Tutorial: Using Mega-cluster Ensemble to Project Great Plains Precipitation

Fig. 1-2: Daily U.S. soil moisture anomalies and the seasonal change.

Discussion: At issue is the Great Plains/Texas drought as mid-meteorological spring approaches (Fig. 1). All western and southern Kansas plus the Oklahoma Panhandle are in D4 (“exceptional drought”) classification. Last week D4-drought in this region expanded sharply and intensified. The drought extends from parts of the northern Great Plains to southern Texas. Florida and parts of the Mid-Atlantic are also in drought. California/Great Basin soil moisture is plentiful given the just-ending stormy winter pattern. However, long-term precipitation deficits maintain moderate drought over many parts of the West. The seasonal soil moisture trend is generally wetter across the southeast Great Plains to the Great Lakes region plus parts of the West while the Gulf States and East Coast are somewhat drier (Fig. 2). To best understand the 15-day outlook, the mega-cluster ensemble is considered to determine rainfall risk in the drought areas.

The most likely upper air scenarios for the 5-day periods through the next 15 days yield an upper-level high-pressure ridge across the North-central U.S. in the 5-day outlook (Fig. 3), a slight eastward shift of the upper ridge to the Great Lakes region in the 6-10-day period (Fig. 4), and a pattern change in the 11-15-day period as the upper ridge shifts into the West and a longwave trough emerges in the East (Fig. 5). Interestingly, the most confident forecast (55%) is the 6-10-day forecast of which 60% of the solution is from the ECMWF model. Although confidence is low (38%) for the 1-5-day projection, other members of the forecast ensemble are very close to the outlook pictured. Not surprisingly, the 34% confidence in the 11-15-day period is more realistically low (34%) and most heavily biased by GFS with 3 other members of the forecast ensemble indicating sharply different opinions on where the North America upper ridge aligns. While a trough in the East is the “most likely” upper air projection for the 11-15-day period, forecast confidence is low! Any 11-15-day forecast made during mid-spring should be used with caution due to diminished skill related to the transition of cold to warm season.

Fig. 3-5: Mega-cluster ensemble “most likely” 15-day upper air forecasts for North America.

The 1-5-day “most likely” percent of normal precipitation forecast is wet across the Gulf States and Northwest U.S. and dry in-between. Each model (ECM, GFS, and CMC) indicates this projection with ECM rendered most skillful at 50% (Fig. 6). The GFS operational forecast is the “caveat” opinion rendering 0.50 to 1.25 in. of rain in the Oklahoma to eastern Nebraska drought area. The ensemble is implying the GFS forecast is too wet.

In the 6-10-day forecast as the upper ridge shifts east, a return flow out of the western Gulf of Mexico develops to provide moisture for an approaching upper trough to cause rainfall across the North-central U.S. with intensity (Fig. 7). Some of that rain will occur in the core of the central and southern Great Plains drought area, but not much. The “most likely” forecast is the wettest of the 4 ensembles.

The 11-15-day outlook is favorable for rainfall in the core of the Great Plains drought zone (Fig. 8). Leading influence is the ECM with a projection of an inch or more rainfall across Kansas and Oklahoma.

Fig. 6-8: Mega-cluster ensemble “most likely” 15-day percent of precipitation forecasts for North America.