Hmmm... I wasn't even aware that such a number was being thrown around and I doubt it's veracity although I am an AGW proponent.
Why? Well, number one, who has ever managed to collate all the names of all the scientists in the world? I doubt that there is a central registry, especially as there isn't even a universally accepted definition of a "scientist". For example, can a scientist only be such if they have a degree recognised by Western academics? What if a person rigourously follows the scientific method but lacks the piece of paper? Does someone with a BSc in political science count? What about someone from the University of Pyongyang?
In other words, if we cannot define the total, how can we define a percentage thereof?
If you then say that you are only talking about a percentage of a group that can be exactly defined (e.g. a faculty or an association) then I am no longer interested in whether they agree in whole, in part, as a majority or a minority...
Again, why? Because any predefined group will, by necessity, be pre-qualified (e.g. all English speakers) which will come with their own biases.
In other words, any such statement can be considred a red herring and not germane to the issue at hand.
The way to judge these things is to use your own critical capacities:
What are the hypotheses? Have they been tested in a number of ways? Have the people who did the testing respected the scientific method? Do they have the experience and/or knowledge to do these investigations appropriately? Have others tried to duplicate, corroborate or refute the findings? Did those others follow the same rigorous process? How successful were they in corrobating or refuting? What do I, as a thinking, intelligent, person think of the results (before people draw conclusions, I should be capable of drawing my own)? What conclusions are drawn (and there will always be multiple and sometimes conflicting)?
How do all of these conclusions, including my own, stack up to real life experience (in other words, dlo they make sense)?
etc, etc.
You will note that nowhere in there do I think it relevant whether 9, 19 or 90% of this group or that agree or disagree. I do think it important that if a group wishes to make a statement or issue a report that they do so as a group but I accept that they are the only ones able to judge what percentage of their group have to agree before it can be deemed a "consensus" - it is not for outsiders to impose our rules on them.
Short answer:
I don't think it is possible to have empirical data regarding the opinion of ALL scientists.
I think different organisations will have different definitions, by-laws, regulations or whatever that determine when they, as a group can issue a group statement or opinion. Thus the second question is also unanswerable.