Add salps study

_quarto.yml

@@ -15,6 +15,8 @@ website:
     contents:
       - href: index.qmd
         text: Home
+      - href: salps.qmd
+        text: Salpsae
       - href: about.qmd
         text: About
 tools:

docs/search.json

@@ -4,10 +4,7 @@
     "href": "index.html",
     "title": "Jellycast",
     "section": "",
-    "text": "here() starts at /mnt/s1/projects/ecocast/projects/jellyscape\n\n\nFor the Jellyscape project observations are drawn from a recent release of unstaged ECOMON data. We choose the following groups to belong to the gelatinous “jellyscape” although the list is easily modified.\n\nSiphonophores (siph)\nHydromedusea (hydrom)\nCoelenterates (coel)\nCtenophores (ctenop)\nSalps (salps)\n\nThese we can read from the ecomon package and aggregate these species into a count per unit area (10m^2) or count per unit volume (100m^3). Accepting the defaults for the read_ecomon_spp() function yields an aggregated (summed) table for areal abundance for the 5 groups listed.\n\n1 The Ecomon observation data\nHere we show the general distribution of records for the 5 groups listed. Later we show more detailed per-group analyses.\n\n\n\n\n\n\n\n\n\nTo study just the listed groups, we reduce the data set by discarding all non-listed groups and provide an aggregate abundance as shown above. For the selected groups the data covers 1977-02-13 through 2021-11-15 with 32693 records of observation.\nOver the years the number of observations has decreased (that isn’t to say that the effort has decreased.) The survey effort is concentrated in spring and autumn months.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n Back to top",
-    "crumbs": [
-      "Home"
-    ]
+    "text": "here() starts at /mnt/s1/projects/ecocast/projects/jellyscape\n\n\nFor the Jellyscape project observations are drawn from a recent release of unstaged ECOMON data. We choose the following groups to belong to the gelatinous “jellyscape” although the list is easily modified.\n\nSiphonophores (siph)\nHydromedusea (hydrom)\nCoelenterates (coel)\nCtenophores (ctenop)\nSalps (salps)\n\nThese we can read from the ecomon package and aggregate these species into a count per unit area (10m^2) or count per unit volume (100m^3). Accepting the defaults for the read_ecomon_spp() function yields an aggregated (summed) table for areal abundance for the 5 groups listed.\n\n1 The Ecomon observation data\nHere we show the general distribution of records for the 5 groups listed. Later we show more detailed per-group analyses.\n\n\n\n\n\nTo study just the listed groups, we reduce the data set by discarding all non-listed groups and provide an aggregate abundance as shown above. For the selected groups the data covers 1977-02-13 through 2021-11-15 with 32693 records of observation.\nOver the years the number of observations has decreased (that isn’t to say that the effort has decreased.) The survey effort is concentrated in spring and autumn months.\n\n\n\n\n\n\n\n\n\n\n\n\n Back to top"
   },
   {
     "objectID": "observations.html",
@@ -21,9 +18,13 @@
     "href": "about.html",
     "title": "About",
     "section": "",
-    "text": "The jellyscape project was developed at Bigelow Laboratory’s Tandy Center for Ocean Forecasting to aid in the development of modeling studies requiring either observation/abundance point data or to modeled rasterized layers for subsequent use in modeling studies. Use the github issues tracker to post questions or report issues.\n\n\n\n Back to top",
-    "crumbs": [
-      "About"
-    ]
+    "text": "The jellyscape project was developed at Bigelow Laboratory’s Tandy Center for Ocean Forecasting to aid in the development of modeling studies requiring either observation/abundance point data or to modeled rasterized layers for subsequent use in modeling studies. Use the github issues tracker to post questions or report issues.\n\n\n\n Back to top"
+  },
+  {
+    "objectID": "salps.html",
+    "href": "salps.html",
+    "title": "Salpsae",
+    "section": "",
+    "text": "here() starts at /mnt/s1/projects/ecocast/projects/jellyscape\n\n\n\nthisgroup = \"Salps\"\nx = read_ecomon_spp(form = 'sf', groups = thisgroup, agg = FALSE) |>\n  ecomon_to_long() |>\n  dplyr::mutate(year = format(date, \"%Y\") ,\n                month = factor(format(date, \"%b\"), levels = month.abb))\nnonz = dplyr::filter(x, value > 0)\nzero = dplyr::filter(x, value <= 0)\nn = nrow(x)\nnz = zero |> nrow()\ncoast = sf::st_crop(coast, x)\n\nThere are 32693 records for salps of which 23202 are zeroes (about 70.97%) leaving 9491 non-zero salp abundance records.\n\n1 Non-zero abundances\nHere’s an abudnance scaled plot of the non-zero abundances.\n\n\nCode\nplot(nonz['value'],\n     pch = \".\",\n     logz = TRUE,\n     main = glue::glue(\"Non-zero abundance of {thisgroup}\"),\n     axes = TRUE,\n     reset = FALSE)\nplot(coast, add = TRUE, col = \"black\")\n\n\n\n\n\n\n\nCode\nggplot(data = dplyr::group_by(nonz, year), \n       mapping = aes(x = year)) +\n  geom_bar() +\n  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + \n  labs(y = \"Record count\", title = glue::glue(\"Non-zero abundance {thisgroup} Records by Year\"))\n\n\n\n\n\nCode\nggplot(data = dplyr::group_by(nonz, month), \n       mapping = aes(x = month)) +\n  geom_bar() +\n  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + \n  labs(y = \"Record count\", title = glue::glue(\"Non-zero abundance {thisgroup} Records by Month\"))\n\n\n\n\n\n\n\n2 Zero Abundance (i.e. absences)\nThe distribution of zero abundance records is much more extensive.\n\n\nCode\nplot(zero['value'] |>\n       dplyr::filter(value <= 0) |>\n       sf::st_geometry(),\n     pch = \".\",\n     main = glue::glue(\"Zero abundance presences {thisgroup}\"),\n     axes = TRUE,\n     reset = FALSE)\nplot(coast, add = TRUE, col = \"black\")\n\n\n\n\n\n\n\nCode\nggplot(data = dplyr::group_by(zero, year), \n       mapping = aes(x = year)) +\n  geom_bar() +\n  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + \n  labs(y = \"Record count\", title = glue::glue(\"Zero abundance {thisgroup} Records by Year\"))\n\n\n\n\n\nCode\nggplot(data = dplyr::group_by(zero, month), \n       mapping = aes(x = month)) +\n  geom_bar() +\n  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + \n  labs(y = \"Record count\", title = glue::glue(\"Zero abudnance {thisgroup} Records by Month\"))\n\n\n\n\n\n\n\n3 A closer look at non-zero abundances\n\nqs = quantile(nonz$value, seq(from = 0.1, to = 1, by = 0.1))\nix = findInterval(nonz$value, qs, all.inside = TRUE)\nnonz = dplyr::mutate(nonz,\n                     percentile = factor(names(qs)[ix], levels = names(qs)))\n\n\n\nCode\nggplot(data = nonz) + \n  geom_sf(alpha = 0.1, size = 0.3) + \n  geom_sf(data = coast, \n          color = \"orange\") +\n  facet_wrap(~percentile)\n\n\n\n\n\n\n\n\n\n Back to top"
   }
 ]

index_cache/html/__packages

@@ -25,3 +25,4 @@ workflows
 workflowsets
 yardstick
 ecomon
+glue

salps.qmd

@@ -0,0 +1,107 @@
+---
+title: "Salpsae"
+format:
+  html:
+    code-fold: true
+---
+
+```{r setup, echo = FALSE}
+source("setup.R")
+groups = names(default_groups())
+coast = read_coast()
+```
+
+```{r read_ecomon, message=FALSE}
+#| code-fold: false
+thisgroup = "Salps"
+x = read_ecomon_spp(form = 'sf', groups = thisgroup, agg = FALSE) |>
+  ecomon_to_long() |>
+  dplyr::mutate(year = format(date, "%Y") ,
+                month = factor(format(date, "%b"), levels = month.abb))
+nonz = dplyr::filter(x, value > 0)
+zero = dplyr::filter(x, value <= 0)
+n = nrow(x)
+nz = zero |> nrow()
+coast = sf::st_crop(coast, x)
+```
+
+There are `r n` records for `salps` of which `r nz` are zeroes (about `r sprintf("%0.2f%%", nz/n*100)`) leaving `r n-nz` non-zero salp abundance records.
+
+# Non-zero abundances
+
+Here's an abudnance scaled plot of the non-zero abundances.
+```{r}
+plot(nonz['value'],
+     pch = ".",
+     logz = TRUE,
+     main = glue::glue("Non-zero abundance of {thisgroup}"),
+     axes = TRUE,
+     reset = FALSE)
+plot(coast, add = TRUE, col = "black")
+```
+
+
+```{r, warning=FALSE}
+ggplot(data = dplyr::group_by(nonz, year), 
+       mapping = aes(x = year)) +
+  geom_bar() +
+  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + 
+  labs(y = "Record count", title = glue::glue("Non-zero abundance {thisgroup} Records by Year"))
+
+ggplot(data = dplyr::group_by(nonz, month), 
+       mapping = aes(x = month)) +
+  geom_bar() +
+  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + 
+  labs(y = "Record count", title = glue::glue("Non-zero abundance {thisgroup} Records by Month"))
+```
+
+
+
+# Zero Abundance (*i.e.* absences)
+
+The distribution of zero abundance records is much more extensive.
+
+```{r}
+plot(zero['value'] |>
+       dplyr::filter(value <= 0) |>
+       sf::st_geometry(),
+     pch = ".",
+     main = glue::glue("Zero abundance presences {thisgroup}"),
+     axes = TRUE,
+     reset = FALSE)
+plot(coast, add = TRUE, col = "black")
+```
+
+```{r, warning=FALSE}
+ggplot(data = dplyr::group_by(zero, year), 
+       mapping = aes(x = year)) +
+  geom_bar() +
+  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + 
+  labs(y = "Record count", title = glue::glue("Zero abundance {thisgroup} Records by Year"))
+
+ggplot(data = dplyr::group_by(zero, month), 
+       mapping = aes(x = month)) +
+  geom_bar() +
+  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) + 
+  labs(y = "Record count", title = glue::glue("Zero abudnance {thisgroup} Records by Month"))
+```
+
+
+# A closer look at non-zero abundances
+
+```{r}
+#| code-fold: false
+qs = quantile(nonz$value, seq(from = 0.1, to = 1, by = 0.1))
+ix = findInterval(nonz$value, qs, all.inside = TRUE)
+nonz = dplyr::mutate(nonz,
+                     percentile = factor(names(qs)[ix], levels = names(qs)))
+```
+
+
+```{r}
+ggplot(data = nonz) + 
+  geom_sf(alpha = 0.1, size = 0.3) + 
+  geom_sf(data = coast, 
+          color = "orange") +
+  facet_wrap(~percentile)
+```

salps_cache/html/__packages

@@ -0,0 +1,28 @@
+RColorBrewer
+here
+remotes
+tidyr
+readr
+sf
+abind
+stars
+glue
+rnaturalearth
+tidymodels
+broom
+scales
+dials
+dplyr
+ggplot2
+infer
+modeldata
+parsnip
+purrr
+recipes
+rsample
+tibble
+tune
+workflows
+workflowsets
+yardstick
+ecomon

setup.R

@@ -12,7 +12,7 @@
 #' `github` packages can be installed with `remotes::install_github()`
 packages = list(
   cran = c("RColorBrewer", "here", "remotes", "tidyr", "readr", "sf", "stars", 
-           "rnaturalearth", "tidymodels", "dplyr"),
+           "glue", "rnaturalearth", "tidymodels", "dplyr"),
   github = c(ecomon = "BigelowLab")
 )