We have developed a technique to quantify the effect of large-scale quasi-stationary wave anomalies on regional climate. The approach allows regional climate models to respond to external forcing associated with large-scale quasi-stationary waves that occur on scales larger than the regional model domain.
Time-averaged perturbations in the large-scale fields are identified by subtracting the averages of dynamic and thermodynamic fields for a specific episode (e.g., monthly averages during regional drought or flood) from long-term climatological means. The perturbation reflects the difference between climatologically averaged fields and those corresponding to the investigated anomalous periods and thus provides the external forcing accounting for the large-scale anomalies. Then integrations of a regional model are carried out in both control mode (the observed situation) and a filtered mode (in which the quasi-stationary waves were removed from the model initial and boundary conditions). Comparisons between both simulations allow quantification of the role of the large-scale forcing of the climatological anomaly as well as the contribution of regional processes to the anomaly.
The methodology is tested by application to one-month simulations for the drought of 1988 and flood of 1993 in the north-central U.S using the MM5 regional model. The control simulations successfully reproduced the anomalous drought and flood conditions. Then the simulations were redone with the observed large-scale quasi-stationary wave anomalies in the dynamic and thermodynamic fields filtered out of the time-dependent lateral boundary conditions (which were derived from the NCEP/NCAR reanalysis). In these filtered simulations the simulated meteorological fields over the drought (or flood) areas tended toward those for normal years. The differences between the control and filtered simulations indicate that the proposed filtering technique successfully identifies the regional response to anomalous external forcing.
Of particular interest is a dipole pattern of difference in the low-level wind between the control and filtered simulations, with opposite directions in the south- and the north-central U.S. For the 1993 flood the large-scale wave anomaly strengthens the low-level southerly flow in the south-central U.S. but weakens it in the north-central U.S., enhancing the moisture flux convergence over the flood region. The opposite dipole pattern in the low-level winds is apparent for the 1988 drought, suppressing the moisture flux divergence. The most likely cause for these opposing patterns is the effect of the large-scale perturbation on the location of the upper jet stream and the track of high frequency transient waves. The difference in low-level flow between the control and filtered simulations was stronger at night than during the daytime, implying that the response to the quasi-stationary wave anomalies is strongly modulated by diurnal boundary-layer processes. Analyses of moisture transport and moisture divergence fields for the various simulations (not shown) support these inferences.