You are using Rweb1.04 on the server at www.nku.edu
Rweb:>
Rweb:> ############## RClimate Script: Compare JAXA_2007_2010.R ###################################
Rweb:> ## ##
Rweb:> ## Download and process JAXA daily SIE ##
Rweb:> ## Developed by D Kelly O'Day to demonstrate use of source() function for climate data ##
Rweb:> ## http:chartsgraphs.wordpress.com 6/26/10 ##
Rweb:> ############################################################################################
Rweb:> library(plotrix)
Rweb:> # ael
Rweb:> library(reshape)
Rweb:> # lea
Rweb:> options(digits=4)
Rweb:> par(mfrow=c(1,1))
Rweb:> par(ps=10); par(oma=c(0,0,0,0));par(mar=c(2,4,2,1))
Rweb:>
Rweb:> layout(matrix(c(1,2)), heights=c(2,1))
Rweb:>
Rweb:> #ael: this is the file, which is kept current (I pulled the data from here)
Rweb:> # link <- "http://www.ijis.iarc.uaf.edu/seaice/extent/plot.csv"
Rweb:> #ael: updated
Rweb:> #ael: link <-"/var/www/html/data/Rweb/RClimate/files/plot.csv.new"
Rweb:> link <-"/var/www/html/data/Rweb/RClimate/files/plot_extent_n_v2.csv"
Rweb:>
Rweb:> j_data <- read.csv(link, header = T,
+ colClasses = rep("numeric",17),
+ comment.char = "#", na.strings = c(-9999),
+ col.names = c("Month","Day","1980's Average","1990's Average","2000's Average","2002","2003","2004","2005","2006","2007","2008","2009","2010","2011","2012","2013","2014","2015","2016","2017","2018","2019","2020","2021","2022") )
Rweb:>
Rweb:> #ael -- this is my approach to getting in all the data, using the melt command
Rweb:> # which I just discovered today, Wed Dec 17 19:26:38 EST 2014:
Rweb:> jdl <- melt(j_data,id.var=c("Month","Day"),variable_name="Year")
Rweb:> names(jdl) <- c("mo", "day","yr","JASIE")
Rweb:> jdl$yr <- gsub(".s.Average", "", jdl$yr)
Rweb:> jdl$yr <- gsub("X", "", jdl$yr)
Rweb:> j_data <- jdl;
Rweb:> #lea
Rweb:>
Rweb:> ## Basic Date calcs
Rweb:> j_date <- paste(j_data$mo,j_data$day, j_data$yr, sep="/")
Rweb:> ## day of year
Rweb:> day_val <- strptime(j_date, "%m/%d/%Y")
Rweb:> doy <- day_val$yday +1
Rweb:> j_date <- as.Date(j_date, format="%m/%d/%Y")
Rweb:>
Rweb:> j_mo <- j_data$mo;
Rweb:> j_yr <- as.numeric(j_data$yr);
Rweb:> yr_frac <- as.numeric(j_yr + (j_mo-1)/12 + (j_data$day/30)/12)
Rweb:> # ael: needed to turn strings into numbers:
Rweb:> j_dy <- as.numeric(format(j_date,"%d"))
Rweb:>
Rweb:> ######################################################################
Rweb:> # ael: just writing out the current plot.csv file, for other files to use:
Rweb:> # I remove the 1980, 1990, and 2000 average data. So we start with daily data from 2002.
Rweb:> # jd <- data.frame(j_mo,j_dy,j_yr,j_data$JASIE)
Rweb:> # write.csv(jd, file = link_out, row.names = F)
Rweb:> # lea
Rweb:> ######################################################################
Rweb:>
Rweb:> ## data.frame of orig data & dates
Rweb:> jd <- data.frame(j_date, yr_frac, j_yr,j_mo,j_dy,doy, j_data$JASIE/10^6)
Rweb:> names(jd) <- c("dt", "yr_frac", "yr", "mo", "day","DOY", "sie")
Rweb:>
Rweb:>
Rweb:> ## JAXA's source file inludes all days of year with NA's for upcomming days. Remove all NAs
Rweb:> jds <- subset(jd, jd$sie != "NA")
Rweb:>
Rweb:> ## Get last day of jds data.frame
Rweb:> no_row <- nrow(jds)
Rweb:> cur_dt <- jds[no_row,1]
Rweb:> cur_day <- as.numeric(format(cur_dt, "%d"))
Rweb:> cur_mo <- as.numeric(format(cur_dt, "%m"))
Rweb:> cur_yr <- as.numeric(format(cur_dt, "%y"))
Rweb:>
Rweb:> ## subset jds to make data.frames for all data for 2007 & 2010
Rweb:> jd_2010 <- subset(jds, jds$yr == 2010)
Rweb:> n_10 <- nrow(jd_2010)
Rweb:> peak_2010 <- max(jd_2010$sie, na.rm=T)
Rweb:> peak_dt_2010 <- format(jd_2010$dt[which(jd_2010$sie== peak_2010)], "%m/%d")
Rweb:> peak_doy_2010 <- jd_2010$DOY[which(jd_2010$sie== peak_2010)]
Rweb:> pit_2010 <- min(jd_2010$sie, na.rm=T)
Rweb:> pit_dt_2010 <- format(jd_2010$dt[which(jd_2010$sie== pit_2010)], "%m/%d")
Rweb:> pit_doy_2010 <- jd_2010$DOY[which(jd_2010$sie== pit_2010)]
Rweb:> last_dt <- format(jd_2010$dt[n_10], "%m/%d")
Rweb:> last_doy <- jd_2010[n_10,6]
Rweb:> last_2010 <- jd_2010[n_10,7]
Rweb:> decr_2010 <- peak_2010 - last_2010
Rweb:>
Rweb:> jd_2007 <- subset(jds, jds$yr== 2007 & jds$DOY <= last_doy)
Rweb:> n_07 <- nrow(jd_2007)
Rweb:> peak_2007 <- max(jd_2007$sie, na.rm=T)
Rweb:> peak_dt_2007 <- format(jd_2007$dt[which(jd_2007$sie== peak_2007)], "%m/%d")
Rweb:> peak_doy_2007 <- jd_2007$DOY[which(jd_2007$sie== peak_2007)]
Rweb:> pit_2007 <- min(jd_2007$sie, na.rm=T)
Rweb:> pit_dt_2007 <- format(jd_2007$dt[which(jd_2007$sie== pit_2007)], "%m/%d")
Rweb:> pit_doy_2007 <- jd_2007$DOY[which(jd_2007$sie== pit_2007)]
Rweb:> last_2007 <- jd_2007[n_07,7]
Rweb:> decr_2007 <- peak_2007 - last_2007
Rweb:>
Rweb:> jd_2012 <- subset(jds, jds$yr == 2012)
Rweb:> n_2012 <- nrow(jd_2012)
Rweb:> peak_2012 <- max(jd_2012$sie, na.rm=T)
Rweb:> peak_dt_2012 <- format(jd_2012$dt[which(jd_2012$sie== peak_2012)], "%m/%d")
Rweb:> peak_doy_2012 <- jd_2012$DOY[which(jd_2012$sie== peak_2012)]
Rweb:> pit_2012 <- min(jd_2012$sie, na.rm=T)
Rweb:> pit_dt_2012 <- format(jd_2012$dt[which(jd_2012$sie== pit_2012)], "%m/%d")
Rweb:> pit_doy_2012 <- jd_2012$DOY[which(jd_2012$sie== pit_2012)]
Rweb:> last_2012 <- jd_2012[n_2012,7]
Rweb:> decr_2012 <- peak_2012 - last_2012
Rweb:>
Rweb:> jd_2016 <- subset(jds, jds$yr == 2016)
Rweb:> n_2016 <- nrow(jd_2016)
Rweb:> peak_2016 <- max(jd_2016$sie, na.rm=T)
Rweb:> peak_dt_2016 <- format(jd_2016$dt[which(jd_2016$sie== peak_2016)], "%m/%d")
Rweb:> peak_doy_2016 <- jd_2016$DOY[which(jd_2016$sie== peak_2016)]
Rweb:> pit_2016 <- min(jd_2016$sie, na.rm=T)
Rweb:> pit_dt_2016 <- format(jd_2016$dt[which(jd_2016$sie== pit_2016)], "%m/%d")
Rweb:> pit_doy_2016 <- jd_2016$DOY[which(jd_2016$sie== pit_2016)]
Rweb:> last_2016 <- jd_2016[n_2016,7]
Rweb:> decr_2016 <- peak_2016 - last_2016
Rweb:>
Rweb:> jd_2021 <- subset(jds, jds$yr == 2021)
Rweb:> n_2021 <- nrow(jd_2021)
Rweb:> peak_2021 <- max(jd_2021$sie, na.rm=T)
Rweb:> peak_dt_2021 <- format(jd_2021$dt[which(jd_2021$sie== peak_2021)], "%m/%d")
Rweb:> peak_doy_2021 <- jd_2021$DOY[which(jd_2021$sie== peak_2021)]
Rweb:> pit_2021 <- min(jd_2021$sie, na.rm=T)
Rweb:> pit_dt_2021 <- format(jd_2021$dt[which(jd_2021$sie== pit_2021)], "%m/%d")
Rweb:> pit_doy_2021 <- jd_2021$DOY[which(jd_2021$sie== pit_2021)]
Rweb:> last_2021 <- jd_2021[n_2021,7]
Rweb:> decr_2021 <- peak_2021 - last_2021
Rweb:>
Rweb:> jd_1980 <- subset(jds, jds$yr == 1980)
Rweb:> n_1980 <- nrow(jd_1980)
Rweb:> peak_1980 <- max(jd_1980$sie, na.rm=T)
Rweb:> peak_dt_1980 <- format(jd_1980$dt[which(jd_1980$sie== peak_1980)], "%m/%d")
Rweb:> peak_doy_1980 <- jd_1980$DOY[which(jd_1980$sie== peak_1980)]
Rweb:> pit_1980 <- min(jd_1980$sie, na.rm=T)
Rweb:> pit_dt_1980 <- format(jd_1980$dt[which(jd_1980$sie== pit_1980)], "%m/%d")
Rweb:> pit_doy_1980 <- jd_1980$DOY[which(jd_1980$sie== pit_1980)]
Rweb:> last_1980 <- jd_1980[n_1980,7]
Rweb:> decr_1980 <- peak_1980 - last_1980
Rweb:>
Rweb:> jd_1990 <- subset(jds, jds$yr == 1990)
Rweb:> n_1990 <- nrow(jd_1990)
Rweb:> peak_1990 <- max(jd_1990$sie, na.rm=T)
Rweb:> peak_dt_1990 <- format(jd_1990$dt[which(jd_1990$sie== peak_1990)], "%m/%d")
Rweb:> peak_doy_1990 <- jd_1990$DOY[which(jd_1990$sie== peak_1990)]
Rweb:> pit_1990 <- min(jd_1990$sie, na.rm=T)
Rweb:> pit_dt_1990 <- format(jd_1990$dt[which(jd_1990$sie== pit_1990)], "%m/%d")
Rweb:> pit_doy_1990 <- jd_1990$DOY[which(jd_1990$sie== pit_1990)]
Rweb:> last_1990 <- jd_1990[n_1990,7]
Rweb:> decr_1990 <- peak_1990 - last_1990
Rweb:>
Rweb:> ### Day of year axis setup
Rweb:> ## Set up basic day of year vectors (mon_names, 1st day of mon)
Rweb:> mon_names <- c("Jan", "Feb", "Mar", "April", "May", "June", "July", "Aug", "Sept", "Oct","Nov","Dec")
Rweb:> mon_doy <- c(1,32,60,91,121,151,182,213,244,274,305,335,365)
Rweb:> mon_pos <- c(16, 46, 75, 106,135, 165, 200, 228, 255, 289, 320, 355)
Rweb:>
Rweb:> y_min <- (min(jd_2010$sie) )*0.6
Rweb:> y_max <- (max(jd_2010$sie))* 1.2
Rweb:>
Rweb:> # plot titles
Rweb:> main_title <- paste("JAXA 2007/10/12/16/21 Sea Ice Extent\nWith averages for decades 1980 and 1990\n",last_dt," Status", sep="")
Rweb:> y_title <- "JAXA Arctic Sea Ice Extent - millions km^2"
Rweb:>
Rweb:> # Plot
Rweb:> plot(
+ jd_2010$DOY, jd_2010$sie,
+ type="l", col = "red",
+ main=main_title,
+ cex.main=0.85,
+ ylab = y_title, las=1,
+ xlim = c(1, last_doy+2),
+ ylim = c(y_min, y_max ),
+ axes=F,
+ xaxs="i",
+ yaxs="i",
+ las=1)
Rweb:>
Rweb:> # custom x & y axes
Rweb:> axis(side=1, at=mon_doy, labels=F, xaxs="i")
Rweb:> axis(side=1, at=mon_pos, labels=mon_names, tick=F, line=F, xaxs="i")
Rweb:> axis(side=2, at=NULL, labels=T, yaxs="i", las=1)
Rweb:> abline(v=1) # need full length y axis line
Rweb:>
Rweb:> # Add 2007 trend & 2007-2010 peak sie
Rweb:> points(jd_2007$DOY, jd_2007$sie, type="l", col = "blue")
Rweb:> points(peak_doy_2010, peak_2010, type = "p", col = "red", pch=16, cex=1.25)
Rweb:> points(peak_doy_2007, peak_2007, type = "p", col = "blue", pch=16, cex=1.25)
Rweb:> points(pit_doy_2007, pit_2007, col = "blue", type="p", pch=18,cex=1.25)
Rweb:> points(pit_doy_2007, pit_2010, col = "red", type="p", pch=18, cex=1.25)
Rweb:>
Rweb:> # Add 2012 trend
Rweb:> points(jd_2012$DOY, jd_2012$sie, type="l", col = "green")
Rweb:> points(peak_doy_2012, peak_2012, type = "p", col = "green", pch=16, cex=1.25)
Rweb:> points(pit_doy_2012, pit_2012, col = "green", type="p", pch=18,cex=1.25)
Rweb:>
Rweb:> # Add 2016 trend
Rweb:> points(jd_2016$DOY, jd_2016$sie, type="l", col = "yellow")
Rweb:> points(peak_doy_2016, peak_2016, type = "p", col = "yellow", pch=16, cex=1.25)
Rweb:> points(pit_doy_2016, pit_2016, col = "yellow", type="p", pch=18,cex=1.25)
Rweb:>
Rweb:> # Add 2021 trend
Rweb:> points(jd_2021$DOY, jd_2021$sie, type="l", col = "magenta")
Rweb:> points(peak_doy_2021, peak_2021, type = "p", col = "magenta", pch=16, cex=1.25)
Rweb:> points(pit_doy_2021, pit_2021, col = "magenta", type="p", pch=18,cex=1.25)
Rweb:>
Rweb:> # Add 1980 trend
Rweb:> points(jd_1980$DOY, jd_1980$sie, type="l", col = "black")
Rweb:> points(peak_doy_1980, peak_1980, type = "p", col = "black", pch=16, cex=1.25)
Rweb:> points(pit_doy_1980, pit_1980, col = "black", type="p", pch=18,cex=1.25)
Rweb:>
Rweb:> # Add 1990 trend
Rweb:> points(jd_1990$DOY, jd_1990$sie, type="l", col = "grey")
Rweb:> points(peak_doy_1990, peak_1990, type = "p", col = "grey", pch=16, cex=1.25)
Rweb:> points(pit_doy_1990, pit_1990, col = "grey", type="p", pch=18,cex=1.25)
Rweb:>
Rweb:> # create peak notes for legend
Rweb:> peak_note_2007 <- paste("2007 Peak @ ", signif(peak_2007, 4)," on ", peak_dt_2007, sep="")
Rweb:> peak_note_2010 <- paste("2010 Peak @ ", signif(peak_2010, 4)," on ", peak_dt_2010, sep="")
Rweb:> peak_note_2012 <- paste("2012 Peak @ ", signif(peak_2012, 4)," on ", peak_dt_2012, sep="")
Rweb:> peak_note_2016 <- paste("2016 Peak @ ", signif(peak_2016, 4)," on ", peak_dt_2016, sep="")
Rweb:> peak_note_2021 <- paste("2021 Peak @ ", signif(peak_2021, 4)," on ", peak_dt_2021, sep="")
Rweb:> peak_note_1980 <- paste("1980 Peak @ ", signif(peak_1980, 4)," on ", peak_dt_1980, sep="")
Rweb:> peak_note_1990 <- paste("1990 Peak @ ", signif(peak_1990, 4)," on ", peak_dt_1990, sep="")
Rweb:>
Rweb:> # create pit notes for legend
Rweb:> pit_note_2007 <- paste("2007 Pit @ ", signif(pit_2007, 4)," on ", pit_dt_2007, sep="")
Rweb:> pit_note_2010 <- paste("2010 Pit @ ", signif(pit_2010, 4)," on ", pit_dt_2010, sep="")
Rweb:> pit_note_2012 <- paste("2012 Pit @ ", signif(pit_2012, 4)," on ", pit_dt_2012, sep="")
Rweb:> pit_note_2016 <- paste("2016 Pit @ ", signif(pit_2016, 4)," on ", pit_dt_2016, sep="")
Rweb:> pit_note_2021 <- paste("2021 Pit @ ", signif(pit_2021, 4)," on ", pit_dt_2021, sep="")
Rweb:> pit_note_1980 <- paste("1980 Pit @ ", signif(pit_1980, 4)," on ", pit_dt_1980, sep="")
Rweb:> pit_note_1990 <- paste("1990 Pit @ ", signif(pit_1990, 4)," on ", pit_dt_1990, sep="")
Rweb:>
Rweb:> # legend
Rweb:> legend(40, 13, c(
+ peak_note_2007, peak_note_2010, peak_note_2012, peak_note_2016, peak_note_2021, peak_note_1980, peak_note_1990
+ ),
+ col = c("blue","red","green","yellow","magenta", "black","grey"),
+ text.col = "black",
+ lty= c(0, 0, 0, 0, 0, 0, 0),
+ pch= c(16,16,16,16,16,16,16),
+ pt.cex=c( 1, 1, 1, 1, 1, 1, 1),
+ merge = F, bg = "white", bty="o", box.col = "white", cex = .8)
Rweb:>
Rweb:> legend(40, 8, c("2007 Trend", "2010 Trend", "2012 Trend", "2016 Trend", "2016 Trend", "1980 Average", "1990 Average"
+ ),
+ col = c("blue","red","green","yellow","magenta", "black","grey"),
+ text.col = "black",
+ lty= c(1,1,1,1,1,1,1),
+ pch= c(0,0,0,0,0,0,0),
+ pt.cex=c(0,0,0,0,0,0,0),
+ merge = F, bg = "white", bty="o", box.col = "white", cex = .8)
Rweb:>
Rweb:> legend(120, 8, c(
+ pit_note_2007, pit_note_2010, pit_note_2012, pit_note_2016, pit_note_2021, pit_note_1980, pit_note_1990
+ ),
+ col = c("blue","red","green","yellow","magenta", "black","grey"),
+ text.col = "black",
+ lty= c(0, 0, 0, 0, 0, 0, 0),
+ pch= c(18,18,18,18,18,18,18),
+ pt.cex=c(1, 1, 1, 1, 1, 1, 1),
+ merge = F, bg = "white", bty="o", box.col = "white", cex = .8)
Rweb:>
Rweb:> delta <- last_2010-last_2007
Rweb:> delta_note <- paste(last_dt, " Delta: \n2010 - 2007 =\n ", signif(delta,3), " million km^2", sep="")
Rweb:> text(340, 16, delta_note, cex=0.75, adj=0.5)
Rweb:>
Rweb:> ## function to calc slope last x days
Rweb:>
Rweb:> tr_func <- function(n) {
+ last_n_2007 <- subset(jd_2007, jd_2007$DOY > last_doy-n)
+ last_n_2010 <- subset(jd_2010, jd_2010$DOY > last_doy-n)
+
+ lm_2007 <- lm( sie ~ DOY, data = last_n_2007)
+ tr_2007 <- coef(lm_2007)[2]
+
+ lm_2010 <- lm( sie ~ DOY, data = last_n_2010)
+ tr_2010 <- coef(lm_2010)[2]
+ my_df <- data.frame(as.integer(n), signif(tr_2007, 3), signif(tr_2010,3), row.names=NULL)
+ return(my_df)
+ }
Rweb:>
Rweb:> x <- c(2,3,5,7,10,20,30,60,90)
Rweb:> tr_df <- data.frame()
Rweb:> how_many <- length(x)
Rweb:> for (i in 1:how_many) {
+ y <- x[i]
+ results <- tr_func(y)
+ tr_df[i,1] <- y #results[1]
+ tr_df[i,2] <- signif(results[2], 3)
+ tr_df[i,3] <- signif(results[3] ,3)
+ tr_df[i,4] <- signif(results[3]/results[2],3)
+ }
Rweb:>
Rweb:> names(tr_df) <- c("Prev days", "2007", "2010", "2010/2007")
Rweb:>
Rweb:> par(mar=c(0,0,0.5,0))
Rweb:>
Rweb:> plot(x=c(1,2,3), y= c(1,2,3), type="n", axes=F,ann=F, xlim=c(0,3), ylim=c(0,3))
Rweb:>
Rweb:>
Rweb:> table_title <- paste(last_dt, ": Arctic SIE Decline Rates for Number of Previous Days", sep="")
Rweb:> addtable2plot(0.825,2.95,tr_df,bty="o",display.rownames=F,hlines=T,cex=0.95,
+ title=table_title, xjust=0, yjust=0)
Rweb:>
Rweb:>
