If you want to end the simulation prematurely, click on the Abort Simulation button or the X button in the upper right hand corner of the window. This will stop everything and take you back to the original opening menu. The simulation window also tells you the matrix size and the current iteration number.
After EcoSim has finished all of the iterations and calculations, the simulation window disappears and is replaced by a series of tab notebooks. Each tab window contains an important element of your output. The number and labels of the tab windows will vary from module to module, but almost all of the modules will have Input, Simulation, and Summary tabs.
The most recent version of EcoSim can be downloaded from:
| http://homepages.together.net/~gentsmin/ecosim.htm |
EcoSim is an interactive computer program for null model analysis in community ecology. EcoSim allows you to test for community patterns with non-experimental data. EcoSim performs Monte Carlo randomizations to create "pseudo-communities" (Pianka 1986), then statistically compares the patterns in these randomized communities with those in the real data matrix. These null model tests have wide applicability in both applied and basic ecology. Some of the kinds of questions you can ask with EcoSim are:
1) Is the species richness and evenness of unpolluted streams significantly different from that of polluted streams?
2) Do the number of "checkerboard" distributions in an island archipelago match the predictions of Diamond's (1975) assembly rules model?
3) Does the taxonomic diversity of an avian island community differ significantly from that of the adjacent mainland source pool?
This help file will introduce you to null model analysis and teach you how to use EcoSim for null model tests with your own data.
Gary L. Entsminger is a writer and computer programmer. He has written over 100 articles for computing journals and eight books about programming, including The Tao of Objects (1990, 1995; M&T Books), Secrets of the Visual Basic Masters (1992, 1994; Howard W. Sams), The Way of Delphi (1996; Prentice-Hall), and The Way of Java (1997; Prentice-Hall).
Nick (guitars) and Gary (guitar and mandolin) are also founding members of the Kesey-Bear Band, a group of acoustic musicians-biologists-computer programmers based in northern Vermont.
Here are the steps to batch processing:
| Gotelli, N.J. and G.L. Entsminger. 2001. EcoSim: Null models software for ecology. Version 7.0. Acquired Intelligence Inc. & Kesey-Bear. http://homepages.together.net/~gentsmin/ecosim.htm. |
We know that journal and book editors can be picky, but we'd really appreciate it if you could tuck the EcoSim website address in your citation. This will allow your readers to find the software more easily. EcoSim is a copyrighted product, but you may freely copy and distribute it to other users.
The most recent version of EcoSim can be downloaded from:
| http://homepages.together.net/~gentsmin/ecosim.htm |
Check this site periodically for updates. We try to respond promptly to your corrections and suggestions.
|
Nick Gotelli Department of Biology University of Vermont Burlington, VT 05405 ngotelli@zoo.uvm.edu |
Gary Entsminger 99 Schillhammer Road Jericho, VT 05465 gentsmin@together.net |
Click the right mouse button to access copy, paste, and cut functions within EcoSim. Most windows that contain text allow simple cutting and pasting from a menu displayed by clicking the right mouse button.
We encourage users to experiment with the different options and to read up on these null model tests. Like any statistical tool, EcoSim can be abused or misused by an ignorant researcher. Don´t rely mindlessly on the default values; they may not be appropriate for your particular data set or question! The "caveats" section of this help file points out important assumptions and common pitfalls for each of the modules.
EcoSim ‘remembers´ the settings that you request from one run to the next. However, EcoSim will provide a fresh random number seed for each run unless you give it a specific random number seed. Press the "reset defaults" button in the preferences window if you want to reestablish all of EcoSim´s default settings for a particular module.
However, the structure of the matrix and the entries in the matrix will vary from one module to the next. For example, in the niche overlap module, each column represents a niche attribute, such as a diet or micro-habitat category. In this case, each entry in the matrix represents the utilization of a particular niche category by a particular species. In the species diversity module, each column represents a different site. Each entry is the abundance of a particular species collected at the site. For each module of EcoSim, you can read about the data structure in the "Data Format" section of the help file.
When you´re finished editing, select the "Close Editor" item from the menu of the Text Editor. EcoSim will then ask if you want to retain your editing changes. EcoSim then returns you to the main EcoSim window where you see that your data have been updated in the main grid.
If you want to make these changes permanent, be sure to save them to disk when you return to the main EcoSim window. Data modified in the Text Editor are not automatically saved to disk.
EcoSim expects imported data to be in a particular format. The first column is reserved for the row titles, and the first row is reserved for the column titles. Titles must be strings of characters that are not separated by breaks. Be sure that species and genus names do not have breaks between them. Use the underline character to join names. Thus:
Chthamalus stellatus
Should be changed to
Chthamalus_stellatus
EcoSim treats blanks as delimiters between data and labels. Thus, a small data set with two rows and three columns should, in correct format, look like:
|
Species Scotland Wales Ireland Balanus_balanoides 0 0 22.2 Chthamalus_stellatus 1 10 155.6 |
Notice that all data and titles are separated by a single space and that "Species" is used as a label over the column headings. Both row and column labels are mandatory. Without them, EcoSim will not align the data properly. Data entries can be either integer or real numbers. Different modules will require different data structures for analysis.
To keep EcoSim running quickly, we´ve included only a minimum amount of error trapping for input data. Please check your data carefully! EcoSim recognizes the characters 'x' or 'X' as exclusion symbols in the user-defined constraints of the Niche Overlap module. In general, how EcoSim handles non-numeric data values depends on the module. In some cases, EcoSim might successfully analyze a data set that contains letters rather than numbers, but don´t bank on the results!
As always, make sure you remember that EcoSim is looking for a single space or a comma as a delimiter between adjacent cells in your data matrix.
Create a directory on your hard disk. Name it whatever you like. We suggest you name it "EcoSim."
Then copy the self-extracting file, EcoSimx.exe on the EcoSim disk to the directory you create.
Then run EcoSimx.exe. It will extract all system, Help, and tutorial data files to the proper directories.
Note that EcoSim requires no special setting of CONFIG.SYS or *.INI files.
Also note that if you have a screen saver installed, and if it isn´t well behaved, you might encounter problems when EcoSim is simulating large data sets. If your screen saver behaves weirdly or crashes, turn it off while you´re running large, time-consuming simulations. Normally there are no conflicts between EcoSim and well-behaved screen savers.
If 1000 iterations are used, we would reject the null hypothesis at p < 0.05 if the observed value was less than or equal to 50 of the 1000 values generated for the null communities. If only 100 iterations were used, the rejection point would be if the observed was less than or equal to the simulated in 5 out of the 100 iterations.
The maximum number of iterations that EcoSim will accept for Niche Overlap, Species Co-occurrence, Macroecology, and Body Size Overlap is 30,000. Most data sets give reliable results with 1000 or 10,000 iterations. In some modules you might initially use as few as 100 or 500 iterations if the run time is too long.
Diamond, J.M. 1975. Assembly of species communities. pp. 342-444 in: Ecology and evolution of communities. M.L. Cody and J.M. Diamond (eds). Harvard University Press, Cambridge.
Entsminger, G.L. 1996. The Way of Delphi. Prentice-Hall.
Gotelli, N.J. 2001. Research frontiers in null model analysis. Global Ecology and Biogeography Letters 10: 337-343.
Gotelli, N.J. and G.L. Entsminger. 2001. EcoSim: Null models software for ecology. Version 7.0. Acquired Intelligence Inc. & Kesey-Bear. http://homepages.together.net/~gentsmin/ecosim.htm
Gotelli, N.J. and G.R. Graves. 1996. Null models in ecology. Smithsonian Institution Press, Washington, DC.
Gurevitch, J., L.L. Morrow, A. Wallace, and J.S. Walsh. 1992. A meta-analysis of field experiments on competition. The American Naturalist 140: 539-572.
Harvey, P.H., R.K. Colwell, J.W. Silvertown and R.M. May. 1983. Null models in ecology. Annual Review of Ecology and Systematics 14: 189-211.
Manly, B.F.J. 1991. Randomization and Monte Carlo methods in biology. Chapman and Hall, London.
Nitecki, M.H. and A. Hoffman. 1987. Neutral models in biology. Oxford University Press, Oxford.
Pianka, E.R. 1986. Ecology and natural history of desert lizards. Princeton University Press, Princeton.
Wiens, J.A. 1989. The ecology of bird communities. Volume 1. Foundations and patterns. Cambridge University Press, Cambridge.
We developed EcoSim with Delphi, a powerful object-oriented programming environment that uses the underlying Object Pascal language. This state-of-the-art development tool is not available for the Mac. Without it, it would have taken us much longer to produce the version of EcoSim you´re using.
However, all is not lost for Mac users who have access to a PC! EcoSim can use data files created on the Mac in Word, Excel, or any other program that allows you to save ASCII text files. These files can then be transferred to a PC disk and imported directly into EcoSim. Similarly, EcoSim output and transition matrices can be saved to ASCII text files that can then be read by a Macintosh computer.
* Create a data set directly within EcoSim
* Import a text file that contains a data set
Once the data is in EcoSim, you can perform basic editing functions within the spreadsheet-style or grid editor, or you can manipulate the data in a notebook-style text window by Editing Data as Text.
File | New This command creates a fresh 5 x 5 data matrix in the main EcoSim screen.
File | Open This command lets you open a previously saved file from disk.
File | Close This command lets you save (and name) your data file to disk.
File | Save As This command lets you save a data file under a new name. This is useful if you´re editing or experimenting with your data set and need to save several different versions.
File | About This command tells you the full file name and the dimensions of the matrix.
File | Batch...on/off This command instructs EcoSim to process automatically process multiple input files and save the summaries to disk.
File | Exit This command closes EcoSim and returns you to Windows.
Edit | Lock Row Title This command is a toggle switch that locks or unlocks row titles. When the titles are locked, the first row and column of the matrix are shown in gray and cannot be edited. When the row title is unlocked, the first row is shown in white and can be edited. Other cells in the matrix can always be edited at any time.
Edit | Lock Column Title This command is a toggle switch that locks or unlocks column titles. When the titles are locked, the first row and column of the matrix are shown in gray and cannot be edited. When the column title is unlocked, the first column is shown in white and can be edited. Other cells in the matrix can always be edited at any time.
Edit | Resize Matrix takes you to a dialog box that lets you specify the dimensions of the matrix. New rows or columns are added as necessary, with default species and site names added. If you specify matrix dimensions that are smaller than the current matrix, EcoSim strips off the extra rows and columns from the end of the matrix.
Edit | Transpose Matrix exchanges all of the rows in the data set with all of the columns.
Edit | Edit Matrix as Text takes you to the text editor, which allows you to edit your data more extensively in a notepad-style window.
Edit | Edit Matrix as Text | Edit | Tabs To Spaces converts tabs in your data set to single spaces, which is a godsend when trying to import Excel spreadsheets.
Analyze | Co-occurrence takes you to the preferences page of the co-occurrence module.
Analyze | Macroecology takes you to the preferences page of the macroecology module.
Analyze | Niche Overlap takes you to the preferences page of the niche overlap module.
Analyze | Size Overlap takes you to the preferences page of the size overlap module.
Analyze | Species Diversity takes you to the preferences page of the species diversity module.
Analyze | Standard tests takes you to the preferences pages of each of the standard tests modules (ANOVA, Chi Square, Regression, Runs Test).
Help | Help Topics takes you to the main window of the help file (which you are currently reading).
Help | About shows you the current version number of EcoSim that you are using.
"A null model is a pattern-generating model that is based on randomization of ecological data or randomly sampling from a known or imagined distribution. The null model is designed with respect to some ecological or evolutionary process of interest. Certain elements of the data are held constant, and others are allowed to vary stochastically to create new assemblage patterns. The randomization is designed to produce a pattern that would be expected in the absence of a particular mechanism." Gotelli and Graves (1996).
In other words, null models include Monte Carlo simulations of community data, as well as some conventional statistical tests that are tailored to particular ecological questions (Gotelli 2001). Null models are based on the principal of the null hypothesis –that patterns in the data do not reflect biological forces, but represent chance variation or sampling effects.
In statistics, a typical null hypothesis would be that observations in two groups were drawn from the same distribution, so that any differences observed in group means were due only to sampling variation. The alternative hypothesis would be that the observations in the two groups were drawn from two different distributions.
In community ecology, the null hypothesis can take on a variety of forms, but typically it would be that species distributions reflect chance colonization and extinction, subject to certain constraints or differences in sampling intensity. The alternative hypothesis would be that species occurrences are non-independent of one another, reflecting biological processes. The null model is used to randomize the occurrence of species, and to compare the patterns in these randomized communities with those in the real data.
For example, Diamond (1975) hypothesized that competitive interactions lead to checkerboard distributions, in which two competing species never occur together on the same island. Connor and Simberloff (1979) tested Diamond´s hypothesis with a null model in which species were distributed randomly and independently of one another. They compared the expected number of checkerboard pairs in these simulated communities with the number observed in the data matrix.
To do this, you would create a large number (typically 1000) of randomly assembled communities. These "pseudo-communities" are generated by randomly reshuffling the observed species occurrences, subject to certain constraints. Next, a histogram would be constructed of the number of co-occurring species pairs in the pseudo-communities. This histogram tells you the range of values you might expect for the number of co-occurring pairs in an assemblage that was not structured by competition. Finally, the number of co-occurring pairs observed in the actual data matrix would be compared to the distribution of co-occurring pairs for the pseudo-communities.
For example, suppose you found in the observed matrix 680 pairs of species that formed checkerboard distributions and never co-occurred. In the simulated matrices, 985 out of 1000 times, the number of exclusive species pairs was less than 680. Therefore, by chance, the probability of obtaining 680 or more exclusive species pairs is 15/1000 = 0.015. This is less than the conventional level of statistical significance of p = 0.05, so we would conclude that the observed data matrix contains more exclusive species pairs than expected by chance, at least compared to the particular null model we used.
The use of null models has been controversial in ecology, and critics have raised both statistical and philosophical objections to the approach. Like cladistic analysis in systematics, null models have both strong advocates and detractors. It is impossible, in the space of this help file, to review the origin and development of these controversies. We encourage users to learn about the history of this debate so they can use null models in an informed and appropriate way. Good entry points into the literature include Harvey et al. (1983), Nitecki and Hoffman (1987), Wiens (1989), Manly (1991), and Gotelli and Graves (1996), and Gotelli (2001) .
Once you select a module to run, the data window disappears, and you´re presented with a "Preferences" window that contains a number of options for the simulation you´re running. You can change these options to customize the simulation or use the defaults that EcoSim provides. EcoSim "remembers" your changes from one run to the next, so you can easily experiment with the options and see how they affect the results. Clicking the "Restore Defaults" button resets the EcoSim default choices.
You can also cancel the simulation or access this help file from the "Preferences" window. Once you have selected the options you want, the "Run" button starts the simulation. The "Preferences" window disappears and a "Simulation" window pops up while the program is running. It contains information about the simulation and keeps you posted about its progress. An abort button lets you terminate the simulation at any point.
When the simulation is completed, the simulation window disappears and is replaced by a series of tabbed notebooks. By clicking on different tabs, you can quickly jump back and forth from one window to the next. The tabs are arranged to be read consecutively from left to right.
The first tab in a module is the Input tab, which simply shows you the input matrix again. The second tab is the Simulation tab, which shows you one of the simulated matrices. The next tab(s) give various output details from the simulation. The final Summary tab contains all the simulation results in one window. You can edit, delete or annotate this material and save it to disk.
The default value for the random number seed is zero, which instructs EcoSim to go the clock for a new value. You can enter a specific seed (an integer from 1 to 2,147,483,647) for a particular run. Each time you run a new simulation, the random number seed is recorded in the summary window, so you have a complete record of your results.
CAUTION: Checking this option will slightly slow the operation of EcoSim, but more important: it will consume substantial disk space! Unless you have some other plans for the simulated matrices, there is no need to generate them and save them to disk.
If you just want to see one of the simulated matrices, you will be able to do that in the output tabs. If you would like to look at a few matrices, you could set the iterations to a modest number, and then save them to disk. It is unlikely you would ever want to save 1000 simulated matrices unless you plan to write your own program that uses these matrices for some other purpose that is not covered by EcoSim´s standard output.
Different preference choices will change the appearance and structure of the simulated matrix. Also note that the contents of the simulated matrix will change each time you run the simulation, unless you´ve re-entered a particular random number seed.
(Observed index - Simulated index)/(Standard deviation of simulated indices)
Thus, your observed index is rescaled as the number of standard deviation units above or below the mean of the simulated values. Negative standardized effect sizes indicate that the observed index was less than the mean of the simulated indices; positive values indicate that the observed index was greater than the mean of the simulated indices. The approximate 95% confidence intervals for the standardized effect size are -1.96 to 1.96. However, you should always report the probability value that EcoSim calculates directly from the histogram itself.
The rationale for the standardized effect size follows directly from meta-analysis (Gurevitch et al. 1992). In meta-analysis, a standardized comparison of experimental data is made between a "control" and a "treatment" group, with an estimate of the pooled variance for these treatments. Of course, in a null model analysis, there is no "experiment". However, we can think of the simulated matrices as a type of "control" for the effects of species interactions, and the observed matrix as the "treatment" that may show the effects of those interactions.
One potential disadvantage of using the standardized effect size is that the histograms of simulated values that EcoSim generates often are not normally distributed, even though this is an assumption underlying the calculation. But this disadvantage is offset by having a standardized index that measures the effect size from different studies on a common scale.
The content of the Summary tab will vary among modules, but it will typically contain the observed metric calculated for the input matrix, and summary statistics (mean, median, variance) on the variables calculated from the simulated matrices. The Summary tab contains its own text editor, so you can delete unneeded results and annotate the output extensively.
In some modules (e.g., species diversity) the Summary tab may be split into two windows, with the lower window giving output information in the form of a spreadsheet. This spreadsheet window cannot be edited as the text window can. However, the information in either or both windows can be saved to a disk file by pressing the appropriate button at the bottom of the window. There are also buttons for accessing help or closing down the output. Finally, there is a time log so you can see exactly how long it took EcoSim to run the simulation.
If you plan to save the simulated matrices, you will need additional disk storage space corresponding to the size of the input matrix times the number of iterations.
The time it takes to run a module will vary greatly depending on the options and number of iterations you specify and the dimensions of the input matrix. However, most runs using default values take a few seconds to a few minutes at most. Large data sets, high numbers of iterations and/or certain simulation options could cause simulations to run for several minutes or even hours. In this case, EcoSim operates in the background of Windows, so you can use your computer for other tasks while EcoSim works.
In addition to computer requirements, we suggest you read:
Gotelli, N.J. and G.R. Graves. 1996. Null models in Ecology. Smithsonian Institution Press, Inc. (1-800-782-4612).
Many of the issues and ideas we explain briefly in this online Help file are discussed in much more detail in Null models in Ecology. The book also contains an extensive bibliography of papers that use null model tests.
The initial Help screen has a panel of topics on the left and a window of text on the right. At the top of the topics list, you will see a list of the current modules available in EcoSim, followed by an alphabetical list of help topics that are common to all modules.
The Help screen for each individual module has the same organization of topics:
|
A navigator link at the bottom of the Help screen in each module will return you to the initial list of topics.
Take a few minutes right now to learn how to control the appearance and operation of your browser (something you should have done some time ago, but may not have!). For example, in Internet Explorer, you can select "Internet Options" under the "View" menu and change the font, pitch, and background color of your screen. There are also options for finding text within an html file, which is very useful for scanning EcoSim Help for information on specific topics. Help files can also be printed out from your browser.