumario: | Software and Data Repository for MEGAN Optimization Experiments Using MHMCMC This repository is associated with the following manuscript, which has been submitted to JGR Biogeosciences: Christian Alexander DiMaria, Dylan B. A. Jones, Valerio Ferracci, et al. Optimizing the Temperature Sensitivity of the Isoprene Emission Model MEGAN in Different Ecosystems Using a Metropolis-Hastings Markov Chain Monte Carlo Method. ESS Open Archive . February 05, 2025. DOI: 10.22541/essoar.173877770.06627485/v1. https://essopenarchive.org/doi/full/10.22541/essoar.173877770.06627485 The repository contains all the data and code required to repeat our MEGAN temperature response optimization experiments. In particular, starting from the raw isoprene concentration or flux time series data, users of this repo can: Filter the observations using ancillary meteorological data (also included in repo) Normalize the filtered observations to extract the temperature response function Perform a Levenberg-Marquardt optimization for any parameter combinations at all field sites Perform MHMCMC optimization for any parameter combination at all field sites Recreate the figures from the associated manuscript. The code is a mixture of Python scripts, MATLAB scripts, and Jupyter Notebooks. All code has been pre-run, and the outputs (including figures) are included in the repository. Citation Policy and Stipulations For Use Please cite this repository and / or the accompanying manuscript if you use or modify the Python scripts, Matlab scripts, or Jupyter Notebooks, for your own research. You may also use anything contained in the /repo/output/ folder, the /repo/data/geos-chem/ folder, and the /repo/data/normalized_measurements/ and /repo/data/filtered_measurements/ folders, with appropriate reference to this repository and / or accompanying manuscript. IMPORTANT: Raw measurements are included for code validation purposes only. In particular, they are required in order to run the data filtering Jupyter Notebook. These data are not to be used for any other purposes without consulting and crediting their original authors (see reference list at the bottom of this document). If you use any of the datasets in /repo/data/raw_measurements/ in your own research, you MUST consult and credit the ORIGINAL authors. Citing this repository is not sufficient in those cases. ### INSTRUCTIONS ### Do not modify the directory structure of this repository. All of the code has been pre-run, so the outputs needed to produce the figures from the manuscript are already present in the repo. If you wish to regenerate the main results on your own, you must run the code in the following order: 1. obs_filter_fig2_fig3_figS8-S17.ipynb (this filters the raw data) 2. normalize.py (this normalizes the filtered observations to extract the temperature response) 3. LM_optimization.py (this performs the LM optimization for all parameter configurations and all sites) 4. config_1d.m (K2 optimization using MCMC shown in Figure 6) and config_2o.m (CT1 and CT2 optimization using MCMC shown in Figure 7) Additional scripts and notebooks are included to reproduce the figures from the manuscript; these are labelled accordingly. Figure S1 was produced by modifying fig5.py to use the normalized STM measurements, and Figure S5 was produced by running config_1d.m and config_1e.py with the observation error set to a constant value of 0.05, then running fig6.py with these output data for the field sites AABC, ALH, and ACM. To run all MCMC configurations at all sites, a separate run_all.m script is included. The output of that script can be visualized using visualize_all_mcmc_results.py. The MATLAB code has been tested with R2013b and R2024b on both Linux and MacOS. The Python code and Jupyter notebooks have been tested and validated with Python version 3.8 on MacOS, provided that the necessary libraries (imported at the top of each script or Jupyter notebook) are installed. If necessary, a working Python library can be created using Anaconda and the libraries listed in "python_environment.txt" with Python 3.8.12 (note that this environment contains many more libraries than are strictly required for running the code in this repository; it is included here for the sake of reproducibility only). |
|---|