Improving High-Resolution Rapid Refresh (HRRR) Forecasts of Low-Level Winds in the Columbia River Basin for Wind Energy Applications

Under the auspices of the Department of Energy Wind Forecast Improvement Project, version 2 (WFIP2), a study of low-level flow kinematics and dynamics is planned for an area of intensive wind-power generation straddling the Columbia River east of the Cascade Ridge. The goal of this project is to improve short range (1–24 h) operational numerical weather prediction wind forecasts at wind-turbine-rotor heights, typically 30–160-m above ground level. Special instrumentation, including Doppler lidars, radar wind profilers, and instrumented towers, will be deployed for the field phase of the project, which is expected to run from late 2015 through 2016.

Industry partners in the project have identified four weather-related regimes that make for particularly challenging forecasts in this area:

- A Pacific cyclone approaching the coast of Washington or British Columbia, with stable air at low levels east of the Cascades. The associated forecast challenge is determining if and when developing southerly flow ahead of the cyclone will mix down to sufficiently low levels to affect wind-power generation (yielding so-called "up ramps").

- West to northwest flow aloft with terrain-modulated flows east of the Cascades caused by vertically propagating mountain waves, trapped lee waves, downstream shadowing by prominent mountain peaks, etc., all of which modulate "up" and "down" ramps and contribute to high spatial and temporal variability in turbine-level winds.

- Onshore marine pushes through the Columbia Gorge, where the timing, intensity, and duration of associated "up" and "down" ramps must be anticipated.

- Outflows from (mainly terrain-modulated) deep convection and their attendant "up" and "down" ramps.

The first two of these are primarily (but not exclusively) cold-season phenomena, while the last two mainly occur in the warm season.

This preliminary study is aimed at qualitatively describing how the High Resolution Rapid Refresh (HRRR, 3-km horizontal grid spacing), the Rapid Refresh (RAP, 13-km grid spacing), and a WRF-ARW 750-m grid-spacing nest within the HRRR perform in these situations. Moreover, we will describe model development strategies that are intended to improve forecasts of low-level winds in complex terrain during WFIP2. Case examples and preliminary findings will be presented.