Attribution of Observed SST Trends and Sub-Continental Land Warming to Anthropogenic Forcing during 1979–2005

AR4 concluded that significant surface air temperature warming since 1950 was very likely attributed to anthropogenic forcing on global and continental scales, but some difficulties and uncertainties remained in attribution of temperature warming in the sub-continental scales and at time scales less than 50-yr. Previous studies indicate that AMIP experiments from some models reproduce observed temperature increase over land and suggest that land warming has occurred largely in response to a worldwide warming of the oceans, while the oceans may themselves have warmed from a combination of natural and anthropogenic influences. The main purpose here is to attribute global and regional SST trends as well as sub-continental land warming since 1979. Land temperature from CRUTEM4, SST from HadSST3 and outputs of all Coupled Model Intercomparison Project Phase 5 (CMIP5) during 1979-2005 are analyzed.

At the 5% level, a two-side consistent test is conducted to determine whether the difference between the observed warming trend and the multi-model ensemble-mean trend from the Historical experiments is significantly different from zero at each of various spatial scales. Overall, the simulated and observed SST trends are statistically consistent at the global, hemispheric, and basin scales at Tropical Indian, Warm pool, Tropical and North Atlantic. The simulated and observed SST evolutions are also highly correlated with the observations at these scales. The difference between observed and simulated SST trends resembles a Pacific-Decadal Oscillation (PDO) pattern, with the observed trends significantly larger (smaller) than the corresponding models trends mostly over the Tropical eastern Pacific (western North and South Pacific). But the fraction of the grid boxes where the observed SST trend is not consistent with the multi-model ensemble-mean response to all forcings over the 27-year period is only 18%, which can be explained by internal climate variations. We thus conclude that the observed SST trends are very likely to be warmed by increased greenhouse gases which are directly related to anthropogenic activities.

At the 95% confidence level, AMIP and Historical experiments simulate (1) consistent trends in surface warming and downward long-wave radiation (Rlds) each other at the global, hemispheric scales, and over all 21 sub-continental regions; and (2) significant warming trends consistent with observations at the global, hemispheric scales, and over 19 of 21 sub-continental regions. At any spatial scales, Rlds plays a dominant role in temperature increase. PDO (AMO) seems to contribute to observed land warming over East Asia (Europe), causing the warming trend in the Historical runs is statistically different from that in the observation. Therefore, in most of sub-continentals, increasing of greenhouse-gas concentrations tends to be the ultimate source of land warming, while the warmed ocean plays an indirect but important role during such progress through hydrodynamic-radiative teleconnections, by moistening and warming the air over land and increasing surface Rlds.