I finally understand the cause.
Those translations are copied from IBM branches to master branches.

The copy action is trigged by the update of pot files.
There are Jenkins jobs automatically uploading pot files by CLI when there are changes.
This caused the translations copied from IBM branches to master.

Refer to Zanata help document[1], it is said:
Copy Trans usually runs automatically when a document is uploaded using the cli-client.

While I did test on the test server ( http://translate-dev.openstack.org ), there was no
such auto jobs running for test server, so I didn't recognize uploading of pot files
could cause the copy action.
I apologize for missing this point.

What we should do now is to figure out how to handle the situation.

Best regards
Ying Chun Guo (Daisy)

[1] http://docs.zanata.org/en/latest/user-guide/translation-reuse/copy-trans/


Andreas Jaeger <aj@suse.com> wrote on 09/22/2015 10:41:00 PM:

> From: Andreas Jaeger <aj@suse.com>
> To: Ying Chun Guo/China/IBM@IBMCN
> Cc: "openstack-i18n@lists.openstack.org" <openstack-i18n@lists.openstack.org>
> Date: 09/22/2015 10:41 PM
> Subject: Re: [Openstack-i18n] IBM translations to contribute:
> waiting for review
>
> On 2015-09-22 15:40, Ying Chun Guo wrote:
> > I think they are not IBM translations.
> > IBM translations are not merged into master branches yet.
> > They are still in pre-review status.
> > After coordinators review, these translations could be merged into
> > master or
> > liberty branches in Zanata.
> > After that, the importing could happen.
> >
>
> Have a look at these:
>
> https://review.openstack.org/#/c/225557/
> https://review.openstack.org/#/c/226186/
>
> That's such a large import, I would be surprised if those are not the
> IBM translations - and they have Tom Cocozzello mentioned as translator,
>
> Andreas
> --
>   Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
>    SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
>     GF: Felix Imendörffer, Jane Smithard, Graham Norton,
>         HRB 21284 (AG Nürnberg)
>      GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
>