Download BIO 263: Niche Modeling Project_all_2016

Environmental Niche Modeling Exercise
Environmental Niche Modeling Exercise
Climate change is expected to have significant impacts on species’ ranges by altering the
distribution of key climatic variables that determine where species can live. Environmental
niche modeling (ENM) is a statistical tool for estimating which environmental variables
are important determinates of species’ ranges and projecting species’ geographic
distributions based on those environmental variables. ENM can also be used to estimate
how species’ geographic ranges will respond to climate change, by using projections of the
geographic distribution of future climates and estimating the probability that species’ will
occur in that climate.
The goal of this module is to predict present and future species ranges by constructing and
comparing “environmental niche models” for single and interacting species under current
and future climate scenarios.
This exercise is divided into two parts. The goal of Part 1 is to compare the present range
of a species to its predicted future range. Although you can work in a group, each student
will prepare a separate write-up presenting the analysis of one species.
The goal of Part 2 is to predict how the overlap in ranges among interacting species is
predicted to change in future climates. For this portion, students will work in groups.
Group members should coordinate their choice of species to analyze according to a
hypothesis regarding how those species interact. Groups will compose one write-up and
one oral presentation of their results.
By the end of this module you should be able to:
Part 1:
a. Articulate and test a hypothesis about a biological responses to climate change
for a particular species
b. Construct an environmental niche model of a species and use it to project a
species distribution in the future.
c. Evaluate the strengths and limitations of the environmental niche modeling
approach while interpreting your results.
Part 2:
a. Articulate and test a hypothesis about a biological responses to climate change
for interacting species
b. Synthesize information across species and consider how interactions between
species may mediate their responses to climate change
c. Evaluate the strengths and limitations of the environmental niche modeling
approach while interpreting your results.
1
Environmental Niche Modeling Exercise
The following are the major components of the exercise:
Choose your species
Download Software (R and Maxent)
Download and prepare species occurrence data
Download current and future climate data (to be supplied to you)
Analysis Part 1: Conduct niche modeling with Maxent, calculate ranges using R,
view ranges in Google Earth.
Analysis Part 2: Calculate species overlap in present and future using R.
The following are the assignments to be turned in:
Niche Modeling Report 1: Changes in a species range in the future
Oral presentation: Changes in species ranges and in species overlap
Niche Modeling Part 2: Changes in species overlap in the future
Descriptions of each step and assignment follows.
2
Environmental Niche Modeling Exercise
CHOOSE YOUR SPECIES
First, choose your group, and choose an assemblage of 3-4 species (one per group member)
that have a well-documented biological interaction. Articulate approximately three subhypotheses that relate to that species assemblage.
a. A sub-hypothesis concerning which bioclimatic variables may be the most
important predictor of the distribution for each species
b. A sub-hypothesis concerning how each species range sizes and/or locations will
change in the future.
c. A sub-hypothesis concerning how species ranges will change in relation to each
other and in terms of niche overlap. Articulate a hypothesis for specific
interactions among species in your assemblage; there is no need to analyze all
combinations of species.
Once you have come up with your biologically motivated hypothesis, each group member
will be responsible for producing MaxEnt analysis for one of these species (Part 1). Then
the group will combine their output to calculate predicted changes in range area and niche
overlap between species and over time (Part 2).
3
Environmental Niche Modeling Exercise
DOWNLOAD SOFTWARE
1) Download R
First, if you don’t already have it you will need download and install R. If you have
Rstudio (a gui for running R) on your computer already, you should be able to use
that instead. If you don’t have R download it from the following website:
http://mirrors.nics.utk.edu/cran/
Once you have selected your operating system, there will be detailed instructions on
how to install this free software. R is a programming language designed primarily for
statistical analysis. It operates by command-line. Once you have installed R, open it
up and install the following packages for niche overlap analysis by typing these
commands into the R Console and hitting return.
install.packages("ape")
install.packages("phyloclim")
install.packages(“maptools”)
install.packages(“sp”)
Next see if you can successfully load these packages into R by typing in the following
commands:
library(ape)
library(phyloclim)
library(maptools)
library(sp)
If these don’t give you an error, you should be all set!
4
Environmental Niche Modeling Exercise
2) Get Maxent set up to run on your computer
We will use the species distribution modeling (ENM) platform Maxent to predict
current and future species distributions from species occurrence data.
Maxent requires the software Java (version 1.4 or later- the most recent version is
JRE8). The Java runtime environment (JRE) can be obtained from
http://www.oracle.com/technetwork/java/javase/downloads/index.html. Click on
the JRE download link in the bottom right corner and corner and pick the version
appropriate to your computer.
You can download the maxent software from a link in the same gbif site you have
already explored http://www.gbif.org/resources/2596. You will want to download
the most recent version of Maxent v3.3.3k. After accepting the terms of agreement,
you will want to download the ‘zip’ file which contains both the maxent.jar (Mac
OSX) and maxent.bat (Windows) files and a readme.txt file that includes
troubleshooting instructions if you are having trouble.
Open Maxent by either clicking on the maxent.bat file (Windows) or clicking on the
maxent.jar file (Mac OSX). The following screen will appear:
If you see this screen appear. You are all set. If not, come talk to me on Monday.
5
Environmental Niche Modeling Exercise
DOWNLOAD SPECIES OCCURRENCE DATA
1) Use GBIF to download global occurrence data for your species of interest.
GBIF is a website that provides free access to a myriad of biodiversity data. Essentially it
catalogues where in the world a species has been found. It currently has data for almost
1.5 million species.
First go to the website: http://www.gbif.org
You will need to register, create a log-in, and then confirm your email address. This is
not so they can send you ads; it is so they can track who is accessing their data and so
they can send you your data file when they finish compiling it. If you do not want to
register for this tool come talk to me and we can find a different way for you to access the
data you need.
After you have logged in go to: http://www.gbif.org/species
In the search bar, type in the scientific name (Genus and species) of one of the organisms
you are interested in. After you hit search you will be shown a list of matches to that
organism. In grey letters above the blue species name you should see words like
“accepted species” or “sub-species”. Make sure you choose a search result that
corresponds to an accepted species unless you are explicitly interested in two sub-species.
Below the blue species name there is the rest of the taxonomic groupings for your
organism. Make sure the higher taxonomic groups refer to the type of organism you are
interested in. For instance many mosquito borne viruses have the Latin name for
mosquito in them and they will pop up in a search for Aedes aegypti.
Once you have determined the correct search result click on the blue species name. You
will see an overview of your species and also a map that shows the geo-referenced data
available for your species. You can get a feeling for how many georeferenced data points
there are based on the blue number under VIEW RECORDS to the right of the map. If
fewer than 100 records you may need to pick another organism. If more than 10,000
records show up, we may need to trim your dataset size to reduce model run length. Talk
to me if you are interested in using a species with more than 10,000 records.
Now you are ready to obtain the data files. Click on the blue button at the upper right of
the page that says “view occurrences”. Now click on the blue letters on the right hand
side of the screen that says “Add a filter”. Use the dropdown menu to select “Location”.
On the next page click the checkbox that says “georeferenced” and the check box that
says “with no known coordinate issues”. Then hit the blue botton that says “Apply”. This
will take out all records that do not have a latitude and longitude or those where there is a
problem with the latitudes and longitudes listed. Now click the button that appears in the
upper right hand corner that says “download”.
You will see a message that your download has started. You will receive an email with a
link to your data when it has finished being compiled for you. Click the link in the email
and save the data to a new folder you create called “input files”. Name it something like
6
Environmental Niche Modeling Exercise
“Species1_raw” Unzip the file. Click on the folder and then open the file occurrence.txt
in Excel.
2) Reduce your data set if necessary
Some of you may have data sets that are too large to run MaxEnt effectively on your
computer. The goal is to have 10,000 data points (or more), but depending on your
computer capabilities you may need to cut down your dataset. For some computers, there
may be diminishing returns above 1,000 occurrence records. If you need to reduce your data
set, you need to cut out a random sample of occurrences, not a biased sample. Below are
instructions for doing this in R and in Excel.
In R run the following code after setting your working directory to the folder that holds
your occurrence data.
data<-read.csv("Exampe_Aedes_data.csv") #read in your data
head(data) # look at your data file
data$random.num<-runif(dim(data)[1]) # create a new column
with a random number in it
data<-data[with(data, order(data$random.num)), ] #reorder the
data based on the random # column
data<-data[1:2000,] #Take the first 2,000 rows of data. You can
change this value to subset more or less rows eg. data[1:1000,]
data<-data[,1:3]
write.csv(data,"Example_Aedes_data_trimmed.csv") #write a new
.csv file of your data
In Excel:
Create a new column to the right of latitude in your occurrence data .csv file. Call it Random
# or something similar. Type =RAND() into the cell below the column header and drag down
the entire column till the end of your data. Highlight this whole column and copy it
(command is in the edit menu). While the column is still highlighted, go to the edit menu,
click paste special, and click "values". This should get rid of the formula defining the cell and
make it a stable number value.
Next sort your data based on this value. Highlight ALL your four columns of data and find the
sort option (likely in the data menu) in excel. Sort based on the Random # column (should be
column D) and make sure the check mark is checked beside the comment "my list has
headers". Click okay. This should reorder all your rows based on the value in the Random #
column.
Now simply scroll down to the row number that corresponds to the # of records you want to
keep. 2,000 probably is reasonable. Delete all cells below this value and delete the Random #
column and save your .csv file.
3) Format your species data for use in MaxEnt
All we need are the columns ‘species’, ‘decimalLatitude’, and ‘decimalLongitude’. Delete
the rest of the columns.
7
Environmental Niche Modeling Exercise
You will need to reformat the raw data downloaded from GBIF to work in MaxEnt. First,
rename the “decimalLatitude” column “latitude” and the “decimalLongitude” column
“longitude” Second, copy and paste the ‘Longitude’ column to the left of the ‘latitude’
column and the “species” column so it is on the far left-hand side. The columns should
now be in the order: ‘species, ‘longitude’, and ‘latitude’. Next, check the species names for
names that do not match your species name. Delete any a record (i.e., the delete the row)
with a species name that does not match your original species’ name. Next, delete any
record that is missing either longitude or latitude data (hopefully we already removed all
of these with the filter).
Save your reformatted Excel file as a comma separated values (.csv) file (File > Save as…
> Format: .csv) not in the default excel format (.xlsx) or as a text file (.txt). If it says that
you may lose data if you save the file in this format, that is fine, click Okay. If you open
your new file in a text editor like Notepad, it should like this:
species,longitude,latitude
Cakile_edentula,-2.59,37.75
Cakile_edentula,-0.1,40.1
Place this file in a new folder you name “input_files”. I have put an example of this file
format in the Niche Modeling folder.
8
Environmental Niche Modeling Exercise
DOWNLOAD CURRENT AND FUTURE BIOCLIMATIC DATA
I will send these data to the class you via an email. It will either be sent via an attachment
or google drive if the file size is too large for the Sakai email server. There will be two
zipped files “current_bioclim” and “future_bioclim”. Put these two files in your
“input_files” folder where you placed your final species.csv data file. Unzip the folders
and delete the .zip files.
The files contain the standard summaries of Temperature and Precipitation measures
called Bioclim variables. This data for current day is available online at
http://www.worldclim.org/download. Note that the mean of moisture and temperature
may not be the only aspect of climate that influences organismal behavior. Below is the
description of all 19 standard Bioclim variables.
BIO1 = Annual Mean Temperature
BIO2 = Mean Diurnal Range (Mean of monthly (max temp - min temp))
BIO3 = Isothermality (BIO2/BIO7) (* 100)
BIO4 = Temperature Seasonality (standard deviation *100)
BIO5 = Max Temperature of Warmest Month
BIO6 = Min Temperature of Coldest Month
BIO7 = Temperature Annual Range (BIO5-BIO6)
BIO8 = Mean Temperature of Wettest Quarter
BIO9 = Mean Temperature of Driest Quarter
BIO10 = Mean Temperature of Warmest Quarter
BIO11 = Mean Temperature of Coldest Quarter
BIO12 = Annual Precipitation
BIO13 = Precipitation of Wettest Month
BIO14 = Precipitation of Driest Month
BIO15 = Precipitation Seasonality (Coefficient of Variation)
BIO16 = Precipitation of Wettest Quarter
BIO17 = Precipitation of Driest Quarter
BIO18 = Precipitation of Warmest Quarter
BIO19 = Precipitation of Coldest Quarter
The future bioclim variables come from a climate model using the RCP 4.5 (the purple
line) carbon emission scenario. This scenario assumes peak carbon emission in 2040
followed by a subsequent decline. You can read more about these new scenarios created
for Assessment report 5 here:
http://en.wikipedia.org/wiki/Representative_Concentration_Pathways
9
Environmental Niche Modeling Exercise
"All forcing agents CO2 equivalent concentration". Via Wikipedia
Note that the temperature data are in °C * 10. This means that a value of 231
represents 23.1 °C. This does lead to some confusion, but it allows for much reduced
file sizes which is important as for many downloading large files remains difficult.
The unit used for the precipitation data is mm (millimeter).
The future prediction is for approximately 2060-2080.
YOU ARE NOW ALL SET TO CONDUCT THE ANALYSES!!!
10
Environmental Niche Modeling Exercise
NICHE MODELING PART 1
PREDICTING THE NICHE OF YOUR SPECIES USING MAXENT
Open MaxEnt by either clicking on the MaxEnt.bat file (Windows) or clicking on the
MaxEnt.jar file (Mac OSX). The following screen will appear:
First, import your species geospatial data. Click the ‘Browse’ button under samples and
select your species .csv file. After you select it, your species name should appear in the
upper left box. Select your species by checking the box beside it.
Second, import the climate data. Click the ‘Browse’ button under ‘Environmental layers’
and select the folder where you have saved your current_bioclim folder (this should contain
both .bil and .hdr files). A list of all the variables will appear in the right panel. Make sure
all of the variables are selected so that they are included in the analysis.
To determine the environmental factors that predict the current geographic distribution of
your species, do the following. Select the boxes ‘Make picture of predictions’, ‘Do
jackknife to determine variable importance’, and ‘Create response curves.’ The first option
will create a heat map of each species’ predicted distribution based on their associated
climate variables; the colors indicate the probability of the species occurring in a given
location. The second option will determine which of the climatic variables best explain the
11
Environmental Niche Modeling Exercise
distribution of a given species. The third option quantifies how the probability of a species’
occurrence changes as a function of each environmental variable.
Designating a percentage of random observation points as ‘test points’ helps the model
‘train’ its predictions. What this means is that it first uses observed climate data and
presence/absence data to characterize the combinations of climatic factors in the locations
in which a species is present, and it then estimates parameters that describe the strength of
the relationship between the factor and the probability that the species is present. It then
uses the rest of the data to test how well the parameters estimated from the “training” data
set actually predicted the probability that a species was present in the locations not used on
the training set. We will designate 25% of your observations to be used as test points in
the training data set. Click the ‘settings’ button and enter 25 next to “Random test
percentage.” Close the settings dialogue box.
To project the geographic distribution of your species given future climatic conditions, do
the following: In the bottom right corner define the “projection layers directory/file” by
selecting the folder that contains the data that will be used for projecting your species into
the future: future_bioclim. This folder contains the .asc files.
Finally, select the folder where you want the output from the model run to be stored. The
box for this is labeled “Output Directory”.
To run your model, press the “Run” button. A progress monitor describes the steps being
taken. This may take awhile depending on your computers capabilities and how much
occurrence data you have for your species.
INTERPRETING RESULTS FROM MAXENT
MaxEnt outputs a series of files encapsulating numerous results. This output is summarized
in one .html file. Open the .html file. I will outline some of the main output below, but
there is also a MaxEnt tutorial posted in the resources/assignments/niche modeling folder
that describes the output in more detail.
Pictures of the Model: This image, like all distribution maps produced by MaxEnt, uses
colors to indicate the predicted probability that conditions are suitable for the species, with
red indicating a high probability of suitable conditions, green indicating conditions typical
of those where the species is found, and lighter shades of blue indicating low probability
of suitable conditions.
Below this picture, is a projection for your species in the future. Click on the picture to see
a bigger, interactive version.
If you want to copy these images, or want to open them with other software, you will find
the .png files in the directory called “plots” that have been created as an output during the
run (These will be an important component of your results).
12
Environmental Niche Modeling Exercise
The maps below these first two show areas of the globe where the future predictions may
not preform well because they are “no analogue” climates. Because these future climate
combinations do not occur currently, it is hard to infer how species will respond from their
current distribution. This is related to the classic problem of extrapolating beyond the range
of your data. Read the descriptions of the various measures in the html file.
Analysis of variable contributions
MaxEnt provides three main tests that an environmental factor predicts a species’ presence.
The first is the “percent contribution,” which indicates the percentage of the variation in
the presence/absence of the species that can be explained by variation in the environmental
factor.
Second, the “permutation importance” is based on a method called “permutation.”
MaxEnt first estimates the strength of the relationship between the presence of a species
and an environmental factor using the training test data points. Next, it scrambles the
values of the environmental factors in that training data set, and re-calculates the strength
of the relationship. If the environmental factor is actually important in predicting a species
presence, then the strength of the relationship between the species presence and that factor
would be higher in the real data set than in the scrambled data set. The “permutation
importance” is a measure of how much the strength of the relationship changes between
the real and the scrambled data sets. The larger the change, the more “important” that
factor is in predicting the species’ presence.
Take note of the three bioclim factors that were identified as important predictors of species
presence/absence through each of these two methods. Look up what those bioclim variables
actually stand for http://www.worldclim.org/bioclim
Percent contribution
Permutation importance
1. _____________________________
____________________________
2. _____________________________
____________________________
3. _____________________________
____________________________
Are these variables the same for the different species? Are they in the same order for the
different species?
Third, while not required for your write-up, consider the results from the jackknife test.
These tests identify the trait that provides the “most useful information by itself” to the
model and the trait that provides the “most information that isn't present in the other
variables.” These can be found listed in the paragraph above the graph labeled “Jackknife
of regularized training gain for Your Species”. [A jackknife analysis is a way to measure
the consistency of a relationship (in this case, between the probability of species’
occurrence and an environmental factor). It randomly deletes observations from the data
set and re-calculates the parameters, and it does this hundreds of times. This gives a range
13
Environmental Niche Modeling Exercise
of estimates of the parameters: one for each data set. This is used as an estimate of the
consistency (or lack thereof) of the relationship. ]
How do these traits compare to the traits identified above as being the most important
variables?
You see here that MaxEnt gives three ways of measuring the strength of the relationship
between an environmental factor and the probability of species’ occurrence. These are
complementary statistical approaches that give slightly different information about the
importance of the environmental factor and the degree of consistency with which it predicts
species’ presence, and researchers often present all three.
Response curves
Finally, examine the of the three traits you identified by their ‘permutation importance.’
Look at the second set of graphs, not the first. Describe how the logistic prediction (y-axis,
which is a measure of the probability of the species occurring) changes with the variable
in question (x-axis). For example, ‘bio1’ is mean annual temperature and ranges from -256
(-25.6 C) to 302 (30.2 C). Under what values of an environmental factor (e.g. mean annual
temperature) is your species most likely to occur? Which environmental factor(s) have a
large effect on the probability of the species occurring?
In your report, present these statistics and describe how environmental factors identified
by these methods as significant might be important to your species. Do they make sense
based on your understanding of the biology of your species?
ANALYSIS OF DATA IN “R”
You will need to use “R” to calculate range sizes and overlaps in species ranges.
Each person will have two files in the folder that represent their species. The first is the
niche predictions under current climate conditions and the second is the expected niche in
the future. The first will be an .asc file with just the species name. The second will have
the species name with the name of the future climate folder appended. Here is an example.
Aedes_aegypti.asc
Aedes_aegypti_future_bioclim.asc
Detailed explanations for the R code below can be found in the
“NicheModelingDemoRcode.r” file in the Sakai Niche Modeling folder. I recommend
using this file as a template for your R analysis.
Open R and change the ‘working directory’ to be the folder that holds the species data.

If you are using the program R directly, this is under the heading ‘Misc’. Click
“change working directory” and navigate to the folder where you saved your
MaxEnt outputs.
14
Environmental Niche Modeling Exercise

If you are using Rstudio, go to the heading “Session”, click “set working directory”,
then “choose directory”. Navigate to the folder where you saved your MaxEnt
outputs and click “choose”.
If you are starting a new R session, load in the four R packages you installed earlier.
library(ape)
library(phyloclim)
library(maptools)
library(sp)
CALCULATING SPECIES RANGE AREA IN R
Load the output file for your species into R. Note that you will need to replace the file
name in quotes in the R code below to reflect the name used for your file.
Species1.current<-readAsciiGrid("Species1.asc")
We now want to estimate the area of the geographic range of your species (in the present
and the future). We can very roughly approximate the range area for each species using
the knowledge that each of our grid cells reflects approximately 10 km2. To put your ranges
in perspective, the whole of earth’s surface is about 510 million km2 and the land area of
the earth is approximately 150 million km2. Copy the following code into R.
Species.raster<-raster(Species1.current)
plot(Species.raster)
Species.raster<-getValues(Species.raster)
Species.raster<-Species.raster[!is.na(Species.raster)]
range.size.km2<length(Species.raster[Species.raster>.33])*10
range.size.km2
prop.landcover<-range.size.km2/150000000
prop.landcover
prop.landcover*100
Repeat this process for your future projection for that species. Record the current and future
range areas for your species and present this in your report. Make sure to indicate how
pervasive non-analogue climates were in your simulations.
15
Environmental Niche Modeling Exercise
PLOTTING OCCURRENCE DATA IN GOOGLE EARTH
You may want to view geo-spatial location data in Google Earth. Unfortunately, Google
Earth does not read or import .csv files of lat./long. data. Instead, we will need to convert
our data to a .kml file to open in Google Earth. Here is how to do that.
1. Format a .csv file for conversion. The format is as follows:
Geo,,,
Name,Longitude,Latitude
SpeciesA,-76.41,37.37045
SpeciesA,-76.33,37.68717
SpeciesA,-76.308,37.81964
SpeciesA,-76.51,38.02398
SpeciesA,-76.512,38.5175
SpeciesA,-76.523,38.6422
SpeciesA,-76.495,38.88219
(etc.)
Note how this is slightly different then the .csv format for MaxEnt. In particular, you need
the 'Geo' header.
2. Second, you need to convert the file using the following website:
http://ingeapps.com/apps/online/kml-file-creator
On the webpage, under Filename, choose your .csv file. Next, click 'create file'. A file titled
'file.kml' will automatically download.
3. Open Google Earth. Under File, select Open and then select the downloaded 'file.kml'
file. You're points will be marked with yellow pins.
16
Environmental Niche Modeling Exercise
NICHE MODELING REPORT 1:
Format the report in the style of a scientific article with an Introduction, Methods, Results,
and Discussion.

The Introduction should include the following. An overview of the main question
you are addressing; Background information on your species; Your hypothesis
about which climate variables are expected to most strongly determine the range of
your species; Your hypothesis regarding how species range area and distributions
may change in response to climate change.

The Methods section should describe how you characterized the niche of your
species. You do not need to present in detail how you did the analysis, as that is
already written in the instructions for the exercise. What you should include are:
the geographic range analyzed, the number of observations used, the climatechange scenario used, which environmental variables were analyzed; what method
you used to evaluate them (indicate which procedure in MaxEnt, but no need to
describe in more detail). Also indicate what specific comparisons you made, and
why. You should indicate which statistics/tests you used to address each of your
questions.

The Results section should provide a systematic and organized presentation of the
results, including the statistical information. Keep in mind the purpose of each
analysis and the question it addresses, and present your results in terms of answers
to those questions. Highlight your major findings. What climate variables matter
most? How does range size change with climate change? (Indicate how pervasive
non-analogue climates were in your simulations.) Be sure to describe the main
result in each table and figure referred to.

The Discussion should interpret your results in terms of how you expect climate
change to affect your species. Did you find what you expect? Did any of your results
surprise you? Discuss limitations of the analysis. Use this exercise to generate
hypotheses that can be tested with other methods. Articulate future questions and
hypotheses that emerged from your results.
Length:
There is no official page requirement for this. Consider 5 single-spaced pages without
the figures as an absolute maximum. A rough guideline is:
1-1.5 pages for the Introduction. Introduce your species, their interactions, and your
hypotheses.
1 page for the Methods. Describe the species used, sample sizes, sources of data, and all
comparisons conducted. You can refer to the appropriate software, but DO NOT
describe in detail the protocol followed in this hand-out.
1 pages for the Results, excluding tables and figures. Be sure to describe the main results
for each table and figure. Present your results according to your hypotheses.
1-1.5 pages for the Discussion. Interpret your results in terms of your hypotheses.
17
Environmental Niche Modeling Exercise
Citations and Bibliography:
This assignment will require outside research and citations of relevant literature to
support your hypothesis of how species respond to environmental factors and how your
species assemblage interacts. Use peer-reviewed literature, and avoid un-reviewed
information on the internet.
Bibliographies should be in the following format:
Last name, first initial, last name, first initial. Year. Title. Journal volume:page-page.
Citations in text should be in the following format and in chronological order:
(Author last name year; Author last name year)
If there are two authors, provide the last names of both. If there are more than two,
provide the first author and then “et al.” Examples:
(Smith 2010, Jones 2012, Black and Brown 2014; Green et al. 2015)
Grading:
Grading will be based on overall quality and on the following specific criteria:
1. Quality and motivation of the hypotheses
a. Hypotheses are clear in their causality
b. Clear and creative motivation of the sub-hypothesis concerning which
bioclimatic variables may be most important for the species
c. Clear and creative motivation of the sub-hypothesis concerning how the species
range size and location are predicted to change in the future
2. Correct identification of necessary comparisons
a. Articulated and included which bioclim variables are most important for the
species, along with relevant statistics
b. Included maps of the distribution of species at present and in the future
c. Included relevant tables or graphs of range size estimates for the species
d. Clearly articulated each major finding in Results section, and related the
results to the stated hypotheses
3. Appropriateness and creativity of interpretation
a. Demonstrated an understanding of and discussed cogently how to interpret
which bioclimatic variable are most important for the species
b. Clearly discussed changes in range and location of the species
c. Clearly discussed of the limitations of the data and the niche modeling analysis
4. Clarity of writing and presentation of results
a. Clarity of composition of sentences and paragraphs
b. Quality of organization
c. Major conclusions are clear
18
Environmental Niche Modeling Exercise
NICHE MODELLING PART 2
CALCULTING NICHE OVERLAP IN R
This portion of the analysis examines changes in the overlap of interacting species. Niche
overlap is a measure of the similarity of niches and is highly correlated with geographic
range overlap. You can use it to compare the similarity of the niches two species. You will
work as a group for this component.
First, combine the data files from all the members of the group into one folder on one
person’s computer that has successfully installed R. As before, load the output files into R
that you wish to compare. Note that you will need to replace the file names in quotes in
the R code below to reflect the names used for your files.
Current niche area overlap of Species one and Species two
First, you can calculate species area overlap in the present. Again, make sure your file
names match those used in R. Use the following code.
Species1 <-readAsciiGrid("Species1.asc")
Species2 <-readAsciiGrid("Species2.asc")
grid.list<-"NA"
grid.list<-as.list(grid.list)
grid.list[[1]]<-Species1
grid.list[[2]]<-Species2
NO.Sp1vsSp2<-niche.overlap(grid.list)
When you type in NO.Sp1vsSp2 a 2x2 matrix will appear containing two measures of niche
overlap: the D statistic (upper right value) and the I statistic (bottom left corner). These two
measures of niche overlap were proposed by Warren et al., 2008; read if you want more
information on these measures. Both statistics scale from 0 (no overlap) to 1 (complete
overlap). For the most part, the two statistics will give similar results. Report both in your
final write up. These values should be calculated and reported for all species.
Species 1 niche overlap with Species 2 in the future
In a similar fashion, you can calculate overlap for species pairs in the future. This analysis
is the same as for above but read in your future files from maxent instead of the current
files.
To load the prediction files for your species based on future climate models and calculate
their predicted niche overlap, use the following code.
Species1<-readAsciiGrid("Species1_future_bioclim.asc")
Species2<-readAsciiGrid("Species2_future_bioclim.asc")
grid.list<-"NA"
grid.list<-as.list(grid.list)
grid.list[[1]]<-Species1
19
Environmental Niche Modeling Exercise
grid.list[[2]]<-Species2
NO.Sp1vsSp2<-niche.overlap(grid.list)
Repeat this process for all species comparisons. Record the current and future range areas
for each of your species and present this as a table or graph in your report.
How does the degree of niche overlap between your species change over time? What could
be the biotic consequences of these changes? Make sure to indicate how pervasive nonanalogue climates were in your simulations.
20
Environmental Niche Modeling Exercise
CLASS PRESENTATION:
Each group will present a 10-minute presentation to the class. The following is a suggested
structure:
First introduce the species you chose and their relationship to each other.
Present a hypothesis for how climate change might alter their interactions.
Present your results as they pertain to each hypothesis.
Interpret your results in terms of how climate change might alter interactions among your
species.
Presentations will be graded according to the following criteria:
1. The time frame of the presentation is respected.
2. Information is presented in an organized way, and hypotheses and biological
motivations are clear.
3. Results are interpreted thoroughly and accurately.
4. Presentation includes some details but clearly emphasizes a few take home
messages.
5. Group uses visual aids such as the white board or power point effectively.
GROUP REPORT:
Format the report in the style of a scientific article with an Introduction, Methods, Results,
and Discussion. This report can use choice text from the individual reports (Part 1) as
appropriate, but be sure to integrate all the components thoroughly.

The Introduction should provide background on each species you chose and how
they interact. Introduce your first sub-hypothesis about which climate variables will
be most vital to determining the range of each species and how they may vary
among species. Next, motivate and introduce your second hypothesis regarding
how species range area and distributions may change for each species. Third,
motivate and introduce your hypotheses regarding how climate change may alter
the interactions (and niche overlap) between these species.

The Methods section should describe how you characterized the niche of your
species. You do not need to present in detail how you did the analysis, as that is
already written in the instructions for the exercise. What you should include for
each species are: the geographic range analyzed, the number of observations used,
the climate-change scenario used, which environmental variables were analyzed;
what method you used to evaluate them (indicate which procedure in MaxEnt and
R, but there is no need to describe in more detail). Also indicate what specific
comparisons you made, and why. You should indicate which statistics/tests you
used to address each of your questions.

The Results section should provide a systematic and organized presentation of the
results, including the statistical information. Keep in mind the purpose of each
analysis and the question it addresses, and present your results in terms of answers
to those questions. Highlight your major findings. What climate variables matter
21
Environmental Niche Modeling Exercise
most? How does range size change with climate change? (Indicate how pervasive
non-analogue climates were in your simulations.) What is the degree of niche
overlap for under current versus future climate conditions? Be sure to describe the
main result in each table and figure referred to.

The Discussion should interpret your results in terms of how you expect climate
change to affect your species and their overlap. Did you find what you expect? Did
any of your results surprise you? Discuss limitations of the analysis. Use this
exercise to generate hypotheses that can be tested with other methods. Articulate
future questions and hypotheses that emerged from your results.
Length:
There is no official page limit for this, but it should be as concise as possible. A rough
guideline is:
1.5-2 pages for the Introduction. Introduce your species, their interactions, and your
hypotheses.
1 page for the Methods. Describe the species used, sample sizes, sources of data, and all
comparisons conducted. You can refer to the appropriate software, but DO NOT
describe in detail the protocol followed in this hand-out.
1-2 pages for the Results, excluding tables and figures. Be sure to describe the main
results for each table and figure. Present your results according to your hypotheses.
2 pages for the Discussion. Interpret your results in terms of your hypotheses.
Consider 8 single-spaced pages without the figures as an absolute maximum.
Citations and Bibliography:
This assignment will require outside research and citations of relevant literature to
support your hypothesis of how species respond to environmental factors and how
species in assemblage interact. Use peer-reviewed literature, and avoid un-reviewed
information on the internet.
Bibliographies should be in the following format:
Last name, first initial, last name, first initial. Year. Title. Journal volume:page-page.
Citations in text should be in the following format and in chronological order:
(Author last name year; Author last name year)
If there are two authors, provide the last names of both. If there are more than two,
provide the first author and then “et al.” Examples:
(Smith 2010, Jones 2012, Black and Brown 2014; Green et al. 2015)
Grading:
Grading will be based on overall quality and on the following specific criteria:
1. Quality and motivation of the hypotheses
a. Hypotheses are clear in their causality
22
Environmental Niche Modeling Exercise
c. Clear and creative motivation of the sub-hypothesis concerning which
bioclimatic variables may be most important for the species
d. Clear and creative motivation of the sub-hypothesis concerning how the species
range sizes and locations will change in the future for each species
e. Clear and creative motivation of the sub-hypothesis concerning how species
overlap will change in the future
2. Correct identification of necessary comparisons
a. Articulated and included which bioclim variables are most important for each
species, along with relevant statistics
b. Included maps of all species at present and in the future
c. Included relevant tables or graphs of range size estimates for each species
d. Included appropriate comparisons of niche overlap to test the stated
hypotheses
e. Clearly articulated each major finding in Results section, and related the results
to the stated hypotheses
3. Appropriateness and creativity of interpretation
a. Demonstrated an understanding of and discussed cogently how to interpret
which bioclimatic variable are most important for each species
b. Clearly discussed changes in range of each species
c. Clearly interpreted the predicted changes in ranges and niche overlap between
species
d. Clearly discussed of the limitations of the data and the niche modeling analysis
4. Clarity of writing and presentation of results
a. Clarity of composition of sentences and paragraphs
b. Quality of organization
c. Major conclusions are clear
23
Environmental Niche Modeling Exercise
Assessing Peer Contributions:
To make sure everyone is contributing, all group members will be asked to assess the
contribution of other group members to the project (rubric is at the end of this document).
This process serves two purposes. First it indicates whether any group members should
not receive full credit for the assignment, and secondly it provides feedback on how
successful you are at group work. Fill out a rubric for each group member.
24
Environmental Niche Modeling Exercise
Rubric for grading other group member’s participation
Criteria
1. Contributed the projection of assigned species and did so
in a timely fashion
Notes:
2. Offered help to others, or sought help when needed
Notes:
3. Asked questions that moved the discussion along
Notes
4. Contributed ideas, opinions
Notes:
5. Provided constructive feedback to other group members
Notes:
Total score:
25
/100