Project

General

Profile

Download (8.21 KB) Statistics
| Branch: | Revision:
1
### Figures and tables for MOD09 Cloud Manuscript
2

    
3
setwd("~/acrobates/adamw/projects/cloud/")
4

    
5

    
6
## libraries
7
library(rasterVis)
8
library(latticeExtra)
9
library(xtable)
10
library(reshape)
11
library(caTools)
12

    
13
## read in data
14
cld=read.csv("data/NDP026D/cld.csv")
15
cldm=read.csv("data/NDP026D/cldm.csv")
16
cldy=read.csv("data/NDP026D/cldy.csv")
17
clda=read.csv("data/NDP026D/clda.csv")
18
st=read.csv("data/NDP026D/stations.csv")
19

    
20

    
21
## add lulc factor information
22
require(plotKML); data(worldgrids_pal)  #load IGBP palette
23
IGBP=data.frame(ID=0:16,col=worldgrids_pal$IGBP[-c(18,19)],stringsAsFactors=F)
24
IGBP$class=rownames(IGBP);rownames(IGBP)=1:nrow(IGBP)
25

    
26
## month factors
27
cld$month2=factor(cld$month,labels=month.name)
28
cldm$month2=factor(cldm$month,labels=month.name)
29

    
30
coordinates(st)=c("lon","lat")
31
projection(st)=CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs")
32

    
33
##make spatial object
34
cldms=cldm
35
coordinates(cldms)=c("lon","lat")
36
projection(cldms)=CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs")
37

    
38
##make spatial object
39
cldys=cldy
40
coordinates(cldys)=c("lon","lat")
41
projection(cldys)=CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs")
42

    
43
#### Evaluate MOD35 Cloud data
44
mod09=brick("data/mod09.nc")
45
mod09c=brick("data/mod09_clim_mean.nc",varname="CF");names(mod09c)=month.name
46
mod09a=brick("data/mod09_clim_mac.nc",varname="CF_annual")#;names(mod09c)=month.name
47

    
48
## derivatives
49
if(!file.exists("data/mod09_std.nc")) {
50
  system("cdo -chname,CF,CFmin -timmin data/mod09_clim_mean.nc data/mod09_min.nc")
51
  system("cdo -chname,CF,CFmax -timmax data/mod09_clim_mean.nc data/mod09_max.nc")
52
  system("cdo -chname,CF,CFsd -timstd data/mod09_clim_mean.nc data/mod09_std.nc")
53
  system("cdo -f nc2 merge data/mod09_std.nc data/mod09_min.nc data/mod09_max.nc data/mod09_metrics.nc") 
54
}
55

    
56
mod09min=raster("data/mod09_metrics.nc",varname="CFmin")
57
mod09max=raster("data/mod09_metrics.nc",varname="CFmax")
58
mod09sd=raster("data/mod09_metrics.nc",varname="CFsd")
59
mod09mean=raster("data/mod09_clim_mac.nc")
60

    
61

    
62
names(mod09d)=c("Mean","Minimum","Maximum","Standard Deviation")
63

    
64
plot(mod09a,layers=1,margin=F,maxpixels=100)
65

    
66
## calculated differences
67
cldm$dif=cldm$mod09-cldm$cld
68
clda$dif=clda$mod09-clda$cld
69

    
70
## read in global coasts for nice plotting
71
library(maptools)
72

    
73
data(wrld_simpl)
74
coast <- unionSpatialPolygons(wrld_simpl, rep("land",nrow(wrld_simpl)), threshold=5)
75
coast=as(coast,"SpatialLines")
76

    
77

    
78
## Figures
79
n=100
80
  at=seq(0,100,length=n)
81
colr=colorRampPalette(c("black","green","red"))
82
cols=colr(n)
83

    
84

    
85
pdf("output/validation.pdf",width=11,height=8.5)
86

    
87
## 4-panel maps
88
#- Annual average
89
levelplot(mod09a,col.regions=colr(100),cuts=100,at=seq(0,100,len=100),colorkey=list(space="bottom",adj=1),
90
  margin=F,maxpixels=1e6,ylab="Latitude",xlab="Longitude",useRaster=T)+
91
  layer(sp.lines(coast,col="black"),under=F)
92
#- Monthly minimum
93
#- Monthly maximum
94
#- STDEV or Min-Max
95

    
96

    
97
### maps of mod09 and NDP
98
## map of stations
99
p_mac=levelplot(mod09a,col.regions=colr(100),cuts=100,at=seq(0,100,len=100),colorkey=list(space="bottom",adj=1),
100
  margin=F,maxpixels=1e6,ylab="Latitude",xlab="Longitude",useRaster=T)+
101
  layer(panel.xyplot(lon,lat,pch=16,cex=.3,col="black"),data=data.frame(coordinates(st)))+
102
  layer(sp.lines(coast,col="black"),under=F)
103

    
104
p_mace=xyplot(lat~dif,data=cldm[cldm$lat>-60,],panel=function(x,y,subscripts=T){
105
  x2=x[order(y)]
106
  y2=y[order(y)]
107
  win=8000
108
  Q50=runquantile(x2,win,probs=c(.25,.5,.75))
109
  ## polygon
110
  panel.polygon(c(Q50[,1],rev(Q50[,3])),c(y2,rev(y2)),type="l",col=grey(.8))
111
### hist
112
  n=150
113
  xbins=seq(-70,50,len=n)
114
  ybins=seq(-60,90,len=n)
115
  tb=melt(as.matrix(table(
116
    x=cut(x,xbins,labels=xbins[-1]),
117
    y=cut(y,ybins,labels=ybins[-1]))))
118
  qat=unique(tb$value)
119
  print(qat)
120
  panel.levelplot(tb$x,tb$y,tb$value,at=qat,col.regions=c("transparent",hmcols(length(qat))),subscripts=1:nrow(tb))
121
###
122
#  panel.xyplot(x,y,pch=16,cex=.1,col=)
123
#  cuts=cut(y,lats,labels=lats[-10])
124
#  qs=do.call(rbind,tapply(x,cuts,quantile,c(.25,.50,.75),na.rm=T))
125
  colnames(qs)=c("Q25","Q50","Q75")
126
  panel.lines(Q50[,1],y2,type="l",col=grey(.5))
127
  panel.lines(Q50[,2],y2,type="l",col="black")
128
  panel.lines(Q50[,3],y2,type="l",col=grey(.5))
129
},asp=1,xlab="Difference (MOD09-Observed)")+layer(panel.abline(v=0,lty="dashed",col="red"))
130

    
131
print(p_mac,position=c(0,0,.75,1),more=T)
132
print(p_mace,position=c(0.75,0,1,1))
133

    
134

    
135
p1=c("MODIS Cloud Frequency and NDP-026D Validation Stations"=p_mac,"Difference (MOD09-NDP026D)"=p_mace,x.same=F,y.same=T,layout=c(2,1))
136
resizePanels(p1,w=c(.75,.25))
137

    
138
quantile(cldm$dif,seq(0,1,len=6),na.rm=T)
139
at=c(-70,-50,-25,-10,-5,0,5,10,25,50,70)
140
bwr=colorRampPalette(c("blue","grey","red"))
141
xyplot(lat~lon|month2,data=cldm,groups=cut(cldm$dif,at),
142
       par.settings=list(superpose.symbol=list(col=bwr(length(at)-1))),pch=16,cex=.25,
143
       auto.key=list(space="right",title="Difference\n(MOD09-NDP026D)",cex.title=1),asp=1,
144
       main="NDP-026D Cloud Climatology Stations",ylab="Latitude",xlab="Longitude")+
145
  layer(sp.lines(coast,col="black",lwd=.1),under=F)
146

    
147

    
148

    
149
### heatmap of mod09 vs. NDP for all months
150
hmcols=colorRampPalette(c("grey","blue","red"))
151
#hmcols=colorRampPalette(c(grey(.8),grey(.3),grey(.2)))
152
tr=c(0,120)
153
colkey <- draw.colorkey(list(col = hmcols(tr[2]), at = tr[1]:tr[2],height=.25))
154

    
155
xyplot(cld~mod09,data=cld[cld$Nobs>10,],panel=function(x,y,subscripts){
156
  n=150
157
  bins=seq(0,100,len=n)
158
  tb=melt(as.matrix(table(
159
    x=cut(x,bins,labels=bins[-1]),
160
    y=cut(y,bins,labels=bins[-1]))))
161
  qat=tr[1]:tr[2]#unique(tb$value)
162
  print(qat)
163
  panel.levelplot(tb$x,tb$y,tb$value,at=qat,col.regions=c("transparent",hmcols(length(qat))),subscripts=subscripts)
164
  },asp=1,scales=list(at=seq(0,100,len=6)),ylab="NDP Mean Cloud Amount (%)",xlab="MOD09 Cloud Frequency (%)",
165
       legend= list(right = list(fun = colkey,title="Station Count")))+
166
  layer(panel.abline(0,1,col="black",lwd=2))
167
#  layer(panel.ablineq(lm(y ~ x), r.sq = TRUE,at = 0.6,pos=1, offset=22,digits=2,col="blue"), style = 1)
168

    
169

    
170
xyplot(cld~mod09|month2,data=cld[cld$Nobs>50,],panel=function(x,y,subscripts){
171
  n=50
172
  bins=seq(0,100,len=n)
173
  tb=melt(as.matrix(table(
174
    x=cut(x,bins,labels=bins[-1]),
175
    y=cut(y,bins,labels=bins[-1]))))
176
  qat=unique(tb$value)
177
  qat=0:78
178
  qat=tr[1]:tr[2]#unique(tb$value)
179
  panel.levelplot(tb$x,tb$y,tb$value,at=qat,col.regions=c("transparent",hmcols(length(qat))),subscripts=1:nrow(tb))
180
  panel.abline(0,1,col="black",lwd=2)
181
#  panel.ablineq(lm(y ~ x), r.sq = TRUE,at = 0.6,pos=1, offset=0,digits=2,col="blue")
182
  panel.text(70,10,bquote(paste(R^2,"=",.(round(summary(lm(y ~ x))$r.squared,2)))),cex=1.2)
183
},asp=1,scales=list(at=seq(0,100,len=6),useRaster=T,colorkey=list(width=.5,title="Number of Stations")),
184
          ylab="NDP Mean Cloud Amount (%)",xlab="MOD09 Cloud Frequency (%)",
185
              legend= list(right = list(fun = colkey)))+ layer(panel.abline(0,1,col="black",lwd=2))
186

    
187

    
188
## Monthly Climatologies
189
for(i in 1:2){
190
 p1=xyplot(cld~mod09|month2,data=cldm,cex=.2,pch=16,subscripts=T,ylab="NDP Mean Cloud Amount",xlab="MOD09 Cloud Frequency (%)")+
191
  layer(panel.lines(1:100,predict(lm(y~x),newdata=data.frame(x=1:100)),col="green"))+
192
  layer(panel.lines(1:100,predict(lm(y~x+I(x^2)),newdata=data.frame(x=1:100)),col="blue"))+
193
  layer(panel.abline(0,1,col="red"))
194
    if(i==2){
195
     p1=p1+layer(panel.segments(mod09[subscripts],cld[subscripts]-cldsd[subscripts],mod09[subscripts],cld[subscripts]+cldsd[subscripts],subscripts=subscripts,col="grey"),data=cldm,under=T,magicdots=T)
196
     p1=p1+layer(panel.segments(mod09[subscripts]-mod09sd[subscripts],cld[subscripts],mod09[subscripts]+mod09sd[subscripts],cld[subscripts],subscripts=subscripts,col="grey"),data=cldm,under=T,magicdots=T) 
197
       }
198
print(p1)
199
}
200

    
201
bwplot(lulcc~dif,data=cldm,horiz=T,xlab="Difference (MOD09-Observed)",varwidth=T,notch=T)+layer(panel.abline(v=0))
202

    
203

    
204
dev.off()
205

    
206

    
207
summary(lm(cld~mod09,data=cld))
208

    
209
## explore validation error
210
cldm$lulcc=as.factor(IGBP$class[match(cldm$lulc,IGBP$ID)])
211

    
212
## Table of RMSE's by lulc by month
213
lulcrmsel=ddply(cldm,c("month","lulc"),function(x) c(count=nrow(x),rmse=sqrt(mean((x$mod09-x$cld)^2,na.rm=T))))
214
lulcrmsel=lulcrmsel[!is.na(lulcrmsel$lulc),]
215
lulcrmsel$lulcc=as.factor(IGBP$class[match(lulcrmsel$lulc,IGBP$ID)])
216

    
217
lulcrmse=cast(lulcrmsel,lulcc~month,value="rmse")
218
lulcrmse
219
print(xtable(lulcrmse,digits=1),"html")
220
  
221
levelplot(rmse~lulc*month,data=lulcrmsel,col.regions=heat.colors(20))
222

    
223

    
224
### Linear models
225
summary(lm(dif~as.factor(lulc)+lat+month2,data=cldm))
226

    
(20-20/50)