I think it's better to make forecasts that are later proved wrong, then acknowledge that they are wrong, make new forecasts and appropriately calibrate the new forecasts based on the lesson learned in humility from the first wrong forecast.

Trenberth in particular seems to be a fairly honest and open climate scientist, in that he made an explicit forecast, then later admitted a change of mind. He's also the person who admitted (within an email leaked by Climategate) that there was a problem with balancing the energy budget, and he later publicly noted the same, and tried to come up with an explanation.

The problem isn't with people making wrong forecasts, it's with people (a) refusing to make forecasts while still implicity doing so by claiming near-certainty about the future and seeking action based on that, or (b) making forecasts and insisting on the forecasts being treated as correct without an external test of validity or a past record of forecasting expertise.

See also:

http://econlog.econlib.org/archives/2012/05/the_bettors_oat.html