[Bug]: problem producing maxent models and their predictions

[mguerreiro24]

Describe the bug

I am getting this error in one of my species (code given bellow):
Error in data.frame(th = rownames, val = ths, fpa = fpa, or = or_train, :
arguments imply differing number of rows: 5, 3
3.
stop(gettextf("arguments imply differing number of rows: %s",
paste(unique(nrows), collapse = ", ")), domain = NA)
2.
data.frame(th = rownames, val = ths, fpa = fpa, or = or_train,
stringsAsFactors = FALSE)
1.
thresholds(model, type = "cloglog", test = test)

And when I calculate predictions rasters from these models, there are only 0's and one value of presence (e.g.: 0,6 or 0,64).

Steps to reproduce the bug

library(SDMtune)
Sys.setenv(JAVA_HOME = "C:/Program Files/Java/jre-1.8/")
setwd('D:/PostDoc/')

species <- 'PheronemaCarpenteri'
#species <- 'DendrophylliaCornigera'
presence_locations <- read.csv(paste(species,'/EnvironmentalData.csv',sep=''))

current <- sqrt(presence_locations$horizontalXCurrents^2+presence_locations$horizontalYCurrents^2)
presence_locations <- cbind(presence_locations,current)
presence_locations <- subset(presence_locations,current<1)
presence_locations <- presence_locations[,c(1:7,9,10,13)]#selecting variables
variables <- variable.names(presence_locations)[3:10]
  
background <- read.csv('backgroundEnvironment.csv')
background <- na.omit(background)
current <- sqrt(background$horizontalXCurrents^2+background$horizontalYCurrents^2)
background <- cbind(background,current)
background <- subset(background,current<1)
background <- background[,c(1:7,9,10,13)]#selecting variables

proportion <- dim(presence_locations)[1]*10



lengthOfVariables <- length(variables)
# "ANN", "BRT", "Maxent", "Maxnet" or "RF"

models <- c('maxent',
            #'BRT',
            'RF',
            'NN'
            )
runs <- 1:5
modelsList <- list()
evaluationList <- list()
TSSs <- c()
AUCs <- c()
threshs <- c()
responseCurves <- list()

for (i in 1:(lengthOfVariables-1)){
for (ii in (i+1):lengthOfVariables){
for (modelName in models){
for (run in runs){
  background <- background[sample(nrow(background),proportion),]
  data_species <- rbind(presence_locations,background)
  df_coords<-data_species[,1:2]#coordinates: presence & absence
  df_data<-data_species[,c(i+2,ii+2)]#environmental data: presence & absence
  pa<- c(rep(1,dim(presence_locations)[1]),rep(0,dim(background)[1]))#vectors of 1 and 0: presence & absence
  data <- SWD(species = species,
              coords = df_coords,
              data = df_data,
              pa = pa)
  # Split presence locations in training (75%) and testing (25%) datasets
  datasets <- trainValTest(data, test = 0.25, only_presence = TRUE, seed = run)
  train <- datasets[[1]]
  test <- datasets[[2]]
  if (modelName=='maxent'){model <- train(method = "Maxent", data=train)}
  if (modelName=='RF'){model <- train(method = "ANN", data=train, size = 99, maxit = 10000)}
  if (modelName=='NN'){model <- train(method = "RF", data=train)}
  lModelList <- length(modelsList)
  modelsList[[lModelList+1]] <- model
  AUCs <- c(AUCs, auc(model,test))
  TSSs <- c(TSSs, tss(model,test))
  if (modelName=='maxent'){
    modelT <-thresholds(model,type = "cloglog",test=test)#<-------------------------------------------Error here
    threshs <- c(threshs,modelT[2,3])
  } else{
    modelT <- thresholds(model,test=test)
    threshs <- c(threshs,modelT[2,5])
  }
  evaluationList[[lModelList+1]] <- varImp(model,permut = 25)
  temp <-list()
  if (modelName=='maxent'){
    my <- plotResponse(model,
                 var = variables[i],
                 type = "cloglog")
    temp[[variables[i]]] <- my@data
    my <- plotResponse(model,
                       var = variables[ii],
                       type = "cloglog")
    temp[[variables[ii]]] <- my@data
    responseCurves[[lModelList+1]] <- temp
    }else {
    my <- plotResponse(model,var = variables[i])
    temp[[variables[i]]] <- my@data
    my <- plotResponse(model,var = variables[ii])
    temp[[variables[ii]]] <- my@data
    responseCurves[[lModelList+1]] <- temp
  }
}}}}

Session information

R version 4.5.1 (2025-06-13 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 10 x64 (build 19045)

Matrix products: default
  LAPACK version 3.12.1

Random number generation:
 RNG:     Mersenne-Twister 
 Normal:  Inversion 
 Sample:  Rounding 
 
locale:
[1] LC_COLLATE=Portuguese_Portugal.utf8  LC_CTYPE=Portuguese_Portugal.utf8   
[3] LC_MONETARY=Portuguese_Portugal.utf8 LC_NUMERIC=C                        
[5] LC_TIME=Portuguese_Portugal.utf8    

time zone: Europe/Berlin
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] SDMtune_1.3.3

loaded via a namespace (and not attached):
 [1] terra_1.8-70       vctrs_0.6.5        cli_3.6.5          rlang_1.1.6       
 [5] stringi_1.8.7      generics_0.1.4     S7_0.2.0           rJava_1.0-11      
 [9] glue_1.8.0         sp_2.2-0           scales_1.4.0       grid_4.5.1        
[13] tibble_3.3.0       lifecycle_1.0.4    stringr_1.5.2      compiler_4.5.1    
[17] dplyr_1.1.4        codetools_0.2-20   RColorBrewer_1.1-3 Rcpp_1.1.0        
[21] pkgconfig_2.0.3    rstudioapi_0.17.1  farver_2.1.2       lattice_0.22-7    
[25] R6_2.6.1           tidyselect_1.2.1   pillar_1.11.1      magrittr_2.0.4    
[29] tools_4.5.1        gtable_0.3.6       dismo_1.3-16       raster_3.6-32     
[33] ggplot2_4.0.0

Additional information

I am running a small sample (n=11; train=8; test=3) species distribution model ensemble.
I believe that the low number may be one of the causes.

Reproducible example

• I have done my best to provide the steps to reproduce the bug