dissolved oxygen.R

#DO data

library(tidyverse)
library(lubridate)
library(DroughtData)
library(viridis)

RRI = read_csv("data/RRI_2008-2021_WQ.csv")
yrs = read_csv("data/yearassignments.csv")
#calculate means, mins, maxes

#Get rid of the data that's flagged bad
RRI2 = filter(RRI, !QAQCFlag %in% c("X")) %>%
  mutate(Date = date(DateTime)) %>%
  group_by(Month, Year, Julian, Date, Depth, Analyte) %>%
  summarize(Mean = mean(Value, na.rm = T), Max = max(Value, na.rm = T), Min = min(Value, na.rm = T)) %>%
  mutate(Year = case_when(Month %in% c("Oct",
                                       "Nov","Dec") ~ Year +1,
                        TRUE ~ Year))

#Add water year type and make water year type a factor
RRI2a = left_join(RRI2, yrs) %>%
  mutate(Yr_type = factor(Yr_type, levels = c("Critical", "Dry", "Below Normal", "Wet")))

#quick plot of all the data
ggplot(RRI2a, aes(x = Yr_type, y = Mean)) + geom_boxplot()+
  facet_grid(Depth~Analyte, scales = "free_y")

#Just the DO data
ggplot(filter(RRI2a, Analyte == "Dissolved Oxygen"), aes(x = Yr_type, y = Mean, fill = Yr_type)) + 
  geom_boxplot()+ drt_color_pal_yrtype()+
  facet_wrap(~Depth, scales = "free_y")+ theme_bw()

#filter out summer DO, when low DO is most problematic
summerDO = filter(RRI2a, Month %in% c("Sep", "Jun", "Jul", "Aug"),  Analyte == "Dissolved Oxygen")

ggplot(summerDO, aes(x = Yr_type, y = Mean)) + geom_boxplot()+
  facet_wrap(~Depth, scales = "free_y")

#Look for changes over time
ggplot(summerDO, aes(x = as.factor(Year), y = Min, fill = Yr_type)) + geom_boxplot()+
  facet_wrap(~Depth, scales = "free_y")

#changes in minimum over time
ggplot(summerDO, aes(x = Year, y = Min)) + geom_point(aes(color = Yr_type))+ geom_smooth(method = lm)+
  facet_wrap(~Depth, scales = "free_y")+ylab("Summer Daily MInimum DO mg/L")

#More changes over time, all days
ggplot(filter(RRI2a, Analyte == "Dissolved Oxygen"), aes(x = Date, y = Mean)) + geom_point()+
  facet_wrap(~Depth, scales = "free_y")

#color code by year type
ggplot(filter(RRI2a, Analyte == "Dissolved Oxygen"), aes(x = Date, y = Min, color = Yr_type)) + geom_point()+
  facet_wrap(~Depth)

#Let's pivot it so I can regress water quality against each otehr
RRwide = pivot_wider(RRI2a, id_cols = c(Month, Year, Julian, Date, Depth, Yr_type), 
                     names_from = Analyte, values_from = Mean)

#plot DO versus temperature
ggplot(RRwide, aes(x = `Temperature`, y = `Dissolved Oxygen`, color = Yr_type)) + geom_point()+
  ylab("Dissolved Oxygen mg/L")+ xlab("Temperature C")+ 
  scale_color_brewer(palette = "Dark2", name = "Water Year\nType")+ theme_bw()

#color code by fluorescence
ggplot(RRwide, aes(x = `Temperature`, y = `Dissolved Oxygen`, color = log(Fluorescence))) + geom_point()+
  scale_color_viridis()


ggplot(RRwide, aes(x = log(Fluorescence), y = `Dissolved Oxygen`, color = Temperature)) + geom_point()+
  scale_color_viridis()

# G - good 
# 
# X - bad
# 
# A - added (no data and  no date stamp, date time added for continuous date)
# 
# M - missing (no data, date stamp present)
# 
# U - unchecked 
# 
# Q - qualified