Hawaii Extended Forecast Discussion NWS Weather Prediction Center College Park MD 710 AM EST Sat Feb 07 2015 Valid 00Z Sun Feb 08 2015 - 00Z Sun Feb 15 2015 At least in principle guidance is fairly agreeable with the forecast scenario through about Thu. Thereafter increasingly significant model/ensemble differences develop from upstream over the Pacific into the islands. Initially bands of mostly light rainfall will soon depart along with a fast moving front. Trailing weak high pressure will track just north of the islands on Sat. Then winds will become southwesterly and strengthen ahead of a potent cold front that should pass through the region in the Mon-Tue time frame. Some locally enhanced rainfall is possible with the front. Also expect another leading area of rainfall closer to the Big Island, with GFS trends from the past couple days leading toward the ECMWF continuity of a separate band of moisture lifting into the Big Island and vicinity from the southeast. Passage of this front will bring a brief drying trend and still brisk westerly or northwesterly winds. Forecast confidence decreases rapidly by Fri-Sat as models/ensembles differ over relative emphasis of amplifying northern stream flow to the northwest of the islands and low latitude waviness originating from west of 160E. The past couple days GFS/GEFS forecasts in varying ways have been emphasizing the low latitude wave, with some eventual strengthening once influenced by northern stream flow. This would lead to a longer period of moist southwesterly flow relative to the ECMWF/ECMWF mean/CMC mean that are northern stream dominant and depict a clean frontal passage with faster timing. The operational 00z CMC is somewhat a hybrid of the two clusters with the result being an intermediate frontal passage. At this time would lean to the 00z ECMWF mean/CMC mean as the best individual solutions given the prevailing amplitude of central Pacific mean flow. Confidence in the deep surface low development of the operational 00z ECMWF is fairly low at this time so would prefer its specifics less than the aforementioned means, but still more than the GFS/GEFS scenario. Rausch