data set for EC 30 mins for the two sites

Dataset Description This dataset and accompanying R scripts support the intensive carbon dynamics observation platform conducted in tropical alpine peatlands of Guatavita, Colombia. The data include half-hourly and cumulative greenhouse-gas fluxes (CO₂, CH₄, N₂O), dissolved organic carbon (DOC) transport, and related hydrological and meteorological measurements, together with model outputs and analysis scripts. All analyses were performed in R (version ≥ 4.2). The repository is organized into two main components: Chamber and Bayesian analysis pipeline (root folder) Tower flux gap-filling and uncertainty analysis (folder golden/) 1. Chamber and Bayesian Workflow This section integrates chamber measurements, water-table data, and modeled fluxes for both conserved and degraded peatland plots. The scripts allow data preparation, prediction of half-hourly fluxes, Bayesian partitioning of net ecosystem exchange (NEE) into gross primary production (GPP) and ecosystem respiration (ER), and generation of publication-quality figures. Main steps: Data preparation – Cleaning and merging chamber and tower data (flux_chamber3.r, flux_wt_guatavita_jc.r, waterlevel.r). Prediction dataset construction – Builds model input datasets (flux predict.R, flux predict2.R). Bayesian flux partitioning – Separates NEE into GPP and ER using hierarchical Bayesian models (bayesian models.r, bayesianflux.r). This step must be run separately for each station (ST1 and ST2) by modifying the station code inside the scripts. Trace gas analyses – Quantifies N₂O and DOC fluxes (N2Oflux.r, DOC_flux.r). Visualization and summaries – Produces the cumulative and seasonal flux figures and summary tables (final plot.r). Primary outputs: Modelled CO₂ and CH₄ fluxes (*_Model_EC_long.csv, *_pred_30min_*.csv) Seasonal and cumulative carbon balance summaries (Final_Cumulative_CO2_CH4_CO2eq_2023_2024_bySeason_Method_Station.csv, Summary_CO2_CH4_CO2eq_byMethod_Station_Season_Year.csv) Mean and confidence-interval tables for each gas (PerGas_CO2_CH4_with_CO2eq_Mg_ha_mean95CI.csv, Totals_CO2eq_across_gases_Mg_ha_mean95CI.csv) Publication figures (figure.png, figure_transparent.png, figure.svg) 2. Tower Flux (Eddy-Covariance) Workflow The folder golden/ contains the workflow used for tower-based fluxes, including gap-filling, uncertainty analysis, and manuscript-quality visualization. These scripts use the REddyProc R package and standard meteorological variables. Scripts: REddyProc_Guatavita_Station1_Gold.R – Gap-filling for Station 1 REddyProc_Guatavita_Station2_Gold.R – Gap-filling for Station 2 Guatavita_gapfilling_uncertainty.R – Quantifies gap-filling uncertainty Guatavita_plot_manuscript.R – Generates final tower flux figures Each station’s eddy-covariance data were processed independently following standard u-star filtering and uncertainty propagation routines. Data Files Input data include chamber fluxes (co2flux.csv, ch4flux.csv, db_gutavita_N2O_all.csv), water-table and hydrological measurements (WaterTable.csv, wtd_martos_21_25.csv), DOC transport (DOC transport.csv), and auxiliary meteorological variables (tower_var.csv). Intermediate model results are stored in .rds files, and cumulative or seasonal summaries are provided in .csv and .xlsx formats. Reproducibility Notes All scripts assume relative paths from the project root. To reproduce the complete analyses: Install required R packages (tidyverse, ggplot2, rjags, coda, REddyProc, among others). Run the chamber workflow in the order listed above. Repeat the Bayesian modeling step for both stations. Execute the tower scripts in the golden/ folder for gap-filling and visualization. Large intermediate .rds files are retained for reproducibility and should not be deleted unless re-running the models from scratch. Citation and Contact Principal Investigator: Juan C. Benavides, Pontificia Universidad Javeriana, Bogotá, Colombia

数据集说明
本数据集及配套R脚本，支撑哥伦比亚瓜塔维塔（Guatavita）热带高山泥炭地的高精度碳动态观测平台研究。数据涵盖半小时尺度与累积态温室气体通量（CO₂、CH₄、N₂O）、溶解性有机碳（Dissolved Organic Carbon, DOC）输运数据，以及相关水文与气象观测资料，同时包含模型输出结果与分析脚本。所有分析均基于R语言（版本≥4.2）完成。
本代码仓库分为两大核心模块：箱室法与贝叶斯分析流程（根目录）、塔基通量间隙填充及不确定性分析（golden/文件夹）

1. 箱室法与贝叶斯分析工作流
本模块整合了原生与退化泥炭地样地的箱室观测数据、地下水位数据及模拟通量结果。配套脚本可实现数据预处理、半小时尺度通量预测、将生态系统净交换（Net Ecosystem Exchange, NEE）贝叶斯拆解为总初级生产力（Gross Primary Production, GPP）与生态系统呼吸（Ecosystem Respiration, ER），以及生成符合期刊出版标准的可视化图表。
主要流程步骤：
数据预处理：清洗并合并箱室与塔基观测数据（对应脚本：flux_chamber3.r、flux_wt_guatavita_jc.r、waterlevel.r）。
预测数据集构建：生成模型输入数据集（对应脚本：flux_predict.R、flux_predict2.R）。
贝叶斯通量拆解：基于层级贝叶斯模型将NEE拆解为GPP与ER（对应脚本：bayesian_models.r、bayesianflux.r）。该步骤需通过修改脚本内的站点代码，分别对ST1与ST2两个站点独立运行。
痕量气体分析：量化N₂O与溶解性有机碳通量（对应脚本：N2Oflux.r、DOC_flux.r）。
可视化与结果汇总：生成累积通量与季节变化通量图及汇总表格（对应脚本：final_plot.r）。

核心输出结果：
模拟CO₂与CH₄通量文件（*_Model_EC_long.csv、*_pred_30min_*.csv）
季节与累积碳平衡汇总文件（Final_Cumulative_CO2_CH4_CO2eq_2023_2024_bySeason_Method_Station.csv、Summary_CO2_CH4_CO2eq_byMethod_Station_Season_Year.csv）
各气体的均值与置信区间表格（PerGas_CO2_CH4_with_CO2eq_Mg_ha_mean95CI.csv、Totals_CO2eq_across_gases_Mg_ha_mean95CI.csv）
出版级可视化图表（figure.png、figure_transparent.png、figure.svg）

2. 塔基通量（涡度协方差，Eddy-Covariance）工作流
golden/文件夹内包含塔基通量的完整工作流程，涵盖间隙填充、不确定性分析及符合论文出版标准的可视化环节。所有脚本均使用REddyProc R包及标准气象变量完成计算。
配套脚本：
REddyProc_Guatavita_Station1_Gold.R：站点1的通量间隙填充脚本
REddyProc_Guatavita_Station2_Gold.R：站点2的通量间隙填充脚本
Guatavita_gapfilling_uncertainty.R：量化间隙填充过程的不确定性
Guatavita_plot_manuscript.R：生成最终塔基通量可视化图表
各站点的涡度协方差数据均按照标准u*滤波与不确定性传递流程独立处理。

数据文件
输入数据包含箱室通量数据（co2flux.csv、ch4flux.csv、db_gutavita_N2O_all.csv）、地下水位与水文观测数据（WaterTable.csv、wtd_martos_21_25.csv）、溶解性有机碳输运数据（DOC transport.csv）及辅助气象变量数据（tower_var.csv）。
模型中间结果存储于.rds格式文件中，累积或季节汇总结果以.csv与.xlsx格式提供。

可复现性说明
所有脚本均默认使用项目根目录下的相对路径。如需完整复现分析流程，请执行以下步骤：
1. 安装所需R包（包括tidyverse、ggplot2、rjags、coda、REddyProc等）。
2. 按照前述顺序运行箱室法工作流脚本，并分别为两个站点重复执行贝叶斯建模步骤；随后运行golden/文件夹内的塔基通量脚本完成间隙填充与可视化环节。
3. 为保障可复现性，大型模型中间.rds文件将保留，除非需要从头重新运行模型，否则请勿删除此类文件。

引用与联系方式
项目负责人：Juan C. Benavides，哥伦比亚波哥大哈维里亚那天主教大学（Pontificia Universidad Javeriana）