-
Notifications
You must be signed in to change notification settings - Fork 0
/
UseApplyModelFunctionsVeg.R
232 lines (165 loc) · 8.61 KB
/
UseApplyModelFunctionsVeg.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
# General Code for applying shrub nutrient models
# Try to avoid R-specific functions
library(spatstat)
library(maptools)
setwd( "C:/Users/elm26/Documents/SERDPproject")
source("SoilProbeSampling/CompassFunctions")
source("SoilProbeSampling/RasterGeometry")
source("SoilProbeSampling/ApplyModelFunctions_stemMeasure.R")
# PICK A DESERT
#DESERT="Sonoran"
DESERT="Mojave"
YEAR="2013Xtra"
##########################################################
# Build Null Grid ------
if (DESERT=="Sonoran") { Xmin=327515.0; Xmax=327598.0; Ymin=3619121.0; Ymax=3619251.0 }
if (DESERT=="Mojave") { Xmin=510386; Xmax=510516; Ymin=3890308; Ymax=3890391 }
xcrds=seq(from=Xmin, to=Xmax, by=0.2)
ycrds=seq(from=Ymin, to=Ymax, by=0.2)
#windows(11,7)
windows(9,11)
xcrds=seq(from=Xmin, to=Xmax, by=0.2)
ycrds=seq(from=Ymin, to=Ymax, by=0.2)
GPredDens=matrix(data=NA, ncol=length(xcrds), nrow=length(ycrds))
GPredDens=im(GPredDens, xcrds, ycrds, unitname=c("meter","meters"))
##########################################################
# Get Shrub locations-------------
if (DESERT=="Sonoran") ThisShrubData=read.csv("./AZLarreaVolume_Complete.csv",header=TRUE)
if (DESERT=="Mojave") ThisShrubData=read.csv("../CALarreaVolume.csv",header=TRUE)
ThisShrubData$Area_Bot = ThisShrubData$Area_Bot/1000
ThisShrubData$Area_Top = ThisShrubData$Area_Top/1000
ThisShrubData$Area_Stem = ThisShrubData$Area_Stem/1000
ThisShrubData$Vol_Bot = ThisShrubData$Vol_Bot/100000
ThisShrubData$Vol_Stem = ThisShrubData$Vol_Stem/100000
ThisShrubData$Vol_Cent = ThisShrubData$Vol_Cent/100000
ThisShrubData$Vol_BotRing = ThisShrubData$Vol_BotRing/100000
ThisShrubData$Vol_StemRing = ThisShrubData$Vol_StemRing/100000
#Read in and munge data -----------------
# PICK A DESERT
#DESERT="Sonoran"
DESERT="Mojave"
YEAR="2013Xtra"
if((DESERT=="Mojave") & (YEAR=="2013Xtra")){
CensusData=read.csv("../Mojave2013Intense24Shrubs.csv", blank.lines.skip=TRUE)
CensusData$Invasives=CensusData$Erodium+CensusData$Schismus+CensusData$Bromus
Shrubs=read.csv("../CALarreaVolume.csv")
Shrubs=subset(Shrubs, ShrubID %in% as.character(1:168))
Shrubs$Shrub=as.numeric(as.character(Shrubs$ShrubID)) #THIS MESSES UP!!!
}
Census=merge(Shrubs[,c("Shrub","Area_Bot","Vol_Bot","Vol_Stem")], CensusData, by.y="Shrub", by.x="Shrub")
Census$Area_Bot=Census$Area_Bot/10000
Census$Vol_Bot=Census$Vol_Bot/1000000
Census$Vol_Stem=Census$Vol_Stem/1000000
Census=Census[order(Census[,"Shrub"], Census[,"MHcode"]), ]
names(Census)[1]="Shrub"
# Setup for elaborate plotting --------------
colorset=sample(rainbow(168))
firecols=c("gray40", "green")
watercols=c("lightblue", "tan")
dircols=c("red","blue")
firepch=c(16,1)
dirwds=c(1,2)
# Chose target species to model this time --------------
TargSpec="ErodCover"; ylabel="% Cover"
#TargSpec="Erodium"; ylabel="# Plants"
Census$Target=round(Census[,TargSpec]) #Round to deal with issue of 0.1 percent cover values
Census$LogitTarg=logit.transform(Census$Target/100)
Census$TargPres=ceiling(Census$Target/100) #Make a new binary variable showing zero or not
# table(Census[c("Fire", "Rain", "TranDir")])
# unique(Census[c("Fire", "Rain", "TranDir")])
# unique(Census[c( "Shrub", "Fire", "Rain", "TranDir")])
# Make training and Testing Data- ------------
# balance out Fire, Rain and TranDir, 7 of each shrub
# Withhold two in each category for testing
set.seed(81)
BurnAmbN=sample(unique(subset(Census, (Fire=="B")&(Rain=="A")&(TranDir=="N"))$Shrub), 2)
BurnAmbS=sample(unique(subset(Census, (Fire=="B")&(Rain=="A")&(TranDir=="S"))$Shrub), 2)
BurnDrtN=sample(unique(subset(Census, (Fire=="B")&(Rain=="D")&(TranDir=="N"))$Shrub), 2)
UnBurnAmbN=sample(unique(subset(Census, (Fire=="UB")&(Rain=="A")&(TranDir=="N"))$Shrub), 2)
UnBurnAmbS=sample(unique(subset(Census, (Fire=="UB")&(Rain=="A")&(TranDir=="S"))$Shrub), 2)
UnBurnDrtN=sample(unique(subset(Census, (Fire=="UB")&(Rain=="D")&(TranDir=="N"))$Shrub), 2)
TestShrubs=c(BurnAmbN, BurnAmbS, BurnDrtN, UnBurnAmbN, UnBurnAmbS, UnBurnDrtN)
TestShrubs
#58 131 99 14 124 120 122 80 93 63 69 26
Census.Test=Census[Census$Shrub %in% TestShrubs,]
Census.Train=Census[!(Census$Shrub %in% TestShrubs),]
# PICK ONE NUTRIENT AT A TIME. Pick the same name as the column heading
SPECIES="ErodCover"; ylabel="% Cover"
VegData=Census.Train[Census.Train$Target>0,c("Shrub", "TranDir", "Area_Bot", "Fire","Rain", "Plot", "PlotDist", SPECIES)]
VegData=Census.Test[Census.Test$Target>0,c("Shrub", "TranDir", "Area_Bot", "Fire","Rain", "Plot", "PlotDist", SPECIES)]
names(VegData)[ncol(VegData)]="Species"
source("./VegSpatialModeling/PlotTransectGraphs.R")
# Sample Shrub is the list of shrubs to model against.
SampleShrub=subset(ShrubData,(Northing<=Ymax)&(Northing>=Ymin)&(Easting<=Xmax)&(Easting>=Xmin), #Only want shrubs within boundaries
select=c("ShrubID","Easting","Northing","Area_Bot","Area_Stem")) #Only want certain info
SampleShrub=unique(VegData[c("Shrub","TranDir","Area_Bot","Fire","Rain")])
SS=SampleShrub #just to shorten the crazy code below...
ModelResults=read.csv("../ErodMojParams.csv",header=T)
ModelResults[, 1:5]
whichmodel=1
ThisModel=ModelResults[whichmodel,]
ThisModel
Model=ThisModel
rownames(SS)=1:nrow(SS)
nrow(SS)
SS$TranDirDummy=as.numeric(SS$TranDir)-1 #Now N=0 and S=1
SS$FireDummy=as.numeric(SS$Fire)-1 #Now B=0 and UB=1
SS$RainDummy=as.numeric(SS$Rain)-1 #Now A=0 and D=1
a.rand=rnorm(nrow(SS),0,Model$a..Intercept..1)
h.rand=rnorm(nrow(SS),0,Model$h..Intercept..1)
k.rand=rnorm(nrow(SS),0,Model$k..Intercept..1)
#Change the random process part to inside the loop???
SS$a=Model$a..Intercept.+ Model$a.Area_Bot * SS$Area_Bot + Model$a.TranDirS*SS$TranDirDummy+ Model$a.Area_Bot.TranDirS*SS$Area_Bot*SS$TranDirDummy+a.rand
SS$h=Model$h..Intercept.+ Model$h.Area_Bot * SS$Area_Bot + Model$h.TranDirS*SS$TranDirDummy+h.rand
SS$k=Model$k..Intercept.+ Model$k.Area_Bot * SS$Area_Bot + Model$k.FireUB*SS$FireDummy+Model$k.RainD*SS$RainDummy + Model$k.Area_Bot.FireUB*SS$Area_Bot*SS$FireDummy+k.rand
SampleShrub=SS
#Stage 1:
Stg1Model=data.frame(t(fixed.effects(hurdle6.lme)))
### Show curves for these shrubs -----------
#windows(7,4)
# Setup for elaborate plotting
colorset=sample(rainbow(168))
firecols=c("gray40", "green")
watercols=c("lightblue", "tan")
dircols=c("red","blue")
firepch=c(16,1)
dirwds=c(1,2)
col=watercols[SampleShrub$Rain[i]]
col=firecols[SampleShrub$Fire[i]]
col=dircols[SampleShrub$TranDir[i]]
windows()
with(VegData, plot(PlotDist, Species/100, xlab="Distance from Shrub Stem", ylab="Cover"))
#Must divide species by 100 so logit transform will work.
with(VegData, plot(PlotDist, logit.transform(Species/100), xlab="Distance from Shrub Stem", ylab="Cover", yaxt="n"))
axis(2, at=logit.transform(seq(0,0.35, by=0.05)), labels=seq(0,0.35, by=0.05))
#plot(c(0, 500), c(0,1), pch=NA, xlab="Distance from Shrub (cm)", ylab="P(plant present)")
for (i in 1:nrow(SampleShrub)){
new.x=seq(0,260, by=20)
#Stage 1-
logitp=Stg1Model$X.Intercept. + Stg1Model$RainD*SampleShrub$RainDummy[i]+Stg1Model$PlotDist*new.x + Stg1Model$TranDirS*SampleShrub$TranDirDummy[i]+ Stg1Model$I.PlotDist.2.*(new.x^2)+ Stg1Model$FireUB*SampleShrub$FireDummy[i] + Stg1Model$PlotDist.TranDirS*SampleShrub$TranDirDummy[i]*new.x+ Stg1Model$TranDirS.I.PlotDist.2.*SampleShrub$TranDirDummy[i]*(new.x^2)+ Stg1Model$PlotDist.FireUB*SampleShrub$FireDummy[i]*new.x+ Stg1Model$I.PlotDist.2..FireUB*SampleShrub$FireDummy[i]*(new.x^2)
p=INVlogit.transform(logitp)
#lines(new.x, p, col=dircols[SampleShrub$TranDir[i]], lwd=3)
#Stage 2- if present, what is %Cover?
cover=INVlogit.transform(SampleShrub$a[i]*((new.x/100)-SampleShrub$h[i])^2+SampleShrub$k[i])
occurs=rbinom(length(new.x), 1, p)
#logit.transform(occur)
#Combine the two
new.y=cover*occurs
lines(new.x, new.y, lwd=3,col=watercols[SampleShrub$Rain[i]])
}
title(main=paste("Curves for", SPECIES, " in the ", DESERT, " Model ", ThisModel$Model), sub="Testing DataSet")
######## Map Making -------------
GPredDens=matrix(data=NA, ncol=length(xcrds), nrow=length(ycrds))
GPredDens.im=im(GPredDens, xcrds, ycrds, unitname=c("meter","meters"))
BlankImage.im=GPredDens.im
ShrubData=ThisShrubData
### Quadratic Vertex Models--------------
#PICK MODEL TO USE- Must match desert and nutrient
DESERT
SPECIES
#ThisIntSpPars=subset(IntSpcParams, (Desert==DESERT)&( Nutrient==NUTRIENT))
Map.im=ApplyQuadVertexModel(ThisModel, ThisShrubData, ThisIntSpPars, GPredNut.im)
save(SonoranN1.im, file="SoilProbeSampling/SonoranN1imageObject")
cols=gray(seq(1,0,length.out=700))
windows(9,9)
plot(Map.im,col=cols,main=paste("estimated map for ", NUTRIENT, " in the ", DESERT, " Model ", ThisModel$Model))