_targets.R

library(targets)
library(tarchetypes)
source("code/packages.R")
source("code/functions.R")

# Workflow Plans
#
# ## Data sources
#
# The raw data are in `.xlsx` files. Some of the data are pre-processed
# (mean and sd of count and length of gemmae, 30 min averages of PPFD).
#
# The files include some additional analsyes and other data that won't
# be used here.


tar_plan(

  # Load and clean data ----

  # ### Microclimate data
  #
  # Microclimate variables include PPFD (photon flux density, in μmol of
  # light per sq m per sec), rel. humidity (%), and temperature (°C)
  # measured once every 30 min.
  #
  # PPFD is 30 min-averages of values taken every 4 minutes
  # (raw 4 minute values not included).
  #
  # There are two sites, Okutama (site of the independent gametophyte
  # colony) and Uratakao (site of the sporophyte population). Additional
  # sites were also measured, but not included in this analysis
  # because of too much missing data due to mechanical failures.
  #
  # For Okutama, the raw data for temperature, rel. hum,
  # and light (PPFD) are in different columns
  # in the `xlsx` file, so read in each variable separately.

  okutama_temp_raw = read_excel(
    "data_raw/datalogger_raw.xlsm",
    sheet = 1,
    range = "A2:B28683",
    col_names = c("date_time", "temp")),

  okutama_rh_raw = read_excel(
    "data_raw/datalogger_raw.xlsm",
    sheet = 1,
    range = "C2:D28683",
    col_names = c("date_time", "rh")),

  okutama_par_raw = read_excel(
    "data_raw/datalogger_raw.xlsm",
    sheet = 1,
    range = "F2:G28204",
    col_names = c("date_time", "par")),

  # Combine raw okutama microclimate variables into a single tibble.
  okutama_microclimate_raw =
    full_join(okutama_temp_raw, okutama_rh_raw) %>%
    full_join(okutama_par_raw),

  # Uratakao raw data have a single column for time of the three microclimate
  # variables, so these can be read in all at once.
  takao_microclimate_raw = read_excel(
    "data_raw/datalogger_raw.xlsm",
    sheet = 2,
    range = "D2:G21459",
    col_names = c("date_time", "par", "temp", "rh")),

  # ### Cover
  #
  # Cover data (area of gametophytes in sq. cm) was
  # measured once per month for
  # four 10 x 10 cm plots at the Okutama site.
  cover_raw = read_excel(
    "data_raw/pleurosoriopsis_data_figs.xlsx",
    sheet = 1,
    skip = 1,
    col_names = c("date", "q_1", "q_2", "q_3", "q_4")
    ),

  # ### Gemmae
  #
  # Gemmae count and length were measured in 10 individuals per
  # monthly census. The data provided are mean values with standard deviation
  # already cacluated.

  gemmae_count_raw = read_xlsx(
    "data_raw/pleurosoriopsis_data_figs.xlsx",
    sheet = 3,
    range = "A2:C62"),

  gemmae_length_raw = read_xlsx(
    "data_raw/pleurosoriopsis_data_figs.xlsx",
    sheet = 3,
    range = "E2:G62",
    col_types = c("text", "numeric", "numeric")),

  # ### Gametophyte size data
  #
  # Size data include length and width of invididuals,
  # split into different sheets based
  # on presence or absence of gametangia.
  #
  # Read in size data for individuals without gametangia
  size_asexual_raw = read_excel(
    "data_raw/okutama_gameto_size.xlsx",
    sheet = 1,
    range = "B4:D94"),

  # Read in size data for individuals with gametangia.
  # (No dates here)
  size_with_archegonia = read_excel(
    "data_raw/okutama_gameto_size.xlsx",
    sheet = 2,
    range = "B4:D36") %>%
    rename(id = `番号`, length = `長さ（㎜）`, width = `　幅（㎜）`) %>%
    mutate(sex = "female"),

  size_with_antheridia = read_excel(
    "data_raw/okutama_gameto_size.xlsx",
    sheet = 2,
    range = "O5:Q9")  %>%
    rename(id = `番号`, length = `長さ（㎜）　`, width = `幅（㎜）`) %>%
    mutate(sex = "male"),

  size_sexual = bind_rows(
    size_with_archegonia,
    size_with_antheridia
  ),

  # ### Clean data
  #
  # Select microclimate variables to use in downstream analyses
  selected_vars = c("rh_min", "par_total", "temp_mean"),

  # Clean microclimate data
  okutama_microclimate = clean_microclimate(
    okutama_microclimate_raw),

  takao_microclimate = clean_microclimate(
    takao_microclimate_raw),

  # Clean gametophyte cover data
  cover = clean_cover_data(cover_raw),

  # Clean gemmae count and length data
  gemmae_data = clean_gemmae_data(
    gemmae_count_raw = gemmae_count_raw,
    gemmae_length_raw = gemmae_length_raw
  ),

  # Clean asexual size data
  size_asexual_clean = clean_asexual_size(
    size_asexual_raw
  ),

  # Combine asexual and sexual size data
  gameto_size = bind_rows(size_asexual_clean, size_sexual),

  # Combine morphology into single tibble
  combined_morph = combine_morph_data (
    cover_data = cover,
    gemmae_data = gemmae_data
  ),

  # Combine morphology and Okutama climate into single tibble
  combined_monthly_morph = combine_okutama_data (
    okutama_microclimate = okutama_microclimate,
    combined_morph = combined_morph,
    selected_vars = selected_vars),

# Calculate daily summary variables ----
#
# In order to compare data across sites, we need to use daily values since
# the measurement times are slightly different between sites.
#
# There are many different ways to calculate
# daily microclimate values (mean, max, min, sd, of temp, humidity, etc.).
#
# We will use three biologically relevant values: mean temp, min RH,
# and DLI (daily light integral, the sum of all PAR measurements for one day).
#
# https://en.wikipedia.org/wiki/Daily_light_integral
#
# To calculate DLI (moles of light per sq m per day), we need to integrate daily PPFD
# (photon flux density, in μmol of light per sq m per sec) values.
#
# PPFD was measured once every 4 minutes in then converted to a mean
# value once every 30 minutes. The values in the raw data are the 30
# minute averages. To get DLI:
#
# μmol PAR per second * 1800 seconds in 30 min = μmol PAR in 30 min
#
# sum of all μmol PAR in 30 min values in one day / 1,000,000 = total mol PAR for the day (DLI)

  # Combine clean microclimata data from Okutama and Uratakao
  # into a single tibble with rows of daily values.
  daily_microclimate = calculate_daily_microclimate (
    site_1_data = okutama_microclimate,
    site_1_name = "okutama",
    site_2_data = takao_microclimate,
    site_2_name = "uratakao"
  ),

  # Subset to only days that are shared between
  # sites for paired t-test.
  paired_microclimate = make_paired_dataset(
    daily_microclimate
  ),

  # Make nested data frame for looping.
  nested_paired_microclimate = nest_microclimate(
    paired_microclimate,
    selected_vars = selected_vars
  ),

# Analysis ----

  t_test_results = compare_microclimate(
    paired_microclimate = paired_microclimate,
    nested_paired_microclimate = nested_paired_microclimate,
    selected_vars = selected_vars
    ),

  linear_model_results = make_climate_morph_lm(
    selected_vars = selected_vars,
    combined_monthly_morph = combined_monthly_morph
  ),

  # Write out MS
  tar_render(
    ms,
    "ms.Rmd"
  )

)