Could this site helpful for you. Distribution data of Rosa spp. proibably this could be a starting point for ecological works and factor analysis? 

http://www.discoverlife.org/mp/20q?search=Rosa&guide=USGAClarke&l=spanish-publisher

I can imagine that there are no real reliable data on Rosa species, just because the taxonomy of Rosa is very confusing. Especially in the sect. Caninae several species concepts are in use. So, even if you have indicator values for the taxa in this section, you have to be very careful in interpreting and using them. If you for instance compare the species in the new Dutch standard list (published in Gorteria 2011), in the new Flora Gallica (published last year), and the ideas in the BSBI handbook series (somewhat older), they all are very different.... And the conclusion of Sell and Murrell in volume 2 of the Flora of the British Isles and Ireland (also published last year) is that Rosa taxonomy is very much corrupted by taxonomist hobby-ism.... 

Dear Attila and Rense, thanks a lot!

In future, I want to research roses in relation to the ecological factors in different locations and habitats. Also, I agree with Rense Haveman that situation with taxonomy of Rosa is very confusing. However, among all roses there are those which are "pure" species. For example, there are Rosa rubiginosa, Rosa villosa, Rosa rugosa etc.

Eventually, in such work at the first step it needed to find limits of ecological variation of factors in which rose's species will be viable. This is why I want to reveal relation of roses to ecological conditions in different locations (countries, habitats etc.) according to well-known (Ellenberg, Landolt etc.) and, probably, newly created ecological scales.

As answer to Attila Haris, relatively to possibility of ecological work, I may say that of course it required extensive work to evaluate the roses in relation to ecological factors. And multivariate analysis should be a good tool in this.

Best regards, Anatoliy

As all pointed out, the taxonomy of Rosa is far from simple, and lots of revision is needed, which is the main reason that the standard lists like Ellenberg have very few species included, or use aggregates. The best way to go about this would be to carefully revise the material you have, and then try to develop your own ecological indicator values, probably best based on the Ellenberg scale, for each species.

Dear Dr. Bussmann, thank you for your advice! I use exactly term "aggregates" in my work for classification of roses. And it is very helpful, especially for section Caninae.

You could check with some of the taxonomic specialists who have just published treatments for the Rosaceae in Flora of North America (htpps: www.fna.org).

Thank you for your advice!

If you are working with well-thought-through groups (aggregates), you can minimize the threats of the confusing taxonomy. However, possibly, nobody knows in Central and Western Europe (where the most appropriate list of indicator values for vascular plants were prepared) the Russian Rosa-taxa and their ecology. Thus, usefulness of western data in your work is rather low. I think, the Hungarian list of indicator values (Borhidi, A., 1995: Social behavior types, the naturalness and relative ecological indicator values of the higher plants in the Hungarian flora. Acta Botanica Hungarica 39, 97–181) is somewhat better to your purposes, but I suppose certain hiatus (compered to your region) in this case as well.

Summary: the best you can do, a correct study on the regional Rosa-taxonomy and, additionally, vegetation-ecological investigations in the same region.

Hi,

You can also see: Zarzycki K., Trzcińska-Tacik H., Różański W., Szeląg Z., Wołek J., Korzeniak U. 2002. Ecological indicator values of vascular plants of Poland. Ekologiczne liczby wskaźnikowe roślin naczyniowych Polski. W. Szafer Inst. PAS, Kraków.

Thanks a lot for all suggestions! I will set these sources.