The Fluxible R package is made to transform any dataset of gas concentration over time into a gas flux dataset. It was originally made to be used with a closed loop chamber system connected to a gas analyzer.
The goal of fluxible is to provide a workflow that removes individual evaluation of each flux, reduces risk of bias, and makes it reproducible. Users set specific data quality standards and selection parameters as function arguments that are applied to the entire dataset. Fluxible offers different methods to estimate fluxes: linear, quadratic, exponential (Zhao et al., 2018), and the original HM model (Hutchinson and Mosier, 1981; Pedersen et al., 2010). The kappamax method (Hüppi et al., 2018) is also included, at the quality control step. The package runs the calculations automatically, without prompting the user to take decisions mid-way, and provides quality flags and plots at the end of the process for a visual check.
This makes it easy to use with large flux datasets and to integrate into a reproducible and automated data processing pipeline such as the targets R package (Landau, 2021). Using the Fluxible R package makes the workflow reproducible, increases compatibility across studies, and is more time efficient.
For a visual overview of the package, see the poster.
Fluxible can be installed from CRAN.
install.packages("fluxible")You can install the development version of fluxible from the GitHub repo with:
# install.packages("devtools")
devtools::install_github("plant-functional-trait-course/fluxible")library(fluxible)
conc_df <- flux_match(
co2_df_short,
record_short,
datetime,
start,
measurement_length = 220
)
slopes_df <- flux_fitting(
conc_df,
conc,
datetime,
fit_type = "exp_zhao18",
end_cut = 60
)
#> Cutting measurements...
#> Estimating starting parameters for optimization...
#> Optimizing fitting parameters...
#> Calculating fits and slopes...
#> Done.
slopes_flag_df <- flux_quality(
slopes_df,
conc
)
#>
#> Total number of measurements: 6
#>
#> ok 6 100 %
#> discard 0 0 %
#> zero 0 0 %
#> force_discard 0 0 %
#> start_error 0 0 %
#> no_data 0 0 %
#> force_ok 0 0 %
#> force_zero 0 0 %
#> force_lm 0 0 %
#> no_slope 0 0 %
flux_plot(
slopes_flag_df,
conc,
datetime,
f_ylim_lower = 390,
f_ylim_upper = 650,
facet_wrap_args = list(
ncol = 3,
nrow = 2,
scales = "free"
)
)
#> Plotting in progress
Output of flux_plot, showing fluxes plotted individually with diagnostics and quality flags.
fluxes_df <- flux_calc(
slopes_flag_df,
f_slope_corr,
datetime,
temp_air,
conc_unit = "ppm",
flux_unit = "mmol/m2/h",
cols_keep = c("turfID", "type"),
cols_ave = c("temp_soil", "PAR"),
setup_volume = 24.575,
atm_pressure = 1,
plot_area = 0.0625
)
#> Cutting data according to 'keep_arg'...
#> Averaging air temperature for each flux...
#> Creating a df with the columns from 'cols_keep' argument...
#> Creating a df with the columns from 'cols_ave' argument...
#> Calculating fluxes...
#> R constant set to 0.082057
#> Concentration was measured in ppm
#> Fluxes are in mmol/m2/h
fluxes_gpp <- flux_gpp(
fluxes_df,
type,
datetime,
id_cols = "turfID",
cols_keep = c("temp_soil_ave")
)
#> Warning in flux_gpp(fluxes_df, type, datetime, id_cols = "turfID", cols_keep = c("temp_soil_ave")):
#> NEE missing for measurement turfID: 156 AN2C 156
fluxes_gpp
#> # A tibble: 9 × 5
#> datetime type f_flux temp_soil_ave turfID
#> <dttm> <chr> <dbl> <dbl> <chr>
#> 1 2022-07-28 23:43:25 ER 51.9 10.9 156 AN2C 156
#> 2 2022-07-28 23:47:12 GPP 9.72 10.7 74 WN2C 155
#> 3 2022-07-28 23:47:12 NEE 32.0 10.7 74 WN2C 155
#> 4 2022-07-28 23:52:00 ER 22.3 10.7 74 WN2C 155
#> 5 2022-07-28 23:59:22 GPP -6.63 10.8 109 AN3C 109
#> 6 2022-07-28 23:59:22 NEE 44.3 10.8 109 AN3C 109
#> 7 2022-07-29 00:03:00 ER 50.9 10.5 109 AN3C 109
#> 8 2022-07-29 00:06:25 GPP NA 12.2 29 WN3C 106
#> 9 2022-07-29 00:06:25 NEE 32.7 12.2 29 WN3C 106The licoread R package, developped in collaboration with LI-COR, provides an easy way to import raw files from Li-COR gas analyzers as R objects that can be used directly with the Fluxible R package.
We are working on a tool to automatically select the window of the measurement on which to fit a model. This selection will be based on environmental variable, such as photosynthetically active radiation (PAR), measured simultaneously.
As we want fluxible to fit the use of as many projects as possible, more fitting expressions will be included in the flux_fitting function. Feel welcome to get in touch if you wish to include yours in fluxible.
Joseph Gaudard, University of Bergen, Norway
If you are running a course and want to talk about Fluxible, feel free to use this two-slides presentation. Of course, you can always reach out if you wish to have more material.
Gaudard, J., Trepel, J., Dawson, H. R., Enquist, B., Halbritter, A. H., Mustri, M., Niittynen, P., Santos-Andrade, P. E., Topper, J. P., Vandvik, V., and Telford, R. J.: Fluxible: an R package to calculate ecosystem gas fluxes from closed loop chamber systems in a reproducible and automated workflow (slides), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12409, https://doi.org/10.5194/egusphere-egu25-12409, 2025.
Gaudard, J., Telford, R., Vandvik, V., and Halbritter, A. H.: Fluxible: an R package to calculate ecosystem gas fluxes in a reproducible and automated workflow. (poster), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-956, https://doi.org/10.5194/egusphere-egu24-956, 2024.
Fluxible builds on the earlier effort from the Plant Functional Traits Course Community co2fluxtent (Brummer et al., 2023).