my main question is, why are the projects still running climate change from the early 1900s to 1990s, instead of running into the future to see what happens in the future. I think thats what this project is about, isn't? Isn't this project to predict the future climate in the 21st century or so? Just wondering of that, most of the projects that I have gotten have the similar timeline, early 1900s to late 1990s.

... To test parameter combinations against history before running into the future. We used to run 160 years in a single model (a few of us ran some 200-year models) but there was a lot of whinging over time by people accustomed to other projects where tasks complete in a few minutes or a few hours, so CPDN tasks were broken into pieces, one consequence of which is much larger downloads and uploads.

The different experiments are documented here (select from list on right side of the page):
http://climateprediction.net/content/experiments

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Actually the models DO predict the future.
But it's the "future" that will happen as a result of starting from a given point in time, and using given values for the many parameters and variables that make up what is known to be a part of "weather".

These models are based on what the individual research groups are working on, and they are located in many different parts of the planet.
The project people just act as co-ordinators and testers for these research groups, and the public in turn just run the individual models.
If a research group wants to investigate a past event, that's up to them. One such research effort was the Seasonal attribution experiment, which you'll find explained in one of the pages pointed to by astroWX. Another was the Millennium experiment.

And to add a bit to what astroWX said, the individual models are all part of a much longer climate run; they get "stitched together" by the researchers to form a very long time line, FOR THAT SET OF STARTING VALUES.
And sometimes it's necessary to get past-time-modules re-run, because none of the computers that ran that one in the past actually completed it.

There is NO "real world" predicting going on. That gets done by weather bureaus in various countries.

---

Backups: Here