The Biogeography of a climate pulse: A global map of forest CH₄:CO₂ emission ratios

On September 17, 2024, we carried out a systematic search of the Web of Science Core Collection with the query string “(greenhouse gas emissions) AND (forest).” The search yielded 4,448 results. Over the initial screening phase, we evaluated the titles and abstracts of all 4,448 articles to assess their eligibility for inclusion (Fig. S1). We then manually selected 336 qualifying studies through application of the following criteria: 1. The study presented primary data (i.e., we excluded data synthesis) that had been assayed under field settings. 2. The study was carried out in a woody ecosystem. 3. The study contained information for either carbon dioxide (CO2) or methane (CH4) fluxes or both. We later (over the analysis of the ratio) filtered out those studies that contained estimates for only one of the emission types. We processed the remaining studies in the dataset by extracting the following variables: Study location: categorical variable. Typically, we abstracted the variable with th country name and if available specific site names are included. Latitude of the study location: continuous variable. When not reported, we extracted the latitude of the central point in the study area based on the location data. Longitude of the study location: continuous variable. When not reported, we extracted the longitude of the central point in the study area based on the location data. Forest successional stage: categorical variable. We divided the data into two classes: pristine natural forest and plantation forest. Forest type: categorical variable. Categorical variable with five levels: evergreen, deciduous, conifer, broadleaf mangroves and mixed. Species name: categorical variable. We extracted the dominant species in the study's forest ecosystem. Soil pH: continuous variable. When given a range of pH values, the median value was selected as the pH value. and retain one decimal place. SOC: continuous variable. When given a range of SOC values, the median value is selected as the SOC value, two decimal places are retained, and the units are harmonized to g kg-1. TN: continuous variable. When given a range of TN values, the median value is selected as the TN value, two decimal places are retained, and the units are harmonized to g kg-1. Bulk density (BD): a continuous variable. We used the median BD value per stud yrounded to two decimal places, and expressed in g cm-3. We calculate the missing bulk density from the sand/clay percentages via the Soil Texture Calculator (https://agricomputing.com/soil-texture-calculator/). Mean CO2 fluxes: continuous variables. We express values in mg m-2 h-1 with two-decimal precision. Mean CH4 fluxes: a continuous variable. We express values in μg m-2 h-1 with two-decimal precision. Sample size: continuous variable. Date: continuous variable. We reported on the year when the study got initiated. Season: categorical variable. We infer seasons from dates, using the “year average” for seasons when the date is a certain year. Note that the seasons are reversed in the Northern and Southern Hemispheres. Measurement approach: categorical variable. Instruments or methods used in research to measure greenhouse gas fluxes. For studies in which the data were presented graphically, we used the plot digitizer (https://plotdigitizer.com).