1 POWER ANALYSES FOR DETECTION
- 1.1 calculate power analyses - traditional and weighted
- 1.2 summarise data by averaging
  - 1.2.1 summarise detection probability across bats
2 WHICH SPECIES ARE WE MOST LIKELY TO DETECT PD ON?
- 2.1 early winter analysis
- 2.2 late winter analysis
3 CATEGORICAL ANALYSIS - HOW DOES DIFFERENTIAL ARRIVAL IMPACT LATE WINTER METRICS?
4 CONTINUOUS ANALYSIS - SAME AS ABOVE BUT CONTINUOUS
5 OVERWINTER MOBILITY
6 NUMEROUS COVARIATES AND THE EFFECT ON TIMING OF DETECTION

1 POWER ANALYSES FOR DETECTION

1.1 calculate power analyses - traditional and weighted

library(tidyverse)
library(dplyr)
data.site.summ = data %>%
  group_by(site,date)%>%
  mutate(tot.bats.sampled = sum(N, na.rm=T))%>%
  mutate(tot.bats.counted = sum(count, na.rm=T))%>%
  mutate(prop.sampled = tot.bats.sampled/tot.bats.counted)%>%
  mutate(weighted.average.prev.early = sum(weighted.average.early.num)/tot.bats.sampled)%>%
  mutate(site.prob.missing.traditional = dbinom(x=0,size=tot.bats.sampled,prob=weighted.average.prev.early))%>%
  mutate(site.prob.missing.all = (1 - weighted.average.prev.early)^(tot.bats.sampled)*(1-(1-weighted.average.prev.early)^(tot.bats.counted - tot.bats.sampled)))%>%
  filter(season=="hiber_earl" & t=="winter")

1.2 summarise data by averaging

missing.value.cals = data.site.summ %>%
  group_by(site)%>%
  summarise(weighted.average.prev.early=mean(weighted.average.prev.early), 
            site.prob.missing.traditional=mean(site.prob.missing.traditional), 
            site.prob.missing.all=mean(site.prob.missing.all))

1.2.1 summarise detection probability across bats

summary(missing.value.cals$site.prob.missing.all)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
0.005171 0.017689 0.038582 0.063044 0.057758 0.326666

2 WHICH SPECIES ARE WE MOST LIKELY TO DETECT PD ON?

2.1 early winter analysis

modS1=glmer(gd~species + (1|site) ,data=mw0.trim,family="binomial", subset=t=="fall");summary(modS1)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: gd ~ species + (1 | site)
   Data: mw0.trim
 Subset: t == "fall"

     AIC      BIC   logLik deviance df.resid 
   581.9    603.0   -286.0    571.9      496 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.3486 -0.9002 -0.1708  0.9136  5.8562 

Random effects:
 Groups Name        Variance Std.Dev.
 site   (Intercept) 2.327    1.525   
Number of obs: 501, groups:  site, 7

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  -1.8225     0.6642  -2.744 0.006075 ** 
speciesMYLU   1.7925     0.3844   4.662 3.12e-06 ***
speciesEPFU   1.4013     0.5117   2.738 0.006175 ** 
speciesMYSE   1.4387     0.3909   3.681 0.000232 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) spMYLU spEPFU
speciesMYLU -0.429              
speciesEPFU -0.351  0.627       
speciesMYSE -0.420  0.807  0.617

2.2 late winter analysis

modS2=glmer(gd~species + (1|site) ,data=mw0.trim,family="binomial", subset=t=="winter");summary(modS2)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: gd ~ species + (1 | site)
   Data: mw0.trim
 Subset: t == "winter"

     AIC      BIC   logLik deviance df.resid 
  1032.6   1057.0   -511.3   1022.6      976 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-1.3321 -0.6038 -0.3941  0.7507  4.6090 

Random effects:
 Groups Name        Variance Std.Dev.
 site   (Intercept) 0.8578   0.9262  
Number of obs: 981, groups:  site, 15

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  -1.6848     0.3090  -5.453 4.94e-08 ***
speciesMYLU   0.9344     0.2221   4.207 2.59e-05 ***
speciesEPFU   0.6558     0.2954   2.220  0.02641 *  
speciesMYSE   0.8375     0.2614   3.204  0.00135 ** 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) spMYLU spEPFU
speciesMYLU -0.458              
speciesEPFU -0.458  0.470       
speciesMYSE -0.405  0.563  0.457

3 CATEGORICAL ANALYSIS - HOW DOES DIFFERENTIAL ARRIVAL IMPACT LATE WINTER METRICS?

3.1 create data subsets

data.trim=subset(data,species!="SUBSTRATE") #remove any substrate data (should be absent, but make sure to remove levels)
data.trim=subset(data.trim,species!="MYSO") #remove this species because only present in two sites

##manipulate and create dataframes that are subsets of larger data, but only from late or early hibernation#
#bring some early metric into late dataframe to determine explore what is the effect of early metrics on late winter metrics
mw0.trim$site.species = paste(mw0.trim$site,mw0.trim$species, sep = ".") #this is my match column
mw0.trim.late = subset(mw0.trim, season=="hiber_late") #subset to late hibernation
mw0.trim.early = subset(mw0.trim, season=="hiber_earl") #subset to early hibernation
mw0.trim.late$early.prev = mw0.trim.early$gd[match(mw0.trim.late$site.species,mw0.trim.early$site.species)]
#bring some early metric into late dataframe to determine explore what is the effect of early metrics on late winter metrics

data.trim$site.species = paste(data.trim$site,data.trim$species, sep=".") #match column
data.trim.late = subset(data.trim, season=="hiber_late") #late hibernation
data.trim.early = subset(data.trim, season=="hiber_earl") #early hibernation

#prevalence
data.trim.late$early.prev = data.trim.early$gd[match(data.trim.late$site.species,data.trim.early$site.species)] #match in early prevalence

#loads
data.trim.late$early.loads = data.trim.early$lgdL[match(data.trim.late$site.species,data.trim.early$site.species)] #match in early loads

#change negative bats to limit of detection in loads column only
data.trim.late$early.loads[is.na(data.trim.late$early.loads)==T] = -6

#add these average loads during early to my larger dataframe with each bat as a point
mw0.trim.late$early.loads = data.trim.late$early.loads[match(mw0.trim.late$site.species,data.trim.late$site.species)]

3.2 prev in late winter

mod2=glmer(gd~species*t + (1|site),weights=N, data=data.trim.late,family="binomial");summary(mod2)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: gd ~ species * t + (1 | site)
   Data: data.trim.late
Weights: N

     AIC      BIC   logLik deviance df.resid 
   258.9    278.9   -120.5    240.9       59 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-3.3342 -0.4236  0.0902  0.5216  2.3033 

Random effects:
 Groups Name        Variance Std.Dev.
 site   (Intercept) 4.47     2.114   
Number of obs: 68, groups:  site, 22

Fixed effects:
                    Estimate Std. Error z value Pr(>|z|)    
(Intercept)           3.4459     1.0207   3.376 0.000735 ***
speciesEPFU          -3.2814     0.7489  -4.382 1.18e-05 ***
speciesMYLU          -0.4431     0.7344  -0.603 0.546278    
speciesPESU          -4.1628     0.7947  -5.238 1.62e-07 ***
twinter              -2.2458     1.2101  -1.856 0.063473 .  
speciesEPFU:twinter   1.5497     0.8804   1.760 0.078363 .  
speciesMYLU:twinter   0.2899     0.8339   0.348 0.728128    
speciesPESU:twinter   2.0029     0.9009   2.223 0.026205 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) spEPFU spMYLU spPESU twintr sEPFU: sMYLU:
speciesEPFU -0.447                                          
speciesMYLU -0.453  0.588                                   
speciesPESU -0.482  0.590  0.563                            
twinter     -0.842  0.376  0.382  0.404                     
spcsEPFU:tw  0.377 -0.848 -0.499 -0.497 -0.444              
spcsMYLU:tw  0.400 -0.518 -0.881 -0.498 -0.449  0.587       
spcsPESU:tw  0.423 -0.519 -0.496 -0.878 -0.467  0.588  0.596

emmeans(mod2, pairwise ~ t|species)

$emmeans
species = MYSE:
 t      emmean    SE  df asymp.LCL asymp.UCL
 fall    3.446 1.021 Inf    1.4454     5.446
 winter  1.200 0.652 Inf   -0.0784     2.479

species = EPFU:
 t      emmean    SE  df asymp.LCL asymp.UCL
 fall    0.164 0.959 Inf   -1.7147     2.044
 winter -0.532 0.611 Inf   -1.7288     0.666

species = MYLU:
 t      emmean    SE  df asymp.LCL asymp.UCL
 fall    3.003 0.949 Inf    1.1419     4.864
 winter  1.047 0.597 Inf   -0.1239     2.218

species = PESU:
 t      emmean    SE  df asymp.LCL asymp.UCL
 fall   -0.717 0.944 Inf   -2.5678     1.134
 winter -0.960 0.606 Inf   -2.1477     0.228

Results are given on the logit (not the response) scale. 
Confidence level used: 0.95 

$contrasts
species = MYSE:
 contrast      estimate   SE  df z.ratio p.value
 fall - winter    2.246 1.21 Inf 1.856   0.0635 

species = EPFU:
 contrast      estimate   SE  df z.ratio p.value
 fall - winter    0.696 1.14 Inf 0.612   0.5404 

species = MYLU:
 contrast      estimate   SE  df z.ratio p.value
 fall - winter    1.956 1.12 Inf 1.746   0.0808 

species = PESU:
 contrast      estimate   SE  df z.ratio p.value
 fall - winter    0.243 1.12 Inf 0.217   0.8286 

Results are given on the log odds ratio (not the response) scale.

3.3 loads in late winter

mod3=lmer(lgdL~species*t + (1|site), data=mw0.trim.late);summary(mod3)

Linear mixed model fit by REML ['lmerMod']
Formula: lgdL ~ species * t + (1 | site)
   Data: mw0.trim.late

REML criterion at convergence: 1348.1

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-2.90453 -0.62385  0.02865  0.58963  3.07815 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 1.021    1.011   
 Residual             1.194    1.093   
Number of obs: 429, groups:  site, 21

Fixed effects:
                    Estimate Std. Error t value
(Intercept)          -1.6562     0.4426  -3.742
speciesPESU          -1.5507     0.3339  -4.644
speciesMYLU          -0.9636     0.1980  -4.866
speciesEPFU          -2.1173     0.3475  -6.092
twinter              -1.4356     0.5450  -2.634
speciesPESU:twinter   1.0992     0.4214   2.609
speciesMYLU:twinter   0.7705     0.2815   2.737
speciesEPFU:twinter   0.5327     0.4265   1.249

Correlation of Fixed Effects:
            (Intr) spPESU spMYLU spEPFU twintr sPESU: sMYLU:
speciesPESU -0.224                                          
speciesMYLU -0.274  0.387                                   
speciesEPFU -0.133  0.153  0.305                            
twinter     -0.812  0.182  0.222  0.108                     
spcsPESU:tw  0.177 -0.792 -0.307 -0.121 -0.271              
spcsMYLU:tw  0.193 -0.272 -0.703 -0.215 -0.328  0.422       
spcsEPFU:tw  0.108 -0.125 -0.249 -0.815 -0.218  0.233  0.374

emmeans(mod3, pairwise ~ t|species)

$emmeans
species = MYSE:
 t      emmean    SE   df lower.CL upper.CL
 fall    -1.66 0.443 20.2    -2.58   -0.734
 winter  -3.09 0.318 32.3    -3.74   -2.443

species = PESU:
 t      emmean    SE   df lower.CL upper.CL
 fall    -3.21 0.492 29.8    -4.21   -2.202
 winter  -3.54 0.330 36.5    -4.21   -2.873

species = MYLU:
 t      emmean    SE   df lower.CL upper.CL
 fall    -2.62 0.433 18.5    -3.53   -1.713
 winter  -3.28 0.298 24.9    -3.90   -2.671

species = EPFU:
 t      emmean    SE   df lower.CL upper.CL
 fall    -3.77 0.525 39.5    -4.84   -2.711
 winter  -4.68 0.319 31.6    -5.33   -4.025

Degrees-of-freedom method: kenward-roger 
Confidence level used: 0.95 

$contrasts
species = MYSE:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter    1.436 0.545 23.5 2.632   0.0147 

species = PESU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter    0.336 0.593 31.7 0.568   0.5743 

species = MYLU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter    0.665 0.525 20.2 1.266   0.2197 

species = EPFU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter    0.903 0.615 37.1 1.468   0.1504 

Degrees-of-freedom method: kenward-roger

3.4 lambda in late winter

mod4=lmer(log.lambda~species*t + (1|site), data=data.trim.late.lambda);summary(mod4)

Linear mixed model fit by REML ['lmerMod']
Formula: log.lambda ~ species * t + (1 | site)
   Data: data.trim.late.lambda

REML criterion at convergence: 31.9

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-1.80892 -0.55960 -0.04786  0.60157  1.81107 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 0.05463  0.2337  
 Residual             0.04941  0.2223  
Number of obs: 60, groups:  site, 20

Fixed effects:
                    Estimate Std. Error t value
(Intercept)          -0.2439     0.1317  -1.852
speciesEPFU           0.5945     0.1645   3.614
speciesMYLU           0.3860     0.1283   3.008
speciesPESU           0.1889     0.1477   1.279
twinter               0.4350     0.1655   2.628
speciesEPFU:twinter  -0.5674     0.1970  -2.880
speciesMYLU:twinter  -0.4112     0.1636  -2.513
speciesPESU:twinter  -0.2431     0.1818  -1.337

Correlation of Fixed Effects:
            (Intr) spEPFU spMYLU spPESU twintr sEPFU: sMYLU:
speciesEPFU -0.380                                          
speciesMYLU -0.487  0.390                                   
speciesPESU -0.423  0.340  0.434                            
twinter     -0.796  0.303  0.388  0.337                     
spcsEPFU:tw  0.317 -0.835 -0.326 -0.284 -0.445              
spcsMYLU:tw  0.382 -0.306 -0.784 -0.341 -0.528  0.443       
spcsPESU:tw  0.344 -0.276 -0.353 -0.812 -0.472  0.394  0.478

emmeans(mod4, pairwise~t|species)

$emmeans
species = MYSE:
 t      emmean     SE   df lower.CL upper.CL
 fall   -0.244 0.1317 33.5  -0.5116   0.0239
 winter  0.191 0.1007 44.2  -0.0118   0.3940

species = EPFU:
 t      emmean     SE   df lower.CL upper.CL
 fall    0.351 0.1679 49.0   0.0132   0.6880
 winter  0.218 0.0966 41.9   0.0232   0.4132

species = MYLU:
 t      emmean     SE   df lower.CL upper.CL
 fall    0.142 0.1317 33.5  -0.1257   0.4099
 winter  0.166 0.0910 37.3  -0.0184   0.3503

species = PESU:
 t      emmean     SE   df lower.CL upper.CL
 fall   -0.055 0.1511 43.6  -0.3597   0.2496
 winter  0.137 0.0970 41.7  -0.0589   0.3327

Degrees-of-freedom method: kenward-roger 
Confidence level used: 0.95 

$contrasts
species = MYSE:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter  -0.4350 0.166 37.4 -2.624  0.0125 

species = EPFU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter   0.1324 0.194 47.5  0.683  0.4977 

species = MYLU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter  -0.0239 0.160 34.7 -0.149  0.8824 

species = PESU:
 contrast      estimate    SE   df t.ratio p.value
 fall - winter  -0.1919 0.180 43.0 -1.068  0.2913 

Degrees-of-freedom method: kenward-roger

4 CONTINUOUS ANALYSIS - SAME AS ABOVE BUT CONTINUOUS

4.1 how are late winter loads affected by early prevalence?

mod6=lmer(lgdL~species+early.prev +(1|site), data=mw0.trim.late);summary(mod6)

Linear mixed model fit by REML ['lmerMod']
Formula: lgdL ~ species + early.prev + (1 | site)
   Data: mw0.trim.late

REML criterion at convergence: 1261.6

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-2.93185 -0.67037  0.01515  0.65539  2.98943 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 1.049    1.024   
 Residual             1.262    1.123   
Number of obs: 394, groups:  site, 19

Fixed effects:
            Estimate Std. Error t value
(Intercept)  -2.7819     0.2877  -9.668
speciesPESU  -0.6462     0.2339  -2.762
speciesMYLU  -0.2910     0.1773  -1.641
speciesEPFU  -1.9322     0.2700  -7.155
early.prev    0.8494     0.2893   2.936

Correlation of Fixed Effects:
            (Intr) spPESU spMYLU spEPFU
speciesPESU -0.401                     
speciesMYLU -0.459  0.520              
speciesEPFU -0.341  0.379  0.438       
early.prev  -0.335  0.393  0.558  0.341

4.2 how is late winter prevalence affected by early prevalence?

mod5=glmer(gd~species+early.prev +(1|site),weights=N, data=data.trim.late,family="binomial");summary(mod5)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: gd ~ species + early.prev + (1 | site)
   Data: data.trim.late
Weights: N

     AIC      BIC   logLik deviance df.resid 
   204.3    216.3    -96.1    192.3       49 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-1.9955 -0.5332  0.1541  0.6788  1.6161 

Random effects:
 Groups Name        Variance Std.Dev.
 site   (Intercept) 4.231    2.057   
Number of obs: 55, groups:  site, 20

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   1.6992     0.6009   2.828  0.00469 ** 
speciesEPFU  -2.7521     0.6720  -4.095 4.21e-05 ***
speciesMYLU  -0.2661     0.3614  -0.736  0.46151    
speciesPESU  -2.5151     0.4114  -6.114 9.71e-10 ***
early.prev   12.4435     6.2987   1.976  0.04820 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) spEPFU spMYLU spPESU
speciesEPFU -0.388                     
speciesMYLU -0.464  0.425              
speciesPESU -0.492  0.445  0.686       
early.prev  -0.288  0.123  0.144  0.220

4.3 how are late winter impacts affected by early prevalence?

mod6=lmer(log.lambda~species+early.prev +(1|site), data=data.trim.late.lambda);summary(mod6)

Linear mixed model fit by REML ['lmerMod']
Formula: log.lambda ~ species + early.prev + (1 | site)
   Data: data.trim.late.lambda

REML criterion at convergence: 24.1

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-2.31360 -0.38344 -0.04979  0.62521  2.23076 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 0.05378  0.2319  
 Residual             0.05162  0.2272  
Number of obs: 49, groups:  site, 18

Fixed effects:
            Estimate Std. Error t value
(Intercept) -0.03862    0.08458  -0.457
speciesEPFU  0.20691    0.11698   1.769
speciesMYLU  0.13551    0.08910   1.521
speciesPESU  0.08589    0.09140   0.940
early.prev   0.48082    0.70962   0.678

Correlation of Fixed Effects:
            (Intr) spEPFU spMYLU spPESU
speciesEPFU -0.433                     
speciesMYLU -0.482  0.398              
speciesPESU -0.516  0.418  0.483       
early.prev  -0.158 -0.045 -0.234  0.006

5 OVERWINTER MOBILITY

5.1 Changes in counts over winter - site

summary(mod.overwinter.movements2)

Linear mixed model fit by REML ['lmerMod']
Formula: log.overwinter.lambda ~ log.early.winter.count + species + (1 |      site)
   Data: preWNS

REML criterion at convergence: 46.4

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-2.83773 -0.48121  0.02284  0.56248  2.99240 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 0.02909  0.1706  
 Residual             0.07031  0.2652  
Number of obs: 88, groups:  site, 16

Fixed effects:
                        Estimate Std. Error t value
(Intercept)             0.277126   0.113676   2.438
log.early.winter.count -0.152472   0.057270  -2.662
speciesMYLU             0.018150   0.090899   0.200
speciesMYSE            -0.103511   0.100143  -1.034
speciesPESU            -0.009488   0.088944  -0.107

Correlation of Fixed Effects:
            (Intr) lg.r.. spMYLU spMYSE
lg.rly.wnt. -0.744                     
speciesMYLU -0.409  0.004              
speciesMYSE -0.556  0.283  0.521       
speciesPESU -0.571  0.208  0.602  0.568

5.2 Does the log transformation of lambda help normality?

5.3 Do model diagnostic plots look reasonable?

5.4 If we preserve structure of data and use Gamma distribution do we get a similar finding?

summary(mod.overwinter.movements2g)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: Gamma  ( log )
Formula: overwinter.lambda ~ log.early.winter.count + species + (1 | site)
   Data: preWNS

     AIC      BIC   logLik deviance df.resid 
   200.0    217.4    -93.0    186.0       81 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-1.4921 -0.5895 -0.0835  0.4068  3.9362 

Random effects:
 Groups   Name        Variance Std.Dev.
 site     (Intercept) 0.2064   0.4544  
 Residual             0.3322   0.5763  
Number of obs: 88, groups:  site, 16

Fixed effects:
                       Estimate Std. Error t value Pr(>|z|)   
(Intercept)             0.91237    0.28271   3.227  0.00125 **
log.early.winter.count -0.38356    0.12130  -3.162  0.00157 **
speciesMYLU            -0.05143    0.19464  -0.264  0.79160   
speciesMYSE            -0.30116    0.22053  -1.366  0.17207   
speciesPESU            -0.14827    0.19353  -0.766  0.44360   
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) lg.r.. spMYLU spMYSE
lg.rly.wnt. -0.654                     
speciesMYLU -0.409  0.070              
speciesMYSE -0.520  0.322  0.580       
speciesPESU -0.537  0.254  0.652  0.637

6 NUMEROUS COVARIATES AND THE EFFECT ON TIMING OF DETECTION

6.1 temperature

summary(gob8)


Call:
glm(formula = pd.arrival ~ mean.temp, family = "binomial", data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.1451  -0.9399  -0.7092   1.4147   1.6224  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)  -2.1649     1.7413  -1.243    0.214
mean.temp     0.2083     0.2313   0.900    0.368

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 26.734  on 20  degrees of freedom
Residual deviance: 25.839  on 19  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 29.839

Number of Fisher Scoring iterations: 4

6.2 vapor pressure deficit

summary(gob9a)


Call:
glm(formula = pd.arrival ~ mean.vpd, family = "binomial", data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.1390  -0.9353  -0.7392   1.2213   1.7577  

Coefficients:
             Estimate Std. Error z value Pr(>|z|)
(Intercept)  -0.04919    0.68436  -0.072    0.943
mean.vpd    -17.61464   15.46895  -1.139    0.255

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 26.734  on 20  degrees of freedom
Residual deviance: 25.079  on 19  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 29.079

Number of Fisher Scoring iterations: 4

6.3 total bats

summary(gob1)


Call:
glm(formula = pd.arrival ~ log10(total), family = "binomial", 
    data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.0612  -0.8776  -0.8166   1.4364   1.5859  

Coefficients:
             Estimate Std. Error z value Pr(>|z|)
(Intercept)   -1.2752     1.3300  -0.959    0.338
log10(total)   0.2384     0.5736   0.416    0.678

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 27.522  on 21  degrees of freedom
Residual deviance: 27.350  on 20  degrees of freedom
AIC: 31.35

Number of Fisher Scoring iterations: 4

6.4 number of MYLU

summary(gob4)


Call:
glm(formula = pd.arrival ~ sum.mylu, family = "binomial", data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-0.8931  -0.8930  -0.8903   1.4917   1.5964  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)
(Intercept) -7.133e-01  4.785e-01  -1.491    0.136
sum.mylu    -6.548e-05  2.087e-04  -0.314    0.754

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 27.522  on 21  degrees of freedom
Residual deviance: 27.409  on 20  degrees of freedom
AIC: 31.409

Number of Fisher Scoring iterations: 4

6.5 number of EPFU

summary(gob6)


Call:
glm(formula = pd.arrival ~ sum.epfu, family = "binomial", data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-0.8916  -0.8908  -0.8832   1.4931   1.6046  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)
(Intercept) -0.7172423  0.4970035  -1.443    0.149
sum.epfu    -0.0008242  0.0037308  -0.221    0.825

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 27.522  on 21  degrees of freedom
Residual deviance: 27.470  on 20  degrees of freedom
AIC: 31.47

Number of Fisher Scoring iterations: 4

6.6 overwinter mobility

summary(gob14)


Call:
glm(formula = pd.arrival ~ log.overwinter.lambda, family = "binomial", 
    data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.0806  -0.7973  -0.7248   0.1033   1.6431  

Coefficients:
                      Estimate Std. Error z value Pr(>|z|)
(Intercept)            -0.8887     0.6109  -1.455    0.146
log.overwinter.lambda  -2.1356     2.6492  -0.806    0.420

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 17.995  on 15  degrees of freedom
Residual deviance: 17.161  on 14  degrees of freedom
  (6 observations deleted due to missingness)
AIC: 21.161

Number of Fisher Scoring iterations: 4

6.7 effect of early winter sampling date

gob15 = glm(pd.arrival~early.pdates, family = "binomial", data=site.dat);summary(gob15)


Call:
glm(formula = pd.arrival ~ early.pdates, family = "binomial", 
    data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.1193  -0.9260  -0.8429   1.3045   1.5832  

Coefficients:
             Estimate Std. Error z value Pr(>|z|)
(Intercept)   -9.9006    17.3967  -0.569    0.569
early.pdates   0.8049     1.5068   0.534    0.593

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 25.898  on 19  degrees of freedom
Residual deviance: 25.613  on 18  degrees of freedom
  (2 observations deleted due to missingness)
AIC: 29.613

Number of Fisher Scoring iterations: 4

6.8 species richness

gob16 = glm(pd.arrival~spec.num, family = "binomial", data=site.dat);summary(gob16)


Call:
glm(formula = pd.arrival ~ spec.num, family = "binomial", data = site.dat)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.0358  -0.8487  -0.8487   1.3259   1.7686  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)  -2.3155     2.0912  -1.107    0.268
spec.num      0.4932     0.6356   0.776    0.438

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 27.522  on 21  degrees of freedom
Residual deviance: 26.871  on 20  degrees of freedom
AIC: 30.871

Number of Fisher Scoring iterations: 4

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

R Notebook - Appendix for Langwig et al. Midwinter Arrival Ms