Package 'FluxSeparator'

FluxSeparator: Separation of Diffusive and Ebullitive Fluxes

Description

Separates diffusive and ebullitive (bubble) fluxes from continuous concentration measurements using a running variance approach.

Author(s)

Jonas Stage Sø [email protected]

---

Diffusive flux

Description

Separates the diffusive and ebullitive fluxes, to calculate the diffusive flux, as a linear function of concentrations over time. This is done by firstly finding all events that are considered ebullitive (for more info see ebullitive_flux).

Several factors can be set to determine what is considered ebullitive events, remove observations before doing diffusive flux, number of observations used in the diffusive flux calculation, cutoffs if concentrations start too high and number of observations needed in the linear model.

Output data is converted to concentration change per hour.

Usage

diffusive_flux(
  data,
  concentration_values = "pred_CH4",
  runvar_cutoff = 0.5,
  remove_observations_prior = 200,
  number_of_observations_used = 400,
  show_plots = TRUE,
  IndexSpan = 30,
  cutoff_start_value = Inf,
  number_of_observations_required = 50,
  number_of_pumpcycles_in_plot = 50,
  smooth_data = FALSE,
  look_for_bubbles = TRUE,
  Hutchinson_Mosier_correction = FALSE,
  volume,
  area,
  min_obs_per_cycle = 100,
  time_gap_seconds = 30,
  runvar_window = 5,
  smooth_window = 10
)

Arguments

data	Your data frame. Must contain columns datetime, PumpCycle, station, and tempC.
concentration_values	Name of your variable representing the concentration.
runvar_cutoff	Cutoff of the running variance, which is used to determine if an increase in concentration is an ebullitive event. Lower values increases number of ebullitive events registered.
remove_observations_prior	Remove n number of observations before calculating the diffusive flux by a linear slope.
number_of_observations_used	Number of observations used to calculate the diffusive flux by a linear slope.
show_plots	Show plots which can assist in the determination of good fits for the model. A boolean variable which should be TRUE or FALSE.
IndexSpan	Number of observations which are included before and after an ebullitive event, to ensure the entire event is determined.
cutoff_start_value	Variable indicating what the maximum starting concentration can be. Defaults to Inf (no filtering).
number_of_observations_required	Number of observations required in each cycle for the function to compute a linear model on the data.
number_of_pumpcycles_in_plot	Number of cycles which are plotted. Used only if show_plots = TRUE.
smooth_data	Computes a running mean on the concentration data five times, to smoothen data if data is low bit resolution. See Sø et al. (2023) for more information.
look_for_bubbles	Can be used for the function to not consider ebullitive events. Can be useful when calculating diffusive CO2 flux.
Hutchinson_Mosier_correction	Can be used to correct flux measurements based on the Hutchinson-Mosier correction (1981). However, fluxes are only calculated of three points. A boolean variable which should be TRUE or FALSE.
volume	Volume of the chamber used for calculating fluxes (L). This is only needed if Hutchinson_Mosier_correction = TRUE.
area	Surface area of the chamber used for calculating fluxes ($m^2$). This is only needed if Hutchinson_Mosier_correction = TRUE.
min_obs_per_cycle	Minimum number of observations per pump cycle to include in analysis (default 100).
time_gap_seconds	Time gap in seconds used to separate event groups when detecting bubbles (default 30).
runvar_window	Window size for the running variance calculation (default 5).
smooth_window	Window size for the running mean smoothing passes (default 10). Only used when smooth_data = TRUE.

Value

A data frame containing the following:

• station - The station from the input data

• PumpCycle - Cycle number

• datetime_start - Start time of the cycle

• datetime_end - End time of the cycle

• slope_concentration_hr - The diffusive flux per hour ($ppm\, h^{-1}$)

• slope_standard_error - The standard error of the flux

• n_obs_included_in_lm - Number of observations used to calculate the diffusive flux

• r2 - Variance explained by the linear model

• temp - Average temperature within the chamber (C)

• hmr_slope - Flux calculated using the HMR package ($ppm\, h^{-1}$). This is only returned if the method selected by the HMR package is either Hutchinson-Mosier correction or No flux, in case of LR used the function will return NA. Only given if the Hutchinson-Mosier correction is calculated.

• hmr_se - Standard error of the hmr_slope. This is only returned if the hmr_slope is calculated.

• hmr_pvalue - The p-value for the null hypothesis of zero flux. This is only returned if the hmr_slope is calculated.

• hmr_lower95 - The lower end-point of the 95\ for the flux. This is only returned if the hmr_slope is calculated.

• hmr_upper95 - The upper end-point of the 95\ for the flux. This is only returned if the hmr_slope is calculated.

• method - The method used by the HMR package to calculate flux. This is only returned if the hmr_slope is calculated.

When show_plots = TRUE, the returned data frame carries an attribute "plots" containing a list of ggplot2 objects.

Author(s)

Jonas Stage Sø [email protected]

References

Sø et al. (2024). Self-Made Equipment for Automatic Methane Diffusion and Ebullition Measurements From Aquatic Environments. doi:10.1029/2024JG008035.

Sø et al. (2023). Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. doi:10.1016/j.scitotenv.2023.162895.

Hutchinson, G.L. and Mosier, A.R. (1981). Improved soil cover method for field measurement of nitrous oxide fluxes. Soil Science Society of America Journal, 45, pp. 311-316.

Pullens, J.W.M., Abalos, D., Petersen, S.O. and Pedersen, A.R. (2023). Identifying criteria for greenhouse gas flux estimation with automatic and manual chambers: A case study for N2O. European Journal of Soil Science, 74, e13340. doi:10.1111/ejss.13340.

See Also

ebullitive_flux, ppm_to_umol

Examples

library(FluxSeparator)

data(DIY_sensor_data)

DIY_sensor_data %>%
  diffusive_flux(cutoff_start_value = 450, show_plots = TRUE)
  # 450 would be good for CO2, while 5 could be good for CH4

---

Data from Lake Lyng

Description

A subset of data from the paper Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. One measurement containing only diffusive flux and one containing ebullitive events.

Usage

DIY_sensor_data

Format

A data frame with 2,201 rows and 12 variables:

datetime

Date and time of measurement

RH

Relative humidity (percent)

tempC

Temperature in degrees Celsius

CH4smV

Methane sensor voltage

K33_RH

Relative humidity measured by the CO2 sensor

K33_Temp

Temperature in degrees Celsius measured by the CO2 sensor

K33_CO2

CO2 concentration in ppm

SampleNumber

Sample number in this pump cycle

PumpCycle

Pump cycle which counts upwards after the chamber has been flushed

pred_CH4

Predicted methane concentration, calculated using the read_CH4_files function and calibration values

station

Station name, needed for diffusive and ebullitive flux calculations

sensor

Sensor identification number

Source

Sø et al. (2023). Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. doi:10.1016/j.scitotenv.2023.162895

Examples

data(DIY_sensor_data)
head(DIY_sensor_data)

---

Ebullitive flux

Description

Separates the diffusive and ebullitive fluxes, to calculate the ebullitive flux, as the change in concentration from ebullitive events. This is done by computing a running variance; if the running variance exceeds a customizable cutoff value it is considered an ebullitive event.

Additional factors can be set to determine what is considered ebullitive events. Output data is converted to concentration change per hour.

Usage

ebullitive_flux(
  data,
  concentration_values = "pred_CH4",
  top_selection = "last",
  runvar_cutoff = 0.5,
  show_plots = TRUE,
  IndexSpan = 30,
  concentration_diffusion_cutoff = 1,
  number_of_pumpcycles_in_plot = 24,
  smooth_data = FALSE,
  min_obs_per_cycle = 100,
  time_gap_seconds = 30,
  runvar_window = 5,
  smooth_window = 10
)

Arguments

data	Your data frame. Must contain columns datetime, PumpCycle, station, and tempC.
concentration_values	Name of your variable representing the concentration.
top_selection	Can be set to "last" or "max" to either use the last or maximum concentration value in each ebullitive event.
runvar_cutoff	Cutoff of the running variance, which is used to determine if an increase in concentration is an ebullitive event. Lower values increase the number of ebullitive events registered.
show_plots	Show diagnostic plots. A logical; defaults to TRUE.
IndexSpan	Number of observations which are included before and after an ebullitive event, to ensure the entire event is captured.
concentration_diffusion_cutoff	Minimum concentration change that is considered an ebullitive event.
number_of_pumpcycles_in_plot	Number of cycles which are plotted. Used only if show_plots = TRUE.
smooth_data	Computes a running mean on the concentration data five times, to smooth data if data is low bit resolution. See Sø et al. (2023) for more information.
min_obs_per_cycle	Minimum number of observations per pump cycle required for processing (default 100).
time_gap_seconds	Time gap in seconds used to separate bubble event groups (default 30).
runvar_window	Window size for the running variance computation (default 5).
smooth_window	Window size for each running-mean smoothing pass (default 10). Only used when smooth_data = TRUE.

Value

A data frame containing the following:

• station
  The station column from the input data

• PumpCycle
  Cycle number

• datetime_start
  Start time of the cycle

• datetime_end
  End time of the cycle

• sum_bubbles_concentration
  The sum of the differences in concentration caused by bubbles

• n_bubbles
  Number of bubbles detected. Bear in mind that this function has difficulties detecting the number of bubbles if they are close to each other

• pumpcycle_duration_hr
  Length of the cycle duration in hours

• temp
  Average temperature within the chamber

• bubbles_per_time
  Amount of bubbles divided by the duration of the cycle in hours

• concentration_per_time
  Ebullitive flux, as the sum of concentration change divided by the duration in hours

When show_plots = TRUE, a "plots" attribute is attached to the result containing a list of ggplot2 objects.

Author(s)

Jonas Stage Sø [email protected]

References

Sø et al. (2024). Self-Made Equipment for Automatic Methane Diffusion and Ebullition Measurements From Aquatic Environments. doi:10.1029/2024JG008035.

Sø et al. (2023). Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. doi:10.1016/j.scitotenv.2023.162895.

See Also

diffusive_flux, ppm_to_umol

Examples

library(FluxSeparator)

data(DIY_sensor_data)

DIY_sensor_data %>%
  ebullitive_flux()

---

FluxSeparator Shiny App

Description

Shiny app to visualize the process of separating diffusive and ebullitive fluxes. The app also gives the user the ability to mark areas where diffusive fluxes should be calculated.

Usage

FluxSeparatorApp(...)

Arguments

...	Additional arguments passed to shinyApp().

Author(s)

Jonas Stage Sø [email protected]

References

Sø et al. (2024). Self-Made Equipment for Automatic Methane Diffusion and Ebullition Measurements From Aquatic Environments. doi:10.1029/2024JG008035.

Sø et al. (2023). Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. doi:10.1016/j.scitotenv.2023.162895.

Hutchinson, G.L. and Mosier, A.R. (1981). Improved soil cover method for field measurement of nitrous oxide fluxes. Soil Science Society of America Journal, 45, pp. 311-316.

Pullens, J.W.M., Abalos, D., Petersen, S.O. and Pedersen, A.R. (2023). Identifying criteria for greenhouse gas flux estimation with automatic and manual chambers: A case study for N2O. European Journal of Soil Science, 74, e13340. doi:10.1111/ejss.13340.

See Also

diffusive_flux, ebullitive_flux, ppm_to_umol

Examples

library(FluxSeparator)

FluxSeparatorApp()

---

ppm_to_µmol

Description

Conversion of $ppmV$ to $\mu mol$ $m^{-2}$ $h^{-1}$ using the ideal gas law.

Usage

ppm_to_umol(pressure, concentration, volume, temperature_C, area)

Arguments

pressure	Air pressure during measurement (Pa).
concentration	Concentration of the gas in ppm (µmol/mol).
volume	Volume of the chamber used for measuring in $m^{3}$.
temperature_C	Temperature in degrees Celsius in the chamber.
area	Surface area of the chamber used in $m^{2}$.

Value

A numeric vector of flux values in $\mu mol$ $m^{-2}$ $h^{-1}$.

Author(s)

Jonas Stage Sø [email protected]

Examples

# Convert a single value
ppm_to_umol(pressure = 101325, concentration = 10,
            volume = 0.01, temperature_C = 20, area = 0.05)

---

read_CH4_files

Description

A function to ease the import of data from DIY sensors, which reads a CSV file, calculates the absolute humidity, V0, RsR0, and the methane concentration.

Usage

read_CH4_files(data, files, join_model_coef = TRUE, model_coef_data = NULL)

Arguments

data	A data frame containing the path, sensor identification, and model coefficients for this specific sensor. Model coefficients can also be read in as a separate data frame and defined in the model_coef_data variable.
files	A vector supplying the path to the file being read.
join_model_coef	Boolean variable. Join data with dataframe model_coef_data to convert sensor voltage signal to methane concentration.
model_coef_data	Data frame consisting of the calibration values used to convert sensor voltage signal to methane concentration. Required when join_model_coef = TRUE.

Value

A data frame output including all the original values, with the exception of the model coefficients.

pred_CH4

Computed from the calibration model. The CH4 concentration calculated from the sensor resistance, expressed in ppm.

Author(s)

Jonas Stage Sø [email protected]

References

Sø et al. (2023). Methane and carbon dioxide fluxes at high spatiotemporal resolution from a small temperate lake. doi:10.1016/j.scitotenv.2023.162895

Sø et al. (2024). Self-Made Equipment for Automatic Methane Diffusion and Ebullition Measurements From Aquatic Environments. doi:10.1029/2024JG008035

See Also

ebullitive_flux, diffusive_flux, ppm_to_umol

Examples

## Not run: 
library(FluxSeparator)

# read in model coef
model_coef <- read_csv("model_coef.csv")

# path to DIY sensors files
path_to_files <- list.files(pattern = ".csv")

# create data frame for path, sensor and station.
data_path <- tibble(path = path_to_files,
                    sensor = c(1, 2, 3, 4),
                    station = c(1, 2, 4, 3))


# join with model_coef and calculate CH4 in ppm.
read_CH4_files(data_path, path,
               model_coef_data = model_coef)

#### Example using join_model_coef = FALSE ####

# join with model_coef.
joined_data_path <- left_join(data_path, model_coef, by = join_by(sensor))

# calculate CH4 in ppm.
read_CH4_files(joined_data_path,
               path,
               join_model_coef = FALSE)

## End(Not run)

---