This piece of work is part of a broader analysis performed on the North Atlantic Climate Variability and the ability of contemporary models to accurately predict it. The project took place under the supervision of Dr. Paulo Ceppi, as part of the Undergaduate Research Opportunity Placement (2021) at the Department of Physics, Imperial College London.
Motivation: Global mean temperature at the Earth’s surface responds both to externally imposed forcings, such as those arising from anthropogenic greenhouse gases, as well as to naturally-occurring patterns of atmospheric and oceanic variability. The latter, such as the North Atlantic Oscillation (NAO), orchestrate variations in weather and climate over much of the globe on inter annual and longer time scales.
The NAO is measured using a south(Azores)-minus-north(Iceland) index of atmospheric surface pressure variation across the NA and is closely connected with changes in the North Atlantic atmospheric polar jet stream and wider changes in atmospheric circulation. The physical, human, and biological impacts of NAO changes extend well beyond weather and climate, with major economic, social, and environmental effects.
Recent studies have highlighted that a large proportion of contemporary climate models exhibit too weak multi-decadal twentieth century NA climate variability when compared to observations. Diagnostics of the regional NA eddy-driven tropospheric westerly jet can help to improve understanding of key drivers and impacts of NAO variability.
Analysis:
Fig. 1: Historical winter mean U850 (line contours) and regression slopes (color fill) of time series of winter mean U850 at each grid point on (left) NAO index, (middle) jet latitude and (right) strength. All time series were linearly detrended prior to calculation of regression slopes.
Figure 1 shows the climatological winter mean lower tropospheric westerly wind structure of the NA region, for which strengthening and poleward shifting of the jet are both associated with a more positive NAO. A key point is that for a given change in the NAO the impacts or drivers of that change depend strongly on the ratio of associated jet shifting and/or strengthening, which exhibit contrasting spatial correlation patterns with local westerly wind across the NA and over Europe (compare Figures 1 (mid) and (right)).
This motivates us to carry out an analysis of NA jet variability across the Coupled Model Intercomparison Project Phase 5 (CMIP5) models to help identify to what extent the observed variations are reproduced across the CMIP5 historical simulations.
Fig. 2: Smoothed time series of winter NAO index, jet latitude and strength anomalies respectively, with CMIP5 ensemble members shown by decadally smoothed thin gray lines and the historical ensemble mean shown by the thick black lines (dashed shows decadally smoothed time series and dotted shows 30-year smoothing). The solid segments highlight periods containing the maximum 30-year linear trend. All available years are shown from each time series, hence the differing start and end dates in some cases. Histograms (Fig. 2d-f) of maximum 30-year linear trends of the smoothed and detrended CMIP5 time series shown in panels (a)–(c). Vertical black lines show maximum 30-year linear trends from each of the 20CRv2c ensemble members (i.e., trends during the periods highlighted by solid black lines in (a)-(c).
From Figures 2a–c there is evidence of contrasting amplitudes of variability in historical data compared to CMIP5 models. A similar contrast is also evident in terms of extreme 30-year linear trends (Figures 2d–f). The 30-year period 1965–1994 has previously been highlighted as exhibiting a large observed NAO increase, so this trend length was chosen in Figure 2 to compare with the frequency of occurrence in the CMIP5 models. Here maxima in rolling 30-year overlapping trends (starting just one year apart) for each CMIP5 realization are calculated from the decadally smoothed time series shown in Figures 2a–2d. The maximum of these overlapping trends is then extracted for each realization and used to populate the histograms shown in Figures 2e–2h. The same procedure is applied to the decadally smoothed historical dataand the resulting trends are shown by the vertical lines. By this measure, 30-year trends matching or exceeding the largest seen in the historical ensemble mean are less likely for jet strength and NAO index than for jet latitude.
Overall, the above results indicate that the too weak twentieth century multidecadal variability in the NA jet stream, as identified through the NAO in previous studies, is even more pronounced in terms of the jet strength. In terms of climate impacts, such biases in characteristics and strength of low-frequency jet variability are highly important.