Count pixels for each class; your area would be the area of one pixel (resolution squared) times the count, and you can convert that (probably meters or feet) to hectares. How you can get them counted depends on the software you're using.

Mr. Paulo given the exact answer.

You can add new attribute column and convert the frequency (number of pixels in each class) into area in ha in the raster format. Otherwise, if you are having the number of pixels of each class then you can use the simple excel sheet for conversion of pixels into ha depends on your resolution of the output data or pixel size.

ArcGIS can do this.  Extract by Attributes allows you to select your ranges and this produces a separate raster for each range.  Convert your floating raster (created from the asc file) to integer raster using Raster calculator (you might have to first multiply up your values by 10000 to eliminate the decimal).  You can now use zonal statistics as table to calculate the area of your rasters.  Make sure your data is in a suitable projection such as UTM rather than WGS84.

Hi Daniel,

Sorry for the rushed reply to your question.  I hope what I said makes sense.  If not, let me know and I can give you more detail.  I have recently done the same calculation from my MaxEnt models.

Kind regards,

Richard

Hi Richard

It would be great if you could explain a bit more.

Hi Daniel,

I hope this makes a bit more sense.  I have tried to make it more concise too!

­   1) In “Conversion Tools” under “To Raster” select “ASCII to Raster” (select asc file from MaxEnt, name your raster, and select Float from the drop down list).

­   2) In ArcCatalog, right click on the raster which was created, click properties and scroll down to “Spatial Reference”.  You will probably have to select your projection, probably WGS84 and then click OK.

­   3) In ArcCatalog, select Spatial Analyst and then “Extract by Attributes”.  Your input raster is from 1 above.  The Where clause would be the values you want so for example if you want everything greater than 0.8 you would type VALUE > 0.8.  Choose a name and location for the Output raster.  This didn’t work for me when the raster was open in ArcGIS but running it through ArcCatalog seemed to work!

­   4) Using “Zonal Statistics as Table” (found under Spatial Analyst Tools then Zonal) select a basic polygon that covers the area of interest.  It doesn’t have to have any information attached and use the polygon’s ID as the Zone Field (in a basic polygon, there will only be one ID but be aware that for each value within the chosen Zone Field, you will get a calculated area), your raster from 3 above as the input value raster.  Now select a name and location for the Output table and keep Statistics type to ALL.

­   5) Once the table has been calculated, you can open it in Arc and Export the data as a text file which can then be imported to Excel.

You can also convert your raster from 3 above to a polygon and calculate areas that way.  To do this you need to convert your floating raster to an integer which is why I was talking about multiplying your values by 10000 etc in my previous answer.  As I said, it was rushed so please ignore that and try what I have suggest here instead.  I hope this works for you.  Let me know if you have any problems.

Kind regards,

Richard

Hi Richard,

Thanks for the explanation. It worked perfectly. Now can you help me understand the resulting figures:

Count = 351

Area = 0.609

Min = 0.800

Max = 0.883

Range = 0.083

Mean 0 0.822

STD = 0.017

Sum = 288.642

I used Africa (the entire continent, without islands) in step 4 (your explanation)

Thanks in advance

Daniel

Hi Daniel,

Area is what you need to look at.  Units will depend on the projection you used so for example, I use UTM which makes my area calculation in square meters.  The projection is very important here.  The min, max, range, and STD are all related to your model values.  Does the area calculation match to what you would of expected?

There is a second way to run this using "Zonal Geometry as Table" which will calculate the area for all your SDM model values at once but you will have to multiply the raster values in point 2 by 10000 (or similar) and then convert that raster to an integer raster.  If you want to do this, be warned it can take a while to run with large datasets.  I have just checked the two methods and they do output the same result, but with more detail in the latter.

I hope this helps.

Kind regards,

Richard

I have a similar issue but I am using BIOMOD2 for individual and ensemble species distribution modelling. I also want to categorise the ensemble model out put into suitability classes (5 classes: very low (0 - 0.2), low (0.2 - 0.4), medium (0.4 - 0.6), high (0.6 - 0.8) and very high (0.8 - 1). Can anyone tell me in a stepwise manner, how to do it. It would be a great help.

Kind regards,

Zafar

Hi Zafar A Reshi,

a bit late, but it might help other people with a similar issue like yours (i.e., categorise species distribution (raster) output into suitability classes in R):

#packages

> library(MASS)

> library(raster)

#create a reproducible example of a species distribution (sdm) raster ranging #from 0 to 1 (suitability)

> set.seed(42)

> ras_sdm ras_sdm ras_sdm crs(ras_sdm) ras_sdm plot(ras_sdm,legend.args=list(text='Suitability'))

#create the desired classes

> r_class = matrix(c(0, 0.2, 0.2, # very low

0.201, 0.4, 0.4, # low

0.401, 0.6, 0.6, # medium

0.601, 0.8, 0.8, # high

0.801,1, 1), # very high

ncol = 3, byrow = TRUE)

#reclassify the raster into a new raster with binned classes

> ras_sdm_recl = reclassify(ras_sdm, rcl = r_class)

# check the new reclassified raster

> plot(ras_sdm_recl, legend=FALSE,

col=c("#f2f2f2","#eeb48f","#e8c22f","#b0d900","#4dc003","#00a601"),xaxt='n', yaxt='n')

> par(xpd=TRUE)

> legend("topright",legend=c("very low","low","medium","high", "very high"),

fill=c("#f2f2f2","#eeb48f","#e8c22f","#b0d900","#4dc003","#00a601"))

# export the new raster

> writeRaster(x = ras_sdm_recl,

filename = "ras_sdm_recl.tif",

datatype = "FLT4S",

options = c("COMPRESS=DEFLATE"),

overwrite = TRUE)

# end---