Hi,

You can use NTSYS-pc program (see link below) for some multivariate analyses, for occurrence data. I have used this program and can recommend it to others.

http://www.exetersoftware.com/cat/ntsyspc/ntsyspc.html

Thank you for your advice Majid.

Hi Heli,

we analyse our vegetation (mainly species rich dry grasslands) data with CANOCO and/or R.

Best Regards,

Sonja 

Hi Heli,

    You can use Ecological Methodology software and STATISTA  software. 

Best regards

Thanks for all of you! Your comments are very informative.

Hi Jutila,

Here are some advices that may help you to avoid the most common misconceptions,  I often encounter as a referee and handling-editor:

Use CA and /or DCA in CANOCO, you do not need to detrend (D-CA) if there is no Arch -effect, the ‘D’ does not mean default, as many seem to think.

You may correlate your axes with env.-var. (e.g. island size) passively in CA /DCA without doing a simultaneously regression and ordination as which is the case in Constrained CA (CCA), in CCA the islands  in sample-plot will be controlled by the env.-var only (env.-space), whereas in CA/DCA you will have the islands located in species –space, which is very illustrative.

If the gradient length expressed in SD-units shown in DCA is below 2.0 to 1.5 SD you may consider to use a PCA. This may be the case because you do not have many islands.

 The procedure in CANOCO is straight forward, and the chance of making mistakes is much less than what may happen by using  the VEGAN pck. in R.

Good luck

Ole R

So much depends on what you want to do.

For ordination, unconstrained gradient analysis, classification, and Indicator Species Analysis, I mostly use PC-ORD. The graphics are not that great, but it is very fast to use and the text output can be imported into better graphics packages to make better graphs. Of course it costs money, but the ease of use may make up for it in time saved. I have been using it since version 2.

The are some analyses that PC-ORD cannot do (CANOCO is somewhat limited) but it is a good general purpose application for vegetation. The data modification module is very handy. 

Thank you Ole and Michael for your very valuable comments and advice.

We use both CANOCO and PAST as teaching tools for multivariate statistics. Students generally pick up CANOCO very quickly, but some of the menu options can get a little confusing. Having said that, CANOCO is very powerful and good tutorials are available. CANOCO v.5 is a lot easier to use than some of the older versions. Make sure you obtain the most recent version of PAST as the development team are always improving it and fixing bugs.

OK. Thank you for your advice.

Julia

If your primary goal is ordination and/or group classification cluster analysis Primer-E is an excellent software package that produces first class graphics and also performs a suite of hypothesis tests.  There is another statistical package that can be purchased with it called PERMANOVA that performs mulitvariate statistical tests equivalent to MANOVA, canonical discriminant analysis, PCA and several others. All of these driven by a graphical user interface. The program imports Excel spreadsheet data sets and other formats.  Output is either ASCII data files, tables you can cut and paste directly from the output window and several different types of graph formats. The user can select from a variety of fonts and symbols for the graphs. User support is very good usually with direct contact with the authors of the manuals which are excellent. USGS would find this particular application extremely useful. The only limitation I can see might be the size of your data set which you said was 555 species  I think. I'm not sure if the software can handle that many variables simultaneously. Here's a link to their website:

http://www.primer-e.com/

Thank you for your answer. I'm currently working as an independet researcher. My USGS affiliation dates back to 15 years.

Heli,

Sorry got your name wrong earlier... Best of luck with your research. The software I suggested is a bit pricey for an individual however if you can find an on-going class somewhere in the states or somewhere abroad you are planning on visiting you can take the class and get the software and class as a package deal usually at a lower price, Best of luck.