Laserfiche WebLink
<br />How to make an experimental climate forecast <br /> <br />Current practice at CPC: ENSO + OCN (trend) + + + <br />Statistical Predictands are based on U.S. "Climate Divisions" <br /> <br />++t++++t++++++++++++++++++++ <br /> <br />Can one improve upon status quo for the interior southwest? <br /> <br />Use better climate divisions / good station data (include. SNOTEL) <br /><Better downscaling of ENSO impacts> <br /> <br />Add customized predictors: regional SST, soil moisture (history), <br />teleconnections (for instance, Indian Ocean), no <br /> <br />Here: apply stepwise linear regression (SLR) with 10% increased <br />variance requirement (more stringent than CLIPER), and decadal <br />cross-validation <br />