In Conte-Ecology/singlePassDetection: Compares detection corrected abundance estimates for single and multiple pass fish surveys

Shenandoah Single vs Multi-pass Sampling

Cleaning before entry into R: Did text to columns by comma to separate coded notes into columns Replaced "MORT" and other non-numeric mortality notes with 7 Made a 'dead' column that was comprised of notes 7 (mortality) and 10 (whole fish collected) Replaced ' and " with blank to avoid problems reading in text columns Replaced ND (No Data) with NA for missing data Changed Species to Code in the fish data headers to be able to merge with the species data (could do in R after import) Added primary key Removed # symbol from GameFish

rm(list=ls())

library(data.table) # fast, efficient reading in of big data
library(dplyr)
library(ggplot2)
library(lubridate)

setwd('/Users/Dan/Documents/Research/Stream_Climate_Change/Brook_Trout/singlePassDetection/')

gameFish <- fread("inst/extdata/W_FI_Gamefish.csv", sep = ",", na.strings = c("NA", "ND", "", " "))

gameFish[gameFish == 99] <- NA

siteVisits <- fread("inst/extdata/R_SiteVisits.csv", sep = ",")

siteVisits[siteVisits == 999] <- NA

species <- fread("inst/extdata/W_FI_zdd_FISH_Species.csv", sep = ",", header = TRUE)
upstream <- fread("inst/extdata/shenandoahUpstreamStats.csv", sep = ",")

siteVisits$SiteID <- siteVisits$SITEID

# I love magrittr and pipes!
fish <- gameFish %>%
  group_by(SiteVisit_ID, RUN, Code, AGE) %>%
  summarise(meanTL = mean(TL), count = n()) %>%
  left_join(species, by = 'Code') %>%
  left_join(siteVisits, by = 'SiteVisit_ID')

Fill out RUN, species, and age combinations

This is not efficient at all but I was having trouble getting the columns to fill out properly, so I just brute forced it.

fish$siteVisitPass <- paste0(fish$SiteVisit_ID, "--", fish$RUN)

fullGrid <- expand.grid(siteVisitPass = unique(fish$siteVisitPass), Code = unique(fish$Code), AGE = 0:1)

fullGrid$siteVisitPass <- as.character(fullGrid$siteVisitPass)
fullGrid$Code <- as.character(fullGrid$Code)

fullDF <- fullGrid %>%
  left_join(fish, by = c("siteVisitPass", "Code", "AGE"))

fullDF$SiteVisit_ID <- matrix(unlist(strsplit(fullDF$siteVisitPass, split = "--", fixed = TRUE)), ncol = 2, byrow = TRUE)[ , 1]

fullDF$RUN <- matrix(unlist(strsplit(fullDF$siteVisitPass, split = "--", fixed = TRUE)), ncol = 2, byrow = TRUE)[ , 2]

foo <- select(siteVisits, SiteVisit_ID, SITEID, sDate)
fullDF2 <- left_join(fullDF, y = foo, by = c("SiteVisit_ID"))
fullDF2$SiteID <- fullDF2$SITEID.y

allDF <- left_join(fullDF2, upstream, by = 'SiteID')
allDF[is.na(allDF$count), ]$count <- 0 # replace count NA with 0 because now all the combinations should represent site visits and passes that were conducted for all species and ages

allDF$date <- ymd_hms(allDF$sDate.y, tz = "US/Eastern")

# reduce to the variables of interest for modeling and rename the variables with consistent format

subDF <- select(allDF,  
                        SiteID,
                        SiteVisit_ID,
                        date,
                        Code,
                        RUN,
                        AGE,
                        count,
                        FI_Shocker_Voltage,
                        FEATUREID,
                        Latitude,
                        Longitude,
                        TotDASqKM,
                        ReachSlope,
                        ReachElev,
                        tmaxann_brkt,
                        forest,
                        developed)

setnames(subDF, old = names(subDF), new = c("site",
                                                            "SiteVisit_ID",
                                                            "date",
                          "spp",
                          "pass",
                          "age",
                          "count",
                          "voltage",
                          "FEATUREID",
                          "lat",
                          "lon",
                          "drainageArea",
                          "slope",
                          "elev",
                          "tmaxAnn",
                          "forest",
                          "developed"))

remove sites missing covariate data and filter to brook trout

bktDF <- subDF %>%
  filter(!is.na(lat)) %>%
  filter(spp == "BKT")

# check 
subDF %>%
  group_by(spp) %>%
  summarise(siteVisitsPasses = n(), sum(count))

gameFish %>%
  group_by(Code) %>%
  summarise(siteVisitsPasses = n())

Prep climate data

Use the same flow and temperature for all sites

load("data/localShenDaymet.RData")

climate <- fread('inst/extdata/Shen_climate_data.csv', sep=',')

setnames(climate, names(climate), c("year", "yday", "tmax", "tmin", "precip", "siteID"))

climate$day.year <- paste0(climate$yday, '-', climate$year)
climate$date <- parse_date_time(climate$day.year, "%j%Y", tz = "US/Eastern")

# move this function to functions in R/
getSeason <- function(DATES) {
    WS <- as.Date("2012-12-15", format = "%Y-%m-%d") # Winter Solstice
    SE <- as.Date("2012-3-15",  format = "%Y-%m-%d") # Spring Equinox
    SS <- as.Date("2012-6-15",  format = "%Y-%m-%d") # Summer Solstice
    FE <- as.Date("2012-9-15",  format = "%Y-%m-%d") # Fall Equinox

    # Convert dates from any year to 2012 dates
    d <- as.Date(strftime(DATES, format="2012-%m-%d"))

    ifelse (d >= WS | d < SE, "Winter",
      ifelse (d >= SE & d < SS, "Spring",
        ifelse (d >= SS & d < FE, "Summer", "Fall")))
}

climate$season <- getSeason(climate$date)
annual.climate <- climate %>%
  group_by(year, season) %>%
  summarise(mean.tmax = mean(tmax), mean.precip = mean(precip))

write.table(annual.climate, file = 'inst/extdata/annual-climate.csv', row.names = FALSE, sep = ",")

masterData$season <- getSeason(masterData$date)

annual.site.climate <- masterData %>%
  group_by(site, year, season) %>%
  summarise(mean.tmean = mean(airTemp), mean.precip = mean(prcp))

save(annual.site.climate, file = 'data/annual-site-climate.RData')

Conte-Ecology/singlePassDetection documentation built on May 6, 2019, 12:51 p.m.