8. Calculating Monthly averages_Council_github.Rmd

---
title: "Council monthly averages" - Same as step 7 for making daily averages, but doing monthly averages (optional)
output: html_document
date: "2024-09-27"
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r}
rm(list = ls())
library(data.table)
library(ggplot2)
library(cowplot)
library(openair)
library(plotrix)
library(signal)
library(svMisc)
library(zoo)
library(stringr)
library(plyr)
library(viridis)
library(lubridate)
library(tidyverse)
library(gridExtra)
library(plotly)
library(RColorBrewer)
library(openair)
```


#read in cleaned and Ameriflux prepped half-hourly dataset
```{r}
#commented out - updated version following adding wind direction to RF in Github folder 
#df = fread(input = "C:/Users/kkent/Documents/Council Data/Council BASE gapfilling/US-NGC2_HH_201701010000_202309010000.csv", na.strings="-9999") 

df = fread(input = "C:/Users/kkent/Documents/Github Flux Network/Council BASE prepping/Council_BASE-data-prepping/US-NGC2_HH_201701010000_202309010000.csv", na.strings = "-9999")
           
##na.strings = c('-9999','NA','NaN','NAN','-7999')


```

#Create useable timestamp variable 
```{r}
df$TIMESTAMP_END = as.character(df$TIMESTAMP_END)
df$TIMESTAMP_START = as.character(df$TIMESTAMP_START)

df$TIMESTAMP_END = as.POSIXct(df$TIMESTAMP_END, tz="UTC", format = "%Y%m%d%H%M")
df$TIMESTAMP_START = as.POSIXct(df$TIMESTAMP_START, tz="UTC", format = "%Y%m%d%H%M")


```

#Create a new df for monthly averages 
```{r}
#using "date" instead of "day" in order to use the openAir and timeAverage packages/functions
df$date = as.Date(df$TIMESTAMP_END)

date = unique(df$date)
df_avg = as.data.frame(date)
```

#Create FC_night 
```{r}
# Filter CO2 Flux data by incoming shortwave values 
df <- df %>%
  mutate(FC_night = ifelse(SW_IN <= 0, FC, NA))

df <- df %>%
  mutate(FC_night_F = ifelse(SW_IN <= 0, FC_F, NA))
  
```

#Average by month 
```{r}
#Units are umol/m2/s for CO2, and nmol/m2/s for CH4
library(openair)
#average entire dataframe with tighter threshold 
df_monthly_avg <- as.data.frame(date)
df_monthly_avg <-timeAverage(df, avg.time = "month", data.thresh = 50)


#average dataframe with lighter threshold to get good nighttime data coverage
df_avg_night <- as.data.frame(date)
df_avg_night <-timeAverage(df, avg.time = "month", data.thresh = 10)


#add nighttime data back into the full dataset dataframe
df_monthly_avg$FC_night = df_avg_night$FC_night
df_monthly_avg$FC_night_F = df_avg_night$FC_night_F

#Optionally, print the resulting dataframe with monthly averages to screen / make sure it looks ok 
print(df_monthly_avg)


```




#Based on Dani's old code using a loop  

```{r}
# Step 1: Create the month column from TIMESTAMP_END
# df$month = format(as.Date(df$TIMESTAMP_END), "%Y-%m")  # Extract month in YYYY-MM format

# Create a dataframe with unique months
# unique_months = unique(df$month)
# df_monthly_avg = data.frame(month = unique_months, stringsAsFactors = FALSE)  # Initialize df_avg with unique months
# 
# # Step 2: Average by month to create monthly average dataframe
# for (i in 1:ncol(df)) {
#   if (class(df[[i]])[1] == 'numeric') {
#     
#     colname = colnames(df)[i]
# 
#     # Calculate monthly averages
#     monthly_avg_val <- aggregate(df[[colname]] ~ month, data = df, FUN = mean, na.action = na.omit)
#     
#     # Rename the average column
#     colnames(monthly_avg_val)[2] <- colname
#     
#     # Join the monthly averages to df_avg
#     df_monthly_avg = left_join(df_monthly_avg, monthly_avg_val, by = 'month')
#     
#   } else {
#     next
#   }
# }
# 


#print the resulting dataframe with monthly averages
#print(df_monthly_avg)

```


#check out the averages to make sure they look ok

```{r}


plot(df_monthly_avg$date, df_monthly_avg$FC_F)
plot(df_monthly_avg$date, df_monthly_avg$GPP_F)
plot(df_monthly_avg$date, df_monthly_avg$RECO_F)
plot(df_monthly_avg$date, df_monthly_avg$FC_night_F)

plot(df_monthly_avg$TS_3_1_1, df_monthly_avg$FC_night_F)
plot(df_monthly_avg$TS_3_1_1, df_monthly_avg$RECO_F)
plot(df_monthly_avg$TS_3_1_1, df_monthly_avg$GPP_F)



# ggplot(df_avg,aes(x=month,y=FC_F))+
#   geom_point()
# 
# ggplot(df_avg,aes(x=month,y=GPP_F))+
#   geom_point()
# 
# 
# ggplot(df_avg,aes(x=month,y=RECO_F))+
#   geom_point()


```

# Save Data

```{r}

write.csv(x = df_monthly_avg,file = 'C:/Users/kkent/Documents/Council Data/Council BASE gapfilling/council_monthly_AVG_gapfilled_clean_2017_2023_for analysis.csv',quote = F,row.names = F)




```