diff --git a/ecoinformatica_2014_2015.Rproj b/ecoinformatica_2014_2015.Rproj new file mode 100644 index 0000000..8e3c2eb --- /dev/null +++ b/ecoinformatica_2014_2015.Rproj @@ -0,0 +1,13 @@ +Version: 1.0 + +RestoreWorkspace: Default +SaveWorkspace: Default +AlwaysSaveHistory: Default + +EnableCodeIndexing: Yes +UseSpacesForTab: Yes +NumSpacesForTab: 2 +Encoding: UTF-8 + +RnwWeave: Sweave +LaTeX: pdfLaTeX diff --git a/sesion_2/retos/Cuaderno_de_campo_digital.accdb b/sesion_2/retos/Cuaderno_de_campo_digital.accdb new file mode 100644 index 0000000..d1a4e52 Binary files /dev/null and b/sesion_2/retos/Cuaderno_de_campo_digital.accdb differ diff --git a/sesion_3/retos/Reto_3_f1.pdf b/sesion_3/retos/Reto_3_f1.pdf new file mode 100644 index 0000000..0a15df8 Binary files /dev/null and b/sesion_3/retos/Reto_3_f1.pdf differ diff --git a/sesion_3/retos/Reto_3_f2.pdf b/sesion_3/retos/Reto_3_f2.pdf new file mode 100644 index 0000000..7a4c347 Binary files /dev/null and b/sesion_3/retos/Reto_3_f2.pdf differ diff --git a/sesion_3/retos/Reto_3_p1.R b/sesion_3/retos/Reto_3_p1.R new file mode 100644 index 0000000..9c2e697 --- /dev/null +++ b/sesion_3/retos/Reto_3_p1.R @@ -0,0 +1,10 @@ +#Algoritmo que multiplique 5 números introducidos por el usuario. + +multiplicacion <- 1 #Creo la variable "multiplicación" +for (k in 1:5){ #Creo un bucle para multiplicar 5 números que introducira el usuario + print ("introduce un numero") #Escribe "introduce un numero" para que el usuario introduzca el número que quiera + x <- scan (nmax=1) #Almaceno el valor introducido + + multiplicacion <- multiplicacion * x #Se multiplica el valor introducido con lo acumulado hasta el momento +} +print (multiplicacion) #Escribe el resultado de multiplicacion de los cinco numeros introducidos diff --git a/sesion_3/retos/Reto_3_p1.pdf b/sesion_3/retos/Reto_3_p1.pdf new file mode 100644 index 0000000..023abcc Binary files /dev/null and b/sesion_3/retos/Reto_3_p1.pdf differ diff --git a/sesion_3/retos/Reto_3_p2.R b/sesion_3/retos/Reto_3_p2.R new file mode 100644 index 0000000..020da0a --- /dev/null +++ b/sesion_3/retos/Reto_3_p2.R @@ -0,0 +1,16 @@ +#Algoritmo que dado un umbral por el usuario, +#dados 10 números por el usuario, +#cuente cuántos de esos números supera el umbral indicado + + +contador <- 0 #Creo la variable "contador" +print("Introduce un umbral") #Escribe "introduce un umbral" para que el usuario introduzca el valor umbral que desee +umbral <- scan(nmax=1) #Almaceno el umbral introducido +for(k in 1:10){ #Creo un bucle para que el usuario introduzca 10 numeros y cuente cuantos de estos numeros superan el umbral previamente introducido + print("Introduce un numero")#Escribe "introduce un número" para que el usuario introduzca el numero que quiera + x <- scan(nmax=1)#Almaceno el numero introducido + if(x > umbral){ #Si el numero introducido es mayor que el valor umbral + contador <- contador+1 #Aumenta en una unidad el contador (Cuenta el numero dado) + } +} +print(paste("Cantidad de numeros que superan el umbral:",contador)) #Mostrar la cantidad de numeros que supero el umbral diff --git a/sesion_3/retos/Reto_3_p2.pdf b/sesion_3/retos/Reto_3_p2.pdf new file mode 100644 index 0000000..e184e12 Binary files /dev/null and b/sesion_3/retos/Reto_3_p2.pdf differ diff --git a/sesion_3/retos/Reto_3_p3.R b/sesion_3/retos/Reto_3_p3.R new file mode 100644 index 0000000..898d018 --- /dev/null +++ b/sesion_3/retos/Reto_3_p3.R @@ -0,0 +1,11 @@ +#Algoritmo que haga la media de 10 temperaturas indicadas por el usuario + + +suma <-0 #Creo la variable suma +for(k in 1:10){ #Creo un bucle para que el usuario introduzca 10 valores de temperatura y realice la media de estos valores + print("Introduce la temperatura") #Escribe "introduce la temperatura" para que el usuario introduzca el valor de temperatura que quiera + temp <-scan(nmax=1) #Almaceno la temperatura introducida + suma <- suma + temp #Se suma el valor introducido con el acumulado +} +resultado <- suma/k #Creo la variable resultado que es el valor de la suma total entre el numero de temperaturas introducidas (media) +print(resultado) #Muestro el resultado obtenido diff --git a/sesion_3/retos/Reto_3_p3.pdf b/sesion_3/retos/Reto_3_p3.pdf new file mode 100644 index 0000000..3e0a219 Binary files /dev/null and b/sesion_3/retos/Reto_3_p3.pdf differ diff --git a/sesion_3/retos/Reto_3_p4.R b/sesion_3/retos/Reto_3_p4.R new file mode 100644 index 0000000..9295226 --- /dev/null +++ b/sesion_3/retos/Reto_3_p4.R @@ -0,0 +1,28 @@ +#Reto_3_p4: +#Dado el conjunto de datos ndvi.zip +#Cada imagen raster se tomó un día, a una hora y minutos concretos +#Queremos hacer un script que muestre la evolución del NDVI medio para las horas del día + +#Cargo las librerias raster y rgdal +library(raster) +library(rgdal) + +horas <- c(12,13,14,15) #Creo el vector horas donde se almacenaran las horas +valores <- c() #Creo el vector valores vacio donde se almacenaran los valores de las medias de cada hora +i <- 1 #Creo un contador para acceder correctamente a la posicion dentro del vector valores donde se deben almacenar las distintas medias para las horas + +#Creo un bucle que busque los tres ficheros para cada hora, los "junte" y realice la media de los valores para cada hora, obteniendo un único valor de ndvi para cada hora +for(h in horas){ + + imagenes <- list.files(path="./ndvi", full.names=TRUE, pattern=paste("_", h, "..\\.jpg\\.asc$", sep="")) + stack_imagen <- stack(imagenes) + media2 <- mean(stack_imagen) + media <- cellStats(media2,mean) + + valores[i] <- media #Almaceno en el vector valores, en la posición correspondiente la media para cada hora + i <- i+1 + + } +plot(horas, valores) #Muestra la gráfica obtenida + + diff --git a/sesion_3/retos/Reto_3_p4.pdf b/sesion_3/retos/Reto_3_p4.pdf new file mode 100644 index 0000000..9113345 Binary files /dev/null and b/sesion_3/retos/Reto_3_p4.pdf differ diff --git a/sesion_3/retos/Reto_3_p4.png b/sesion_3/retos/Reto_3_p4.png new file mode 100644 index 0000000..fc752b1 Binary files /dev/null and b/sesion_3/retos/Reto_3_p4.png differ diff --git a/sesion_6/retos/producto_1/Mapa de tendencias de la nieve.png b/sesion_6/retos/producto_1/Mapa de tendencias de la nieve.png new file mode 100644 index 0000000..69b7673 Binary files /dev/null and b/sesion_6/retos/producto_1/Mapa de tendencias de la nieve.png differ diff --git a/sesion_6/retos/producto_1/Mapa de tendencias del ndvi.png b/sesion_6/retos/producto_1/Mapa de tendencias del ndvi.png new file mode 100644 index 0000000..7b08798 Binary files /dev/null and b/sesion_6/retos/producto_1/Mapa de tendencias del ndvi.png differ diff --git a/sesion_6/retos/producto_1/Mapa de tendencias significativas de la nieve.png b/sesion_6/retos/producto_1/Mapa de tendencias significativas de la nieve.png new file mode 100644 index 0000000..c6ca03d Binary files /dev/null and b/sesion_6/retos/producto_1/Mapa de tendencias significativas de la nieve.png differ diff --git a/sesion_6/retos/producto_1/Mapa de tendencias significativas del ndvi.png b/sesion_6/retos/producto_1/Mapa de tendencias significativas del ndvi.png new file mode 100644 index 0000000..54133ae Binary files /dev/null and b/sesion_6/retos/producto_1/Mapa de tendencias significativas del ndvi.png differ diff --git a/sesion_6/retos/producto_1/Producto_1.pdf b/sesion_6/retos/producto_1/Producto_1.pdf new file mode 100644 index 0000000..48b3674 Binary files /dev/null and b/sesion_6/retos/producto_1/Producto_1.pdf differ diff --git a/sesion_6/retos/producto_1/ndvi_robledal.csv b/sesion_6/retos/producto_1/ndvi_robledal.csv index d557721..a530fc2 100644 --- a/sesion_6/retos/producto_1/ndvi_robledal.csv +++ b/sesion_6/retos/producto_1/ndvi_robledal.csv @@ -23580,4 +23580,4 @@ 197847;2009;0.51;-3.13026053962652;37.0276617611953 197847;2010;0.54;-3.13026053962652;37.0276617611953 197847;2011;0.51;-3.13026053962652;37.0276617611953 -197847;2012;0.53;-3.13026053962652;37.0276617611953 +197847;2012;0.53;-3.13026053962652;37.0276617611953 \ No newline at end of file diff --git a/sesion_6/retos/producto_1/producto_1.R b/sesion_6/retos/producto_1/producto_1.R new file mode 100644 index 0000000..7974aa3 --- /dev/null +++ b/sesion_6/retos/producto_1/producto_1.R @@ -0,0 +1,127 @@ +#Cargo las librerias Kendall, sp, rgdal, classInt y RColorBrewer +library('Kendall') +library(sp) +library(rgdal) +library(classInt) +library(RColorBrewer) + + +v <- c()#Creo un vector (v) que almacena los valores del ndvi de un pixel para todos los años +m <- c()#Almacena los valores del Analisis MannKendall de robledal +m2 <-c()#Almacena los valores del Analisis MannKendall de nieve +tau <-c() +pvalue <- c() +lat <- c() +lng <- c() + +#######ROBLEDAL######## +#Leo la table de ndvi_robledal +ndvi <- read.table("ndvi_robledal.csv",header = TRUE,sep=';') + +#Creo un bucle +for(k in 1:1814){ + + #Almaceno la latitud y longitud de cada pixel + lat[k] <-ndvi$lat[(k-1)*13+1] + lng[k] <-ndvi$lng[(k-1)*13+1] + + + for(i in 1:13){ + + #Almaceno el valor del ndvi de cada año para cada pixel + v[i]=ndvi$ndvi_i[(k-1)*13+i] + } + #Realizo el analisis MannKendall + m <- MannKendall(v) + #Extraigo tau y pvalor + tau[k] <-m$tau[1] + pvalue[k] <- m$sl[1] +} + +#Uso script pintar# + +#Pintar mapas +tendencias=data.frame(tau,pvalue,lat,lng) +## definimos las coordenadas de los puntos +coordinates(tendencias) =~lng+lat +## definimos el sistema de coordenadas WGS84 +proj4string(tendencias)=CRS("+init=epsg:4326") + +## partimos los valores de tau en 5 clases +clases <- classIntervals(tendencias$tau, n = 5) +## obtenemos cinco colores para una paleta de colores que se llama "Spectral" +plotclr <- rev(brewer.pal(5, "Spectral")) +## Asociamos los valores de tau a su valor correspondiente +colcode <- findColours(clases, plotclr) + +## plot sin tener en cuenta +plot(tendencias, col=colcode, pch=19, cex = .6, main = "Mapa de tendencias del ndvi") +## mostramos la leyenda +legend("topright", legend=names(attr(colcode, "table")), fill=attr(colcode, "palette"), bty="n") + +###### pintar el mapa + +tendencias$significativa <- ifelse(tendencias$pvalue < 0.05, 1, 2) +## plot sin tener en cuenta +plot(tendencias, col=colcode, pch=c(17, 19)[as.numeric(tendencias$significativa)], cex = .6, main = "Mapa de tendencias significativas del ndvi") +## mostramos la leyenda +legend("topright", legend=names(attr(colcode, "table")), fill=attr(colcode, "palette"), bty="n") + + + + + + +#########NIEVE####### +#Leo la tabla nieve_robledal +nieve <- read.table("nieve_robledal.csv",header =TRUE, sep=';') + +#Creo un bucle +for(k in 1:593){ + #Almaceno la latitud y longitud de cada pixel + lat[k] <-nieve$lat[(k-1)*12+1] + lng[k] <-nieve$lng[(k-1)*12+1] + + for(i in 1:12){ + #Alamaceno el valor del scd de cada año para cada pixel + v[i]=nieve$scd[(k-1)*12+i] + } + #Realizo el Analisis MannKendall + m2<- MannKendall(v) + #Extraigo los valores de tau y pvalor + tau[k] <-m2$tau[1] + pvalue[k] <- m2$sl[1] + +} + +#Uso el script pintar# + +tendencias=data.frame(tau,pvalue,lat,lng) +## definimos las coordenadas de los puntos +coordinates(tendencias) =~lng+lat +## definimos el sistema de coordenadas WGS84 +proj4string(tendencias)=CRS("+init=epsg:4326") + +## partimos los valores de tau en 5 clases +clases <- classIntervals(tendencias$tau, n = 5) +## obtenemos cinco colores para una paleta de colores que se llama "Spectral" +plotclr <- rev(brewer.pal(5, "Spectral")) +## Asociamos los valores de tau a su valor correspondiente +colcode <- findColours(clases, plotclr) + +## plot sin tener en cuenta +plot(tendencias, col=colcode, pch=19, cex = .6, main = "Mapa de tendencias de la nieve") +## mostramos la leyenda +legend("topright", legend=names(attr(colcode, "table")), fill=attr(colcode, "palette"), bty="n") + +###### pintar el mapa + +tendencias$significativa <- ifelse(tendencias$pvalue < 0.05, 1, 2) +## plot sin tener en cuenta +plot(tendencias, col=colcode, pch=c(17, 19)[as.numeric(tendencias$significativa)], cex = .6, main = "Mapa de tendencias significativas de la nieve") +## mostramos la leyenda +legend("topright", legend=names(attr(colcode, "table")), fill=attr(colcode, "palette"), bty="n") + + + + diff --git a/sesion_6/retos/producto_2/Clasificacion del robledal en 3 subtipos.png b/sesion_6/retos/producto_2/Clasificacion del robledal en 3 subtipos.png new file mode 100644 index 0000000..84791f8 Binary files /dev/null and b/sesion_6/retos/producto_2/Clasificacion del robledal en 3 subtipos.png differ diff --git a/sesion_6/retos/producto_2/Clasificacion del robledal en 4 subtipos.png b/sesion_6/retos/producto_2/Clasificacion del robledal en 4 subtipos.png new file mode 100644 index 0000000..0002e33 Binary files /dev/null and b/sesion_6/retos/producto_2/Clasificacion del robledal en 4 subtipos.png differ diff --git a/sesion_6/retos/producto_2/Clasificacion del robledal en 5 subtipos.png b/sesion_6/retos/producto_2/Clasificacion del robledal en 5 subtipos.png new file mode 100644 index 0000000..2c852ad Binary files /dev/null and b/sesion_6/retos/producto_2/Clasificacion del robledal en 5 subtipos.png differ diff --git a/sesion_6/retos/producto_2/Producto_2.pdf b/sesion_6/retos/producto_2/Producto_2.pdf new file mode 100644 index 0000000..1968cc0 Binary files /dev/null and b/sesion_6/retos/producto_2/Producto_2.pdf differ diff --git a/sesion_6/retos/producto_2/producto_2.R b/sesion_6/retos/producto_2/producto_2.R new file mode 100644 index 0000000..2f01b53 --- /dev/null +++ b/sesion_6/retos/producto_2/producto_2.R @@ -0,0 +1,53 @@ +#Cargo las librerias sp, rgdal, classInt y RColorBrewer +library(sp) +library(rgdal) +library(classInt) +library(RColorBrewer) + +n <- 3:5 #Numero de clases que voy a usar en los diferentes clusters + +#Leo la tabla robles_ecoinfo +info <- read.table("robles_ecoinfo.csv",header = TRUE,sep=',') + +info2 <- info[,-1]#Quito la primera columna +info2 <-info2[,-1]#Quito la segunda columna + +#Creo un bucle +for(n_cluster in n){ + +#Realizo el cluster +mi_cluster <- kmeans (info2, n_cluster, iter.max = 200, nstart = 3) +#Extraigo el campo cluster del cluster +cluster <- mi_cluster$cluster +#Cojo las coordenadas x e y de "robles_ecoinfo" +x <- info$x +y <- info$y +#Hago una tabla con los valores del cluster, x e y +robles<-data.frame(cluster,x,y) + +### Uso el script pintar_mapa ### + +#Cargamos libreria raster y obtenemos los ficheros de imágenes para una hora determinada +### Para el siguiente fragmento de script, suponemos que: +### - Tienes una variable que se llama robles +### - La variable robles es un data.frame +### - que el data.frame tiene una columna que se llama pvalue +### - que tiene dos columnas, x e y +### - que tiene una columna que se llama "cluster" que indica a qué cluster pertenece. + + +## OJO: esta variable debería de estar en vuestro script +## variable con el número de cluster con el que estamos probando +## definimos las coordenadas de los puntos +coordinates(robles) =~x+y +## definimos el sistema de coordenadas WGS84 +proj4string(robles)=CRS("+init=epsg:23030") + +## obtenemos n_cluster colores para una paleta de colores que se llama "Spectral", para cada cluster creado +plotclr <- rev(brewer.pal(n_cluster, "Spectral")) + +## plot, asignando el color en función del cluster al que pertenece +plot(robles, col=plotclr[robles$cluster], pch=19, cex = .6, main = paste("Poblaciones de robledal con Clusters=",n_cluster)) +} + + diff --git a/sesion_6/retos/producto_3/__MACOSX/._presencia b/sesion_6/retos/producto_3/__MACOSX/._presencia new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/._presencia differ diff --git a/sesion_6/retos/producto_3/__MACOSX/._resultados b/sesion_6/retos/producto_3/__MACOSX/._resultados new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/._resultados differ diff --git a/sesion_6/retos/producto_3/__MACOSX/._variables b/sesion_6/retos/producto_3/__MACOSX/._variables new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/._variables differ diff --git a/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia.xls b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia.xls new file mode 100644 index 0000000..0f4e2e9 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia.xls differ diff --git a/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_geograficas.xls b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_geograficas.xls new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_geograficas.xls differ diff --git a/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_q_pyrenaica.csv b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_q_pyrenaica.csv new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/presencia/._presencia_q_pyrenaica.csv differ diff --git a/sesion_6/retos/producto_3/__MACOSX/resultados/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/resultados/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/resultados/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/._futuro_B2 b/sesion_6/retos/producto_3/__MACOSX/variables/._futuro_B2 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/._futuro_B2 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/._presente b/sesion_6/retos/producto_3/__MACOSX/variables/._presente new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/._presente differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2010 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2010 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2010 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2020 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2020 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2020 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2030 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2030 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2030 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2040 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2040 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2040 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2050 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2050 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2050 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2060 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2060 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2060 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2070 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2070 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2070 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2080 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2080 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2080 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2090 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2090 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2090 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2100 b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2100 new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/._2100 differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2020/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2020/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2020/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2030/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2030/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2030/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2040/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2040/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2040/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2050/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2050/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2050/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2060/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2060/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2060/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2070/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2070/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2070/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2080/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2080/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2080/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2090/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2090/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2090/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._PA.asc.aux.xml b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._PA.asc.aux.xml new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._PA.asc.aux.xml differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._TMNI.asc.aux.xml b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._TMNI.asc.aux.xml new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._TMNI.asc.aux.xml differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._huella_humana.asc.aux.xml b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._huella_humana.asc.aux.xml new file mode 100644 index 0000000..4d238e0 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._huella_humana.asc.aux.xml differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._maxent.cache b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._maxent.cache new file mode 100644 index 0000000..233c500 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/._maxent.cache differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PA.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._PV.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNI.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMNV.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXI.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.mxe new file mode 100644 index 0000000..4d238e0 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._TMXV.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._huella_humana.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_horas_ran.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_horas_ran.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_horas_ran.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._sol_rad_sum.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.mxe new file mode 100644 index 0000000..4d238e0 Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_pend.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.info b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.info new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.info differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.mxe b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.mxe new file mode 100644 index 0000000..626e16b Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/futuro_B2/2100/maxent.cache/._topo_posic.mxe differ diff --git a/sesion_6/retos/producto_3/__MACOSX/variables/presente/._.DS_Store b/sesion_6/retos/producto_3/__MACOSX/variables/presente/._.DS_Store new file mode 100644 index 0000000..b91719d Binary files /dev/null and b/sesion_6/retos/producto_3/__MACOSX/variables/presente/._.DS_Store differ diff --git a/sesion_6/retos/producto_3/presencia/presencia.xls b/sesion_6/retos/producto_3/presencia/presencia.xls new file mode 100755 index 0000000..39c5d1f Binary files /dev/null and b/sesion_6/retos/producto_3/presencia/presencia.xls differ diff --git a/sesion_6/retos/producto_3/presencia/presencia_geograficas.xls b/sesion_6/retos/producto_3/presencia/presencia_geograficas.xls new file mode 100755 index 0000000..c972543 Binary files /dev/null and b/sesion_6/retos/producto_3/presencia/presencia_geograficas.xls differ diff --git a/sesion_6/retos/producto_3/presencia/presencia_q_pyrenaica.csv b/sesion_6/retos/producto_3/presencia/presencia_q_pyrenaica.csv new file mode 100755 index 0000000..bc8dd17 --- /dev/null +++ b/sesion_6/retos/producto_3/presencia/presencia_q_pyrenaica.csv @@ -0,0 +1,42 @@ +species,x,y +Quercus pyrenaica,-3.4592728,37.090360 +Quercus pyrenaica,-3.4425072,37.103936 +Quercus pyrenaica,-3.4366969,37.099468 +Quercus pyrenaica,-3.4368059,37.094943 +Quercus pyrenaica,-3.4311936,37.099425 +Quercus pyrenaica,-3.5126549,37.085607 +Quercus pyrenaica,-3.2901411,36.987112 +Quercus pyrenaica,-3.2901131,36.982704 +Quercus pyrenaica,-3.2955989,36.960091 +Quercus pyrenaica,-3.2956725,36.964589 +Quercus pyrenaica,-3.2956849,36.991038 +Quercus pyrenaica,-3.3966371,36.950827 +Quercus pyrenaica,-3.3966605,36.955335 +Quercus pyrenaica,-3.4023723,36.975643 +Quercus pyrenaica,-3.4023708,36.971100 +Quercus pyrenaica,-3.4079099,36.973280 +Quercus pyrenaica,-3.4137523,36.973314 +Quercus pyrenaica,-3.4136521,36.975559 +Quercus pyrenaica,-3.4164485,36.973287 +Quercus pyrenaica,-3.4192188,36.964235 +Quercus pyrenaica,-3.4814641,37.058641 +Quercus pyrenaica,-3.4758634,37.058735 +Quercus pyrenaica,-3.4646709,37.067758 +Quercus pyrenaica,-3.4816563,37.067699 +Quercus pyrenaica,-3.4813782,37.036142 +Quercus pyrenaica,-3.3808810,37.135790 +Quercus pyrenaica,-3.3750942,37.131292 +Quercus pyrenaica,-3.3809367,37.131183 +Quercus pyrenaica,-3.3638698,37.131327 +Quercus pyrenaica,-3.3466400,37.117965 +Quercus pyrenaica,-3.3415149,37.122217 +Quercus pyrenaica,-3.3469166,37.126780 +Quercus pyrenaica,-3.3468960,37.122273 +Quercus pyrenaica,-3.2566979,37.190109 +Quercus pyrenaica,-3.2530087,37.202160 +Quercus pyrenaica,-3.2570577,37.176605 +Quercus pyrenaica,-3.2514513,37.181088 +Quercus pyrenaica,-3.2514776,37.185613 +Quercus pyrenaica,-3.2564000,37.185305 +Quercus pyrenaica,-3.2458027,37.190132 +Quercus pyrenaica,-3.2514363,37.190139 \ No newline at end of file diff --git a/sesion_6/retos/producto_3/resultados/Producto_3.pdf b/sesion_6/retos/producto_3/resultados/Producto_3.pdf new file mode 100644 index 0000000..2aa3e0a Binary files /dev/null and b/sesion_6/retos/producto_3/resultados/Producto_3.pdf differ diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.html b/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.html new file mode 100644 index 0000000..13019ec --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:08:27 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2020\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2020. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2020\Quercus_pyrenaica_2020_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2020. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2020 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2020
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2020
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2020
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2020 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2020 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.qgs b/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.qgs new file mode 100644 index 0000000..1963032 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2020.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.891688 + -1.477742 + 38.833203 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208213931642 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2020/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_202020150208214241373 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2020/Quercus_pyrenaica_2020.asc + Quercus_pyrenaica_2020 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.html b/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.html new file mode 100644 index 0000000..a9a3f3d --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Thu Jan 29 18:02:33 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.7980.140Equate entropy of thresholded and original distributions0.0150.033

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in H:\Documents and Settings\alumno\Escritorio\andalucia\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in H:\Documents and Settings\alumno\Escritorio\andalucia\variables\futuro_A2\2030. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in H:\Documents and Settings\alumno\Escritorio\andalucia\Quercus_pyrenaica_2030_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in H:\Documents and Settings\alumno\Escritorio\andalucia\variables\futuro_A2\2030. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2030 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in H:\Documents and Settings\alumno\Escritorio\andalucia\variables\futuro_A2\2030
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (9 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: linear quadratic hinge
+responsecurves: true
+outputdirectory: H:\Documents and Settings\alumno\Escritorio\andalucia
+projectionlayers: H:\Documents and Settings\alumno\Escritorio\andalucia\variables\futuro_A2\2030
+samplesfile: H:\Documents and Settings\alumno\Escritorio\andalucia\presencia\presencia_q_pyrenaica.csv
+environmentallayers: H:\Documents and Settings\alumno\Escritorio\andalucia\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves "outputdirectory=H:\Documents and Settings\alumno\Escritorio\andalucia" "projectionlayers=H:\Documents and Settings\alumno\Escritorio\andalucia\variables\futuro_A2\2030" "samplesfile=H:\Documents and Settings\alumno\Escritorio\andalucia\presencia\presencia_q_pyrenaica.csv" "environmentallayers=H:\Documents and Settings\alumno\Escritorio\andalucia\variables\presente" -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.qgs b/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.qgs new file mode 100644 index 0000000..f5b4ac8 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2030.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.891688 + -1.477742 + 38.833203 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208214403946 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2030/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_203020150208214409129 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2030/Quercus_pyrenaica_2030.asc + Quercus_pyrenaica_2030 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.html b/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.html new file mode 100644 index 0000000..3909409 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 22:55:26 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2040\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2040. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2040\Quercus_pyrenaica_2040_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2040. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2040 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2040
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2040
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2040
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2040 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2040 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.qgs b/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.qgs new file mode 100644 index 0000000..a67f6c7 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2040.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.891688 + -1.477742 + 38.833203 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208214524079 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2040/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_204020150208214529414 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2040/Quercus_pyrenaica_2040.asc + Quercus_pyrenaica_2040 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.html b/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.html new file mode 100644 index 0000000..8aa66fb --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:21:02 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2050\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2050. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2050\Quercus_pyrenaica_2050_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2050. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2050 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2050
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (22 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2050
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2050
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_A2_2050 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_A2\2050 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.qgs b/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.qgs new file mode 100644 index 0000000..d606480 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_A2_2050.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.891688 + -1.477742 + 38.833203 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208214630413 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2050/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_205020150208214635210 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_A2_2050/Quercus_pyrenaica_2050.asc + Quercus_pyrenaica_2050 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.html new file mode 100644 index 0000000..23ced58 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:14:49 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2020\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2020. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2020\Quercus_pyrenaica_2020_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2020. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2020 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2020
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2020
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2020
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2020 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2020 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.qgs new file mode 100644 index 0000000..85932dc --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2020.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.891688 + -1.477742 + 38.833203 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208214756308 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2020/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_202020150208214801275 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2020/Quercus_pyrenaica_2020.asc + Quercus_pyrenaica_2020 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.html new file mode 100644 index 0000000..f5dab54 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:17:02 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2030\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2030. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2030\Quercus_pyrenaica_2030_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2030. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2030 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2030
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2030
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2030
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2030 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2030 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.qgs new file mode 100644 index 0000000..b166053 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2030.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -8.546600 + 35.390006 + -0.603636 + 39.334885 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208214903382 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2030/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_203020150208214908411 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2030/Quercus_pyrenaica_2030.asc + Quercus_pyrenaica_2030 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.html new file mode 100644 index 0000000..b44a4f7 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:25:47 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2040\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2040. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2040\Quercus_pyrenaica_2040_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2040. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2040 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2040
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2040
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2040
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2040 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2040 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.qgs new file mode 100644 index 0000000..89e2207 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2040.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215047150 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2040/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_204020150208215052363 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2040/Quercus_pyrenaica_2040.asc + Quercus_pyrenaica_2040 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.html new file mode 100644 index 0000000..6587a7b --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:27:57 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2050\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2050. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2050\Quercus_pyrenaica_2050_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2050. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2050 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2050
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2050
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2050
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2050 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2050 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.qgs new file mode 100644 index 0000000..05c5fb1 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2050.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215202522 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2050/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_205020150208215208311 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2050/Quercus_pyrenaica_2050.asc + Quercus_pyrenaica_2050 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.html new file mode 100644 index 0000000..6d62de9 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:29:54 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2060\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2060. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2060\Quercus_pyrenaica_2060_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2060. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2060 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2060
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2060
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2060
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2060 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2060 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.qgs new file mode 100644 index 0000000..c497111 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2060.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215352721 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2060/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_206020150208215400791 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2060/Quercus_pyrenaica_2060.asc + Quercus_pyrenaica_2060 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.html new file mode 100644 index 0000000..97818a0 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:32:30 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2070\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2070. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2070\Quercus_pyrenaica_2070_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2070. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2070 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2070
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2070
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2070
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2070 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2070 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.qgs new file mode 100644 index 0000000..7ca846e --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2070.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215501844 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2070/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_207020150208215507593 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2070/Quercus_pyrenaica_2070.asc + Quercus_pyrenaica_2070 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.html new file mode 100644 index 0000000..b22b7ba --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:34:48 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2080\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2080. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2080\Quercus_pyrenaica_2080_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2080. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2080 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2080
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2080
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2080
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2080 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2080 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.qgs new file mode 100644 index 0000000..c87c630 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2080.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215608682 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2080/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_208020150208215614275 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2080/Quercus_pyrenaica_2080.asc + Quercus_pyrenaica_2080 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.html new file mode 100644 index 0000000..39948df --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:36:52 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2090\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2090. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2090\Quercus_pyrenaica_2090_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2090. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2090 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2090
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2090
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2090
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2090 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2090 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.qgs new file mode 100644 index 0000000..e480eba --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2090.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215719129 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2090/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_209020150208215726191 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2090/Quercus_pyrenaica_2090.asc + Quercus_pyrenaica_2090 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.html b/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.html new file mode 100644 index 0000000..f589999 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.html @@ -0,0 +1,80 @@ +Maxent model for Quercus_pyrenaica +

Maxent model for Quercus_pyrenaica

+
This page contains some analysis of the Maxent model for Quercus_pyrenaica, created Sun Feb 01 23:38:57 CET 2015 using Maxent version 3.3.3k. If you would like to do further analyses, the raw data used here is linked to at the end of this page.
+

Analysis of omission/commission

+The following picture shows the omission rate and predicted area as a function of the cumulative threshold. The omission rate is is calculated both on the training presence records, and (if test data are used) on the test records. The omission rate should be close to the predicted omission, because of the definition of the cumulative threshold. +

+
The next picture is the receiver operating characteristic (ROC) curve for the same data. Note that the specificity is defined using predicted area, rather than true commission (see the paper by Phillips, Anderson and Schapire cited on the help page for discussion of what this means). This implies that the maximum achievable AUC is less than 1. If test data is drawn from the Maxent distribution itself, then the maximum possible test AUC would be 0.991 rather than 1; in practice the test AUC may exceed this bound. +

+
+
+Some common thresholds and corresponding omission rates are as follows. If test data are available, binomial probabilities are calculated exactly if the number of test samples is at most 25, otherwise using a normal approximation to the binomial. These are 1-sided p-values for the null hypothesis that test points are predicted no better than by a random prediction with the same fractional predicted area. The "Balance" threshold minimizes 6 * training omission rate + .04 * cumulative threshold + 1.6 * fractional predicted area.
+
Cumulative thresholdLogistic thresholdDescriptionFractional predicted areaTraining omission rate
1.0000.005Fixed cumulative value 10.0600.000
5.0000.045Fixed cumulative value 50.0260.000
10.0000.117Fixed cumulative value 100.0160.000
11.6880.137Minimum training presence0.0150.000
30.1630.34210 percentile training presence0.0060.100
11.6880.137Equal training sensitivity and specificity0.0150.000
11.6880.137Maximum training sensitivity plus specificity0.0150.000
1.1050.006Balance training omission, predicted area and threshold value0.0580.000
11.6880.137Equate entropy of thresholded and original distributions0.0150.000

+

Pictures of the model

+This is a representation of the Maxent model for Quercus_pyrenaica. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+

+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2100\Quercus_pyrenaica_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+This is the projection of the Maxent model for Quercus_pyrenaica onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2100. Warmer colors show areas with better predicted conditions. White dots show the presence locations used for training, while violet dots show test locations. Click on the image for a full-size version.
+
+
Click here to interactively explore this prediction using the Explain tool. If clicking from your browser does not succeed in starting the tool, try running the script in D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2100\Quercus_pyrenaica_2100_explain.bat directly. This tool requires the environmental grids to be small enough that they all fit in memory.

+The following picture shows where the prediction is most affected by variables being outside their training range, while projecting the Maxent model onto the environmental variables in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2100. The values shown in the picture give the absolute difference in predictions when using vs not using clamping. (Clamping means that environmental variables and features are restricted to the range of values encountered during training.) Warmer colors show areas where the treatment of variable values outside their training ranges is likely to have a large effect on predicted suitability. +
+
The following two pictures compare the environmental similarity of variables in 2100 to the environmental data used for training the model. In the first picture (MESS), areas in red have one or more environmental variables outside the range present in the training data, so predictions in those areas should be treated with strong caution. The second picture (MoD) shows the most dissimilar variable, i.e., the one that is furthest outside its training range. For details, see Elith et al., Methods in Ecology and Evolution, 2010 +

+

+

Response curves

+
These curves show how each environmental variable affects the Maxent prediction. +The +curves show how the logistic prediction changes as each environmental variable is varied, keeping all other environmental variables at their average sample value. Click on a response curve to see a larger version. Note that the curves can be hard to interpret if you have strongly correlated variables, as the model may depend on the correlations in ways that are not evident in the curves. In other words, the curves show the marginal effect of changing exactly one variable, whereas the model may take advantage of sets of variables changing together.

+ + + + + + + + + + +
+
In contrast to the above marginal response curves, each of the following curves represents a different model, namely, a Maxent model created using only the corresponding variable. These plots reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. They may be easier to interpret if there are strong correlations between variables.

+ + + + + + + + + + +
+

Analysis of variable contributions


+The following table gives estimates of relative contributions of the environmental variables to the Maxent model. To determine the first estimate, in each iteration of the training algorithm, the increase in regularized gain is added to the contribution of the corresponding variable, or subtracted from it if the change to the absolute value of lambda is negative. For the second estimate, for each environmental variable in turn, the values of that variable on training presence and background data are randomly permuted. The model is reevaluated on the permuted data, and the resulting drop in training AUC is shown in the table, normalized to percentages. As with the variable jackknife, variable contributions should be interpreted with caution when the predictor variables are correlated.
+
VariablePercent contributionPermutation importance
TMXV46.528.5
topo_posic25.31.6
TMXI15.30.1
TMNV5.355.8
PV3.67.5
TMNI2.85.6
sol_rad_sum0.50.1
PA0.40.6
topo_pend0.20.1
huella_humana0.20


+

Raw data outputs and control parameters


+The data used in the above analysis is contained in the next links. Please see the Help button for more information on these.
+The model applied to the training environmental layers
+The model applied to the environmental layers in D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2100
+The coefficients of the model
+The omission and predicted area for varying cumulative and raw thresholds
+The prediction strength at the training and (optionally) test presence sites
+Results for all species modeled in the same Maxent run, with summary statistics and (optionally) jackknife results
+

+Regularized training gain is 4.204, training AUC is 0.997, unregularized training gain is 4.589.
+Algorithm terminated after 500 iterations (21 seconds).
+
+The follow settings were used during the run:
+30 presence records used for training.
+10027 points used to determine the Maxent distribution (background points and presence points).
+Environmental layers used (all continuous): PA PV TMNI TMNV TMXI TMXV huella_humana sol_rad_sum topo_pend topo_posic
+Regularization values: linear/quadratic/product: 0.250, categorical: 0.250, threshold: 1.700, hinge: 0.500
+Feature types used: hinge linear quadratic
+responsecurves: true
+outputdirectory: D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2100
+projectionlayers: D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2100
+samplesfile: D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv
+environmentallayers: D:\ECOINFORMATICA\andalucia1\variables\presente
+Command line used:
+
+Command line to repeat this species model: java density.MaxEnt nowarnings noprefixes -E "" -E Quercus_pyrenaica responsecurves outputdirectory=D:\ECOINFORMATICA\andalucia1\RESULTADOS_B2_2100 projectionlayers=D:\ECOINFORMATICA\andalucia1\variables\futuro_B2\2100 samplesfile=D:\ECOINFORMATICA\andalucia1\presencia\presencia_q_pyrenaica.csv environmentallayers=D:\ECOINFORMATICA\andalucia1\variables\presente -N ._PA -N ._PV -N ._TMNI -N ._TMNV -N ._TMXI -N ._TMXV -N ._huella_humana -N ._sol_rad_sum -N ._topo_pend -N ._topo_posic
diff --git a/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.qgs b/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.qgs new file mode 100644 index 0000000..c9ddbf9 --- /dev/null +++ b/sesion_6/retos/producto_3/resultados/Resultado_B2_2100.qgs @@ -0,0 +1,157 @@ + + + + + degrees + + -7.672494 + 35.824131 + -1.477742 + 38.900760 + + 0 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + + + + + + + + + + + + Quercus_pyrenaica20150208215841776 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2100/Quercus_pyrenaica.asc + Quercus_pyrenaica + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + Quercus_pyrenaica_210020150208215846799 + D:/ECOINFORMATICA/andalucia1/RESULTADOS_B2_2100/Quercus_pyrenaica_2100.asc + Quercus_pyrenaica_2100 + + + +proj=longlat +datum=WGS84 +no_defs + 3452 + 4326 + EPSG:4326 + WGS 84 + longlat + WGS84 + true + + + 255 + + gdal + + SingleBandPseudoColor + PseudoColorShader + + Not Set + Not Set + Not Set + Banda 1 + 0 + + + + + NoEnhancement + + + -3.40282e+38 + 3.40282e+38 + + + -9999.000000 + + + + + + + + + + + + EPSG:4326 + + + false + + + 0 + 255 + 255 + 255 + 255 + 255 + 255 + + + 2 + true + + + diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/PA.asc.aux.xml b/sesion_6/retos/producto_3/variables/futuro_B2/2100/PA.asc.aux.xml new file mode 100755 index 0000000..e9917a0 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/PA.asc.aux.xml @@ -0,0 +1,10 @@ + + + + 1502 + 440.10678231196 + 61 + 152.7062815289 + + + diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/TMNI.asc.aux.xml b/sesion_6/retos/producto_3/variables/futuro_B2/2100/TMNI.asc.aux.xml new file mode 100755 index 0000000..8407893 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/TMNI.asc.aux.xml @@ -0,0 +1,10 @@ + + + + 111 + 52.694582681644 + -48 + 21.603717219418 + + + diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/huella_humana.asc.aux.xml b/sesion_6/retos/producto_3/variables/futuro_B2/2100/huella_humana.asc.aux.xml new file mode 100755 index 0000000..6885af1 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/huella_humana.asc.aux.xml @@ -0,0 +1,10 @@ + + + + 42 + 11.978532840489 + 0 + 6.5683969337754 + + + diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.info new file mode 100755 index 0000000..40874d6 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/PA.asc +modified 1305263438000 +length 1148075 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.mxe new file mode 100755 index 0000000..529727c Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PA.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.info new file mode 100755 index 0000000..52cc12a --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/PV.asc +modified 1305263438000 +length 1024736 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.mxe new file mode 100755 index 0000000..950beab Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/PV.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.info new file mode 100755 index 0000000..0cf23a2 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/TMNI.asc +modified 1305263438000 +length 1021835 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.mxe new file mode 100755 index 0000000..0a61184 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNI.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.info new file mode 100755 index 0000000..7dc9559 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/TMNV.asc +modified 1305263438000 +length 1147746 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.mxe new file mode 100755 index 0000000..23103af Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMNV.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.info new file mode 100755 index 0000000..da6ea8c --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/TMXI.asc +modified 1305263440000 +length 1141970 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.mxe new file mode 100755 index 0000000..cda88ea Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXI.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.info new file mode 100755 index 0000000..4fd7177 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/TMXV.asc +modified 1305263440000 +length 1147746 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.mxe new file mode 100755 index 0000000..6217d61 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/TMXV.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.info new file mode 100755 index 0000000..57b8b7c --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/huella_humana.asc +modified 1305263488000 +length 979942 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.mxe new file mode 100755 index 0000000..8790d97 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/huella_humana.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_horas_ran.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_horas_ran.mxe new file mode 100755 index 0000000..74777f5 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_horas_ran.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.info new file mode 100755 index 0000000..413ef32 --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/sol_rad_sum.asc +modified 1305263486000 +length 2008878 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.mxe new file mode 100755 index 0000000..01e1ccf Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/sol_rad_sum.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.info new file mode 100755 index 0000000..0f1fa0b --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/topo_pend.asc +modified 1305263486000 +length 1263126 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.mxe new file mode 100755 index 0000000..6332515 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_pend.mxe differ diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.info b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.info new file mode 100755 index 0000000..8a6313a --- /dev/null +++ b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.info @@ -0,0 +1,3 @@ +file /Users/fjbonet/Documents/4_docencia/master_ugr_cons_bio/2014_2015/ecoinformatica_2014_2015/sesiones/4_analisis_sintesis/3_modelos_distribucion/andalucia/variables/futuro_B2/2100/topo_posic.asc +modified 1305263486000 +length 1080216 diff --git a/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.mxe b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.mxe new file mode 100755 index 0000000..162a1b3 Binary files /dev/null and b/sesion_6/retos/producto_3/variables/futuro_B2/2100/maxent.cache/topo_posic.mxe differ