library(knitr)
library(dismo)
library(gbm)
library(randomForest)
library(ranger) # for the Random Forest
library(terra)
terraOptions(tempdir='big_data/temp')
library(sf)
library(tidyverse)Eira barbara Variable Selection
Variable Selection for Eira barbara, using the data generated in the previous step.
- R Libraries
28 Predictors
bio1: Annual Mean Temperature, from Bioclimatic variables (WorldClim V2.1)bio2: Mean Diurnal Range (Mean of monthly (max temp - min temp)), from Bioclimatic variables (WorldClim V2.1)bio3: Isothermality (BIO2/BIO7) (×100), from Bioclimatic variables (WorldClim V2.1)bio4: Temperature Seasonality (standard deviation ×100), from Bioclimatic variables (WorldClim V2.1)bio5: Max Temperature of Warmest Month, from Bioclimatic variables (WorldClim V2.1)bio6: Min Temperature of Coldest Month, from Bioclimatic variables (WorldClim V2.1)bio7: Temperature Annual Range (BIO5-BIO6), from Bioclimatic variables (WorldClim V2.1)bio8: Mean Temperature of Wettest Quarter, from Bioclimatic variables (WorldClim V2.1)bio9: Mean Temperature of Driest Quarter, from Bioclimatic variables (WorldClim V2.1)bio10: Mean Temperature of Warmest Quarter, from Bioclimatic variables (WorldClim V2.1)bio11: Mean Temperature of Coldest Quarter, from Bioclimatic variables (WorldClim V2.1)bio12: Annual Precipitation, from Bioclimatic variables (WorldClim V2.1)bio13: Precipitation of Wettest Month, from Bioclimatic variables (WorldClim V2.1)bio14: Precipitation of Driest Month, from Bioclimatic variables (WorldClim V2.1)bio15: Precipitation Seasonality (Coefficient of Variation), from Bioclimatic variables (WorldClim V2.1)bio16: Precipitation of Wettest Quarter, from Bioclimatic variables (WorldClim V2.1)bio17: Precipitation of Driest Quarter, from Bioclimatic variables (WorldClim V2.1)bio18: Precipitation of Warmest Quarter, from Bioclimatic variables (WorldClim V2.1)bio19: Precipitation of Coldest Quarter, from Bioclimatic variables (WorldClim V2.1)elev: Elevation (WorldClim V2.1 SRTM elevation data), from Bioclimatic variables (WorldClim V2.1)urban: Urban and Built-up Lands, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))barren: Barren, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))water: Water Bodies, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))savanna: Savannas, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))woodysavanna: Woody savannas, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))wetland: Permanent Wetlands, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))grass: Grasslands, from Land cover LC1 (MODIS TERRA LandCover_Type_Yearly_500m (MCD12Q1))npp: Net Primary Production (NPP) (MODIS TERRA Net_PP_GapFil_Yearly_500m (M*D17A3HGF))tree: Percentage of Tree Cover, from Vegetation Continuous Fields (MODIS TERRA Veg_Cont_Fields_Yearly_250m (MOD44B))nontree: Percentage of No Tree Cover, from Vegetation Continuous Fields (MODIS TERRA Veg_Cont_Fields_Yearly_250m (MOD44B))nonveg: Percentage of Non Tree Vegetation Cover, from Vegetation Continuous Fields (MODIS TERRA Veg_Cont_Fields_Yearly_250m (MOD44B))
Code
bio <- rast('big_data/bio_high.tif')
elev <- rast('big_data/elev_high.tif')
urban <- rast('big_data/urban_high.tif') %>% resample(., elev)
barren <- rast('big_data/barren_high.tif') %>% resample(., elev)
water <- rast('big_data/water_high.tif') %>% resample(., elev)
savanna <- rast('big_data/savanna_high.tif') %>% resample(., elev)
woodysavanna <- rast('big_data/woodysavanna_high.tif') %>% resample(., elev)
wetland <- rast('big_data/wetland_high.tif') %>% resample(., elev)
grass <- rast('big_data/grass_high.tif') %>% resample(., elev)
npp <- rast('big_data/npp_high.tif') %>% resample(., elev)
names(npp) <- 'npp'
tree <- rast('big_data/tree_high.tif') %>% resample(., elev)
nontree <- rast('big_data/nontree_high.tif') %>% resample(., elev)
nonveg <- rast('big_data/nonveg_high.tif') %>% resample(., elev)
env <- c(bio, elev, urban, barren, water,
savanna, woodysavanna, wetland,
grass, npp, tree, nontree, nonveg) %>% scale()
rm(bio, elev, urban, barren, water,
savanna, woodysavanna, wetland,
grass, npp, tree, nontree, nonveg)
gc()Eira barbara’ preferences
We will use the data from both periods
Code
PA_time1 <- readRDS('data/species_POPA_data/PA_ebarbara_time1_blobs.rds')%>% ungroup()
PA_time2 <- readRDS('data/species_POPA_data/PA_ebarbara_time2_blobs.rds')%>% ungroup()
PA_time1 %>% st_drop_geometry() %>% head() %>% kable()| ID | presence | temporalSpan | effort | blobArea |
|---|---|---|---|---|
| 1 | 0 | 1365 days | 1043 | 241873943 [m^2] |
| 2 | 1 | 1769 days | 2404 | 537081878 [m^2] |
| 3 | 1 | 6222 days | 11078 | 2241827859 [m^2] |
| 4 | 0 | 720 days | 1291 | 162791738 [m^2] |
| 5 | 0 | 1769 days | 1779 | 756798182 [m^2] |
| 6 | 0 | 304 days | 3300 | 19990863 [m^2] |
Code
PA_time2 %>% st_drop_geometry() %>% head() %>% kable()| ID | presence | temporalSpan | effort | blobArea |
|---|---|---|---|---|
| 1 | 1 | 510 days | 11095 | 2.229446e+09 [m^2] |
| 2 | 1 | 150 days | 12276 | 1.542295e+03 [m^2] |
| 3 | 1 | 407 days | 1585 | 1.049520e-01 [m^2] |
| 4 | 1 | 869 days | 1440 | 7.494574e+01 [m^2] |
| 5 | 1 | 70 days | 400 | 2.546836e+01 [m^2] |
| 6 | 1 | 756 days | 7164 | 2.398904e-01 [m^2] |
Presence-absence data for the second period
Preparation of data for the tests
Code
# combine pre and pos datasets
PA.data <- st_join(PA_time1, PA_time2 %>% dplyr::select(presence), left = T) %>%
group_by(ID) %>%
mutate(presence=max(presence.x, presence.y, na.rm = T))
# calculate area, coordinates, and extract env predictors for each blob
PA.coords <- st_coordinates(st_centroid(PA.data)) %>% as_tibble()
PA.area <- as.numeric(PA.data$blobArea)
PA.env <- terra::extract(x = env, y = vect(PA.data),
fun = mean, rm.na=T) %>%
mutate(across(where(is.numeric), ~ifelse(is.nan(.), NA, .)))
## the data
PA <- data.frame(PA.coords,
area = PA.area,
presabs = PA.data$presence,
env = PA.env)Correlation between variables
Code
PA %>% filter(!if_any(everything(), is.na)) %>% dplyr::select(-c(1:5)) %>% cor() %>% kable()| env.bio_1 | env.bio_2 | env.bio_3 | env.bio_4 | env.bio_5 | env.bio_6 | env.bio_7 | env.bio_8 | env.bio_9 | env.bio_10 | env.bio_11 | env.bio_12 | env.bio_13 | env.bio_14 | env.bio_15 | env.bio_16 | env.bio_17 | env.bio_18 | env.bio_19 | env.elev | env.urban | env.barren | env.water | env.savanna | env.woodysavanna | env.wetland | env.grass | env.npp | env.tree | env.nontree | env.nonveg | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| env.bio_1 | 1.0000000 | 0.9296441 | 0.9683961 | 0.2245524 | 0.3650713 | -0.0974913 | 0.2598362 | 0.4133726 | -0.1088920 | -0.2171595 | 0.1956382 | -0.2049978 | 0.6915905 | -0.6813646 | 0.8361199 | 0.9132283 | -0.5533539 | 0.8696582 | 0.9064849 | -0.6579087 | -0.2201133 | -0.0977410 | 0.2893823 | -0.0024650 | 0.0280840 | 0.1505896 | -0.1029622 | -0.5423448 | 0.2241057 | 0.1944363 | 0.3216861 |
| env.bio_2 | 0.9296441 | 1.0000000 | 0.8201186 | 0.1070940 | 0.1263286 | -0.0188141 | 0.1075105 | 0.1704404 | -0.0270968 | -0.2734955 | 0.1868085 | -0.1426003 | 0.4381869 | -0.3771548 | 0.9194218 | 0.7840670 | -0.3309997 | 0.8667005 | 0.7893311 | -0.8088566 | -0.2416197 | -0.1129521 | 0.2673516 | 0.0269795 | 0.0437261 | 0.2158389 | -0.1911302 | -0.5957826 | 0.2394162 | 0.2026016 | 0.3128536 |
| env.bio_3 | 0.9683961 | 0.8201186 | 1.0000000 | 0.3145337 | 0.4841739 | -0.0902739 | 0.3019023 | 0.5416489 | -0.1019188 | -0.2194668 | 0.2423853 | -0.2688312 | 0.8182906 | -0.8385889 | 0.7284579 | 0.9531144 | -0.6805953 | 0.7863656 | 0.9351986 | -0.5387853 | -0.1778445 | -0.0766928 | 0.3146130 | -0.0212454 | 0.0151050 | 0.0830030 | -0.0941385 | -0.4287304 | 0.2510269 | 0.1763470 | 0.3326884 |
| env.bio_4 | 0.2245524 | 0.1070940 | 0.3145337 | 1.0000000 | 0.7642354 | 0.7267723 | -0.4858212 | 0.7801517 | 0.7257054 | 0.3332077 | 0.7446606 | -0.5595604 | 0.5005643 | -0.4148608 | -0.0696912 | 0.4462757 | -0.6123342 | 0.0898132 | 0.3375557 | -0.2702648 | 0.0226405 | -0.1373670 | 0.3077364 | -0.4152737 | -0.1495779 | -0.1146293 | -0.1512988 | 0.2195306 | 0.5598284 | -0.0677038 | 0.2675771 |
| env.bio_5 | 0.3650713 | 0.1263286 | 0.4841739 | 0.7642354 | 1.0000000 | 0.1933128 | 0.1171199 | 0.9784670 | 0.1835373 | 0.3598166 | 0.4346112 | -0.3516015 | 0.6733107 | -0.6659381 | -0.0034808 | 0.5031871 | -0.6367356 | 0.2330854 | 0.4266622 | -0.0644255 | 0.0457209 | -0.1574742 | 0.2459976 | -0.3587865 | -0.0619659 | -0.1330635 | 0.1157962 | 0.0741961 | 0.3196852 | 0.0073694 | 0.2150625 |
| env.bio_6 | -0.0974913 | -0.0188141 | -0.0902739 | 0.7267723 | 0.1933128 | 1.0000000 | -0.8648488 | 0.1913458 | 0.9965494 | 0.1837194 | 0.7740640 | -0.5300579 | 0.0892696 | 0.1186764 | -0.1996677 | 0.1249108 | -0.3001855 | -0.1378122 | 0.0099549 | -0.3037383 | -0.0295608 | -0.0459756 | 0.2895092 | -0.3454885 | -0.1352915 | -0.0004524 | -0.3242425 | 0.2813411 | 0.5902548 | -0.0208942 | 0.2560326 |
| env.bio_7 | 0.2598362 | 0.1075105 | 0.3019023 | -0.4858212 | 0.1171199 | -0.8648488 | 1.0000000 | 0.1220172 | -0.8782263 | -0.1345391 | -0.5756748 | 0.4185941 | 0.2440087 | -0.3840589 | 0.2496259 | 0.1033349 | 0.0482055 | 0.2443413 | 0.1945389 | 0.2985608 | 0.0225795 | -0.0051382 | -0.1606119 | 0.2264854 | 0.1850820 | -0.0391527 | 0.2617882 | -0.2260104 | -0.4273476 | 0.0504557 | -0.1488980 |
| env.bio_8 | 0.4133726 | 0.1704404 | 0.5416489 | 0.7801517 | 0.9784670 | 0.1913458 | 0.1220172 | 1.0000000 | 0.1819016 | 0.3180265 | 0.4533295 | -0.4161798 | 0.7057661 | -0.7191324 | 0.0339333 | 0.5715997 | -0.6962200 | 0.2568050 | 0.4948495 | -0.1003456 | 0.0642175 | -0.1537781 | 0.2693363 | -0.3170436 | -0.0375880 | -0.1611847 | 0.0346935 | 0.0835534 | 0.3395489 | -0.0039337 | 0.2253595 |
| env.bio_9 | -0.1088920 | -0.0270968 | -0.1019188 | 0.7257054 | 0.1835373 | 0.9965494 | -0.8782263 | 0.1819016 | 1.0000000 | 0.1840244 | 0.7748265 | -0.5272262 | 0.0770351 | 0.1295226 | -0.2067214 | 0.1137978 | -0.2912206 | -0.1496451 | -0.0013905 | -0.3043808 | -0.0294760 | -0.0483047 | 0.2836730 | -0.3479909 | -0.1389403 | -0.0005642 | -0.3307918 | 0.2925463 | 0.5897787 | -0.0290542 | 0.2499643 |
| env.bio_10 | -0.2171595 | -0.2734955 | -0.2194668 | 0.3332077 | 0.3598166 | 0.1837194 | -0.1345391 | 0.3180265 | 0.1840244 | 1.0000000 | -0.1092807 | -0.1343382 | -0.0811383 | 0.1021154 | -0.4429341 | -0.1997582 | -0.0648619 | -0.0145843 | -0.3385074 | 0.1685297 | 0.2582501 | -0.1263971 | -0.1456711 | -0.2570984 | 0.0182359 | -0.0638922 | 0.0956723 | 0.2583383 | -0.0288362 | -0.1901205 | -0.1873301 |
| env.bio_11 | 0.1956382 | 0.1868085 | 0.2423853 | 0.7446606 | 0.4346112 | 0.7740640 | -0.5756748 | 0.4533295 | 0.7748265 | -0.1092807 | 1.0000000 | -0.6357030 | 0.4318087 | -0.2273811 | -0.0001334 | 0.4547914 | -0.5733417 | 0.0224191 | 0.3619071 | -0.3934250 | -0.0896497 | -0.0345635 | 0.4128212 | -0.3398411 | -0.0967487 | -0.0340330 | -0.2794210 | 0.1254955 | 0.6191653 | 0.0444401 | 0.3673542 |
| env.bio_12 | -0.2049978 | -0.1426003 | -0.2688312 | -0.5595604 | -0.3516015 | -0.5300579 | 0.4185941 | -0.4161798 | -0.5272262 | -0.1343382 | -0.6357030 | 1.0000000 | -0.3325137 | 0.3049036 | 0.1966475 | -0.5250598 | 0.8023981 | -0.0861157 | -0.3309889 | 0.3821173 | -0.0890791 | 0.0167159 | -0.4386322 | 0.3434812 | 0.0467406 | -0.0727980 | 0.3469681 | -0.1737652 | -0.5951082 | -0.0668483 | -0.3700583 |
| env.bio_13 | 0.6915905 | 0.4381869 | 0.8182906 | 0.5005643 | 0.6733107 | 0.0892696 | 0.2440087 | 0.7057661 | 0.0770351 | -0.0811383 | 0.4318087 | -0.3325137 | 1.0000000 | -0.9118016 | 0.3261453 | 0.8255327 | -0.8073928 | 0.4789765 | 0.7641805 | -0.2005044 | -0.1124490 | -0.0021708 | 0.3366226 | -0.1551412 | -0.0112436 | -0.0695311 | -0.0158723 | -0.1472418 | 0.2873553 | 0.1459596 | 0.3206495 |
| env.bio_14 | -0.6813646 | -0.3771548 | -0.8385889 | -0.4148608 | -0.6659381 | 0.1186764 | -0.3840589 | -0.7191324 | 0.1295226 | 0.1021154 | -0.2273811 | 0.3049036 | -0.9118016 | 1.0000000 | -0.3065325 | -0.8003439 | 0.7896644 | -0.4447706 | -0.7650724 | 0.1075528 | 0.0491570 | 0.0151001 | -0.2590556 | 0.0693126 | 0.0098816 | 0.0774311 | -0.0168947 | 0.1143688 | -0.1926304 | -0.0852376 | -0.2445389 |
| env.bio_15 | 0.8361199 | 0.9194218 | 0.7284579 | -0.0696912 | -0.0034808 | -0.1996677 | 0.2496259 | 0.0339333 | -0.2067214 | -0.4429341 | -0.0001334 | 0.1966475 | 0.3261453 | -0.3065325 | 1.0000000 | 0.6102757 | -0.0543381 | 0.7628378 | 0.7072778 | -0.6630353 | -0.2740025 | -0.1002805 | 0.1226871 | 0.1672372 | 0.0462821 | 0.1653903 | -0.0906072 | -0.6102193 | 0.0503882 | 0.1630034 | 0.1918470 |
| env.bio_16 | 0.9132283 | 0.7840670 | 0.9531144 | 0.4462757 | 0.5031871 | 0.1249108 | 0.1033349 | 0.5715997 | 0.1137978 | -0.1997582 | 0.4547914 | -0.5250598 | 0.8255327 | -0.8003439 | 0.6102757 | 1.0000000 | -0.8241800 | 0.7110844 | 0.9320639 | -0.6213394 | -0.1491519 | -0.0630686 | 0.4173645 | -0.1202489 | 0.0029894 | 0.1034344 | -0.2159955 | -0.3303117 | 0.4202314 | 0.1793595 | 0.4141109 |
| env.bio_17 | -0.5533539 | -0.3309997 | -0.6805953 | -0.6123342 | -0.6367356 | -0.3001855 | 0.0482055 | -0.6962200 | -0.2912206 | -0.0648619 | -0.5733417 | 0.8023981 | -0.8073928 | 0.7896644 | -0.0543381 | -0.8241800 | 1.0000000 | -0.3509483 | -0.6692034 | 0.3091761 | -0.0078816 | 0.0078058 | -0.4383634 | 0.2711251 | 0.0293186 | -0.0121391 | 0.2074795 | -0.0198950 | -0.4936629 | -0.1095460 | -0.3848206 |
| env.bio_18 | 0.8696582 | 0.8667005 | 0.7863656 | 0.0898132 | 0.2330854 | -0.1378122 | 0.2443413 | 0.2568050 | -0.1496451 | -0.0145843 | 0.0224191 | -0.0861157 | 0.4789765 | -0.4447706 | 0.7628378 | 0.7110844 | -0.3509483 | 1.0000000 | 0.6106291 | -0.5930138 | -0.1337635 | -0.0844168 | 0.1904426 | -0.0413308 | 0.0667723 | 0.1977084 | -0.0870639 | -0.5298813 | 0.1649284 | 0.1344020 | 0.2359080 |
| env.bio_19 | 0.9064849 | 0.7893311 | 0.9351986 | 0.3375557 | 0.4266622 | 0.0099549 | 0.1945389 | 0.4948495 | -0.0013905 | -0.3385074 | 0.3619071 | -0.3309889 | 0.7641805 | -0.7650724 | 0.7072778 | 0.9320639 | -0.6692034 | 0.6106291 | 1.0000000 | -0.5894832 | -0.2207150 | -0.0580746 | 0.3417830 | 0.0150607 | 0.0015716 | 0.0673108 | -0.1438877 | -0.4089540 | 0.2748842 | 0.1941162 | 0.3490410 |
| env.elev | -0.6579087 | -0.8088566 | -0.5387853 | -0.2702648 | -0.0644255 | -0.3037383 | 0.2985608 | -0.1003456 | -0.3043808 | 0.1685297 | -0.3934250 | 0.3821173 | -0.2005044 | 0.1075528 | -0.6630353 | -0.6213394 | 0.3091761 | -0.5930138 | -0.5894832 | 1.0000000 | 0.1776808 | 0.0973969 | -0.2664306 | 0.0973135 | 0.0062860 | -0.1804121 | 0.3347777 | 0.3623028 | -0.3968661 | -0.0874815 | -0.3009313 |
| env.urban | -0.2201133 | -0.2416197 | -0.1778445 | 0.0226405 | 0.0457209 | -0.0295608 | 0.0225795 | 0.0642175 | -0.0294760 | 0.2582501 | -0.0896497 | -0.0890791 | -0.1124490 | 0.0491570 | -0.2740025 | -0.1491519 | -0.0078816 | -0.1337635 | -0.2207150 | 0.1776808 | 1.0000000 | -0.0109690 | -0.0607542 | -0.0470701 | 0.0318901 | -0.0606497 | -0.0753713 | 0.1344147 | -0.0366332 | -0.0745566 | -0.0958183 |
| env.barren | -0.0977410 | -0.1129521 | -0.0766928 | -0.1373670 | -0.1574742 | -0.0459756 | -0.0051382 | -0.1537781 | -0.0483047 | -0.1263971 | -0.0345635 | 0.0167159 | -0.0021708 | 0.0151001 | -0.1002805 | -0.0630686 | 0.0078058 | -0.0844168 | -0.0580746 | 0.0973969 | -0.0109690 | 1.0000000 | -0.0116914 | -0.0361940 | -0.0154035 | -0.0116054 | -0.0235731 | -0.0790421 | -0.0543190 | -0.0802550 | 0.0973328 |
| env.water | 0.2893823 | 0.2673516 | 0.3146130 | 0.3077364 | 0.2459976 | 0.2895092 | -0.1606119 | 0.2693363 | 0.2836730 | -0.1456711 | 0.4128212 | -0.4386322 | 0.3366226 | -0.2590556 | 0.1226871 | 0.4173645 | -0.4383634 | 0.1904426 | 0.3417830 | -0.2664306 | -0.0607542 | -0.0116914 | 1.0000000 | -0.1898651 | -0.0920385 | 0.1375010 | -0.1324898 | -0.3658775 | 0.7287007 | 0.7473932 | 0.9149495 |
| env.savanna | -0.0024650 | 0.0269795 | -0.0212454 | -0.4152737 | -0.3587865 | -0.3454885 | 0.2264854 | -0.3170436 | -0.3479909 | -0.2570984 | -0.3398411 | 0.3434812 | -0.1551412 | 0.0693126 | 0.1672372 | -0.1202489 | 0.2711251 | -0.0413308 | 0.0150607 | 0.0973135 | -0.0470701 | -0.0361940 | -0.1898651 | 1.0000000 | -0.1167850 | -0.1579102 | -0.1249173 | -0.1437237 | -0.4725033 | 0.1212126 | -0.2018991 |
| env.woodysavanna | 0.0280840 | 0.0437261 | 0.0151050 | -0.1495779 | -0.0619659 | -0.1352915 | 0.1850820 | -0.0375880 | -0.1389403 | 0.0182359 | -0.0967487 | 0.0467406 | -0.0112436 | 0.0098816 | 0.0462821 | 0.0029894 | 0.0293186 | 0.0667723 | 0.0015716 | 0.0062860 | 0.0318901 | -0.0154035 | -0.0920385 | -0.1167850 | 1.0000000 | -0.0847233 | -0.1104016 | 0.0006886 | -0.0777231 | -0.1006558 | -0.1058182 |
| env.wetland | 0.1505896 | 0.2158389 | 0.0830030 | -0.1146293 | -0.1330635 | -0.0004524 | -0.0391527 | -0.1611847 | -0.0005642 | -0.0638922 | -0.0340330 | -0.0727980 | -0.0695311 | 0.0774311 | 0.1653903 | 0.1034344 | -0.0121391 | 0.1977084 | 0.0673108 | -0.1804121 | -0.0606497 | -0.0116054 | 0.1375010 | -0.1579102 | -0.0847233 | 1.0000000 | -0.1258117 | -0.3612828 | 0.2205404 | 0.2479464 | 0.2796946 |
| env.grass | -0.1029622 | -0.1911302 | -0.0941385 | -0.1512988 | 0.1157962 | -0.3242425 | 0.2617882 | 0.0346935 | -0.3307918 | 0.0956723 | -0.2794210 | 0.3469681 | -0.0158723 | -0.0168947 | -0.0906072 | -0.2159955 | 0.2074795 | -0.0870639 | -0.1438877 | 0.3347777 | -0.0753713 | -0.0235731 | -0.1324898 | -0.1249173 | -0.1104016 | -0.1258117 | 1.0000000 | -0.1337146 | -0.4646464 | 0.2073061 | -0.1003525 |
| env.npp | -0.5423448 | -0.5957826 | -0.4287304 | 0.2195306 | 0.0741961 | 0.2813411 | -0.2260104 | 0.0835534 | 0.2925463 | 0.2583383 | 0.1254955 | -0.1737652 | -0.1472418 | 0.1143688 | -0.6102193 | -0.3303117 | -0.0198950 | -0.5298813 | -0.4089540 | 0.3623028 | 0.1344147 | -0.0790421 | -0.3658775 | -0.1437237 | 0.0006886 | -0.3612828 | -0.1337146 | 1.0000000 | -0.0468240 | -0.5293037 | -0.4653541 |
| env.tree | 0.2241057 | 0.2394162 | 0.2510269 | 0.5598284 | 0.3196852 | 0.5902548 | -0.4273476 | 0.3395489 | 0.5897787 | -0.0288362 | 0.6191653 | -0.5951082 | 0.2873553 | -0.1926304 | 0.0503882 | 0.4202314 | -0.4936629 | 0.1649284 | 0.2748842 | -0.3968661 | -0.0366332 | -0.0543190 | 0.7287007 | -0.4725033 | -0.0777231 | 0.2205404 | -0.4646464 | -0.0468240 | 1.0000000 | 0.2718941 | 0.7398183 |
| env.nontree | 0.1944363 | 0.2026016 | 0.1763470 | -0.0677038 | 0.0073694 | -0.0208942 | 0.0504557 | -0.0039337 | -0.0290542 | -0.1901205 | 0.0444401 | -0.0668483 | 0.1459596 | -0.0852376 | 0.1630034 | 0.1793595 | -0.1095460 | 0.1344020 | 0.1941162 | -0.0874815 | -0.0745566 | -0.0802550 | 0.7473932 | 0.1212126 | -0.1006558 | 0.2479464 | 0.2073061 | -0.5293037 | 0.2718941 | 1.0000000 | 0.7880283 |
| env.nonveg | 0.3216861 | 0.3128536 | 0.3326884 | 0.2675771 | 0.2150625 | 0.2560326 | -0.1488980 | 0.2253595 | 0.2499643 | -0.1873301 | 0.3673542 | -0.3700583 | 0.3206495 | -0.2445389 | 0.1918470 | 0.4141109 | -0.3848206 | 0.2359080 | 0.3490410 | -0.3009313 | -0.0958183 | 0.0973328 | 0.9149495 | -0.2018991 | -0.1058182 | 0.2796946 | -0.1003525 | -0.4653541 | 0.7398183 | 0.7880283 | 1.0000000 |
Variable Importance analyses
Simple GLM
Code
presabs.glm <- PA %>%
dplyr::select(-c(1,2,3,5)) %>%
filter(!is.na(env.elev)&!is.na(env.bio_1))
glm.full<- glm(presabs ~.,
family = "binomial",
data = presabs.glm)
summary(glm.full)
Call:
glm(formula = presabs ~ ., family = "binomial", data = presabs.glm)
Deviance Residuals:
Min 1Q Median 3Q Max
-2.8609 -0.7373 -0.3816 0.6826 2.4389
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -6.566e+00 1.122e+01 -0.585 0.558439
env.bio_1 2.660e+01 7.881e+00 3.375 0.000739 ***
env.bio_2 1.148e+01 7.799e+00 1.472 0.140916
env.bio_3 -4.750e+01 1.581e+01 -3.003 0.002669 **
env.bio_4 2.697e+00 2.039e+00 1.323 0.185931
env.bio_5 -5.395e+00 1.539e+00 -3.504 0.000458 ***
env.bio_6 -2.581e+00 1.608e+00 -1.605 0.108445
env.bio_7 1.325e+00 6.730e-01 1.970 0.048894 *
env.bio_8 2.647e+00 2.030e+00 1.304 0.192111
env.bio_9 3.252e+00 2.080e+00 1.564 0.117858
env.bio_10 2.311e-01 4.050e-01 0.571 0.568335
env.bio_11 2.377e-01 5.418e-01 0.439 0.660853
env.bio_12 1.991e+00 1.804e+00 1.104 0.269786
env.bio_13 -1.078e+00 1.361e+00 -0.792 0.428301
env.bio_14 -2.235e+01 8.404e+00 -2.660 0.007821 **
env.bio_15 -5.314e+05 1.012e+06 -0.525 0.599594
env.bio_16 8.907e+05 1.697e+06 0.525 0.599597
env.bio_17 7.569e+05 1.442e+06 0.525 0.599598
env.bio_18 1.103e+00 9.738e-01 1.132 0.257560
env.bio_19 -1.566e+00 1.600e+00 -0.979 0.327550
env.elev -9.271e-01 9.762e-01 -0.950 0.342270
env.urban 3.173e-01 2.034e-01 1.560 0.118775
env.barren -2.435e+01 5.659e+01 -0.430 0.667026
env.water -9.433e-02 2.250e-01 -0.419 0.675005
env.savanna 4.191e-01 2.203e-01 1.903 0.057100 .
env.woodysavanna 1.069e-01 2.224e-01 0.480 0.630872
env.wetland -3.750e-02 1.104e-01 -0.340 0.734146
env.grass -1.940e-02 3.162e-01 -0.061 0.951064
env.npp 6.397e-01 3.139e-01 2.038 0.041558 *
env.tree 5.212e-01 5.012e-01 1.040 0.298342
env.nontree -4.879e-01 4.521e-01 -1.079 0.280524
env.nonveg 2.303e-01 5.771e-01 0.399 0.689776
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 742.01 on 558 degrees of freedom
Residual deviance: 514.62 on 527 degrees of freedom
AIC: 578.62
Number of Fisher Scoring iterations: 11Code
step(glm.full) # step might not work with gam so glmStart: AIC=578.62
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_11 + env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 +
env.bio_16 + env.bio_17 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.water + env.savanna + env.woodysavanna +
env.wetland + env.grass + env.npp + env.tree + env.nontree +
env.nonveg
Df Deviance AIC
- env.grass 1 514.62 576.62
- env.wetland 1 514.74 576.74
- env.nonveg 1 514.78 576.78
- env.water 1 514.79 576.79
- env.bio_11 1 514.81 576.81
- env.woodysavanna 1 514.85 576.85
- env.bio_17 1 514.89 576.89
- env.bio_16 1 514.89 576.89
- env.bio_15 1 514.89 576.89
- env.bio_10 1 514.94 576.94
- env.bio_13 1 515.24 577.24
- env.barren 1 515.25 577.25
- env.elev 1 515.49 577.49
- env.bio_19 1 515.66 577.66
- env.tree 1 515.71 577.71
- env.nontree 1 515.78 577.78
- env.bio_12 1 515.82 577.82
- env.bio_18 1 515.93 577.93
- env.bio_8 1 516.32 578.32
- env.bio_4 1 516.44 578.44
<none> 514.62 578.62
- env.bio_2 1 516.85 578.85
- env.bio_9 1 517.10 579.10
- env.bio_6 1 517.21 579.21
- env.urban 1 517.34 579.34
- env.savanna 1 518.34 580.34
- env.bio_7 1 518.75 580.75
- env.npp 1 519.03 581.03
- env.bio_14 1 522.02 584.02
- env.bio_3 1 524.18 586.18
- env.bio_1 1 527.08 589.08
- env.bio_5 1 527.61 589.61
Step: AIC=576.62
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_11 + env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 +
env.bio_16 + env.bio_17 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.water + env.savanna + env.woodysavanna +
env.wetland + env.npp + env.tree + env.nontree + env.nonveg
Df Deviance AIC
- env.wetland 1 514.74 574.74
- env.nonveg 1 514.78 574.78
- env.water 1 514.79 574.79
- env.bio_11 1 514.81 574.81
- env.woodysavanna 1 514.88 574.88
- env.bio_17 1 514.90 574.90
- env.bio_16 1 514.90 574.90
- env.bio_15 1 514.90 574.90
- env.bio_10 1 514.95 574.95
- env.bio_13 1 515.25 575.25
- env.barren 1 515.25 575.25
- env.elev 1 515.56 575.56
- env.bio_19 1 515.67 575.67
- env.nontree 1 515.82 575.82
- env.bio_12 1 515.82 575.82
- env.bio_18 1 515.93 575.93
- env.tree 1 516.08 576.08
- env.bio_4 1 516.45 576.45
- env.bio_8 1 516.58 576.58
<none> 514.62 576.62
- env.bio_2 1 516.85 576.85
- env.bio_9 1 517.16 577.16
- env.bio_6 1 517.26 577.26
- env.urban 1 517.41 577.41
- env.bio_7 1 518.78 578.78
- env.npp 1 519.04 579.04
- env.savanna 1 520.44 580.44
- env.bio_14 1 522.03 582.03
- env.bio_3 1 524.18 584.18
- env.bio_1 1 527.10 587.10
- env.bio_5 1 529.74 589.74
Step: AIC=574.74
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_11 + env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 +
env.bio_16 + env.bio_17 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.water + env.savanna + env.woodysavanna +
env.npp + env.tree + env.nontree + env.nonveg
Df Deviance AIC
- env.water 1 514.83 572.83
- env.nonveg 1 514.90 572.90
- env.bio_11 1 514.92 572.92
- env.bio_17 1 515.02 573.02
- env.bio_16 1 515.02 573.02
- env.bio_15 1 515.02 573.02
- env.bio_10 1 515.04 573.04
- env.woodysavanna 1 515.06 573.06
- env.bio_13 1 515.28 573.28
- env.barren 1 515.33 573.33
- env.elev 1 515.68 573.68
- env.bio_19 1 515.71 573.71
- env.bio_12 1 515.87 573.87
- env.bio_18 1 516.07 574.07
- env.tree 1 516.09 574.09
- env.nontree 1 516.21 574.21
- env.bio_4 1 516.53 574.53
<none> 514.74 574.74
- env.bio_8 1 516.75 574.75
- env.bio_2 1 517.08 575.08
- env.bio_9 1 517.18 575.18
- env.bio_6 1 517.28 575.28
- env.urban 1 517.66 575.66
- env.bio_7 1 518.78 576.78
- env.npp 1 519.72 577.72
- env.savanna 1 521.24 579.24
- env.bio_14 1 522.14 580.14
- env.bio_3 1 524.21 582.21
- env.bio_1 1 527.10 585.10
- env.bio_5 1 529.85 587.85
Step: AIC=572.83
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_11 + env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 +
env.bio_16 + env.bio_17 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.savanna + env.woodysavanna +
env.npp + env.tree + env.nontree + env.nonveg
Df Deviance AIC
- env.nonveg 1 514.98 570.98
- env.bio_11 1 514.99 570.99
- env.bio_17 1 515.10 571.10
- env.bio_16 1 515.10 571.10
- env.bio_15 1 515.10 571.10
- env.bio_10 1 515.11 571.11
- env.woodysavanna 1 515.14 571.14
- env.barren 1 515.38 571.38
- env.bio_13 1 515.38 571.38
- env.elev 1 515.75 571.75
- env.bio_19 1 515.81 571.81
- env.bio_12 1 516.00 572.00
- env.tree 1 516.09 572.09
- env.bio_18 1 516.16 572.16
- env.bio_4 1 516.63 572.63
- env.nontree 1 516.77 572.77
- env.bio_8 1 516.78 572.78
<none> 514.83 572.83
- env.bio_2 1 517.16 573.16
- env.bio_9 1 517.26 573.26
- env.bio_6 1 517.34 573.34
- env.urban 1 517.84 573.84
- env.bio_7 1 518.84 574.84
- env.npp 1 520.10 576.10
- env.savanna 1 521.30 577.30
- env.bio_14 1 522.20 578.20
- env.bio_3 1 524.28 580.28
- env.bio_1 1 527.22 583.22
- env.bio_5 1 529.89 585.89
Step: AIC=570.98
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_11 + env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 +
env.bio_16 + env.bio_17 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.savanna + env.woodysavanna +
env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_11 1 515.14 569.14
- env.bio_17 1 515.25 569.25
- env.bio_16 1 515.25 569.25
- env.bio_15 1 515.25 569.25
- env.woodysavanna 1 515.28 569.28
- env.bio_10 1 515.30 569.30
- env.bio_13 1 515.44 569.44
- env.barren 1 515.46 569.46
- env.bio_12 1 516.00 570.00
- env.bio_19 1 516.05 570.05
- env.elev 1 516.15 570.15
- env.bio_18 1 516.25 570.25
- env.bio_4 1 516.81 570.81
<none> 514.98 570.98
- env.bio_8 1 517.00 571.00
- env.bio_2 1 517.33 571.33
- env.bio_9 1 517.40 571.40
- env.bio_6 1 517.52 571.52
- env.urban 1 517.84 571.84
- env.bio_7 1 518.98 572.98
- env.npp 1 520.36 574.36
- env.savanna 1 521.62 575.62
- env.nontree 1 522.09 576.09
- env.bio_14 1 522.85 576.85
- env.tree 1 523.52 577.52
- env.bio_3 1 524.74 578.74
- env.bio_1 1 527.68 581.68
- env.bio_5 1 530.51 584.51
Step: AIC=569.14
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_10 +
env.bio_12 + env.bio_13 + env.bio_14 + env.bio_15 + env.bio_16 +
env.bio_17 + env.bio_18 + env.bio_19 + env.elev + env.urban +
env.barren + env.savanna + env.woodysavanna + env.npp + env.tree +
env.nontree
Df Deviance AIC
- env.bio_10 1 515.31 567.31
- env.bio_17 1 515.43 567.43
- env.bio_16 1 515.43 567.43
- env.bio_15 1 515.43 567.43
- env.woodysavanna 1 515.44 567.44
- env.bio_13 1 515.48 567.48
- env.barren 1 515.64 567.64
- env.bio_12 1 516.16 568.16
- env.bio_19 1 516.16 568.16
- env.elev 1 516.32 568.32
- env.bio_18 1 516.44 568.44
- env.bio_4 1 516.96 568.96
<none> 515.14 569.14
- env.bio_8 1 517.46 569.46
- env.bio_2 1 517.68 569.68
- env.bio_9 1 517.82 569.82
- env.bio_6 1 517.88 569.88
- env.urban 1 518.01 570.01
- env.bio_7 1 518.98 570.98
- env.npp 1 520.37 572.37
- env.savanna 1 521.91 573.91
- env.nontree 1 523.18 575.18
- env.bio_14 1 523.29 575.29
- env.tree 1 523.63 575.63
- env.bio_3 1 525.85 577.85
- env.bio_1 1 528.24 580.24
- env.bio_5 1 530.52 582.52
Step: AIC=567.31
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_13 + env.bio_14 + env.bio_15 + env.bio_16 + env.bio_17 +
env.bio_18 + env.bio_19 + env.elev + env.urban + env.barren +
env.savanna + env.woodysavanna + env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_17 1 515.59 565.59
- env.bio_16 1 515.59 565.59
- env.bio_15 1 515.59 565.59
- env.woodysavanna 1 515.68 565.68
- env.barren 1 515.78 565.78
- env.bio_13 1 515.83 565.83
- env.bio_19 1 516.49 566.49
- env.elev 1 516.71 566.71
- env.bio_12 1 516.82 566.82
- env.bio_18 1 516.85 566.85
<none> 515.31 567.31
- env.bio_8 1 517.62 567.62
- env.bio_4 1 517.71 567.71
- env.bio_2 1 517.81 567.81
- env.bio_9 1 517.97 567.97
- env.bio_6 1 518.02 568.02
- env.urban 1 518.50 568.50
- env.bio_7 1 520.25 570.25
- env.npp 1 520.97 570.97
- env.savanna 1 522.19 572.19
- env.nontree 1 523.21 573.21
- env.bio_14 1 523.36 573.36
- env.tree 1 523.67 573.67
- env.bio_3 1 526.10 576.10
- env.bio_1 1 529.67 579.67
- env.bio_5 1 531.63 581.63
Step: AIC=565.59
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_13 + env.bio_14 + env.bio_15 + env.bio_16 + env.bio_18 +
env.bio_19 + env.elev + env.urban + env.barren + env.savanna +
env.woodysavanna + env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_16 1 515.74 563.74
- env.woodysavanna 1 515.90 563.90
- env.bio_13 1 515.99 563.99
- env.barren 1 516.02 564.02
- env.bio_19 1 516.73 564.73
- env.bio_12 1 516.93 564.93
- env.elev 1 517.02 565.02
- env.bio_18 1 517.16 565.16
<none> 515.59 565.59
- env.bio_4 1 517.79 565.79
- env.bio_2 1 517.98 565.98
- env.bio_8 1 518.01 566.01
- env.bio_9 1 518.28 566.28
- env.bio_6 1 518.30 566.30
- env.urban 1 519.03 567.03
- env.bio_7 1 520.30 568.30
- env.bio_15 1 520.60 568.60
- env.npp 1 521.35 569.35
- env.savanna 1 522.35 570.35
- env.nontree 1 523.38 571.38
- env.bio_14 1 523.45 571.45
- env.tree 1 524.21 572.21
- env.bio_3 1 526.13 574.13
- env.bio_1 1 529.77 577.77
- env.bio_5 1 531.70 579.70
Step: AIC=563.74
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_13 + env.bio_14 + env.bio_15 + env.bio_18 + env.bio_19 +
env.elev + env.urban + env.barren + env.savanna + env.woodysavanna +
env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_13 1 516.00 562.00
- env.woodysavanna 1 516.12 562.12
- env.barren 1 516.16 562.16
- env.bio_19 1 516.73 562.73
- env.elev 1 517.18 563.18
- env.bio_18 1 517.46 563.46
<none> 515.74 563.74
- env.bio_4 1 517.81 563.81
- env.bio_2 1 518.12 564.12
- env.bio_8 1 518.21 564.21
- env.bio_6 1 518.43 564.43
- env.bio_9 1 518.46 564.46
- env.bio_12 1 518.58 564.58
- env.urban 1 519.72 565.72
- env.bio_7 1 520.50 566.50
- env.bio_15 1 520.93 566.93
- env.npp 1 522.08 568.08
- env.savanna 1 523.10 569.10
- env.bio_14 1 523.52 569.52
- env.nontree 1 523.84 569.84
- env.tree 1 525.09 571.09
- env.bio_3 1 526.23 572.23
- env.bio_1 1 529.78 575.78
- env.bio_5 1 531.82 577.82
Step: AIC=562
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_14 + env.bio_15 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.savanna + env.woodysavanna +
env.npp + env.tree + env.nontree
Df Deviance AIC
- env.woodysavanna 1 516.45 560.45
- env.barren 1 516.46 560.46
- env.bio_19 1 517.21 561.21
- env.bio_18 1 517.63 561.63
- env.elev 1 517.90 561.90
<none> 516.00 562.00
- env.bio_2 1 518.21 562.21
- env.bio_4 1 518.41 562.41
- env.bio_8 1 518.48 562.48
- env.bio_9 1 518.49 562.49
- env.bio_12 1 518.62 562.62
- env.bio_6 1 518.64 562.64
- env.urban 1 520.06 564.06
- env.bio_7 1 520.51 564.51
- env.bio_15 1 521.09 565.09
- env.npp 1 522.23 566.23
- env.savanna 1 523.52 567.52
- env.bio_14 1 523.61 567.61
- env.nontree 1 524.62 568.62
- env.tree 1 526.05 570.05
- env.bio_3 1 526.73 570.73
- env.bio_1 1 530.57 574.57
- env.bio_5 1 532.86 576.86
Step: AIC=560.45
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_14 + env.bio_15 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.barren + env.savanna + env.npp + env.tree +
env.nontree
Df Deviance AIC
- env.barren 1 516.94 558.94
- env.bio_19 1 517.61 559.61
- env.bio_18 1 518.20 560.20
<none> 516.45 560.45
- env.elev 1 518.54 560.54
- env.bio_4 1 518.60 560.60
- env.bio_2 1 518.96 560.96
- env.bio_9 1 519.03 561.03
- env.bio_6 1 519.06 561.06
- env.bio_12 1 519.24 561.24
- env.bio_8 1 519.68 561.68
- env.urban 1 520.32 562.32
- env.bio_7 1 521.10 563.10
- env.bio_15 1 521.58 563.58
- env.npp 1 522.36 564.36
- env.savanna 1 523.52 565.52
- env.bio_14 1 524.67 566.67
- env.nontree 1 525.49 567.49
- env.tree 1 526.14 568.14
- env.bio_3 1 527.78 569.78
- env.bio_1 1 530.86 572.86
- env.bio_5 1 535.84 577.84
Step: AIC=558.94
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_14 + env.bio_15 + env.bio_18 + env.bio_19 + env.elev +
env.urban + env.savanna + env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_19 1 518.17 558.17
- env.bio_18 1 518.65 558.65
- env.elev 1 518.91 558.91
<none> 516.94 558.94
- env.bio_4 1 519.10 559.10
- env.bio_2 1 519.40 559.40
- env.bio_9 1 519.57 559.57
- env.bio_6 1 519.58 559.58
- env.bio_12 1 519.69 559.69
- env.bio_8 1 520.12 560.12
- env.urban 1 520.97 560.97
- env.bio_7 1 521.71 561.71
- env.bio_15 1 522.06 562.06
- env.npp 1 523.81 563.81
- env.savanna 1 524.31 564.31
- env.bio_14 1 525.08 565.08
- env.nontree 1 525.89 565.89
- env.tree 1 526.67 566.67
- env.bio_3 1 528.22 568.22
- env.bio_1 1 531.82 571.82
- env.bio_5 1 536.16 576.16
Step: AIC=558.17
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_14 + env.bio_15 + env.bio_18 + env.elev + env.urban +
env.savanna + env.npp + env.tree + env.nontree
Df Deviance AIC
- env.elev 1 520.09 558.09
<none> 518.17 558.17
- env.bio_9 1 520.35 558.35
- env.bio_6 1 520.69 558.69
- env.bio_8 1 520.83 558.83
- env.bio_4 1 521.17 559.17
- env.bio_2 1 521.37 559.37
- env.urban 1 522.34 560.34
- env.bio_7 1 522.43 560.43
- env.bio_12 1 523.07 561.07
- env.bio_18 1 523.82 561.82
- env.savanna 1 525.47 563.47
- env.npp 1 525.73 563.73
- env.bio_14 1 526.32 564.32
- env.bio_15 1 527.09 565.09
- env.nontree 1 527.22 565.22
- env.tree 1 528.31 566.31
- env.bio_3 1 529.64 567.64
- env.bio_1 1 531.85 569.85
- env.bio_5 1 536.64 574.64
Step: AIC=558.09
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_8 + env.bio_9 + env.bio_12 +
env.bio_14 + env.bio_15 + env.bio_18 + env.urban + env.savanna +
env.npp + env.tree + env.nontree
Df Deviance AIC
- env.bio_8 1 521.77 557.77
- env.bio_9 1 521.81 557.81
<none> 520.09 558.09
- env.bio_6 1 522.54 558.54
- env.bio_7 1 523.49 559.49
- env.bio_2 1 524.01 560.01
- env.bio_4 1 524.41 560.41
- env.urban 1 524.68 560.68
- env.bio_12 1 524.79 560.79
- env.bio_18 1 525.09 561.09
- env.bio_14 1 527.51 563.51
- env.savanna 1 528.32 564.32
- env.nontree 1 529.67 565.67
- env.npp 1 529.79 565.79
- env.tree 1 530.46 566.46
- env.bio_15 1 530.64 566.64
- env.bio_3 1 530.67 566.67
- env.bio_1 1 532.25 568.25
- env.bio_5 1 536.84 572.84
Step: AIC=557.77
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_9 + env.bio_12 + env.bio_14 +
env.bio_15 + env.bio_18 + env.urban + env.savanna + env.npp +
env.tree + env.nontree
Df Deviance AIC
- env.bio_9 1 522.88 556.88
<none> 521.77 557.77
- env.bio_6 1 523.90 557.90
- env.bio_12 1 525.87 559.87
- env.bio_2 1 526.32 560.32
- env.bio_18 1 526.37 560.37
- env.urban 1 526.62 560.62
- env.bio_7 1 528.10 562.10
- env.bio_14 1 529.43 563.43
- env.nontree 1 530.57 564.57
- env.savanna 1 531.84 565.84
- env.tree 1 531.86 565.86
- env.npp 1 532.37 566.37
- env.bio_3 1 532.42 566.42
- env.bio_15 1 533.50 567.50
- env.bio_1 1 533.52 567.52
- env.bio_4 1 537.68 571.68
- env.bio_5 1 541.15 575.15
Step: AIC=556.88
presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 + env.bio_5 +
env.bio_6 + env.bio_7 + env.bio_12 + env.bio_14 + env.bio_15 +
env.bio_18 + env.urban + env.savanna + env.npp + env.tree +
env.nontree
Df Deviance AIC
<none> 522.88 556.88
- env.bio_6 1 525.14 557.14
- env.bio_18 1 526.81 558.81
- env.urban 1 527.38 559.38
- env.bio_7 1 528.18 560.18
- env.bio_2 1 528.67 560.67
- env.bio_12 1 530.18 562.18
- env.bio_14 1 530.22 562.22
- env.nontree 1 531.09 563.09
- env.savanna 1 532.39 564.39
- env.bio_3 1 532.77 564.77
- env.tree 1 533.10 565.10
- env.bio_1 1 533.52 565.52
- env.npp 1 534.93 566.93
- env.bio_15 1 540.43 572.43
- env.bio_4 1 544.45 576.45
- env.bio_5 1 546.02 578.02
Call: glm(formula = presabs ~ env.bio_1 + env.bio_2 + env.bio_3 + env.bio_4 +
env.bio_5 + env.bio_6 + env.bio_7 + env.bio_12 + env.bio_14 +
env.bio_15 + env.bio_18 + env.urban + env.savanna + env.npp +
env.tree + env.nontree, family = "binomial", data = presabs.glm)
Coefficients:
(Intercept) env.bio_1 env.bio_2 env.bio_3 env.bio_4 env.bio_5
-1.4632 20.7750 16.7683 -43.6013 5.2812 -4.2671
env.bio_6 env.bio_7 env.bio_12 env.bio_14 env.bio_15 env.bio_18
-0.7010 1.0693 1.5955 -21.0583 -7.0540 1.4240
env.urban env.savanna env.npp env.tree env.nontree
0.3717 0.4765 0.8007 0.6285 -0.4219
Degrees of Freedom: 558 Total (i.e. Null); 542 Residual
Null Deviance: 742
Residual Deviance: 522.9 AIC: 556.9Boosted regression trees
Boosted regression trees were fitted using packages: gbm and dismo — specifically the gbm.step() function (Hijmans et al., 2017)
Code
# cross-validation optimisation of a boosted regression tree model
brt <- gbm.step(data = PA,
gbm.x = 6:ncol(PA),
gbm.y = "presabs",
family = "bernoulli")
GBM STEP - version 2.9
Performing cross-validation optimisation of a boosted regression tree model
for NA and using a family of bernoulli
Using 565 observations and 31 predictors
creating 10 initial models of 50 trees
folds are stratified by prevalence
total mean deviance = 1.3269
tolerance is fixed at 0.0013
ntrees resid. dev.
50 1.203
now adding trees...
100 1.1241
150 1.0716
200 1.0349
250 1.0102
300 0.9909
350 0.975
400 0.9623
450 0.9522
500 0.9433
550 0.9355
600 0.9291
650 0.924
700 0.9184
750 0.9137
800 0.9093
850 0.9062
900 0.9028
950 0.8986
1000 0.8945
1050 0.8916
1100 0.8879
1150 0.8844
1200 0.8818
1250 0.8791
1300 0.8768
1350 0.8751
1400 0.872
1450 0.8697
1500 0.8676
1550 0.8659
1600 0.864
1650 0.8635
1700 0.8618
1750 0.8595
1800 0.8596
1850 0.8582
1900 0.8569
1950 0.8566
2000 0.8567
2050 0.8559
2100 0.8563
2150 0.8553
2200 0.8536
2250 0.8533
2300 0.8527
2350 0.8527
2400 0.8523
2450 0.8515
2500 0.8515
2550 0.852
2600 0.8515
2650 0.851
2700 0.8505
2750 0.8496
2800 0.8489
2850 0.8488
2900 0.8475
2950 0.8472
3000 0.8477
3050 0.8483
3100 0.8483
3150 0.848
3200 0.8474
3250 0.8485
3300 0.8486
3350 0.8489
3400 0.8491
3450 0.8488 
mean total deviance = 1.327
mean residual deviance = 0.606
estimated cv deviance = 0.847 ; se = 0.038
training data correlation = 0.795
cv correlation = 0.647 ; se = 0.024
training data AUC score = 0.952
cv AUC score = 0.87 ; se = 0.011
elapsed time - 0.17 minutes Code
summary(brt)
var rel.inf
env.npp env.npp 16.36571599
env.bio_10 env.bio_10 14.55387294
env.woodysavanna env.woodysavanna 13.42653096
env.nontree env.nontree 9.13332502
env.urban env.urban 6.08203761
env.bio_18 env.bio_18 3.63438713
env.bio_4 env.bio_4 3.41292200
env.bio_8 env.bio_8 3.24998923
env.bio_15 env.bio_15 2.96201858
env.bio_17 env.bio_17 2.90581272
env.bio_12 env.bio_12 2.69728096
env.tree env.tree 2.69636127
env.grass env.grass 2.58599753
env.bio_14 env.bio_14 2.15810462
env.bio_9 env.bio_9 2.06431763
env.bio_16 env.bio_16 1.98543142
env.bio_7 env.bio_7 1.36512507
env.wetland env.wetland 1.29455125
env.bio_19 env.bio_19 1.21964255
env.nonveg env.nonveg 1.19108472
env.bio_13 env.bio_13 0.77959866
env.bio_2 env.bio_2 0.76835885
env.elev env.elev 0.76328742
env.bio_6 env.bio_6 0.75682912
env.savanna env.savanna 0.61223648
env.bio_5 env.bio_5 0.54200687
env.bio_3 env.bio_3 0.41660876
env.bio_11 env.bio_11 0.28012471
env.bio_1 env.bio_1 0.09643991
env.barren env.barren 0.00000000
env.water env.water 0.00000000Code
variables_brt <- brt$contributions[1:6,] %>% pull(var)
#exploration of shape of relationships
#gbm.plot(brt, n.plots = 12, plot.layout=c(6, 2))Random forest
Code
presabs.rf <- PA %>%
dplyr::select(-c(1,2,3,5)) %>%
mutate(presabs = as.factor(presabs))
rf <- randomForest(presabs ~ .,
data=presabs.rf,
importance=T,
nperm=2, # two permutations per tree to estimate importance
na.action=na.omit,
mtry= 1/3*ncol(presabs.rf)-1)
varImpPlot(rf, type=2)
Code
variables_rf <- rf$importance %>% as_tibble(rownames = 'var') %>% arrange(desc(MeanDecreaseGini)) %>% head(n=6) %>% pull(var)Ranger
Code
presabs.ranger <- PA %>%
dplyr::select(-c(1,2,3,5)) %>%
filter(!if_any(everything(), is.na)) %>%
mutate(presabs = as.factor(presabs))
## Learn the model:
ranger <- ranger(presabs ~ .,
data = presabs.ranger,
num.trees = 150,
mtry = 1/3*ncol(presabs.ranger)-1,
min.node.size = 5,
max.depth = NULL,
write.forest = TRUE,
importance = "impurity")
# Get the variable importance
importance(ranger) env.bio_1 env.bio_2 env.bio_3 env.bio_4
3.9705229 7.1674218 4.8764798 7.9957307
env.bio_5 env.bio_6 env.bio_7 env.bio_8
5.6504988 6.5724679 6.1964904 7.3501020
env.bio_9 env.bio_10 env.bio_11 env.bio_12
8.1038185 25.5522344 6.2327071 6.7432816
env.bio_13 env.bio_14 env.bio_15 env.bio_16
6.5339720 8.1198586 9.2777852 5.7880864
env.bio_17 env.bio_18 env.bio_19 env.elev
6.8298731 5.5505851 4.4749455 7.2100926
env.urban env.barren env.water env.savanna
12.4759667 0.2073320 0.6722292 4.4616751
env.woodysavanna env.wetland env.grass env.npp
18.5623381 1.7660608 6.0611602 21.8425596
env.tree env.nontree env.nonveg
9.0818534 17.6400638 7.1887874 Code
importance(ranger) %>%
as.data.frame(row.names = names(.)) %>%
setNames(c("Importance")) %>%
rownames_to_column("covariate") %>%
# mutate(covariate = case_when(
# covariate == "env.bio_1" ~ "Annual Mean Temperature",
# )) %>%
ggplot(aes(x = Importance, y = reorder(covariate, Importance))) +
ylab('')+
geom_point() + theme_bw()
Code
variables_ranger <- importance(ranger) %>% as_tibble(rownames = 'var') %>% arrange(desc(value)) %>% head(n=6) %>% pull(var)Correlation
Correlation between the six more important variables detected.
- Boosted regression tree: env.npp, env.bio_10, env.woodysavanna, env.nontree, env.urban, env.bio_18
- Random forest: env.npp, env.bio_10, env.woodysavanna, env.nontree, env.urban, env.bio_15
- Ranger: env.bio_10, env.npp, env.woodysavanna, env.nontree, env.urban, env.bio_15
Code
selectedVariables <- unique(c(variables_brt, variables_rf, variables_ranger))
PA %>%
filter(!if_any(everything(), is.na)) %>%
dplyr::select(selectedVariables) %>%
pairs()
Code
PA %>%
filter(!if_any(everything(), is.na)) %>%
dplyr::select(selectedVariables) %>%
cor() %>% kable()| env.npp | env.bio_10 | env.woodysavanna | env.nontree | env.urban | env.bio_18 | env.bio_15 | |
|---|---|---|---|---|---|---|---|
| env.npp | 1.0000000 | 0.2583383 | 0.0006886 | -0.5293037 | 0.1344147 | -0.5298813 | -0.6102193 |
| env.bio_10 | 0.2583383 | 1.0000000 | 0.0182359 | -0.1901205 | 0.2582501 | -0.0145843 | -0.4429341 |
| env.woodysavanna | 0.0006886 | 0.0182359 | 1.0000000 | -0.1006558 | 0.0318901 | 0.0667723 | 0.0462821 |
| env.nontree | -0.5293037 | -0.1901205 | -0.1006558 | 1.0000000 | -0.0745566 | 0.1344020 | 0.1630034 |
| env.urban | 0.1344147 | 0.2582501 | 0.0318901 | -0.0745566 | 1.0000000 | -0.1337635 | -0.2740025 |
| env.bio_18 | -0.5298813 | -0.0145843 | 0.0667723 | 0.1344020 | -0.1337635 | 1.0000000 | 0.7628378 |
| env.bio_15 | -0.6102193 | -0.4429341 | 0.0462821 | 0.1630034 | -0.2740025 | 0.7628378 | 1.0000000 |
Code
tmpFiles(remove=T)