data set, HadISST1, and the nighttime marine air temperature (NMAT) data set, HadMAT1. HadISST1 replaces the global sea ice and sea surface temperature (GISST) data sets and is a unique combination of monthly globally complete fields of SST and sea ice concentration on a 1 ° latitude-longitude grid from 1871. The companion HadMAT1 runs monthly from 1856 on a 5 ° latitude-longitude grid and incorporates new corrections for the effect on NMAT of increasing deck (and hence measurement) heights. HadISST1 and HadMAT1 temperatures are reconstructed using a two-stage reducedspace optimal interpolation procedure, followed by superposition of quality-improved gridded observations onto the reconstructions to restore local detail. The sea ice fields are made more homogeneous by compensating satellite microwave-based sea ice concentrations for the impact of surface melt effects on retrievals in the Arctic and for algorithm deficiencies in the Antarctic and by making the historical in situ concentrations consistent with the satellite data. SSTs near sea ice are estimated using statistical relationships between SST and sea ice concentration. HadISST1 compares well with other published analyses, capturing trends in global, hemispheric, and regional SST well,

Estimating the mean and the covariance matrix of an incomplete dataset and filling in missing values with imputed values is generally a nonlinear problem, which must be solved iteratively. The expectation maximization (EM) algorithm for Gaussian data, an iterative method both for the estimation of mean values and covariance matrices from incomplete datasets and for the imputation of missing values, is taken as the point of departure for the development of a regularized EM algorithm. In contrast to the conventional EM algorithm, the regularized EM algorithm is applicable to sets of climate data, in which the number of variables typically exceeds the sample size. The regularized EM algorithm is based on iterated analyses of linear regressions of variables with missing values on variables with available values, with regression coefficients estimated by ridge regression, a regularized regression method in which a continuous regularization parameter controls the filtering of the noise in the data. The regularization parameter is determined by generalized cross-validation, such as to minimize, approximately, the expected mean squared error of the imputed values. The regularized EM algorithm can estimate, and exploit for the imputation of missing values, both synchronic and diachronic covariance matrices, which may contain information on spatial covariability, stationary temporal covariability, or cyclostationary temporal covariability. A test of the regularized EM algorithm with simulated surface temperature data demonstrates that the algorithm is applicable to typical sets of climate data and that it leads to more accurate estimates of the missing values than a conventional non-iterative imputation technique.

In this report, global coupled ocean–atmosphere models are used to explore possible mechanisms for observed decadal variability and trends in Pacific Ocean SSTs over the past century. The leading mode of internally generated decadal (�7 yr) variability in the model resembles the observed decadal variability in terms of pattern and amplitude. In the model, the pattern and time evolution of tropical winds and oceanic heat content are similar for the decadal and ENSO timescales, suggesting that the decadal variability has a similar ‘‘delayed oscillator’ ’ mechanism to that on the ENSO timescale. The westward phase propagation of the heat content anomalies, however, is slower and centered slightly farther from the equator (�12 � vs 9�N) for the decadal variability. Cool SST anomalies in the midlatitude North Pacific during the warm tropical phase of the decadal variability are induced in the model largely by oceanic advection anomalies. An index of observed SST over a broad triangular region of the tropical and subtropical Pacific indicates a warming rate of �0.41�C (100 yr) �1 since 1900, �1.2�C (100 yr) �1 since 1949, and �2.9�C (100 yr) �1 since 1971. All three warming trends are highly unusual in terms of their duration, with occurrence rates of less than 0.5 % in a 2000-yr simulation of internal climate variability using a low-resolution model. The most unusual is the trend since 1900 (96-yr duration): the longest simulated duration of a trend of this magnitude is 85 yr. This

Abstract. Analyses are conducted to assess whether simulated trends in SST and land surface air temperature from two versions of a coupled ocean-atmosphere model are consistent with the geographical distribution of observed trends over the period 1949-199'7. The simulated trends are derived from model experiments with both constant and time-varying radiative forcing. The models analyzed are low-resolution (R15,-4°) and medium-resolution (R30,-2°) versions of the Geophysical Fluid Dynamics Laboratory (GFDL) coupled climate mcodel. Internal climate variability is estimated from long control integrations of the models ~ith no change of external forcing. The radiatively forced trends are based on ensembles of integrations using estimated past concentrations of greenhouse gases and direct effects of anthropogenic sulfate aerosols (G+S). For the regional assessment, the observed trends at each grid point with adequate temporal coverage during 1949-1997 are first compared with the RI5 and R30 model unj'orced internal variability. Nearly 50 % of the analyzed areas have observed warming trends exceeding the 95th percentile of trends from the control simulations. These results suggest that regional warming trends over much of the globe during 1949-1997 are very unlikely to have occurred due to internal climate variability alone and suggest a role for a sustained positive thermal forcing such as increasing greenhouse gases.

Atlantic decadal climate variations are studied using marine meteorological observations. To remove artificial interhemispheric correlation, we perform empirical orthogonal function (EOF) analysis of sea surface temperature (SST) variability separately for the North and South Atlantic. The first EOF for the North (South) Atlantic in the decadal (8–16 years) band features a meridional tripole (dipole). In the tropics, the northern and southern leading EOFs form a meridional dipole with a center of action at 15 on either side of the equator. The leading sea level pressure (SLP) EOFs for the North and South Atlantic each feature a center of action that is displaced poleward of the tropical SST extreme, at 30 latitude. The SLP center of action in the North Atlantic has a barotropic structure and contributes significantly to surface wind variability in the tropics. Despite being derived from statistically independent data samples, the principle components for the leading SST and SLP EOFs (four in total) are significantly correlated with one another, indicative of the existence of an interhemispheric mode spanning the entire Atlantic Ocean. The same analysis for a longer SST record suggests that this pan-Atlantic decadal variability exists throughout the 20th century. In the North Atlantic, composite analysis of wind velocity and heat fluxes based on the PCs of the

Current attribution analyses that seek to determine the relative contributions of different forcing agents to observed near-surface temperature changes underestimate the importance of weak signals, such as that due to changes in solar irradiance. Here a new attribution method is applied that does not have a systematic bias against weak signals. It is found that current climate models underestimate the observed climate response to solar forcing over the twentieth century as a whole, indicating that the climate system has a greater sensitivity to solar forcing than do models. The results from this research show that increases in solar irradiance are likely to have had a greater influence on global-mean temperatures in the first half of the twentieth century than the combined effects of changes in anthropogenic forcings. Nevertheless the results confirm previous analyses showing that greenhouse gas increases explain most of the global warming observed in the second half of the twentieth century. 1.

Surface air temperatures from a 1000-yr integration of a coupled atmosphere–ocean model with constant forcing are analyzed by using a method that decomposes temperature variations into a component associated with a characteristic spatial structure and a residual. The structure function obtained from the coupled model output is almost identical to the so-called cold ocean–warm land (COWL) pattern based on observations, in which above-average spatial mean temperature is associated with anomalously cold oceans and anomalously warm land. This pattern features maxima over the high-latitude interiors of Eurasia and North America. The temperature fluctuations at the two continental centers exhibit almost no temporal correlation with each other. The temperature variations at the individual centers are related to teleconnection patterns in sea level pressure and 500-mb height that are similar to those identified in previous observational and modeling studies. As in observations, variations in the polarity and amplitude of this structure function are an important source of spatially averaged surface air temperature variability. Results from parallel integrations of models with more simplified treatments of the ocean confirm that the contrast in thermal inertia between land and ocean is the primary factor for the existence of the COWL pattern, whereas dynamical air–sea interactions do not play a significant role. The internally generated variability in

[1] Understanding Arctic temperature variability is essential for assessing possible future melting of the Greenland ice sheet, Arctic sea ice and Arctic permafrost. Temperature trend reversals in 1940 and 1970 separate two Arctic warming periods (1910–1940 and 1970–2008) by a significant 1940– 1970 cooling period. Analyzing temperature records of the Arctic meteorological stations we find that (a) the Arctic amplification (ratio of the Arctic to global temperature trends) is not a constant but varies in time on a multi-decadal time scale, (b) the Arctic warming from 1910–1940 proceeded at a significantly faster rate than the current 1970–2008 warming, and (c) the Arctic temperature changes are highly correlated with the Atlantic Multi-decadal Oscillation (AMO) suggesting the Atlantic Ocean thermohaline circulation is linked to the Arctic temperature variability on a multi-decadal time scale. Citation: Chylek, P., C. K. Folland,

An approach to identifying climate changes is presented that does not hinge on simulations of natural climate variations or anthropogenic changes. Observed interdecadal climate variations are decomposed into several discriminants, mutually uncorrelated spatiotemporal components with a maximal ratio of interdecadal-to-intradecadal variance. The dominant discriminants of twentieth-century variations in surface temperature exhibit large-scale warming in which, particularly in the Northern Hemisphere summer months, localized cooling is embedded. The structure of the large-scale warming is consistent with expected effects of increases in greenhouse gas concentrations. The localized cooling, with maxima on scales of 1000–2000 km over East Asia, eastern Europe, and North America, is suggestive of radiative effects of anthropogenic sulfate aerosols. 1.

We present a probabilistic factor analysis model which can be used for studying spatio-temporal datasets. The spatial and temporal structure is modeled by using Gaussian process priors both for the loading matrix and the factors. The posterior distributions are approximated using the variational Bayesian framework. High computational cost of Gaussian process modeling is reduced by using sparse approximations. The model is used to compute the reconstructions of the global sea surface temperatures from a historical dataset. The results suggest that the proposed model can outperform the state-of-the-art reconstruction systems. 1