Analysis of Meteorological Conditions for the Yakima Smoke Intrusion Case Study, 28 September 2009

On 28 September 2009, the Naches Ranger District on the Okanogan-Wenatchee National Forest in south-central Washington State ignited an 800-ha prescribed fire. Later that afternoon, elevated PM2.5 concentrations and visible smoke were reported in Yakima, Washington, about 40 km east of the burn unit. The U.S. National Weather Service forecast for the day had predicted good dispersion conditions and winds that would carry the smoke to the less populated area north of Yakima. We undertook a case study of this event to determine whether conditions leading to the intrusion of the smoke plume into Yakima could have been predicted before the burn was ignited, either from forecasts and model output available on the day of the burn or from higher resolution model output made available only after the event. We evaluated three different meteorological model predictions: (1) 4-km resolution hourly weather predictions from the Weather Research and Forecasting (WRF) model that were available to forecasters on the day of the burn; (2) 4-km resolution WRF predictions at 10-minute intervals; and (3) 1.33-km resolution WRF predictions at 10-minute intervals. We found that predicted winds from the 4- and 1.33-km model resolutions compared well with each other, whereas there were some differences in the predicted planetary boundary layer height over Yakima. We also used the high-resolution 1.33-km WRF output to generate smoke dispersion predictions using the BlueSky Smoke Modeling Framework. Results showed that forecasters and regulators using either the model output available on the day of the burn or the higher-resolution model output generated afterward would not have anticipated the meteorological conditions that resulted in the smoke intrusion that day.