Land use change converts temperate dryland landscape into a net methane source
Data files
Mar 12, 2024 version files 206.54 KB
-
ghg_raw.csv
-
ghgstats.csv
-
landcover.xlsx
-
n_raw.csv
-
README.md
-
soilcn.csv
Abstract
Drylands cover approximately 40% of the global land surface and are thought to contribute significantly to the soil methane sink. However, large-scale methane budgets have not fully considered the influence of agricultural land use change in drylands, which often includes irrigation to create land cover types that support hay or grains for livestock production. These land cover types may represent a small proportion of the landscape but could disproportionately contribute to greenhouse gas exchange and are currently omitted in estimates of dryland methane fluxes. We measured greenhouse gas fluxes among big sagebrush, introduced wetlands, and hay meadows in a semi-arid temperate dryland in Wyoming, USA to investigate how these small-scale irrigated land cover types contributed to landscape-scale methane dynamics. Big sagebrush ecosystems dominated the landscape while the introduced wetlands and hay meadows represented 1% and 12% respectively. Methane uptake was consistent in the big sagebrush ecosystems, emissions and uptake were variable in the hay meadows, and emissions were consistent in the introduced wetlands. Despite making up 1% of the total land area, methane production in the introduced wetlands overwhelmed consumption occurring throughout the rest of the landscape, making this region a net methane source. Our work suggests that introduced wetlands and other irrigated land cover types created for livestock production may represent a significant, previously overlooked source of anthropogenic methane in this region and perhaps in drylands globally.
README: Land use change converts temperate dryland landscape into a net methane source
Raw flux data for methane, carbon dioxide, and other species were measured using a paired Picarro-Licor trace gas analyzer from June – August 2021 (flux data is in ghg_raw.csv
, data for statistical analysis in ghg_stats.csv
). Net nitrogen mineralization data was collected through ion exchange resins (data is in n_raw.csv
). Bulk density, soil texture, pH, and soil carbon and nitrogen were completed on soil samples analyzed at Yale University (data is in soilcn.csv
). Carbon and nitrogen were measured on an elemental analyzer. Soil texture was estimated using particle size analysis. pH was measured using a bench top pH meter. Land cover classification was done using the NatureServe database for Sublette County, WY (data is in landcover.xlsx
). In 'included' tab on landcover.xlsx, there are two sections of data on the left and right, not one rectangular table of data.
Description of the data and file structure
ghg_raw.csv:
- date: date of sampling (mm/dd/yy)
- week: week of sampling (Week 1 - Week 11)
- id: soil collar ID
- ranch: site name (site1-3)
- location: one of the four land cover types measured [meadow (hay meadow); bog (introduced wetland); slsage (sloping sagebrush); upsage (upland sagebrush)]
- sh/is: indicates whether sample was taken under shrub or in the interspace
- stemp: soil temperature ('b0C)
- pcent_sand: % sand content
- ph: soil pH
- chamber offset: height of soil collar from the surface of soil (measured in cm)
- volume add: volume added through extension of chamber attachment (cm3)
- start time: start time of sampling (hour:minute:second)
- end time: end time of sampling (hour:minute:second)
- time off: offset from the time read by the two instruments (s)
- co2 flux: carbon dioxide flux measured over 8 minute period (umol co2 m-2 s-1)
- co2 rsq: R-squared of exponential fit of values over 8 minute period
- ch4 flux: methane flux measured over 8 minute period (umol ch4 m-2 s-1)
- ch4 rsq: R-squared of exponential fit of values over 8 minute period
- n2o flux: nitrous oxide flux measured over 8 minute period (umol n2o m-2 s-1)
- n2o rsq: R-squared of exponential fit of values over 8 minute period
- nh3 flux: ammonia flux measured over 8 minute period (umol nh3 m-2 s-1)
- nh3 rsq: R-squared of exponential fit of values over 8 minute period
- h2o flux: h2o flux (% water vapor)
- h2o rsq: R-squared of exponential fit of values over 8 minute period
- notes: anything important to note from data collection of that day
ghgstats.csv:
- date: date of sampling (mm/dd/yy)
- week: week of sampling (Week 1 - Week 11)
- id: soil collar ID
- ranch: site name (site1-3)
- location: one of the four land cover types measured [meadow (hay meadow); bog (introduced wetland); slsage (sloping sagebrush); upsage (upland sagebrush)]
- stemp: soil temperature ('b0C)
- h2o flux: h2o flux (% water vapor)
- h2o rsq: R-squared of exponential fit of values over 8 minute period
- sand: % sand content (0-5 cm)
- clay: % clay content (0-5 cm)
- pH: soil pH (0-5 cm)
- carbon: % soil carbon (0-5 cm)
- nitrogen: % soil nitrogen (0-5 cm)
- co2 flux: carbon dioxide flux measured over 8 minute period (umol co2 m-2 s-1)
- co2 rsq: R-squared of exponential fit of values over 8 minute period
- ch4 flux: methane flux measured over 8 minute period (umol ch4 m-2 s-1)
- ch4 rsq: R-squared of exponential fit of values over 8 minute period
n_raw.csv:
- ranch: site name (site1-3)
- location: one of the four land cover types measured [meadow (hay meadow); bog (introduced wetland); slsage (sloping sagebrush); upsage (upland sagebrush)]
- no3: nitrate concentrations (ug no3/10cm2/burial length)
- nh4: ammonium concentrations (ug nh4/10cm2/burial length)
- fe: iron concentrations (ug fe/10cm2/burial length)
- s: sulfur concentrations (ug s/10cm2/burial length)
soilcn.csv:
- ranch: site name (site1-3)
- location: one of the four land cover types measured [meadow (hay meadow); bog (introduced wetland); slsage (sloping sagebrush); upsage (upland sagebrush)]
- depth: depth of soil sample (cm)
- finebulkd: fine soil bulk density (g soil cm-3)
- c: % soil carbon (0-5 cm)
- n: % soil N (0-5 cm)
- cpool: total soil carbon for 0-5 cm (g C m-2)
- npool: total soil N for 0-5 cm (g N m-2)
landcover.xlsx:
Sheet1 (ecosystems):
- gridcode: NatureServe grid code assignment
- Shape Area: area covered by a given land cover type (m2)
- Value: grid code assignment
- Count: Number of land cover types classified under a given name within the county boundary
- ECOLSYS_LU: Land cover type classification under NatureServe categories
Sheet2 (included):
- ecotype: Land cover type classification under NatureServe categories
- shape area: area covered by a given land cover type (m2)
- Total shape area: area covered by landscape of included cover types (m2)
- % introduced wetland: % of area represented by introduced wetland cover types within landscape
- % hay meadow: % of area represented by hay meadow cover types within landscape
- % big sagebrush: % of area represented by big sagebrush cover types within landscape
Code/Software
methanefigs.R:
used to make the primary figures for the manuscript and supplementary figures
Completed in R (version 4.3.2)
Packages: ggplot2
, wesanderson
, stringr
, zoo
, tidyverse
, scales
, ggpubr
, cowplot
, gt
, dplyr
methanestats.R:
-used to do the statistical analysis for the manuscript
-Completed in R (version 4.3.2)
-Packages: tidyverse
, ggpubr
, rstatix
, nlme
, lme4
, emmeans
, MASS
, glmm
Methods
Raw flux data for methane, carbon dioxide, and other species were measured using a paired Picarro-Licor trace gas analyzer from June – August 2021 (flux data is in ghg_raw.csv, data for statistical analysis in ghg_stats.csv). Net nitrogen mineralization data was collected through ion exchange resins (data is in n_raw.csv). Bulk density, soil texture, pH, and soil carbon and nitrogen were completed on soil samples analyzed at Yale University (data is in soilcn.csv). Carbon and nitrogen were measured on an elemental analyzer. Soil texture was estimated using particle size analysis. pH was measured using a bench top pH meter. Land cover classification was done using the NatureServe database for Sublette County, WY (data is in landcover.xlsx). In 'included' tab on landcover.xlsx, there are two sections of data on the left and right, not one rectangular table of data.