Model_Damb_bio_27feb23.Rmd

---
title: "Damb_climatic-only_Niche_models"
author: "Laura J. Giraldo Kalil"
date: '2023-february-27'
output: html_document and several raster(tiff), csv, txt and graphic files
---

Wallace Session 2021-10-015

Edited and personalized by Laura J. Giraldo Kalil 
27 february 2023


This file contains species niche ecological models from the manuscript "Ecological niche comparison among closely related tree species of Lauraceae using climatic and edaphic data." Submited for its publication in Frontiers of Biogeography


The models and the main structure of the file were built with 
Wallace v2.0.0 for Damburneya ambigens, D. colorata, D. gentlei and D.salicifolia
usign five climate 
variables(Bio1,Bio4,Bio12,Bio 14,Bio 15) at 1km resolution

The Environmental raster files are named with the following format: 'Bio_1.tif'

This script contains the steps for MaxEnt niche modeling, and niche overlap analyses with ecospat and NicheRover. Environmental layers and species occurrence data must be provided and prepared previously by the user. MaxEnt response curves are generated only for the chosen model.

Occurrence data of all species used here can be found in a single file. It is available in the Appendix S2, and it was named here as: "TODAS_filtradas_feb23.csv"

More information: <https://wallaceecomod.github.io/>


The function "sdm_threshold" (by Cecina Babish) was used to apply 10 percentile threshold to the predicted model raster. The function is available at: <https://babichmorrowc.github.io/post/2019-04-12-sdm-threshold/>
================


### Package installation

Wallace uses the following R packages that must be installed and loaded
before starting.

```{r}
library(spocc)
library(spThin)
library(dismo)
library(rgeos)
library(ENMeval)
library(wallace)
library(here)
library(ggplot2)
library(gridExtra)
library(terra)
library(ade4)
library(nicheROVER)
library(dplyr)
library(gridtext)
library(cowplot)
library(ggpubr)
library(scales)

```


------------------------------------------------------------------------

## Analysis for *Damburneya ambigens* (Da)

User CSV path with occurrence data. If the CSV file is not in the
current workspace, change to the correct file path
(e.g. “/Users/darwin/Documents/occs/”).

```{r}
# NOTE: provide the folder path of the .csv file
occs_path <- here("Distr_Nectandras","Ocurrencias_filtradas")
occs_path <- file.path(occs_path,"TODAS_filtradas_feb23.csv")

# get a list of species occurrence data
userOccs_Da <- occs_userOccs(
  txtPath = occs_path, 
  txtName = "TODAS_filtradas_feb23.csv", 
  txtSep = ",", 
  txtDec = ".")
occs_Da <- userOccs_Da$Damburneya_ambigens$cleaned
```


### Save cleaned occurrences data
Save cleaned occurrences

```{r}
write.csv(occs_Da,
          here("Distr_Nectandras","Ocurrencias_filtradas",
               "cleanoccs_Dambigens.csv"))
```



### Obtain environmental data

Using user-specified variables.

```{r}
## Specify the directory with the environmental variables
dir_envs_Da <- here("Rasters","Seleccionados","cortados_biosoils")
envs_path <- file.path(dir_envs_Da, 
                       c('Bio_1.tif' , 'Bio_4.tif', 'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'))

# Create environmental object 
envs_Da <- envs_userEnvs(
  rasPath = envs_path,
  rasName = c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'),
  doBrick = FALSE)

occs_xy_Da <- occs_Da[c('longitude', 'latitude')]

occs_vals_Da <- as.data.frame(raster::extract(envs_Da, occs_xy_Da))

# remove occurrence records with NA environmental values
occs_Da <- occs_Da[!(rowSums(is.na(occs_vals_Da)) > 1), ]

# also remove variable value rows with NA environmental values
occs_vals_Da <- na.omit(occs_vals_Da)

# add columns for env variable values for each occurrence record
occs_Da <- cbind(occs_Da, occs_vals_Da)
```

### Process environmental data

Sampling of 10000 background points and corresponding environmental data
using a user provided background extent with a 0.009 degree buffer.

```{r}
# Load the user provided shapefile or csv file with the desired extent.
  ##User must input the path to shapefile or csv file and the file name 
# Define path
bgPath_Da <- list.files(path='C:/Users/Laura G/Documents/SIG_DAMBURNEYA/Poligonos/M_pol/Dambigens_M',  full.names = TRUE)

bgExt_Da <- penvs_userBgExtent(
  bgShp_path = paste0(bgPath_Da,"M_Dambigens", ".shp"),
  bgShp_name = paste0("M_Dambigens", c(".shp", ".shx", ".dbf")),
  userBgBuf = 0.009,
  occs = occs_Da)

# Mask environmental data to provided extent
bgMask_Da <- penvs_bgMask(
  occs = occs_Da,
  envs = envs_Da,
  bgExt = bgExt_Da)

# Sample background points from the provided area
bgSample_Da <- penvs_bgSample(
  occs = occs_Da,
  bgMask =  bgMask_Da,
  bgPtsNum = 10000)

# Extract values of environmental layers for each background point
bgEnvsVals_Da <- as.data.frame(raster::extract(bgMask_Da,  bgSample_Da))

##Add extracted values to background points table
bgEnvsVals_Da <- cbind(scientific_name = paste0("bg_", "Damburneya ambigens"), bgSample_Da,
                            occID = NA, year = NA, institution_code = NA, country = NA,
                            state_province = NA, locality = NA, elevation = NA,
                            record_type = NA, bgEnvsVals_Da)
```

### Save background points 

Save the 10000 generated background points with the extracted environmental information

```{r}
write.csv(bgEnvsVals_Da,
          here("background","background_bio",
               "Damburneya_ambigens_bgPoints_B.csv"))
```


### Partition occurrence data

Partition occurrences and background points for model training and
validation using “hierarchical checkerboard”, a spatial partition method
with an aggregation factor of 2.

```{r}
# R code to get partitioned data
groups_Da <- part_partitionOccs(
  occs = occs_Da ,
  bg =  bgSample_Da, 
  method = "cb2",
  bgMask = bgMask_Da,
  aggFact = 2) 
```


### Build and Evaluate Niche Model

Generating a species distribution model using the maxent.jar algorithm
as implemented in ENMeval V2.0 (with clamping = FALSE). For tuning using
L, LQ, P feature classes and regularization multipliers in the 1, 4
range increasing by 1. Not using any categorical predictor variables.


```{r}
# Run maxent model for the selected species
model_Da <- model_maxent(
  occs = occs_Da,
  bg = bgEnvsVals_Da,
  user.grp = groups_Da, 
  bgMsk = bgMask_Da,
  rms = c(1, 4), 
  rmsStep =  0.5,
  fcs = c('L', 'LQ','LQH','LQP','LQHP'),
  clampSel = FALSE,
  algMaxent = "maxent.jar",
  parallel = TRUE,
  numCores = 3)


#Save models for future work
saveRDS(model_Da,
        here("Modelos","bio_modelos","ambigens","model_Da_bio.rds"))


#For loading model:
#model_Da <- readRDS(here("Modelos","bio_modelos","ambigens","model_Da_bio.rds"))

```

#Export model results tables

Export tables with evaluation statistics for model evaluation

```{r}
write.csv(model_Da@results,
          here("Modelos","bio_modelos","ambigens",
               "Damburneya_ambigens_eval_Tbl_B.csv"))

write.csv(model_Da@results.partitions,
          here("Modelos","bio_modelos","ambigens",
               "Damburneya_ambigens__evalTblBins_B.csv"))
write.csv(model_Da@variable.importance,
          here("Modelos","bio_modelos","ambigens",
               "D_ambigens_var_importance_B.csv"))

```


### Visualize

Generate Maxent evaluation plots

```{r}


# Generate an evaluation plot

# create a colorblind friendly palette for evaluation plots

EVAL_palette <- c( '#4477AA', '#EE6677', '#228833', '#CCBB44', '#66CCEE', '#AA3377', '#BBBBBB')


#_____________AUC.val
maxentEvalPlot_Da_AUCval<- ENMeval::evalplot.stats(
   model_Da,"auc.val","rm","fc")+
  ylab("AUC value")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#____________AUC.diff
maxentEvalPlot_Da_AUCdiff<- ENMeval::evalplot.stats(
   model_Da,"auc.diff","rm","fc")+
  ylab("AUC diff.")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#_________OR.mtp


maxentEvalPlot_Da_ORmtp<- ENMeval::evalplot.stats(
   model_Da, "or.mtp", x.var="rm", color="fc")+
    ylab("OR.mtp")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)


#_______OR.10p

maxentEvalPlot_Da_OR10p<- ENMeval::evalplot.stats(
   model_Da,"or.10p",
   x.var="rm", color="fc")+
    ylab("OR.10p")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#______________delta AIC

maxentEvalPlot_Da_AIC<- ENMeval::evalplot.stats(
   model_Da,"delta.AICc","rm", "fc")+
  ylab("delta AIC")+
    theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)



#Extract legend for multiple plot

get_only_legend <- function(plot) {
    
plot_table <- ggplot_gtable(ggplot_build(plot)) # get tabular interpretation of plot
    
legend_plot <- which(sapply(plot_table$grobs, function(x) x$name) == "guide-box")      #  Mark only legend in plot

legend <- plot_table$grobs[[legend_plot]]   # extract legend
return(legend) # return legend
}


#Temporary plot to extract legend
for_legend_EVAL_Da<-maxentEvalPlot_Da_AIC+theme(legend.position = "bottom")

legend_EVAL_Da <- get_only_legend(for_legend_EVAL_Da)  

x_axis_title_EVAL_Da <- grid.text("Regularization multiplier",
                          just="bottom",
                          gp=gpar(fontsize=9)
                          )
EVAL_Da_title <- text_grob("D. ambigens", size = 9, face = "italic") 


tmp_combined_EVAL_Da<-grid.arrange(maxentEvalPlot_Da_AUCval,
                                maxentEvalPlot_Da_AUCdiff,
                                maxentEvalPlot_Da_ORmtp,
                                maxentEvalPlot_Da_OR10p,
                                maxentEvalPlot_Da_AIC,
                                ncol=1, nrow=5,
                                heights=c(1,1,1,1,1),
                                widths=1.2)   #temporary combined plot


#Save evaluation plots

pdf(here::here("Modelos","bio_modelos","ambigens",
         "EvalPlots_Da_B.pdf"),width = 2.5)

EVAL_Da<-grid.arrange(tmp_combined_EVAL_Da, 
             x_axis_title_EVAL_Da,legend_EVAL_Da,
             top=EVAL_Da_title,
             nrow = 3, ncol=1,
             heights=c(5,0.2,0.1),
             widths=1.2)

dev.off()

```

### Visualize
First, MaxEnt evaluation plots are analized and the best model is chosen.
Generate response plots for the selected model: LQ_rm1

```{r}
# Generate response plots


pdf(here("Modelos","bio_modelos","ambigens","responsePlots_Da_B.pdf"))
dismo::response(
   model_Da@models$fc.LQ_rm.1,col="#CC3311",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_1_resp_Da <- dismo::response(
   model_Da@models$fc.LQ_rm.1,var="Bio_1",col="#CC3311",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_4_resp_Da <- dismo::response(
   model_Da@models$fc.LQ_rm.1,var="Bio_4",col="#CC3311",expand=100,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_12_resp_Da <- dismo::response(
   model_Da@models$fc.LQ_rm.1,var="Bio_12",col="#CC3311",expand=1000,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_14_resp_Da <- dismo::response(
   model_Da@models$fc.LQ_rm.1,var="Bio_14",col="#CC3311",expand=50,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_15_resp_Da <- dismo::response(
   model_Da@models$fc.LQ_rm.1,var="Bio_15",col="#CC3311",expand=100,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

dev.off()

```


### Visualize

Generate a map of the Maxent generated model with a “p10” threshold rule (cloglog transformation)


```{r}
# Select current model and obtain raster prediction
m_Da <- model_Da@models[["fc.LQ_rm.1"]]
predSel_Da <- dismo::predict(
  m_Da, bgMask_Da,
  args = c(paste0("outputformat=", "cloglog"), 
           paste0("doclamp=", tolower(as.character(TRUE)))), 
  na.rm = TRUE)

#Save the continuous cloglog prediction raster
writeRaster(predSel_Da,here("Modelos","bio_modelos",
                            "ambigens", 
                            "cloglog_Dambigens_B.tif"), 
            format="GTiff",
            overwritte=TRUE)
#Apply the function "sdm_threshold" (by Cecina Babish) to apply 10 percentile threshold to the predicted model raster the function is available at: https://babichmorrowc.github.io/post/2019-04-12-sdm-threshold/

#Call function
sdm_threshold <- function(sdm, occs, type = "mtp", binary = FALSE){
  occPredVals <- raster::extract(sdm, occs)
  if(type == "mtp"){
    thresh <- min(na.omit(occPredVals))
  } else if(type == "p10"){
    if(length(occPredVals) < 10){
      p10 <- floor(length(occPredVals) * 0.9)
    } else {
      p10 <- ceiling(length(occPredVals) * 0.9)
    }
    thresh <- rev(sort(occPredVals))[p10]
  }
  sdm_thresh <- sdm
  sdm_thresh[sdm_thresh < thresh] <- NA
  if(binary){
    sdm_thresh[sdm_thresh >= thresh] <- 1
  }
  return(sdm_thresh)
}


#Calculate 10 percentile training presence threshold
predSel_Da_trhld<-sdm_threshold(predSel_Da,occs_Da[,3:4],"p10",binary=TRUE)

### Save binary raster prediction based on 10 percentile presence classification

writeRaster(predSel_Da_trhld,here("Modelos","bio_modelos",
                            "ambigens", 
                            "binp10_cloglog_Dambigens_B.tif"), 
            format="GTiff",
            overwritte=TRUE)

```

------------------------------------------------------------------------

## Analysis for *Damburneya colorata* (Dc)

User CSV path with occurrence data. If the CSV file is not in the
current workspace, change to the correct file path
(e.g. “/Users/darwin/Documents/occs/”).

```{r}
# NOTE: provide the folder path of the .csv file
occs_path <- here("Distr_Nectandras","Ocurrencias_filtradas")
occs_path <- file.path(occs_path,"TODAS_filtradas_feb23.csv")

# get a list of species occurrence data
userOccs_Dc <- occs_userOccs(
  txtPath = occs_path, 
  txtName = "TODAS_filtradas_feb23.csv", 
  txtSep = ",", 
  txtDec = ".")

occs_Dc <- userOccs_Dc$Damburneya_colorata$cleaned
```


### Save cleaned occurrences data
Save cleaned occurrences

```{r}
write.csv(occs_Dc,
          here("Distr_Nectandras","Ocurrencias_filtradas",
               "cleanoccs_Dcolorata_new_.csv"))
```




### Obtain environmental data

Using user-specified variables.

```{r}
## Specify the directory with the environmental variables
dir_envs_Dc <- here("Rasters","Seleccionados","cortados_biosoils")
envs_path <- file.path(dir_envs_Dc, c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'))

# Create environmental object 
envs_Dc <- envs_userEnvs(
  rasPath = envs_path,
  rasName = c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'),
  doBrick = FALSE)

occs_xy_Dc <- occs_Dc[c('longitude', 'latitude')]

occs_vals_Dc <- as.data.frame(raster::extract(envs_Dc, occs_xy_Dc))

# remove occurrence records with NA environmental values
occs_Dc <- occs_Dc[!(rowSums(is.na(occs_vals_Dc)) > 1), ]

# also remove variable value rows with NA environmental values
occs_vals_Dc <- na.omit(occs_vals_Dc)

# add columns for env variable values for each occurrence record
occs_Dc <- cbind(occs_Dc, occs_vals_Dc)
```

### Process environmental data

Sampling of 10000 background points and corresponding environmental data
using a user provided background extent with a 0.009 degree buffer.

```{r}
# Load the user provided shapefile or csv file with the desired extent.
  ##User must input the path to shapefile or csv file and the file name 
# Define path
bgPath_Dc <- list.files(path='C:/Users/Laura G/Documents/SIG_DAMBURNEYA/Poligonos/M_pol/Dcolorata_M',  full.names = TRUE)

bgExt_Dc <- penvs_userBgExtent(
  bgShp_path = paste0(bgPath_Dc, "M_Dcol_ok", ".shp"),
  bgShp_name = paste0("M_Dcol_ok", c(".shp", ".shx", ".dbf")),
  userBgBuf = 0.009,
  occs = occs_Dc)

# Mask environmental data to provided extent
bgMask_Dc <- penvs_bgMask(
  occs = occs_Dc,
  envs = envs_Dc,
  bgExt = bgExt_Dc)

# Sample background points from the provided area
bgSample_Dc <- penvs_bgSample(
  occs = occs_Dc,
  bgMask =  bgMask_Dc,
  bgPtsNum = 10000)

# Extract values of environmental layers for each background point
bgEnvsVals_Dc <- as.data.frame(raster::extract(bgMask_Dc,  bgSample_Dc))

##Add extracted values to background points table
bgEnvsVals_Dc <- cbind(scientific_name = paste0("bg_", "Damburneya colorata"), bgSample_Dc,
                            occID = NA, year = NA, institution_code = NA, country = NA,
                            state_province = NA, locality = NA, elevation = NA,
                            record_type = NA, bgEnvsVals_Dc)
```

### Save background points 

Save the 10000 generated background points with the extracted environmental information

```{r}
write.csv(bgEnvsVals_Dc,
          here("background","background_bio",
               "Damburneya_colorata_new_bgPoints_B.csv"))
```




### Partition occurrence data

Partition occurrences and background points for model training and
validation using jackknife, a non-spatial partition method.

```{r}
# R code to get partitioned data
groups_Dc <- part_partitionOccs(
  occs = occs_Dc ,
  bg =  bgSample_Dc, 
  method = "jack",
  kfolds = 10) 
```

### Build and Evaluate Niche Model

Generating a species distribution model using the maxent.jar algorithm
as implemented in ENMeval V2.0 (with clamping = FALSE). For tuning using
L, LQ, P feature classes and regularization multipliers in the 1, 4
range increasing by 1. Not using any categorical predictor variables.


```{r}
# Run maxent model for the selected species
model_Dc <- model_maxent(
  occs = occs_Dc,
  bg = bgEnvsVals_Dc,
  user.grp = groups_Dc, 
  bgMsk = bgMask_Dc,
  rms = c(1, 4), 
  rmsStep =  0.5,
  fcs = c('L', 'LQ','LQH','LQP','LQHP'),
  clampSel = FALSE,
  algMaxent = "maxent.jar",
  parallel = TRUE,
  numCores = 3)

#Save models for future work
saveRDS(model_Dc,
        here("Modelos","bio_modelos","colorata","model_Dc_new_bio.rds"))

#For loading model:
#model_Dc <- readRDS(here("Modelos","bio_modelos","colorata","model_Dcnew_bio.rds"))

```


#Export model results tables

Export tables with evaluation statistics for model evaluation

```{r}
write.csv(model_Dc@results,
          here("Modelos","bio_modelos","colorata",
               "Damburneya_colorata_new_eval_Tbl_B.csv"))

write.csv(model_Dc@results.partitions,
          here("Modelos","bio_modelos","colorata",
               "Damburneya_colorata_new_evalTblBins_B.csv"))

write.csv(model_Dc@variable.importance,
          here("Modelos","bio_modelos","colorata",
               "D_colorata_new_var_importance_B.csv"))

```


### Visualize

Generate a Maxent evaluation plot using “auc.diff” as evaluation
statistic.

```{r}


# Generate an evaluation plot

# create a colorblind friendly palette for evaluation plots

EVAL_palette <- c( '#4477AA', '#EE6677', '#228833', '#CCBB44', '#66CCEE', '#AA3377', '#BBBBBB')


#_____________AUC.val
maxentEvalPlot_Dc_AUCval<- ENMeval::evalplot.stats(
   model_Dc,"auc.val","rm","fc")+
  ylab("AUC value")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#____________AUC.diff
maxentEvalPlot_Dc_AUCdiff<- ENMeval::evalplot.stats(
   model_Dc,"auc.diff","rm","fc")+
  ylab("AUC diff.")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#_________OR.mtp


maxentEvalPlot_Dc_ORmtp<- ENMeval::evalplot.stats(
   model_Dc, "or.mtp", x.var="rm", color="fc")+
    ylab("OR.mtp")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)


#_______OR.10p

maxentEvalPlot_Dc_OR10p<- ENMeval::evalplot.stats(
   model_Dc,"or.10p",
   x.var="rm", color="fc")+
    ylab("OR.10p")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#______________delta AIC

maxentEvalPlot_Dc_AIC<- ENMeval::evalplot.stats(
   model_Dc,"delta.AICc","rm", "fc")+
  ylab("delta AIC")+
    theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)



#Extract legend for multiple plot

get_only_legend <- function(plot) {
    
plot_table <- ggplot_gtable(ggplot_build(plot)) # get tabular interpretation of plot
    
legend_plot <- which(sapply(plot_table$grobs, function(x) x$name) == "guide-box")      #  Mark only legend in plot

legend <- plot_table$grobs[[legend_plot]]   # extract legend
return(legend) # return legend
}


#Temporary plot to extract legend
for_legend_EVAL_Dc<-maxentEvalPlot_Dc_AIC+theme(legend.position = "bottom")

legend_EVAL_Dc <- get_only_legend(for_legend_EVAL_Dc)  

x_axis_title_EVAL_Dc <- grid.text("Regularization multiplier",
                          just="bottom",
                          gp=gpar(fontsize=9)
                          )
EVAL_Dc_title <- text_grob("D. colorata", size = 9, face = "italic") 


tmp_combined_EVAL_Dc<-grid.arrange(maxentEvalPlot_Dc_AUCval,
                                maxentEvalPlot_Dc_AUCdiff,
                                maxentEvalPlot_Dc_ORmtp,
                                maxentEvalPlot_Dc_OR10p,
                                maxentEvalPlot_Dc_AIC,
                                ncol=1, nrow=5,
                                heights=c(1,1,1,1,1),
                                widths=1.2)   #temporary combined plot


#Save evaluation plots

pdf(here::here("Modelos","bio_modelos","colorata",
         "EvalPlots_Dc_new_B.pdf"),width = 2.5)

EVAL_Dc<-grid.arrange(tmp_combined_EVAL_Dc, 
             x_axis_title_EVAL_Dc,legend_EVAL_Dc,
             top=EVAL_Dc_title,
             nrow = 3, ncol=1,
             heights=c(5,0.2,0.1),
             widths=1.2)

dev.off()
```


### Visualize

Generate response plots for the model LQP_rm.1

```{r}
pdf(here("Modelos","bio_modelos","colorata","responsePlots_Dc_new_B_.pdf"))
dismo::response(
   model_Dc@models$fc.LQP_rm.1,col="#4477AA",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_1_resp_Dc <- dismo::response(
   model_Dc@models$fc.LQP_rm.1,var="Bio_1",col="#4477AA",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_4_resp_Dc <- dismo::response(
   model_Dc@models$fc.LQP_rm.1,var="Bio_4",col="#4477AA",expand=300,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_12_resp_Dc <- dismo::response(
   model_Dc@models$fc.LQP_rm.1,var="Bio_12",col="#4477AA",expand=1000,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_14_resp_Dc <- dismo::response(
   model_Dc@models$fc.LQP_rm.1,var="Bio_14",col="#4477AA",expand=50,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_15_resp_Dc <- dismo::response(
   model_Dc@models$fc.LQP_rm.1,var="Bio_15",col="#4477AA",expand=100,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

dev.off()

```



### Visualize

Generate a map of the Maxent generated model with a “p10” threshold rule (cloglog transformation)

```{r}
# Select current model and obtain raster prediction
m_Dc <- model_Dc@models[["fc.LQP_rm.1"]]
predSel_Dc <- dismo::predict(
  m_Dc, bgMask_Dc,
  args = c(paste0("outputformat=", "cloglog"), 
           paste0("doclamp=", tolower(as.character(TRUE)))), 
  na.rm = TRUE)

#Save the continuous cloglog prediction raster
writeRaster(predSel_Dc,here("Modelos","bio_modelos",
                            "colorata", 
                            "cloglog_Dcolorata__new_B.tif"), 
            format="GTiff",
            overwritte=TRUE)


#Apply the function "sdm_threshold" (by Cecina Babish) to apply 10 percentile threshold to the predicted model raster 

predSel_Dc_trhld<-sdm_threshold(predSel_Dc,occs_Dc[,3:4],"p10",binary=TRUE)

### Save binary raster prediction based on 10 percentile presence classification

writeRaster(predSel_Dc_trhld,here("Modelos","bio_modelos",
                            "colorata", 
                            "binp10_cloglog_Dcolorata_new_B.tif"), 
            format="GTiff",
            overwritte=TRUE)

```

------------------------------------------------------------------------

## Analysis for *Damburneya gentlei* (Dg)

User CSV path with occurrence data. If the CSV file is not in the
current workspace, change to the correct file path
(e.g. “/Users/darwin/Documents/occs/”).

```{r}
# NOTE: provide the folder path of the .csv file
occs_path <- here("Distr_Nectandras","Ocurrencias_filtradas")

occs_path <- file.path(occs_path,"TODAS_filtradas_feb23.csv")

# get a list of species occurrence data
userOccs_Dg <- occs_userOccs(
  txtPath = occs_path, 
  txtName = "TODAS_filtradas_feb23.csv", 
  txtSep = ",", 
  txtDec = ".")

occs_Dg <- userOccs_Dg$Damburneya_gentlei$cleaned
```

### Save cleaned occurrences data
Save cleaned occurrences

```{r}
write.csv(occs_Dg,
          here("Distr_Nectandras","Ocurrencias_filtradas",
               "cleanoccs_Dgentlei.csv"))
```


### Obtain environmental data

Using user-specified variables.

```{r}
## Specify the directory with the environmental variables
dir_envs_Dg <- here("Rasters","Seleccionados","cortados_biosoils")

envs_path <- file.path(dir_envs_Dg, c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'))

# Create environmental object 
envs_Dg <- envs_userEnvs(
  rasPath = envs_path,
  rasName = c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'),
  doBrick = FALSE)

occs_xy_Dg <- occs_Dg[c('longitude', 'latitude')]

occs_vals_Dg <- as.data.frame(raster::extract(envs_Dg, occs_xy_Dg))

# remove occurrence records with NA environmental values
occs_Dg <- occs_Dg[!(rowSums(is.na(occs_vals_Dg)) > 1), ]

# also remove variable value rows with NA environmental values
occs_vals_Dg <- na.omit(occs_vals_Dg)

# add columns for env variable values for each occurrence record
occs_Dg <- cbind(occs_Dg, occs_vals_Dg)
```

### Process environmental data

Sampling of 10000 background points and corresponding environmental data
using a user provided background extent with a 0.009 degree buffer.

```{r}
# Load the user provided shapefile or csv file with the desired extent.
  ##User must input the path to shapefile or csv file and the file name 
# Define path
bgPath_Dg <- list.files(path='C:/Users/Laura G/Documents/SIG_DAMBURNEYA/Poligonos/M_pol/Dgentlei_M',  full.names = TRUE)

bgExt_Dg <- penvs_userBgExtent(
  bgShp_path = paste0(bgPath_Dg, "M_Dgentlei_ok", ".shp"),
  bgShp_name = paste0("M_Dgentlei_ok", c(".shp", ".shx", ".dbf")),
  userBgBuf = 0.009,
  occs = occs_Dg)

# Mask environmental data to provided extent
bgMask_Dg <- penvs_bgMask(
  occs = occs_Dg,
  envs = envs_Dg,
  bgExt = bgExt_Dg)

# Sample background points from the provided area
bgSample_Dg <- penvs_bgSample(
  occs = occs_Dg,
  bgMask =  bgMask_Dg,
  bgPtsNum = 10000)

# Extract values of environmental layers for each background point
bgEnvsVals_Dg <- as.data.frame(raster::extract(bgMask_Dg,  bgSample_Dg))

##Add extracted values to background points table
bgEnvsVals_Dg <- cbind(scientific_name = paste0("bg_", "Damburneya gentlei"), bgSample_Dg,
                            occID = NA, year = NA, institution_code = NA, country = NA,
                            state_province = NA, locality = NA, elevation = NA,
                            record_type = NA, bgEnvsVals_Dg)
```

### Save background points 

Save the 10000 generated background points with the extracted environmental information

```{r}
write.csv(bgEnvsVals_Dg,
          here("background","background_bio",
               "Damburneya_gentlei_bgPoints_B.csv"))
```



### Partition occurrence data

Partition occurrences and background points for model training and
validation using “hierarchical checkerboard”, a spatial partition method
with an aggregation factor of 2.

```{r}
# R code to get partitioned data
groups_Dg <- part_partitionOccs(
  occs = occs_Dg ,
  bg =  bgSample_Dg, 
  method = "cb2",
  bgMask = bgMask_Dg,
  aggFact = 2) 
```


### Build and Evaluate Niche Model

Generating a species distribution model using the maxent.jar algorithm
as implemented in ENMeval V2.0 (with clamping = TRUE). For tuning using
L, LQ, P, feature classes and regularization multipliers in the 1, 4
range increasing by 1. Not using any categorical predictor variables.

```{r}
# Run maxent model for the selected species
model_Dg <- model_maxent(
  occs = occs_Dg,
  bg = bgEnvsVals_Dg,
  user.grp = groups_Dg, 
  bgMsk = bgMask_Dg,
  rms = c(1, 4), 
  rmsStep =  0.5,
  fcs = c('L', 'LQ','LQH','LQP','LQHP'),
  clampSel = FALSE,
  algMaxent = "maxent.jar",
  parallel = TRUE,
  numCores = 3)

saveRDS(model_Dg,
        here("Modelos","bio_modelos","gentlei","model_Dg_bio.rds"))


#For loading model:
#model_Dg <- readRDS(here("Modelos","bio_modelos","gentlei","model_Dg_bio.rds"))


```


#Export model results tables

Export tables with evaluation statistics for model evaluation

```{r}
write.csv(model_Dg@results,
          here("Modelos","bio_modelos","gentlei",
               "Damburneya_gentlei_eval_Tbl_B.csv"))

write.csv(model_Dg@results.partitions,
          here("Modelos","bio_modelos","gentlei",
               "Damburneya_gentlei__evalTblBins_B.csv"))

write.csv(model_Dg@variable.importance,
          here("Modelos","bio_modelos","gentlei",
               "D_gentlei_var_importance_B.csv"))



```


### Visualize

Generate a Maxent evaluation plot 

```{r}
# Generate an evaluation plot

# create a colorblind friendly palette for evaluation plots

EVAL_palette <- c( '#4477AA', '#EE6677', '#228833', '#CCBB44', '#66CCEE', '#AA3377', '#BBBBBB')


#_____________AUC.val
maxentEvalPlot_Dg_AUCval<- ENMeval::evalplot.stats(
   model_Dg,"auc.val","rm","fc")+
  ylab("AUC value")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#____________AUC.diff
maxentEvalPlot_Dg_AUCdiff<- ENMeval::evalplot.stats(
   model_Dg,"auc.diff","rm","fc")+
  ylab("AUC diff.")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#_________OR.mtp


maxentEvalPlot_Dg_ORmtp<- ENMeval::evalplot.stats(
   model_Dg, "or.mtp", x.var="rm", color="fc")+
    ylab("OR.mtp")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)


#_______OR.10p

maxentEvalPlot_Dg_OR10p<- ENMeval::evalplot.stats(
   model_Dg,"or.10p",
   x.var="rm", color="fc")+
    ylab("OR.10p")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#______________delta AIC

maxentEvalPlot_Dg_AIC<- ENMeval::evalplot.stats(
   model_Dg,"delta.AICc","rm", "fc")+
  ylab("delta AIC")+
    theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)



#Extract legend for multiple plot

get_only_legend <- function(plot) {
    
plot_table <- ggplot_gtable(ggplot_build(plot)) # get tabular interpretation of plot
    
legend_plot <- which(sapply(plot_table$grobs, function(x) x$name) == "guide-box")      #  Mark only legend in plot

legend <- plot_table$grobs[[legend_plot]]   # extract legend
return(legend) # return legend
}


#Temporary plot to extract legend
for_legend_EVAL_Dg<-maxentEvalPlot_Dg_AIC+theme(legend.position = "bottom")

legend_EVAL_Dg <- get_only_legend(for_legend_EVAL_Dg)  

x_axis_title_EVAL_Dg <- grid.text("Regularization multiplier",
                          just="bottom",
                          gp=gpar(fontsize=9)
                          )
EVAL_Dg_title <- text_grob("D. gentlei", size = 9, face = "italic") 


tmp_combined_EVAL_Dg<-grid.arrange(maxentEvalPlot_Dg_AUCval,
                                maxentEvalPlot_Dg_AUCdiff,
                                maxentEvalPlot_Dg_ORmtp,
                                maxentEvalPlot_Dg_OR10p,
                                maxentEvalPlot_Dg_AIC,
                                ncol=1, nrow=5,
                                heights=c(1,1,1,1,1),
                                widths=1.2)   #temporary combined plot


#Save evaluation plots

pdf(here::here("Modelos","bio_modelos","gentlei",
         "EvalPlots_Dg_B.pdf"),width = 2.5)

EVAL_Dg<-grid.arrange(tmp_combined_EVAL_Dg, 
             x_axis_title_EVAL_Dg,legend_EVAL_Dg,
             top=EVAL_Dg_title,
             nrow = 3, ncol=1,
             heights=c(5,0.2,0.1),
             widths=1.2)

dev.off()
```



### Visualize

Generate response plots for the model LQHP_rm.2

```{r}
pdf(here("Modelos","bio_modelos","gentlei","responsePlots_Dg_B.pdf"))
dismo::response(
   model_Dg@models$fc.LQHP_rm.2,col="#DDAA33",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_1_resp_Dg <- dismo::response(
   model_Dg@models$fc.LQHP_rm.2,var="Bio_1",col="#DDAA33",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_4_resp_Dg <- dismo::response(
   model_Dg@models$fc.LQHP_rm.2,var="Bio_4",col="#DDAA33",expand=100,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_12_resp_Dg <- dismo::response(
   model_Dg@models$fc.LQHP_rm.2,var="Bio_12",col="#DDAA33",expand=1000,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_14_resp_Dg <- dismo::response(
   model_Dg@models$fc.LQHP_rm.2,var="Bio_14",col="#DDAA33",expand=50,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

Bio_15_resp_Dg <- dismo::response(
   model_Dg@models$fc.LQHP_rm.2,var="Bio_15",col="#DDAA33",expand=100,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

dev.off()

```




### Visualize

Generate a map of the Maxent generated model with a “p10” threshold rule


```{r}
# Select current model and obtain raster prediction
m_Dg <- model_Dg@models[["fc.LQHP_rm.2"]]

predSel_Dg <- dismo::predict(
  m_Dg, bgMask_Dg,
  args = c(paste0("outputformat=", "cloglog"), 
           paste0("doclamp=", tolower(as.character(TRUE)))), 
  na.rm = TRUE)

#Save the continuous cloglog prediction raster
writeRaster(predSel_Dg,here("Modelos","bio_modelos",
                            "gentlei", 
                            "cloglog_Dgentlei_B.tif"), 
            format="GTiff",
            overwritte=TRUE)

#Apply the function "sdm_threshold" (by Cecina Babish) to apply 10 percentile threshold to the predicted model raster 

predSel_Dg_trhld<-sdm_threshold(predSel_Dg,occs_Dg[,3:4],"p10",binary=TRUE)

### Save binary raster prediction based on 10 percentile presence classification

writeRaster(predSel_Dg_trhld,here("Modelos","bio_modelos",
                            "gentlei", 
                            "binp10_cloglog_Dgentlei_B.tif"), 
            format="GTiff",
            overwritte=TRUE)



```




------------------------------------------------------------------------

## Analysis for *Damburneya salicifolia* (Ds)

User CSV path with occurrence data. If the CSV file is not in the
current workspace, change to the correct file path
(e.g. “/Users/darwin/Documents/occs/”).

```{r}
# NOTE: provide the folder path of the .csv file
occs_path <- here("Distr_Nectandras","Ocurrencias_filtradas")

occs_path <- file.path(occs_path,"TODAS_filtradas_feb23.csv")

# get a list of species occurrence data
userOccs_Ds <- occs_userOccs(
  txtPath = occs_path, 
  txtName = "TODAS_filtradas_feb23.csv", 
  txtSep = ",", 
  txtDec = ".")

occs_Ds <- userOccs_Ds$Damburneya_salicifolia$cleaned
```

### Obtain environmental data

Using user-specified variables.

```{r}
## Specify the directory with the environmental variables
dir_envs_Ds <- here("Rasters","Seleccionados","cortados_biosoils")

envs_path <- file.path(dir_envs_Ds, c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'))

# Create environmental object 
envs_Ds <- envs_userEnvs(
  rasPath = envs_path,
  rasName = c('Bio_1.tif' , 'Bio_4.tif',  'Bio_12.tif', 'Bio_14.tif', 'Bio_15.tif'),
  doBrick = FALSE)

occs_xy_Ds <- occs_Ds[c('longitude', 'latitude')]

occs_vals_Ds <- as.data.frame(raster::extract(envs_Ds, occs_xy_Ds))

# remove occurrence records with NA environmental values
occs_Ds <- occs_Ds[!(rowSums(is.na(occs_vals_Ds)) > 1), ]

# also remove variable value rows with NA environmental values
occs_vals_Ds <- na.omit(occs_vals_Ds)

# add columns for env variable values for each occurrence record
occs_Ds <- cbind(occs_Ds, occs_vals_Ds)
```

### Save cleaned occurrences data
Save cleaned occurrences

```{r}
write.csv(occs_Ds,
          here("Distr_Nectandras","Ocurrencias_filtradas",
               "cleanoccs_Dsalicifolia.csv"))
```

### Process environmental data

Sampling of 10000 background points and corresponding environmental data
using a user provided background extent with a 0.009 degree buffer.

```{r}
# Load the user provided shapefile or csv file with the desired extent.
  ##User must input the path to shapefile or csv file and the file name 
# Define path
bgPath_Ds <- list.files(path='C:/Users/Laura G/Documents/SIG_DAMBURNEYA/Poligonos/M_pol/Dsalicifolia_M',  full.names = TRUE)

bgExt_Ds <- penvs_userBgExtent(
  bgShp_path = paste0(bgPath_Ds, "M_Dsalicifolia_ok", ".shp"),
  bgShp_name = paste0("M_Dsalicifolia_ok", c(".shp", ".shx", ".dbf")),
  userBgBuf = 0.009,
  occs = occs_Ds)

# Mask environmental data to provided extent
bgMask_Ds <- penvs_bgMask(
  occs = occs_Ds,
  envs = envs_Ds,
  bgExt = bgExt_Ds)

# Sample background points from the provided area
bgSample_Ds <- penvs_bgSample(
  occs = occs_Ds,
  bgMask =  bgMask_Ds,
  bgPtsNum = 10000)

# Extract values of environmental layers for each background point
bgEnvsVals_Ds <- as.data.frame(raster::extract(bgMask_Ds,  bgSample_Ds))

##Add extracted values to background points table
bgEnvsVals_Ds <- cbind(scientific_name = paste0("bg_", "Damburneya salicifolia"), bgSample_Ds,
                            occID = NA, year = NA, institution_code = NA, country = NA,
                            state_province = NA, locality = NA, elevation = NA,
                            record_type = NA, bgEnvsVals_Ds)
```

### Save background points 

Save the 10000 generated background points with the extracted environmental information

```{r}
write.csv(bgEnvsVals_Ds,
          here("background","background_bio",
               "Damburneya_salicifolias_bgPoints_B.csv"))
```

### Partition occurrence data

Partition occurrences and background points for model training and
validation using “hierarchical checkerboard”, a spatial partition method
with an aggregation factor of 2.

```{r}
# R code to get partitioned data
groups_Ds <- part_partitionOccs(
  occs = occs_Ds ,
  bg =  bgSample_Ds, 
  method = "cb2",
  bgMask = bgMask_Ds,
  aggFact = 2) 
```

### Build and Evaluate Niche Model

Generating a species distribution model using the maxent.jar algorithm
as implemented in ENMeval V2.0 (with clamping = TRUE). For tuning using
L, LQ, H, feature classes and regularization multipliers in the 1, 4
range increasing by 1. Not using any categorical predictor variables.

```{r}
#First, increase R memory

memory.limit(24382)

# Run maxent model for the selected species
model_Ds <- model_maxent(
  occs = occs_Ds,
  bg = bgEnvsVals_Ds,
  user.grp = groups_Ds, 
  bgMsk = bgMask_Ds,
  rms = c(1, 4), 
  rmsStep =  0.5,
  fcs = c('L', 'LQ','LQH','LQP','LQHP'),
  clampSel = FALSE,
  algMaxent = "maxent.jar",
  parallel = FALSE)

#Save models for future work
saveRDS(model_Ds,
        here("Modelos","bio_modelos","salicifolia","model_Ds_bio.rds"))


#For loading model:
#model_Ds <- readRDS(here("Modelos","bio_modelos","salicifolia","model_Ds_bio.rds"))


```

------------------------------------------------------------------------
#Export model results tables

Export tables with evaluation statistics for model evaluation

```{r}
write.csv(model_Ds@results,
          here("Modelos","bio_modelos","salicifolia",
               "Damburneya_salicifolia_eval_Tbl_B.csv"))

write.csv(model_Ds@results.partitions,
          here("Modelos","bio_modelos","salicifolia",
               "Damburneya_salicifolia__evalTblBins_B.csv"))

write.csv(model_Ds@variable.importance,
          here("Modelos","bio_modelos","salicifolia",
               "D_salicifolia_var_importance_B.csv"))

```
------------------------------------------------------------------------


### Visualize

Generate a Maxent evaluation plots

```{r}
# Generate an evaluation plot

# create a colorblind friendly palette for evaluation plots

EVAL_palette <- c( '#4477AA', '#EE6677', '#228833', '#CCBB44', '#66CCEE', '#AA3377', '#BBBBBB')


#_____________AUC.val
maxentEvalPlot_Ds_AUCval<- ENMeval::evalplot.stats(
   model_Ds,"auc.val","rm","fc")+
  ylab("AUC value")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#____________AUC.diff
maxentEvalPlot_Ds_AUCdiff<- ENMeval::evalplot.stats(
   model_Ds,"auc.diff","rm","fc")+
  ylab("AUC diff.")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#_________OR.mtp


maxentEvalPlot_Ds_ORmtp<- ENMeval::evalplot.stats(
   model_Ds, "or.mtp", x.var="rm", color="fc")+
    ylab("OR.mtp")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)


#_______OR.10p

maxentEvalPlot_Ds_OR10p<- ENMeval::evalplot.stats(
   model_Ds,"or.10p",
   x.var="rm", color="fc")+
    ylab("OR.10p")+
  xlab("Regularization multiplier")+
  theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text.x = element_blank(),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)

#______________delta AIC

maxentEvalPlot_Ds_AIC<- ENMeval::evalplot.stats(
   model_Ds,"delta.AICc","rm", "fc")+
  ylab("delta AIC")+
    theme_classic()+
  theme(legend.position="none",
          legend.title=element_blank(),
          legend.key.size = unit(0.8,"line"),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.x = element_blank(),
          legend.text= element_text (size=8),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA),
          strip.text.x = element_blank(),
          strip.background = element_rect(colour=NA),
          strip.placement = "outside")+
  scale_color_manual(values=EVAL_palette)



#Extract legend for multiple plot

get_only_legend <- function(plot) {
    
plot_table <- ggplot_gtable(ggplot_build(plot)) # get tabular interpretation of plot
    
legend_plot <- which(sapply(plot_table$grobs, function(x) x$name) == "guide-box")      #  Mark only legend in plot

legend <- plot_table$grobs[[legend_plot]]   # extract legend
return(legend) # return legend
}


#Temporary plot to extract legend
for_legend_EVAL_Ds<-maxentEvalPlot_Ds_AIC+theme(legend.position = "bottom")

legend_EVAL_Ds <- get_only_legend(for_legend_EVAL_Ds)  

x_axis_title_EVAL_Ds <- grid.text("Regularization multiplier",
                          just="bottom",
                          gp=gpar(fontsize=9)
                          )
EVAL_Ds_title <- text_grob("D. salicifolia", size = 9, face = "italic") 


tmp_combined_EVAL_Ds<-grid.arrange(maxentEvalPlot_Ds_AUCval,
                                maxentEvalPlot_Ds_AUCdiff,
                                maxentEvalPlot_Ds_ORmtp,
                                maxentEvalPlot_Ds_OR10p,
                                maxentEvalPlot_Ds_AIC,
                                ncol=1, nrow=5,
                                heights=c(1,1,1,1,1),
                                widths=1.2)   #temporary combined plot


#Save evaluation plots

pdf(here::here("Modelos","bio_modelos","salicifolia",
         "EvalPlots_Ds_B.pdf"),width = 2.5)

EVAL_Ds<-grid.arrange(tmp_combined_EVAL_Ds, 
             x_axis_title_EVAL_Ds,legend_EVAL_Ds,
             top=EVAL_Ds_title,
             nrow = 3, ncol=1,
             heights=c(5,0.2,0.1),
             widths=1.2)

dev.off()



```

### Visualize

Generate response plots for the model LQP_rm.1

```{r}
pdf(here("Modelos","bio_modelos","salicifolia","responsePlots_Ds_new_B.pdf"))
dismo::response(
   model_Ds@models$fc.LQP_rm.1,col="#44AA99",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))
Bio_1_resp_Ds <- dismo::response(
   model_Ds@models$fc.LQP_rm.1,var="Bio_1",col="#44AA99",expand=20,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))
Bio_4_resp_Ds <- dismo::response(
   model_Ds@models$fc.LQP_rm.1,var="Bio_4",col="#44AA99",fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))
Bio_12_resp_Ds <- dismo::response(
   model_Ds@models$fc.LQP_rm.1,var="Bio_12",col="#44AA99",expand=1000,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))
Bio_14_resp_Ds <- dismo::response(
   model_Ds@models$fc.LQP_rm.1,var="Bio_14",col="#44AA99",fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))
Bio_15_resp_Ds <- dismo::response(
   model_Ds@models$fc.LQP_rm.1,var="Bio_15",col="#44AA99",expand=50,fun=function(x, y, ...) predict(x, y, args="doclamp=FALSE", ...))

dev.off()


```


### Visualize

Generate a map of the Maxent generated model with a “p10” threshold rule

```{r}
# Select current model and obtain raster prediction
m_Ds <- model_Ds@models[["fc.LQP_rm.1"]]
predSel_Ds <- dismo::predict(
  m_Ds, bgMask_Ds,
  args = c(paste0("outputformat=", "cloglog"), 
           paste0("doclamp=", tolower(as.character(TRUE)))), 
  na.rm = TRUE)

#Save the continuous cloglog prediction raster
writeRaster(predSel_Ds,here("Modelos","bio_modelos",
                            "salicifolia", 
                            "cloglog_Dsalicifolia_new_B.tif"), 
            format="GTiff",
            overwritte=TRUE)


#Apply the function "sdm_threshold" (by Cecina Babish) to apply 10 percentile threshold to the predicted model raster

predSel_Ds_trhld<-sdm_threshold(predSel_Ds,occs_Ds[,3:4],"p10",binary=TRUE)

### Save binary raster prediction based on 10 percentile presence classification

writeRaster(predSel_Ds_trhld,here("Modelos","bio_modelos",
                            "salicifolia", 
                            "binp10_cloglog_Dsalicifolia_new_B.tif"), 
            format="GTiff",
            overwritte=TRUE)


```


## *Response plots edition*
In ggplot and mixing all the species response curves in a single plot per variable

```{r}
# BIOCLIM VARIABLES

my_palette<-c('#CC3311', '#4477AA','#DDAA33','#44AA99')

Bio_1_da<-as.data.frame(Bio_1_resp_Da)
Bio_1_da$sp<-"D. ambigens"
Bio_1_dc<-as.data.frame(Bio_1_resp_Dc)
Bio_1_dc$sp<-"D. colorata"
Bio_1_dg<-as.data.frame(Bio_1_resp_Dg)
Bio_1_dg$sp<-"D. gentlei"
Bio_1_ds<-as.data.frame(Bio_1_resp_Ds)
Bio_1_ds$sp<-"D. salicifolia"

r_Bio_1<-rbind(Bio_1_da,Bio_1_dc,Bio_1_dg,Bio_1_ds)
names(r_Bio_1)[1]<-"Bio_1"

plot_Bio_1<- ggplot(r_Bio_1, aes(x=Bio_1, y=p, color=sp)) +
    geom_line(size=1)+
    labs(colour = "Species", x="Bio 1")+
    theme_classic()+ 
    coord_cartesian(ylim=c(0,1))+
    xlim(0,40)+
    scale_y_continuous(labels = label_number(accuracy = 0.1),
                     breaks = c(0, 0.5, 1),)+
    scale_colour_manual(values=my_palette)+
    theme(legend.position = "none",
          legend.title=element_blank(),
          legend.text = element_text(face="italic", size=8),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.y = element_blank(),
          plot.margin = unit(c(0,0,0,0), "cm"),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA))

Bio_4_da<-as.data.frame(Bio_4_resp_Da)
Bio_4_da$sp<-"D. ambigens"
Bio_4_dc<-as.data.frame(Bio_4_resp_Dc)
Bio_4_dc$sp<-"D. colorata"
Bio_4_dg<-as.data.frame(Bio_4_resp_Dg)
Bio_4_dg$sp<-"D. gentlei"
Bio_4_ds<-as.data.frame(Bio_4_resp_Ds)
Bio_4_ds$sp<-"D. salicifolia"

r_Bio_4<-rbind(Bio_4_da,Bio_4_dc,Bio_4_dg,Bio_4_ds)
names(r_Bio_4)[1]<-"Bio_4"

plot_Bio_4<- ggplot(r_Bio_4, aes(x=Bio_4, y=p, color=sp)) +
    geom_line(size=1)+
    labs(colour = "Species", x="Bio 4")+
    theme_classic()+   
    coord_cartesian(ylim=c(0,1))+
    xlim(0,500)+
    scale_y_continuous(labels = label_number(accuracy = 0.1),
                     breaks = c(0, 0.5, 1),)+
    scale_colour_manual(values=my_palette)+
    theme(legend.position = "none",
          legend.title=element_blank(),
          legend.text = element_text(face="italic", size=8),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.y = element_blank(),
          plot.margin = unit(c(0,0,0,0), "cm"),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA))


Bio_12_da<-as.data.frame(Bio_12_resp_Da)
Bio_12_da$sp<-"D. ambigens"
Bio_12_dc<-as.data.frame(Bio_12_resp_Dc)
Bio_12_dc$sp<-"D. colorata"
Bio_12_dg<-as.data.frame(Bio_12_resp_Dg)
Bio_12_dg$sp<-"D. gentlei"
Bio_12_ds<-as.data.frame(Bio_12_resp_Ds)
Bio_12_ds$sp<-"D. salicifolia"

r_Bio_12<-rbind(Bio_12_da,Bio_12_dc,Bio_12_dg,Bio_12_ds)
names(r_Bio_12)[1]<-"Bio_12"

plot_Bio_12<- ggplot(r_Bio_12, aes(x=Bio_12, y=p, color=sp)) +
    geom_line(size=1)+
  labs(colour = "Species", x="Bio 12")+
    theme_classic()+   
    coord_cartesian(ylim=c(0,1))+
    scale_y_continuous(labels = label_number(accuracy = 0.1),
                     breaks = c(0, 0.5, 1),)+
    scale_colour_manual(values=my_palette)+
    xlim(0,5000)+
    theme(legend.position = "none",
          legend.title=element_blank(),
          legend.text = element_text(face="italic", size=8),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.y = element_blank(),
          plot.margin = unit(c(0,0,0,0), "cm"),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA))

Bio_14_da<-as.data.frame(Bio_14_resp_Da)
Bio_14_da$sp<-"D. ambigens"
Bio_14_dc<-as.data.frame(Bio_14_resp_Dc)
Bio_14_dc$sp<-"D. colorata"
Bio_14_dg<-as.data.frame(Bio_14_resp_Dg)
Bio_14_dg$sp<-"D. gentlei"
Bio_14_ds<-as.data.frame(Bio_14_resp_Ds)
Bio_14_ds$sp<-"D. salicifolia"

r_Bio_14<-rbind(Bio_14_da,Bio_14_dc,Bio_14_dg,Bio_14_ds)
names(r_Bio_14)[1]<-"Bio_14"

plot_Bio_14<- ggplot(r_Bio_14, aes(x=Bio_14, y=p, color=sp)) +
    geom_line(size=1)+
    labs(colour = "Species", x="Bio 14")+
    theme_classic()+    
    coord_cartesian(ylim=c(0,1))+
    xlim(0,175)+
    scale_y_continuous(labels = label_number(accuracy = 0.1),
                     breaks = c(0, 0.5, 1),)+
    scale_colour_manual(values=my_palette)+
    theme(legend.position = "none",
          legend.title=element_blank(),
          legend.text = element_text(face="italic", size=8),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.y = element_blank(),
          plot.margin = unit(c(0,0,0,0), "cm"),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA))

Bio_15_da<-as.data.frame(Bio_15_resp_Da)
Bio_15_da$sp<-"D. ambigens"
Bio_15_dc<-as.data.frame(Bio_15_resp_Dc)
Bio_15_dc$sp<-"D. colorata"
Bio_15_dg<-as.data.frame(Bio_15_resp_Dg)
Bio_15_dg$sp<-"D. gentlei"
Bio_15_ds<-as.data.frame(Bio_15_resp_Ds)
Bio_15_ds$sp<-"D. salicifolia"

r_Bio_15<-rbind(Bio_15_da,Bio_15_dc,Bio_15_dg,Bio_15_ds)
names(r_Bio_15)[1]<-"Bio_15"

plot_Bio_15<- 
  ggplot(r_Bio_15, aes(x=Bio_15, y=p, color=sp)) +
    geom_line(size=1)+
    labs(colour = "Species", x="Bio 15")+
    theme_classic()+
    coord_cartesian(ylim=c(0,1))+
    xlim(0,150)+
    scale_y_continuous(labels = label_number(accuracy = 0.1),
                     breaks = c(0, 0.5, 1),)+
    scale_colour_manual(values=my_palette)+
    theme(legend.position = "none",
          legend.title=element_blank(),
          legend.text = element_text(face="italic", size=8),
          axis.text = element_text(size = 8),
          axis.title= element_text(size= 9),
          axis.title.y = element_blank(),
          plot.margin = unit(c(0,0,0,0), "cm"),
          legend.box = "horizontal",
          panel.border = element_rect(colour = "black", fill=NA))

#===================== FIGURE LAYOUT =================================
white_palette<-c("white","white","white","white","white")

empty_plot <- plot_Bio_15 + 
  theme_void()+
  geom_blank()+
  theme(legend.position="left",
        legend.text = element_text(face="italic", size=8))+
  scale_colour_manual(values=white_palette)+
  scale_fill_manual(values=white_palette)


library(cowplot)

legend_summ <- get_legend(
  plot_Bio_15 + 
    guides(color = guide_legend(nrow = 4)) +
    theme(legend.position = "bottom")
)

pdf(here("Modelos","RESP_EVAL_figures","bioclim_resp_new_fig.pdf"),width = 3,height=9)

summary_fig <-plot_grid(plot_Bio_1,
                        plot_Bio_4,
                        plot_Bio_12,
                        plot_Bio_14,
                        plot_Bio_15,
                   legend_summ,
                   ncol = 1,
                   nrow=5,
                   axis="l",
                   rel_widths=1,
                   rel_heights=1,
                   align="hv",
                   labels=c("a","b","c","d","e"),
                   label_size=8, 
                   label_fontface="plain",
                   hjust=3,
                   vjus=0.2)

annotate_figure(summary_fig,
                left = text_grob("Predicted value", 
                                 rot = 90))
dev.off()

```


# Evaluation plots layout

```{r}

# create a colorblind friendly palette for evaluation plots

EVAL_palette <- c( '#4477AA', '#EE6677', '#228833', '#CCBB44', '#66CCEE', '#AA3377', '#BBBBBB')


# for D. ambigens


#_____________AUC.val
maxentEvalPlot_Da_AUCval_lyt<- maxentEvalPlot_Da_AUCval+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#____________AUC.diff
maxentEvalPlot_Da_AUCdiff_lyt<- maxentEvalPlot_Da_AUCdiff+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))
#_________OR.mtp
maxentEvalPlot_Da_ORmtp_lyt<- maxentEvalPlot_Da_ORmtp+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#_______OR.10p

maxentEvalPlot_Da_OR10p_lyt<- maxentEvalPlot_Da_OR10p+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#______________delta AIC

maxentEvalPlot_Da_AIC_lyt<- maxentEvalPlot_Da_AIC+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

EVAL_Da_title <- text_grob("D. ambigens", size = 9, face = "italic") 


tmp_combined_EVAL_Da_lyt<-grid.arrange(EVAL_Da_title,
                                       maxentEvalPlot_Da_AUCval_lyt,
                                       maxentEvalPlot_Da_AUCdiff_lyt,
                                       maxentEvalPlot_Da_ORmtp_lyt,
                                       maxentEvalPlot_Da_OR10p_lyt,
                                       maxentEvalPlot_Da_AIC_lyt,
                                       ncol=1, nrow=6,
                                       heights=c(0.1,1,1,1,1,1),
                                       widths=1.2)

# for D. colorata



#_____________AUC.val
maxentEvalPlot_Dc_AUCval_lyt<- maxentEvalPlot_Dc_AUCval+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#____________AUC.diff
maxentEvalPlot_Dc_AUCdiff_lyt<- maxentEvalPlot_Dc_AUCdiff+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))
#_________OR.mtp
maxentEvalPlot_Dc_ORmtp_lyt<- maxentEvalPlot_Dc_ORmtp+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#_______OR.10p

maxentEvalPlot_Dc_OR10p_lyt<- maxentEvalPlot_Dc_OR10p+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#______________delta AIC

maxentEvalPlot_Dc_AIC_lyt<- maxentEvalPlot_Dc_AIC+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

EVAL_Dc_title <- text_grob("D. colorata", size = 9, face = "italic") 


tmp_combined_EVAL_Dc_lyt<-grid.arrange(EVAL_Dc_title,
                                       maxentEvalPlot_Dc_AUCval_lyt,
                                       maxentEvalPlot_Dc_AUCdiff_lyt,
                                       maxentEvalPlot_Dc_ORmtp_lyt,
                                       maxentEvalPlot_Dc_OR10p_lyt,
                                       maxentEvalPlot_Dc_AIC_lyt,
                                       ncol=1, nrow=6,
                                       heights=c(0.1,1,1,1,1,1),
                                       widths=1.2)

# for D. gentlei



#_____________AUC.val
maxentEvalPlot_Dg_AUCval_lyt<- maxentEvalPlot_Dg_AUCval+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#____________AUC.diff
maxentEvalPlot_Dg_AUCdiff_lyt<- maxentEvalPlot_Dg_AUCdiff+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))
#_________OR.mtp
maxentEvalPlot_Dg_ORmtp_lyt<- maxentEvalPlot_Dg_ORmtp+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#_______OR.10p

maxentEvalPlot_Dg_OR10p_lyt<- maxentEvalPlot_Dg_OR10p+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#______________delta AIC

maxentEvalPlot_Dg_AIC_lyt<- maxentEvalPlot_Dg_AIC+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

EVAL_Dg_title <- text_grob("D. gentlei", size = 9, face = "italic") 


tmp_combined_EVAL_Dg_lyt<-grid.arrange(EVAL_Dg_title,
                                       maxentEvalPlot_Dg_AUCval_lyt,
                                       maxentEvalPlot_Dg_AUCdiff_lyt,
                                       maxentEvalPlot_Dg_ORmtp_lyt,
                                       maxentEvalPlot_Dg_OR10p_lyt,
                                       maxentEvalPlot_Dg_AIC_lyt,
                                       ncol=1, nrow=6,
                                       heights=c(0.1,1,1,1,1,1),
                                       widths=1.2)




# for D. salicifolia:

#_____________AUC.val
maxentEvalPlot_Ds_AUCval_lyt<- maxentEvalPlot_Ds_AUCval+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#____________AUC.diff
maxentEvalPlot_Ds_AUCdiff_lyt<- maxentEvalPlot_Ds_AUCdiff+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))
#_________OR.mtp
maxentEvalPlot_Ds_ORmtp_lyt<- maxentEvalPlot_Ds_ORmtp+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#_______OR.10p

maxentEvalPlot_Ds_OR10p_lyt<- maxentEvalPlot_Ds_OR10p+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#______________delta AIC

maxentEvalPlot_Ds_AIC_lyt<- maxentEvalPlot_Ds_AIC+
  theme(axis.title.y = element_blank(),
        plot.margin = unit(c(0,0,0,0), "cm"))

#_____combine plots per species


#Extract legend for multiple plot

get_only_legend <- function(plot) {
  
  plot_table <- ggplot_gtable(ggplot_build(plot)) # get tabular interpretation of plot
  
  legend_plot <- which(sapply(plot_table$grobs, function(x) x$name) == "guide-box")      #  Mark only legend in plot
  
  legend <- plot_table$grobs[[legend_plot]]   # extract legend
  return(legend) # return legend
}


#ADD to multiple plot


#Temporary plot to extract legend
for_legend_EVAL_Ds<-maxentEvalPlot_Ds_AIC+theme(legend.position = "bottom")

legend_EVAL_Ds <- get_only_legend(for_legend_EVAL_Ds)  

x_axis_title_EVAL_Ds <- grid.text("Regularization multiplier",
                                  just="bottom",
                                  gp=gpar(fontsize=9)
)

EVAL_Ds_title <- text_grob("D. salicifolia", size = 9, face = "italic") 

tmp_combined_EVAL_Ds_lyt<-grid.arrange(EVAL_Ds_title,
                                   maxentEvalPlot_Ds_AUCval_lyt,
                                   maxentEvalPlot_Ds_AUCdiff_lyt,
                                   maxentEvalPlot_Ds_ORmtp_lyt,
                                   maxentEvalPlot_Ds_OR10p_lyt,
                                   maxentEvalPlot_Ds_AIC_lyt,
                                   ncol=1, nrow=6,
                                   heights=c(0.1,1,1,1,1,1),
                                   widths=1.2)   #temporary combined plot

# Join the summary of allplots
EVAL_summ_lyt<-grid.arrange(tmp_combined_EVAL_Da_lyt,
                        tmp_combined_EVAL_Dc_lyt,
                        tmp_combined_EVAL_Dg_lyt,
                        tmp_combined_EVAL_Ds_lyt, 
                        nrow = 1, ncol=4,
                        heights=1,
                        widths=c(1,1,1,1))


#Save evaluation plots

pdf(here::here("Modelos","RESP_EVAL_figures",
               "EvalPlots_BIOCLIM_new_fig.pdf"))



EVAL_summ_lyt_edited <- grid.arrange(EVAL_summ_lyt,
                                 x_axis_title_EVAL_Ds,
                                 legend_EVAL_Ds,
                                 nrow=3, ncol=1,
                                 heights=c(1.8,0.07,0.05),
                                 widths=0.7)
dev.off()
```


------------------------------------------------------------------------


##### Niche overlap calculated for background points with ecospat


## ESPACE analysis for *Damburneya ambigens* and *Damburneya colorata* (Da\_Dc)

### Environmental space for background data

Performing and plotting principal component analysis to reduce
dimensionality of environmental space for *Damburneya ambigens* &
  *Damburneya colorata*. PCA done for background conditions.


```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Da_Dc <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15')


#Run the pca
espace_pca_bg_Da_Dc <- espace_pca(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya colorata", 
  occs.z1 = occs_Da[,pcaSel_bg_Da_Dc],
  occs.z2 = occs_Dc[,pcaSel_bg_Da_Dc],
  bgPts.z1 = bgEnvsVals_Da[,pcaSel_bg_Da_Dc],
  bgPts.z2 = bgEnvsVals_Dc[,pcaSel_bg_Da_Dc])


inertia_pca_bg_Da_Dc<-inertia.dudi(espace_pca_bg_Da_Dc,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
          "pcasumm_bg_Da_Dc_new_B.txt"))

"summary"
summary(espace_pca_bg_Da_Dc)

"eigenvalues"
espace_pca_bg_Da_Dc$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Da_Dc$c1

"variable coordinates"
espace_pca_bg_Da_Dc$co

"inertia"
sink()

##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
         "scatter_bg_Da_Dc_new_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Da_Dc$scores[espace_pca_bg_Da_Dc$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Da_Dc$scores[espace_pca_bg_Da_Dc$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#CC3311', '#4477AA'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Da_Dc<-paste("PC1= ",round(inertia_pca_bg_Da_Dc$tot.inertia[1,3],2),"%")

pca2cont_bg_Da_Dc<-paste("PC2= ", round(inertia_pca_bg_Da_Dc$tot.inertia[2,3]-inertia_pca_bg_Da_Dc$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Da_Dc<-paste(pca1cont_bg_Da_Dc,", ",pca2cont_bg_Da_Dc)

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
         "corcircle_Da_Dc_new_B.pdf"),width=14.5, height=10)   

ade4::s.corcircle(espace_pca_bg_Da_Dc$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Da_Dc, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Da_Dc, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
         "screeplot_Da_Dc_new_B.pdf"))

screeplot(espace_pca_bg_Da_Dc, main = NULL,col="gray")

dev.off()

```

### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of background environmental conditions
for *Damburneya ambigens* & *Damburneya colorata*
```{r}
# Create density grid
espace_occDens_bg_Da_Dc <- espace_occDens(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya colorata", 
  pca = espace_pca_bg_Da_Dc) 

#plots

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc","dens_grid_Da_Dc_new_B.pdf"), height=4, width =7.25)

par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Dc[["Damburneya ambigens"]], title = "Damburneya ambigens")
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Dc[["Damburneya colorata"]], title = "Damburneya colorata")

dev.off()
```



### Environmental space

Evaluating niche overlap between *Damburneya ambigens* & *Damburneya
colorata* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE

```{r}
##Run tests 
espace_nicheOv_bg_Da_Dc <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Dc[["Damburneya ambigens"]],
  z2 = espace_occDens_bg_Da_Dc[["Damburneya colorata"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Dc_Da <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Dc[["Damburneya colorata"]],
  z2 = espace_occDens_bg_Da_Dc[["Damburneya ambigens"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 



# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
         "overlap_Da_Dc_new_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Da_Dc[["Damburneya ambigens"]],
  espace_occDens_bg_Da_Dc[["Damburneya colorata"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya ambigens vs Damburneya colorata",
  colz1 = '#CC3311', #red
  colz2 = '#4477AA', #blue
  colinter = 'grey18', #grey
  colZ1 = '#CC3311', #red
  colZ2 = '#4477AA' #blue
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
         "similarity_Da_Dc_new_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Dc $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Dc$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Da_Dc",
          "Noverlap_Da_Dc_new_B.txt"))
"Niche overlap, equivalency and similarity for D. ambigens and D. colorata"

espace_nicheOv_bg_Da_Dc

"/n Niche overlap, equivalency and similarity for D. colorata and D. ambigens"    
espace_nicheOv_bg_Dc_Da

sink()


```

------------------------------------------------------------------------
  
  ## ESPACE analysis for *Damburneya ambigens* and *Damburneya gentlei* (Da\_Dg)
  
  ### Environmental space
  
  Performing and plotting principal component analysis to reduce
dimensionality of environmental space for *Damburneya ambigens* &
  *Damburneya gentlei*. PCA done for background environmental conditions.


```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Da_Dg <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15')



#Run the pca
espace_pca_bg_Da_Dg <- espace_pca(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya gentlei", 
  occs.z1 = occs_Da[,pcaSel_bg_Da_Dg],
  occs.z2 = occs_Dg[,pcaSel_bg_Da_Dg],
  bgPts.z1 = bgEnvsVals_Da[,pcaSel_bg_Da_Dg],
  bgPts.z2 = bgEnvsVals_Dg[,pcaSel_bg_Da_Dg])


inertia_pca_bg_Da_Dg<-inertia.dudi(espace_pca_bg_Da_Dg,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
          "pcasumm_bg_Da_Dg_B.txt"))

"summary"
summary(espace_pca_bg_Da_Dg)

"eigenvalues"
espace_pca_bg_Da_Dg$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Da_Dg$c1

"variable coordinates"
espace_pca_bg_Da_Dg$co

"inertia"
inertia_pca_bg_Da_Dg

sink()



##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
         "scatter_bg_Da_Dg_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Da_Dg$scores[espace_pca_bg_Da_Dg$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Da_Dg$scores[espace_pca_bg_Da_Dg$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#CC3311', '#DDAA33'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Da_Dg<-paste("PC1= ",round(inertia_pca_bg_Da_Dg$tot.inertia[1,3],2),"%")
pca2cont_bg_Da_Dg<-paste("PC2= ", round(inertia_pca_bg_Da_Dg$tot.inertia[2,3]-inertia_pca_bg_Da_Dg$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Da_Dg<-paste(pca1cont_bg_Da_Dg,", ",pca2cont_bg_Da_Dg)

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
         "corcircle_Da_Dg_B.pdf"),width=14.5, height=10)   

ade4::s.corcircle(espace_pca_bg_Da_Dg$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Da_Dg, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Da_Dg, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
         "screeplot_Da_Dg_B.pdf"))

screeplot(espace_pca_bg_Da_Dg, main = NULL,col="gray")

dev.off()

```


### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of background environmental conditions 
for *Damburneya ambigens* & *Damburneya gentlei*
  
```{r}
# Create density grid
espace_occDens_bg_Da_Dg <- espace_occDens(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya gentlei", 
  pca = espace_pca_bg_Da_Dg) 

#plots
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg","dens_grid_Da_Dg_B.pdf"), height=4, width =7.25)

par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Dg[["Damburneya ambigens"]], title = "Damburneya ambigens")
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Dg[["Damburneya gentlei"]], title = "Damburneya gentlei")

dev.off()
```

### Environmental space

Evaluating niche overlap between *Damburneya ambigens* & *Damburneya
gentlei* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE

```{r}
##Run tests 
espace_nicheOv_bg_Da_Dg <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Dg[["Damburneya ambigens"]],
  z2 = espace_occDens_bg_Da_Dg[["Damburneya gentlei"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Dg_Da <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Dg[["Damburneya gentlei"]],
  z2 = espace_occDens_bg_Da_Dg[["Damburneya ambigens"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 


# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
         "overlap_Da_Dg_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Da_Dg[["Damburneya ambigens"]],
  espace_occDens_bg_Da_Dg[["Damburneya gentlei"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya ambigens vs Damburneya gentlei",
  colz1 = '#CC3311', #red
  colz2 = '#DDAA33', #yellow
  colinter = 'grey18', #grey
  colZ1 = '#CC3311', #red
  colZ2 = '#DDAA33' #yellow
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
         "similarity_Da_Dg_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Dg $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Dg$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Da_Dg",
          "Noverlap_Da_Dg_B.txt"))
"Niche overlap, equivalency and similarity for D. ambigens and D. gentlei"

espace_nicheOv_bg_Da_Dg

"/n Niche overlap, equivalency and similarity for D. gentlei and D. ambigens"    
espace_nicheOv_bg_Dg_Da

sink()


```

------------------------------------------------------------------------
  
  ## ESPACE analysis for *Damburneya ambigens* and *Damburneya salicifolia* (Da\_Ds)
  
  
  ### Environmental space
  
  Performing and plotting principal component analysis to reduce
dimensionality of environmental space for *Damburneya ambigens* &
  *Damburneya salicifolia*. PCA done for background environmental conditions.

```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Da_Ds <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15' )



#Run the pca
espace_pca_bg_Da_Ds <- espace_pca(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya salicifolia", 
  occs.z1 = occs_Da[,pcaSel_bg_Da_Ds],
  occs.z2 = occs_Ds[,pcaSel_bg_Da_Ds],
  bgPts.z1 = bgEnvsVals_Da[,pcaSel_bg_Da_Ds],
  bgPts.z2 = bgEnvsVals_Ds[,pcaSel_bg_Da_Ds])


inertia_pca_bg_Da_Ds<-inertia.dudi(espace_pca_bg_Da_Ds,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
          "pcasumm_bg_Da_Ds_B.txt"))

"summary"
summary(espace_pca_bg_Da_Ds)

"eigenvalues"
espace_pca_bg_Da_Ds$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Da_Ds$c1

"variable coordinates"
espace_pca_bg_Da_Ds$co

"inertia"


sink()



##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
         "scatter_bg_Da_Ds_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Da_Ds$scores[espace_pca_bg_Da_Ds$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Da_Ds$scores[espace_pca_bg_Da_Ds$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#CC3311', '#44AA99'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Da_Ds<-paste("PC1= ",round(inertia_pca_bg_Da_Ds$tot.inertia[1,3],2),"%")
pca2cont_bg_Da_Ds<-paste("PC2= ", round(inertia_pca_bg_Da_Ds$tot.inertia[2,3]-inertia_pca_bg_Da_Ds$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Da_Ds<-paste(pca1cont_bg_Da_Ds,", ",pca2cont_bg_Da_Ds)

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
         "corcircle_Da_Ds_B.pdf"),width=14.5, height=10)   

ade4::s.corcircle(espace_pca_bg_Da_Ds$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Da_Ds, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Da_Ds, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
         "screeplot_Da_Ds_B.pdf"))

screeplot(espace_pca_bg_Da_Ds, main = NULL,col="gray")

dev.off()

```




### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of environmental conditions in the background
for *Damburneya ambigens* & *Damburneya salicifolia*
  
```{r}
# Create density grid
espace_occDens_bg_Da_Ds <- espace_occDens(
  sp.name1 = "Damburneya ambigens",
  sp.name2 = "Damburneya salicifolia", 
  pca = espace_pca_bg_Da_Ds) 

#plots
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds","dens_grid_Da_Ds_B.pdf"), height=4, width =7.25)

par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Ds[["Damburneya ambigens"]], title = "Damburneya ambigens")
ecospat::ecospat.plot.niche(espace_occDens_bg_Da_Ds[["Damburneya salicifolia"]], title = "Damburneya salicifolia")
dev.off()
```

### Environmental space

Evaluating niche overlap between *Damburneya ambigens* & *Damburneya salicifolia* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE

```{r}
##Run tests 
espace_nicheOv_bg_Da_Ds <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Ds[["Damburneya ambigens"]],
  z2 = espace_occDens_bg_Da_Ds[["Damburneya salicifolia"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Ds_Da <- espace_nicheOv(
  z1 = espace_occDens_bg_Da_Ds[["Damburneya salicifolia"]],
  z2 = espace_occDens_bg_Da_Ds[["Damburneya ambigens"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 



# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
         "overlap_Da_Ds_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Da_Ds[["Damburneya ambigens"]],
  espace_occDens_bg_Da_Ds[["Damburneya salicifolia"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya ambigens vs Damburneya salicifolia",
  colz1 = '#CC3311', #red
  colz2 = '#44AA99', #green
  colinter = 'grey18', #grey
  colZ1 = '#CC3311', #red
  colZ2 = '#44AA99' #green
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
         "similarity_Da_Ds_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Ds $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Da_Ds$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Da_Ds",
          "Noverlap_Da_Ds_B.txt"))
"Niche overlap, equivalency and similarity for D. ambigens and D. salicifolia"

espace_nicheOv_bg_Da_Ds

"/n Niche overlap, equivalency and similarity for D. salicifolia and D. ambigens"    
espace_nicheOv_bg_Ds_Da

sink()


```
------------------------------------------------------------------------
  
  ## ESPACE analysis for *Damburneya colorata* and *Damburneya gentlei* (Dc\_Dg)
  
  ### Environmental space
  
  Calculating the part of environmental space more densly populated by
species & the availability of environmental conditions in the background
for *Damburneya colorata* & *Damburneya gentlei*
  
```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Dc_Dg <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15' )


#Run the pca
espace_pca_bg_Dc_Dg <- espace_pca(
  sp.name1 = "Damburneya colorata",
  sp.name2 = "Damburneya gentlei", 
  occs.z1 = occs_Dc[,pcaSel_bg_Dc_Dg],
  occs.z2 = occs_Dg[,pcaSel_bg_Dc_Dg],
  bgPts.z1 = bgEnvsVals_Dc[,pcaSel_bg_Dc_Dg],
  bgPts.z2 = bgEnvsVals_Dg[,pcaSel_bg_Dc_Dg])


inertia_pca_bg_Dc_Dg<-inertia.dudi(espace_pca_bg_Dc_Dg,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
          "pcasumm_bg_Dc_Dg_new_B.txt"))

"summary"
summary(espace_pca_bg_Dc_Dg)

"eigenvalues"
espace_pca_bg_Dc_Dg$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Dc_Dg$c1

"variable coordinates"
espace_pca_bg_Dc_Dg$co

"inertia"
inertia_pca_bg_Dc_Dg

sink()



##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
         "scatter_bg_Dc_Dg_new_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Dc_Dg$scores[espace_pca_bg_Dc_Dg$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Dc_Dg$scores[espace_pca_bg_Dc_Dg$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#4477AA', '#DDAA33'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Dc_Dg<-paste("PC1= ",round(inertia_pca_bg_Dc_Dg$tot.inertia[1,3],2),"%")
pca2cont_bg_Dc_Dg<-paste("PC2= ", round(inertia_pca_bg_Dc_Dg$tot.inertia[2,3]-inertia_pca_bg_Dc_Dg$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Dc_Dg<-paste(pca1cont_bg_Dc_Dg,", ",pca2cont_bg_Dc_Dg)

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
         "corcircle_Dc_Dg_new_B.pdf"),width=14.5, height=10)   

ade4::s.corcircle(espace_pca_bg_Dc_Dg$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Dc_Dg, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Dc_Dg, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
         "screeplot_Dc_Dg_new_B.pdf"))

screeplot(espace_pca_bg_Dc_Dg, main = NULL,col="gray")

dev.off()

```

### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of environmental conditions in the background
for *Damburneya colorata* & *Damburneya gentlei*
  
```{r}
# Create density grid
espace_occDens_bg_Dc_Dg <- espace_occDens(
  sp.name1 = "Damburneya colorata",
  sp.name2 = "Damburneya gentlei", 
  pca = espace_pca_bg_Dc_Dg) 
#plots
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg","dens_grid_Dc_Dg_new_B.pdf"), height=4, width =7.25)

par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Dc_Dg[["Damburneya colorata"]], title = "Damburneya colorata")
ecospat::ecospat.plot.niche(espace_occDens_bg_Dc_Dg[["Damburneya gentlei"]], title = "Damburneya gentlei")
dev.off()
```
### Environmental space

Evaluating niche overlap between *Damburneya colorata* & *Damburneya gentlei* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE
```{r}
##Run tests 
espace_nicheOv_bg_Dc_Dg <- espace_nicheOv(
  z1 = espace_occDens_bg_Dc_Dg[["Damburneya colorata"]],
  z2 = espace_occDens_bg_Dc_Dg[["Damburneya gentlei"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Dg_Dc <- espace_nicheOv(
  z1 = espace_occDens_bg_Dc_Dg[["Damburneya gentlei"]],
  z2 = espace_occDens_bg_Dc_Dg[["Damburneya colorata"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 



# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
         "overlap_Dc_Dg_new_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Dc_Dg[["Damburneya colorata"]],
  espace_occDens_bg_Dc_Dg[["Damburneya gentlei"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya colorata vs Damburneya gentlei",
  colz1 = '#4477AA', #blue
  colz2 = '#DDAA33', #yellow
  colinter = 'grey18', #grey
  colZ1 = '#4477AA', #blue
  colZ2 = '#DDAA33' #yellow
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
         "similarity_Dc_Dg_new_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dc_Dg $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dc_Dg$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Dc_Dg",
          "Noverlap_Dc_Dg_new_B.txt"))
"Niche overlap, equivalency and similarity for D. colorata and D. gentlei"

espace_nicheOv_bg_Dc_Dg

"/n Niche overlap, equivalency and similarity for D. gentlei and D. colorata"    
espace_nicheOv_bg_Dg_Dc

sink()


```
------------------------------------------------------------------------
  
  ## ESPACE analysis for *Damburneya colorata* and *Damburneya salicifolia* (Dc\_Ds)
  
  ### Environmental space
  
  Performing and plotting principal component analysis to reduce
dimensionality of environmental space for *Damburneya colorata* &
  *Damburneya salicifolia*. PCA done for background environmental conditions.

```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Dc_Ds <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15' )



#Run the pca
espace_pca_bg_Dc_Ds <- espace_pca(
  sp.name1 = "Damburneya colorata",
  sp.name2 = "Damburneya salicifolia", 
  occs.z1 = occs_Dc[,pcaSel_bg_Dc_Ds],
  occs.z2 = occs_Ds[,pcaSel_bg_Dc_Ds],
  bgPts.z1 = bgEnvsVals_Dc[,pcaSel_bg_Dc_Ds],
  bgPts.z2 = bgEnvsVals_Ds[,pcaSel_bg_Dc_Ds])


inertia_pca_bg_Dc_Ds<-inertia.dudi(espace_pca_bg_Dc_Ds,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
          "pcasumm_bg_Dc_Ds_new_B.txt"))

"summary"
summary(espace_pca_bg_Dc_Ds)

"eigenvalues"
espace_pca_bg_Dc_Ds$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Dc_Ds$c1

"variable coordinates"
espace_pca_bg_Dc_Ds$co

"inertia"
inertia_pca_bg_Dc_Ds

sink()



##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
         "scatter_bg_Dc_Ds_new_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Dc_Ds$scores[espace_pca_bg_Dc_Ds$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Dc_Ds$scores[espace_pca_bg_Dc_Ds$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#4477AA', '#44AA99'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Dc_Ds<-paste("PC1= ",round(inertia_pca_bg_Dc_Ds$tot.inertia[1,3],2),"%")
pca2cont_bg_Dc_Ds<-paste("PC2= ", round(inertia_pca_bg_Dc_Ds$tot.inertia[2,3]-inertia_pca_bg_Dc_Ds$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Dc_Ds<-paste(pca1cont_bg_Dc_Ds,", ",pca2cont_bg_Dc_Ds)

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
         "corcircle_Dc_Ds_new_B.pdf"),width=14, height=10)   

ade4::s.corcircle(espace_pca_bg_Dc_Ds$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Dc_Ds, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Dc_Ds, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
         "screeplot_Dc_Ds_new_B.pdf"))

screeplot(espace_pca_bg_Dc_Ds, main = NULL,col="gray")

dev.off()

```



### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of environmental conditions in the background
for *Damburneya colorata* & *Damburneya salicifolia*
  
```{r}
# Create density grid
espace_occDens_bg_Dc_Ds <- espace_occDens(
  sp.name1 = "Damburneya colorata",
  sp.name2 = "Damburneya salicifolia", 
  pca = espace_pca_bg_Dc_Ds) 
#plots
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds","dens_grid_Dc_Ds_new_B.pdf"), height=4, width =7.25)
par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Dc_Ds[["Damburneya colorata"]], title = "Damburneya colorata")
ecospat::ecospat.plot.niche(espace_occDens_bg_Dc_Ds[["Damburneya salicifolia"]], title = "Damburneya salicifolia")

dev.off()
```

### Environmental space

Evaluating niche overlap between *Damburneya colorata* & *Damburneya
salicifolia* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE

```{r}
##Run tests 
espace_nicheOv_bg_Dc_Ds <- espace_nicheOv(
  z1 = espace_occDens_bg_Dc_Ds[["Damburneya colorata"]],
  z2 = espace_occDens_bg_Dc_Ds[["Damburneya salicifolia"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Ds_Dc <- espace_nicheOv(
  z1 = espace_occDens_bg_Dc_Ds[["Damburneya salicifolia"]],
  z2 = espace_occDens_bg_Dc_Ds[["Damburneya colorata"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 



# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
         "overlap_Dc_Ds_new_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Dc_Ds[["Damburneya colorata"]],
  espace_occDens_bg_Dc_Ds[["Damburneya salicifolia"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya colorata vs Damburneya salicifolia",
  colz1 = '#4477AA', #blue
  colz2 = '#44AA99', #green
  colinter = 'grey18', #grey
  colZ1 = '#4477AA', #blue
  colZ2 = '#44AA99' #green
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
         "similarity_Dc_Ds_new_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dc_Ds $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dc_Ds$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Dc_Ds",
          "Noverlap_Dc_Ds_new_B.txt"))
"Niche overlap, equivalency and similarity for D. colorata and D. salicifolia"

espace_nicheOv_bg_Dc_Ds

"/n Niche overlap, equivalency and similarity for D. salicifolia and D. colorata"    
espace_nicheOv_bg_Ds_Dc

sink()


```



------------------------------------------------------------------------
  
  ## ESPACE analysis for *Damburneya gentlei* and *Damburneya salicifolia* (Dg\_Ds)
  
  ### Environmental space
  
  Performing and plotting principal component analysis to reduce
dimensionality of environmental space for *Damburneya gentlei* &
  *Damburneya salicifolia*. PCA done for background environmental conditions.

```{r}
####Extract environmental information

#Determine the variables to use
pcaSel_bg_Dg_Ds <- c('Bio_1' , 'Bio_4', 'Bio_12',  'Bio_14', 'Bio_15' )


#Run the pca
espace_pca_bg_Dg_Ds <- espace_pca(
  sp.name1 = "Damburneya gentlei",
  sp.name2 = "Damburneya salicifolia", 
  occs.z1 = occs_Dg[,pcaSel_bg_Dg_Ds],
  occs.z2 = occs_Ds[,pcaSel_bg_Dg_Ds],
  bgPts.z1 = bgEnvsVals_Dg[,pcaSel_bg_Dg_Ds],
  bgPts.z2 = bgEnvsVals_Ds[,pcaSel_bg_Dg_Ds])


inertia_pca_bg_Dg_Ds<-inertia.dudi(espace_pca_bg_Dg_Ds,col.inertia=TRUE)

#Print PCA summary of results
sink(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
          "pcasumm_bg_Dg_Ds_B.txt"))

"summary"
summary(espace_pca_bg_Dg_Ds)

"eigenvalues"
espace_pca_bg_Dg_Ds$eig

"the column normed scores i.e. the principal axes"
espace_pca_bg_Dg_Ds$c1

"variable coordinates"
espace_pca_bg_Dg_Ds$co

"inertia"


sink()



##generate plots

pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
         "scatter_bg_Dg_Ds_B.pdf"))

# PCA Scatter Plot
if ("occsBg" == "occs") {
  x <- espace_pca_bg_Dg_Ds$scores[espace_pca_bg_Dg_Ds$scores$bg == 'sp', ]
  x.f <- factor(x$sp)
} else if ("occsBg" == "occsBg") {
  x <- espace_pca_bg_Dg_Ds$scores[espace_pca_bg_Dg_Ds$scores$sp == 'bg', ]
  x.f <- factor(x$bg)
}
ade4::s.class(x, x.f, xax = 1, yax = 2,
              col = c('#DDAA33', '#44AA99'), cstar = 0, cpoint = 0.1)

dev.off()


# PCA Correlation circle

#axis explained variance (inertia)
pca1cont_bg_Dg_Ds<-paste("PC1= ",round(inertia_pca_bg_Dg_Ds$tot.inertia[1,3],2),"%")
pca2cont_bg_Dg_Ds<-paste("PC2= ", round(inertia_pca_bg_Dg_Ds$tot.inertia[2,3]-inertia_pca_bg_Dg_Ds$tot.inertia[1,3],2),
                      "%")

axis_cont_bg_Dg_Ds<-paste(pca1cont_bg_Dg_Ds,", ",pca2cont_bg_Dg_Ds)

pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
         "corcircle_Dg_Ds_B.pdf"),width=14.5, height=10)   

ade4::s.corcircle(espace_pca_bg_Dg_Ds$co, xax = 1, yax = 2,grid=TRUE,
                  lab = pcaSel_bg_Dg_Ds, full = TRUE, box = TRUE,
                  clabel=1.5)
mtext(axis_cont_bg_Dg_Ds, side=3, cex=1.5, adj=0.9)
dev.off()

# PCA screeplot
pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
         "screeplot_Dg_Ds_B.pdf"))

screeplot(espace_pca_bg_Dg_Ds, main = NULL,col="gray")

dev.off()

```

### Environmental space

Calculating the part of environmental space more densly populated by
species & the availability of environmental conditions in the background
for *Damburneya gentlei* & *Damburneya salicifolia*
  
```{r}
# Create density grid
espace_occDens_bg_Dg_Ds <- espace_occDens(
  sp.name1 = "Damburneya gentlei",
  sp.name2 = "Damburneya salicifolia", 
  pca = espace_pca_bg_Dg_Ds) 
#plots
pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds","dens_grid_Dg_Ds_B.pdf"), height=4, width =7.25)
par(mfrow = c(1,2))
ecospat::ecospat.plot.niche(espace_occDens_bg_Dg_Ds[["Damburneya gentlei"]], title = "Damburneya gentlei")
ecospat::ecospat.plot.niche(espace_occDens_bg_Dg_Ds[["Damburneya salicifolia"]], title = "Damburneya salicifolia")
dev.off()
```

### Environmental space

Evaluating niche overlap between *Damburneya gentlei* & *Damburneya
salicifolia* for which the occurrence density grid was computed based on background environmental conditions. Running
equivalence test (FALSE) and similarity test TRUE


```{r}
##Run tests 
espace_nicheOv_bg_Dg_Ds <- espace_nicheOv(
  z1 = espace_occDens_bg_Dg_Ds[["Damburneya gentlei"]],
  z2 = espace_occDens_bg_Dg_Ds[["Damburneya salicifolia"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 

espace_nicheOv_bg_Ds_Dg <- espace_nicheOv(
  z1 = espace_occDens_bg_Dg_Ds[["Damburneya salicifolia"]],
  z2 = espace_occDens_bg_Dg_Ds[["Damburneya gentlei"]],
  iter = 1000,
  equivalency = FALSE,
  similarity = TRUE) 



# plots
layout(matrix(c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 3), 4, 3, byrow = F))

# layout.show(3)
#Plot niche overlap

pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
         "overlap_Dg_Ds_B.pdf"))   

ecospat::ecospat.plot.niche.dyn(
  espace_occDens_bg_Dg_Ds[["Damburneya gentlei"]],
  espace_occDens_bg_Dg_Ds[["Damburneya salicifolia"]],
  0.5,
  name.axis1="PC1", name.axis2="PC2",
  title = "Damburneya gentlei vs Damburneya salicifolia",
  colz1 = '#DDAA33', #yellow
  colz2 = '#44AA99', #green
  colinter = 'grey18', #grey
  colZ1 = '#DDAA33', #yellow
  colZ2 = '#44AA99' #green
)
dev.off()

#plot niche similarity
pdf(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
         "similarity_Dg_Ds_B.pdf"))   
if (FALSE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dg_Ds $equiv,
                                     "D", "Equivalency test")
if (TRUE==TRUE)
  ecospat::ecospat.plot.overlap.test(espace_nicheOv_bg_Dg_Ds$simil,
                                     "D", "Similarity test")
dev.off()

#### write results

sink(here("Modelos","bio_modelos","comparados_bg","Dg_Ds",
          "Noverlap_Dg_Ds_B.txt"))
"Niche overlap, equivalency and similarity for D. gentlei and D. salicifolia"

espace_nicheOv_bg_Dg_Ds

"/n Niche overlap, equivalency and similarity for D. salicifolia and D. gentlei"    
espace_nicheOv_bg_Ds_Dg

sink()
```



### Testing nicheoverlap and size 

Niche overlap and size are calculated and visualized with NicheROVER. A dataframe is necessary, with
columns = environmental variables+ 1 (species or data source name)


# Niche overlap for background
Using only the 5 most important variables and background data


```{r}
# 2-d projections of 10 niche regions

library(nicheROVER)

nsamples <- 10

#For specified route:
#temp<-read_B.csv(here("Modelos","bio_modelos","overlap_nROVER","data_env_spp_B.csv"))
#tmp_bg_spp <-temp[2:8]


#tmp_vec<-c("bg_Damburneya ambigens","bg_Damburneya colorata","bg_Damburneya gentlei","bg_Damburneya salicifolia")

#tmp_bg_spp <- tmp_bg_spp[tmp_bg_spp$scientific_name %in% tmp_vec,]

#Rename species
#tmp_bg_spp$scientific_name<-sub("bg_Damburneya ", "D.",         tmp_bg_spp$scientific_name)

#levels(as.factor(tmp_bg_spp$scientific_name))


# Build a reduced dataset with only 1000 random bg points
# per species



#create a temporary dataframe of environmental data for species occurrences 

tmp_Da <- occs_Da[,c(2,15:19)]
tmp_Dc <- occs_Dc[,c(2,15:19)]
tmp_Dg <- occs_Dg[,c(2,15:19)]
tmp_Ds <- occs_Ds[,c(2,15:19)]

tmp_spp_occs <- rbind(tmp_Da, tmp_Dc, tmp_Dg, tmp_Ds)


#create a temporary dataframe of environmental data for each species' M background points 

tmp_bg_Da <-bgEnvsVals_Da[,c(1,12:16)]
tmp_bg_Dc <-bgEnvsVals_Dc[,c(1,12:16)]
tmp_bg_Dg <-bgEnvsVals_Dg[,c(1,12:16)]
tmp_bg_Ds <-bgEnvsVals_Ds[,c(1,12:16)]

tmp_bg_spp <- rbind(tmp_bg_Da, tmp_bg_Dc, tmp_bg_Dg, tmp_bg_Ds)


# Sample rows of data with Base R
small_bg_Da <- tmp_bg_Da[sample(1:nrow(tmp_bg_Da), 1000), ]  
small_bg_Dc <- tmp_bg_Dc[sample(1:nrow(tmp_bg_Dc), 1000), ]  
small_bg_Dg <- tmp_bg_Dg[sample(1:nrow(tmp_bg_Dg), 1000), ]  
small_bg_Ds <- tmp_bg_Ds[sample(1:nrow(tmp_bg_Ds), 1000), ] 

small_bg_spp <- 
 rbind(small_bg_Da, small_bg_Dc, small_bg_Dg, small_bg_Ds)

small_bg_spp$scientific_name <- as.factor(small_bg_spp$scientific_name)

#Rename species
small_bg_spp$scientific_name<-sub("bg_Damburneya ", "D.",         small_bg_spp$scientific_name)



#Change factor order to improve visualiation and avoid graphic overlap

small_bg_spp$scientific_name<-factor(small_bg_spp$scientific_name,                                     levels=c("D.salicifolia", 
                                         "D.ambigens", 
                                         "D.gentlei",
                                         "D.colorata"))

#Select only the 5 most important variables and scientific names

small_bg_spp<-small_bg_spp[,1:6]
head(small_bg_spp)

#Plot only the most important variables chosen
clrs <- c('#44AA99','#CC3311', '#DDAA33' ,'#4477AA' ) # colors for each species

data.par <- tapply(1:nrow(small_bg_spp), small_bg_spp$scientific_name,
                  function(ii) niw.post(nsamples = 10, X = small_bg_spp[ii,2:6]))

# format data for plotting function
ov.data <- tapply(1:nrow(small_bg_spp), small_bg_spp$scientific_name, function(ii) X = small_bg_spp[ii,2:6])



pdf(here("Modelos","bio_modelos",
               "overlap_nROVER","bg_spp_ovlp_VARs_B.pdf"))
niche.plot(niche.par = data.par, niche.data = ov.data, pfrac = .05,
            col = clrs, ndens=10)
dev.off()

```