-
Notifications
You must be signed in to change notification settings - Fork 1
/
README.Rmd
240 lines (190 loc) · 14.3 KB
/
README.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
---
output:
rmarkdown::github_document:
html_preview: no
---
<!--- README.md is generated from README.Rmd. Please edit that file -->
# surveyvoi: Survey Value of Information
[](https://lifecycle.r-lib.org/articles/stages.html)
[](https://github.com/prioritizr/surveyvoi/actions)
[](https://github.com/prioritizr/surveyvoi/actions)
[](https://github.com/prioritizr/surveyvoi/actions)
[](https://github.com/prioritizr/surveyvoi/actions)
[](https://github.com/prioritizr/surveyvoi/actions)
[](https://app.codecov.io/gh/prioritizr/surveyvoi/branch/master)
[](https://CRAN.R-project.org/package=surveyvoi)
```{r "knitr_config", include = FALSE}
knitr::opts_chunk$set(fig.path = "man/figures/README-", fig.align = "center")
```
```{r "initialization", include = FALSE}
devtools::load_all()
h = 3.5
w = 5.0
ow = "350"
knitr::opts_chunk$set(
fig.height = h,
fig.width = w,
out.width = ow,
fig.path = "man/figures/README-",
fig.align = "center"
)
```
The _surveyvoi_ package is a decision support tool for prioritizing sites for ecological surveys based on their potential to improve plans for conserving biodiversity (e.g. plans for establishing protected areas). Given a set of sites that could potentially be acquired for conservation management -- wherein some sites have previously been surveyed and other sites have not -- this package provides functionality to generate and evaluate plans for additional surveys. Specifically, plans for ecological surveys can be generated using various conventional approaches (e.g. maximizing expected species richness, geographic coverage, diversity of sampled environmental conditions) and by maximizing value of information. After generating plans for surveys, they can also be evaluated using value of information analysis. Please note that several functions depend on the Gurobi optimization software (available from https://www.gurobi.com). Additionally, the JAGS software (available from https://mcmc-jags.sourceforge.io/) is required to fit hierarchical generalized linear models.
## Installation
The latest official version can be installed from the [Comprehensive R Archive Network (CRAN)](https://cran.r-project.org/) using the following _R_ code.
```{r, eval = FALSE}
install.packages("surveyvoi", repos = "https://cran.rstudio.com/")
```
Alternatively, the latest development version can be installed from [GitHub](https://github.com/prioritizr/surveyvoi) using the following code. Please note that while developmental versions may contain additional features not present in the official version, they may also contain coding errors.
```{r, eval = FALSE}
if (!require(remotes)) install.packages("remotes")
remotes::install_github("prioritizr/surveyvoi")
```
#### Windows
The [Rtools](https://cran.r-project.org/bin/windows/Rtools/) software needs to be installed to install the _surveyvoi R_ package from source. This software provides system requirements from [rwinlib](https://github.com/rwinlib/).
#### Ubuntu
The `gmp`, `fftw3`, `mpfr`, and `symphony` libraries need to be installed to install the _surveyvoi R_ package. Although the `fftw3` and `symphony` libraries are not used directly, they are needed to successfully install dependencies. For recent versions of Ubuntu (18.04 and later), these libraries are available through official repositories. They can be installed using the following system commands:
```
apt-get -y update
apt-get install -y libgmp3-dev libfftw3-dev libmpfr-dev coinor-libsymphony-dev
```
#### Linux
For Unix-alikes, `gmp` (>= 4.2.3), `mpfr` (>= 3.0.0), `fftw3` (>= 3.3), and `symphony` (>= 5.6.16) are required.
#### macOS
The `gmp`, `fftw`, `mpfr`, and `symphony` libraries are required. Although the `fftw3` and `symphony` libraries are not used directly, they are needed to successfully install dependencies. The easiest way to install these libraries is using [HomeBrew](https://brew.sh/). After installing HomeBrew, these libraries can be installed using the following commands in the system terminal:
```
brew tap coin-or-tools/coinor
brew install symphony
brew install pkg-config
brew install gmp
brew install fftw
brew install mpfr
```
## Citation
Please cite the _surveyvoi R_ package when using it in publications. To cite the package, please use:
> Hanson, JO, McCune JL, Chadès I, Proctor CA, Hudgins EJ, & Bennett JR (2023) Optimizing ecological surveys for conservation. Journal of Applied Ecology, 60: 41--51. Available at <https://doi.org/10.1111/1365-2664.14309>.
## Usage
Here we provide a short example showing how to use the _surveyvoi R_ package to prioritize funds for ecological surveys. In this example, we will generate plans for conducting ecological surveys (termed "survey schemes") using simulated data for six sites and three conservation features (e.g. bird species). To start off, we will set the seed for the random number generator for reproducibility and load some R packages.
```{r, message = FALSE}
set.seed(502) # set RNG for reproducibility
library(surveyvoi) # package for value of information analysis
library(dplyr) # package for preparing data
library(tidyr) # package for preparing data
library(ggplot2) # package for plotting data
```
Now we will load some datasets that are distributed with the package. First, we will load the `sim_sites` object. This spatially explicit dataset (i.e. `sf` object) contains information on the sites within our study area. Critically, it contains (i) sites that have already been surveyed, (ii) candidate sites for additional surveys, (iii) sites that have already been protected, and (iv) candidate sites that could be protected in the future. Each row corresponds to a different site, and each column describes different properties associated with each site. In this table, the `"management_cost"` column indicates the cost of protecting each site; `"survey_cost"` column indicates the cost of conducting an ecological survey within each site; and `"e1"` and `"e2"` columns contain environmental data for each site (not used in this example). The remaining columns describe the existing survey data and the spatial distribution of the features across the sites. The `"n1"`, `"n2"`, and `"n3"` columns indicate the number of surveys conducted within each site that looked for each of the three features (respectively); and `"f1"`, `"f2"`, and `"f3"` columns describe the proportion of surveys within each site that looked for each feature where the feature was detected (respectively). For example, if `"n1"` has a value of 2 and `"f1"` has a value of 0.5 for a given site, then the feature `"f1"` was detected in only one of the two surveys conducted in this site that looked for the feature. Finally, the `"p1"`, `"p2"`, and `"p3"` columns contain modeled probability estimates of each species being present in each site (see `fit_hglm_occupancy_models()` and `fit_xgb_occupancy_models()` to generate such estimates for your own data).
```{r "load_site_data"}
# load data
data(sim_sites)
# print table
print(sim_sites, width = Inf)
```{r "management_cost_plot"}
# plot cost of protecting each site
ggplot(sim_sites) +
geom_sf(aes(color = management_cost), size = 4) +
ggtitle("management_cost") +
theme(legend.title = element_blank(), text = element_text(size = 16))
```
```{r "survey_cost_plot"}
# plot cost of conducting an additional survey in each site
# note that these costs are much lower than the protection costs
ggplot(sim_sites) +
geom_sf(aes(color = survey_cost), size = 4) +
ggtitle("survey_cost") +
theme(legend.title = element_blank(), text = element_text(size = 16))
```
```{r "n_plot", fig.height = h, fig.width = w * 2.4, out.width = "800"}
# plot survey data
## n1, n2, n3: number of surveys in each site that looked for each feature
sim_sites %>%
select(n1, n2, n3) %>%
gather(name, value, -geometry) %>%
ggplot() +
geom_sf(aes(color = value), size = 4) +
facet_wrap(~name, nrow = 1) +
theme(text = element_text(size = 16))
```
```{r "f_plot", fig.height = h, fig.width = w * 2.4, out.width = "800"}
# plot survey results
## f1, f2, f3: proportion of surveys in each site that looked for each feature
## that detected the feature
sim_sites %>%
select(f1, f2, f3) %>%
gather(name, value, -geometry) %>%
ggplot() +
geom_sf(aes(color = value), size = 4) +
facet_wrap(~name, nrow = 1) +
scale_color_continuous(limits = c(0, 1)) +
theme(text = element_text(size = 16))
```
```{r "p_plot", fig.height = h, fig.width = w * 2.4, out.width = "800"}
# plot modeled probability of occupancy data
sim_sites %>%
select(p1, p2, p3) %>%
gather(name, value, -geometry) %>%
ggplot() +
geom_sf(aes(color = value), size = 4) +
facet_wrap(~name, nrow = 1) +
scale_color_continuous(limits = c(0, 1)) +
theme(text = element_text(size = 16))
```
Next, we will load the `sim_features` object. This table contains information on the conservation features (e.g. species). Specifically, each row corresponds to a different feature, and each column contains information associated with the features. In this table, the `"name"` column contains the name of each feature; `"survey"` column indicates whether future surveys would look for this species; `"survey_sensitivity"` and `"survey_specificity"` columns denote the sensitivity (true positive rate) and specificity (true negative rate) for the survey methodology for correctly detecting the feature; `"model_sensitivity"` and `"model_specificity"` columns denote the sensitivity (true positive rate) and specificity (true negative rate) for the species distribution models fitted for each feature; and `"target"` column denotes the required number of protected sites for each feature (termed "representation target", each feature has a target of `r sim_features$target[1]` site).
```{r "load_feature_data"}
# load data
data(sim_features)
# print table
print(sim_features, width = Inf)
```
After loading the data, we will now generate an optimized ecological survey scheme. To achieve this, we will use `approx_optimal_survey_scheme()` function. This function uses a greedy heuristic algorithm to maximize value of information. Although other functions can return solutions that are guaranteed to be optimal (i.e. `optimal_survey_scheme()`), they can take a very long time to complete because they use a brute-force approach. This function also uses an approximation routine to reduce computational burden.
To perform the optimization, we will set a total budget for (i) protecting sites and (ii) surveying sites. Although you might be hesitant to specify a budget, please recall that you would make very different plans depending on available funds. For instance, if you have near infinite funds then you wouldn't bother conducting any surveys and simply protect everything. Similarly, if you had very limited funds, then you wouldn't survey any sites to ensure that at least one site could be protected. Generally, conservation planning problems occur somewhere between these two extremes---but the optimization process can't take that into account if you don't specify a budget. For brevity, here we will set the total budget as 90% of the total costs for protecting sites.
```{r "generate_optimal_survey_scheme", message = FALSE, results = FALSE}
# calculate budget
budget <- sum(0.4 * sim_sites$management_cost)
# generate optimized survey scheme
opt_scheme <-
approx_optimal_survey_scheme(
site_data = sim_sites,
feature_data = sim_features,
site_detection_columns = c("f1", "f2", "f3"),
site_n_surveys_columns = c("n1", "n2", "n3"),
site_probability_columns = c("p1", "p2", "p3"),
site_management_cost_column = "management_cost",
site_survey_cost_column = "survey_cost",
feature_survey_column = "survey",
feature_survey_sensitivity_column = "survey_sensitivity",
feature_survey_specificity_column = "survey_specificity",
feature_model_sensitivity_column = "model_sensitivity",
feature_model_specificity_column = "model_specificity",
feature_target_column = "target",
total_budget = budget,
survey_budget = budget,
verbose = TRUE
)
```
```{r "validation", include = FALSE}
# ensure that at least one site is selected
assertthat::assert_that(
sum(c(opt_scheme[1, ])) >= 1,
msg = "no sites selected, not a compelling example"
)
```
```{r "survey_scheme_plot"}
# the opt_scheme object is a matrix that contains the survey schemes
# each column corresponds to a different site,
# and each row corresponds to a different solution
# in the event that there are multiple near-optimal survey schemes, then this
# matrix will have multiple rows
print(str(opt_scheme))
# let's add the first solution (row) in opt_scheme to the site data to plot it
sim_sites$scheme <- c(opt_scheme[1, ])
# plot scheme
# TRUE = selected for an additional ecological survey
# FALSE = not selected
ggplot(sim_sites) +
geom_sf(aes(color = scheme), size = 4) +
ggtitle("scheme") +
theme(text = element_text(size = 16))
```
This has just been a taster of the _surveyvoi R_ package. In addition to this functionality, it can be used to evaluate survey schemes using value of information analysis. Furthermore, it can be used to generate survey schemes using conventional approaches (e.g. sampling environmental gradients, and selecting places with highly uncertain information). For more information, see the [package vignette](https://prioritizr.github.io/surveyvoi/articles/surveyvoi.html).
## Getting help
If you have any questions about using the _surveyvoi R_ package or suggestions for improving it, please [file an issue at the package's online code repository](https://github.com/prioritizr/surveyvoi/issues/new).