gf <-
gradientForest(
cbind(mean_env, mean_fauna),
predictor.vars = colnames(mean_env),
response.vars = colnames(mean_fauna),
ntree = 500,
transform = NULL,
compact = T,
nbin = 201,
maxLevel = lev,
corr.threshold = 0.5
)
gf
## A forest of 500 regression trees for each of 59 species
##
## Call:
##
## gradientForest(data = cbind(mean_env, mean_fauna), predictor.vars = colnames(mean_env),
## response.vars = colnames(mean_fauna), ntree = 500, transform = NULL,
## maxLevel = lev, corr.threshold = 0.5, compact = T, nbin = 201)
##
##
##
## Important variables:
## [1] mud AFDW Pb TP Cu
gf$result %>%
data.frame()
## .
## glyceridae 0.102260724
## goniadidae 0.030215161
## nephtyidae 0.228548127
## nereididae 0.129890309
## syllidae 0.154807850
## owenia 0.093362531
## sabellidae 0.213561262
## magelona 0.165026628
## aonides 0.332308205
## polydorid complex 0.017294828
## prionospio 0.255902773
## scolecolepides 0.231008147
## microspio 0.164708610
## spionidae 0.233519612
## heteromastusfiliformisandbarantollalepte 0.165409162
## capitella 0.065745229
## maldanidae 0.212437136
## orbiniidae 0.120670263
## paraonidae 0.188053327
## scalibregmatidae 0.105007266
## amphipoda 0.269030031
## cumacea 0.254137568
## halicarcinus 0.148599733
## hemiplax hirtipes 0.088186626
## palaemonidae 0.177142200
## hemigrapsus 0.048913078
## isopoda 0.030978366
## tanaidacea 0.090693819
## ostracoda 0.257318847
## copepoda 0.186336373
## actiniidae 0.233260190
## edwardsia 0.307865544
## anthozoa 0.012994504
## hiatula 0.160359116
## mytilidae 0.011719877
## nuculidae 0.276481474
## arthritica bifurca 0.172376721
## lasaea 0.282957566
## diplodonta 0.226335900
## zeacumantus 0.190148899
## amphibola crenata 0.081504490
## potamopyrgus 0.096387491
## notoacmea 0.133970411
## diloma 0.008723071
## nemertea 0.170600815
## sipuncula 0.002587112
## oligochaeta 0.246524698
## scolelepis 0.029710189
## scoloplos cylindrifer 0.182271422
## orbinia papillosa 0.216949832
## austrohelice crassa 0.349386705
## macomona liliana 0.326257595
## austrovenus stutchburyi 0.323894877
## paphies australis 0.260029620
## melanopsis 0.044753969
## halopyrgus pupoides 0.096256532
## cominella glandiformis 0.015123378
## nematoda 0.202207047
## nicona estuariensis 0.261942642
gf$overall.imp %>%
data.frame()
## .
## AFDW 0.04093269
## Cu 0.02355860
## mud 0.04502443
## Pb 0.03312813
## TN 0.02325936
## TP 0.03027223
## Zn 0.02056801
gf$overall.imp2 %>%
data.frame()
## .
## AFDW 17.29308
## Cu 11.76664
## mud 18.35672
## Pb 15.93648
## TN 11.24796
## TP 15.93607
## Zn 14.76125
gf$species.pos.rsq
## [1] 59
##Plot predictors overall importance
plot(gf, plot.type = "O")
most_important <-
names(importance(gf))[1:6]##get most important predictors names
###Split density plot
par(mgp = c(2, 0.75, 0))
plot(
gf,
plot.type = "S",
imp.vars = most_important,
leg.posn = "topright",
cex.legend = 0.8,
cex.axis = 0.9,
cex.lab = 0.9,
line.ylab = 0.9,
par.args = list(mgp = c(1.5, 0.5, 0), mar = c(3.1, 1.5, 0.1, 1))
)
##Species cummulative plot
plot(
gf,
plot.type = "C",
imp.vars = most_important,
show.overall = F,
legend = T,
leg.posn = "topleft",
leg.nspecies = 5,
cex.lab = 1,
cex.legend = 1,
cex.axis = 1,
line.ylab = 1,
par.args = list(
mgp = c(1.5, 0.5, 0),
mar = c(2.5, 1, 0.1, 0.5),
omi = c(0, 0.3, 0, 0)
)
)
##Predictor cummulative plot across all species
plot(
gf,
plot.type = "C",
imp.vars = most_important,
show.species = F,
common.scale = T,
cex.axis = 1,
cex.lab = 1,
line.ylab = 0.9,
par.args = list(
mgp = c(1.5, 0.5, 0),
mar = c(2.5, 1, 0.1, 0.5),
omi = c(0, 0.3, 0, 0)
)
)
Boosted regression tress on selected taxa shown as important in the gradient forest analysis
Macomona liliana
###gbm macomona #####
par(
mfrow = c(2, 4),
mgp = c(1.5, .5, 0),
mar = c(2.5, 1.5, 0.5, 0.5),
omi = c(.2, 0.1, 0, 0)
)
summary(gbm_maco,n.trees=best.iter)
## var rel.inf
## mud mud 47.712980
## AFDW AFDW 25.800480
## TP TP 11.169760
## Pb Pb 5.487283
## Zn Zn 4.431558
## TN TN 3.713017
## Cu Cu 1.684922
plot(gbm_maco, 1, best.iter)
plot(gbm_maco, 2, best.iter)
plot(gbm_maco, 3, best.iter)
plot(gbm_maco, 4, best.iter)
plot(gbm_maco, 5, best.iter)
plot(gbm_maco, 6, best.iter)
plot(gbm_maco, 7, best.iter)
Paphies australis
par(
mfrow = c(2, 4),
mgp = c(1.5, .5, 0),
mar = c(2.5, 1.5, 0.5, 0.5),
omi = c(.2, 0.1, 0, 0)
)
summary(gbm_pipi,n.trees=best.iter)
## var rel.inf
## mud mud 43.6866225
## AFDW AFDW 26.2150916
## TP TP 14.7807624
## Zn Zn 9.1099272
## Pb Pb 4.5143656
## TN TN 1.3215681
## Cu Cu 0.3716628
plot(gbm_pipi,1,best.iter)
plot(gbm_pipi,2,best.iter)
plot(gbm_pipi,3,best.iter)
plot(gbm_pipi,4,best.iter)
plot(gbm_pipi,5,best.iter)
plot(gbm_pipi,6,best.iter)
plot(gbm_pipi,7,best.iter)
Oligochaeta
par(
mfrow = c(2, 4),
mgp = c(1.5, .5, 0),
mar = c(2.5, 1.5, 0.5, 0.5),
omi = c(.2, 0.1, 0, 0)
)
summary(gbm_oligo,n.trees=best.iter)
## var rel.inf
## AFDW AFDW 50.271477
## TP TP 34.096233
## mud mud 8.054585
## Pb Pb 3.053457
## Cu Cu 2.578858
## Zn Zn 1.945389
## TN TN 0.000000
plot(gbm_oligo,1,best.iter)
plot(gbm_oligo,2,best.iter)
plot(gbm_oligo,3,best.iter)
plot(gbm_oligo,4,best.iter)
plot(gbm_oligo,5,best.iter)
plot(gbm_oligo,6,best.iter)
plot(gbm_oligo,7,best.iter)
Austrovenus stutchburyi
par(
mfrow = c(2, 4),
mgp = c(1.5, .5, 0),
mar = c(2.5, 1.5, 0.5, 0.5),
omi = c(.2, 0.1, 0, 0)
)
summary(gbm_cockle,best.iter)
## var rel.inf
## mud mud 36.211703
## AFDW AFDW 34.640564
## Pb Pb 8.591573
## Zn Zn 8.348554
## TP TP 8.241761
## TN TN 2.429671
## Cu Cu 1.536173
plot(gbm_cockle,1,best.iter)
plot(gbm_cockle,2,best.iter)
plot(gbm_cockle,3,best.iter)
plot(gbm_cockle,4,best.iter)
plot(gbm_cockle,5,best.iter)
plot(gbm_cockle,6,best.iter)
plot(gbm_cockle,7,best.iter)
