CDEC vs Gustavo's Livneh.Rmd

```{r}
#Loading packages to be used
library(tidyverse)
library(readr)
library(extrafont)
library(ggplot2)
library(ggthemes)
library(stringr)
library(dplyr)
library(grid)
library(gridExtra)
library(lubridate)
library(cder)

#Getting observed data from CDEC for each basin using CDEC's package
#Stanislaus
#SNS.obs <- cdec_query("SNS", "65", "M", "1950-01-01", "2013-12-01")
#SNS.obs$DateTime <- as.Date(SNS.obs$DateTime, format = "%Y/%m/%d")
#Merced
Obs <- cdec_query("MRC", 65, "M",  "1950-01-01", "2013-12-01")
Obs
Obs$DateTime <- as.Date(Obs$DateTime, format = "%Y/%m/%d")
#Tuolumne
#TLG.obs <- cdec_query("TLG", 65, "M", "1950-01-01", "2013-12-01")
#TLG.obs$DateTime <- as.Date(TLG.obs$DateTime, format = "%Y/%m/%d")
#Upper San Joaquin
#SJF.obs <- data.frame(cdec_query("SJF", 65, "M", "1950-01-01", "2013-12-01"))
#SJF.obs$DateTime <- as.Date(SJF.obs$DateTime, format = "%Y/%m/%d")

#Converting the flows in Acre-feet to million cubic meters
#SJF.obs$Value <- SJF.obs$Value/810.71318210885
#SNS.obs$Value <- SNS.obs$Value/810.71318210885

Obs$Value <- Obs$Value/810.71318210885
#TLG.obs$Value <- TLG.obs$Value/810.71318210885
Obs <- Obs %>%mutate(month = month(DateTime), year = year(DateTime)) #%>%
#  group_by(month, year)
Obs
```

```{r}
#Binding the 4 stations together into one object
#Obs <- rbind(SNS.obs, TLG.obs, SJF.obs, MRC.obs)
 #checking if it is all right

#Reading in Livneh data of each basin from each csv file
#SNS.sim <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/Livneh/SNS_natural_flow_monthly_mcm.csv")
Sim1 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb01_mcm.csv")
Sim2 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb02_mcm.csv") %>%
  mutate(flow2 = flow)
Sim3 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb03_mcm.csv") %>%
  mutate(flow3 = flow)
Sim4 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb04_mcm.csv") %>%
  mutate(flow4 = flow)
Sim5 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb05_mcm.csv") %>%
  mutate(flow5 = flow)
Sim6 <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/GCMs/Merced/historical/Livneh/total_runoff_sb06_mcm.csv") %>%
  mutate(flow6 = flow)

Sim <- cbind(Sim1, Sim2[3], Sim3[3], Sim4[3], Sim5[3], Sim6[3])
#TLG.sim <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/Livneh/TLG_natural_flow_monthly_mcm.csv")
#SJF.sim <- read_csv("C:/Users/gusta/Desktop/PhD/CERCWET/Livneh/sJF_natural_flow_monthly_mcm.csv")
Sim
#Binding the 4 stations together into one object
#Sim <- rbind(SNS, TLG, SJF, MRC)
Sim <- Sim %>%
  mutate(sum = rowSums(.[2:7]))#checking if it is all right
#Creating one tibble with the whole data set from both objects
```

```{r}
Sim2 <- Sim[c(1,8)] %>%
  mutate(monthyear = as.character(format(Date, "%Y-%m"))) %>%
          arrange(Date) %>% 
          group_by(monthyear) %>%
          summarise(Date=Date[1], flow = sum(sum))
Sim2
```

```{r}
fulldataset <- cbind(Sim2[2:3], Obs[7])
fulldataset
```

```{r}

#Creating function to calculate RSR, SNE and PBIAS
model.assess <- function(Sim, Obs) { #the inputs are numeric simulated and observed data
  if(!is.numeric(Sim)) { #verifies if the simulated data is really numeric
    stop('This function only works for numeric input!\n', #throws an error if it is not
         'You have provided an object of class: ', class(Sim)[1])
  }
  if(!is.numeric(Obs)) { #verifies if the observed data is really numeric
    stop('This function only works for numeric input!\n', #throws an error if it is not
         'You have provided an object of class: ', class(Obs)[1])
  }    
  rmse = sqrt( mean( (Sim - Obs)^2, na.rm = TRUE) ) #Formula to calculate RMSE
  RSR <- rmse / sd(Obs) #object producing RSR test from the RMSE formula
  PBIAS <- 100 *(sum((Sim - Obs)/sum(Obs), na.rm =TRUE)) #object producing PBIAS test
  NSE <- 1 - sum((Obs - Sim)^2)/sum((Obs - mean(Obs))^2, na.rm =TRUE) #object producing NSE test
  Values <- c(RSR, PBIAS, NSE) #Creates a column for the results of the tests
  Test <- c('RSR', 'PBIAS', 'NSE') #Creates a column for the name of the tests
  stats <- tibble(Test, Values) #Creates a tibble of the final products of the tests, giving numeric values for each test in this order: RSR, PBIAS, NSE
  return(stats) #returns a tibble as a result, which can be used in graphs or tables
}
```

```{r}
#Cleaning the data

#Selecting only the data that is interesting into
dataset <- fulldataset[c("StationID", "DateTime", "Value", "flow")] %>% #Columns with the name of the river, date, observed and simulated flows
  ungroup %>%
  rename(Sim_flow = flow, Obs_flow = Value, Date = DateTime) %>% #renaming columns to identify them easily
  mutate(Date = as.Date(Date, format = "%Y/%m/%d")) %>% #reading date as date
  mutate(Month = factor(format(as.Date(Date, format = "%Y/%m/%d"), "%B"))) #creating a Month column to facilitate creating facets by month
head(dataset)
```

```{r}

dataset <- fulldataset %>%
  mutate(Date = as.Date(Date, format = "%Y/%m/%d")) %>% #reading date as date
  mutate(Month = factor(format(as.Date(Date, format = "%Y/%m/%d"), "%B")))

dataset$Month <- factor(dataset$Month, #set months as factors, to be able to create facets by month
                        level = c("January", "February", "March", "April", "May", "June", "July",
                                  "August", "September", "October", "November", "December"))
```


```{r}
#Model assessment of each basin
#Merced
MRC.ma <- data.frame(model.assess(dataset$flow, dataset$Value))
MRC.ma
```
#Tuolumne
TLG.ma <- data.frame(model.assess(TLG.sim$flow, TLG.obs$Value))
TLG.ma
#Stanislaus
SNS.ma <- data.frame(model.assess(SNS.sim$flow, SNS.obs$Value))
SNS.ma 
#Upper San Joaquin
SJF.ma <- data.frame(model.assess(SJF.sim$flow, SJF.obs$Value))
SJF.ma

```{r}
#Creating figures

#Merced time series
MRC <-  dataset %>%
  filter(StationID == "MRC") %>% #selecting the data only from Merced
  ggplot() +  #plotting it
  theme_bw(base_size=12, base_family='Times New Roman') + #changing font to Times New Roman, 12pt, Bold
  geom_line(aes(x = Date, y = Obs_flow, colour = "#00AFBB"))+ #selecting observed data in greenish blue
  geom_line(aes(x = Date, y = Sim_flow, colour = "#FC4E07"))+ #selecting simulated data in orange
  scale_y_continuous(expand = c(0, 0))+ #removing blank space in the borders of the plotting area
  scale_x_date(date_breaks = "2 years", #setting the tick values of x axis
               date_labels = "%b/%Y", #setting label to written abbreviation of months and 4-digit years
               limits = as.Date(c('1985-01-01','2000-12-31')), #setting limits for x axis
               expand = c(0, 0)) + #removing empty space in the plotting area
  scale_color_identity(name = element_blank(), #no name for legend
                       labels = c("Observed (CDEC)","Simulated (Livneh)"), #select the variables for each line
                       guide = "legend")+ #creating legend
  labs(title = "Merced River", #creating title
       x = element_blank(), #empty x label
       y = "Monthly full natural Flow (mcm)") + #name of y axis
  theme(legend.position = "bottom", #position of the legend
        legend.direction = "horizontal", #direction of the legend
        legend.box.margin = margin(t = -19), #location of the legend
        plot.title = element_text(hjust = 0.5)) #centering the title

MRC + annotate("text", as.Date("1987-07-01"), y = 600, #adding where to add text 
               label = "RSR = 0.32\nPBIAS = 0.03\nNSE = 0.90",  #adding manually results of statistical tests
               collapse = "\n", hjust = 0, size=4, family= "Times New Roman") +

png("mer_timeseries.png", units ="in", width=8, height=5, res = 300) #creating the figure

#To use these commands for other rivers, just change the name of the object, the StationID in line 103, and the title in line 116, and the test results in line 125

#Merced full time series
#Same comands as the above, just changing date_breaks interval and date_lablels (lines 109-110), and removing data limits (line 111)
MRC.full <- dataset %>%
  filter(StationID == "MRC") %>% #selecting the data only from Merced
  ggplot() + #plotting it
  theme_bw(base_size=12, base_family='Times New Roman') + #change font to Times New Roman, 12pt, Bold
  geom_line(aes(x = Date, y = Obs_flow, colour = "#00AFBB"))+ #selecting observed data in greenish blue
  geom_line(aes(x = Date, y = Sim_flow, colour = "#FC4E07"))+ #selecting simulated data in orange
  scale_y_continuous(expand = c(0, 0))+ #removing empty space in the plotting area
  scale_x_date(date_breaks = "10 years", #setting the tick values of x axis
               date_labels = "%Y", #setting label to 4-digit years
               expand = c(0, 0)) + #removing empty space in the plotting area
  scale_color_identity(name = element_blank(), #no name for legend
                       labels = c("Observed (CDEC)","Simulated (Livneh)"), #select the variables for each line
                       guide = "legend")+ #creatinng legend
  labs(title = "Merced River", #creating title
       x = element_blank(), #empty x label
       y = "Monthly full natural flow (mcm)") + #name of x axis
  theme(legend.position = "bottom", #position of the legend
        legend.direction = "horizontal",#direction of the legend
        legend.box.margin = margin(t = -19),#location of the legend
        plot.title = element_text(hjust = 0.5)) #centering the title

MRC + annotate("text", as.Date("1987-07-01"), y = 600, #adding where to add text 
               label = "RSR = 0.32\nPBIAS = 0.03\nNSE = 0.90",  #adding manually results of statistical tests
               collapse = "\n", hjust = 0, size=4, family= "Times New Roman") +

png("mer_full_timeseries.png", units ="in", width=8, height=5, res = 300) #creating the figure

#To use these commands for other rivers, just change the name of the object, the StationID in line 133, and the title in line 145, and the test results in line 156


#Creating a figure with facets by month

#Merced full monthly time series
#Same comands as the above, just changing date_breaks interval (line 141), and without plotting test results (lines 155-157)
MRC.monthly <- dataset %>% 
  filter(StationID == "MRC") %>% #selecting the data only from Merced
  ggplot() + #plotting it
  theme_bw(base_size=12, base_family='Times New Roman') + #change font to Times New Roman, 12pt, Bold
  geom_line(aes(x = Date, y = Obs_flow, colour = "#00AFBB"))+ #selecting observed data in greenish blue
  geom_line(aes(x = Date, y = Sim_flow, colour = "#FC4E07"))+ #selecting simulated data in orange
  scale_y_continuous(expand = c(0, 0))+ #removing empty space in the plotting area
  scale_x_date(date_breaks = "15 years", #setting the tick values of x axis
               date_labels = "%Y", #setting label to 4-digit years
               expand = c(0, 0)) +#removing empty space in the plotting area
  scale_color_identity(name = element_blank(), #no name for legend
                       labels = c("Observed (CDEC)","Simulated (Livneh)"), #select the variables for each line
                       guide = "legend")+ #creatinng legend
  labs(title = "Merced River", #creating title
       x = element_blank(), #empty x label
       y = "Monthly full natural flow (mcm)") + #name of x axis
  theme(legend.position = "bottom", #position of the legend
        legend.direction = "horizontal",#direction of the legend
        legend.box.margin = margin(t = -19),#location of the legend
        plot.title = element_text(hjust = 0.5)) #centering the title

MRC.monthly + facet_wrap(~ Month, ncol=4, scales = "free") + #creating facets by month
  
png("mer_montlhy_timeseries.png", units ="in", width=8, height=5, res = 300) #creating the figure

#To use these commands for other rivers, just change the name of the object, the StationID in line 167


#Full monthly time series of all basins
#Same comands as the above, just removing filter by StationID (line 197) and removing the title (line 210)
all.basins <- dataset %>%
  filter(StationID == "MRC") %>% #selecting the data only from Merced
  ggplot() + #plotting it
  theme_bw(base_size=12, base_family='Times New Roman') + #change font to Times New Roman, 12pt, Bold
  geom_line(aes(x = Date, y = Obs_flow, colour = "#00AFBB"))+ #selecting observed data in greenish blue
  geom_line(aes(x = Date, y = Sim_flow, colour = "#FC4E07"))+ #selecting simulated data in orange
  scale_y_continuous(expand = c(0, 0))+ #removing empty space in the plotting area
  scale_x_date(date_breaks = "15 years", #setting the tick values of x axis
               date_labels = "%Y", #setting label to 4-digit years
               expand = c(0, 0)) + #removing empty space in the plotting area
  scale_color_identity(name = element_blank(), #no name for legend
                       labels = c("Observed (CDEC)","Simulated (Livneh)"), #select the variables for each line
                       guide = "legend")+ #creatinng legend
  labs(title = "Merced River", #creating title
       x = element_blank(), #empty x label
       y = "Monthly full natural flow (mcm)") + #name of x axis
  theme(legend.position = "bottom", #position of the legend
        legend.direction = "horizontal",#direction of the legend
        legend.box.margin = margin(t = -19),#location of the legend
        plot.title = element_text(hjust = 0.5)) #centering the title

all.basins + facet_wrap(~ Month, ncol=4, scales = "free") + #creating facets by month
  
  png("mer_montlhy_timeseries.png", units ="in", width=8, height=5, res = 300) #creating the figure

```