I tend to look at several models (NAM, ECMWF, HRDPS) and see if there is agreement first...in SE AK, I have noticed the ECMWF model skewing towards the upper end of wind forecasts and the others being more median in their calls. For that reason, I tend to plan around the ECMWF and hope for the others.
If you dont use it already, the spotwx.com forecast has varying models with different spatial/temporal coverage. The HRDPS model has proven fairly solid in terms of localized, near-term weather for me.