Laserfiche WebLink
<br />Preliminary Conclusions on Forecast Skill <br /> <br />1. Seasonal forecast skill for the interior southwestern U.S. derives <br />not only from ENSO, but also from SST regions closer and farther <br />to this region. In fact, predictive skill in the ENSO context derives <br />from the 'flavors of ENSO' rather than mainstream ENSO indices <br />such as Nino 3.4 ('ONI') or ME!. <br /> <br />2. Cross-validated skill appears to be largest in the winter season, and <br />lowest in the spring, consistent with CPC's experience. However, <br />independent verification for the last five years shows the opposite <br />to be the case: JFM skill is actually negative, while AMJ is best. <br />Nevertheless, two regions of skill in the winter season are <br />arguably the most important: the mountains of northern UT <br />and CO. <br /> <br />3. Unresolved issues: how to optimally present forecast skill vs. <br />areal coverage (which is already larger than CPC's), how to best <br />estimate expected forecast skill, and how to best represent <br />preceding soil moisture and snowcover/-pack conditions. <br />