Search

Information in decadal climate prediction arises from a well-initialized ocean state and from the predicted response to an external forcing. The length of time over which the initial conditions benefit the decadal forecast depends on the start date of the forecast. We characterize this state-dependent predictability for decadal forecasts of upper ocean heat content in the Community Earth System Model. We find regionally dependent initial condition predictability, with extended predictability generally observed in the extratropics. We also detect state-dependent predictability, with the year of loss of information from the initialization varying between start dates. The decadal forecasts in the North Atlantic show substantial information from the initial conditions beyond the 10-yr forecast window, and a high degree of state-dependent predictability. We find some evidence for state-dependent predictability in the ensemble spread in this region, similar to that seen in weather and subseasonal-to-seasonal forecasts. For some start dates, an increase of information with lead time is observed, for which the initialized forecasts predict a growing phase of the Atlantic multidecadal oscillation. Finally we consider the information in the forecast from the initial conditions relative to the forced response, and quantify the crossover time scale after which the forcing provides more information. We demonstrate that the climate change signal projects onto different patterns than the signal from the initial conditions. This means that even after the crossover time scale has been reached in a basin-averaged sense, the benefits of initialization can be felt locally on longer time scales. [ABSTRACT FROM AUTHOR]

The time scales of the Paris Climate Agreement indicate urgent action is required on climate policies over the next few decades, in order to avoid the worst risks posed by climate change. On these relatively short time scales the combined effect of climate variability and change are both key drivers of extreme events, with decadal time scales also important for infrastructure planning. Hence, in order to assess climate risk on such time scales, we require climate models to be able to represent key aspects of both internally driven climate variability and the response to changing forcings. In this paper we argue that we now have the modeling capability to address these requirements—specifically with global models having horizontal resolutions considerably enhanced from those typically used in previous Intergovernmental Panel on Climate Change (IPCC) and Coupled Model Intercomparison Project (CMIP) exercises. The improved representation of weather and climate processes in such models underpins our enhanced confidence in predictions and projections, as well as providing improved forcing to regional models, which are better able to represent local-scale extremes (such as convective precipitation). We choose the global water cycle as an illustrative example because it is governed by a chain of processes for which there is growing evidence of the benefits of higher resolution. At the same time it comprises key processes involved in many of the expected future climate extremes (e.g., flooding, drought, tropical and midlatitude storms). [ABSTRACT FROM AUTHOR]

Forecast verification is important across scientific disciplines, as it provides a framework for evaluating the performance of a forecasting system. In the atmospheric sciences, probabilistic skill scores are often used for verification, as they provide a way of ranking the performance of different probabilistic forecasts unambiguously. In order to be useful, a skill score must be proper: it must encourage honesty in the forecaster and reward forecasts that are reliable and have good resolution. A new score, the error-spread score (ES), is proposed, which is particularly suitable for evaluation of ensemble forecasts. It is formulated with respect to the moments of the forecast. The ES is confirmed to be a proper score and is therefore sensitive to both resolution and reliability. The ES is tested on forecasts made using the Lorenz ′96 system and found to be useful for summarizing the skill of the forecasts. The European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble prediction system (EPS) is evaluated using the ES. Its performance is compared with a perfect statistical probabilistic forecast: the ECMWF high-resolution deterministic forecast dressed with the observed error distribution. This generates a forecast that is perfectly reliable if considered over all time, but does not vary from day to day with the predictability of the atmospheric flow. The ES distinguishes between the dynamically reliable EPS forecasts and the statically reliable dressed deterministic forecasts. Other skill scores are tested and found to be comparatively insensitive to this desirable forecast quality. The ES is used to evaluate seasonal range ensemble forecasts made with the ECMWF System 4. The ensemble forecasts are found to be skilful when compared with climatological or persistence forecasts, though this skill is dependent on the region and time of year. [ABSTRACT FROM AUTHOR]

Important questions concerning parameterization of tropical convection are how should subgrid-scale variability be represented and which large-scale variables should be used in the parameterizations? Here the statistics of observational data in Darwin, Australia, are compared with those of short-term forecasts of convection made by the European Centre for Medium-Range Weather Forecasts Integrated Forecast System. The forecasts use multiplicative-noise stochastic physics (MNSP) that has led to many improvements in weather forecast skill. However, doubts have recently been raised about whether MNSP is consistent with observations of tropical convection. It is shown that the model can reproduce the variability of convection intensity for a given large-scale state, both with and without MNSP. Therefore MNSP is not inconsistent with observations, and much of the modeled variability arises from nonlinearity of the deterministic part of the convection scheme. It is also shown that the model can reproduce the lack of correlation between convection intensity and large-scale CAPE and an entraining CAPE, even though the convection parameterization assumes that deep convection is more intense when the vertical temperature profile is more unstable, with entrainment taken into account. Relationships between convection and large-scale convective inhibition and vertical velocity are also correctly captured. [ABSTRACT FROM AUTHOR]