Methodology for generating the U.S. spring forecast
Climate discussion: Springtime is typically the most difficult climate forecast as the transition from cold to warm season is usually volatile and changeable. Climate signals such as ENSO are also frequently in transition contributing to forecast uncertainty. Spring 2019 will contain these uncertainties.
As always, we begin with ENSO to generate a climate forecast. Analogs are taken from the past ~20 years due to the strong influence on climate for all seasons by the constricting polar ice cap caused by increased CO2 emissions (Fig. 1). From the past 20 years there are 5 ENSO analogs which show a neutral to La Nina signature 1 year ago transitioning to weak El Nino the following winter and similar to the 2018-19 ENSO regime (Fig. 2). In 4 of these analog years ENSO transitioned to neutral phase or La Nina. In one analog year El Nino re-emerged and intensified.
Dynamic models are mostly in favor of El Nino in 2019 despite short-term signs of weakening to neutral phase. Additional warming of the equatorial Pacific subsurface near the Dateline now is recognized as fuel for a potential new effort to generate El Nino later this spring. Therefore the forecast is based on a consensus of analogs which indicate neutral to weak El Nino during spring.
The MAR/APR/MAY 500 MB anomaly pattern produced by the analogs has similarities to the NCEP CFS V2 model (Fig. 3-4). Most important is agreement on emergence of an upper ridge pattern in central North America. The analog is stronger with an upper trough south of Greenland while the CFS V2 is weaker. The analog and CFS V2 are in agreement on a ridge pattern over Europe. The least confident portion of the projected 500 MB anomaly pattern is the (CFS V2) upper trough over Alaska to Siberia. If this trough is weaker, the North Atlantic trough and Europe ridge will be stronger.
The forecast in this report combines persistence (previous outlooks) with the analog forecast and certainly considers the CFS V2 output. Forecast highlights include
Fig. 1: The 1953-2018 arctic sea ice extent identifies significant constriction the past 20 years due to CO2 emissions. The influence on climate is sufficiently evident to limit selection of ENSO analogs to the past 20 years.
Fig. 2: The ENSO analog years selected from the past 2 decades in which La Nina transitioned top weak El Nino and what followed.
Fig. 3-4: The 500 MB anomalies for meteorological spring generated by the analog years compared to the NCEP CFS V2 forecast model have similarities but vary on intensity.