Change in Midlatitude Flow Patterns 1948–2011

Change in Midlatitude Flow Patterns 1948-2011

Support for climate change has been increasing through a range of studies and observations in recent decades. One of the implications of climate change is the possibility of an increase in the frequency and/or intensity of extratropical cyclones. This paper examines trends in mid-latitude upper atmospheric data over 1948-2012, using the NCEP NCAR Reanalysis. A technique is derived to evaluate changes in the character of mid-latitude storms.

The Norwegian model for extratropical cyclone development suggests that storms increase in strength when they succeed in advecting cold air from polar regions equatorward upwind of the system and warm air from equatorial regions poleward downwind of the system. In the upper troposphere, the hypsometric and thermal wind effects result in the amplification of ridge/trough system associated with the polar and subtropical jet streams. The amplification of the jet stream can be quantified through a ratio of the meridional and zonal components of the jet stream across a region. Thus, the strongest extratropical cyclones are associated with highly amplified jet streams with a dominant meridional component.

The analysis of the jet stream flow pattern is performed on the 500 hPa pressure level. The measurement of the amplified nature of the wind flow is calculated as the ratio

                   (1)

This ratio of flow, like all ratios, suffers from an asymmetric nature. A dominant denominator has a domain from zero to one while a dominant numerator has a domain from one to infinity. To make the ratio symmetric, the ratio is adjusted as

                   (2)

This adjusted flow equation preserves the main features of the original ratio, but the domain is now -1 to +1 and is symmetric about zero. Negative (Positive) values correspond to predominately zonal (meridional) wind.

A region between 20N and 70N latitude, 60W and 125W longitude was analyzed at each time step (every 6 hours) for the duration of the data set (1948-2012). At each time step, the zonal nature of the 500 hPa wind flow was analyzed through equation (2). Performed throughout the data set, this produced 93,440 analyzed time periods. A histogram of the data across all years shows a bell-curve with the most frequent ratios being in a range of -0.49 and -0.44 (Fig. 1). Ratios higher and lower than the maximum reduce to zero at a minimum (maximum) adjusted ratio of -0.86 (0.55). As expected, a majority of flow regimes produce predominately zonal flow. It is of note that 5.2% of the time periods show a predominately meridional flow regime.

A linear regression analysis is performed to record the change in frequency of each ratio per year through the dataset. The trend in data show increasing (decreasing) frequency of ratios up to 0.25 greater (less) than the maximum frequency. Of note are the trends in the tails of the ratio histogram. Ratios indicating highly meridional wind flow between -0.15 and 0.25 increase in frequency with time. This range comprises 13.6% of the data, and so avoids errors associated with small datasets. While the pattern of linear change is not symmetrical with the left tail, the predominately zonal wind patterns are almost uniformly decreasing in frequency.

An analysis of the change in specific ratios shows the relationship of observed values to the linear rate of change (Figure 2). The adjusted ratio of -0.53 (-0.38) is representative of the most decreasing (increasing) frequency with time. Analysis of the observed data show the frequencies being nearly identical in the beginning of the trend and diverging in the 1990s and 2000s. 

The analyzed 500 hPa midlatitude wind flow indicates a trend toward increasing frequency of highly meridional flow, indicative of trough/ridge amplification associated with extratropical cyclones. The observations support an increasing frequency of the potential for mature extratropical cyclones over recent decades. This conclusion assumes no systematic changes in data evaluation by the NCEP NCAR Reanalysis in later decades.

Figure 1 – Filled area - Histogram of adjusted ratio frequency over the duration of the dataset. Solid line – Linear change of each ratio frequency per year. Dashed line – Centered average of change in frequency for each ratio and the four ratios higher and lower than the average. For example, the centered average at the ratio 0.10 is the average of the frequency changes for 0.06 through 0.14.

Figure 2 – Black line – Frequency of -0.53 adjusted ratio each year from 1948 through 2011. Thick black line – Linear regression line for the -0.53 time series. Thin gray line – Frequency of -0.38 ratio each year from 1948 through 2011. Thick gray line – Linear regression line for the -0.38 time series.