|Title||Contribution of the Interdecadal Pacific Oscillation to twentieth-century global surface temperature trends|
|Publication Type||Journal Article|
|Year of Publication||2016|
|Authors||Meehl G.A, Hu AX, Santer B.D, Xie SP|
|Journal||Nature Climate Change|
|Type of Article||Article|
|Keywords||20th; aerosol; Climate variability; decadal variability; hiatus periods|
Longer-term externally forced trends in global mean surface temperatures (GMSTs) are embedded in the background noise of internally generated multidecadal variability(1). A key mode of internal variability is the Interdecadal Pacific Oscillation (IPO), which contributed to a reduced GMST trend during the early 2000s(1-3). We use a novel, physical phenomenon-based approach to quantify the contribution from a source of internally generated multidecadal variability-the IPO-to multidecadal GMST trends. Here we show that the largest IPO contributions occurred in its positive phase during the rapidwarming periods from 1910-1941 and 1971-1995, with the IPO contributing 71% and 75%, respectively, to the difference between the median values of the externally forced trends and observed trends. The IPO transition from positive to negative in the late-1990s contributed 27% of the discrepancy between model median estimates of the forced part of the GMST trend and the observed trend from 1995 to 2013, with additional contributions that are probably due to internal variability outside of the Pacific(4) and an externally forced response from small volcanic eruptions(5). Understanding and quantifying the contribution of a specific source of internally generated variability-the IPO-to GMST trends is necessary to improve decadal climate prediction skill.
|Short Title||Nat. Clim. Chang.|
n this period, therefore, the CCSM4-based estimate of Interdecadal Pacific Oscillation (IPO) variability can account for 27% of the difference between the observed trends and the median of the unadjusted externally forced model trends. Similar results are obtained for a slightly different definition of the negative IPO period from 2000–2013 (Fig. 2b). We note that both the unadjusted and adjusted model results do not include the effects from a series of moderate volcanic eruptions in the early twenty-first century, and thus are likely to be biased warm. Accounting for these effects would probably bring both the unadjusted and adjusted multi-model median trends in Fig. 1b closer to the observed values. Thus, the results shown here indicate that discrepancies between simulated and observed trends over the early twenty-first century warming slowdown are likely to be attributable not only to the IPO, but also to volcanic (and other) external forcings, as well as to possible contributions from Atlantic SSTs associated with the Atlantic Multidecadal Oscillation (AMO, see Supplementary Information).