20. Calculating most probable distribution of a species with Niche of Occurrence (NOO) in ModestR (Version ModestR 5.6 or higher)

MODESTR QUICK TUTORIALS HTTP://WWW.IPEZ.ES/MODESTR/
Step by step tutorial:
Calculating most probable
distribution of a species with
Niche of Occurrence (NOO)
in ModestR

What do you need for this tutorial:
1. ModestR v.5.6 or higher installed
2. Internet connection
3. About 25 minutes
ModestR software can be freely downloaded from http://www.ipez.es/ModestR

We’ll describe how to calculate species
distribution with Niche of Occurrence
(NOO2D) in ModestR.
Follow the next steps!

To calculate species distribution, we’ll need species occurrences. You can download them from GBIF (see tutorials 1 and 2
in http://www.ipez.es/ModestR/Manual_Tutorial.html). You can also import them from a CSV file. For this tutorial we’ll
later provide a sample CSV file with a virtual species generated using the “virtualspecies” package from R, and the
environmental variables from WorldClim.
They can come from GBIF, or from a CSV file like this one

To calculate species distribution we’ll also need the environmental data we want to work with.
There are different datasets freely available, such as WorldClim, NOOA’s World Ocean Atlas or Copernicus
For this tutorial we’ll use WorldClim data version 2.
To download and integrate these data in ModestR, follow
“Tutorial 14- Adding WorldClim data to ModestR”
You can find it in the ModestR tutorials page:
http://www.ipez.es/ModestR/Manual_Tutorial.html

1) Select menu “Options / Preferences”
2) To use Spanish format (default):
 Decimal separator  “ , “
 Field separator  “ ; “
To use English format :
 Decimal separator  “ . “
 Field separator  “ , “
For this tutorial you’ll find a sample file “Single_Virtual_Species_Occurrences_for_Tutorial_20” among the sample files
provided with this tutorial. You can choose it between two CSV options: Format Spanish / Format English
You may select the format depending on the configuration in the "Options / Preferences" menu of the "CSV Exportation
Options" section of any of the ModestR applications.
For example in ModestR DataBase Manager you can find it here:

Download the file “Single_Virtual_Species_Occurrences_for_Tutorial_20” from next link:
CSV in Spanish Format / CSV in English Format
Then import it from MapMaker
2) Go to menu “Import/Occurrence data from
CSV file” then select the downloaded CSV file
1) Run ModestR MapMaker

When importing occurrences to ModestR, you have to select the valid habitats for the species
2) In the dialog box just
check the “Land” habitat
3)Then click on “Accept”
4) Import completed
Then click on “Accept”
1)Be sure of skipping the first column (species name)
and of adding Longitude and Latitude as the second
and third columns.
The columns to be imported and their order will
depend on the format for the specific CSV file
Pay also attention to select the correct decimal and
field separators, according to the CSV file format
(English-like format uses “.” and “,” respectively, while
Spanish-like format uses “,” and “;” respectively)

Once occurrences imported, ModestR checks the habitats and indicates when occurrences are in an invalid habitat for
the species using another color. Those occurrences are shown but are not used in any analysis or calculation
Invalid occurrences (here in black) are
occurrences located in habitats not valid for the
species. As we indicated that only land was valid
for this species, invalid occurrences are those
located in sea or freshwater habitats.

NOO can be used to calculate a density map that shows species suitability, or to calculate a distribution.
We’ll first start with the density map
1) Go to menu Mapping/Niche of occurrence/Density Map

Next step is creating or selecting an existing 2D or 3D CEL (Compounded Environmental Layer).
A CEL will bind together several environmental variables in a single multidimensional variable. In this case we’ll
create a new 2D CEL using the WorldClim data we have imported previously (see Tutorial 14)
1) Select “Create a new 2D CEL”
More details about what are compounded
environmental layers can be found in ModestR User’s
Manual on ModestR website.
A tutorial about compounded environmental layers
creation can also be found at
2) Click “Accept”
If other previously created and stored
CEL’s exist, they will be show here
A 2D CEL has only longitude and latitude dimensions.
For 3D CEL’s that can include a 3th dimension such as
time or depth, see tutorials about 3DNOO in

When creating a 2D CEL you have to select the whished variables you want to work with.
2) For this example, in the
dialog box that will be shown,
select all the variables from
WorldClim
3) We’ll change the
geographical extent (GE) to
Extent of Occurrence (EOO),
to limit the distribution to be
calculated to the EOO of the
current species
You can use “Preview” button to see the EOO in
the world map (you probably have to move this
window to be able to see the world map below)
4) Click on “Continue”
We’ll keep the default mode to
calculate EOO, which is the
alpha shape (see “Tutorial 4- Data
cleaning and hulling” for more details
about EOO calculation)

5) Click on “Calculate VIF “
In the next step you can calculate VIF (Variance Inflation Factor)
of the variables to detect collinearity and, optionally, delete
some variables that may be redundant

You can select a variable an delete it from the list.
VIF will be recalculated with the remaining variables. You
can continue deleting other variables, until VIF is considered
acceptable (typically when VIF<30)
For this example we’ll keep only the variables
wc2.0_bio_5m_02, 03, 04, 08, 09, 12, 14, 15, 18 and 19.
We’ll delete the other variables.
VIF of each variable is shown
VIF >30 is shown in red.
6) Select a variable and click on “Delete selected“
Note.- to select several variables keep Ctrl key pressed while selecting.

In the next step you can calculate the relative relevance of each variable regarding the species distribution using a
Instability Index (see https://doi.org/10.1016/j.ecoinf.2016.11.004 for more information).
This can help you selecting the most relevant variables for the species
8) Click on “Analyse” and Contribution
of each variable will appear
When you press on “Analyse”
the most relevant variables
will be automatically selected.
This step is not mandatory. You can just click on “Continue”
without performing this analysis neither the variable selection.
In that case all the variables will be used.

In the next step variables are optimally ordered in the polar space using they correlation matrix.
You can eventually reorder variables, but the most of the time you’ll just continue

In the next step, the 2DCEL is generated and shown in the polar coordinates system
11) Click on “Show in map” to show the CEL in the world map.
Take into account that a CEL contains only environmental variables. It will
be employed to calculate the species distribution, but it is not related
with a specific species.

In the last step, you can optionally save this 2D CEL to be used later. Or you can just load it for the current work
session. For this example we’ll save it to use it later
1) Enter a name
2) Click on “Save layer”
3) When save completed,
click on “Continue”

2) Click “Run” to calculate density
1) We’ll leave “Smoothing” to 1.
But you can also test increasing it to 2 or 3
Once a 2DCEL selected or built, we can perform the NOO2D using this CEL and
the species data.
NOO2D uses a kernel density function to calculate the suitability of the
species. Therefore a smoothing parameter can be adjusted to take into account
the environmental tolerance of the species. The more a species is tolerant, the
higher we may adjust smoothing. Typical values range between 1 and 3
A NOO2D has only longitude and latitude
dimensions. For species distribution niche in 3D, that
can include a 3th dimension such as time or depth,
see tutorials about 3DNOO in

Results will be displayed both in the world map, showing geographic density, and in the dialog, showing polar density map
Use the Clear button to delete
current results, change parameters
(such as smoothing) an run again
the NOO calculation.
Suitability and density rasters can
be saved to files using this button

Going back to MapMaker, you can move the dialog box to be able to show the world map, and see the species
suitability.
1) Drag the dialog box to a side
to be able to see the world map
You can see in the world map the suitability

Close the dialog box and go back to main MapMaker window. If many occurrences are shown on the map, they may
sometimes hide parts of the density map. Try to hide them to see the whole density map.
Uncheck valid and invalid occurrences to only
see the density map.
Then check them again to show the occurrences
before continuing to the next step.

We’ll now calculate Distribution map
1) Go to menu
“Mapping/Niche of occurrence/ Distribution map”

Thanks to having saved previously the 2DCEL we built, we’ll now just have to select it to calculate NOO2D distribution
2) Select the previously saved CEL
3) Click on “Accept”

5) Click on “Run”. Calculation will start
Range can be “Full layer” or “EOO”. In this case we can
leave Full layer as the 2DCEL we are using is already
limited to the EOO of the species
“Add area” option (by default) indicates that the
estimated area will be added to the distribution map. The
other option, “Hull simulation”, adds just a shape only
for visualization purposes, not as part of the species
distribution
4) Leave “Smoothing” to 1
The parameters to set are the almost the same than for the density map

If needed click on “Clear raster”
button to clear density map and see
the species distribution.
The distribution of the species is
shown in the map.
You may save this map in a database
or as a standalone file.

The procedure previously described is useful when working with one or few species, and when we want to see the results
of each step, test several parameters, etc.
But when working with many species, it may be more useful to use DataManager application from ModestR, that
allows working with databases of hundreds or thousands of species. In DataManager you can perform NOO2D in batch
mode and obtain for example the distribution of many species. In the next steps we explain how.

Download database “VirtualSpecies for tutorial 20.MODESTRDB” from THIS LINK.
Then open DataMananager and open this database.
1) Select menu “File / Open Database” and then
select the downloaded database file

The first step is processing all maps if they still aren’t (that is, converting them to an internal format that allows ModestR
to work with them more efficiently)
3) Click on “Process button”, or on menu
“Process/Process checked pending maps”
2) “Check all” the species checking the
root node checkbox
4) For this tutorial we can uncheck all
processing options, in order to make it fast
5) Click “OK”

If map processing is needed, it will start immediately. If it is not needed, a message will indicate that no maps have to be
processed (the necessity of map processing depends on the ModestR and world map versions you have).

Now we are ready to calculate NOO2D for all the species in the database
2) Go to menu “Mapping/Apply niche of
occurrence to checked maps/Using a per-
species layer/2D”
1) “Check all” the species
checking the root node checkbox
The “per-species layer” option will select the more relevant variables for each
species, create an on-the-fly 2D CEL with them and calculate the NOO2D of the
species. That are the same steps we already explained at the beginning of this tutorial,
but interactively in MapMaker. Here they will be done in batch mode.
The other option, using a “stored layer” will use the same 2DCEL, previously built and
stored in MapMaker, to calculate the NOO2D for all species.

Now we are ready to calculate NOO2D for all the species in the database.
We’ll use the same options we already used in MapMaker
1) For this example we’ll use
only the variables:
- wc2.0_bio_5m_02, 03, 04,
08, 09, 12, 14, 15, 18 and 19.
2) We’ll change the geographical
extent (GE) to “Extent of Occurrence”
(EOO), to limit the distribution to be
calculated to the EOO of the current
species.
We’ll keep the default mode to calculate
EOO, which is the “Alpha shape”
(See “Tutorial 4- Data cleaning and hulling”
for more details about EOO calculation)

In the next step you can calculate VIF (Variance Inflation Factor) of
the variables to detect collinearity and, optionally, delete some
variables that may be redundant.
But as we already have done this step in MapMaker and then we
now have just selected the variables with a low VIF, we can just
continue.

In the next step you can set how the relative relevance of each variable regarding the species distribution will be
calculated (see https://doi.org/10.1016/j.ecoinf.2016.11.004 for more information). We already described this step in
MapMaker. The difference is that here we’ll not see the results immediately. They will be calculated for each species
when the batch process will be started.
Here this step is mandatory, as we selected the option of using a “per-
species layer” to calculate the NOO of each species.
The “per-species layer” option will select the more relevant variables
for each species, create an on-the-fly 2D CEL with them and calculate
the NOO2D of the species. That are the same steps we already
explained at the beginning of this tutorial, but interactively in
MapMaker.
The other option, using a “stored layer” will use the same 2DCEL,
previously built and stored in MapMaker, to calculate the NOO2D for all
species. Thus, it will not apply the step to select most relevant variables
for each species, as the same 2DCEL will be used for all of them.

In the next step you can adjust NOO2D parameters such as smoothing, calculation range and other ones.
2) Click on “Start”
1) Check “Export detailed data”. This will export a density
map as an image, as well as density and suitability rasters
for each processed map.
Suitability rasters are the typical output of SDM’s, and they
may be used for later analyses, such as estimating species
composition in one or more regions with KnowBR and
ModestR (see “Tutorial 18- Estimating species composition in one
or more regions with KnowBR and ModestR” for more details)

Select a folder to save detailed result data

1) Click on “Continue without backup”
NOO2D will calculate species distributions and save them in the current database. Therefore, ModestR will propose you
to make a backup of the current database before continuing. But as this is just a sample database, we’ll continue
without doing backup.

NOO2D processing will start, This can take some time!
1) Click on “Si”/”Yes”
2) Click on “Open file” to open the saved
report. As it is a CSV file, it is usually
opened in Excel or the Notepad.
Once ended, you can
save a report about
the process. Save it!

This report contains the contribution (relative importance) of each variable for each species, the selected
variables used to calculate NOO for each species, the AUC, the Minimum Training Presence Threshold (MTPT)
and some other values for each species

Now go to the folder you selected to save detailed results.
You will find those subfolders:

“CEL_maps” contains images of the 2DCELs created for each species

“Density_rasters” contains rasters of the density map for each species
To display a raster in MapMaker, go to menu
“Layers/Rasters/Load raster from ESRI ASC-GeoTIFF file”

“Image_maps” contains images of the species distributions. Anyway, species distribution maps can be
exported as images from MapMaker, with the advantage of being able to zoom specific areas, add grids, etc.

“Polar_maps” contains images of the polar maps of each species

“Suitability_rasters” contains rasters of the suitability map for each species
Suitability rasters are the typical
output of SDM’s, and they may
be used for later analyses, such as
estimating species composition in
one or more regions with
KnowBR and ModestR (see
“Tutorial 18- Estimating species
composition in one or more
regions with KnowBR and
ModestR” for more details)
To display a raster in MapMaker, go to menu
“Layers/Rasters/Load raster from ESRI ASC-GeoTIFF file”

Going back to DataManager, now the species maps have been modified, because the result of the NOO has
been added to each species.
1) “Right-click” over any species
and select “Open map in
Mapmaker” to see the resulting
map after the NOO

Uncheck valid and invalid
occurrences to only see
the areas added to the
map by the NOO process
Those areas have been added
as result of the NOO

Here we explained how to perform NOO2D. But NOO can also be performed in 3D, including a 3th dimension such as
time or depth. See tutorials about 3DNOO in http://www.ipez.es/ModestR/Manual_Tutorial.html, such as:
• Tutorial 10- Creating environmental layers 3D
• Tutorial 11- Calculating species niche for different periods with ModestR
• Tutorial 17- An example using ModestR to assess species distribution in different climatic change scenarios
• Tutorial 19- Calculating 3D species niche in marine habitats with ModestR

It was the Step by step tutorial:
Calculating species niche with Niche of
Occurrence (NOO) in ModestR
Thank you for your interest.
You can find this one and other tutorials in http://www.ipez.es/ModestR
By the ModestR team

20. Calculating most probable distribution of a species with Niche of Occurrence (NOO) in ModestR (Version ModestR 5.6 or higher)

Recommended

Recommended

More Related Content

What's hot

What's hot (12)

Similar to 20. Calculating most probable distribution of a species with Niche of Occurrence (NOO) in ModestR (Version ModestR 5.6 or higher)

Similar to 20. Calculating most probable distribution of a species with Niche of Occurrence (NOO) in ModestR (Version ModestR 5.6 or higher) (18)

Recently uploaded

Recently uploaded (20)

20. Calculating most probable distribution of a species with Niche of Occurrence (NOO) in ModestR (Version ModestR 5.6 or higher)