Rosie Eade

Skillful long‐range prediction of European and North American winters

This work was supported by the Joint DECC/Defra Met Office Hadley Centre Climate Programme (GA01101), the UK Public Weather Service research program, and the European Union Framework 7 SPECS project. Leon Hermanson was funded as part of his Research Fellowship by Willis as part of Willis Research Network (WRN).

Do seasonal‐to‐decadal climate predictions underestimate the predictability of the real world?

Seasonal-to-decadal predictions are inevitably uncertain, depending on the size of the predictable signal relative to unpredictable chaos. Uncertainties can be accounted for using ensemble techniques, permitting quantitative probabilistic forecasts. In a perfect system, each ensemble member would represent a potential realization of the true evolution of the climate system, and the predictable components in models and reality would be equal. However, we show that the predictable component is sometimes lower in models than observations, especially for seasonal forecasts of the North Atlantic Oscillation and multiyear forecasts of North Atlantic temperature and pressure. In these cases the forecasts are underconfident, with each ensemble member containing too much noise. Consequently, most deterministic and probabilistic measures underestimate potential skill and idealized model experiments underestimate predictability. However, skilful and reliable predictions may be achieved using a large ensemble to reduce noise and adjusting the forecast variance through a postprocessing technique proposed here.

A comparison of full-field and anomaly initialization for seasonal to decadal climate prediction

There are two main approaches for dealing with model biases in forecasts made with initialized climate models. In full-field initialization, model biases are removed during the assimilation process by constraining the model to be close to observations. Forecasts drift back towards the model’s preferred state, thereby re-establishing biases which are then removed with an a posterior lead-time dependent correction diagnosed from a set of historical tests (hindcasts). In anomaly initialization, the model is constrained by observed anomalies and deviates from its preferred climatology only by the observed variability. In theory, the forecasts do not drift, and biases may be removed based on the difference between observations and independent model simulations of a given period. Both approaches are currently in use, but their relative merits are unclear. Here we compare the skill of each approach in comprehensive decadal hindcasts starting each year from 1960 to 2009, made using the Met Office decadal prediction system. Both approaches are more skilful than climatology in most regions for temperature and some regions for precipitation. On seasonal timescales, full-field initialized hindcasts of regional temperature and precipitation are significantly more skilful on average than anomaly initialized hindcasts. Teleconnections associated with the El Niño Southern Oscillation are stronger with the full-field approach, providing a physical basis for the improved precipitation skill. Differences in skill on multi-year timescales are generally not significant. However, anomaly initialization provides a better estimate of forecast skill from a limited hindcast set.

Earth's energy imbalance since 1960 in observations and CMIP5 models

Observational analyses of running 5 year ocean heat content trends (Ht) and net downward top of atmosphere radiation (N) are significantly correlated (r ∼ 0.6) from 1960 to 1999, but a spike in Ht in the early 2000s is likely spurious since it is inconsistent with estimates of N from both satellite observations and climate model simulations. Variations in N between 1960 and 2000 were dominated by volcanic eruptions and are well simulated by the ensemble mean of coupled models from the Fifth Coupled Model Intercomparison Project (CMIP5). We find an observation-based reduction in N of − 0.31 ± 0.21 W m−2 between 1999 and 2005 that potentially contributed to the recent warming slowdown, but the relative roles of external forcing and internal variability remain unclear. While present-day anomalies of N in the CMIP5 ensemble mean and observations agree, this may be due to a cancelation of errors in outgoing longwave and absorbed solar radiation. Key Points Observed maximum in ocean heat content trend in early 2000s is likely spurious Net incoming radiation (N) reduced by 0.31 ± 0.21 W m−2 during the warming pause Present-day estimates of N may contain opposing errors in radiative components

Forecast cooling of the Atlantic subpolar gyre and associated impacts

Decadal variability in the North Atlantic and its subpolar gyre (SPG) has been shown to be predictable in climate models initialized with the concurrent ocean state. Numerous impacts over ocean and land have also been identified. Here we use three versions of the Met Office Decadal Prediction System to provide a multimodel ensemble forecast of the SPG and related impacts. The recent cooling trend in the SPG is predicted to continue in the next 5 years due to a decrease in the SPG heat convergence related to a slowdown of the Atlantic Meridional Overturning Circulation. We present evidence that the ensemble forecast is able to skilfully predict these quantities over recent decades. We also investigate the ability of the forecast to predict impacts on surface temperature, pressure, precipitation, and Atlantic tropical storms and compare the forecast to recent boreal summer climate.

Examining reliability of seasonal to decadal sea surface temperature forecasts: The role of ensemble dispersion

Useful probabilistic climate forecasts on decadal timescales should be reliable (i.e. forecast probabilities match the observed relative frequencies) but this is seldom examined. This paper assesses a necessary condition for reliability, that the ratio of ensemble spread to forecast error being close to one, for seasonal to decadal sea surface temperature retrospective forecasts from the Met Office Decadal Prediction System (DePreSys). Factors which may affect reliability are diagnosed by comparing this spread-error ratio for an initial condition ensemble and two perturbed physics ensembles for initialized and uninitialized predictions. At lead times less than 2 years, the initialized ensembles tend to be under-dispersed, and hence produce overconfident and hence unreliable forecasts. For longer lead times, all three ensembles are predominantly over-dispersed. Such over-dispersion is primarily related to excessive inter-annual variability in the climate model. These findings highlight the need to carefully evaluate simulated variability in seasonal and decadal prediction systems.Useful probabilistic climate forecasts on decadal timescales should be reliable (i.e. forecast probabilities match the observed relative frequencies) but this is seldom examined. This paper assesses a necessary condition for reliability, that the ratio of ensemble spread to forecast error being close to one, for seasonal to decadal sea surface temperature retrospective forecasts from the Met Office Decadal Prediction System (DePreSys). Factors which may affect reliability are diagnosed by comparing this spread-error ratio for an initial condition ensemble and two perturbed physics ensembles for initialized and uninitialized predictions. At lead times less than 2 years, the initialized ensembles tend to be under-dispersed, and hence produce overconfident and hence unreliable forecasts. For longer lead times, all three ensembles are predominantly over-dispersed. Such over-dispersion is primarily related to excessive inter-annual variability in the climate model. These findings highlight the need to carefully evaluate simulated variability in seasonal and decadal prediction systems.

A Bayesian Framework for Verification and Recalibration of Ensemble Forecasts: How Uncertain is NAO Predictability?

Predictability estimates of ensemble prediction systems are uncertain due to limited numbers of past forecasts and observations. To account for such uncertainty, this paper proposes a Bayesian inferential framework that provides a simple 6-parameter representation of ensemble forecasting systems and the corresponding observations. The framework is probabilistic, and thus allows for quantifying uncertainty in predictability measures such as correlation skill and signal-to-noise ratios. It also provides a natural way to produce recalibrated probabilistic predictions from uncalibrated ensembles forecasts. The framework is used to address important questions concerning the skill of winter hindcasts of the North Atlantic Oscillation for 1992-2011 issued by the Met Office GloSea5 climate prediction system. Although there is much uncertainty in the correlation between ensemble mean and observations, there is strong evidence of skill: the 95% credible interval of the correlation coefficient of [0.19,0.68] does not overlap zero. There is also strong evidence that the forecasts are not exchangeable with the observations: With over 99% certainty, the signal-to-noise ratio of the forecasts is smaller than the signal-to-noise ratio of the observations, which suggests that raw forecasts should not be taken as representative scenarios of the observations. Forecast recalibration is thus required, which can be coherently addressed within the proposed framework.

Predictions of Climate Several Years Ahead Using an Improved Decadal Prediction System

AbstractDecadal climate predictions are now established as a source of information on future climate alongside longer-term climate projections. This information has the potential to provide key evidence for decisions on climate change adaptation, especially at regional scales. Its importance implies that following the creation of an initial generation of decadal prediction systems, a process of continual development is needed to produce successive versions with better predictive skill. Here, a new version of the Met Office Hadley Centre Decadal Prediction System (DePreSys 2) is introduced, which builds upon the success of the original DePreSys. DePreSys 2 benefits from inclusion of a newer and more realistic climate model, the Hadley Centre Global Environmental Model version 3 (HadGEM3), but shares a very similar approach to initialization with its predecessor. By performing a large suite of reforecasts, it is shown that DePreSys 2 offers improved skill in predicting climate several years ahead. Differenc...

Skilful prediction of Sahel summer rainfall on inter-annual and multi-year timescales

Summer rainfall in the Sahel region of Africa exhibits one of the largest signals of climatic variability and with a population reliant on agricultural productivity, the Sahel is particularly vulnerable to major droughts such as occurred in the 1970s and 1980s. Rainfall levels have subsequently recovered, but future projections remain uncertain. Here we show that Sahel rainfall is skilfully predicted on inter-annual and multi-year (that is, >5 years) timescales and use these predictions to better understand the driving mechanisms. Moisture budget analysis indicates that on multi-year timescales, a warmer north Atlantic and Mediterranean enhance Sahel rainfall through increased meridional convergence of low-level, externally sourced moisture. In contrast, year-to-year rainfall levels are largely determined by the recycling rate of local moisture, regulated by planetary circulation patterns associated with the El Niño-Southern Oscillation. Our findings aid improved understanding and forecasting of Sahel drought, paramount for successful adaptation strategies in a changing climate.

Comments on “Multiyear Predictions of North Atlantic Hurricane Frequency: Promise and Limitations”

Vecchi et al. (2013, hereafter V13) show that retrospective decadal predictions (reforecasts) of multiyear North Atlantic hurricane frequency have high correlations with observations, in agreement with an earlier study (Smith et al. 2010, hereafter S10). However, V13 state that ‘‘the skill in the initialized forecasts comes in large part from the persistence of a mid-1990s shift by the initialized forecasts, rather than from predicting its evolution.’’ Here, we provide a different interpretation of the Met Office Decadal Prediction System (DePreSys) reforecasts, showing that these would have provided clear evidence for an impending reversal to a period of above average hurricane frequency had they been available in 1994, before the observed increase occurred. This is illustrated in Fig. 1a, which shows the information that would have been available in 1994. DePreSys reforecasts starting from 1991 onward clearly predict an increase in hurricane numbers, in fact to levels higher than ever simulated before by this modeling system, while observed counts remained low (including each individual year from 1991 to 1994; not shown). The conclusion in V13 that DePreSys did not predict the 1995 shift is partly based on their analysis of the difference in storm counts averaged over years 2–6 minus the first year of each forecast (Fig. 7 in V13; cf. Fig. 1c). V13 argue that observations straddling the 1995 shift (green triangles in Fig. 1c) are unusually large in this statistic, whereas the DePreSys forecasts are not (blue histogram in Fig. 1c). However, in DePreSys, this statistic is particularly sensitive to the forecast initialized in 1990, which erroneously predicted a very active hurricane season for 1991. Furthermore, this forecast was unaware of the eruption of Mount Pinatubo in June 1991, which likely suppressed hurricane numbers in that year (Evan 2012) and of course was unpredictable. If we exclude 1990 and consider the forecasts starting after Pinatubo, between 1991 and 1993, then DePreSys (red histogram in Fig. 1c) hindcasts predicted an increase similar to that observed. How much confidence could we have had in the DePreSys forecasts of a shift in hurricane frequency? Assessment of previous reforecasts is inconclusive: the decline from the mid-1960s was successfully captured, but the maximum in the late 1970s was not predicted (Fig. 1a) and forecasts beginning in the late 1970s incorrectly predicted an increase (although these were unaware of the impending eruption of El Chich on, which likely decreased hurricane numbers; Evan 2012). We therefore examine the physical mechanisms driving the increased hurricane numbers predicted during Corresponding author address: Doug Smith, Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, United Kingdom. E-mail: doug.smith@metoffice.gov.uk 1 JANUARY 2014 CORRES PONDENCE 487