First thing first: check out the model results and look for interesting species pairs

Relative importance of covariates

p_relimp_pois

Groupers

# Relative importance
grps_pois_relimp
## # A tibble: 4 x 6
##   species env_rel_imp anthro_rel_imp biotic_rel_imp env_bio_rel_imp
##   <chr>         <dbl>          <dbl>          <dbl>           <dbl>
## 1 Epinep~       0.201         0.0108         0.426           0.0598
## 2 Epinep~       0             0.731          0.152           0     
## 3 Serran~       0.933         0.0147         0.0262          0     
## 4 Serran~       0.879         0.104          0               0     
## # ... with 1 more variable: mpa_bio_rel_imp <dbl>
# And coefficients
grps_coefs
## # A tibble: 16 x 4
##    species                coefficient_type   value direction
##    <chr>                  <chr>              <dbl> <chr>    
##  1 Epinephelus.costae     bio               0.179  pos      
##  2 Epinephelus.marginatus bio               0.0888 pos      
##  3 Serranus.cabrilla      bio               0.0906 pos      
##  4 Epinephelus.costae     bio_env           0.0475 pos      
##  5 Epinephelus.costae     bio_mpa          -0.100  neg      
##  6 Epinephelus.marginatus bio_mpa          -0.0483 neg      
##  7 Epinephelus.costae     env               0.0871 pos      
##  8 Serranus.cabrilla      env              -0.498  neg      
##  9 Serranus.cabrilla      env               0.277  pos      
## 10 Serranus.scriba        env              -0.390  neg      
## 11 Epinephelus.costae     mpa              -0.0202 neg      
## 12 Epinephelus.marginatus mpa               0.195  pos      
## 13 Serranus.cabrilla      mpa              -0.0679 neg      
## 14 Serranus.scriba        mpa               0.0998 pos      
## 15 Serranus.cabrilla      temp             -0.498  neg      
## 16 Serranus.scriba        temp             -0.260  neg

Species pairs for visualisations

Find out which species pairs are effected by covarites interaction

Temperature

lapply(grps_pois$key_coefs, function(x) x %>% 
         filter(str_detect(string = Variable, pattern = "temp_")))
## $Epinephelus.costae
##                 Variable Rel_importance Standardised_coef Raw_coef
## 1 temp_Serranus.cabrilla     0.05981719          0.047545 0.047545
## 
## $Epinephelus.marginatus
## [1] Variable          Rel_importance    Standardised_coef Raw_coef         
## <0 rows> (or 0-length row.names)
## 
## $Serranus.cabrilla
## [1] Variable          Rel_importance    Standardised_coef Raw_coef         
## <0 rows> (or 0-length row.names)
## 
## $Serranus.scriba
## [1] Variable          Rel_importance    Standardised_coef Raw_coef         
## <0 rows> (or 0-length row.names)

Epinephelus.costae ~ Serranus.cabrilla x Temperature (RI = 0.08; Coef = 0.06) Serranus.cabrilla ~ Epinephelus.costae x Temperature (RI = 0.01; Coef = 0.06)

MPA

lapply(grps_pois$key_coefs, function(x) x %>% 
         filter(str_detect(string = Variable, pattern = "mpa_")))
## $Epinephelus.costae
##                     Variable Rel_importance Standardised_coef    Raw_coef
## 1      mpa_Serranus.cabrilla     0.07097833       -0.05179105 -0.05179105
## 2 mpa_Epinephelus.marginatus     0.06184192       -0.04834296 -0.04834296
## 
## $Epinephelus.marginatus
##                 Variable Rel_importance Standardised_coef    Raw_coef
## 1 mpa_Epinephelus.costae     0.04490714       -0.04834296 -0.04834296
## 
## $Serranus.cabrilla
## [1] Variable          Rel_importance    Standardised_coef Raw_coef         
## <0 rows> (or 0-length row.names)
## 
## $Serranus.scriba
## [1] Variable          Rel_importance    Standardised_coef Raw_coef         
## <0 rows> (or 0-length row.names)

Epinephelus.costae ~ Serranus.cabrilla x MPA (RI = 0.07; Coef = -0.05) Epinephelus.costae ~ Epinephelus.marginatus x MPA (RI = 0.07; Coef = -0.05) Epinephelus.marginatus ~ Epinephelus.costae x MPA (RI = 0.06; Coef = -0.05)

Visualise abundance of species i

  1. Using MRF_predict to predict the abundance of species i along temperature gradient / MPA status in two scenarios:
  1. When species j is absent
  2. When species j is present (at its 90th percentile of abundance)
  1. Using raw data (transformed abundance of species i) and smoothed GAM function along temperature gradient / MPA status in two scenarios:
  1. When species j is absent
  2. When species j is present (at any abundance)

Epinephelus.costae ~ Serranus.cabrilla x Temperature RI = 0.08 Coef = 0.06

Serranus.cabrilla ~ Epinephelus.costae x Temperature RI = 0.01 Coef = 0.06

Epinephelus.costae ~ Serranus.cabrilla x MPA RI = 0.07 Coef = -0.05

Epinephelus.costae ~ Epinephelus.marginatus x MPA RI = 0.07 Coef = -0.05

Epinephelus.marginatus ~ Epinephelus.costae x MPA RI = 0.06 Coef = -0.05

Seabreams

# Relative importance
dip_pois_relimp
## # A tibble: 4 x 6
##   species env_rel_imp anthro_rel_imp biotic_rel_imp env_bio_rel_imp
##   <chr>         <dbl>          <dbl>          <dbl>           <dbl>
## 1 Diplod~       0.818          0             0.0653          0.0714
## 2 Diplod~       0.102          0.341         0.182           0.0110
## 3 Diplod~       0.499          0             0.210           0.0399
## 4 Diplod~       0.364          0             0.204           0.0551
## # ... with 1 more variable: mpa_bio_rel_imp <dbl>
# And coefficients
dip_coefs
## # A tibble: 24 x 4
##    species            coefficient_type   value direction
##    <chr>              <chr>              <dbl> <chr>    
##  1 Diplodus.annularis bio              -0.0315 neg      
##  2 Diplodus.annularis bio               0.0521 pos      
##  3 Diplodus.puntazzo  bio               0.141  pos      
##  4 Diplodus.sargus    bio               0.349  pos      
##  5 Diplodus.vulgaris  bio               0.244  pos      
##  6 Diplodus.annularis bio_env           0.0636 pos      
##  7 Diplodus.puntazzo  bio_env          -0.0272 neg      
##  8 Diplodus.sargus    bio_env           0.116  pos      
##  9 Diplodus.vulgaris  bio_env           0.173  pos      
## 10 Diplodus.annularis bio_mpa           0.0459 pos      
## # ... with 14 more rows

Species pairs for visualisations

Find out which species pairs are effected by covarites interaction

Temperature

lapply(dip_pois$key_coefs, function(x) x %>% 
         filter(str_detect(string = Variable, pattern = "temp_")))
## $Diplodus.annularis
##                 Variable Rel_importance Standardised_coef   Raw_coef
## 1 temp_Diplodus.vulgaris     0.07138733        0.06358139 0.06358139
## 
## $Diplodus.puntazzo
##                 Variable Rel_importance Standardised_coef    Raw_coef
## 1 temp_Diplodus.vulgaris     0.01104622       -0.02722366 -0.02722366
## 
## $Diplodus.sargus
##                 Variable Rel_importance Standardised_coef  Raw_coef
## 1 temp_Diplodus.vulgaris     0.03994941         0.1158576 0.1158576
## 
## $Diplodus.vulgaris
##                  Variable Rel_importance Standardised_coef   Raw_coef
## 1    temp_Diplodus.sargus     0.04128987        0.10984683 0.11585760
## 2 temp_Diplodus.annularis     0.01383338        0.06358139 0.06358139

Diplodus.annularis ~ Diplodus.vulgaris x Temperature (RI = 0.1; Coef = 0.08) Diplodus.puntazzo ~ Diplodus.vulgaris x Temperature (RI = 0.01; Coef = -0.03) Diplodus.sargus ~ Diplodus.vulgaris x Temperature (RI = 0.05; Coef = 0.13) Diplodus.vulgaris ~ Diplodus.sargus x Temperature (RI = 0.05; Coef = 0.13) Diplodus.vulgaris ~ Diplodus.annularis x Temperature (RI = 0.02; Coef = 0.08)

MPA

lapply(dip_pois$key_coefs, function(x) x %>% 
         filter(str_detect(string = Variable, pattern = "mpa_")))
## $Diplodus.annularis
##                Variable Rel_importance Standardised_coef   Raw_coef
## 1 mpa_Diplodus.vulgaris     0.03718557        0.04588879 0.04588879
## 
## $Diplodus.puntazzo
##                Variable Rel_importance Standardised_coef  Raw_coef
## 1 mpa_Diplodus.vulgaris      0.3320568         0.1492608 0.1492608
## 
## $Diplodus.sargus
##                Variable Rel_importance Standardised_coef  Raw_coef
## 1 mpa_Diplodus.vulgaris      0.1918045         0.2538626 0.2538626
## 
## $Diplodus.vulgaris
##                Variable Rel_importance Standardised_coef  Raw_coef
## 1   mpa_Diplodus.sargus      0.2205292         0.2538626 0.2538626
## 2 mpa_Diplodus.puntazzo      0.0762360         0.1492608 0.1492608

Diplodus.annularis ~ Diplodus.vulgaris x MPA (RI = 0.05; Coef = 0.06) Diplodus.puntazzo ~ Diplodus.vulgaris x MPA (RI = 0.34; Coef = 0.15) Diplodus.sargus ~ Diplodus.vulgaris x MPA (RI = 0.2; Coef = 0.26) Diplodus.vulgaris ~ Diplodus.sargus x MPA (RI = 0.21; Coef = 0.26) Diplodus.vulgaris ~ Diplodus.puntazzo x MPA (RI = 0.07; Coef = 0.15) Diplodus.vulgaris ~ Diplodus.annularis x MPA (RI = 0.01; Coef = 0.06)

Visualise abundance of species i

  1. Using MRF_predict to predict the abundance of species i along temperature gradient / MPA status in two scenarios:
  1. When species j is absent
  2. When species j is present (at its 90th percentile of abundance)
  1. Using raw data (transformed abundance of species i) and smoothed GAM function along temperature gradient / MPA status in two scenarios:
  1. When species j is absent
  2. When species j is present (at any abundance)

Diplodus.annularis ~ Diplodus.vulgaris x Temperature RI = 0.1 Coef = 0.08 Diplodus.vulgaris ~ Diplodus.annularis x Temperature RI = 0.02 Coef = 0.08

Diplodus.puntazzo ~ Diplodus.vulgaris x Temperature RI = 0.01 Coef = -0.03

Diplodus.sargus ~ Diplodus.vulgaris x Temperature RI = 0.05 Coef = 0.13

Diplodus.vulgaris ~ Diplodus.sargus x Temperature RI = 0.05 Coef = 0.13

Diplodus.annularis ~ Diplodus.vulgaris x MPA RI = 0.05 Coef = 0.06

Diplodus.vulgaris ~ Diplodus.annularis x MPA RI = 0.01; Coef = 0.06

Diplodus.puntazzo ~ Diplodus.vulgaris x MPA RI = 0.34 Coef = 0.15

Diplodus.vulgaris ~ Diplodus.puntazzo x MPA RI = 0.07 Coef = 0.15

Diplodus.sargus ~ Diplodus.vulgaris x MPA RI = 0.2; Coef = 0.26

Diplodus.vulgaris ~ Diplodus.sargus x MPA RI = 0.21; Coef = 0.26