Marten Space-time
Marten (Martes americana), U.P., Michigan
1 Get County Boundaries
Read county boundaries shapefile.
Counties = readOGR(dsn = "./Counties",
layer = "Counties_v17a",
stringsAsFactors = FALSE)## OGR data source with driver: ESRI Shapefile
## Source: "E:\HarvMod\HarvestModeling\Counties", layer: "Counties_v17a"
## with 83 features
## It has 15 fields
## Integer64 fields read as strings: OBJECTID FIPSNUMUP = subset(Counties, PENINSULA == "upper")
UP.union = gUnaryUnion(UP)
#Create a rasterized version
Ras = raster(res = 0.01, ext = extent(UP.union),
crs = proj4string(UP.union))
Domain.r = rasterize(UP.union, Ras,
field = 0,
background = NA)
#Creat a point grid version
Grd.pnt = rasterToPoints(Domain.r, spatial = TRUE)
Grd.pnt@data = Grd.pnt@data %>%
mutate(Long = Grd.pnt@coords[,1],
Lat = Grd.pnt@coords[,2]) %>%
select(-layer)2 Load Marten Harvest Data
Second data set (with age and sex)
#Mart.df = read.csv("C:/Users/Talesha Dokes/Michigan State University/Roloff, Gary - HarvestModeling/MartenDataJun2019/Roloff_marten_registrations.csv",
# header = TRUE, sep=",")
Mart.df = read.csv("./MartenDataJun2019/Roloff_marten_registrations.csv",
header = TRUE, sep=",")
#Table with PLSS abd coordinates for Michigan
#PLSS = read.csv("F://MartenPhDWork//Hecology//MI_PLSS.csv", #FYI - This file wasn't on the OneDrive
# header = TRUE, sep=",")
PLSS = read.csv("./PLSS/MI_PLSS.csv",
header = TRUE, sep=",")
Mart.df$Long = with(PLSS,
Long[match(
Mart.df$CorrectedTRSec,
TWNRNGSEC)])
Mart.df$Lat = with(PLSS,
Lat[match(
Mart.df$CorrectedTRSec,
TWNRNGSEC)])
Mart.df = Mart.df %>%
mutate(Year = Season_Year,
#Month = month(TakeDate),
Age = Lab_Age,
Sex = Lab_Sex,
Spp = "Marten",
OBS = 1,
Source = "MDNR") %>%
filter(is.na(Year) == FALSE,
is.na(Long) == FALSE,
is.na(Lat) == FALSE) %>%
select( Year, Age, Sex, Spp, OBS, Year, Long, Lat, Source)
dim(Mart.df)## [1] 3479 8#Count records per year
Mart.df %>%
group_by(Year) %>%
summarise(Count = length(Year))## # A tibble: 19 x 2
## Year Count
## <int> <int>
## 1 2000 67
## 2 2001 63
## 3 2002 64
## 4 2003 114
## 5 2004 134
## 6 2005 125
## 7 2006 146
## 8 2007 233
## 9 2008 218
## 10 2009 214
## 11 2010 253
## 12 2011 178
## 13 2012 276
## 14 2013 251
## 15 2014 302
## 16 2015 223
## 17 2016 95
## 18 2017 170
## 19 2018 353Mart.df = Mart.df %>% filter(Year < 2019)
Mart.df %>%
group_by(Year) %>%
summarise(Count = length(Year))## # A tibble: 19 x 2
## Year Count
## <int> <int>
## 1 2000 67
## 2 2001 63
## 3 2002 64
## 4 2003 114
## 5 2004 134
## 6 2005 125
## 7 2006 146
## 8 2007 233
## 9 2008 218
## 10 2009 214
## 11 2010 253
## 12 2011 178
## 13 2012 276
## 14 2013 251
## 15 2014 302
## 16 2015 223
## 17 2016 95
## 18 2017 170
## 19 2018 353dim(Mart.df)## [1] 3479 8Convert to Points
Marten.pnt = SpatialPointsDataFrame(Mart.df[,c("Long", "Lat")], Mart.df)
proj4string(Marten.pnt) = proj4string(UP)
#Keep only those in the UP
Marten.pnt$UP = is.na(over(Marten.pnt, UP.union))
Marten.pnt = subset(Marten.pnt, UP == FALSE)Quick Plot
UP_map = fortify(UP)## Regions defined for each Polygonstmp.df = Marten.pnt@data %>%
select(Long, Lat, Year)
ggplot(UP_map, aes(long,lat, group=group)) +
geom_polygon(fill = "tan", col="black") +
geom_point(data=tmp.df,
aes(Long, Lat,
group=NULL, fill=NULL),
col = "red", size=1) +
xlab("Longitude") +
ylab("Latitude") +
ggtitle("Marten Harvest") +
guides(color=guide_legend(title = "Year")) +
facet_wrap(~Year, ncol = 3) +
coord_equal() +
theme(panel.grid.minor = element_blank(),
panel.grid.major = element_blank(),
panel.background = element_blank(),
plot.background = element_blank(),
panel.border = element_blank(),
legend.title = element_blank(),
axis.title.x = element_text(size=16, face="bold"),
axis.title.y = element_text(size=16, face="bold"),
plot.title = element_text(size=20, face="bold", hjust = 0.5))
Change Projection
nProj = "+proj=utm +zone=16 +datum=NAD83 +units=km +no_defs +ellps=GRS80 +towgs84=0,0,0"
UP.prj = spTransform(UP, nProj)
UP.union.prj = spTransform(UP.union, nProj)
Marten.pnt.prj = spTransform(Marten.pnt, nProj)
Marten.pnt.prj@data = Marten.pnt.prj@data %>%
mutate(Longp = Marten.pnt.prj@coords[,1],
Latp = Marten.pnt.prj@coords[,2])3 Construct Mesh
max.edge = 2
bound.outer = 30
bdry = inla.sp2segment(UP.union.prj)
mesh = inla.mesh.2d(boundary = bdry,
loc=cbind(Marten.pnt.prj),
max.edge = c(1, 12)*max.edge,
cutoff = 5,
min.angle = 23,
offset = c(max.edge, bound.outer))
plot(mesh, lwd=0.5, main=" ",
rgl=F, draw.vertices = TRUE,
vertex.color = "blue")
mesh$n## [1] 1984Get Mesh Nodes
dd = as.data.frame(cbind(mesh$loc[,1],
mesh$loc[,2]))
names(dd) = c("Longp", "Latp")
dd = dd %>%
mutate(OBS = 0,
Source = "Mesh")
##Make a copy for each year
Year.lvls = levels(factor(Marten.pnt.prj$Year))
for(i in 1:length(Year.lvls)){
tmp.dd = dd %>%
mutate(Year = Year.lvls[i])
if(i==1){dd.df = tmp.dd}
else{dd.df = rbind(dd.df, tmp.dd)}}
Month.lvls = levels(factor(Marten.pnt.prj$Month))
Age.lvls = levels(factor(Marten.pnt.prj$Age))
dd.length = dim(dd.df)[1]
#Random values for background points
#dd.df$Month = sample(Month.lvls, dd.length, replace = T)
dd.df$Sex = sample(c("F","M"), dd.length, replace = T)
dd.df$Age = sample(Age.lvls, dd.length, replace = T)
levels(factor(dd.df$Year))## [1] "2000" "2001" "2002" "2003" "2004" "2005" "2006" "2007" "2008" "2009"
## [11] "2010" "2011" "2012" "2013" "2014" "2015" "2016" "2017" "2018"head(dd.df)## Longp Latp OBS Source Year Sex Age
## 1 522.7310 5050.779 0 Mesh 2000 M 6.5
## 2 700.1876 5066.286 0 Mesh 2000 M 8.5
## 3 489.1956 5057.737 0 Mesh 2000 F 2.5
## 4 677.8654 5151.228 0 Mesh 2000 F 0.5
## 5 412.3404 5201.961 0 Mesh 2000 M 1.5
## 6 656.7944 5149.811 0 Mesh 2000 F 5.5Combine Marten and Nodes Also converting to points.
Mod.pnts = Marten.pnt.prj@data %>%
select(Longp, Latp, OBS, Source, Year, Sex, Age)
Mod.pnts = rbind(Mod.pnts, dd.df)
Mod.pnts = SpatialPointsDataFrame(Mod.pnts[,c("Longp", "Latp")], Mod.pnts)
proj4string(Mod.pnts) = proj4string(Marten.pnt.prj)4 Monthly PRISM Data
This will download PRISM data to specified folder and then convert it to a raster. Matching climate to month of harvest.
Max Temperature (TMAX)
library(prism)
range(Marten.pnt$Year)## [1] 2000 2018#range(Marten.pnt$Month)
options(prism.path = "./PRISM/TMAX") #Where to save the data; seperate folder for temp
get_prism_monthlys(type = 'tmax', #climate variable wanted, options: "ppt", "tmean", "tmin", "tmax"
years=2000:2018, #Years wanted
mon = 10:12, #Months wanted
keepZip = FALSE) #Don't keep zip file##
|
| | 0%
|
|= | 2%
|
|== | 4%
|
|=== | 5%
|
|===== | 7%
|
|====== | 9%
|
|======= | 11%
|
|======== | 12%
|
|========= | 14%
|
|========== | 16%
|
|=========== | 18%
|
|============= | 19%
|
|============== | 21%
|
|=============== | 23%
|
|================ | 25%
|
|================= | 26%
|
|================== | 28%
|
|=================== | 30%
|
|===================== | 32%
|
|====================== | 33%
|
|======================= | 35%
|
|======================== | 37%
|
|========================= | 39%
|
|========================== | 40%
|
|=========================== | 42%
|
|============================= | 44%
|
|============================== | 46%
|
|=============================== | 47%
|
|================================ | 49%
|
|================================= | 51%
|
|================================== | 53%
|
|=================================== | 54%
|
|==================================== | 56%
|
|====================================== | 58%
|
|======================================= | 60%
|
|======================================== | 61%
|
|========================================= | 63%
|
|========================================== | 65%
|
|=========================================== | 67%
|
|============================================ | 68%
|
|============================================== | 70%
|
|=============================================== | 72%
|
|================================================ | 74%
|
|================================================= | 75%
|
|================================================== | 77%
|
|=================================================== | 79%
|
|==================================================== | 81%
|
|====================================================== | 82%
|
|======================================================= | 84%
|
|======================================================== | 86%
|
|========================================================= | 88%
|
|========================================================== | 89%
|
|=========================================================== | 91%
|
|============================================================ | 93%
|
|============================================================== | 95%
|
|=============================================================== | 96%
|
|================================================================ | 98%
|
|=================================================================| 100%GetFiles = ls_prism_data(name=TRUE) #File names
#Count files
dim(GetFiles)## [1] 57 2#View (2 columns)
head(GetFiles)## files
## 1 PRISM_tmax_stable_4kmM2_200010_bil
## 2 PRISM_tmax_stable_4kmM2_200011_bil
## 3 PRISM_tmax_stable_4kmM2_200012_bil
## 4 PRISM_tmax_stable_4kmM2_200110_bil
## 5 PRISM_tmax_stable_4kmM2_200111_bil
## 6 PRISM_tmax_stable_4kmM2_200112_bil
## product_name
## 1 Oct 2000 - 4km resolution - Maximum temperature
## 2 Nov 2000 - 4km resolution - Maximum temperature
## 3 Dec 2000 - 4km resolution - Maximum temperature
## 4 Oct 2001 - 4km resolution - Maximum temperature
## 5 Nov 2001 - 4km resolution - Maximum temperature
## 6 Dec 2001 - 4km resolution - Maximum temperatureMyMonths = substr(GetFiles$product_name, 1, 3) #Pull month abbreviation
unique(MyMonths)## [1] "Oct" "Nov" "Dec"MyYear = substr(GetFiles$product_name, 6, 9) #Pull year
unique(MyYear)## [1] "2000" "2001" "2002" "2003" "2004" "2005" "2006" "2007" "2008" "2009"
## [11] "2010" "2011" "2012" "2013" "2014" "2015" "2016" "2017" "2018"#stack rasters
TMAX.stk = prism_stack(ls_prism_data())
#Simplified names
names(TMAX.stk) = paste(MyMonths, MyYear, sep="_")
#Plot one
plot(TMAX.stk[[1]])
##Create TMAX Dataframe
TMAX.df = as.data.frame(
extract(TMAX.stk, #Finding values for each location
spTransform(Mod.pnts,
proj4string(TMAX.stk)),
method = "simple"))
#Some points for the mesh were over water and coded as NA, impute by month and year
for(i in 1:dim(TMAX.df)[2]){
TMAX.df[,i][is.na(TMAX.df[,i])] = mean(TMAX.df[,i], na.rm=TRUE)
}
TMAX.df$RowID = 1:nrow(TMAX.df)
dim(TMAX.df)## [1] 41145 58head(TMAX.df)## Oct_2000 Nov_2000 Dec_2000 Oct_2001 Nov_2001 Dec_2001 Oct_2002 Nov_2002
## 1 15.37 3.37 -7.70 12.42 9.86 0.12 7.73 1.74
## 2 15.19 3.10 -8.01 12.25 9.63 -0.06 7.41 1.50
## 3 15.59 3.89 -7.56 11.78 8.40 -0.27 8.40 1.10
## 4 15.02 4.34 -6.17 11.04 8.54 0.68 9.03 2.13
## 5 14.73 4.89 -5.55 12.46 8.49 1.61 9.64 2.55
## 6 14.47 5.48 -5.17 11.98 7.81 2.07 9.48 2.84
## Dec_2002 Oct_2003 Nov_2003 Dec_2003 Oct_2004 Nov_2004 Dec_2004 Oct_2005
## 1 -1.42 11.74 2.60 -0.56 12.95 5.46 -3.74 13.80
## 2 -1.71 11.53 2.33 -0.84 12.69 5.18 -4.03 13.59
## 3 -1.85 11.61 2.24 -1.20 12.00 4.31 -4.70 13.85
## 4 -0.96 11.66 3.35 -0.23 13.10 5.55 -2.85 13.68
## 5 -0.87 11.09 4.25 0.19 12.94 5.77 -2.55 13.88
## 6 -0.09 10.89 5.05 0.93 12.64 6.16 -1.75 13.84
## Nov_2005 Dec_2005 Oct_2006 Nov_2006 Dec_2006 Oct_2007 Nov_2007 Dec_2007
## 1 3.92 -4.01 9.80 5.24 -0.48 14.84 2.58 -4.45
## 2 3.69 -4.29 9.58 5.00 -0.74 14.64 2.33 -4.74
## 3 2.97 -4.47 9.62 4.66 -0.93 14.80 2.22 -4.60
## 4 3.96 -3.38 9.87 5.51 0.18 14.99 3.24 -3.17
## 5 4.73 -3.27 10.53 5.61 0.47 15.22 4.14 -2.52
## 6 5.33 -2.49 10.19 5.74 1.26 15.14 4.19 -1.60
## Oct_2008 Nov_2008 Dec_2008 Oct_2009 Nov_2009 Dec_2009 Oct_2010 Nov_2010
## 1 11.58 2.97 -6.27 8.22 7.65 -4.73 13.98 4.82
## 2 11.37 2.72 -6.50 7.95 7.44 -4.98 13.79 4.59
## 3 11.80 2.63 -6.31 8.06 7.17 -4.84 14.33 4.64
## 4 11.55 3.95 -4.59 7.96 8.29 -3.58 14.35 5.55
## 5 12.10 4.44 -4.12 9.09 8.00 -2.85 14.24 5.85
## 6 11.66 4.74 -3.42 8.91 8.06 -1.98 13.87 6.30
## Dec_2010 Oct_2011 Nov_2011 Dec_2011 Oct_2012 Nov_2012 Dec_2012 Oct_2013
## 1 -4.36 14.04 5.18 -1.21 11.06 3.89 -1.40 11.83
## 2 -4.59 13.81 4.93 -1.45 10.84 3.68 -1.63 11.59
## 3 -4.19 13.92 4.38 -1.58 11.16 3.29 -1.67 11.67
## 4 -3.10 14.14 5.49 -0.80 11.23 4.15 -0.15 12.07
## 5 -2.82 14.25 6.00 -0.02 12.15 4.57 0.07 12.71
## 6 -2.27 13.91 6.76 0.70 12.22 5.21 0.70 12.58
## Nov_2013 Dec_2013 Oct_2014 Nov_2014 Dec_2014 Oct_2015 Nov_2015 Dec_2015
## 1 2.09 -7.83 10.50 -1.01 -2.29 12.58 7.38 0.79
## 2 1.85 -8.10 10.32 -1.27 -2.53 12.39 7.14 0.54
## 3 1.68 -8.00 9.96 -1.60 -2.42 12.25 6.68 0.67
## 4 2.78 -6.53 10.04 -0.40 -1.39 12.21 7.59 2.16
## 5 3.48 -6.18 10.76 0.51 -1.13 12.57 7.78 2.75
## 6 4.01 -5.78 10.79 1.13 -0.32 12.38 8.13 3.50
## Oct_2016 Nov_2016 Dec_2016 Oct_2017 Nov_2017 Dec_2017 Oct_2018 Nov_2018
## 1 13.203 8.715 -3.665 14.229 1.526 -5.771 7.995 -1.016
## 2 13.031 8.476 -3.932 13.993 1.285 -6.056 7.865 -1.245
## 3 12.878 8.246 -3.955 14.158 0.925 -6.052 7.732 -1.376
## 4 13.124 9.050 -2.271 14.103 2.398 -4.367 8.467 0.320
## 5 14.163 9.015 -1.810 14.475 2.823 -4.787 9.035 0.607
## 6 14.365 9.112 -0.527 14.775 3.600 -4.163 9.176 1.213
## Dec_2018 RowID
## 1 -1.461 1
## 2 -1.669 2
## 3 -1.813 3
## 4 -0.454 4
## 5 -0.562 5
## 6 0.330 6Find the average of October, November, and December for each year.
Melt.Temp.df = melt(TMAX.df, "RowID")
head(Melt.Temp.df) #Now in long format## RowID variable value
## 1 1 Oct_2000 15.37
## 2 2 Oct_2000 15.19
## 3 3 Oct_2000 15.59
## 4 4 Oct_2000 15.02
## 5 5 Oct_2000 14.73
## 6 6 Oct_2000 14.47Melt.Temp.df$Year = substr(Melt.Temp.df$variable, 5, 9)
head(Melt.Temp.df)## RowID variable value Year
## 1 1 Oct_2000 15.37 2000
## 2 2 Oct_2000 15.19 2000
## 3 3 Oct_2000 15.59 2000
## 4 4 Oct_2000 15.02 2000
## 5 5 Oct_2000 14.73 2000
## 6 6 Oct_2000 14.47 2000range(Melt.Temp.df$Year) #double checking## [1] "2000" "2018"TMAX.df.avg = as.data.frame(
Melt.Temp.df %>%
group_by(Year) %>% #only Oct-Dec are in dataset
summarise(Average = mean(value)))
head(TMAX.df.avg)## Year Average
## 1 2000 4.380668
## 2 2001 7.594855
## 3 2002 3.657639
## 4 2003 5.188226
## 5 2004 5.441967
## 6 2005 5.154181#Match to marten records by Year
Mod.pnts$TMAX = with(TMAX.df.avg,
Average[match(
Mod.pnts$Year,
Year)])Total Precipitation (PPT) As with TMAX
options(prism.path = "./PRISM/PPT") #Where to save the data; seperate folder for precip
get_prism_monthlys(type = 'ppt', #climate variable wanted, options: "ppt", "tmean", "tmin", "tmax"
years=2000:2018, #Years wanted
mon = 10:12, #Months wanted
keepZip = FALSE) #Don't keep zip file##
|
| | 0%
|
|= | 2%
|
|== | 4%
|
|=== | 5%
|
|===== | 7%
|
|====== | 9%
|
|======= | 11%
|
|======== | 12%
|
|========= | 14%
|
|========== | 16%
|
|=========== | 18%
|
|============= | 19%
|
|============== | 21%
|
|=============== | 23%
|
|================ | 25%
|
|================= | 26%
|
|================== | 28%
|
|=================== | 30%
|
|===================== | 32%
|
|====================== | 33%
|
|======================= | 35%
|
|======================== | 37%
|
|========================= | 39%
|
|========================== | 40%
|
|=========================== | 42%
|
|============================= | 44%
|
|============================== | 46%
|
|=============================== | 47%
|
|================================ | 49%
|
|================================= | 51%
|
|================================== | 53%
|
|=================================== | 54%
|
|==================================== | 56%
|
|====================================== | 58%
|
|======================================= | 60%
|
|======================================== | 61%
|
|========================================= | 63%
|
|========================================== | 65%
|
|=========================================== | 67%
|
|============================================ | 68%
|
|============================================== | 70%
|
|=============================================== | 72%
|
|================================================ | 74%
|
|================================================= | 75%
|
|================================================== | 77%
|
|=================================================== | 79%
|
|==================================================== | 81%
|
|====================================================== | 82%
|
|======================================================= | 84%
|
|======================================================== | 86%
|
|========================================================= | 88%
|
|========================================================== | 89%
|
|=========================================================== | 91%
|
|============================================================ | 93%
|
|============================================================== | 95%
|
|=============================================================== | 96%
|
|================================================================ | 98%
|
|=================================================================| 100%GetFiles = ls_prism_data(name=TRUE) #File names
#Count files
dim(GetFiles)## [1] 57 2#View (2 columns)
head(GetFiles)## files
## 1 PRISM_ppt_stable_4kmM3_200010_bil
## 2 PRISM_ppt_stable_4kmM3_200011_bil
## 3 PRISM_ppt_stable_4kmM3_200012_bil
## 4 PRISM_ppt_stable_4kmM3_200110_bil
## 5 PRISM_ppt_stable_4kmM3_200111_bil
## 6 PRISM_ppt_stable_4kmM3_200112_bil
## product_name
## 1 Oct 2000 - 4km resolution - Precipitation
## 2 Nov 2000 - 4km resolution - Precipitation
## 3 Dec 2000 - 4km resolution - Precipitation
## 4 Oct 2001 - 4km resolution - Precipitation
## 5 Nov 2001 - 4km resolution - Precipitation
## 6 Dec 2001 - 4km resolution - PrecipitationMyMonths = substr(GetFiles$product_name, 1, 3) #Pull month abbreviation
MyYear = substr(GetFiles$product_name, 6, 9) #Pull year
#stack rasters
PPT.stk = prism_stack(ls_prism_data())
#Simplified names
names(PPT.stk) = paste(MyMonths, MyYear, sep="_")
#Plot one
plot(PPT.stk[[1]])
##Create TMAX Dataframe
PPT.df = as.data.frame(
extract(PPT.stk, #Finding values for each location
spTransform(Mod.pnts,
proj4string(PPT.stk)),
method = "simple"))
for(i in 1:dim(PPT.df)[2]){
PPT.df[,i][is.na(PPT.df[,i])] = mean(PPT.df[,i], na.rm=TRUE)
}
PPT.df$RowID = 1:nrow(PPT.df)
dim(PPT.df)## [1] 41145 58head(PPT.df)## Oct_2000 Nov_2000 Dec_2000 Oct_2001 Nov_2001 Dec_2001 Oct_2002 Nov_2002
## 1 35.23 68.64 59.40 70.85 56.38 46.87 140.93 42.62
## 2 33.88 59.09 53.95 66.62 55.18 44.57 152.46 35.17
## 3 51.79 66.46 73.69 87.62 57.93 54.60 170.89 49.66
## 4 68.00 98.02 90.95 99.44 70.53 55.96 160.45 74.15
## 5 55.36 76.43 76.33 126.48 65.01 65.38 134.16 44.55
## 6 40.99 57.05 42.53 147.86 98.53 60.38 88.01 26.71
## Dec_2002 Oct_2003 Nov_2003 Dec_2003 Oct_2004 Nov_2004 Dec_2004 Oct_2005
## 1 13.06 58.09 94.05 41.44 87.92 35.08 53.50 150.22
## 2 11.50 55.22 86.63 40.02 89.44 32.11 53.35 149.34
## 3 11.91 74.02 72.57 53.71 125.62 46.47 96.05 138.99
## 4 24.85 79.12 75.21 72.59 133.70 67.82 84.19 147.32
## 5 20.48 64.00 87.14 67.48 169.62 67.73 97.36 110.26
## 6 23.56 55.57 122.05 58.28 129.21 75.60 89.22 56.30
## Nov_2005 Dec_2005 Oct_2006 Nov_2006 Dec_2006 Oct_2007 Nov_2007 Dec_2007
## 1 84.30 38.54 65.89 34.45 62.60 154.93 36.14 60.55
## 2 78.66 36.73 64.23 33.19 61.59 163.23 34.34 58.40
## 3 86.03 30.61 64.05 29.75 51.61 129.86 67.04 66.02
## 4 123.71 45.12 64.60 42.56 64.29 126.61 54.47 75.03
## 5 105.97 39.65 91.19 174.70 70.11 127.09 57.82 67.61
## 6 109.45 32.46 83.26 62.77 74.40 117.33 62.17 45.04
## Oct_2008 Nov_2008 Dec_2008 Oct_2009 Nov_2009 Dec_2009 Oct_2010 Nov_2010
## 1 57.03 44.59 56.57 134.40 32.66 58.91 39.70 52.98
## 2 56.87 42.37 53.41 132.85 30.79 55.01 41.47 50.61
## 3 63.85 56.49 60.74 119.87 15.86 54.66 61.07 41.72
## 4 66.59 67.34 91.52 140.55 32.13 72.62 61.73 50.63
## 5 55.81 76.84 108.77 140.73 30.92 67.82 59.02 53.92
## 6 56.97 70.38 118.40 165.26 34.64 59.42 60.92 76.84
## Dec_2010 Oct_2011 Nov_2011 Dec_2011 Oct_2012 Nov_2012 Dec_2012 Oct_2013
## 1 40.03 60.24 46.64 40.68 113.64 52.00 22.41 96.27
## 2 38.21 59.96 43.89 38.31 120.09 47.56 21.97 99.82
## 3 54.41 73.26 45.26 38.97 111.51 43.41 28.53 85.61
## 4 86.02 65.07 53.99 47.83 125.61 49.51 40.47 90.69
## 5 49.52 82.14 69.66 48.54 133.84 46.74 49.51 100.94
## 6 41.93 69.42 87.50 41.08 120.32 26.88 59.42 78.06
## Nov_2013 Dec_2013 Oct_2014 Nov_2014 Dec_2014 Oct_2015 Nov_2015 Dec_2015
## 1 93.79 44.86 115.52 101.75 58.89 60.81 88.71 99.479
## 2 89.88 41.66 111.69 99.34 55.18 65.16 85.55 99.671
## 3 80.38 32.39 112.27 80.46 50.61 64.44 60.07 87.878
## 4 120.79 60.37 165.89 117.36 58.13 71.53 89.21 125.580
## 5 176.53 87.10 174.43 148.20 58.09 61.57 115.53 133.260
## 6 173.93 63.04 229.87 127.88 64.02 90.23 142.60 151.865
## Oct_2016 Nov_2016 Dec_2016 Oct_2017 Nov_2017 Dec_2017 Oct_2018 Nov_2018
## 1 61.122 40.455 64.623 168.823 58.181 55.480 174.724 59.695
## 2 59.744 42.982 59.779 166.483 53.320 52.893 171.800 54.773
## 3 81.642 37.667 47.970 153.757 47.921 51.701 156.593 61.696
## 4 178.877 49.675 64.092 168.305 46.266 72.738 168.485 66.063
## 5 117.555 65.505 88.331 173.269 63.590 122.602 202.388 67.014
## 6 87.207 53.003 74.624 173.178 69.833 80.728 213.381 82.058
## Dec_2018 RowID
## 1 38.523 1
## 2 34.709 2
## 3 41.682 3
## 4 51.381 4
## 5 43.737 5
## 6 46.274 6Find the average of October, November, and December for each year.
Melt.PPT.df = melt(PPT.df, "RowID")
head(Melt.PPT.df) #Now in long format## RowID variable value
## 1 1 Oct_2000 35.23
## 2 2 Oct_2000 33.88
## 3 3 Oct_2000 51.79
## 4 4 Oct_2000 68.00
## 5 5 Oct_2000 55.36
## 6 6 Oct_2000 40.99Melt.PPT.df$Year = substr(Melt.PPT.df$variable, 5, 9)
head(Melt.PPT.df)## RowID variable value Year
## 1 1 Oct_2000 35.23 2000
## 2 2 Oct_2000 33.88 2000
## 3 3 Oct_2000 51.79 2000
## 4 4 Oct_2000 68.00 2000
## 5 5 Oct_2000 55.36 2000
## 6 6 Oct_2000 40.99 2000range(Melt.PPT.df$Year) #double checking## [1] "2000" "2018"PPT.df.avg = as.data.frame(
Melt.PPT.df %>%
group_by(Year) %>% #only Oct-Dec are in dataset
summarise(Average = mean(value)))
head(PPT.df.avg)## Year Average
## 1 2000 56.86325
## 2 2001 76.81729
## 3 2002 61.61163
## 4 2003 60.23360
## 5 2004 83.30350
## 6 2005 79.25578#Match to marten records by Year
Mod.pnts$PPT = with(PPT.df.avg,
Average[match(
Mod.pnts$Year,
Year)])
range(Mod.pnts$PPT)## [1] 51.65295 100.514485 SUm of Road Lengths
#suppressMessages(library(spatstat))#Rds = readOGR(dsn = "F://MartenPhDWork//Hecology//Allroads",
# layer = "output",
# stringsAsFactors = FALSE)
#Rds = spTransform(Rds, nProj) #convert projection
#Road.psp = as.psp(Rds) #format roads
#load("F://MartenPhDWork//Hecology//Rds.RData")
loc.road = raster("./Roads/locrdssecsum")
maj.road = raster("./Roads/majrdssecsum")#M.pp = as.ppp(Mod.pnts) #format marten points
#Distance from each point to nearest road
#Mod.pnts$nRoad = nncross(M.pp, Road.psp)[,"dist"]
Mod.pnts$loc= as.numeric(
extract(loc.road, #Finding values for each location
spTransform(Mod.pnts,
proj4string(loc.road)),
method = "simple"))
Mod.pnts$loc[Mod.pnts$loc < 0] = NA
Mod.pnts$loc[is.na(Mod.pnts$loc)] = 0 #these aren't random (mostly mesh points over water)
Mod.pnts$maj = as.numeric(
extract(maj.road, #Finding values for each location
spTransform(Mod.pnts,
proj4string(maj.road)),
method = "simple"))
Mod.pnts$maj[Mod.pnts$maj < 0] = NA
Mod.pnts$maj[is.na(Mod.pnts$maj)] = 06 LandFire
LandFire are factor-type attributes, not continuous.
VDISTURB: 1. Extract data.
2. Create a “Look.up” value based on closest matching year.
3. Match individual records to available landfore years.
#LOad the data
VDISTURB <- list.files(path="./LANDFIRE/VDISTURB", pattern = "secmaj$", full.names = TRUE)
VDISTURB.LF <- raster::stack(VDISTURB)
#Extract to a data frame
VDISTURB.df = as.data.frame(
extract(VDISTURB.LF,
spTransform(Mod.pnts,
proj4string(VDISTURB.LF), method = "simple")))
head(VDISTURB.df)## v2008secmaj v2010secmaj v2012secmaj v2014secmaj
## 1 0 0 0 0
## 2 0 0 0 0
## 3 0 0 0 0
## 4 0 0 0 0
## 5 0 0 0 0
## 6 0 0 0 0#create year column
VDISTURB.mlt = melt(VDISTURB.df)## No id variables; using all as measure variablesVDISTURB.mlt$Year = substr(VDISTURB.mlt$variable, 2, 5)
head(VDISTURB.mlt)## variable value Year
## 1 v2008secmaj 0 2008
## 2 v2008secmaj 0 2008
## 3 v2008secmaj 0 2008
## 4 v2008secmaj 0 2008
## 5 v2008secmaj 0 2008
## 6 v2008secmaj 0 2008Available.yrs = as.numeric(unique(VDISTURB.mlt$Year)) #ID available years in dataset, convert to number
Available.yrs## [1] 2008 2010 2012 2014Mod.pnts$Year = as.integer(Mod.pnts$Year) #Convert to number
Close.match = function(x){which.min(abs(x - Available.yrs))} #what is the position of the number with minimal absolute difference.
#Test (demo)
Close.match(2010) #test with 2010 that is available## [1] 2Available.yrs[Close.match(2010)]## [1] 2010Close.match(2011) #test with 2011 (not available)## [1] 2Available.yrs[Close.match(2011)]## [1] 2010Close.match(2013) #test with 2013 (not available)## [1] 3Available.yrs[Close.match(2013)]## [1] 2012Mod.pnts$Look.up = Available.yrs[sapply(Mod.pnts$Year, Close.match)]
head(Mod.pnts[,c("Year","Look.up")]) #2014 is most recent## Year Look.up
## 1 2017 2014
## 2 2017 2014
## 3 2017 2014
## 4 2017 2014
## 5 2017 2014
## 6 2017 2014tail(Mod.pnts[,c("Year","Look.up")])## Year Look.up
## 41140 2018 2014
## 41141 2018 2014
## 41142 2018 2014
## 41143 2018 2014
## 41144 2018 2014
## 41145 2018 2014#Assign column based on Look.up year
Mod.pnts$VDISTURB = ifelse(Mod.pnts$Look.up == 2014, VDISTURB.df$v2014secmaj,
ifelse(Mod.pnts$Look.up == 2008, VDISTURB.df$v2008secmaj,
ifelse(Mod.pnts$Look.up == 2010, VDISTURB.df$v2010secmaj,
ifelse(Mod.pnts$Look.up == 2012, VDISTURB.df$v2012secmaj,
NA))))
Mod.pnts$VDISTURB[is.na(Mod.pnts$VDISTURB)] = 0CBD
#LOad the data
CBD <- list.files(path="./LANDFIRE/CBD",
pattern = "secmaj$", full.names = TRUE)
CBD.LF <- raster::stack(CBD)
#Extract to a data frame
CBD.df = as.data.frame(
extract(CBD.LF,
spTransform(Mod.pnts,
proj4string(CBD.LF), method = "simple")))
head(CBD.df)## cbd2001secmaj cbd2008secmaj cbd2010secmaj cbd2012secmaj cbd2014secmaj
## 1 1 1 1 1 1
## 2 1 1 1 1 1
## 3 1 1 1 1 1
## 4 1 1 1 1 1
## 5 1 1 1 1 1
## 6 1 1 1 1 1#create year column
CBD.mlt = melt(CBD.df)## No id variables; using all as measure variablesCBD.mlt$Year = substr(CBD.mlt$variable, 4, 7)
head(CBD.mlt)## variable value Year
## 1 cbd2001secmaj 1 2001
## 2 cbd2001secmaj 1 2001
## 3 cbd2001secmaj 1 2001
## 4 cbd2001secmaj 1 2001
## 5 cbd2001secmaj 1 2001
## 6 cbd2001secmaj 1 2001Available.yrs = as.numeric(unique(CBD.mlt$Year)) #ID available years in dataset, convert to number
Available.yrs## [1] 2001 2008 2010 2012 2014Mod.pnts$Year = as.integer(Mod.pnts$Year) #Convert to number
Mod.pnts$Look.up = Available.yrs[sapply(Mod.pnts$Year, Close.match)] #Overwrite prior look up year
#Assign column based on Look.up year
Mod.pnts$CBD = ifelse(Mod.pnts$Look.up == 2014, CBD.df$cbd2014secmaj,
ifelse(Mod.pnts$Look.up == 2001, CBD.df$cbd2001secmaj,
ifelse(Mod.pnts$Look.up == 2010, CBD.df$cbd2010secmaj,
ifelse(Mod.pnts$Look.up == 2012, CBD.df$cbd2012secmaj,
ifelse(Mod.pnts$Look.up == 2008, CBD.df$cbd2008secmaj,
NA)))))
Mod.pnts$CBD[is.na(Mod.pnts$CBD)] = 0EVC
#LOad the data
EVC <- list.files(path="./LANDFIRE/EVC",
pattern = "secmaj$", full.names = TRUE)
EVC.LF <- raster::stack(EVC)
#Extract to a data frame
EVC.df = as.data.frame(
extract(EVC.LF,
spTransform(Mod.pnts,
proj4string(EVC.LF), method = "simple")))
head(EVC.df)## evc2001secmaj evc2008secmaj evc2010secmaj evc2012secmaj evc2014secmaj
## 1 108 108 108 108 107
## 2 108 107 107 107 107
## 3 107 107 107 107 107
## 4 107 107 107 107 107
## 5 108 108 108 107 107
## 6 106 107 107 107 107#create year column
EVC.mlt = melt(EVC.df)## No id variables; using all as measure variablesEVC.mlt$Year = substr(EVC.mlt$variable, 4, 7)
head(EVC.mlt)## variable value Year
## 1 evc2001secmaj 108 2001
## 2 evc2001secmaj 108 2001
## 3 evc2001secmaj 107 2001
## 4 evc2001secmaj 107 2001
## 5 evc2001secmaj 108 2001
## 6 evc2001secmaj 106 2001Available.yrs = as.numeric(unique(EVC.mlt$Year)) #ID available years in dataset, convert to number
Available.yrs## [1] 2001 2008 2010 2012 2014Mod.pnts$Year = as.integer(Mod.pnts$Year) #Convert to number
Mod.pnts$Look.up = Available.yrs[sapply(Mod.pnts$Year, Close.match)] #Overwrite prior look up year
#Assign column based on Look.up year
Mod.pnts$EVC = ifelse(Mod.pnts$Look.up == 2014, EVC.df$evc2014secmaj,
ifelse(Mod.pnts$Look.up == 2001, EVC.df$evc2001secmaj,
ifelse(Mod.pnts$Look.up == 2010, EVC.df$evc2010secmaj,
ifelse(Mod.pnts$Look.up == 2012, EVC.df$evc2012secmaj,
ifelse(Mod.pnts$Look.up == 2008, EVC.df$evc2008secmaj,
NA)))))
Mod.pnts$EVC[is.na(Mod.pnts$EVC)] = 0
Mod.pnts$EVC[Mod.pnts$EVC < 0] = 0EVH
#Load the data
EVH <- list.files(path="./LANDFIRE/EVH", pattern = "secmaj$", full.names = TRUE)
EVH.LF <- raster::stack(EVH)
#Extract to a data frame
EVH.df = as.data.frame(
extract(EVH.LF,
spTransform(Mod.pnts,
proj4string(EVH.LF), method = "simple")))
head(EVH.df)## evh2001secmaj evh2008secmaj evh2010secmaj evh2012secmaj
## 1 110 110 110 110
## 2 110 110 110 110
## 3 110 110 110 110
## 4 110 110 110 110
## 5 110 110 110 110
## 6 110 110 110 110#create year column
EVH.mlt = melt(EVH.df)## No id variables; using all as measure variablesEVH.mlt$Year = substr(EVH.mlt$variable, 4, 7)
head(EVH.mlt)## variable value Year
## 1 evh2001secmaj 110 2001
## 2 evh2001secmaj 110 2001
## 3 evh2001secmaj 110 2001
## 4 evh2001secmaj 110 2001
## 5 evh2001secmaj 110 2001
## 6 evh2001secmaj 110 2001Available.yrs = as.numeric(unique(EVH.mlt$Year)) #ID available years in dataset, convert to number
Available.yrs## [1] 2001 2008 2010 2012Mod.pnts$Year = as.integer(Mod.pnts$Year) #Convert to number
Mod.pnts$Look.up = Available.yrs[sapply(Mod.pnts$Year, Close.match)] #Overwrite prior look up year
#Assign column based on Look.up year
Mod.pnts$EVH = ifelse(Mod.pnts$Look.up == 2001, EVH.df$evh2001secmaj,
ifelse(Mod.pnts$Look.up == 2010, EVH.df$evh2010secmaj,
ifelse(Mod.pnts$Look.up == 2012, EVH.df$evh2012secmaj,
ifelse(Mod.pnts$Look.up == 2008, EVH.df$evh2008secmaj,
NA))))
Mod.pnts$EVH[is.na(Mod.pnts$EVH)] = 0EVT
#Load the data
EVT <- list.files(path="./LANDFIRE/EVT", pattern = "secmaj$", full.names = TRUE)
EVT.LF <- raster::stack(EVT)
#Extract to a data frame
EVT.df = as.data.frame(
extract(EVT.LF,
spTransform(Mod.pnts,
proj4string(EVT.LF), method = "simple")))
head(EVT.df)## evt2001secmaj evt2008secmaj evt2010secmaj evt2012secmaj evt2014secmaj
## 1 2302 2302 3302 3302 3302
## 2 2302 2302 3302 3302 3302
## 3 2302 2302 3302 3302 3302
## 4 2302 2302 3302 3302 3302
## 5 2302 2302 3302 3302 3302
## 6 2302 2302 3302 3302 3302#create year column
EVT.mlt = melt(EVT.df)## No id variables; using all as measure variablesEVT.mlt$Year = substr(EVT.mlt$variable, 4, 7)
head(EVT.mlt)## variable value Year
## 1 evt2001secmaj 2302 2001
## 2 evt2001secmaj 2302 2001
## 3 evt2001secmaj 2302 2001
## 4 evt2001secmaj 2302 2001
## 5 evt2001secmaj 2302 2001
## 6 evt2001secmaj 2302 2001Available.yrs = as.numeric(unique(EVT.mlt$Year)) #ID available years in dataset, convert to number
Available.yrs## [1] 2001 2008 2010 2012 2014Mod.pnts$Year = as.integer(Mod.pnts$Year) #Convert to number
Mod.pnts$Look.up = Available.yrs[sapply(Mod.pnts$Year, Close.match)] #Overwrite prior look up year
#Assign column based on Look.up year
Mod.pnts$EVT = ifelse(Mod.pnts$Look.up == 2014, EVT.df$evt2014secmaj,
ifelse(Mod.pnts$Look.up == 2001, EVT.df$evt2001secmaj,
ifelse(Mod.pnts$Look.up == 2010, EVT.df$evt2010secmaj,
ifelse(Mod.pnts$Look.up == 2012, EVT.df$evt2012secmaj,
ifelse(Mod.pnts$Look.up == 2008, EVT.df$evt2008secmaj,
NA)))))
Mod.pnts$EVT[is.na(Mod.pnts$EVT)] = 07 Data Scaling
Tmax.sc = scale(Mod.pnts$TMAX, scale=T, center=T)
Mod.pnts$TMAX.sc = as.numeric(Tmax.sc)
Ppt.sc = scale(Mod.pnts$PPT, scale=T, center=T)
Mod.pnts$PPT.sc = as.numeric(Ppt.sc)
loc.sc = scale(Mod.pnts$loc, scale=T, center = T)
Mod.pnts$loc.sc = as.numeric(loc.sc)
maj.sc= scale(Mod.pnts$maj, scale=T, center = T)
Mod.pnts$maj.sc = as.numeric(maj.sc)Check Data
Chk.frm = as.data.frame(
Mod.pnts@data %>%
group_by(Year, Source, OBS) %>%
summarise(Count = length(Year)))
Chk.frm## Year Source OBS Count
## 1 2000 MDNR 1 67
## 2 2000 Mesh 0 1984
## 3 2001 MDNR 1 63
## 4 2001 Mesh 0 1984
## 5 2002 MDNR 1 63
## 6 2002 Mesh 0 1984
## 7 2003 MDNR 1 113
## 8 2003 Mesh 0 1984
## 9 2004 MDNR 1 134
## 10 2004 Mesh 0 1984
## 11 2005 MDNR 1 125
## 12 2005 Mesh 0 1984
## 13 2006 MDNR 1 144
## 14 2006 Mesh 0 1984
## 15 2007 MDNR 1 231
## 16 2007 Mesh 0 1984
## 17 2008 MDNR 1 216
## 18 2008 Mesh 0 1984
## 19 2009 MDNR 1 214
## 20 2009 Mesh 0 1984
## 21 2010 MDNR 1 253
## 22 2010 Mesh 0 1984
## 23 2011 MDNR 1 177
## 24 2011 Mesh 0 1984
## 25 2012 MDNR 1 273
## 26 2012 Mesh 0 1984
## 27 2013 MDNR 1 249
## 28 2013 Mesh 0 1984
## 29 2014 MDNR 1 299
## 30 2014 Mesh 0 1984
## 31 2015 MDNR 1 218
## 32 2015 Mesh 0 1984
## 33 2016 MDNR 1 95
## 34 2016 Mesh 0 1984
## 35 2017 MDNR 1 169
## 36 2017 Mesh 0 1984
## 37 2018 MDNR 1 346
## 38 2018 Mesh 0 1984#Reduce data
Mod.pnts$UP = is.na(over(Mod.pnts, UP.union.prj))
Mod.pnts = subset(Mod.pnts, UP == FALSE)8 Construct Model
Select Time Steps and Knots (time) This model uses an autoregressive term for temporal correlation.
k1 = length(levels(factor(Mod.pnts$Year))) #Total Years in the data
Mod.pnts$Step = as.integer(as.factor(Mod.pnts$Year)) #Integer version
range(Mod.pnts$Step)## [1] 1 19#Reduce temporal dimenstion
gtime1 = seq(1, k1, 2) #Group all available years by 2
mesh.t1 = inla.mesh.1d(gtime1, degree = 1) #Converting to matrix
locs1 = cbind(Mod.pnts$Longp, Mod.pnts$Latp) #Marten locations
A.mart <- inla.spde.make.A(mesh = mesh,
loc = locs1,
group.mesh = mesh.t1, #Time knots
group = Mod.pnts$Step) #All data yearsSpatial Prior and Index
FE.df = Mod.pnts@data
spde0 = inla.spde2.pcmatern(mesh,
alpha = 2,
prior.range=c(500, 0.01),
prior.sigma=c(1, 0.01),
constr = TRUE)
field1 = inla.spde.make.index("field1",
spde0$n.spde,
n.group=mesh.t1$n) Organize data
FE.df$Age2 = FE.df$Age
FE.df$Age2[is.na(FE.df$Age2)] = "Unknown"
FE.lst = list(c(field1, #Spatial index (created in step above)
list(intercept1 = 1)), #intercept
list(TMAX = FE.df[,"TMAX.sc"], #Max Temp covariate
PPT = FE.df[,"PPT.sc"], #Precip covariate
loc.sc = FE.df[,"loc.sc"], #Road variable 1
maj.sc = FE.df[,"maj.sc"], #Road variable 2
Age = FE.df[,"Age2"],
vDist = FE.df[,"VDISTURB"], #Veg disturbance
CBD = FE.df[,"CBD"], #bulk density
EVC = FE.df[,"EVC"], #cover type
EVH = FE.df[,"EVH"],
EVT = FE.df[,"EVT"],
Year = FE.df[,"Year"],
Sex = FE.df[,"Sex"]))
Stack1 = inla.stack(data = list(PA = FE.df$OBS), #PA = either a 1 (marten) or zero (mesh point/node)
A = list(A.mart, 1), #created 2 steps above, loaction info
effects = FE.lst,
tag = "mart.0") #just an arbitrary name for the file
#Temporal Reduced
#save(list=c("Stack1", "spde0"), file="./HPCC/FMart_082419.RData") #New One 8/24/199 Non Spatial Model
These were previoluy run: load results:
load("E:/HarvMod/HarvestModeling/HPCC/Results_082619.RData")h.spec = list(theta=list(prior='pccor1', param=c(0, 0.9)))
YearPrior = list(prec = list(prior="pc.prec", param = c(1, 0.0001)))
iid.prior = list(theta=list(prior = "normal", param=c(0, 5))) #LandFire prior
Frm0 = PA ~ -1 + intercept1 +
#f(field1, #Turn off space-time effect
# model=spde,
# group = field1.group,
# control.group=list(model="ar1"),
# hyper=h.spec) +
f(vDist,
model="iid",
hyper = iid.prior) +
f(CBD,
model="iid",
hyper = iid.prior) +
f(EVC,
model="iid",
hyper = iid.prior) +
f(EVH,
model="iid",
hyper = iid.prior) +
f(EVT,
model="iid",
hyper = iid.prior) +
f(Year,
model="iid",
hyper = iid.prior) +
f(Sex,
model="iid",
hyper = iid.prior) +
f(Age,
model="iid",
hyper = iid.prior) +
TMAX + PPT + loc.sc + maj.sc
Model0 = inla(Frm0,
data = inla.stack.data(Stack1, spde=spde0),
family = "binomial",
verbose = TRUE,
control.fixed = list(prec = 0.01,
prec.intercept = 0.001),
control.predictor = list(
A = inla.stack.A(Stack1),
compute = TRUE,
link = 1),
control.inla = list(int.strategy = "eb"),
control.results = list(return.marginals.random = TRUE,
return.marginals.predictor = TRUE),
control.compute=list(dic = TRUE, cpo = TRUE, waic = TRUE)) summary(Model0)##
## Call:
## c("inla(formula = Frm0, family = \"binomial\", data = inla.stack.data(Stack1, ", " spde = spde0), verbose = TRUE, control.compute = list(dic = TRUE, ", " cpo = TRUE, waic = TRUE), control.predictor = list(A = inla.stack.A(Stack1), ", " compute = TRUE, link = 1), control.inla = list(int.strategy = \"eb\"), ", " control.results = list(return.marginals.random = TRUE, return.marginals.predictor = TRUE), ", " control.fixed = list(prec = 0.01, prec.intercept = 0.001))" )
##
## Time used:
## Pre-processing Running inla Post-processing Total
## 2.3389 1975.3925 1.9697 1979.7011
##
## Fixed effects:
## mean sd 0.025quant 0.5quant 0.975quant mode kld
## intercept1 -2.1889 1.2260 -4.5974 -2.1885 0.2146 -2.1875 0
## TMAX 0.2018 0.2399 -0.2691 0.2017 0.6725 0.2017 0
## PPT -0.0538 0.2421 -0.5292 -0.0538 0.4212 -0.0538 0
## loc.sc 0.1052 0.0212 0.0634 0.1053 0.1468 0.1054 0
## maj.sc -0.0100 0.0213 -0.0520 -0.0099 0.0315 -0.0097 0
##
## Random effects:
## Name Model
## vDist IID model
## CBD IID model
## EVC IID model
## EVH IID model
## EVT IID model
## Year IID model
## Sex IID model
## Age IID model
##
## Model hyperparameters:
## mean sd 0.025quant 0.5quant 0.975quant mode
## Precision for vDist 1.2292 0.5543 0.4756 1.1222 2.6110 0.9341
## Precision for CBD 0.9813 0.4157 0.4118 0.9016 2.0127 0.7633
## Precision for EVC 1.2985 0.4792 0.6143 1.2141 2.4692 1.0646
## Precision for EVH 1.0346 0.4434 0.4288 0.9495 2.1334 0.8016
## Precision for EVT 1.6129 0.5131 0.8320 1.5378 2.8286 1.3984
## Precision for Year 1.0279 0.3146 0.5405 0.9848 1.7673 0.9039
## Precision for Sex 0.6813 0.3419 0.2543 0.6038 1.5582 0.4820
## Precision for Age 0.2374 0.0703 0.1302 0.2270 0.4033 0.2078
##
## Expected number of effective parameters(std dev): 102.74(0.00)
## Number of equivalent replicates : 286.00
##
## Deviance Information Criterion (DIC) ...............: 11719.47
## Deviance Information Criterion (DIC, saturated) ....: 11719.47
## Effective number of parameters .....................: 100.24
##
## Watanabe-Akaike information criterion (WAIC) ...: 11713.76
## Effective number of parameters .................: 90.93
##
## Marginal log-Likelihood: -6007.32
## CPO and PIT are computed
##
## Posterior marginals for linear predictor and fitted values computed10 Spatial-temporal Model
h.spec = list(theta=list(prior='pccor1', param=c(0, 0.9)))
YearPrior = list(prec = list(prior="pc.prec", param = c(1, 0.0001)))
iid.prior = list(theta=list(prior = "normal", param=c(0, 5))) #LandFire prior
Frm1 = PA ~ -1 + intercept1 +
f(field1,
model=spde0,
group = field1.group,
control.group=list(model="ar1"),
hyper=h.spec) +
f(vDist,
model="iid",
hyper = iid.prior) +
f(CBD,
model="iid",
hyper = iid.prior) +
f(EVC,
model="iid",
hyper = iid.prior) +
f(EVH,
model="iid",
hyper = iid.prior) +
f(EVT,
model="iid",
hyper = iid.prior) +
f(Year,
model="iid",
hyper = iid.prior) +
f(Age,
model="iid",
hyper = iid.prior) +
f(Sex,
model="iid",
hyper = iid.prior) +
TMAX + PPT + loc.sc + maj.sc
#theta1 = Model1$internal.summary.hyperpar$mean
theta1 = c(4.556413, 0.969241, 5.673482, 0.116404, -0.383791, 0.367687, -0.260188, 0.592449, -0.019105, -1.560902, -0.425632)
Model1 = inla(Frm1,
data = inla.stack.data(Stack1),
family = "binomial",
verbose = TRUE,
control.fixed = list(prec = 0.01,
prec.intercept = 0.001),
control.predictor = list(
A = inla.stack.A(Stack1),
compute = TRUE,
link = 1),
control.mode = list(restart = TRUE, theta = theta1),
control.inla = list(strategy = "gaussian", int.strategy = "eb"),
control.results = list(return.marginals.random = TRUE,
return.marginals.predictor = TRUE),
control.compute=list(dic = TRUE, cpo = TRUE, waic = TRUE)) summary(Model1)##
## Call:
## c("inla(formula = Frm1, family = \"binomial\", data = inla.stack.data(Stack1), ", " verbose = TRUE, control.compute = list(dic = TRUE, cpo = TRUE, ", " waic = TRUE), control.predictor = list(A = inla.stack.A(Stack1), ", " compute = TRUE, link = 1), control.inla = list(strategy = \"gaussian\", ", " int.strategy = \"eb\"), control.results = list(return.marginals.random = TRUE, ", " return.marginals.predictor = TRUE), control.fixed = list(prec = 0.01, ", " prec.intercept = 0.001), num.threads = 8)")
##
## Time used:
## Pre-processing Running inla Post-processing Total
## 2.2321 56054.0006 4.4843 56060.7170
##
## Fixed effects:
## mean sd 0.025quant 0.5quant 0.975quant mode kld
## intercept1 -4.5470 1.3050 -7.1092 -4.5471 -1.9870 -4.5470 0
## TMAX 0.1778 0.2459 -0.3051 0.1778 0.6602 0.1778 0
## PPT -0.0027 0.2529 -0.4992 -0.0027 0.4934 -0.0027 0
## loc.sc 0.2031 0.0296 0.1450 0.2031 0.2612 0.2031 0
## maj.sc 0.0284 0.0286 -0.0277 0.0284 0.0845 0.0284 0
##
## Random effects:
## Name Model
## field1 SPDE2 model
## vDist IID model
## CBD IID model
## EVC IID model
## EVH IID model
## EVT IID model
## Year IID model
## Age IID model
## Sex IID model
##
## Model hyperparameters:
## mean sd 0.025quant 0.5quant 0.975quant mode
## Range for field1 93.2774 13.8092 69.3530 92.1533 123.4641 89.8641
## Stdev for field1 2.7089 0.3534 2.0850 2.6846 3.4717 2.6357
## GroupRho for field1 0.9829 0.0063 0.9676 0.9840 0.9921 0.9860
## Precision for vDist 1.2034 0.5161 0.4802 1.1109 2.4707 0.9416
## Precision for CBD 0.8161 0.3468 0.3361 0.7516 1.6748 0.6373
## Precision for EVC 1.6720 0.6580 0.7416 1.5548 3.2851 1.3462
## Precision for EVH 0.8816 0.3664 0.3696 0.8147 1.7816 0.6955
## Precision for EVT 1.9236 0.6945 0.8814 1.8209 3.5714 1.6243
## Precision for Year 1.0191 0.3285 0.5315 0.9665 1.8083 0.8706
## Precision for Age 0.2104 0.0593 0.1144 0.2039 0.3459 0.1913
## Precision for Sex 0.7599 0.3886 0.2801 0.6704 1.7601 0.5316
##
## Expected number of effective parameters(std dev): 471.33(0.00)
## Number of equivalent replicates : 62.34
##
## Deviance Information Criterion (DIC) ...............: 9486.29
## Deviance Information Criterion (DIC, saturated) ....: NULL
## Effective number of parameters .....................: 490.26
##
## Watanabe-Akaike information criterion (WAIC) ...: 9413.73
## Effective number of parameters .................: 396.72
##
## Marginal log-Likelihood: -5043.66
## CPO and PIT are computed
##
## Posterior marginals for linear predictor and fitted values computed11 Results
11.1 Age Effect
mic.df = as.data.frame(Model1$summary.random$Age)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("Age Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.2 Sex Effect
mic.df = as.data.frame(Model1$summary.random$Sex)[2:3,]
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=ID, y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("Sex Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.3 Year Effect
mic.df = as.data.frame(Model1$summary.random$Year)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("Year Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.4 VDIST
mic.df = as.data.frame(Model1$summary.random$vDist)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("VDIST Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.5 CBD
mic.df = as.data.frame(Model1$summary.random$CBD)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("CBD Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.6 EVC
mic.df = as.data.frame(Model1$summary.random$EVC)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("EVC Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.7 EVH
mic.df = as.data.frame(Model1$summary.random$EVH)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("EVH Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.8 EVT
mic.df = as.data.frame(Model1$summary.random$EVT)
names(mic.df) = c("ID", "Mean", "sd", "Q025", "Q50", "Q975")
ggplot(mic.df, aes(x=factor(ID), y=Mean)) +
geom_point(size=2, pch=19, col = "red") +
geom_linerange(aes(ymin=Q025, ymax=Q975), colour="black") +
geom_hline(yintercept = 0,
linetype = "dotted",
colour = "red",
size = 1) +
theme_classic() +
xlab("EVT Effect") +
ylab("Effect (logit)") +
theme(axis.title.y = element_text(face="bold", size=16),
axis.title.x = element_text(face="bold", size=14),
axis.text.y = element_text(face="bold", size=16),
axis.text.x = element_text(face="bold", size=10,
vjust=1, hjust=1, angle=45),
strip.text = element_text(face="bold", size = 20))
11.9 View Random Fields
F.s1 = cbind(Model1$summary.random$field1$mean,
field1$field1.group)
s1xmean = list()
s1xmean = split(F.s1[,1], F.s1[,2])
Pred.pnts = spTransform(Grd.pnt, nProj)
Grd_A = inla.spde.make.A(mesh, loc=Pred.pnts@coords)
for(i in 1:10){
Pred.pnts$m1RE = drop(Grd_A %*% s1xmean[[i]])
tmp.r = rasterize(spTransform(Pred.pnts, proj4string(UP)),
Domain.r,
"m1RE",
background = NA)
if(i == 1){M0.stk = tmp.r}
else{M0.stk = stack(M0.stk, tmp.r)}
}
names(M0.stk) = paste("Knot", 1:10, sep="_")range(M0.stk)## class : RasterBrick
## dimensions : 317, 698, 221266, 2 (nrow, ncol, ncell, nlayers)
## resolution : 0.01, 0.01 (x, y)
## extent : -90.41829, -83.43829, 45.09269, 48.26269 (xmin, xmax, ymin, ymax)
## coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0
## data source : in memory
## names : range_min, range_max
## min values : -4.209383, -3.752171
## max values : 4.449773, 5.016270rng = seq(-4.25, 5.21, 0.01)
mCols = c("#A50026", "#D73027", "#F46D43", "#FDAE61", "#E0F3F8", "#ABD9E9", "#74ADD1", "#4575B4","#313695")
#brewer.pal(11, "RdBu")
cr = colorRampPalette(c(rev(mCols)),
bias = 1.4, space = "rgb")
RF0osi = levelplot(M0.stk,
margin = FALSE,
layout=c(2, 5),
xlab = " ", ylab = NULL,
col.regions = cr, at = rng,
colorkey = list(labels=list(cex=1.5),
space = "right"),
par.settings = list(axis.line = list(col = "black"),
strip.background = list(col = 'transparent'),
strip.border = list(col = 'transparent')),
scales = list(cex = 1.5)) +
latticeExtra::layer(sp.polygons(UP, col = "black", lwd = 1))
RF0osi