I always concerned about the "cost-benefit" of regional climate projections where ensemble data is not available (e.g. South America). Since the changing signal is more reliable than quantitative estimations, I wonder if using global models ensemble isn't more recommended than a couple of regional projections.