Grand claims are needed (if you're referring to the claim "anthropogenic air pollution very probably results in significant warming of Earth's climate", which is pretty much the biggest claim scientists made). There's a reason to think that way, and that reason has been questioned by tens of thousands of capable minds over the course of decades. Intelligent humans would listen to the warning, and act even when the above is not completely certain. Nothing can ever be completely certain, but scientific results will always be the closest thing.
By the way, we *are* able to predict the weather to a known level of accuracy, which is also rather high for short-term forecasts. Climate simulation is the same thing but much simpler (because we don't care about where and when it will be what temperature, only the average), but of course more difficult because of other reasons. That said, there are many uncertainties, some so uncertain that no value is given, but their range *is* known. The possible ranges can be read in the IPCC documents from 2007. This-and-that effect cannot be bigger than some limit, and these values are quite trustworthy, because if some effect was HUGE, then it would necessarily also be evident. The sun's impact is actually pretty well known - the changes in power output have been much too small to account for detected changes.
The temperature measurement network isn't grand, but it's also not giving out random numbers. We know that. The numbers don't look random. The signal-to-noise ratio is big enough that we can use those numbers, and other effects are accounted for (right, some thermometers are next to asphalt, but guess what - asphalt warms up the ground, there's now more asphalt than 50 years ago, thus asphalt indeed contributes to global warming (I suppose these effects go under the label of "land use" in IPCC documents if you want to look it up)).
Scientists are sceptics and continue to be that. But this means more than just questioning findings. Turns out the scientists have long ago researched the problem of how good their results are, and the 2007 report was groundbreaking indeed because then, for the first time ever, scientists concluded their results are "very probably correct". And mind you, their result was that the humans cause warming in the range of 0.6 - 2.4 watts per square meter. Of course there's always a tradeoff between dependability and accuracy of some result, now the numbers add up such that scientists can very confidently say something that's very approximate, but still useful.
By the way, the biggest uncertainty in climate forecasts is the amount of pollution humans spew out in the future. How would they know that? They wouldn't. We might be able to cut pollution by 50% in 20 years, or we might quadruple it in the same time frame. No way to know.