Respected all

I have run the RegCM-4 model for future with ECHAM5 data as input .. after comparing the results of the decade 2001-2010 with the observed data i got large biases (with an average decadal bias of 200-250 mm in the precipitation for one of the study stations) i have done this analysis for 10 station in the same region with almost same results.. so please suggest me how to incorporate this large bias for future analysis .. i want to present these results ...but they don't seem realistic..?

how can I predict the next decade2011-2020, 2010-2030 using these biases.... if there are methods please tell.
Also can any body tell me about the reason for the large biases...

Kind regards
______