Search

A new detection and attribution method is presented and applied to the global mean surface air temperature (GSAT) from 1900 to 2014. The method aims at attributing the climate changes to the variations of greenhouse gases, anthropogenic aerosols, and natural forcings. A convolutional neural network (CNN) is trained using the simulated GSAT from historical and single‐forcing simulations of 12 climate models. Then, we perform a backward optimization with the CNN to estimate the attributable GSAT changes. Such a method does not assume additivity in the effects of the forcings. The uncertainty in the attributable GSAT is estimated by sampling different starting points from single‐forcing simulations and repeating the backward optimization. To evaluate this new method, the attributable GSAT changes are also calculated using the regularized optimal fingerprinting (ROF) method. Using synthetic non‐additive data, we first find that the neural network‐based method estimates attributable changes better than ROF. When using GSAT data from climate model, the attributable anomalies are similar for both methods, which might reflect that the influence of forcing is mainly additive for the GSAT. However, we found that the uncertainties given both methods are different. The new method presented here can be adapted and extended in future work, to investigate the non‐additive changes found at the local scale or on other physical variables. Plain Language Summary: In order to design effective adaptation policies, it is essential to have reliable estimates of the effect of anthropogenic activities on the climate. For that purpose, a new attribution method based on a neural network is designed and evaluated. The method estimates the past global mean surface air temperatures anomalies caused by the changes in the greenhouse gases concentration, the variation of anthropogenic aerosols, and the variations driven by naturally occurring phenomena. To build this estimation, the data from observations and climate models are used. This methodology is compared with another state‐of‐the‐art method. The results of both methods are evaluated and discussed. The proposed method provides better estimations in the case of large non‐additivity of the causes of climate change and can be applied to other physical variables or at the regional scale. In the case of the global mean surface air temperature, the method presented provides estimation similar to other methods. Key Points: We present a non linear method based on neural network to attribute the global mean surface air temperature variability to different forcingsWe use a convolutional neural network associated with a backward optimization to estimate the climate response to the different external forcingsThe attributable forcings are consistent with those obtained using another state‐of‐the‐art method [ABSTRACT FROM AUTHOR]

Climate reconstructions by data assimilation need to accommodate for sensitivities to proxies and prior estimates because models are uncertain and proxies are spatiotemporally limited. This study examines these sensitivities using multiple climate model simulations and different combinations of proxies (i.e., corals, ice cores, and tree‐ring cellulose). Experiments were conducted using an offline data assimilation approach; the results showed annual variations in the global distribution of surface air temperature and precipitation amount from 850 to 2000. Standard deviations of surface air temperature and precipitation amount during the entire period differed by up to 36% due to prior estimates. Experiments with different types of proxies showed that the El Niño‐like distribution of positive anomalies in the central to eastern tropical Pacific may only be adequately reproduced in experiments with corals, and not experiments without corals. The correlation coefficient of the NINO.3 index from 1971 to 2000 between experiments with corals and the Japanese 55‐year Reanalysis (JRA‐55) was 0.81 at maximum. By contrast, the correlation coefficient between experiments without corals and JRA‐55 was a maximum of 0.19. Key Points: Climate reconstruction by data assimilation illustrates annual variations in surface air temperature and precipitation from 850 to 2,000Spatiotemporal differences in climate reconstruction were compared based on prior estimatesThe impact of proxies on the results of past reproduced El Niño cases from 1971 to 2000 was validated [ABSTRACT FROM AUTHOR]

Plain Language Summary: Melting of the Antarctic ice sheet and shelf in the future will be influenced by interannual changes in the surface air temperature (SAT) in Antarctica. The SAT changes in Antarctica are related to variations in the Southern Hemisphere Annular Mode (SAM) during the austral summer. The SAM is a dominant pattern of atmospheric variability in the Southern Hemisphere and influences the Antarctic SAT with opposite changes between the northern Antarctic Peninsula (AP) and Eastern Antarctica (EA). To project future changes in the Antarctic SAT, we analyzed historical and future simulations from the Climate Model Intercomparison Project 5 models. We found that the degree of opposite interannual SAT changes between EA and the AP increases in the future due to intensified magnitude of the SAM‐related circulation anomalies, and summers of warmer SAT in the northern AP and cooler SAT in EA increase by 4% in the future compared to the historical period. This finding has major consequences for glacier melting in the northern AP in the future because more days of extremely high SAT in the northern AP may occur in the future. In this study, using the Climate Model Intercomparison Project 5 (CMIP5) simulations and by empirical orthogonal function (EOF) analysis, the first mode of variability in interannual surface air temperature (SAT) in Antarctica (EOF1) was examined for the period between 1979–2004 and 2051–2099 during the austral summer. The ensemble mean of EOF1 of the CMIP5 models shows a positive SAT anomaly over the northern Antarctic Peninsula (AP) and a negative SAT anomaly over Eastern Antarctica (EA) in both periods. A poleward expansion of the AP positive anomaly and an increase in the negative anomaly over interior EA are expected in 2051–2099, resulting in a larger difference of interannual SAT between interior EA and the AP in 2051–2099 than in 1979–2004. The increasing difference in the interannual SAT is consistent with a larger magnitude of the SAM‐related circulation anomalies in the future. Key Points: An increasing summer interannual surface air temperature (SAT) variability in interior Eastern Antarctica (EA) and the Antarctic Peninsula (AP) simulated in 2051–2099An increasing difference in summer interannual SAT between EA and the AP simulated in 2051–2099 than in 1979–2004The increasing difference in summer interannual SAT caused by a larger Southern Hemisphere Annular Mode‐related circulation anomaly in 2051–2099 than in 1979–2004 [ABSTRACT FROM AUTHOR]

We investigate the impacts of external forcings and internal climate variability (ICV) on global‐mean surface air temperature (GMST) variations (1850–2014) in observation and an ensemble of sixth Coupled Model Intercomparison Project (CMIP6) historical simulations. Using multiple regression models, we estimate that 83% of the observed GMST variance is explained by global‐scale anthropogenic forcing. Most CMIP6 models underestimate this anthropogenic contribution (median explained variance is 67%), but compare better with observation when the North Atlantic regional contribution is included in the anthropogenic forcing (median 77%). The ICV contributions to the GMST variance are ∼8% for observation and ∼7% for the models (median). The models' GMST responses to the external forcings and ICV vary widely, with the two responses being oppositely related. The North Atlantic predominantly drives the free GMST variability in observations, whereas the Pacific drives this in most CMIP6 models due to stronger than observed simulated GMST‐Pacific ICV relationships. Plain Language Summary: Global surface temperature has significantly warmed over the last one and a‐half centuries due to the combined effects of human produced greenhouse gases and aerosols and naturally occurring volcanic eruptions and solar output changes. These externally forced temperature changes are modulated by the climate system's internal variability. Most climate models simulate these different temperature changes, but with significant differences between the models' simulations. Here, we analyze data from observations and from simulations by 42 state‐of‐the‐art climate models to quantify the effects of external forcings and internal variability on surface temperature change. We find that most of the climate models underestimate the observed warming arising from global climate forcing but overestimate the multidecadal variations arising from North Atlantic aerosol forcing. The effect on temperature change from internal variability is much smaller than that from the combined external forcing, in both observations and climate models. Also, the models showing the largest temperature responses to external forcing show the smallest responses to internal variability, and vice‐versa. Understanding the reasons for these differences between climate simulations may help improve the confidence in climate change projections. Key Points: Most models underestimate (overestimate) the surface temperature response to global‐scale anthropogenic (North Atlantic regional) forcingThe models' diverse responses to external forcing have a strong negative relationship with the responses to internal variabilityThe North Atlantic (Pacific) internal variability dominates the unforced surface temperature variations in observations (most models) [ABSTRACT FROM AUTHOR]

Arctic sea‐ice extent (SIE) has declined drastically in recent decades, yet its evolution prior to the satellite era is highly uncertain. Studies using SIE observations find little variability prior to the 1970s; however, these reconstructions are based on limited data, especially prior to the 1950s. We use ensemble Kalman filter data assimilation of surface air temperature observations with Last Millennium climate model simulations to create a fully gridded Arctic sea‐ice concentration reconstruction from 1850 to 2018 and investigate the evolution of Arctic SIE during this period. We find a decline of ∼1.25×106 km2 during the early 20th‐century warming (1910–1940). The 25‐year trends during this period are ∼33–38% smaller than the satellite era (1979–2018) but almost twice as large as previous estimates. Additionally, we find that variability of SIE on decadal timescales prior to the satellite era is ∼40% greater than previously estimated. Plain Language Summary: Arctic sea ice is an important part of the climate system, serving as the interface between the ocean–atmosphere system. Arctic sea ice has undergone a rapid decline in recent decades, prompting the question of whether there have been changes of similar magnitude in the past. To answer such questions, a long record of sea ice is necessary, but spatially and temporally complete satellite observations are only available starting in 1979. Previous studies combining sea‐ice observations from various sources during the Instrumental Era (1850–2014) found little variability in sea‐ice extent prior to the satellite era, but data availability is limited prior to the 1950s. Here, we create an independent estimate of Arctic sea ice from 1850 to 2018 using a data assimilation approach that blends more abundant temperature observations with data from climate models. Our results show substantial loss of sea ice between 1910 and 1940, with a trend that is about ∼33–38% less than what has been observed in satellite observations. These results reinforce previous findings not only that the current trend is unprecedented in duration since 1850 but also that sea‐ice variability prior to 1979 is ∼40% larger than previously estimated. Key Points: Data assimilation is a skillful technique for reconstructing Arctic sea‐ice extent during the satellite eraReconstructed sea ice shows large declines in total Arctic sea‐ice extent during the early 20th‐century warming (1910–1940)Trends in total Arctic sea‐ice extent during the satellite era are ∼33–38% greater than during the early 20th‐century warming [ABSTRACT FROM AUTHOR]

In energetic equilibrium, the atmosphere's net radiative divergence ( [ABSTRACT FROM AUTHOR]

Retrospectively comparing future model projections to observations provides a robust and independent test of model skill. Here we analyze the performance of climate models published between 1970 and 2007 in projecting future global mean surface temperature (GMST) changes. Models are compared to observations based on both the change in GMST over time and the change in GMST over the change in external forcing. The latter approach accounts for mismatches in model forcings, a potential source of error in model projections independent of the accuracy of model physics. We find that climate models published over the past five decades were skillful in predicting subsequent GMST changes, with most models examined showing warming consistent with observations, particularly when mismatches between model‐projected and observationally estimated forcings were taken into account. Plain Language Summary: Climate models provide an important way to understand future changes in the Earth's climate. In this paper we undertake a thorough evaluation of the performance of various climate models published between the early 1970s and the late 2000s. Specifically, we look at how well models project global warming in the years after they were published by comparing them to observed temperature changes. Model projections rely on two things to accurately match observations: accurate modeling of climate physics and accurate assumptions around future emissions of CO2 and other factors affecting the climate. The best physics‐based model will still be inaccurate if it is driven by future changes in emissions that differ from reality. To account for this, we look at how the relationship between temperature and atmospheric CO2 (and other climate drivers) differs between models and observations. We find that climate models published over the past five decades were generally quite accurate in predicting global warming in the years after publication, particularly when accounting for differences between modeled and actual changes in atmospheric CO2 and other climate drivers. This research should help resolve public confusion around the performance of past climate modeling efforts and increases our confidence that models are accurately projecting global warming. Key Points: Evaluation of uninitialized multidecadal climate model future projection performance provides a concrete test of model skillThe quasi‐linear relationship between model/observed forcings and temperature change is used to control for errors in projected forcingModel simulations published between 1970 and 2007 were skillful in projecting future global mean surface warming [ABSTRACT FROM AUTHOR]