Last updated: 2023-11-15

Checks: 7 0

Knit directory: bgc_argo_r_argodata/

This reproducible R Markdown analysis was created with workflowr (version 1.7.0). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

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.

Environment: empty

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.

Seed: set.seed(20211008)

The command set.seed(20211008) 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.

Session information: recorded

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

Cache: none

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

File paths: relative

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

Repository version: 3eba518

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 3eba518. 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:    output/

Unstaged changes:
    Modified:   code/start_background_job.R

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/temp_core_cluster_analysis.Rmd) and HTML (docs/temp_core_cluster_analysis.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
Rmd 3eba518 ds2n19 2023-11-15 Introduction of vertical alignment and cluster analysis to github website.

Tasks

This markdown file carries out cluster analysis on previously created core temperature anomaly profiles.

The cluster_analysis_determine_k chunk is used to give an indication of an appropriate number of clusers and this can then set the option opt_num_clusters. Chunk cluster_analysis_cluster_details carried out the cluster analysis and the results are used in subsequent figures.

Set directories

location of pre-prepared data

Set options

Define options that are used to determine what analysis is done

# Options

# opt_analysis_type
# opt_analysis_type = 1 do analysis to determine number of clusters to use
# opt_analysis_type = 2 do cluster analysis and further analysis on the identified clusters clusters
opt_analysis_type <- 2

# opt_num_clusters
# How many clusters are used in teh cluster analysis
opt_num_clusters <- 7

# opt_measure_label, opt_xlim and opt_xbreaks are associated formatting
opt_measure_label <- "temperature anomaly (°C)"
opt_xlim <- c(-4, 4)
opt_xbreaks <- c(-4, -2, 0, 2, 4)

# options relating to cluster analysis
opt_n_start <- 15
opt_max_iterations <- 500
opt_n_clusters <- 14 # Max number of clusters to try when determining optimal number of clusters
theme_set(theme_bw())

map <-
  read_rds(paste(path_emlr_utilities,
                 "map_landmask_WOA18.rds",
                 sep = ""))

Cluster analysis

Preperation

Prepare data for cluster analysis

temp_anomaly_va <- read_rds(file = paste0(path_core_preprocessed, "/temp_anomaly_va.rds"))


# select just 1500m profiles
temp_anomaly_va <- temp_anomaly_va %>% 
  filter(profile_range == 3)

# Simplified table ready to pivot
temp_anomaly_va_id <- temp_anomaly_va %>%
  select(file_id,
         depth,
         anomaly,
         year, 
         month, 
         lat, 
         lon)

# Check duplicate values and remove those profiles
check_duplicates <- temp_anomaly_va_id %>%
  group_by(file_id, depth) %>%
  summarise(count_measures = n()) %>%
  filter(count_measures > 1) %>%
  ungroup()

check_duplicates <- check_duplicates %>%
  select (file_id, count_measures)

temp_anomaly_va_id2 <-
  full_join(temp_anomaly_va_id, check_duplicates)

temp_anomaly_va_id <- temp_anomaly_va_id2 %>%
  filter(is.na(count_measures)) %>%
  select(file_id,
         depth,
         anomaly,
         year,
         month,
         lat,
         lon)

rm(temp_anomaly_va_id2)
                            
# wide table with each depth becoming a column
temp_anomaly_va_wide <- temp_anomaly_va_id %>%
  select(file_id, depth, anomaly) %>%
  pivot_wider(names_from = depth, values_from = anomaly)

# Drop any rows with missing values N/A caused by gaps in climatology data
temp_anomaly_va_wide <- temp_anomaly_va_wide %>% 
  drop_na()

# Table for cluster analysis
points <- temp_anomaly_va_wide %>%
  column_to_rownames(var = "file_id")

Number of clusters

if (opt_analysis_type == 1) {


  # cluster analysis - What k? try between 1 and opt_n_clusters clusters
  kclusts <- 
  tibble(k = 1:opt_n_clusters) %>%
  mutate(
    kclust = map(k, ~kmeans(points, .x, iter.max = opt_max_iterations, nstart = opt_n_start)),
    tidied = map(kclust, tidy),
    glanced = map(kclust, glance),
    augmented = map(kclust, augment, points)
  )
  

  # cluster analysis data
  clusters <-
    kclusts %>%
    unnest(cols = c(tidied))
  
  assignments <-
    kclusts %>%
    unnest(cols = c(augmented))
  
  clusterings <-
    kclusts %>%
    unnest(cols = c(glanced))
  
  # What cluster works best?
  clusterings %>%
  ggplot(aes(k, tot.withinss)) +
    geom_line() +
    geom_point() +
    scale_x_continuous(breaks = c(2, 4, 6, 8, 10, 12, 14))

}

Cluster analysis

if (opt_analysis_type == 2) {

  set.seed(1)
  kclusts <-
    tibble(k = opt_num_clusters) %>%
    mutate(
      kclust = map(k, ~ kmeans(points, .x, iter.max = opt_max_iterations, nstart = opt_n_start)),
      tidied = map(kclust, tidy),
      glanced = map(kclust, glance),
      augmented = map(kclust, augment, points)
    )
  
  profile_id <-
    kclusts %>%
    unnest(cols = c(augmented)) %>%
    select(file_id = .rownames,
           cluster = .cluster) %>% 
    mutate(file_id = as.numeric(file_id),
           cluster = as.character(cluster))
  
  # Add cluster to temp_anomaly_va
  temp_anomaly_cluster <- full_join(temp_anomaly_va, profile_id)
  
  # Plot cluster mean
  temp_anomaly_cluster <- temp_anomaly_cluster %>% 
    filter(!is.na(cluster))
  
  # Plot cluster mean
  anomaly_cluster_mean <- temp_anomaly_cluster %>%
    group_by(cluster, depth) %>%
    summarise(
      count_cluster = n(),
      anomaly_mean = mean(anomaly, na.rm = TRUE),
      anomaly_sd = sd(anomaly, na.rm = TRUE)
    ) %>%
    ungroup()
  
  anomaly_cluster_mean_year <- temp_anomaly_cluster %>%
    group_by(cluster, depth, year) %>%
    summarise(
      count_cluster = n(),
      anomaly_mean = mean(anomaly, na.rm = TRUE),
      anomaly_sd = sd(anomaly, na.rm = TRUE)
    ) %>%
    ungroup()
  
  anomaly_year_mean <- temp_anomaly_cluster %>%
    group_by(cluster, year) %>%
    summarise(
      count_cluster = n(),
      anomaly_mean = mean(anomaly, na.rm = TRUE),
      anomaly_sd = sd(anomaly, na.rm = TRUE)
    ) %>%
    ungroup()
  
  anomaly_year_mean <- anomaly_year_mean %>%
    group_by(year) %>%
    summarise(anomaly_mean = mean(anomaly_mean, na.rm = TRUE)) %>%
    ungroup ()
  
  # create figure
  anomaly_cluster_mean %>%
    ggplot() +
    geom_path(aes(x = anomaly_mean,
                  y = depth)) +
    geom_ribbon(aes(
      xmax = anomaly_mean + anomaly_sd,
      xmin = anomaly_mean - anomaly_sd,
      y = depth
    ),
    alpha = 0.2) +
    geom_vline(xintercept = 0) +
    scale_y_reverse() +
    facet_wrap( ~ cluster) +
    coord_cartesian(xlim = opt_xlim) +
    scale_x_continuous(breaks = opt_xbreaks) +
    labs(
      title = paste0('Overall mean anomaly profiles by cluster'),
      x = opt_measure_label,
      y = 'depth (m)'
    )
  
  anomaly_cluster_mean_year %>%
    mutate(year = as.factor(year)) %>%
    ggplot() +
    geom_path(aes(x = anomaly_mean,
                  y = depth,
                  col = year)) +
    geom_vline(xintercept = 0) +
    scale_y_reverse() +
    facet_wrap( ~ cluster) +
    coord_cartesian(xlim = opt_xlim) +
    scale_x_continuous(breaks = opt_xbreaks) +
    scale_color_viridis_d() +
    labs(
      title = paste0('Overall mean anomaly profiles by cluster'),
      x = opt_measure_label,
      y = 'depth (m)'
    )
  
}

Cluster by year

count of each cluster by year

if (opt_analysis_type == 2) {

  # Determine profile count by cluster and year
  cluster_by_year <- temp_anomaly_cluster %>% 
    count(cluster, year,
          name = "count_cluster")

  year_min <- min(cluster_by_year$year)
  year_max <- max(cluster_by_year$year)
  
  # create figure
  cluster_by_year %>% 
    ggplot(aes(x = year, y = count_cluster, col = cluster, group=cluster))+
    geom_point() +
    geom_line() +
    scale_x_continuous(breaks = seq(year_min, year_max, 2)) +
    scale_color_brewer(palette = 'Dark2')+
    labs(x = 'year', 
         y = 'number of profiles', 
         col = 'cluster',
         title = 'Count of profiles by year and cluster')

}

Cluster by month

count of each cluster by month of year

if (opt_analysis_type == 2) {

  # Determine profile count by cluster and year
  cluster_by_year <- temp_anomaly_cluster %>% 
    count(cluster, month,
          name = "count_cluster")

  # create figure
  cluster_by_year %>% 
    ggplot(aes(x = month, y = count_cluster, col = cluster, group=cluster))+
    geom_point() +
    geom_line() +
    scale_x_continuous(breaks = seq(1, 12, 2)) +
    scale_color_brewer(palette = 'Dark2')+
    labs(x = 'month', 
         y = 'number of profiles', 
         col = 'cluster',
         title = 'Count of profiles by month and cluster')

}

Cluster spatial

location of each cluster on map, spatial analysis

if (opt_analysis_type == 2) {

  # create figure
map +
  geom_tile(data = temp_anomaly_cluster, 
            aes(x = lon, 
                y = lat, 
                fill = cluster))+
  lims(y = c(-85, -30))+
  scale_fill_brewer(palette = 'Dark2')+
  labs(title = 'cluster spatial distribution')

  # create figure
map +
  geom_tile(data = temp_anomaly_cluster,
            aes(x = lon,
                y = lat,
                fill = cluster)) +
  lims(y = c(-85, -30)) +
  scale_fill_brewer(palette = 'Dark2') +
  facet_wrap( ~ cluster, ncol = 2) +
  labs(title = 'cluster spatial distribution')

  # create figure
map +
  geom_tile(data = temp_anomaly_cluster %>% 
              count(lat, lon, cluster),
            aes(x = lon,
                y = lat,
                fill = n)) +
  lims(y = c(-85, -30)) +
  scale_fill_viridis_c(option = "cividis", direction = -1, trans = "log10") +
  facet_wrap( ~ cluster, ncol = 2) +
  labs(title = 'cluster spatial distribution')

}

Cluster spatial year

location of each cluster on map, spatial analysis by year

if (opt_analysis_type == 2) {

  # create figure
  temp_anomaly_cluster %>%
    group_split(year) %>%
    map(
      ~ map +
        geom_tile(data = .x,
                  aes(
                    x = lon,
                    y = lat,
                    fill = cluster
                  )) +
        #lims(y = c(-85, -30))+
        scale_fill_brewer(palette = 'Dark2') +
        facet_wrap( ~ cluster, ncol = 2) +
        labs(title = paste0(
          'cluster spatial distribution ', unique(.x$year)
        ))
    )

}
[[1]]


[[2]]


[[3]]


[[4]]


[[5]]


[[6]]


[[7]]


[[8]]


[[9]]


[[10]]


[[11]]


sessionInfo()
R version 4.2.2 (2022-10-31)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: openSUSE Leap 15.4

Matrix products: default
BLAS:   /usr/local/R-4.2.2/lib64/R/lib/libRblas.so
LAPACK: /usr/local/R-4.2.2/lib64/R/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

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

other attached packages:
 [1] yardstick_1.2.0    workflowsets_1.0.1 workflows_1.1.3    tune_1.1.2        
 [5] rsample_1.2.0      recipes_1.0.8      parsnip_1.1.1      modeldata_1.2.0   
 [9] infer_1.0.5        dials_1.2.0        scales_1.2.1       broom_1.0.5       
[13] tidymodels_1.1.1   ggpmisc_0.5.4-1    ggpp_0.5.5         ggforce_0.4.1     
[17] gsw_1.1-1          gridExtra_2.3      lubridate_1.9.0    timechange_0.1.1  
[21] argodata_0.1.0     forcats_0.5.2      stringr_1.5.0      dplyr_1.1.3       
[25] purrr_1.0.2        readr_2.1.3        tidyr_1.3.0        tibble_3.2.1      
[29] ggplot2_3.4.4      tidyverse_1.3.2   

loaded via a namespace (and not attached):
  [1] googledrive_2.0.0   colorspace_2.0-3    ellipsis_0.3.2     
  [4] class_7.3-20        rprojroot_2.0.3     fs_1.5.2           
  [7] rstudioapi_0.15.0   furrr_0.3.1         listenv_0.8.0      
 [10] farver_2.1.1        MatrixModels_0.5-1  prodlim_2019.11.13 
 [13] fansi_1.0.3         xml2_1.3.3          codetools_0.2-18   
 [16] splines_4.2.2       cachem_1.0.6        knitr_1.41         
 [19] polyclip_1.10-4     polynom_1.4-1       jsonlite_1.8.3     
 [22] workflowr_1.7.0     dbplyr_2.2.1        compiler_4.2.2     
 [25] httr_1.4.4          backports_1.4.1     assertthat_0.2.1   
 [28] Matrix_1.5-3        fastmap_1.1.0       gargle_1.2.1       
 [31] cli_3.6.1           later_1.3.0         tweenr_2.0.2       
 [34] htmltools_0.5.3     quantreg_5.94       tools_4.2.2        
 [37] gtable_0.3.1        glue_1.6.2          Rcpp_1.0.10        
 [40] cellranger_1.1.0    jquerylib_0.1.4     RNetCDF_2.6-1      
 [43] DiceDesign_1.9      vctrs_0.6.4         iterators_1.0.14   
 [46] timeDate_4021.106   xfun_0.35           gower_1.0.0        
 [49] globals_0.16.2      rvest_1.0.3         lifecycle_1.0.3    
 [52] googlesheets4_1.0.1 future_1.29.0       MASS_7.3-58.1      
 [55] ipred_0.9-13        hms_1.1.2           promises_1.2.0.1   
 [58] parallel_4.2.2      SparseM_1.81        RColorBrewer_1.1-3 
 [61] yaml_2.3.6          sass_0.4.4          rpart_4.1.19       
 [64] stringi_1.7.8       highr_0.9           foreach_1.5.2      
 [67] lhs_1.1.6           hardhat_1.3.0       lava_1.7.0         
 [70] rlang_1.1.1         pkgconfig_2.0.3     evaluate_0.18      
 [73] lattice_0.20-45     labeling_0.4.2      tidyselect_1.2.0   
 [76] parallelly_1.32.1   magrittr_2.0.3      R6_2.5.1           
 [79] generics_0.1.3      DBI_1.1.3           pillar_1.9.0       
 [82] haven_2.5.1         whisker_0.4         withr_2.5.0        
 [85] survival_3.4-0      nnet_7.3-18         future.apply_1.10.0
 [88] modelr_0.1.10       crayon_1.5.2        utf8_1.2.2         
 [91] tzdb_0.3.0          rmarkdown_2.18      grid_4.2.2         
 [94] readxl_1.4.1        git2r_0.30.1        reprex_2.0.2       
 [97] digest_0.6.30       httpuv_1.6.6        GPfit_1.0-8        
[100] munsell_0.5.0       viridisLite_0.4.1   bslib_0.4.1