Last updated: 2021-02-11

Checks: 7 0

Knit directory: BloomSail/

This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.


Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

The command set.seed(20191021) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version 4bdeb1f. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    data/
    Ignored:    output/Plots/Figures_publication/.tmp.drivedownload/

Untracked files:
    Untracked:  output/Plots/response_time/all_plots.pdf
    Untracked:  output/Plots/response_time/profiles_pCO2.pdf
    Untracked:  output/Plots/response_time/profiles_pCO2_delta_grid.pdf
    Untracked:  output/Plots/response_time/profiles_pCO2_delta_rel_grid.pdf
    Untracked:  output/Plots/response_time/profiles_pCO2_grid.pdf
    Untracked:  output/Plots/response_time/time_series_depth_pCO2_corr_by_profile.pdf

Unstaged changes:
    Deleted:    analysis/CT_dynamics.Rmd
    Modified:   analysis/_site.yml
    Modified:   code/Workflowr_project_managment.R
    Modified:   output/Plots/Figures_publication/Appendix/Fig_A1.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_A1.png
    Modified:   output/Plots/Figures_publication/Appendix/Fig_A2.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_B1.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_B2.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_C1.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_C2.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_C3.pdf
    Modified:   output/Plots/Figures_publication/Appendix/Fig_C4.pdf
    Modified:   output/Plots/Figures_publication/Article/Fig_1.pdf
    Modified:   output/Plots/Figures_publication/Article/Fig_2.pdf
    Modified:   output/Plots/Figures_publication/Article/Fig_3.pdf
    Modified:   output/Plots/Figures_publication/Article/Fig_4.pdf
    Modified:   output/Plots/Figures_publication/Article/Fig_5.pdf
    Modified:   output/Plots/response_time/tau_determination_pCO2_corr_flushperiods_nls.pdf

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.


These are the previous versions of the repository in which changes were made to the R Markdown (analysis/read-in.Rmd) and HTML (docs/read-in.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File Version Author Date Message
html 4dfc4f8 jens-daniel-mueller 2021-02-11 Build site.
Rmd 8dc3356 jens-daniel-mueller 2021-02-11 cleaning
html 8e5e6d9 jens-daniel-mueller 2021-02-11 Build site.
Rmd 40f9a07 jens-daniel-mueller 2021-02-11 cleaning
html 8051798 jens-daniel-mueller 2021-02-10 Build site.
Rmd 9477a18 jens-daniel-mueller 2021-02-10 rerun all with empty folders
html ed2a1b2 jens-daniel-mueller 2021-02-08 Build site.
Rmd 833d0f6 jens-daniel-mueller 2021-02-08 code and output cleaning
html 7027b12 jens-daniel-mueller 2021-02-08 Build site.
Rmd 3d6eaa1 jens-daniel-mueller 2021-02-08 code and output cleaning
html c5fc34c jens-daniel-mueller 2021-01-22 Build site.
html 4277235 jens-daniel-mueller 2021-01-05 Build site.
html 9a3f42a jens-daniel-mueller 2020-10-24 Build site.
html 05248bf jens-daniel-mueller 2020-10-20 Build site.
html 1c4fe8e jens-daniel-mueller 2020-10-20 table with time series in depth intervals added
html a43fcac jens-daniel-mueller 2020-10-13 Build site.
Rmd a23fb5d jens-daniel-mueller 2020-10-13 removed unnecessary output prints
html 6896725 jens-daniel-mueller 2020-10-01 Build site.
Rmd f2f9226 jens-daniel-mueller 2020-10-01 tower data converted to UTC
html 9f66019 jens-daniel-mueller 2020-10-01 Build site.
html 27c5431 jens-daniel-mueller 2020-09-29 Build site.
Rmd 2e0f902 jens-daniel-mueller 2020-09-29 all parameters separate, rebuild
html 1d01685 jens-daniel-mueller 2020-09-28 Build site.
Rmd d28129f jens-daniel-mueller 2020-09-28 republish after tau factor set to 1 and using final pCO2 data
html 02a1609 jens-daniel-mueller 2020-09-25 Build site.
Rmd 99e69cf jens-daniel-mueller 2020-09-25 activated read-in of th and ts data
Rmd 616c27f jens-daniel-mueller 2020-09-25 updated repo manually
html 904f0f7 jens-daniel-mueller 2020-09-23 Build site.
html 66bf52a jens-daniel-mueller 2020-09-23 Build site.
Rmd 0c8eed6 jens-daniel-mueller 2020-09-23 included postprocessed cleaned data
html c919fb7 jens-daniel-mueller 2020-06-29 Build site.
Rmd 1461cb6 jens-daniel-mueller 2020-06-29 Fig update for talk
html 603af23 jens-daniel-mueller 2020-05-25 Build site.
html 3414c23 jens-daniel-mueller 2020-05-25 Build site.
html 9ccd9a3 jens-daniel-mueller 2020-05-25 Build site.
Rmd 9bedac5 jens-daniel-mueller 2020-05-25 revised pp time series plot
html dd3bd89 jens-daniel-mueller 2020-05-07 Build site.
Rmd ad98da2 jens-daniel-mueller 2020-05-07 harmonized parameter labeling
html 3832733 jens-daniel-mueller 2020-04-30 Build site.
Rmd 4f4ab08 jens-daniel-mueller 2020-04-30 harmonized code until RT determination
html 1b6480f jens-daniel-mueller 2020-04-30 Build site.
Rmd fe72316 jens-daniel-mueller 2020-04-30 revised variable and object names, used temp-dependent tau only, rerun code
html d9248a6 jens-daniel-mueller 2020-04-29 Build site.
Rmd 70bd3f0 jens-daniel-mueller 2020-04-29 correct interpolation, new d pco2 plot
html aa52c73 jens-daniel-mueller 2020-04-28 Build site.
Rmd 3044ec0 jens-daniel-mueller 2020-04-28 completely revised
html 57f7231 jens-daniel-mueller 2020-04-28 Build site.
Rmd 5ebd364 jens-daniel-mueller 2020-04-28 revised
html b5722a7 jens-daniel-mueller 2020-04-28 Build site.
html 472c2b4 jens-daniel-mueller 2020-04-21 Build site.
html f8fcf50 jens-daniel-mueller 2020-04-19 created pub figures for time series
html 87658c3 jens-daniel-mueller 2020-04-14 Build site.
Rmd 5c96a65 jens-daniel-mueller 2020-04-14 temperature penetration depth
html 624835e jens-daniel-mueller 2020-04-02 Build site.
Rmd a7ac65d jens-daniel-mueller 2020-04-02 BloomSail data 1-5m and sd in time series plots
html 26cf733 jens-daniel-mueller 2020-04-02 Build site.
Rmd 57b77af jens-daniel-mueller 2020-04-02 corrected Finnmaid lat borders and plotted fm track in map
html a6c4c22 jens-daniel-mueller 2020-03-30 Build site.
html 80c78b3 jens-daniel-mueller 2020-03-30 Build site.
html 5f8ca30 jens-daniel-mueller 2020-03-20 Build site.
Rmd 1ebd01a jens-daniel-mueller 2020-03-20 reorganitzed filenames and navbar

library(tidyverse)
library(data.table)
library(lubridate)
library(DataExplorer)
library(leaflet)
library(readxl)
library(gsubfn)

1 Scope of this script

For each data source:

  • read raw data files
  • harmonize column names
  • clean obvious erroneous data
  • merged into one file
  • write merged file

2 Sea-Bird SBE 16 sensor data (ts)

CTD sensor data including recordings from the analog output of auxiliary pH, O2, Chla and pCO2 sensors were recorded with a measurement frequency of 15 sec. (In addition, pCO2 data were also internally recorded on the Contros HydroC instrument with higher temporal resolution and will later be used for further analysis after merging with CTD data.)

2.1 Read regular profiles and transects

files <-
  list.files(path = "data/input/TinaV/Sensor/Profiles_Transects/", pattern = "[.]cnv$")

for (file in files) {
  start_date <-
    data.table(read.delim(
      here::here("data/input/TinaV/Sensor/Profiles_Transects/", file),
      sep = "#",
      nrows = 160
    ))[[78, 1]]
  start_date <- substr(start_date, 15, 34)
  start_date <- mdy_hms(start_date, tz = "UTC")
  
  temp <-
    read.delim(
      here::here("data/input/TinaV/Sensor/Profiles_Transects/", file),
      sep = "",
      skip = 160,
      header = FALSE
    )
  temp <- data.table(temp[, c(2, 3, 4, 5, 6, 7, 9, 11, 13)])
  names(temp) <-
    c("date_time",
      "dep",
      "tem",
      "sal",
      "V_pH",
      "pH",
      "Chl",
      "O2",
      "pCO2_analog")
  temp$start_date <- start_date
  temp$date_time <- temp$date_time + temp$start_date
  
  temp$ID <- substr(file, 1, 6)
  temp$type <- substr(file, 8, 8)
  temp$station <- substr(file, 8, 10)
  
  temp$cast <- "up"
  temp[date_time < mean(temp[dep == max(temp$dep)]$date_time)]$cast <-
    "down"
  
  if (exists("dataset")) {
    dataset <- rbind(dataset, temp)
  }
  
  if (!exists("dataset")) {
    dataset <- temp
  }
  
  rm(start_date)
  rm(temp)
}


ts <- dataset
rm(dataset, file, files)

2.2 Read profiles and transects around Östergarnsholm

files <-
  list.files(path = "data/input/TinaV/Sensor/Ostergarnsholm/", pattern = "[.]cnv$")

for (file in files) {
  start_date <-
    data.table(read.delim(
      here::here("data/input/TinaV/Sensor/Ostergarnsholm/", file),
      sep = "#",
      nrows = 160
    ))[[78, 1]]
  start_date <- substr(start_date, 15, 34)
  start_date <- mdy_hms(start_date, tz = "UTC")
  
  temp <-
    read.delim(
      here::here("data/input/TinaV/Sensor/Ostergarnsholm/", file),
      sep = "",
      skip = 160,
      header = FALSE
    )
  temp <- data.table(temp[, c(2, 3, 4, 5, 6, 7, 9, 11, 13)])
  names(temp) <-
    c("date_time",
      "dep",
      "tem",
      "sal",
      "V_pH",
      "pH",
      "Chl",
      "O2",
      "pCO2_analog")
  temp$start_date <- start_date
  temp$date_time <- temp$date_time + temp$start_date
  
  temp$ID <- substr(file, 1, 6)
  temp$type <- substr(file, 8, 8)
  temp$station <- substr(file, 11, 12)
  
  temp$cast <- "up"
  temp[date_time < mean(temp[dep == max(temp$dep)]$date_time)]$cast <-
    "down"
  
  if (exists("dataset")) {
    dataset <- rbind(dataset, temp)
  }
  
  if (!exists("dataset")) {
    dataset <- temp
  }
  
  rm(start_date)
  rm(temp)
}


ts_OGB <- dataset
rm(dataset, file, files)

ts_OGB <- ts_OGB %>%
  mutate(
    type = if_else(station == "bo", "P", "T"),
    station = if_else(station == "bo", "P14", station),
    station = if_else(station == "in", "T14", station),
    station = if_else(station == "ou", "T15", station)
  )
ts <- bind_rows(ts, ts_OGB) %>%
  arrange(date_time)

rm(ts_OGB)

2.3 EDA raw data

source("code/eda.R")
eda(ts, "ts-raw")
rm(eda)

The output of an automated Exploratory Data Analysis (EDA) performed on the raw data with the package DataExplorer can be accessed here:

Link to EDA report of CTD raw data

2.4 Clean data set

Sensor recordings were cleaned from obviously erroneous readings, by setting suspicious values to NA.

# running the commented code for plotting
# before and after the cleaning steps
# allows to visualize the removal of errorneous readings

ts <- data.table(ts)

# Profiling data

# temperature

# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(tem, dep, col=station, linetype = cast))+
#   geom_line()+
#   scale_y_reverse()+
#   geom_vline(xintercept = c(10, 20))+
#   facet_wrap(~ID)

ts[ID == "180723" & station == "P07" & dep < 2 & cast == "up"]$tem <- NA

# salinity

# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(sal, dep, col=station, linetype = cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)

ts[sal < 6]$sal <- NA

# pH

# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(pH, dep, col=station, linetype=cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)
# 
# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(V_pH, dep, col=station, linetype=cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)

ts[pH < 7.5]$V_pH <- NA
ts[pH < 7.5]$pH <- NA

ts[ID == "180709" & station == "P03" & dep < 5 & cast == "down"]$pH <- NA
ts[ID == "180709" & station == "P05" & dep < 10 & cast == "down"]$pH <- NA
ts[ID == "180718" & station == "P10" & dep < 3 & cast == "down"]$pH <- NA
ts[ID == "180815" & station == "P03" & dep < 2 & cast == "down"]$pH <- NA
ts[ID == "180820" & station == "P11" & dep < 15 & cast == "down"]$pH <- NA

ts[ID == "180709" & station == "P03" & dep < 5 & cast == "down"]$V_pH <- NA
ts[ID == "180709" & station == "P05" & dep < 10 & cast == "down"]$V_pH <- NA
ts[ID == "180718" & station == "P10" & dep < 3 & cast == "down"]$V_pH <- NA
ts[ID == "180815" & station == "P03" & dep < 2 & cast == "down"]$V_pH <- NA
ts[ID == "180820" & station == "P11" & dep < 15 & cast == "down"]$V_pH <- NA


# pCO2

# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(pCO2, dep, col=station, linetype = cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)

ts[ID == "180616"]$pCO2_analog <- NA

# O2

# ts %>% 
#   filter(type == "P") %>% 
#   ggplot(aes(O2, dep, col=station, linetype = cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)

# Chlorophyll

# ts %>%
#   filter(type == "P") %>%
#   ggplot(aes(Chl, dep, col=station, linetype = cast))+
#   geom_path()+
#   scale_y_reverse()+
#   facet_wrap(~ID)

ts[Chl > 100]$Chl <- NA


# Surface transect data

# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, dep, col=station))+
#   geom_point()+
#   scale_y_reverse()+
#   facet_wrap(~ID, scales = "free_x")
# 
# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, tem, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")
# 
# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, sal, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")
# 
# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, pCO2, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")
# 
# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, pH, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")
# 
# ts %>%
#   filter(type == "T") %>%
#   ggplot(aes(date, Chl, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")

ts[type == "T" & Chl > 10]$Chl <- NA


# ts %>% 
#   filter(type == "T") %>% 
#   ggplot(aes(date, O2, col=station))+
#   geom_point()+
#   facet_wrap(~ID, scales = "free_x")

2.5 Write summary file

Relevant columns were selected and renamed, only observations from regular stations (P01-P13) and transects (T01-T13) were selected and summarized data were written to file.

ts <- ts %>% 
  select(date_time,
         ID,
         type,
         station,
         dep,
         sal,
         tem,
         pCO2_analog)

ts %>% 
  write_csv(here::here("data/intermediate/_summarized_data_files", "ts.csv"))

2.6 EDA clean data

source("code/eda.R")
eda(ts, "ts_clean")

rm(eda)

The output of an automated Exploratory Data Analysis (EDA) performed on the cleaned data with the package DataExplorer can be accessed here:

Link to EDA report of CTD clean data

2.7 Overview plots

ts %>%
  arrange(date_time) %>%
  filter(type == "P",!(station %in% c("PX1", "PX2"))) %>%
  ggplot(aes(tem, dep, col = ymd(ID), group = ID)) +
  geom_path() +
  scale_y_reverse() +
  scale_color_viridis_c(trans = "date", name = "") +
  facet_wrap( ~ station)
Temperature profiles by stations. Color refers to the starting date of each cruise.

Temperature profiles by stations. Color refers to the starting date of each cruise.

ts %>%
  arrange(date_time) %>%
  filter(type == "P",!(station %in% c("PX1", "PX2"))) %>%
  ggplot(aes(pCO2_analog, dep, col = ymd(ID), group = ID)) +
  geom_path() +
  scale_y_reverse() +
  scale_color_viridis_c(trans = "date", name = "") +
  facet_wrap( ~ station)
pCO~2~ (analog signal) profiles by stations. Color refers to the starting date of each cruise.

pCO2 (analog signal) profiles by stations. Color refers to the starting date of each cruise.

3 HydroC CO2 data (th)

3.1 Read data

Originally, HydroC pCO2 data were provided by KM Contros after applying a drift correction to the raw data, which was based on pre- and post-deployment calibration results. Those preliminary data are read-in here and is referred to as V1. However, some data recorded during testing and configuration of the sensor were later on removed, and the post-processing was repeated based on a cleaned data set. This revised post-processed file is referred to as V2 and used in the merging script.

# Read Contros corrected data file, based on all recordings
th <-
  read_csv2(here::here("data/input/TinaV/Sensor/HydroC-pCO2/corrected_Contros",
                       "parameter&pCO2s(method 43).txt"),
            col_names = c("date_time", "Zero", "Flush", "p_NDIR",
                          "p_in", "T_control", "T_gas", "%rH_gas",
                          "Signal_raw", "Signal_ref", "T_sensor",
                          "pCO2_corr", "Runtime", "nr.ave")) %>% 
  mutate(date_time = dmy_hms(date_time),
         Flush = as.factor(as.character(Flush)),
         Zero = as.factor(as.character(Zero)))

3.2 Deployment identification and subsetting

Individual deployments (periods of observations with less than 30 sec between recordings) were identified and relevant deployment periods were selected. This procedure removes some data recorded during sensor testing and set-up.

# identify individual deployments
th <- th %>%
  arrange(date_time) %>%
  mutate(deployment = cumsum(c(TRUE, diff(date_time) >= 30)))

# write pre-cleaning file for later comparison
th %>%
  select(date_time, pCO2_corr, deployment) %>%
  write_csv(here::here(
    "data/intermediate/_summarized_data_files",
    "th_pre_cleaning.csv"
  ))

# filter relevant deployments
th <- th %>%
  filter(deployment %in% c(2, 6, 9, 14, 17, 21, 23, 27, 31, 33, 34, 35, 37))

3.3 Removal of duplicated time stamps

A low number of the recorded HydroC data revealed the exact same time stamp. This was corrected either by filling a corresponding gap before or after the duplicate, or by removing one of the duplicated rows if no such gap existed.

# add counter for date_time observations
th <- th %>%
  add_count(date_time)

# find triplicated time stamp and select only first observation, and merge

th_no_triple <- th %>%
  filter(n <= 2)

th_triple_clean <- th %>%
  filter(n > 2) %>%
  slice(1)

th <- full_join(th_no_triple, th_triple_clean)

rm(list = setdiff(ls(), c("th", "parameters")))


# find duplicated time stamps and shift first by one second backward, and merge

# th %>%
#   distinct(date_time)

th <- th %>%
  select(-n) %>%
  add_count(date_time)

# unique(th$n)

th_no_duplicated <- th %>%
  filter(n == 1)

th_duplicated <- th %>%
  filter(n == 2)

th_duplicated_first <- th_duplicated %>%
  group_by(date_time) %>%
  slice(1) %>%
  ungroup() %>%
  mutate(date_time = date_time - 1)

th_duplicated_second <- th_duplicated %>%
  group_by(date_time) %>%
  slice(2) %>%
  ungroup()

th_duplicated_clean <-
  full_join(th_duplicated_first, th_duplicated_second) %>%
  arrange(date_time)

th <- full_join(th_no_duplicated, th_duplicated_clean)

# th %>%
#   distinct(date_time)

rm(list = setdiff(ls(), c("th", "parameters")))



# find duplicated time stamps and shift first by two seconds forward, and merge

# th %>%
#   distinct(date_time)

th <- th %>%
  select(-n) %>%
  add_count(date_time)

# unique(th$n)

th_no_duplicated <- th %>%
  filter(n == 1)

th_duplicated <- th %>%
  filter(n == 2)

th_duplicated_first <- th_duplicated %>%
  group_by(date_time) %>%
  slice(1) %>%
  ungroup() %>%
  mutate(date_time = date_time + 2)

th_duplicated_second <- th_duplicated %>%
  group_by(date_time) %>%
  slice(2) %>%
  ungroup()

th_duplicated_clean <-
  full_join(th_duplicated_first, th_duplicated_second) %>%
  arrange(date_time)

th <- full_join(th_no_duplicated, th_duplicated_clean)

# th %>%
#   distinct(date_time)

rm(list = setdiff(ls(), c("th", "parameters")))

# remaining duplicates are observations where other observations with a +/- 1 sec timestamp exist
# for those cases, only the first duplicated observation is selected (similar to triplicate treatment)


# th %>%
#   distinct(date_time)

th <- th %>%
  select(-n) %>%
  add_count(date_time)

# unique(th$n)


th_still_no_duplicated <- th %>%
  filter(n == 1)

th_still_duplicated_first <- th %>%
  filter(n == 2) %>%
  group_by(date_time) %>%
  slice(1)

th <- full_join(th_still_no_duplicated, th_still_duplicated_first)

# th %>%
#   distinct(date_time)

rm(list = setdiff(ls(), c("th", "parameters")))

th <- th %>%
  select(-n)

3.4 Flush and Zeroing identification

Flush and zeroing periods of the sensor are identified and assigned with unique IDs.

# Zeroing ID labeling

th <- th %>%
  arrange(date_time) %>%
  group_by(Zero) %>%
  mutate(Zero_counter = as.factor(cumsum(c(
    TRUE, diff(date_time) >= 30
  )))) %>%
  ungroup()

# Flush: Identification

th <- th %>%
  mutate(Flush = 0) %>%
  group_by(Zero, Zero_counter) %>%
  mutate(
    start = min(date_time),
    duration = date_time - start,
    Flush = if_else(Zero == 0 &
                      duration < parameters$HC_flush_duration, "1", "0")
  ) %>%
  ungroup()


#  Flush: Identify equilibration and internal gas mixing periods

th <- th %>%
  mutate(mixing = if_else(
    duration < parameters$HC_mixing_duration,
    "mixing",
    "equilibration"
  ))

3.5 Deployment plots

pdf(file=here::here("output/Plots/read_in",
    "th_deployments.pdf"), onefile = TRUE, width = 7, height = 4)

for (i in unique(th$deployment)) {
  
  #i <- unique(th$deployment)[3]
  
  sub <-  th %>%
      filter(deployment == i)
  start_date <- min(sub$date_time)
  
  print(
    sub %>% 
      ggplot(aes(date_time, pCO2_corr, col=Zero_counter))+
      geom_line()+
      labs(title = paste("Deployment: ",i, "| Start time: ", start_date))
)
  
}

dev.off()

rm(sub, start_date, i)

A pdf with pCO2 timeseries plots of all individual deployments can be found here:

Link to pCO2 timeseries plots

source("code/eda.R")
eda(th, "th")
rm(eda)

The output of an automated Exploratory Data Analysis (EDA) performed with the package DataExplorer can be accessed here:

Link to EDA report of HydroC pCO2 data data

3.6 Write summary file

Summarized clean pCO2 data were written to file.

th %>%
  select(date_time,
         Zero,
         Flush,
         pCO2_corr,
         deployment,
         Zero_counter,
         duration,
         mixing) %>%
  write_csv(here::here("data/intermediate/_summarized_data_files",
                       "th.csv"))

rm(th)

4 Bottle data CO2 (tb)

Discrete samples were collected with a Niskin bottle and analyzed for CT and AT at IOW’s CO2 lab.

4.1 Read data

# Read CO2 system bottle data
tb <-
  read_csv(
    here::here(
      "data/input/TinaV/Bottle/Tracegases",
      "BloomSail_bottle_CO2_all.csv"
    ),
    col_types = list("c", "c", "n", "n", "n", "n", "n")
  )

# select and rename relevant columns
tb <- tb %>%
  select(
    ID = transect.ID,
    station = label,
    dep = Dep,
    sal = Sal,
    CT,
    AT
  )

4.2 Write summary file

tb %>% write_csv(here::here("data/intermediate/_summarized_data_files",
                            "tb.csv"))
rm(tb)

5 Bottle data plankton (tp)

Discrete samples were collected with a Niskin bottle and analysed for phytoplankton composition and biomass at IOW’s phytoplankton lab (Norbert Wasmund).

5.1 Read data

tp <- read_csv(
  here::here(
    "data/input/TinaV/Bottle/Phytoplankton",
    "181205_BloomSail_Plankton_counts.csv"
  )
)

# delete colomns that contain counts, not calculated biomass
tp <- tp[, -seq(4, 21, 1)]

# assign new column names
# for species: 
# nr = size class, 
# HV = Heterocyst per Volume, 
# Hl = Heterocyst per length, 
# t = total

names(tp) <-
  c(
    "date",
    "station",
    "dep",
    "Aphanizomenon.1",
    "Aphanizomenon.2",
    "Aphanizomenon.3",
    "Aphanizomenon.t",
    "Aphanizomenon.HV",
    "Aphanizomenon.Hl",
    "Dolichospermum.1",
    "Dolichospermum.2",
    "Dolichospermum.3",
    "Dolichospermum.4",
    "Dolichospermum.t",
    "Dolichospermum.HV",
    "Dolichospermum.Hl",
    "Nodularia.1",
    "Nodularia.2",
    "Nodularia.3",
    "Nodularia.t",
    "Nodularia.HV",
    "Nodularia.Hl",
    "Nodulariadead.1",
    "Nodulariadead.2",
    "Nodulariadead.3",
    "Nodulariadead.t",
    "total.t"
  )


# change format of data table and separate into 2 columns for species and class

tp <-
  gather(tp, para, value, Aphanizomenon.1:total.t, factor_key = TRUE)

tp <- separate(tp, col = para, into = c("Species", "class"))

# change class of columns
tp <- tp %>%
  mutate(ID = date,
         date = ymd(date))

5.2 Write summary file

tp %>% write_csv(here::here("data/intermediate/_summarized_data_files",
                            "tp.csv"))
rm(tp)

6 GPS track (tt)

GPS track data were recorded with a Samsung Galaxy tablet.

6.1 Read data

files <-
  list.files(path = "data/input/TinaV/Track/GPS_Logger_Track/", pattern = "[.]txt$")

for (file in files) {
  # if the merged dataset does exist, append to it
  if (exists("dataset")) {
    temp <-
      data.table(read.delim(
        here::here("data/input/TinaV/Track/GPS_Logger_Track", file),
        sep = ","
      )[, c(2, 3, 4)])
    names(temp) <- c("date_time", "lat", "lon")
    temp$date_time <- ymd_hms(temp$date, tz = "UTC")
    
    dataset <- rbind(dataset, temp)
    rm(temp)
  }
  
  # if the merged dataset doesn't exist, create it
  if (!exists("dataset")) {
    dataset <-
      data.table(read.delim(
        here::here("data/input/TinaV/Track/GPS_Logger_Track", file),
        sep = ","
      )[, c(2, 3, 4)])
    names(dataset) <- c("date_time", "lat", "lon")
    dataset$date_time <- ymd_hms(dataset$date_time, tz = "UTC")
    
  }
}

tt <- dataset
rm(dataset, file, files)

6.2 Write summary file

tt %>%
  write_csv(here::here("data/intermediate/_summarized_data_files",
                       "tt.csv"))

rm(tt)

7 Atmospheric observations

Atmospheric data were recorded at the ICOS station on Östergarnsholm.

7.1 Read data

og <-
  read_delim(
    here::here(
      "data/input/Ostergarnsholm/Tower",
      "Oes_Jens_atm_water_June_to_August_2018.csv"
    ),
    delim = ";"
  )

og <- og %>%
  mutate(date_time = ymd_hms(paste(
    paste(year, month, day, sep = "/"),
    paste(hour, min, sec, sep = ":")
  ))) %>%
  select(
    "date_time",
    "CO2 12m [ppm]",
    "w_c [ppm m/s]",
    "WS 12m [m/s]",
    "WD 12m [degrees]",
    "T 12m [degrees C]",
    "RIS [W/m^2]"
  )

# conversion from GMT+1 to UTC
og <- og %>%
  mutate(date_time = date_time - 60 ^ 2)

og <- og %>%
  select(date_time, pCO2_atm = "CO2 12m [ppm]", wind = "WS 12m [m/s]")

7.2 Write summary file

og %>%
  write_csv(here::here("data/intermediate/_summarized_data_files",
                       "og.csv"))

rm(og)

8 SOOP Finnmaid

Here, we read in pCO2 and SST data recorded on SOOP Finnmaid in June-August 2018.

8.1 Read data

# LI-COR data

files <-
  list.files(path = "data/input/Finnmaid_2018", pattern = "[.]xls$")

for (file in files) {
  temp <- read_excel(here::here("data/input/Finnmaid_2018", file))
  temp <- temp[c(1, 2, 3, 12, 7, 4, 15, 8, 5, 17)]
  names(temp) <-
    c("date_time",
      "lon",
      "lat",
      "pCO2",
      "sal",
      "tem",
      "cO2",
      "patm",
      "Teq",
      "xCO2")
  temp <- temp[-c(1), ]
  temp$date_time <-
    as.POSIXct(as.numeric(temp$date_time) * 60 * 60 * 24,
               origin = "1899-12-30",
               tz = "GMT")
  temp$lon <- as.numeric(as.character(temp$lon))
  temp$lat <- as.numeric(as.character(temp$lat))
  temp$pCO2 <- as.numeric(as.character(temp$pCO2))
  temp$sal <- as.numeric(as.character(temp$sal))
  temp$tem <- as.numeric(as.character(temp$tem))
  temp$cO2 <- as.numeric(as.character(temp$cO2))
  temp$patm <- as.numeric(as.character(temp$patm))
  temp$Teq <- as.numeric(as.character(temp$Teq))
  temp$xCO2 <- as.numeric(as.character(temp$xCO2))
  temp <- data.table(temp)
  
  temp$route <-
    strapplyc(as.character(file), ".*(.).xls*", simplify = TRUE)
  temp$ID <- substr(as.character(file), 3, 10)
  
  if (exists("dataset")) {
    dataset <- rbind(dataset, temp)
  } else{
    dataset <- temp
  }
  
}


rm(temp, files, file)
dataset <- dataset[pCO2 != 0]


# Los Gatos Research (LGR) data
# Please note that the LGR data were corrected manually before
# The correction procedure is outlined in the Appendix of the ms

files <-
  list.files(path = "data/input/Finnmaid_2018/LGR", pattern = "[.]xls$")

for (file in files) {
  temp <- read_excel(here::here("data/input/Finnmaid_2018/LGR", file))
  temp <- temp[c(2, 3, 4, 8, 6, 5, 14, 7, 15, 9)]
  names(temp) <-
    c("date_time",
      "lon",
      "lat",
      "pCO2",
      "sal",
      "tem",
      "cO2",
      "patm",
      "Teq",
      "xCO2")
  temp <- temp[-c(1), ]
  temp$date_time <- dmy_hms(temp$date_time)
  temp <- data.table(temp)
  
  temp$route <- substr(as.character(file), 12, 12)
  temp$ID <- substr(as.character(file), 3, 10)
  
  if (exists("dataset.LGR")) {
    dataset.LGR <- rbind(dataset.LGR, temp)
  } else{
    dataset.LGR <- temp
  }
  
}

rm(temp, files, file)
# This code can be used to convert O2 units
# but is not applied, because O2 data are not used in this study

source(here::here("code", "O2stoO2c.R"))

dataset.LGR <- dataset.LGR %>%
  filter() %>%
  mutate(cO2 = O2stoO2c(
    O2sat = cO2,
    T = tem,
    S = sal,
    P = 3 / 10,
    p_atm = 1013.5
  ))

rm(O2stoO2c, pH2Osat, sca_T, Scorr, TCorr, R, Vm)
dataset$sensor <- "LICOR"
dataset.LGR$sensor <- "LosGatos"

fm <- bind_rows(dataset, dataset.LGR)

rm(dataset, dataset.LGR)

8.2 Write summary file

fm %>% 
  write_csv(here::here("data/intermediate/_summarized_data_files",
                       "fm.csv"))
rm(fm)

9 Interactive map

fm <-
  read_csv(here::here("data/intermediate/_summarized_data_files",
                      "fm.csv"))

fm_sub <- fm %>%
  arrange(date_time) %>%
  slice(which(row_number() %% 20 == 1))

tt <-
  read_csv(here::here("data/intermediate/_summarized_data_files", "tt.csv"))

tt_sub <- tt %>%
  slice(which(row_number() %% 20 == 1))

rm(tt, fm)

leaflet() %>%
  setView(lng = 20, lat = 57.3, zoom = 8) %>%
  addLayersControl(
    baseGroups = c("Ocean Basemap",
                   "Satellite"),
    overlayGroups = c("BloomSail", "Finnmaid"),
    options = layersControlOptions(collapsed = FALSE),
    position = 'topright'
  ) %>%
  addProviderTiles("Esri.WorldImagery", group = "Satellite") %>%
  addProviderTiles(providers$Esri.OceanBasemap, group = "Ocean Basemap") %>%
  addScaleBar(position = 'topright') %>%
  addMeasure(
    primaryLengthUnit = "kilometers",
    secondaryLengthUnit = 'miles',
    primaryAreaUnit = "sqmeters",
    secondaryAreaUnit = "acres",
    position = 'topleft'
  ) %>%
  addCircles(data = fm_sub,
             ~ lon,
             ~ lat,
             color = "white",
             group = "Finnmaid") %>%
  addPolylines(data = tt_sub,
               ~ lon,
               ~ lat,
               color = "red",
               group = "BloomSail")

Overview map of SOOP Finnmaid and SV Tina V tracks. Please note that only a subet of the recorded track data is plotted.

rm(fm_sub, tt_sub)

sessionInfo()
R version 4.0.3 (2020-10-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)

Matrix products: default

locale:
[1] LC_COLLATE=English_Germany.1252  LC_CTYPE=English_Germany.1252   
[3] LC_MONETARY=English_Germany.1252 LC_NUMERIC=C                    
[5] LC_TIME=English_Germany.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] gsubfn_0.7         proto_1.0.0        readxl_1.3.1       leaflet_2.0.3     
 [5] DataExplorer_0.8.2 lubridate_1.7.9.2  data.table_1.13.6  forcats_0.5.0     
 [9] stringr_1.4.0      dplyr_1.0.2        purrr_0.3.4        readr_1.4.0       
[13] tidyr_1.1.2        tibble_3.0.4       ggplot2_3.3.3      tidyverse_1.3.0   
[17] workflowr_1.6.2   

loaded via a namespace (and not attached):
 [1] httr_1.4.2              jsonlite_1.7.2          viridisLite_0.3.0      
 [4] here_1.0.1              modelr_0.1.8            assertthat_0.2.1       
 [7] highr_0.8               cellranger_1.1.0        yaml_2.2.1             
[10] pillar_1.4.7            backports_1.2.1         glue_1.4.2             
[13] digest_0.6.27           promises_1.1.1          rvest_0.3.6            
[16] leaflet.providers_1.9.0 colorspace_2.0-0        htmltools_0.5.0        
[19] httpuv_1.5.4            pkgconfig_2.0.3         broom_0.7.3            
[22] haven_2.3.1             scales_1.1.1            whisker_0.4            
[25] later_1.1.0.1           git2r_0.27.1            generics_0.1.0         
[28] farver_2.0.3            ellipsis_0.3.1          withr_2.3.0            
[31] cli_2.2.0               magrittr_2.0.1          crayon_1.3.4           
[34] evaluate_0.14           ps_1.5.0                fs_1.5.0               
[37] fansi_0.4.1             xml2_1.3.2              tools_4.0.3            
[40] hms_0.5.3               lifecycle_0.2.0         munsell_0.5.0          
[43] reprex_0.3.0            networkD3_0.4           compiler_4.0.3         
[46] rlang_0.4.10            grid_4.0.3              rstudioapi_0.13        
[49] htmlwidgets_1.5.3       crosstalk_1.1.0.1       igraph_1.2.6           
[52] tcltk_4.0.3             labeling_0.4.2          rmarkdown_2.6          
[55] gtable_0.3.0            DBI_1.1.0               R6_2.5.0               
[58] gridExtra_2.3           knitr_1.30              rprojroot_2.0.2        
[61] stringi_1.5.3           parallel_4.0.3          Rcpp_1.0.5             
[64] vctrs_0.3.6             dbplyr_2.0.0            tidyselect_1.1.0       
[67] xfun_0.19