510_DataAnalysis/EcozoneContinent.R at master · sschooler/510_DataAnalysis · GitHub

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
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
# Sarah Schooler
# 2/16/19
# Continent vs. Ecozone comparisons
# 10 species
require(rgdal)
require(rgeos)
require(raster)
require(enmSdm)
require(dismo)
species.df.full <- read.csv("Species_Table_021619.csv", row.names = 1, stringsAsFactors = FALSE)
species.df <- read.csv("simple.species.csv", row.names = 1, stringsAsFactors = FALSE)

# Sample DF
sample.df <- species.df[c(154, 376, 393, 19, 409),]
sample.df

# Read get species points, buffer them based around continents and ecoregions.
# This loop will also create the maxent data frame, so we need climate as well.
ecoregion <- readOGR(dsn = "./SpatialFiles/", "Sm_Ecoregions")
ecoregion.raster <- raster("ecoregionraster10m.gri")
ecoregions.df <- levels(ecoregion.raster)[[1]]

clim <- brick("clim10m")
require(raster)
require(dplyr)
for (i in 1:nrow(sample.df)){
   species <- sample.df[i, 1] #ith species
   print(species)
   print(i)
   # spatial points file that I made earlier
   species.pts.spdf <- readOGR(dsn = paste0("./Species_Pts/", gsub(" ", ".", species), "/"),
                                         layer = paste0(gsub(" ", ".", species),"sp.pts"))
   species.pts.thinned <- elimCellDups(SpatialPoints(coordinates(species.pts.orig),
                                                  proj4string = crs(clim)), clim[[5]])
   #species.pts <- over(species.pts.thinned, species.pts.orig)
   #species.pts.spdf <- SpatialPointsDataFrame(coords = data.frame(species.pts$dcmlLng, species.pts$dcmlLtt),
   #                                          proj4string=gbif.prj, data=species.pts)
   # get information about ecoregions that each point lies within, creating a buffer of
   # 10000 to avoid getting NAs because of coarse borders and islands

   #sp.buffer <- buffer(species.pts.spdf, width = 10000, dissolve = FALSE, progress= TRUE)
   #print("Buffered, overlapping")
   sp.over.df <- as.data.frame(extract(y = SpatialPoints(coordinates(species.pts.spdf), proj4string=gbif.prj),
                      buffer = 10000, x = ecoregion.raster, small = TRUE, df = TRUE,
                      progress = "text"))

   sp.over.sort <- sp.over.df[order(sp.over.df$layer),]
   sp.over <- sp.over.sort[!duplicated(sp.over.sort$ID),]
   sp.over <- merge(x = sp.over, y = ecoregions.df[,1:3], all.x = TRUE, by.x = "layer", by.y = "ID")
   #sp.over <- over(sp.buffer, ecoregion[,c("region", "biome")])
   print("Extract 1 Finished")
   # need to do this stuff to keep the data frame nice
   species.pts.spdf$ecoregion <- sp.over$region.names
   species.pts.spdf$biome <- sp.over$biome.names
   # which regions/biomes do the species reside in?
   regions <- unique(species.pts.spdf$ecoregion)
   biomes <- unique(species.pts.spdf$biome)
   # clipping the ecoregions layer so we only have the regions/biomes
   # where the species has been found
   these.regions.spdf <- ecoregion[as.vector(!is.na(match(ecoregion$region,regions))),]
   these.biomes.spdf <- ecoregion[as.vector(!is.na(match(ecoregion$biome,biomes))),]
   # generate random points, same number as observations
   print("Generating random points")
   random.pts.region <- spsample(these.regions.spdf, n=10000, type="random")
   random.pts.biome <- spsample(these.biomes.spdf, n=nrow(species.pts.spdf), type = "random")

   # I want to save this info now
   species.pts.df <- as.data.frame(species.pts.spdf@data)
   write.csv(species.pts.df, paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                 gsub(" ", ".", species), ".editeddata.csv", sep = ""))
   print("Extracting...")

   # making a two simple data frames for maxent; region and biome
   simpledata.df <- cbind(rep(1, nrow(species.pts.df)),species.pts.df[,1:2])
   colnames(simpledata.df) <- c("resp", "x", "y")

   random.df.region <- data.frame("x"=random.pts.region@coords[,1], "y"=random.pts.region@coords[,2],"resp"=0)
   for.maxent.pts.region <- rbind(simpledata.df, random.df.region)

   random.df.biome <- data.frame("x"=random.pts.biome@coords[,1], "y"=random.pts.biome@coords[,2],"resp"=0)
   for.maxent.pts.biome <- rbind(simpledata.df, random.df.biome)

   # Extract climate data

   df.clim.region <- as.data.frame(extract(clim, y = SpatialPoints(coordiantes(for.maxent.pts.region), proj4string=gbif.prj)))
   df.clim.biome <- as.data.frame(extract(clim, cbind(for.maxent.pts.biome$x, for.maxent.pts.biome$y)))
   print(paste(as.character(species), "climate extraction completed"))
   # gotta rename the climate variable names for the columns.
   bionames <- c("annualmeantemp", "meandiurnalrange", "isothermality", "tempseasonality",
              "maxtempwarmestmonth", "mintempcoldestmonth", "tempannualrange",
              "meantempwettestquart", "meantempdriestquart",
              "meantempwarmestquart", "meantempcoldestquart",
              "annualprecip", "precipwettestmo", "precipdriestmo", "precipseasonality",
              "precipwettestquart", "precipdriestquart", "precipwarmestquart",
              "precipcoldestquart")
   colnames(df.clim.region) <- bionames
   colnames(df.clim.biome) <- bionames
   # Binds climate data and long lat and pres together, in a way that the
   # maxent function likes
   for.maxent.full.region <- cbind(for.maxent.pts.region, df.clim.region)
   for.maxent.full.biome <- cbind(for.maxent.pts.biome, df.clim.biome)
   # Save 'em!
   write.csv(for.maxent.full.region, paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                    gsub(" ", ".", species), ".maxentdatafile.region.csv", sep = ""))

   write.csv(for.maxent.full.biome, paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                    gsub(" ", ".", species), ".maxentdatafile.biome.csv", sep = ""))

}


# Want to compare temperatures at occupied and random locations
par(mfrow = c(2, 5), oma = c(0,2,2,0), mar = c(2, 1, 3, .5), mgp = c(3,.5,0), font.main = 1)
for (i in 1:nrow(sample.df)){
   species <- sample.df[i, 1]
   for.maxent.full.region <- read.csv(paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                            gsub(" ", ".", species), ".maxentdatafile.region.csv", sep = ""), row.names = 1)
   for.maxent.full.biomes <- read.csv(paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                            gsub(" ", ".", species), ".maxentdatafile.biome.csv", sep = ""), row.names = 1)
   compare <- data.frame("Region.Maxtemp" = for.maxent.full.region$maxtempwarmestmonth,
                         "Biome.Maxtemp" = for.maxent.full.biomes$maxtempwarmestmonth,
                         "Region.Mintemp" = for.maxent.full.region$mintempcoldestmonth,
                         "Biome.Mintemp" = for.maxent.full.biomes$mintempcoldestmonth)
   boxplot.matrix(as.matrix(compare[,1:2]), main = paste0(sample.df[i, 4], "\n", species),
                  sub = sample.df[i, 3],  names = c("Region", "Biome"))
   if (i == 1 | i == 6) {mtext(side = 2, text = "Temperature (ºC)", line = 1.5, cex = .7)}
   if (i == 1) {mtext("Random Points Maximum Temperatures", outer = TRUE)}

}


par(mfrow = c(2, 5), oma = c(0,2,2,0), mar = c(2, 1, 3, .5), mgp = c(3,.5,0), font.main = 1)
for (i in 1:nrow(sample.df)){
   species <- sample.df[i, 1]
   for.maxent.full.region <- read.csv(paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                            gsub(" ", ".", species), ".maxentdatafile.region.csv", sep = ""), row.names = 1)
   for.maxent.full.biomes <- read.csv(paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                            gsub(" ", ".", species), ".maxentdatafile.biome.csv", sep = ""), row.names = 1)
   compare <- data.frame("Region.Maxtemp" = for.maxent.full.region$maxtempwarmestmonth,
                         "Biome.Maxtemp" = for.maxent.full.biomes$maxtempwarmestmonth,
                         "Region.Mintemp" = for.maxent.full.region$mintempcoldestmonth,
                         "Biome.Mintemp" = for.maxent.full.biomes$mintempcoldestmonth)
   boxplot.matrix(as.matrix(compare[,3:4]), main = paste0(sample.df[i, 4], "\n", species),
                  sub = sample.df[i, 3],  names = c("Region", "Biome"))
   if (i == 1 | i == 6) {mtext(side = 2, text = "Temperature (ºC)", line = 1.5, cex = .7)}
   if (i == 1) {mtext("Random Points Minimum Temperatures", outer = TRUE)}

}


species.df.full <- read.csv("Species_Table_021619.csv", row.names = 1, stringsAsFactors = FALSE)
species.df <- read.csv("simple.species.csv", row.names = 1, stringsAsFactors = FALSE)
compare <- data.frame("species" = character(), "thresholdRegion" = numeric(), "SuitRegion" = numeric(),
                      "AreaRegion" = numeric(), "thresholdBiome" = numeric(), "SuitBiome" = numeric(),
                      "AreaBiome" = numeric())
gbif.prj <- CRS("+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0")


require(DescTools)

species.df <- read.csv("simple.species.csv", row.names = 1, stringsAsFactors = FALSE)
species.df.full <- read.csv("Species_Table_021619.csv", row.names = 1, stringsAsFactors = FALSE)
df <- data.frame("species" = character(), "full.area" = numeric(), "above.lower.crit" = numeric(), "below.lower.crit" = numeric(),
                 "above.upper.crit" = numeric(), "below.upper.crit" = numeric(),
                 "area.between.crits" = numeric())
pts.climate.df <- data.frame("species" = character(), "below.maxtemp" = numeric(),
                             "above.maxtemp" = numeric(), "below.mintemp" = numeric(),
                             "above.mintemp" = numeric(), "maxtemp.loc" = numeric(),
                             "mintemp.loc" = numeric(), "maxcrit" = numeric(), "mincrit" = numeric(),
                             "nobs" = integer())
sample.df <- species.df[c(154, 376, 393, 19, 409, 361, 131, 26, 281, 133),]
# type in biome or region here to get appropriate points
bor = "biome"
for (i in 1:nrow(sample.df)){
   species = sample.df[i, 1]
   print(species)
   for.maxent.full <- read.csv(paste("./Species_Pts/", gsub(" ", ".", species), "/",
                                     gsub(" ", ".", species), ".maxentdatafile.",bor,".csv", sep = ""), row.names = 1)
   upper.crit <- species.df.full$UCT...C.[species.df.full$Species==species]
   lower.crit <-species.df.full$LCT...C.[species.df.full$Species==species]
   predsnums <- c(1, 2, 5, 6, 7)
   preds <- for.maxent.full[,predsnums+3]
   resp <- for.maxent.full[,"resp"]
   thisRegMult <- 1
   params <- c(paste("betamultiplier=", thisRegMult, sep=""), "jackknife=false")
   dir.create(paste0("./MaxentModels/",gsub(" ", ".", species)))
   maxent.model <- maxent(x = preds, p = resp, path = paste0("./MaxentModels/",gsub(" ", ".", species)),
                          args = params, silent = FALSE)
   #########
   curPredsRaster <- clim[[predsnums]]
   names(curPredsRaster) <- colnames(preds)
   sp.raster <- predict(model = maxent.model, object = curPredsRaster, progress = "text",
                        filename = paste0("./MaxentModels/",gsub(" ", ".", species),"/", gsub(" ", ".", species),
                                          ".full.tif"), overwrite = TRUE)
   par(mfrow = c(1,1))
   plot(sp.raster, main = "Region Amadina fasciata")
   species.pts.spdf <- readOGR(dsn = paste0("./Species_Pts/", gsub(" ", ".", species), "/"),
                               layer = paste0(gsub(" ", ".", species),"sp.pts"))
   random.pts.spdf <- SpatialPoints(coordinates(for.maxent.full[for.maxent.full$resp==0,][,2:3]), proj4string=gbif.prj)
   plot(random.pts.spdf, add = TRUE)
   plot(species.pts.spdf, add = TRUE, col = "red")
   ######
   par(mfrow = c(2, 2))
   areas.row <- AreaUnderResponseCurves(x = maxent.model, var = c("annualmeantemp", "maxtempwarmestmonth", "mintempcoldestmonth"), at = median,
                                        expand = 10, data = NULL, fun = predict, upper.crit = upper.crit, lower.crit = lower.crit)
   df <- rbind(cbind("species" = species, areas.row), df)
   write.csv(df, paste0("./SummaryTables/response.suitability.",bor,".csv"))


   #Evaluating climate at occurence points
   species.pts.climate <- dplyr::filter(for.maxent.full, for.maxent.full$resp == 1)
   species.pts.climate <- species.pts.climate[!is.na(species.pts.climate$maxtempwarmestmonth),]
   below.maxtemp <- sum(species.pts.climate$maxtempwarmestmonth <= upper.crit)
   above.mintemp <- sum(species.pts.climate$mintempcoldestmonth >= lower.crit)
   above.maxtemp <- sum(species.pts.climate$maxtempwarmestmonth > upper.crit)
   below.mintemp <- sum(species.pts.climate$mintempcoldestmonth < lower.crit)


   maxtemp.loc <- mean(species.pts.climate$maxtempwarmestmonth)
   mintemp.loc <- mean(species.pts.climate$mintempcoldestmonth)
   meantemp.loc <- mean(species.pts.climate$annualmeantemp)
   nobs <- nrow(species.pts.climate)

   pts.climate.df <- rbind(data.frame("species" = species, "below.maxtemp" = below.maxtemp,
                                      "above.maxtemp" = above.maxtemp, "below.mintemp" = below.mintemp,
                                      "above.mintemp" = above.mintemp, "maxtemp.loc" = maxtemp.loc,
                                      "mintemp.loc" = mintemp.loc, "meantemp.loc" = meantemp.loc,
                                      "nobs" = nobs, "maxcrit" = upper.crit, "mincrit" = lower.crit), pts.climate.df)
   write.csv(pts.climate.df, "./SummaryTables/Continent V Ecoregion/ptssuitability.csv")
}

### misc code
#region.spdf <- SpatialPointsDataFrame(coords = data.frame(for.maxent.full.region$x, for.maxent.full.region$y),
#                                   proj4string=gbif.prj, data=for.maxent.full.region)

plot(region.spdf, add = TRUE)

ptsuitability.df <- read.csv("./SummaryTables/Continent V Ecoregion/ptssuitability.csv", row.names = 1)
respsuitability.df <- read.csv("./SummaryTables/Continent V Ecoregion/response.suitability.csv", row.names=1)
boxplot(data = ptsuitability.df, as.matrix(data.frame("Pts Max Temp" = ptsuitability.df$maxtemp.loc,
                                                      "Pts Min Temp" = ptsuitability.df$mintemp.loc,
                                                      "Maximum Critical" = ptsuitability.df$maxcrit,
                                                      "Minimum Critical" = ptsuitability.df$mincrit)),
        ylab = "Mean Temperature Accross Points (ªC)", main = "Mean Values for 10 Species")

# Percent of observations that exist in maximum/minimum temperature observations less than or
# equal to critical temperatures

boxplot(data = ptsuitability.df, as.matrix(data.frame("perc.obs.lt.mintemp" = ptsuitability.df$below.mintemp/ptsuitability.df$nobs,
                                                      "perc.obs.gt.maxtemp" = ptsuitability.df$above.maxtemp/ptsuitability.df$nobs)),
        ylab = "Mean Percent Observations", main = "Threshold Comparison", names = c("% <= Min Crit", "% >= Max Crit"))

par(mfrow = c(1, 3), oma = c(0, 1, 2, 0))
boxplot(data = respsuitability.df, below.lower.crit ~ region.biome, ylab = "Area Under Curve", main = "Below Min Crit")
boxplot(data = respsuitability.df, above.upper.crit ~ region.biome, main = "Above Max Crit")
boxplot(data = respsuitability.df, area.between.crits ~ region.biome, main = "Between Crit Temps")
mtext("Area Under Response Curves", outer = TRUE, cex = 1)