INTERNAL CLIMATE VARIABILITY

Posted by: chaamjamal on: July 16, 2020

• In: Uncategorized
 
• 3 Comments

THIS POST IS A LITERATURE REVIEW OF RECENT STUDIES ON THE INTERNAL VARIABILITY OF CLIMATE AS SEPARATE FROM THE IMPACT OF FOSSIL FUEL EMISSIONS. THE REVIEW IS PRESENTED AS A RELEVANT BIBLIOGRAPHY ALONG WITH A MORE DETAILED LOOK AT THESE TWO RECENT PAPERS ON THE SUBJECT.

(1) Insights from Earth system model initial-condition large ensembles and future prospects, Clara Deser,et al: Nature Climate Change, 2020.   

ABSTRACT: Internal variability in the climate system confounds assessment of human-induced climate change and imposes irreducible limits on the accuracy of climate change projections, especially at regional and decadal scales. A new collection of initial-condition large ensembles generated with seven Earth system models under historical and future radiative forcing scenarios provides new insights into uncertainties due to internal variability versus model differences. These data enhance the assessment of climate change risks, including extreme events, and offer a powerful test-bed for new methodologies aimed at separating forced signals from internal variability in the observational record. Opportunities and challenges confronting the design and dissemination of future LEs, including increased spatial resolution and model complexity alongside emerging Earth system applications, are discussed.

(2) Quantifying the role of internal variability in the temperature we expect to observe in the coming decades Nicola Maher et al, 2020. 

ABSTRACT: On short (15-year) to mid-term (30-year) time-scales how the Earth’s surface temperature evolves can be dominated by internal variability as demonstrated by the global-warming pause or ‘hiatus‘. In this study, we use six single model initial-condition large ensembles (SMILEs) and the Coupled Model Intercomparison Project 5 (CMIP5) to visualise the role of internal variability in controlling possible observable surface temperature trends in the short-term and mid-term projections from 2019 onwards. We confirm that in the short-term, surface temperature trend projections are dominated by internal variability, with little influence of structural model differences or warming pathway. Additionally we demonstrate that this result is independent of the model-dependent estimate of the magnitude of internal variability. Indeed, and perhaps counter intuitively, in all models a lack of warming, or even a cooling trend could be observed at all individual points on the globe, even under the largest greenhouse gas emissions. The near-equivalence of all six SMILEs and CMIP5 demonstrates the robustness of this result to the choice of models used. On the mid-term time-scale, we confirm that structural model differences and scenario uncertainties play a larger role in controlling surface temperature trend projections than they did on the shorter time-scale. In addition we show that whether internal variability still dominates, or whether model uncertainties and internal variability are a similar magnitude, depends on the estimate of internal variability, which differs among the SMILEs. Finally we show that even out to thirty years large parts of the globe (or most of the globe in MPI-GE and CMIP5) could still experience no-warming due to internal variability.  (FULL TEXT POSTED BELOW).

RELATED POST ON CORRELATION OF CMIP5 FORCINGS WITH TEMPERATURE

RELATED POST: HOLOCENE WARMING & COOLING CYCLES

RELATED POST: TIDAL CYCLES AND KEELING & WHORF 2000

RELATED POST: CARL WUNSCH ON UNCERTAIN DATA  

RELATED POST ON TEMPERATURE DATA  

IN THE LAST RELATED POST LINKED ABOVE WE SHOW THAT EVEN STRONG LONG TERM WARMING TRENDS REVEAL VIOLENT DECADAL VARIABILITY THAT CONTAIN WARMING, COOLING AND NO TREND CONDITIONS. THIS DEMONSTRATION PROVIDES EMPIRICAL EVIDENCE THAT DECADAL VARIABILITY MUST BE UNDERSTOOD AS INTERNAL CLIMATE VARIABILITY. ONE OF THESE CHARTS IS REPRODUCED BELOW.

CRITICAL COMMENTARY

The essential findings of the internal climate variability papers presented above and cited below is that although climate models programmed with AGW climate forcings perform reasonably well at longer time spans greater than 60 years, their short term performance at smaller time spans is not statistically significant and the predictions are therefore subject to large uncertainties. These results from a related post [LINK]  are summarized in the tables below.

Comparing the correlation of the HADCRUT temperature reconstructions with AGW forcings at three different time spans, we find a very strong statistically significant detrended correlation at the full span of the data of 162 years. The detrended correlation values are ρ=0.593 for well mixed greenhouse gas forcings (WMGHG) and oddly, somewhat lower but still strong and statistically significant detrended correlation of ρ=0.513 for ALLFORCINGS that includes natural drivers of climate.

Figure 1: Full Span: Correlation between AGW climate forcings and temperature data

In the half span of the data of 81 years, we find a lower detrended correlation of ρ=0.252 for WMGHG and ρ=0.307 for ALLFORCINGS. These 81-year correlations are much higher in the second half and comparable with the full span detrended correlations with values of  ρ=0.548 for WMGHG and ρ=0.510 for ALLFORCINGS. The much higher and statistically significant correlations at the 81-year time span in the second half of the time span implies better support for AGW theory in the 2nd half of the time span beginning in 1932 than in the first half that begins in 1851. This comparison implies that the AGW theory is evident at the 81-year time span and that therefore there is not much evidence for it in the first half.

The effect of the length of the time span on correlation is seen clearly in Figure 3. Here, in a moving 30-year window, the average detrended correlation for WMGHG is ρ=0.278 for WMGHG and ρ=0.372 for ALLFORCINGS. Neither correlation is statistically significant.

Figure 2: First Half: Correlation between AGW climate forcings and temperature data

Figure 3: 2nd Half: Correlation between AGW climate forcings and temperature data

Figure 4: 30yr moving window: Correlation between AGW climate forcings & temperature

The internal variability literature in climate science presented in the two reference papers above and in the bibliography below is a generalization of the phenomenon seen in the correlation analysis that the strong statistically significant correlation between forcings and temperature seen at long time spans is not found in short time spans, particularly so at a time span of 30 years. In fact the findings in the internal variability papers also identifies the 30-year time span as the demarcation where the time span is too short to describe climate in terms of AGW forcings and is therefore driven by internal variability.

Therefore, what we find at the smaller time spans, is that climate variables such as temperature and precipitation contain large variances that impose a large range in the forecast. The large range in temperature forecast may for example also include cooling at the low end. These forecasts cannot be taken literally. Large variances and their large confidence intervals do not mean that we know what will happen and that those extremes are in the forecast. Large variance means that we have reached the limit of what we know and what we can forecast and that these forecasts have no information value. Specifically, a large internal variability forecast that includes cooling at the low end is not a global cooling forecast but an expression of the limit of useful information.

The internal variability constraint in climate science described in these papers applies not only to to the brevity of time spans but also the limit in geographical spans. The internal variability limit of AGW forcings to make climate forecasts is thus found in both short time spans and small geographical extents of the globe for which the forecast is being made. This constraint marks the limit of climate models to make forecasts of how AGW climate change will evolve in the future.

The essence of the internal variability issue is that AGW climate science is a system of making long range forecasts for global mean temperature and its extension to shorter time spans or regional climate is not possible because shorter time spans and regional climate are driven mostly by internal climate variability that is beyond AGW climate science.

WITH REGARD TO EVENT ATTRIBUTION ANALYSIS AND THE EXTREME WEATHER HYPOTHESIS  [LINK],  Internal climate variability limits the ability of climate science to attribute localized extreme weather events to anthropogenic global warming. Localized means geographically limited and Event means time span limited. 

As for example, was the hot dry weather in Australia that’s credited with the bush fires of 2019, a global warming event?  Before climate science acknowledged the internal variability issue, the answer was yes. After climate science acknowledged the internal variability issue, the answer is we don’t know.

The same goes for other event attribution studies [LINK] in which all regional bad and harmful weather events such as forest fires, heat waves, droughts, and floods have been attributed to WMGHG driven global warming. Because these climate/weather events are both time and geography constrained, the Internal Climate Variability finding limits their interpretation as creations of anthropogenic global warming. 

RELEVANT BIBLIOGRAPHY

1. Wigley, T. M. L., and S. C. B. Raper. “Natural variability of the climate system and detection of the greenhouse effect.” Nature 344.6264 (1990): 324-327. Abstract:  Global mean temperatures show considerable variability at all timescales. The causes of this variability are usually classified as external or internal, and the variations themselves may be usefully subdivided into low-frequency variability (timescale ≳= 10 years) and high-frequency variability (≲=10 years). Virtually nothing is known about the nature or magnitude of internally generated, low-frequency variability. There is some evidence from models, however, that this variability may be quite large, possibly causing fluctuations in global mean temperature of up to 0.4 °C over periods of thirty years or more. Here we show how the ocean may produce low-frequency climate variability by passive modulation of natural forcing, to produce substantial trends in global mean temperature on the century timescale. Simulations with a simple climate model are used to determine the main controls on internally generated low-frequency variability, and show that natural trends of up to 0.3 °C may occur over intervals of up to 100 years. Although the magnitude of such trends is unexpectedly large, it is insufficient to explain the observed global warming during the twentieth century.
2. Christensen, O. B., et al. “Internal variability of regional climate models.” Climate Dynamics 17.11 (2001): 875-887.  Abstract:  Two regional climate models have been applied to the task of generating an ensemble of realizations of the year 1982 with observed boundary conditions in areas covering parts of the Mediterranean countries. These realizations were generated by applying boundary conditions from the ECMWF ERA reanalysis project consecutively, carrying over the soil variables from the regional models from one iteration to the next. Monthly mean fields for six iterations of each model have been used as statistical ensembles in order to investigate the internal variability of the regional model dynamics. This internal variability is a necessary consequence of the non-linear physical feedback mechanisms of the RCM being active. A small value of internal variability will give better statistics for climate sensitivity signals, but will make these results less credible. The internal variability is important for the quantitative assessment of a climate sensitivity signal. With the present choice of models and integration domains the internal variabilities of surface fields and precipitation do reach levels that are less than, but in summer of comparable order of magnitude to, corresponding atmospheric variabilities of an atmospheric general circulation model.
3. Reichert, Bernhard K., Lennart Bengtsson, and J. Oerlemans. “Recent glacier retreat exceeds internal variability.” Journal of Climate 15.21 (2002): 3069-3081.  Abstract:  Glacier fluctuations exclusively due to internal variations in the climate system are simulated using downscaled integrations of the ECHAM4/OPYC coupled general circulation model (GCM). A process-based modeling approach using a mass balance model of intermediate complexity and a dynamic ice flow model considering simple shearing flow and sliding are applied. Multimillennia records of glacier length fluctuations for Nigardsbreen (Norway) and Rhonegletscher (Switzerland) are simulated using autoregressive processes determined by statistically downscaled GCM experiments. Return periods and probabilities of specific glacier length changes using GCM integrations excluding external forcings such as solar irradiation changes, volcanic, or anthropogenic effects are analyzed and compared to historical glacier length records. Preindustrial fluctuations of the glaciers as far as observed or reconstructed, including their advance during the “Little Ice Age,” can be explained by internal variability in the climate system as represented by a GCM. However, fluctuations comparable to the present-day glacier retreat exceed any variation simulated by the GCM control experiments and must be caused by external forcing, with anthropogenic forcing being a likely candidate.
4. Yoshimori, Masakazu, et al. “Externally forced and internal variability in ensemble climate simulations of the Maunder Minimum.” Journal of Climate 18.20 (2005): 4253-4270.  Abstract:  The response of the climate system to natural, external forcing during the Maunder Minimum (ca. a.d. 1645–1715) is investigated using a comprehensive climate model. An ensemble of six transient simulations is produced in order to examine the relative importance of externally forced and internally generated variability. The simulated annual Northern Hemisphere and zonal-mean near-surface air temperature agree well with proxy-based reconstructions on decadal time scales. A mean cooling signal during the Maunder Minimum is masked by the internal unforced variability in some regions such as Alaska, Greenland, and northern Europe. In general, temperature exhibits a better signal-to-noise ratio than precipitation. Mean salinity changes are found in basin averages. The model also shows clear response patterns to volcanic eruptions. In particular, volcanic forcing is projected onto the winter North Atlantic Oscillation index following the eruptions. It is demonstrated that the significant spread of ensemble members is possible even on multidecadal time scales, which has an important implication in coordinating comparisons between model simulations and regional reconstructions.  [FULL TEXT]
5. Alexandru, Adelina, Ramon de Elia, and René Laprise. “Internal variability in regional climate downscaling at the seasonal scale.” Monthly Weather Review 135.9 (2007): 3221-3238. Abstract: To study the internal variability of the model and its consequences on seasonal statistics, large ensembles of twenty 3-month simulations of the Canadian Regional Climate Model (CRCM), differing only in their initial conditions, were generated over different domain sizes in eastern North America for a summer season. The degree of internal variability was measured as the spread between the individual members of the ensemble during the integration period. Results show that the CRCM internal variability depends strongly on synoptic events, as is seen by the pulsating behavior of the time evolution of variance during the period of integration. The existence of bimodal solutions for the circulation is also noted. The geographical distribution of variance depends on the variables; precipitation shows maximum variance in the southern United States, while 850-hPa geopotential height exhibits maximum variance in the northeast part of the domain. Results suggest that strong precipitation events in the southern United States may act as a triggering mechanism for the 850-hPa geopotential height spread along the storm track, which reaches its maximum toward the northeast of the domain. This study reveals that successive reductions of the domain size induce a general decrease in the internal variability of the model, but an important variation in its geographical distribution and amplitude was detected. The influence of the internal variability at the seasonal scale was evaluated by computing the variance between the individual member seasonal averages of the ensemble. Large values of internal variability for precipitation suggest possible repercussions of internal variability on seasonal statistics.
6. Deser, Clara, et al. “Uncertainty in climate change projections: the role of internal variability.” Climate dynamics 38.3-4 (2012): 527-546.  Abstract: Uncertainty in future climate change presents a key challenge for adaptation planning. In this study, uncertainty arising from internal climate variability is investigated using a new 40-member ensemble conducted with the National Center for Atmospheric Research Community Climate System Model Version 3 (CCSM3) under the SRES A1B greenhouse gas and ozone recovery forcing scenarios during 2000–2060. The contribution of intrinsic atmospheric variability to the total uncertainty is further examined using a 10,000-year control integration of the atmospheric model component of CCSM3 under fixed boundary conditions. The global climate response is characterized in terms of air temperature, precipitation, and sea level pressure during winter and summer. The dominant source of uncertainty in the simulated climate response at middle and high latitudes is internal atmospheric variability associated with the annular modes of circulation variability. Coupled ocean-atmosphere variability plays a dominant role in the tropics, with attendant effects at higher latitudes via atmospheric teleconnections. Uncertainties in the forced response are generally larger for sea level pressure than precipitation, and smallest for air temperature. Accordingly, forced changes in air temperature can be detected earlier and with fewer ensemble members than those in atmospheric circulation and precipitation. Implications of the results for detection and attribution of observed climate change and for multi-model climate assessments are discussed. Internal variability is estimated to account for at least half of the inter-model spread in projected climate trends during 2005–2060 in the CMIP3 multi-model ensemble. [FULL TEXT]
7. Collins, Matthew, et al. “Long-term climate change: projections, commitments and irreversibility.” Climate Change 2013-The Physical Science Basis: Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, 2013. 1029-1136.   (CHAPTER IN A BOOK): This chapter assesses long-term projections of climate change for the end of the 21st century and beyond, where the forced signal depends on the scenario and is typically larger than the internal variability of the climate system. Changes are expressed with respect to a baseline period of 1986–2005, unless otherwise stated.
8. Deser, Clara, et al. “Projecting North American climate over the next 50 years: Uncertainty due to internal variability.” Journal of Climate 27.6 (2014): 2271-2296.  Abstract: This study highlights the relative importance of internally generated versus externally forced climate trends over the next 50 yr (2010–60) at local and regional scales over North America in two global coupled model ensembles. Both ensembles contain large numbers of integrations (17 and 40): each of which is subject to identical anthropogenic radiative forcing (e.g., greenhouse gas increase) but begins from a slightly different initial atmospheric state. Thus, the diversity of projected climate trends within each model ensemble is due solely to intrinsic, unpredictable variability of the climate system. Both model ensembles show that natural climate variability superimposed upon forced climate change will result in a range of possible future trends for surface air temperature and precipitation over the next 50 yr. Precipitation trends are particularly subject to uncertainty as a result of internal variability, with signal-to-noise ratios less than 2. Intrinsic atmospheric circulation variability is mainly responsible for the spread in future climate trends, imparting regional coherence to the internally driven air temperature and precipitation trends. The results underscore the importance of conducting a large number of climate change projections with a given model, as each realization will contain a different superposition of unforced and forced trends. Such initial-condition ensembles are also needed to determine the anthropogenic climate response at local and regional scales and provide a new perspective on how to usefully compare climate change projections across models. [FULL TEXT]
9. Wang, Zhiyuan, et al. “Global climate internal variability in a 2000-year control simulation with Community Earth System Model (CESM).” Chinese Geographical Science 25.3 (2015): 263-273. Abstract: Using the low-resolution (T31, equivalent to 3.75° × 3.75°) version of the Community Earth System Model (CESM) from the National Center for Atmospheric Research (NCAR), a global climate simulation was carried out with fixed external forcing factors (1850 Common Era. (C.E.) conditions) for the past 2000 years. Based on the simulated results, spatio-temporal structures of surface air temperature, precipitation and internal variability, such as the El Niño-Southern Oscillation (ENSO), the Atlantic Multi-decadal Oscillation (AMO), the Pacific Decadal Oscillation (PDO), and the North Atlantic Oscillation (NAO), were compared with reanalysis datasets to evaluate the model performance. The results are as follows: 1) CESM showed a good performance in the long-term simulation and no significant climate drift over the past 2000 years; 2) climatological patterns of global and regional climate changes simulated by the CESM were reasonable compared with the reanalysis datasets; and 3) the CESM simulated internal natural variability of the climate system performs very well. The model not only reproduced the periodicity of ENSO, AMO and PDO events but also the 3–8 years variability of the ENSO. The spatial distribution of the CESM-simulated NAO was also similar to the observed. However, because of weaker total irradiation and greenhouse gas concentration forcing in the simulation than the present, the model performances had some differences from the observations. Generally, the CESM showed a good performance in simulating the global climate and internal natural variability of the climate system. This paves the way for other forced climate simulations for the past 2000 years by using the CESM.
10. Thompson, David WJ, Clara Deser, et al. “Quantifying the role of internal climate variability in future climate trends.” Journal of Climate 28.16 (2015): 6443-6456. Abstract:  Internal variability in the climate system gives rise to large uncertainty in projections of future climate. The uncertainty in future climate due to internal climate variability can be estimated from large ensembles of climate change simulations in which the experiment setup is the same from one ensemble member to the next but for small perturbations in the initial atmospheric state. However, large ensembles are susceptible to model bias. Here the authors outline an alternative approach for assessing the role of internal variability in future climate based on a simple analytic model and the statistics of the unforced climate variability. The analytic model is derived from the standard error of the regression and assumes that the statistics of the internal variability are roughly Gaussian and stationary in time. When applied to the statistics of an unforced control simulation, the analytic model provides a remarkably robust estimate of the uncertainty in future climate indicated by a large ensemble of climate change simulations. To the extent that observations can be used to estimate the amplitude of internal climate variability, it is argued that the uncertainty in future climate trends due to internal variability can be robustly estimated from the statistics of the observed climate. [FULL TEXT] .

FULL TEXT OF MAHER ETAL 2020

1. Introduction: Short-term trends in climate indices, such as global-mean surface temperature are significantly influenced by internal variability (e.g. Hawkins and Sutton 2009, Marotzke and Forster 2015). This means that although greenhouse gas emissions are ever increasing, we may observe a global cooling trend over the coming decade, as demonstrated by the recent global warming slowdown or hiatus. Conversely, we could also observe a decade of accelerated warming that overshoots what we would expect due to the current emissions (Meehl et al 2013). In this paper we visually demonstrate the role of internal variability in the temperatures that will be observed at each point on the globe in the coming decades and confirm the dominance of internal variability in the short-term trends. To do this we use a combination of six single model initial-condition large ensembles (SMILEs) and the Coupled Model Intercomparison Project 5 (CMIP5) archive to investigate the range of projected temperature trends from 2019 onwards. Unlike previous studies, before the availability of many SMILEs, we are able to additionally demonstrate the effect of the uncertainty in the magnitude of internal variability itself on our results.

Internal variability, or chaotic variability of the climate system (Hasselmann 1976) is a difficult concept to communicate (Deser et al 2012a). It is often explained in terms of the “Butterfly Effect”, where a small change in the present can result in a much larger change in the future state. It is also a difficult concept to study due to the short, and spatially inconsistent observations. Indeed, to truly study the observed internal variability of Earth’s surface temperature one must have long observational records under many different climate conditions, so as to be able to sample the internal variability.

Practically, internal variability can be quantified and studied using climate models, with SMILEs effective tools to quantify the role of small perturbations in changing the short and long-term trajectory of the climate system. Individual SMILEs have been used in previous studies to investigate the role of internal variability in driving surface temperature projections, mainly on 35-60 year time-scales (Deser et al 2012b, Kay et al 2015, Deser et al 2016, Bengtsson and Hodges 2018), with few studies investigating the shorter time-scales (e.g. Marotzke 2019).

To date only one study, which focuses on North America, has investigated 60-year surface temperature trends from multiple SMILEs (Deser et al 2020). Importantly this study has demonstrated that the internal variability of these trends differs between SMILEs (Deser et al 2020). Indeed both Hawkins and Sutton (2009) and Kumar and Ganguly (2017) demonstrated that model differences dominate temperature trends on longer time-scales, with internal variability dominating on shorter time-scales. As such using many SMILEs is key to identifying uncertainties in both the magnitude of internal variability and the forced response due to model differences.

It is also unclear how the rate of anthropogenic greenhouse gas emissions will evolve over the coming decades. The last generation of climate models were run with four different possible futures; RCP2.6, 4.5, 6.0, and 8.5. Scientists have suggested that these scenarios cover the likely range of the possible greenhouse gas emissions for the coming century, however the true pathway will depend on the policy changes made by governments. This pathway is known to be important for long-term projections, however, it has been found to be less important on short-term time-scales (Hawkins and Sutton 2009). Indeed when changes in RCP2.6 were compared to a RCP4.5 scenario in the Max Planck Institute Grand Ensemble (MPI-GE) a large overlap in global temperatures in the short-term projections was found (Marotzke 2019). When extreme temperatures were considered in the Community Earth System Model Large Ensemble (CESM-LE) under RCP4.5 and RCP8.5 scenarios, statistically significant differences were found only in 2050 (Lehner et al 2016, Tebaldi and Wehner 2018), again demonstrating that pathway differences can be less important on short to mid-term time-scales.

Previous studies have focused on detecting the AGW signal as a way of attribution to fossil fuel emissions (e.g. Stone et al 2007). Other studies use large ensembles to identify when a signal will emerge from the noise or internal variability. This is known as the time of emergence (e.g. Hawkins and Sutton 2012, Tebaldi and Friedlingstein 2013). In this study we turn this concept around to look not at when a signal will emerge or when it can be detected and attributed, but how the simulated internal variability can influence observed climate in the coming decades. This has importance for policy makers in determining the range of possible futures that could be observed.

The purpose of this paper is threefold. We will: (a) Visually demonstrate the role of internal variability in driving the observed climate. (b)Illustrate the maximum and minimum trends possible at each point of the globe on both short-term (15-year) and mid-term (30-year) time-scales. (c) Investigate the point-wise relative importance of internal variability, scenario uncertainties and model differences in controlling temperature trends on both the short and mid-term time-scales. Here, we include a new estimate of how model differences in the quantification of internal variability affect the relative importance of these quantities. 2. Models In this study we use 6 SMILEs to investigate internal variability of surface temperature trends (skin temperature; ts). The SMILEs are all CMIP5 class models run with CMIP5 forcing:

The Max Planck institute Grand Ensemble (MPI-GE) (Maher et al 2019). This model has 100 ensemble members available for RCP2.6, RCP4.5 and RCP8.5 scenarios. The Canadian Earth System Model Large Ensembles (CanESM2-LE) (Kirchmeier-Young et al 2017). This model has 50 ensemble members available for RCP8.5. The Large Ensemble Community Project (CESM-LE) (Kay et al 2015). This model has 40 members available for RCP8.5.
The Commonwealth Scientific and Industrial Research Organisation Large Ensemble (CSIRO-Mk3.6-LE). (Jeffrey et al 2012). This model has 30 members for RCP8.5.
Geophysical Fluid Dynamics Laboratory Earth System Model Large Ensemble (GFDL-ESM2M-LE). (Rodgers et al 2015). This model has 30 members for RCP8.5.
Geophysical Fluid Dynamics Laboratory Large Ensemble (GFDL-CM3-LE) (Sun et al 2018). This model has 20 members for RCP8.5.

We additionally use all available ensemble members from the CMIP5 archive (Supplementary table 1), which have the field surface temperature (in CMIP; ts) available. We do not apply any model weighting. We do not select for ensemble members that exist for all scenarios (RCP2.6, RCP4.5 and RCP8.5 in this study), which means that different models and different numbers of ensemble members may be used for each scenario.

Short-term projections (2019-2034):  The mean short-term trend at each point on the globe (2019-2034) and the trend at each grid-point when the global surface temperature trend is both maximum and minimum is demonstrated for the mean of the SMILEs and CMIP5 in figure 1. We find that the SMILEs broadly replicate the CMIP5 response, despite consisting of only 6 models. This highlights the large role of internal variability in driving the CMIP5 spread on short time-scales. The main differences between CMIP5 and the SMILEs are found at high-latitudes when the global surface temperature trend is minimum. In this case CMIP5 shows larger cooling than the SMILEs, suggesting that in this case the SMILEs do not quite cover the range of possible model results at high-latitudes. When the global surface temperature trend is minimum, both CMIP5 and the SMILEs show a negative Interdecadal Pacific Oscillation (IPO) like pattern (figure 1; top row, individual models in Supplementary figures 2 and 3), while when the global surface temperature trend is maximum all maps show a positive IPO-like pattern (figure 1; middle row, individual models in Supplementary figures 2 and 3). This result agrees well with Meehl et al (2013) and Maher et al (2014), who demonstrated for CCSM4 and CMIP5, respectively, that decades of cooling resemble a negative IPO-like pattern and decades of accelerated warming tend to resemble a positive IPO-like pattern.
We next investigate the relative importance of internal variability, model structural differences and scenario uncertainty by completing a decomposition similar to Hawkins and Sutton (2009) (figure 2; left column). Figure 2 demonstrates that internal variability dominates the short-term trend in temperature at all grid points, confirming the results of Hawkins and Sutton (2009), with both a newer generation of models and at a higher resolution. The near equivalence of each of the SMILEs and CMIP5 (figure 1; Supplementary figure 2 and 3) confirms the robustness of this result and demonstrates that the conclusions drawn from the SMILEs can be extended to the larger CMIP5 archive. Building on this confirmation, when we investigate the sensitivity of this result to the uncertainty in internal variability itself, we find that the dominance of internal variability in comparison to the other two uncertainties holds if we sample for either the maximum or minimum variability estimate from the SMILEs (figure 3). This emphasises the robustness of the dominance of internal variability on the short-term time-scale.

We also visualise what the largest and smallest trends at any given location on the globe could be (note that these trends are very unlikely to occur at the same time across the globe), and determine the likelihood of warming occurring on a short-term time-scale at each location (figure 4). We present these results using MPI-GE (RCP2.6 and RCP8.5) and CESM-LE (RCP8.5), as these models represent the spread of all of the SMILEs with RCP2.6 and RCP8.5 covering the scenario spread (individual model results; Supplementary figures 4 and 5). We confirm, again that on short time-scales it is not model differences or scenario uncertainties that dominate what temperature trend might be observed at each location. What will be observed in the coming 15 years is largely determined by internal variability. Counter-intuitively to what one might expect given ever increasing greenhouse gas emissions, figure 4 visually demonstrates that at all locations a cooling trend (or lack of warming trend) could be observed due to the large internal variability on short-term time-scales. We do, however, show that all locations, besides the Southern Ocean and the North Atlantic Ocean are much more likely to warm, than cool, demonstrating the role of increasing greenhouse gases (the forced response). This likelihood increases with increasing greenhouse gas emissions as demonstrated by the differences between the two MPI-GE scenarios and the two CMIP5 scenarios. Again we emphasise the robustness of these results given the near-identical results found in all individual SMILEs and CMIP5.