Chena Drift temperature

Website last updated 2020-10-14 14:13:51 by Benjamin Meyer (bemeyer@alaska.edu)

Introduction

This document will summarize and plot observed temperature values collected with HOBO TempPro v2 loggers in the Chena River. Data is primarily from summers 2019-2020.

Logger locations

# make dataframe of coordinates of HOBO sites


# create map
#leaflet(data = coords) %>%
#  addTiles() %>%  # Add default OpenStreetMap map tiles
#  fitBounds(-147.9, 65.08,-146.03, 64.7) %>%
#  addMarkers(~longitude, ~latitude, popup = ~as.character(site), label = ~as.character(site))

Data import and preparation

Import 2019 data

# Read in the 2019 temp logger deployment data
deploy19 <- read_excel(paste0(user,"Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/TempLoggerDeploy.xlsx"), sheet = "TempLoggerDeploy2019", col_types = c("text", "numeric", "text", "text", "text", "text", "text", "text", "numeric","numeric","numeric", "numeric", "numeric", "text","text", "text", "text", "text", "text", "text", "text","numeric", "text"))

# Read in temp data from 2019 csv files
# specify csv folder location
csv19 <- paste0(user,"/Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/2019_HOBO_Temperature/csv")

# # Read in one  csv file
# file <- "10886779.csv"
# SerialNumber <- str_sub(file, 1, 8)
# temp1 <- read_csv(paste(logger, ".csv", sep = ""), skip = 2, 
#                   col_names = c("RecordNumber", "DateTimeText", "Temp"),
#                   col_types = "icd-----")
# # Add a column with the logger SN
# temp2 <- temp1 %>%
#   mutate(SerialNumber = SerialNumber,
#          DateTime = mdy_hms(DateTimeText)) %>%
#   left_join(deploy, by = "SerialNumber")

# Read in all 2019 csv files in csv folder directory
csvFiles <- list.files(csv19)

# initialize a dataframe to hold all the 2019 temp data
rawTemp19 <- NULL

# Read in each 2019 csv file and append it to the bottom of the temp dataframe
for(i in 1:length(csvFiles)) {
  # specify csv folder location for read-in function (necessary in Markdown files)
  setwd(csv19) 
  file <- csvFiles[i]
  SerialNumber <- str_sub(file, 1, 8)
  # Read in the csv file
  thisTemp <- read_csv(file, skip = 2, 
                    col_names = c("RecordNumber", "DateTimeText", "Temp"),
                    col_types = "icd-----")
  # Add a column with the logger SN
  thisTemp <- thisTemp %>%
    mutate(SerialNumber = SerialNumber)
  rawTemp19 <- rbind(rawTemp19, thisTemp)
}

# Clean up 2019 deployment data -------------
deploy19 <- deploy19 %>%
  mutate(DeployDate = mdy(DeployDate),
         DeployTime = hms(DeployTime),
         RetrieveDate = mdy(RetrieveDate))

# 2019 data: clean up and prep for plotting
temp19 <- rawTemp19 %>%
  left_join(deploy19, by = "SerialNumber") %>%
  mutate(DateTime = mdy_hms(DateTimeText),
         Site = factor(Site)) %>%
  select(FieldSeason, Site, SerialNumber, DateTime, Temp, DeployDate:DownloadComments) %>%
  filter(!is.na(Site)) %>%
  # Trim temp readings outside the deployment window
  filter(DateTime > DeployDate + days(1)) %>%
  filter(DateTime < RetrieveDate - days(1)) %>%
  # When both loggers recorded temps, average them
  group_by(Site, DateTime) %>%
  summarize(Temp = mean(Temp, na.rm = T)) %>%
  # Specify the mainstem sites
  # BM 10/1/20 will need to adjust this code to incorporate ID for Headwaters tribs
  mutate(ChannelType = ifelse(Site == "Third Bridge" | Site == "First Bridge" | 
                                Site == "Moose Creek Dam" | Site == "Nordale" |
                                Site == "Loftus Road", "Mainstem",
                              ifelse(Site == "First Bridge Slough", "Slough", "Tributary")))

Import 2020 data

### 2020
# replicate process for 2020 water temperature data

# Read in the 2020 temp logger deployment data
deploy20 <- read_excel(paste0(user,"Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/TempLoggerDeploy.xlsx"), sheet = "TempLoggerDeploy2020") %>%
  transform(SerialNumber = as.character(SerialNumber))

# Read in temp data from 2020 csv files
# specify csv folder location
csv20 <- paste0(user,"/Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/2020_HOBO_Temperature/csv")

# Read in all 2020 csv files in csv folder directory
csvFiles <- list.files(csv20)

# initialize a dataframe to hold all the 2020 temp data
rawTemp20 <- NULL

# Read in each 2020 csv file and append it to the bottom of the temp dataframe
for(i in 1:length(csvFiles)) {
  setwd(csv20) 
  file <- csvFiles[i]
  SerialNumber <- str_sub(file, 1, 8)
  # Read in the csv file
  thisTemp <- read_csv(file, skip = 2, 
                       col_names = c("RecordNumber", "DateTimeText", "Temp"),
                       col_types = "icd-----")
  # Add a column with the logger SN
  thisTemp <- thisTemp %>%
    mutate(SerialNumber = SerialNumber)
  rawTemp20 <- rbind(rawTemp20, thisTemp)
}


# loss of most data somewhere in this pipe, why?


# 2020 data: clean up and prep for plotting
temp20 <- rawTemp20 %>%
  left_join(deploy20, by = "SerialNumber") %>%
  filter(!is.na(Temp)) %>%
  # received date/time format is error-prone; re-arrange to make unambiguous for conversion to POSIXct class
  separate(DateTimeText,sep = " ", into = c("Date","Time"), remove = F) %>%
  separate(Date,sep="/",into=c("month","day","year")) %>%
  # some old files from 2019 were found on the shuttles
  filter(year == "20") %>%
  mutate(Date = date(paste0(paste0("20",year),"-",month,"-",day))) %>%
  mutate(DateTime = (paste(Date,Time))) %>%
  mutate(DateTime = anytime(DateTime))%>%
  
  select(-month,day,year) %>%
  mutate(Site = factor(Site)) %>%
  select(FieldSeason, Site, SerialNumber, DateTime, Temp, DeployDate:DownloadComments) %>%
  filter(!is.na(Site)) %>%
  # transform
  transform(DeployDate = date(DeployDate)) %>%
  # Trim temp readings outside the deployment window
  mutate(year = year(DateTime)) %>%
  #filter(DateTime > DeployDate + days(1)) %>%
  #filter(DateTime < RetrieveDate - days(1)) %>%
  # When multiple paired loggers recorded temps, average them
  group_by(Site, DateTime) %>%
  summarize(Temp = mean(Temp)) 

## prep for plotting
# assign each site a ChannelType
# Read in the 2020 temp logger deployment data
channeltypes2020 <- read_excel(paste0(user,"Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/TempLoggerDeploy.xlsx"), sheet = "ChannelType") %>%
  transform(SerialNumber = as.character(SerialNumber))

# assign channel types to each 2020 site
temp20 <- left_join(temp20,channeltypes2020,by = "Site") %>%
  mutate(day = yday(DateTime),
         year = year(DateTime)) 

Annual plots by site

Examine annual datasets individually to visually scan for signs of logger malfunction or de-watering.

# plot original 2019 data
temp.faceted.2019 <- ggplot(data = temp19, aes(x = DateTime, y = Temp)) +
  geom_line() +
  facet_wrap(. ~ Site) +
  scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  ggtitle("2019 Original Water Temperature Data")
temp.faceted.2019

# plot original 2020 data
temp20 %>%
  ggplot(aes(DateTime,Temp)) +
  geom_point() +
  facet_wrap(.~Site) +
  scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  ggtitle("2020 Original Water Temperature Data")

Both 2019 and 2020 logger data have erroneous data. Solution: remove data between manually specified dates at specified sites.

Remove erroneous data from 2019 dataset and re-visualize results

# read in manually generated 2019 excise table
excisetemps2019 <- read_excel(paste0(user,"Dropbox/Chena_Data_2020/UAF_Data/2020_Analysis_And_Results/Analysis Scripts/chena_hobos_excise.xlsx"), sheet = "2019")

# create table of observations to be removed
excisetemps2019 <- left_join(temp19,excisetemps2019) %>%
  filter(DateTime >= DateTime_Start & DateTime <= DateTime_End)

# remove manually identified observations from overall dataset
temp19 <- anti_join(temp19,excisetemps2019)

# re-visualize results
temp.faceted.2019 <- ggplot(data = temp19, aes(x = DateTime, y = Temp)) +
  geom_point(size = 0.5) +
  facet_wrap(. ~ Site) +
  # scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  ggtitle("2019 Corrected Water Temperature Data")
temp.faceted.2019

Remove erroneous data from 2020 dataset and re-visualize results

# read in 2020 excise table
excisetemps2020 <- read_excel(paste0(user,"Dropbox/Chena_Data_2020/UAF_Data/2020_Analysis_And_Results/Analysis Scripts/chena_hobos_excise.xlsx"), sheet = "2020")

# create table of observations to be removed
excisetemps2020 <- left_join(temp20,excisetemps2020) %>%
  mutate(day = yday(DateTime)) %>%
  filter(day >= day_start & day <= day_end)

# remove manually identified observations from overall dataset
temp20 <- anti_join(temp20,excisetemps2020)

# plot corrected data
temp20 %>%
  ggplot(aes(DateTime,Temp)) +
  geom_point(size = 0.5) +
  facet_wrap(. ~ Site) +
  # scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  ggtitle("2020 Corrected Water Temperature Data")

Notes:

• Logger at “Unburned2” site was found exposed to air because creek had dried up (at unknown date). Estimated logger exposure date based on sudden increase in daily variability and excised data beyond that date.

• Partial 2020 results missing from Rock Creek, Upper Mastadon Burned, and Upper Mastadon burned due to HOBO shuttle time sync malfunction. Communicated with Onset customer service, who stated that the corrupted files were fix-able. Awaiting email response as of 10/15/20.

Results

Summary statistics

Calculate summary stats

# overall 2019-2020 dt
dt <- temp20 %>%
  select(-SerialNumber) %>%
  bind_rows(temp19) %>%
  mutate(year = year(DateTime)) %>%
  group_by(Site,ChannelType,Fire,year) %>%
  summarize(meanTemp = mean(Temp, na.rm = T),
            maxTemp = max(Temp, na.rm = T),
            minTemp = min(Temp, na.rm = T),
            sdTemp = sd(Temp, na.rm = T))

dt %>%
  datatable() %>%
  formatRound(columns=c("meanTemp","maxTemp","minTemp","sdTemp"), digits=2)

Burned vs Unburned Headwaters Tributaries

Compare 2020 burned to unburned sites

# set theme
plot_theme <- theme(
  axis.text = element_text(size = 14),
  axis.title = element_text(size = 16, face = "bold" ),
  plot.title = element_text(size = 16, face = "bold"),
  legend.background = element_blank(),
  legend.box.background = element_rect(colour = "black"))

# plot corrected data
# plot A
temp20 %>%
  filter(ChannelType == "Headwaters") %>%
  ggplot(aes(DateTime,Temp, color = Fire)) +
  geom_point(size = 0.5) +
  guides(color = guide_legend(override.aes = list(size = 5))) +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  xlab("") +
  scale_fill_discrete(name="Watershed\nCondition") +
  ggtitle("Headwaters Stream Temperatature") +
  plot_theme +
  theme(legend.justification=c(1,1), legend.position=c(1,1))

ggsave(paste0(user,"Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/2020_HOBO_Temperature/2020 Temperture Results/watertemps2020_1.jpg"))

# plot B
## set trendlines span
sp = 0.3
temp20 %>%
  filter(ChannelType == "Headwaters") %>%
  ggplot(aes(DateTime, Temp, fill = Fire)) + 
  geom_point(aes(color = Fire),size = 0.5, alpha = 0.01) +
  stat_smooth(data=subset(temp20,Fire=="Burned"),
                                  method = "lm", se = T, color = "red3", span = sp) +
  stat_smooth(data=subset(temp20,Fire=="Unburned"),
                                method = "loess", se = T, color = "green4", span = sp) +
  scale_color_manual(values=c(Burned="red3",Unburned = "green4")) +
  xlab("") +
  scale_y_continuous(name = "Water Temperature (˚C)") +
  scale_fill_discrete(name="Watershed\nCondition") +
  theme(legend.justification=c(0,1), legend.position=c(1,0)) +
  #guides(color = F) +
  ggtitle("Headwaters Stream Temperatature") +
  plot_theme

ggsave(paste0(user,"Dropbox/Chena_Data_2020/Chena Drift Project Data_SFR/UAF Chena Drift Data/Temperature/2020_HOBO_Temperature/2020 Temperture Results/watertemps2020_2.jpg"))

Further results pending

Working here to modify remaining code to markdown doc.

# specify location to save save results
# setwd("./Temperature")

# Plot the data
temp.allsites.2019 <- ggplot(data = temp, aes(x = DateTime, y = Temp, color = Site,
                                              linetype = ChannelType)) +
                               geom_line() +
  scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)")
temp.allsites.2019
ggsave("Temperature Results/2019 Temperature Results/TempAllSites.2019.pdf")
ggsave("Temperature Results/2019 Temperature Results/TempAllSites.2019.png")

# plot
temp.faceted.2019 <- ggplot(data = temp, aes(x = DateTime, y = Temp)) +
  geom_line() +
  facet_wrap(. ~ Site) +
  scale_x_datetime(name = "Date") +
  scale_y_continuous(name = "Water Temperature (˚C)")
temp.faceted.2019

# save
ggsave("Temperature Results/2019 Temperature Results/TempFaceted.2019.pdf")
ggsave("Temperature Results/2019 Temperature Results/TempFaceted.2019.png")



# save summary stats
write_csv(temp.summary, "Temperature Results/2019 Temperature Results/TempSummary2019.csv")