Birds_Maquis_analysis_by_plot_no_subunit_interactions
Ron Chen
2024-01-15
## Loading required package: readxl## Loading required package: lubridate##
## Attaching package: 'lubridate'## The following objects are masked from 'package:data.table':
##
## hour, isoweek, mday, minute, month, quarter, second, wday, week,
## yday, year## The following objects are masked from 'package:base':
##
## date, intersect, setdiff, union## Loading required package: solartime## - Winter data excluded.
## - The counts of all passing / wintering species (not breeding in the sampled unit) were set to zero. Abreed_and_non_breed contains the non-breeders as well.
## - Species less likely to be interacting with sampling location were EXCLUDED
## - observations with counts greater than or equal to 10 standard deviations AND greater than or equal to 60 individuals were set to 1,
## under the assumption that these are migrating or local wandering phenomena.
## No outliers found in unit Arid South.
## 2 outliers set equal to 1 in Herbaceous and Dwarf-Shrub Vegetation unit.
## unit point_name year
## 1: Herbaceous and Dwarf-Shrub Vegetation Gamla Near 3 2014
## 2: Herbaceous and Dwarf-Shrub Vegetation Yavneel Far 2 2018
## SciName count_under_250 Z_score
## 1: Passer domesticus 62 11.90316
## 2: Passer hispaniolensis 60 11.35733
## No outliers found in unit Inland Sands.
## No outliers found in unit Loess Covered Areas in the Northern Negev.
## No outliers found in unit Mediterranean-Desert Transition Zone.
## 5 outliers set equal to 1 in Mediterranean Maquis unit.
## unit point_name year SciName
## 1: Mediterranean Maquis Beit Oren Near 1 2019 Curruca curruca
## 2: Mediterranean Maquis Beit Oren Near 3 2019 Curruca curruca
## 3: Mediterranean Maquis Kerem Maharal Near 3 2021 Chloris chloris
## 4: Mediterranean Maquis Kerem Maharal Near 3 2021 Garrulus glandarius
## 5: Mediterranean Maquis Nir Etzion Far 11 2021 Columba livia
## count_under_250 Z_score
## 1: 99 11.94535
## 2: 142 17.16927
## 3: 80 19.19266
## 4: 150 20.50138
## 5: 100 14.85118
## 2 outliers set equal to 1 in Negev Highlands unit.
## unit point_name year SciName
## 1: Negev Highlands Yeruham Far Slope 5 2018 Carpospiza brachydactyla
## 2: Negev Highlands Bislach Near Slope 6 2020 Passer hispaniolensis
## count_under_250 Z_score
## 1: 70 14.34802
## 2: 100 10.61343
## No outliers found in unit Planted Conifer Forest.Mediterranean Maquis
Unit is divided into 3 subunits: Judea, Carmel and Galilee. Factors are proximity to settlements and time. Sampling started in spring 2012 (pilot year), but during T0 sampling was performed only in winter of 2014, and the next spring sampling was done in T1 (2015). Therefore 2012 will not be considered as pilot here but as T0. Total 5 campaigns, 3subunits per campaign, 5 sites per subunit, with 6 plots per site (total of 450 plot-campaign combinations).
Raw data Total abundance: 21855 Number of observations: 5795 Total richness: 108
Filtered data Total abundance: 13230 Number of observations: 3576 Total richness: 49
Model gma, abundance and richness
Richness done with rare species. Abundance and mean abundance done without rare species. Full models include cosine and sine of the time difference from June 21st (in radians).
richness
Explore data and plot mean-variance plot. There is a strong relationship, indicating that employing GLMs is the proper way to analyze, rather than OLS (assumption of homogeneity is violated).
## [1] "RICHNESS WITH RARE SPECIES"
| richness | year_ct | site | settlements | subunit | td_sc | cos_td_rad | sin_td_rad | h_from_sunrise | cos_hsun | sin_hsun | monitors_name | wind | precipitation | temperature | clouds | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Min. : 0.000 | Min. :0.0 | Nir Etzion : 31 | Far :224 | Judean Highlands:149 | Min. :-1.9882 | Min. :-0.1671 | Min. :-1.0000 | Length:444 | Min. :-0.3289 | Min. :-0.1253 | Sassi Haham : 53 | 0 : 33 | 0 :119 | 0 : 4 | 0 : 32 | |
| 1st Qu.: 6.000 | 1st Qu.:3.0 | Ein Yaakov : 30 | Near:220 | Carmel :150 | 1st Qu.:-0.2713 | 1st Qu.: 0.5702 | 1st Qu.:-0.8215 | Class :difftime | 1st Qu.: 0.6855 | 1st Qu.: 0.2563 | Eyal Shochat: 47 | 1 : 42 | 3 : 2 | 1 : 38 | 1 : 0 | |
| Median : 8.000 | Median :5.0 | Givat Yearim : 30 | NA | Galilee :145 | Median : 0.4918 | Median : 0.8140 | Median :-0.5808 | Mode :numeric | Median : 0.8738 | Median : 0.4863 | Eran Banker : 30 | 2 : 12 | NA’s:323 | 2 : 37 | 2 : 5 | |
| Mean : 8.054 | Mean :4.8 | Givat Yeshayahu: 30 | NA | NA | Mean : 0.3205 | Mean : 0.7098 | Mean :-0.5864 | NA | Mean : 0.7997 | Mean : 0.4836 | Asaf Mayrose: 18 | 3 : 2 | NA | 3 : 18 | 3 : 13 | |
| 3rd Qu.:10.000 | 3rd Qu.:7.0 | Goren : 30 | NA | NA | 3rd Qu.: 1.1023 | 3rd Qu.: 0.9413 | 3rd Qu.:-0.3375 | NA | 3rd Qu.: 0.9666 | 3rd Qu.: 0.7281 | Ohad Sharir : 18 | NA’s:355 | NA | NA’s:347 | NA’s:394 | |
| Max. :18.000 | Max. :9.0 | Kerem Maharal : 30 | NA | NA | Max. : 1.7127 | Max. : 0.9976 | Max. :-0.0688 | NA | Max. : 1.0000 | Max. : 0.9998 | (Other) : 58 | NA | NA | NA | NA | |
| NA | NA | (Other) :263 | NA | NA | NA | NA | NA | NA | NA’s :89 | NA’s :89 | NA’s :220 | NA | NA | NA | NA |
no observation for all 4 weather variables. many NAs for sampling time of day variables.exclude from model. An extreme observation of richness>20 in Judean highlands near settlements.
| richness | year_ct | site | settlements | subunit | td_sc | cos_td_rad | sin_td_rad | h_from_sunrise | cos_hsun | sin_hsun | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| richness | 1.0000000 | -0.0647635 | 0.0926545 | 0.4762118 | -0.1936858 | 0.0617968 | 0.0751315 | 0.0460403 | -0.0582024 | 0.0933868 | -0.0342460 |
| year_ct | -0.0647635 | 1.0000000 | -0.0460962 | -0.0041782 | 0.0046005 | -0.6240340 | -0.5599811 | -0.6642968 | 0.0244310 | -0.0070326 | 0.0341395 |
| site | 0.0926545 | -0.0460962 | 1.0000000 | 0.0038640 | -0.1472070 | 0.0103683 | 0.0017250 | 0.0296647 | -0.0105111 | -0.0092740 | -0.0240278 |
| settlements | 0.4762118 | -0.0041782 | 0.0038640 | 1.0000000 | -0.0167082 | 0.0029695 | 0.0016285 | 0.0037459 | 0.0341209 | -0.0071595 | 0.0507160 |
| subunit | -0.1936858 | 0.0046005 | -0.1472070 | -0.0167082 | 1.0000000 | -0.0056370 | -0.0129685 | -0.0136922 | 0.1193296 | -0.1166460 | 0.1174210 |
| td_sc | 0.0617968 | -0.6240340 | 0.0103683 | 0.0029695 | -0.0056370 | 1.0000000 | 0.9788766 | 0.9693667 | -0.1975082 | 0.1811635 | -0.2043768 |
| cos_td_rad | 0.0751315 | -0.5599811 | 0.0017250 | 0.0016285 | -0.0129685 | 0.9788766 | 1.0000000 | 0.9008724 | -0.1878186 | 0.1700185 | -0.1970275 |
| sin_td_rad | 0.0460403 | -0.6642968 | 0.0296647 | 0.0037459 | -0.0136922 | 0.9693667 | 0.9008724 | 1.0000000 | -0.2005880 | 0.1868900 | -0.2040813 |
| h_from_sunrise | -0.0582024 | 0.0244310 | -0.0105111 | 0.0341209 | 0.1193296 | -0.1975082 | -0.1878186 | -0.2005880 | 1.0000000 | -0.9598144 | 0.9818950 |
| cos_hsun | 0.0933868 | -0.0070326 | -0.0092740 | -0.0071595 | -0.1166460 | 0.1811635 | 0.1700185 | 0.1868900 | -0.9598144 | 1.0000000 | -0.8930078 |
| sin_hsun | -0.0342460 | 0.0341395 | -0.0240278 | 0.0507160 | 0.1174210 | -0.2043768 | -0.1970275 | -0.2040813 | 0.9818950 | -0.8930078 | 1.0000000 |
Fit Poisson glm, check for existence of overdispersion
## [1] "Estimating overdispersion parameter phi: (Res. Dev.)/(n-p) where n=number of observations; p=number of parameters in the model."## [1] "od = 0.917251305665046"Overdispersion parameter is < 1. Choose Poisson.
## GVIF Df GVIF^(1/(2*Df))
## settlements 3.275497 1 1.809834
## year_ct 3.277089 1 1.810273
## subunit 1.001389 2 1.000347
## cos_td_rad 5.439321 1 2.332235
## sin_td_rad 6.692207 1 2.586930
## settlements:year_ct 4.690440 1 2.165742go with mixed model, attempt model selection yet.
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula:
## richness ~ settlements * year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 2115.0 2151.8 -1048.5 2097.0 435
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.04280 -0.60906 -0.07994 0.57865 2.87628
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.009067 0.09522
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.575932 0.201233 7.831 4.83e-15 ***
## settlementsNear 0.393867 0.061554 6.399 1.57e-10 ***
## year_ct -0.004072 0.009709 -0.419 0.67493
## subunitCarmel 0.082972 0.071979 1.153 0.24902
## subunitGalilee -0.198418 0.072660 -2.731 0.00632 **
## cos_td_rad 0.299159 0.148180 2.019 0.04350 *
## sin_td_rad -0.228252 0.169323 -1.348 0.17765
## settlementsNear:year_ct -0.002007 0.010854 -0.185 0.85329
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sttlmN yer_ct sbntCr sbntGl cs_td_ sn_td_
## settlmntsNr -0.190
## year_ct -0.027 0.552
## subunitCrml -0.195 -0.002 0.005
## subunitGall -0.199 0.003 -0.001 0.501
## cos_td_rad -0.931 0.008 -0.080 0.019 0.019
## sin_td_rad 0.866 -0.010 0.328 -0.006 -0.018 -0.853
## sttlmntsN:_ 0.158 -0.833 -0.660 0.003 0.003 -0.006 0.009perform stepwise model selection of poisson mixed model.
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.0035507 (tol = 0.002, component 1)## Single term deletions
##
## Model:
## richness ~ settlements * year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 2115.0
## subunit 2 2122.2
## cos_td_rad 1 2117.1
## sin_td_rad 1 2114.8
## settlements:year_ct 1 2113.0remove settlements X year.
## Single term deletions
##
## Model:
## richness ~ settlements + year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 2113.0
## settlements 1 2240.8
## year_ct 1 2111.5
## subunit 2 2120.3
## cos_td_rad 1 2115.1
## sin_td_rad 1 2112.8drop year
## Single term deletions
##
## Model:
## richness ~ settlements + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 2111.5
## settlements 1 2239.4
## subunit 2 2118.7
## cos_td_rad 1 2113.3
## sin_td_rad 1 2110.8drop sine.
## Single term deletions
##
## Model:
## richness ~ settlements + subunit + cos_td_rad + (1 | site)
## npar AIC
## <none> 2110.8
## settlements 1 2238.5
## subunit 2 2118.0
## cos_td_rad 1 2113.5Settlements, subunit and cosin time of year remain. Final model:
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: richness ~ settlements + subunit + cos_td_rad + (1 | site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 2110.8 2135.4 -1049.4 2098.8 438
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.09026 -0.59675 -0.07933 0.56203 2.96581
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.009619 0.09808
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.80687 0.07258 24.896 <2e-16 ***
## settlementsNear 0.38421 0.03401 11.296 <2e-16 ***
## subunitCarmel 0.08257 0.07349 1.124 0.2612
## subunitGalilee -0.20004 0.07405 -2.702 0.0069 **
## cos_td_rad 0.13535 0.06305 2.147 0.0318 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sttlmN sbntCr sbntGl
## settlmntsNr -0.279
## subunitCrml -0.526 0.002
## subunitGall -0.508 0.009 0.502
## cos_td_rad -0.635 -0.002 0.021 -0.003## $site
## (Intercept)
## Abirim -7.321848e-02
## Aderet 6.137108e-02
## Beit Oren -2.233669e-02
## Ein Yaakov -2.970003e-02
## Givat Yearim -7.623271e-02
## Givat Yeshayahu 9.364091e-02
## Goren -7.232404e-03
## Iftach 1.773011e-01
## Kerem Maharal 9.669359e-02
## Kfar Shamai -5.892993e-02
## Margaliot -1.725282e-03
## Nehusha -4.299542e-02
## Nir Etzion -1.521589e-01
## Ofer 8.260917e-02
## Ramat Raziel -3.076398e-02
## Yagur 2.684868e-05
##
## with conditional variances for "site"## $site
attempt to add year, see if significant:
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: richness ~ year_ct + settlements + subunit + cos_td_rad + (1 |
## site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 2112.8 2141.5 -1049.4 2098.8 437
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0946 -0.6000 -0.0782 0.5691 2.9651
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.009581 0.09788
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.8147139 0.0962891 18.847 <2e-16 ***
## year_ct -0.0008103 0.0065499 -0.124 0.9015
## settlementsNear 0.3841894 0.0340127 11.295 <2e-16 ***
## subunitCarmel 0.0824132 0.0733925 1.123 0.2615
## subunitGalilee -0.2001589 0.0739537 -2.707 0.0068 **
## cos_td_rad 0.1298876 0.0769263 1.688 0.0913 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) yer_ct sttlmN sbntCr sbntGl
## year_ct -0.658
## settlmntsNr -0.214 0.005
## subunitCrml -0.407 0.016 0.002
## subunitGall -0.391 0.013 0.009 0.502
## cos_td_rad -0.769 0.574 0.001 0.027 0.005year not significant, rightfully dropped. center time of year variable, highly correlated with intercept.
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: richness ~ settlements + subunit + cos_td_rad_c + (1 | site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 2110.8 2135.4 -1049.4 2098.8 438
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.09026 -0.59674 -0.07933 0.56201 2.96580
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.009619 0.09808
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.90293 0.05611 33.916 <2e-16 ***
## settlementsNear 0.38421 0.03401 11.296 <2e-16 ***
## subunitCarmel 0.08256 0.07349 1.124 0.2612
## subunitGalilee -0.20004 0.07404 -2.702 0.0069 **
## cos_td_rad_c 0.13534 0.06305 2.147 0.0318 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sttlmN sbntCr sbntGl
## settlmntsNr -0.363
## subunitCrml -0.663 0.002
## subunitGall -0.660 0.009 0.502
## cos_td_rd_c -0.023 -0.002 0.021 -0.003## Registered S3 methods overwritten by 'broom':
## method from
## tidy.glht jtools
## tidy.summary.glht jtools| Observations | 444 |
| Dependent variable | richness |
| Type | Mixed effects generalized linear model |
| Family | poisson |
| Link | log |
| AIC | 2110.815 |
| BIC | 2135.390 |
| Pseudo-R² (fixed effects) | 0.292 |
| Pseudo-R² (total) | 0.345 |
| exp(Est.) | S.E. | z val. | p | |
|---|---|---|---|---|
| (Intercept) | 6.706 | 0.056 | 33.916 | 0.000 |
| settlementsNear | 1.468 | 0.034 | 11.296 | 0.000 |
| subunitCarmel | 1.086 | 0.073 | 1.124 | 0.261 |
| subunitGalilee | 0.819 | 0.074 | -2.702 | 0.007 |
| cos_td_rad_c | 1.145 | 0.063 | 2.147 | 0.032 |
| Group | Parameter | Std. Dev. |
|---|---|---|
| site | (Intercept) | 0.098 |
| Group | # groups | ICC |
|---|---|---|
| site | 16 | 0.010 |
## Loading required package: Cairo## Loading required package: extrafont## Registering fonts with R## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.
## 2
## 3.141228## 2
## 46.84533## 1 2
## 6.705529 9.846757## [1] 0.4684534## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands 2.10 0.0524 Inf 1.99 2.20
## Carmel 2.18 0.0518 Inf 2.08 2.28
## Galilee 1.89 0.0526 Inf 1.79 2.00
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95## [1] 33.64275## [1] 8.328707## [1] 23.36781## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel -0.0826 0.0735 Inf -1.124 0.2612
## Judean Highlands - Galilee 0.2000 0.0740 Inf 2.702 0.0103
## Carmel - Galilee 0.2826 0.0736 Inf 3.839 0.0004
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsstatistically significant lower richness in Galilee and far from settlements. No significant change in richness over time.
geometric mean of abundance
Explore data. Exclude time of day because of high number of NAs.
## [1] "GEOMETRIC MEAN ABUNDANCE WITHOUT RARE SPECIES"## Warning: Removed 1 rows containing non-finite values (`stat_boxplot()`).
## Removed 1 rows containing non-finite values (`stat_boxplot()`).
## Removed 1 rows containing non-finite values (`stat_boxplot()`).
## Warning: Removed 1 rows containing non-finite values (`stat_boxplot()`).
## Removed 1 rows containing non-finite values (`stat_boxplot()`).
## Removed 1 rows containing non-finite values (`stat_boxplot()`).
| gma | year_ct | site | settlements | subunit | td_sc | cos_td_rad | sin_td_rad | |
|---|---|---|---|---|---|---|---|---|
| Min. : 1.000 | Min. :0.0 | Nir Etzion : 31 | Far :224 | Judean Highlands:149 | Min. :-1.9882 | Min. :-0.1671 | Min. :-1.0000 | |
| 1st Qu.: 2.074 | 1st Qu.:3.0 | Ein Yaakov : 30 | Near:220 | Carmel :150 | 1st Qu.:-0.2713 | 1st Qu.: 0.5702 | 1st Qu.:-0.8215 | |
| Median : 2.551 | Median :5.0 | Givat Yearim : 30 | NA | Galilee :145 | Median : 0.4918 | Median : 0.8140 | Median :-0.5808 | |
| Mean : 2.784 | Mean :4.8 | Givat Yeshayahu: 30 | NA | NA | Mean : 0.3205 | Mean : 0.7098 | Mean :-0.5864 | |
| 3rd Qu.: 3.248 | 3rd Qu.:7.0 | Goren : 30 | NA | NA | 3rd Qu.: 1.1023 | 3rd Qu.: 0.9413 | 3rd Qu.:-0.3375 | |
| Max. :10.198 | Max. :9.0 | Kerem Maharal : 30 | NA | NA | Max. : 1.7127 | Max. : 0.9976 | Max. :-0.0688 | |
| NA’s :1 | NA | (Other) :263 | NA | NA | NA | NA | NA |
Fit glm, compare gamma, gaussian (poisson inappropriate because response is not discrete)
Remove rows 211, 376, 408. Fit fixed and mixed models.
Mixed model converged
## Linear mixed model fit by REML ['lmerMod']
## Formula: gma ~ settlements * year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## Data: P.anal
##
## REML criterion at convergence: 1251.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0107 -0.6237 -0.1601 0.4508 6.1788
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.07354 0.2712
## Residual 0.93681 0.9679
## Number of obs: 440, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 1.132556 0.546614 2.072
## settlementsNear 0.332169 0.169117 1.964
## year_ct -0.005497 0.025215 -0.218
## subunitCarmel 0.077810 0.205187 0.379
## subunitGalilee -0.082583 0.201837 -0.409
## cos_td_rad 1.113885 0.401654 2.773
## sin_td_rad -1.307029 0.471776 -2.770
## settlementsNear:year_ct -0.033993 0.029582 -1.149
##
## Correlation of Fixed Effects:
## (Intr) sttlmN yer_ct sbntCr sbntGl cs_td_ sn_td_
## settlmntsNr -0.157
## year_ct 0.022 0.486
## subunitCrml -0.195 -0.001 0.007
## subunitGall -0.207 0.003 0.005 0.508
## cos_td_rad -0.933 0.003 -0.107 0.011 0.015
## sin_td_rad 0.869 -0.004 0.374 0.001 -0.014 -0.856
## sttlmntsN:_ 0.132 -0.838 -0.581 0.003 0.001 -0.004 0.003perform stepwise model selection of gaussian model.
## Single term deletions
##
## Model:
## gma ~ settlements * year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 1249.7
## subunit 2 1246.5
## cos_td_rad 1 1255.4
## sin_td_rad 1 1255.6
## settlements:year_ct 1 1249.0drop subunit
## Single term deletions
##
## Model:
## gma ~ settlements * year_ct + cos_td_rad + sin_td_rad + (1 |
## site)
## npar AIC
## <none> 1246.5
## cos_td_rad 1 1252.2
## sin_td_rad 1 1252.4
## settlements:year_ct 1 1245.8drop year X settlements
## Single term deletions
##
## Model:
## gma ~ settlements + year_ct + cos_td_rad + sin_td_rad + (1 |
## site)
## npar AIC
## <none> 1245.8
## settlements 1 1247.2
## year_ct 1 1245.0
## cos_td_rad 1 1251.5
## sin_td_rad 1 1251.7drop year
## Single term deletions
##
## Model:
## gma ~ settlements + cos_td_rad + sin_td_rad + (1 | site)
## npar AIC
## <none> 1245.0
## settlements 1 1246.4
## cos_td_rad 1 1250.0
## sin_td_rad 1 1249.7This is the final model:
## Linear mixed model fit by REML ['lmerMod']
## Formula: gma ~ settlements + cos_td_rad + sin_td_rad + (1 | site)
## Data: P.anal
##
## REML criterion at convergence: 1240.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.1096 -0.6337 -0.1613 0.4701 6.2240
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.06469 0.2543
## Residual 0.93758 0.9683
## Number of obs: 440, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 1.28362 0.52079 2.465
## settlementsNear 0.16956 0.09237 1.836
## cos_td_rad 1.05182 0.39719 2.648
## sin_td_rad -1.07442 0.41716 -2.576
##
## Correlation of Fixed Effects:
## (Intr) sttlmN cs_td_
## settlmntsNr -0.088
## cos_td_rad -0.964 0.000
## sin_td_rad 0.957 -0.002 -0.904## $site
## (Intercept)
## Abirim 0.02512544
## Aderet 0.39914017
## Beit Oren 0.11475809
## Ein Yaakov 0.01990634
## Givat Yearim -0.35014787
## Givat Yeshayahu 0.17737192
## Goren 0.17829055
## Iftach -0.29778493
## Kerem Maharal -0.02281586
## Kfar Shamai -0.11844217
## Margaliot -0.08711515
## Nehusha -0.15429168
## Nir Etzion -0.17312429
## Ofer 0.06568644
## Ramat Raziel -0.06810568
## Yagur 0.29154868
##
## with conditional variances for "site"## $site
Not a great fit. settlement is not significant.
| Observations | 440 |
| Dependent variable | gma |
| Type | Mixed effects linear regression |
| AIC | 1252.880 |
| BIC | 1277.401 |
| Pseudo-R² (fixed effects) | 0.024 |
| Pseudo-R² (total) | 0.087 |
| Est. | S.E. | t val. | d.f. | p | |
|---|---|---|---|---|---|
| (Intercept) | 1.284 | 0.525 | 2.447 | 406.954 | 0.015 |
| settlementsNear | 0.170 | 0.092 | 1.836 | 421.695 | 0.067 |
| cos_td_rad | 1.052 | 0.400 | 2.629 | 427.973 | 0.009 |
| sin_td_rad | -1.074 | 0.420 | -2.555 | 417.556 | 0.011 |
| p values calculated using Kenward-Roger standard errors and d.f. |
| Group | Parameter | Std. Dev. |
|---|---|---|
| site | (Intercept) | 0.254 |
| Residual | 0.968 |
| Group | # groups | ICC |
|---|---|---|
| site | 16 | 0.065 |
## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.## Warning: Removed 14 rows containing missing values (`geom_point()`).
## Removed 14 rows containing missing values (`geom_point()`).
## Warning: Removed 14 rows containing missing values (`geom_point()`).
No significant effect in GMA.
abundance
Explore data
## [1] "ABUNDANCE WITHOUT RARE SPECIES"
| abundance | year_ct | site | settlements | subunit | td_sc | cos_td_rad | sin_td_rad | |
|---|---|---|---|---|---|---|---|---|
| Min. : 0.00 | Min. :0.0 | Nir Etzion : 31 | Far :224 | Judean Highlands:149 | Min. :-1.9882 | Min. :-0.1671 | Min. :-1.0000 | |
| 1st Qu.: 15.75 | 1st Qu.:3.0 | Ein Yaakov : 30 | Near:220 | Carmel :150 | 1st Qu.:-0.2713 | 1st Qu.: 0.5702 | 1st Qu.:-0.8215 | |
| Median : 24.00 | Median :5.0 | Givat Yearim : 30 | NA | Galilee :145 | Median : 0.4918 | Median : 0.8140 | Median :-0.5808 | |
| Mean : 29.17 | Mean :4.8 | Givat Yeshayahu: 30 | NA | NA | Mean : 0.3205 | Mean : 0.7098 | Mean :-0.5864 | |
| 3rd Qu.: 36.00 | 3rd Qu.:7.0 | Goren : 30 | NA | NA | 3rd Qu.: 1.1023 | 3rd Qu.: 0.9413 | 3rd Qu.:-0.3375 | |
| Max. :221.00 | Max. :9.0 | Kerem Maharal : 30 | NA | NA | Max. : 1.7127 | Max. : 0.9976 | Max. :-0.0688 | |
| NA | NA | (Other) :263 | NA | NA | NA | NA | NA |
Some outliers with total abundance >100. Examine:
## subunit point_name datetime monitors_name
## 1: Judean Highlands Givat Yeshayahu Far 2 2017-04-20 10:20:00 Eran Banker
## 2: Judean Highlands Givat Yeshayahu Near 1 2017-04-20 06:25:00 Eran Banker
## 3: Judean Highlands Givat Yeshayahu Near 2 2017-04-20 06:50:00 Eran Banker
## 4: Judean Highlands Givat Yeshayahu Near 3 2017-04-20 07:15:00 Eran Banker
## 5: Judean Highlands Aderet Near 2 2019-05-26 06:52:00 Eran Banker
## 6: Carmel Beit Oren Near 3 2019-04-19 07:35:00 Other
## 7: Carmel Kerem Maharal Near 3 2021-04-17 07:41:00 Eliraz Dvir
## richness gma abundance
## 1: 11 5.356635 115
## 2: 13 9.170560 140
## 3: 14 7.652753 138
## 4: 16 9.098580 221
## 5: 15 5.408017 106
## 6: 10 4.471017 113
## 7: 11 4.385841 117Exclude 1 plot with high abundance (>150) to improve model fit.
PHI>1, hence choose negative binomial. Fit fixed and mixed models. Choose mixed model if possible, otherwise choose a model with fixed-effects only.
## [1] 3623.798## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00347226 (tol = 0.002, component 1)## [1] 3613.284mixed model converged, slightly better AIC. Perform stepwise model selection of mixed model.
## Single term deletions
##
## Model:
## abundance ~ settlements * year_ct + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 3613.3
## subunit 2 3616.5
## cos_td_rad 1 3628.9
## sin_td_rad 1 3627.4
## settlements:year_ct 1 3611.7drop settlements X year
## Single term deletions
##
## Model:
## abundance ~ settlements + subunit + year_ct + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 3611.7
## settlements 1 3708.2
## subunit 2 3614.9
## year_ct 1 3611.0
## cos_td_rad 1 3627.4
## sin_td_rad 1 3625.8drop year
## Single term deletions
##
## Model:
## abundance ~ settlements + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## npar AIC
## <none> 3611.0
## settlements 1 3707.6
## subunit 2 3614.2
## cos_td_rad 1 3625.7
## sin_td_rad 1 3624.5Subunit, settlement and time of year remain. The final model:
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: Negative Binomial(3.845) ( log )
## Formula: abundance ~ settlements + subunit + cos_td_rad + sin_td_rad +
## (1 | site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 3611.0 3643.7 -1797.5 3595.0 436
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8301 -0.7127 -0.2336 0.4548 6.7444
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.02135 0.1461
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.89136 0.30491 6.203 5.54e-10 ***
## settlementsNear 0.51868 0.05206 9.962 < 2e-16 ***
## subunitCarmel 0.07702 0.11192 0.688 0.4914
## subunitGalilee -0.23938 0.11080 -2.160 0.0307 *
## cos_td_rad 0.94098 0.22562 4.171 3.04e-05 ***
## sin_td_rad -0.92974 0.23312 -3.988 6.65e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sttlmN sbntCr sbntGl cs_td_
## settlmntsNr -0.114
## subunitCrml -0.229 0.026
## subunitGall -0.234 0.032 0.507
## cos_td_rad -0.940 0.018 0.047 0.039
## sin_td_rad 0.933 -0.023 -0.039 -0.056 -0.899## $site
## (Intercept)
## Abirim -0.116224848
## Aderet 0.171176492
## Beit Oren 0.066853757
## Ein Yaakov 0.022356791
## Givat Yearim -0.205123412
## Givat Yeshayahu 0.206056383
## Goren -0.014560285
## Iftach 0.084554316
## Kerem Maharal 0.106352145
## Kfar Shamai 0.021701734
## Margaliot 0.004312994
## Nehusha -0.099031455
## Nir Etzion -0.232721747
## Ofer 0.033580067
## Ramat Raziel -0.071674915
## Yagur 0.027259196
##
## with conditional variances for "site"## $site
Not a great fit, high residuals. center time of year variables, highly correlated with intercept.
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: Negative Binomial(3.845) ( log )
## Formula: abundance ~ settlements + subunit + cos_td_rad_c + sin_td_rad_c +
## (1 | site)
## Data: P.anal
##
## AIC BIC logLik deviance df.resid
## 3611.0 3643.7 -1797.5 3595.0 436
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8301 -0.7127 -0.2336 0.4548 6.7444
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.02135 0.1461
## Number of obs: 444, groups: site, 16
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.10445 0.08421 36.863 < 2e-16 ***
## settlementsNear 0.51868 0.05206 9.962 < 2e-16 ***
## subunitCarmel 0.07702 0.11192 0.688 0.4914
## subunitGalilee -0.23938 0.11081 -2.160 0.0307 *
## cos_td_rad_c 0.94099 0.22562 4.171 3.04e-05 ***
## sin_td_rad_c -0.92975 0.23311 -3.988 6.65e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sttlmN sbntCr sbntGl cs_t__
## settlmntsNr -0.340
## subunitCrml -0.674 0.026
## subunitGall -0.683 0.032 0.507
## cos_td_rd_c -0.043 0.018 0.047 0.039
## sin_td_rd_c 0.044 -0.023 -0.039 -0.056 -0.899Interpretation of abundance model:
| Observations | 444 |
| Dependent variable | abundance |
| Type | Mixed effects generalized linear model |
| Family | Negative Binomial(3.845) |
| Link | log |
| AIC | 3610.983 |
| BIC | 3643.750 |
| Pseudo-R² (fixed effects) | 0.637 |
| Pseudo-R² (total) | 0.775 |
| exp(Est.) | S.E. | z val. | p | |
|---|---|---|---|---|
| (Intercept) | 22.297 | 0.084 | 36.863 | 0.000 |
| settlementsNear | 1.680 | 0.052 | 9.962 | 0.000 |
| subunitCarmel | 1.080 | 0.112 | 0.688 | 0.491 |
| subunitGalilee | 0.787 | 0.111 | -2.160 | 0.031 |
| cos_td_rad_c | 2.563 | 0.226 | 4.171 | 0.000 |
| sin_td_rad_c | 0.395 | 0.233 | -3.988 | 0.000 |
| Group | Parameter | Std. Dev. |
|---|---|---|
| site | (Intercept) | 0.146 |
| Group | # groups | ICC |
|---|---|---|
| site | 16 | 0.065 |
## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.
## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.
## 1 2
## 22.29694 37.45466## 2
## 67.98118## Confidence intervals for merMod models is an experimental feature. The
## intervals reflect only the variance of the fixed effects, not the random
## effects.
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands 3.36 0.0792 Inf 3.21 3.52
## Carmel 3.44 0.0790 Inf 3.29 3.60
## Galilee 3.12 0.0774 Inf 2.97 3.28
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95## [1] "Carmel vs Galilee"## [1] 37.21787## [1] "Carmel vs Judea "## [1] 8.006405## [1] "Judea vs Galilee"## [1] 27.04604## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel -0.077 0.112 Inf -0.688 0.4914
## Judean Highlands - Galilee 0.239 0.111 Inf 2.160 0.0461
## Carmel - Galilee 0.316 0.111 Inf 2.861 0.0127
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testssignificantly lower abundance in Galilee subunit and far from settlements. No significant temporal trend.
community analysis using package MVabund
## Overlapping points were shifted along the y-axis to make them visible.##
## PIPING TO 2nd MVFACTOR## Only the variables Curruca.melanocephala, Turdus.merula, Streptopelia.decaocto, Passer.domesticus, Pycnonotus.xanthopygos, Parus.major, Columba.livia, Cinnyris.osea, Spilopelia.senegalensis, Garrulus.glandarius, Cecropis.daurica, Prinia.gracilis were included in the plot
## (the variables with highest total abundance).
## Overlapping points were shifted along the y-axis to make them visible.##
## PIPING TO 2nd MVFACTOR## Only the variables Curruca.melanocephala, Turdus.merula, Streptopelia.decaocto, Passer.domesticus, Pycnonotus.xanthopygos, Parus.major, Columba.livia, Cinnyris.osea, Spilopelia.senegalensis, Garrulus.glandarius, Cecropis.daurica, Prinia.gracilis were included in the plot
## (the variables with highest total abundance).
There are few observations with counts of >60. Examine these:
## point_name datetime SciName monitors_name
## 1: Givat Yeshayahu Near 3 2017-04-20 07:15:00 Passer domesticus Eran Banker
## 2: Ein Yaakov Near 31 2019-04-05 10:27:00 Passer domesticus Sassi Haham
## count_under_250
## 1: 65
## 2: 68Both observations are of P. domesticus, it is likely to be seen in large groups.
Look for species that were completely absent in certain subunits or settlemnt proximities:
| SciName | subunit | V1 | is_zero |
|---|---|---|---|
| Curruca curruca | Judean Highlands | 0 | TRUE |
| SciName | settlements | V1 | is_zero |
|---|---|---|---|
Curruca curruca was not observed in Judean Highlands at all. Validate model specifically.
start model specification:
## nb po
## 977.1228 1350.8779## [1] "POISSON"
## [1] "NEGATIVE BINOMIAL"
negative binomial model is better than poisson according to residuals and AIC comparison.
## nb po nb1
## 977.1228 1350.8779 966.4717The addition of the explanatory variable ‘site’ is somewhat improving the AIC of the model. Prefer to exclude site, for simplification. stepwise selection of model:
## Single term deletions
##
## Model:
## spp_no_rare ~ settlements * year_ct + subunit + cos_td_rad +
## sin_td_rad
## Df AIC
## <none> 23451
## subunit 48 23814
## cos_td_rad 24 23500
## sin_td_rad 24 23513
## settlements:year_ct 24 23434drop settlements X year.
## Single term deletions
##
## Model:
## spp_no_rare ~ settlements + subunit + year_ct + cos_td_rad +
## sin_td_rad
## Df AIC
## <none> 23434
## settlements 24 24225
## subunit 48 23797
## year_ct 24 23450
## cos_td_rad 24 23483
## sin_td_rad 24 23497final model includes settlements, year, subunit, sampling time of year.
##
## Test statistics:
## wald value Pr(>wald)
## (Intercept) 11.948 0.001 ***
## settlementsNear 28.844 0.001 ***
## subunitCarmel 14.252 0.001 ***
## subunitGalilee 15.421 0.001 ***
## year_ct 8.548 0.001 ***
## cos_td_rad 10.026 0.001 ***
## sin_td_rad 10.846 0.001 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Test statistic: 38.31, p-value: 0.001
## Arguments:
## Test statistics calculated assuming response assumed to be uncorrelated
## P-value calculated using 999 resampling iterations via pit.trap resampling (to account for correlation in testing).## Analysis of Deviance Table
##
## Model: spp_no_rare ~ settlements + subunit + year_ct + cos_td_rad + sin_td_rad
##
## Multivariate test:
## Res.Df Df.diff Dev Pr(>Dev)
## (Intercept) 443
## settlements 442 1 775.4 0.01 **
## subunit 440 2 435.9 0.01 **
## year_ct 439 1 131.2 0.01 **
## cos_td_rad 438 1 47.5 0.03 *
## sin_td_rad 437 1 110.6 0.01 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Univariate Tests:
## Acridotheres.tristis Alectoris.chukar
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 79.707 0.01 38.366 0.01
## subunit 39.138 0.01 10.607 0.03
## year_ct 29.844 0.01 9.298 0.04
## cos_td_rad 3.002 0.86 2.276 0.92
## sin_td_rad 5.64 0.49 0.596 0.99
## Cinnyris.osea Columba.livia Curruca.curruca
## Dev Pr(>Dev) Dev Pr(>Dev) Dev
## (Intercept)
## settlements 41.948 0.01 18.361 0.01 1.694
## subunit 17.867 0.01 17.503 0.01 28.427
## year_ct 3.939 0.57 12.046 0.04 14.39
## cos_td_rad 1.651 0.95 0.274 1.00 7.887
## sin_td_rad 2.01 0.96 11.357 0.05 5.831
## Falco.tinnunculus Passer.domesticus
## Pr(>Dev) Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 0.59 1.849 0.59 166.539 0.01
## subunit 0.01 9.122 0.06 21.305 0.01
## year_ct 0.01 0.543 0.99 1.781 0.97
## cos_td_rad 0.21 0.078 1.00 0.485 1.00
## sin_td_rad 0.49 0.641 0.98 2.962 0.85
## Prinia.gracilis Pycnonotus.xanthopygos
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 48.787 0.01 45.339 0.01
## subunit 1.864 0.73 25.29 0.01
## year_ct 16.036 0.01 2.87 0.71
## cos_td_rad 0.633 1.00 11.394 0.03
## sin_td_rad 1.004 0.98 0.499 0.99
## Spilopelia.senegalensis Streptopelia.decaocto
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 157.013 0.01 1.955 0.59
## subunit 27.046 0.01 12.168 0.02
## year_ct 2.224 0.86 0.068 0.99
## cos_td_rad 0.333 1.00 6.672 0.34
## sin_td_rad 0.603 0.99 14.008 0.01
## Streptopelia.turtur Carduelis.carduelis
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 1.175 0.59 8.13 0.10
## subunit 6.566 0.23 21.019 0.01
## year_ct 18.089 0.01 6.527 0.22
## cos_td_rad 3.045 0.86 0.389 1.00
## sin_td_rad 1.847 0.97 1.61 0.97
## Cecropis.daurica Chloris.chloris Corvus.cornix
## Dev Pr(>Dev) Dev Pr(>Dev) Dev
## (Intercept)
## settlements 33.414 0.01 23.721 0.01 22.932
## subunit 1.066 0.73 10.394 0.03 47.477
## year_ct 1.257 0.98 0.769 0.99 0.695
## cos_td_rad 0.127 1.00 0.224 1.00 1.013
## sin_td_rad 3.044 0.85 0.465 0.99 0.157
## Corvus.monedula Curruca.melanocephala
## Pr(>Dev) Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 0.01 0.428 0.59 21.24 0.01
## subunit 0.01 22.714 0.01 41.067 0.01
## year_ct 0.99 0.301 0.99 0.747 0.99
## cos_td_rad 1.00 0.184 1.00 0.3 1.00
## sin_td_rad 0.99 0.053 0.99 0.429 0.99
## Dendrocopos.syriacus Garrulus.glandarius
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 9.285 0.09 3.632 0.45
## subunit 3.648 0.48 25.056 0.01
## year_ct 0.002 0.99 4.786 0.45
## cos_td_rad 1.417 0.98 1.146 0.99
## sin_td_rad 3.35 0.77 1.103 0.97
## Parus.major Psittacula.krameri
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 3.908 0.45 15.594 0.01
## subunit 25.606 0.01 14.393 0.01
## year_ct 0.176 0.99 1.082 0.98
## cos_td_rad 3.376 0.86 0.793 1.00
## sin_td_rad 7.701 0.24 1.724 0.97
## Troglodytes.troglodytes Turdus.merula
## Dev Pr(>Dev) Dev Pr(>Dev)
## (Intercept)
## settlements 26.472 0.01 3.926 0.45
## subunit 4.411 0.39 2.142 0.73
## year_ct 0.576 0.99 3.137 0.64
## cos_td_rad 0.48 1.00 0.347 1.00
## sin_td_rad 10.786 0.05 33.19 0.01
## Arguments:
## Test statistics calculated assuming uncorrelated response (for faster computation)
## P-value calculated using 99 iterations via PIT-trap resampling.Factors settlements, year, subunit, time of year have a statistically significant effect on community composition.
Check Curruca curruca model:
##
## Call:
## glm.nb(formula = Curruca.curruca ~ settlements + subunit + year_ct +
## cos_td_rad + sin_td_rad, data = d.curcur, init.theta = 0.1967702774,
## link = log)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8887 -0.4746 -0.1231 0.0000 2.9000
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -4.252e+01 4.837e+06 0.000 1.0000
## settlementsNear -6.070e-01 4.396e-01 -1.381 0.1674
## subunitCarmel 3.401e+01 4.837e+06 0.000 1.0000
## subunitGalilee 3.494e+01 4.837e+06 0.000 1.0000
## year_ct 5.414e-02 1.052e-01 0.515 0.6067
## cos_td_rad 1.792e+00 2.100e+00 0.853 0.3935
## sin_td_rad -7.121e+00 3.136e+00 -2.271 0.0232 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for Negative Binomial(0.1968) family taken to be 1)
##
## Null deviance: 161.416 on 443 degrees of freedom
## Residual deviance: 88.683 on 437 degrees of freedom
## AIC: 261.84
##
## Number of Fisher Scoring iterations: 1
##
##
## Theta: 0.1968
## Std. Err.: 0.0612
##
## 2 x log-likelihood: -245.8440subunit factor still significant, year variable not significant. replace p_value for Curruca curruca in manyglm results with p-value obtained from individual glm.
## Loading required package: ggnewscale## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.## Warning: The `size` argument of `element_line()` is deprecated as of ggplot2 3.4.0.
## ℹ Please use the `linewidth` argument instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.## [1] "black arrow is median of synanthrope/invasive\n\n"
## [1] "black arrow is median of synanthrope/invasive\n\n"
## [[1]]
##
## [[2]]
##
## [[3]]
Alectoris chukar - חוגלת סלעים
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -3.3297204 -1.8142174 1.0363737 0.3667818 0.1915418
## cos_td_rad sin_td_rad
## 1.7562492 -1.0706266
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -3.3297134 -1.8142178 1.0363750 0.3667830 0.1915409
## cos_td_rad sin_td_rad
## 1.7562446 -1.0706260
## [1] "RESULTS FOR Alectoris chukar"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.03
## 4: year_ct 0.04
## 5: cos_td_rad 0.92
## 6: sin_td_rad 0.99
## [1] "change in Alectoris chukar abundance over monitoring period: 460.617421331887 %"
## [1] "Alectoris chukar abundance in high proximity is 513.627439919791 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -1.443 0.258 Inf -1.948 -0.9383
## Carmel -0.407 0.216 Inf -0.831 0.0171
## Galilee -1.076 0.243 Inf -1.552 -0.6010
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] 95.34401
## [1] "Carmel vs Judea (% difference)"
## [1] 181.898
## [1] "Galilee vs Judea (% difference)"
## [1] 44.30847
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel -1.036 0.330 Inf -3.141 0.0050
## Judean Highlands - Galilee -0.367 0.344 Inf -1.067 0.2859
## Carmel - Galilee 0.670 0.318 Inf 2.104 0.0531
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsAcridotheres tristis - מיינה מצויה
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.81322891 3.04949199 -0.06360271 -2.84839762 0.14854946
## cos_td_rad sin_td_rad
## 1.04586057 -2.82696864
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.81322905 3.04949186 -0.06360287 -2.84839757 0.14854959
## cos_td_rad sin_td_rad
## 1.04586061 -2.82696797
## [1] "RESULTS FOR Acridotheres tristis"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.01
## 4: year_ct 0.01
## 5: cos_td_rad 0.86
## 6: sin_td_rad 0.49
## [1] "change in Acridotheres tristis abundance over monitoring period: 280.739909808961 %"
## [1] "Acridotheres tristis abundance in high proximity is 2010.46174933252 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -1.18 0.216 Inf -1.60 -0.753
## Carmel -1.24 0.220 Inf -1.67 -0.808
## Galilee -4.02 0.430 Inf -4.87 -3.180
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] 1519.649
## [1] "Carmel vs Judea (% difference)"
## [1] -6.162242
## [1] "Galilee vs Judea (% difference)"
## [1] -94.20629
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel 0.0636 0.255 Inf 0.249 0.8032
## Judean Highlands - Galilee 2.8484 0.435 Inf 6.547 <.0001
## Carmel - Galilee 2.7848 0.433 Inf 6.436 <.0001
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsStreptopelia turtur - תור מצוי
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -4.0568827 -0.3138816 0.8126151 0.2758993 -0.1499697
## cos_td_rad sin_td_rad
## 2.8254657 -1.6813797
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -4.0568824 -0.3138816 0.8126152 0.2758993 -0.1499697
## cos_td_rad sin_td_rad
## 2.8254655 -1.6813795
## [1] "RESULTS FOR Streptopelia turtur"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.59
## 3: subunit 0.23
## 4: year_ct 0.01
## 5: cos_td_rad 0.86
## 6: sin_td_rad 0.97
## [1] "change in Streptopelia turtur abundance over monitoring period: -74.0688984897411 %"Columba livia - יונת סלעים (ותת-מין יונת הבית)
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -8.17508962 2.03815291 0.07336363 -1.60376098 0.13085672
## cos_td_rad sin_td_rad
## 4.99588496 -5.41099832
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -8.17507599 2.03815322 0.07337114 -1.60375004 0.13085565
## cos_td_rad sin_td_rad
## 4.99587573 -5.41098508
## [1] "RESULTS FOR Columba livia"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.01
## 4: year_ct 0.04
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.05
## [1] "change in Columba livia abundance over monitoring period: 224.690047628383 %"
## [1] "Columba livia abundance in high proximity is 667.641942989646 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands 0.191 0.243 Inf -0.286 0.668
## Carmel 0.264 0.243 Inf -0.211 0.740
## Galilee -1.413 0.298 Inf -1.996 -0.829
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] 435.0132
## [1] "Carmel vs Judea (% difference)"
## [1] 7.612985
## [1] "Galilee vs Judea (% difference)"
## [1] -79.88592
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel -0.0734 0.342 Inf -0.215 0.8301
## Judean Highlands - Galilee 1.6038 0.380 Inf 4.218 <.0001
## Carmel - Galilee 1.6771 0.379 Inf 4.423 <.0001
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsPasser domesticus - דרור הבית
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.2174230 4.9808727 -1.2576557 0.3878957 -0.1176087
## cos_td_rad sin_td_rad
## 1.4382456 -2.3168731
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.2174237 4.9808727 -1.2576557 0.3878958 -0.1176087
## cos_td_rad sin_td_rad
## 1.4382462 -2.3168737
## [1] "RESULTS FOR Passer domesticus"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.01
## 4: year_ct 0.97
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.85
## [1] "Passer domesticus abundance in high proximity is 14460.1398241022 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -0.912 0.270 Inf -1.44 -0.3838
## Carmel -2.170 0.304 Inf -2.76 -1.5744
## Galilee -0.524 0.260 Inf -1.03 -0.0151
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] -80.70939
## [1] "Carmel vs Judea (% difference)"
## [1] -71.56802
## [1] "Galilee vs Judea (% difference)"
## [1] 47.38762
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel 1.258 0.313 Inf 4.022 0.0001
## Judean Highlands - Galilee -0.388 0.296 Inf -1.311 0.1899
## Carmel - Galilee -1.646 0.315 Inf -5.230 <.0001
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsSpilopelia senegalensis - צוצלת
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -2.48717563 2.85687260 -0.13160712 -1.14301540 -0.04214909
## cos_td_rad sin_td_rad
## 0.63738395 -0.60406669
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -2.48716893 2.85687304 -0.13160658 -1.14301634 -0.04214947
## cos_td_rad sin_td_rad
## 0.63737835 -0.60406475
## [1] "RESULTS FOR Spilopelia senegalensis"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.01
## 4: year_ct 0.86
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.99
## [1] "Spilopelia senegalensis abundance in high proximity is 1640.70107787319 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -0.454 0.154 Inf -0.756 -0.153
## Carmel -0.586 0.157 Inf -0.895 -0.277
## Galilee -1.597 0.198 Inf -1.985 -1.210
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] 174.9474
## [1] "Carmel vs Judea (% difference)"
## [1] -12.33142
## [1] "Galilee vs Judea (% difference)"
## [1] -68.11442
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel 0.132 0.190 Inf 0.692 0.4887
## Judean Highlands - Galilee 1.143 0.220 Inf 5.195 <.0001
## Carmel - Galilee 1.011 0.222 Inf 4.560 <.0001
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsCecropis daurica - סנונית מערות
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## 0.98095288 2.16869649 -0.45368366 -0.08874088 0.10836920
## cos_td_rad sin_td_rad
## -2.15241967 2.63508635
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## 0.98094089 2.16869257 -0.45368117 -0.08874169 0.10836879
## cos_td_rad sin_td_rad
## -2.15240660 2.63507576
## [1] "RESULTS FOR Cecropis daurica"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.73
## 4: year_ct 0.98
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.85
## [1] "Cecropis daurica abundance in high proximity is 774.684066550356 % higher than low proximity."Psittacula krameri - דררה
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.269088823 1.668902826 1.190144622 -0.452394245 -0.009682361
## cos_td_rad sin_td_rad
## 0.995954283 -2.204367675
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -5.269099503 1.668904042 1.190146788 -0.452393130 -0.009682248
## cos_td_rad sin_td_rad
## 0.995960536 -2.204373874
## [1] "RESULTS FOR Psittacula krameri"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.01
## 4: year_ct 0.98
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.97
## [1] "Psittacula krameri abundance in high proximity is 430.634907481498 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -2.48 0.347 Inf -3.16 -1.801
## Carmel -1.29 0.263 Inf -1.81 -0.777
## Galilee -2.93 0.404 Inf -3.72 -2.143
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] 416.828
## [1] "Carmel vs Judea (% difference)"
## [1] 228.7564
## [1] "Galilee vs Judea (% difference)"
## [1] -36.3896
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel -1.190 0.419 Inf -2.839 0.0068
## Judean Highlands - Galilee 0.452 0.510 Inf 0.887 0.3750
## Carmel - Galilee 1.643 0.465 Inf 3.533 0.0012
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 testsPrinia gracilis - פשוש
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -0.12702277 1.51717609 0.26228731 -0.06852479 -0.08504256
## cos_td_rad sin_td_rad
## -0.36574149 0.92278481
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -0.12702190 1.51717688 0.26228733 -0.06852537 -0.08504253
## cos_td_rad sin_td_rad
## -0.36574233 0.92278602
## [1] "RESULTS FOR Prinia gracilis"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.73
## 4: year_ct 0.01
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.98
## [1] "change in Prinia gracilis abundance over monitoring period: -53.4844158204341 %"
## [1] "Prinia gracilis abundance in high proximity is 355.933546091388 % higher than low proximity."Curruca curruca - סבכי טוחנים
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -20.28530352 -0.60703439 11.77820245 12.70057714 0.05413925
## cos_td_rad sin_td_rad
## 1.79200052 -7.12131418## (Intercept) settlementsNear subunitGalilee year_ct cos_td_rad
## -8.5072544 -0.6070405 0.9223838 0.0541405 1.7920523
## sin_td_rad
## -7.1214608## [1] "RESULTS FOR Curruca curruca"## [1] "P-values"## term p
## 1: (Intercept) NA
## 2: settlements 0.5900000
## 3: subunit 0.0100000
## 4: year_ct 0.6066676
## 5: cos_td_rad 0.2100000
## 6: sin_td_rad 0.4900000
## subunit emmean SE df asymp.LCL asymp.UCL
## Carmel -3.08 0.463 Inf -3.99 -2.18
## Galilee -2.16 0.398 Inf -2.94 -1.38
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] -60.24298
## [1] "Carmel vs Judea (% difference)"
## numeric(0)
## [1] "Galilee vs Judea (% difference)"
## numeric(0)
## contrast estimate SE df z.ratio p.value
## Carmel - Galilee -0.922 0.443 Inf -2.080 0.0375
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.Chloris chloris - ירקון
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -0.4660311 1.2206656 -0.9393722 -0.1550147 0.0592256
## cos_td_rad sin_td_rad
## -0.3608733 0.7592962
## (Intercept) settlementsNear subunitCarmel subunitGalilee year_ct
## -0.46602851 1.22066509 -0.93937302 -0.15501467 0.05922575
## cos_td_rad sin_td_rad
## -0.36087540 0.75929856
## [1] "RESULTS FOR Chloris chloris"
## [1] "P-values"
## term p
## 1: (Intercept) NA
## 2: settlements 0.01
## 3: subunit 0.03
## 4: year_ct 0.99
## 5: cos_td_rad 1.00
## 6: sin_td_rad 0.99
## [1] "Chloris chloris abundance in high proximity is 238.944126900858 % higher than low proximity."
## subunit emmean SE df asymp.LCL asymp.UCL
## Judean Highlands -0.273 0.190 Inf -0.646 0.1001
## Carmel -1.212 0.223 Inf -1.650 -0.7747
## Galilee -0.428 0.197 Inf -0.814 -0.0417
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
## [1] "Carmel vs Galilee (% difference)"
## [1] -54.35875
## [1] "Carmel vs Judea (% difference)"
## [1] -60.91272
## [1] "Galilee vs Judea (% difference)"
## [1] -14.35974
## contrast estimate SE df z.ratio p.value
## Judean Highlands - Carmel 0.939 0.291 Inf 3.227 0.0038
## Judean Highlands - Galilee 0.155 0.273 Inf 0.569 0.5697
## Carmel - Galilee -0.784 0.296 Inf -2.651 0.0120
##
## Results are averaged over the levels of: settlements
## Results are given on the log (not the response) scale.
## P value adjustment: fdr method for 3 tests