In mkoohafkan/rremat: Russian River Estuary data collected by the Largier Lab

Closure Analysis

#library(rremat)
library(ggplot2)
library(dplyr)
library(tidyr)
library(knitr)
library(reshape2)
library(cluster)
library(fpc)
opts_chunk$set(fig.width = 12, fig.height = 8, echo = FALSE)

Closure referencing

The code below adds a reference to CTD transect dates indicating the closure ID associated with the transect and the number of days since closure began. In this analysis, closures are identified based on the photographic record.

data(closures)
data(ctd)
#closurecol = "closure.id"
#durationcol = "days.since"
#ctd[closurecol] = NA
#ctd[durationcol] = NA
#for(i in seq(nrow(closures))){
#  idmask = ctd$date > closures$initiation[i] & ctd$date < closures$breach[i]
#  ctd[idmask, closurecol] = closures$id[i]
#  ctd[idmask, durationcol] = ctd[idmask, "date"] - closures$initiation[i]
#}

Constriction index

The constriction index is intended to provide an indicator of the degree of mouth constriction/closure. The Constriction index is the ratio of the daily ocean water level range to the estuary water level range.

We perform a cluster analysis to determine if the constriction index can be used to classify mouth condition. We test clustering with 4 clusters (closed, perched, constricted, open); 3 clusters (closed, partially-open, open); and 2 clusters (open, closed).

data(tides)
tiderange = summarize(group_by(tides, datetime = as.Date(datetime)), 
  min = min(height), max = max(height), range = max - min)
data(wll)
rrerange = summarize(group_by(wll, site, 
  mtime = as.Date(mtime)), min = min(depth), max = max(depth), 
  range = max - min)
# get water level from next nearest gauge if data at river mouth is missing
range1 = spread(rrerange[c("site", "mtime", "range")], site, range)
range1["source"] = "river mouth"
range1["range"] = range1[["river mouth"]]
range1[is.na(range1$range), "source"] = "jenner"
range1[is.na(range1$range), "range"] = range1[is.na(range1$range), "jenner"]
range1[is.na(range1$range), "source"] = "sheephouse creek"
range1[is.na(range1$range), "range"] = range1[is.na(range1$range), "sheephouse creek"]
range1[is.na(range1$range), "source"] = "heron rookery"
range1[is.na(range1$range), "range"] = range1[is.na(range1$range), "heron rookery"]
range1[is.na(range1$range), "source"] = "moscow bridge"
range1[is.na(range1$range), "range"] = range1[is.na(range1$range), "moscow bridge"]

rrerange = range1[c("mtime", "range", "source")]

data(ctdmeta)
rreratio = data.frame(ctdmeta, date = as.Date(ctdmeta$start), rrerange = NA, 
  tiderange = NA)
for(i in seq(nrow(rreratio))){
  thisdate = rreratio$date[i]
  if(thisdate %in% rrerange$mtime)
    rreratio[i, "rrerange"] = rrerange[rrerange$mtime == thisdate, "range"]
  if(thisdate %in% tiderange$datetime)
    rreratio[i, "tiderange"] = tiderange[tiderange$datetime == thisdate, "range"]
}  
rreratio["ci"]= rreratio$tiderange/rreratio$rrerange

rreratio[c("cluster4", "cluster3", "cluster2")] = NA
rreratio[!is.na(rreratio$ci), "cluster4"] = pam(scale(na.omit(rreratio$ci)), 4, 
  cluster.only = TRUE)
rreratio[!is.na(rreratio$ci), "cluster3"] = pam(scale(na.omit(rreratio$ci)), 3, 
  cluster.only = TRUE)
rreratio[!is.na(rreratio$ci), "cluster2"] = pam(scale(na.omit(rreratio$ci)), 2, 
  cluster.only = TRUE)

clustertest = pamk(rreratio$ci,krange=1:4, scaling=TRUE)