2.1 Get the ohi-global repo

We’ll be playing with weights and reference points for the global OHI assessments; we calculate scores each year for 221 coastal nations and territories. Let’s get it on your computer.

1. Go to github.com/OHI-Science/ohi-global. This is the ohi-global repository (called ‘repo’).
2. Click on the green ‘clone or download’ button.
3. Clone or download somewhere you can find on your computer. If you are familiar with GitHub, clone this repository. Otherwise click ‘Download ZIP’.
4. Navigate to where you saved the ohi-global repo on your computer and double-click the ohi-global.Rproj file. This will open RStudio (which is how you should always interface with R).
5. Within the RStudio file directory pane (bottom right), navigate to the ESM_270 folder. Inside is a mix of .R scripts and .csv files. We will just focus on a few of them.

2.2 Setup

Since we’re in RStudio with a .Rproj file, our current working directory is wherever/you/saved/it/ohi-global. All of our work for the assignment happens in the ESM_270 folder (what you are reading is from ohi-global/ESM_270/ESM_270_OHI.Rmd). Let’s change the working directory to the ESM_270 folder, and install the packages we need to get going.

Troubleshooting: If you have any problems, check that R and RStudio are up-to-date. The most recent version of R is 3.3 and RStudio is 0.99. If needed, update R here and RStudio here.

Run the following code:

## set working directory to ohi-global/ESM_270 with setwd()
setwd('ESM_270')

## delete any existing version of `ohicore`
for (p in c('ohicore')){
  if (p %in% rownames(installed.packages())){
    lib = subset(as.data.frame(installed.packages()), Package==p, LibPath, drop=T)
    remove.packages(p, lib)
  }
}

## install dependencies
for (p in c('devtools', 'git2r')){
  if (!require(p, character.only=T)){
    install.packages(p)
    require(p, character.only=T)
  }
}

## install most current version of ohicore -- don't worry about the warnings. But make sure there are no errors.
devtools::install_github('ohi-science/ohicore@dev')

## load ohicore library -- again, don't worry about the warnings. But ensure the last line is not an error.
library(ohicore)

## source the function to make flower plots; this won't do anything yet.
source('PlotFlowerMulti.r')

2.3 Calculate scores

We can calculate OHI scores by sourcing calculate_scores.r (below). This uses the following files, which are all in our current working directory, ohi-global/ESM_270:

• input data layers in the layers folder
• goal models and reference points in conf/functions.r
• goal weighs in conf/goals.csv
• calculated scores for all goals and all 221 regions are saved in scores.csv.

Note: because the Toolbox relies on these file names for calculation and for visualizing, we’ll be overwriting some of these files (after saving the originals) to use in each of the steps as we go forward.

## calculate scores. The Toolbox will describe each step as it calculates
source('calculate_scores.r')

## save copies of original files
file.copy('conf/goals.csv',   'conf/goals_orig.csv', overwrite = TRUE)
file.copy('conf/functions.r', 'conf/functions_orig.r', overwrite = TRUE)
file.copy('scores.csv',       'scores_orig.csv', overwrite = TRUE)

2.4 Visualize OHI scores for your region

We just calculated scores for all 221 regions. But you’ll just be inspecting one region, which do you choose? Each region has a unique numeric identifier. Run this to view the full list of regions and choose one to inspect: View(read.csv('layers/rgn_global.csv')). Use use the search bar in the top right of the View pane to find your country and note its rgn_id.

You’ll assign it in R now as my_rgn:

## assign your rgn_id
my_rgn <-  163 # United States; change this to your region of choice

2.5 Flower plot for your region

Now, let’s visualize your region’s scores in a flower plot. The PlotFlowerMulti() function will make figures for you, saving them as .pngs in ohi-global/ESM_270/reports/figures and displaying them in the RStudio Viewer (they will look a bit cut-off in the Viewer but that is just a vizualization glitch; the .pngs are fine).

Run the following – it will save a file with your region’s name, for example ohi-global/ESM_270/reports/figures/flower_orig_United_States.png.:

## parameters for the flower plot
scores_orig <- read.csv('scores.csv')
conf_orig <- read.csv('conf/goals.csv')
name_fig_orig <- 'reports/figures/flower_orig'   # my_rng is appended to name_fig

## function call for the flower plot -- original
PlotFlowerMulti(scores       = scores_orig, 
                rgns_to_plot = my_rgn, 
                conf         = conf_orig, 
                name_fig     = name_fig_orig)

2.6 Change goal weighting

Now you will change the goal weighting for one of the goals.

Refer to the Introduction above to see the list of goals. Currently, goals are weighted equally. This means goals have a weight of 1; if a goal has sub-goals, those are weighted each as 0.5. View this with View(conf_orig %>% dplyr::select(goal, name, weight)). Identify the 2- or 3- letter code for the goal or sub-goal you want to change.

Run the following after updating the weight; the example below shows changing the weight of the Artisanal Fishing Opportunity goal (‘AO’) to 3. Note that all of these weights are relative; so you can just change one goal weight without worrying about what it adds up to. This will save a file with your region’s name, for example ohi-global/ESM_270/reports/figures/flower_weights_new_United_States.png.

## create a new variable conf_new_wt
conf_new_wt <- conf_orig

## change the weight of a goal, eg AO to 3 and save file
conf_new_wt$weight[conf_new_wt$goal == 'AO'] = 3. # the 2- or 3- letter code must be in quotes. 
write.csv(conf_new_wt, 'conf/goals.csv', row.names = FALSE) 

## re-calculate scores with new weight
source('calculate_scores.r')
scores_new_wt <- read.csv('scores.csv')

## parameter for the flower plot -- new weight
name_fig_weights <- 'reports/figures/flower_weights_new'

## function call for the flower plot
PlotFlowerMulti(scores       = scores_new_wt, 
                rgns_to_plot = my_rgn, 
                conf         = conf_new_wt, 
                name_fig     = name_fig_weights)

2.7 Change reference point

The Lasting Special Places sub-goal represents geographic locations that hold particular value for aesthetic, spiritual, cultural, recreational or existence reasons, and assesses how well they are protected. In global assessments, we model this with the proportion of the offshore area that is a marine protected area (MPAs), with the reference point set to 30%.

You can see this by navigating to conf/functions.r, and looking at line 1578: you can also search for LSP =. Right now it’s set to 30% of waters should be in MPAs. What if we altered that? The example below changes it to 100%.

Go ahead and change it and run the following. It will save a file with your region’s name, for example ohi-global/ESM_270/reports/figures/flower_ref_new_United_States.png.

## set a new reference point
ref_pct_new <- 40 # change this with your reference point

## run this to insert your new reference point into conf/functions.r
readLines('conf/functions.r') %>%
  str_replace("LSP = .*", sprintf("LSP = function(layers, ref_pct_cmpa=%s, ref_pct_cp=%s, status_year){", 
                                  ref_pct_new, ref_pct_new)) %>%
  writeLines('conf/functions.r')

## reset original conf/goals.csv file
write.csv(conf_orig, 'conf/goals.csv', row.names = FALSE) # use original conf

## re-calculate scores with new reference point
source('calculate_scores.r')
scores_new_ref <- read.csv('scores.csv')

## parameter for the flower plot -- new reference point
name_fig_ref <- 'reports/figures/flower_ref_new'

PlotFlowerMulti(scores       = scores_new_ref, 
                rgns_to_plot = my_rgn, 
                conf         = conf_orig, 
                name_fig     = name_fig_ref)

How does the new reference point flower plot compare with the original? How much would you have expected the overall score to change? And how would you communicate this using the flower plots?

Run this to compare how much the LSP goal score changed vs. the overall Index score (displayed in the middle of the flower plot):

## investigate the LSP scores: 
scores_orig %>%
  filter(region_id == my_rgn, 
         goal == 'LSP', 
         dimension == 'score')

scores_new_ref %>%
  filter(region_id == my_rgn, 
         goal == 'LSP', 
         dimension == 'score')

## investigate the overall Index scores 
scores_orig %>%
  filter(region_id == my_rgn, 
         goal == 'Index', 
         dimension == 'score')

scores_new_ref %>%
  filter(region_id == my_rgn, 
         goal == 'Index', 
         dimension == 'score')