Welcome to the Q&A Forum

Hi, That sounds like an issue with the cms itself, it is creating the dashed urls but also keeps a record of the original urls and delivers them if requested. If the urls with blanks in them are not being linked to internally and not being linked to by external sites then really you have no problem, only you are looking at the spaced urls nobody else (including the search engines) know they exist, so no issues with duplicate content.

You may just be experiencing the delay that happens between crawl time and analysis time in GWT.  I haven't seen anything official about this, but in my clients' sites and my own sites it appears to me that it can take as much as a month after a recrawl of a page for Google to respect a 301 or rel=canonical in terms of what it reports in Webmaster Tools (and also what it shows in the search results). If you've carefully checked the complete set of HTTP responses and only see a single 301, then I would let it site for another 2 weeks and see if the problem goes away.  FYI I like to use HTTPfox to check HTTP responses, as I can see the complete chain of responses--not just the last or first in the chain.

Yes James you're referencing HTML that's incorrect

You said  "If I am not mistaken noindex nofollow on the internal links could be seen as page rank sculpting where as onpage meta robots noindex nofolow is more of a command like your robots.txt" This used to be an issue, but now nofollowing a link throws away the PR. It used to be that if you had 4 links on a page an nofollowed 2 of them all the PR gets pushed to the 2 links. Now if you nofollow two links, the PR still is divided by 4 and you're throwing away half of your PR. So Google could care less now what you are nofollowing internally. Nofollowing links to your pages is not enough to keep your pages from getting indexed, you need the noindex, follow meta tag on the page. Even robots.txt disallow sometimes isn't enough because Google could come in from a deep link and index the page. Why aren't you using URL parameter handling in WMT to tell Google to not index pages with the ?filter parameter in the URLs?

I'm not a fan of doubling up, but only because it makes the results really hard to measure. If you implement both, you won't know which one worked, ultimately. I'm not sure it's actually harmful - it just can be hard to track. If you're just trying to prevent future problems (and don't have any immediate issues), I'd probably pick one and give it a few weeks.

Hi Irving Fully aware of that, but if I am not mistaken a URL with a - in and one with %20 in is still seen as 2 different URLs and is duplicate content if they both have the same content on? Here is an example of the SERP, you can check on it is the URL that is indexed with the - but once you click it, it goes off to the %20 version, however, if you navigate through the site to this blog post it shows as the - URL, now comes the interesting part, if you use the site search and search for "Southern Right Whale" this blog post comes up first and it is indexed in the site search as the %20 version. Here is the serps let me know your thoughts. Chris.

Hi Joe Thanks for your advice, however, if you look at the websites, gansbaai.com it is really bad, only has 1 page, I really only bought it for the domian age and couple of links it has (and because I want to launch my new sit under this domain)....so there are not enough pages to 301 all danger-point, and I want to loose as little link juice as possible....Any feedback would be greatly appreciated. Thanks