Predictability characteristics of land-falling cyclones along the North American west coast

North Pacific storms frequently make landfall along the west coast of North America, from California to southeast Alaska, and bring strong winds, precipitation and large mountain snowfall that result in significant societal and economic impacts. The success of numerical forecasts of these storms can vary widely, from highly accurate prediction of storm location and intensity to storm position errors on the order of 100's of km and intensity errors of 10's of hPa. Past studies have shown that short-term forecast errors of sea level pressure along the west coast are large and are statistically related to large-scale upper-level flow regime. We expand on this prior work by using other methods to define predictability and relate the predictability to cyclone characteristics. We use a WRF-model ensemble Kalman (EnKF) filter to investigate the predictability, ensemble sensitivity, and spread characteristics of land-falling cyclones on the west coast of North America over two winter seasons (2008 – 2009 and 2009 – 2010). For the EnKF model runs, observations are assimilated on a 6-hr cycle and include: satellite cloud-track winds; aircraft temperature and wind observations; rawinsonde temperature, wind and humidity; and surface wind, moisture and altimeter observations. In this study, we define predictability in terms of ensemble spread at the final forecast time (24-h), where low predictability exhibits large ensemble spread and high predictability exhibits small ensemble spread.

We find that storms that are deepening and that track from the SW exhibit the lowest predictability in terms of the highest ensemble spread and largest ensemble initial condition sensitivity. Decaying storms and storms that track from the NW have the highest predictability. The latitude of the storm when it makes landfall is also related to predictability. Storms that end south of 40°N exhibit higher predictability than storms ending further north regardless of whether they are deepening or decaying over the 24-h period. Composites of 500 hPa heights and sea level pressure at initial and 24-h forecast times were made for storms that are characterized by particular combinations of ensemble spread and ensemble sensitivity. Among our results, we found that the 500 hPa heights and sea level pressure composites for storms that have large ensemble spread and high initial condition ensemble sensitivity (case A) are quite different from storms that have large ensemble spread and low initial condition ensemble sensitivity (case B). The case A storms showed amplification of the upper level trough and deepening of the surface low over the 24-h period. In contrast, case B storms were already rather deep at the initial time and did not undergo further amplification over the forecast period and were further north than case A storms. These results imply that the low predictability (high ensemble spread at forecast time) for case B storms could be due to either large errors at the initial time or due to small errors in cyclone position at the forecast time which appears as large ensemble spread since the surface lows are deep. These results highlight particular synoptic situations and cyclone characteristics that are associated with low predictability and can be used to improve operational modeling/data assimilation systems.