Historical and idealized climate model experiments: an intercomparison of Earth system models of intermediate complexity

Both historical and idealized climate model experiments are performed with a variety of Earth system models of intermediate complexity (EMICs) as part of a community contribution to the Intergovernmental Panel on Climate Change Fifth Assessment Report. Historical simulations start at 850 CE and continue through to 2005. The standard simulations include changes in forcing from solar luminosity, Earth's orbital configuration, CO2, additional greenhouse gases, land use, and sulphate and volcanic aerosols. In spite of very different modelled pre-industrial global surface air temperatures, overall 20th century trends in surface air temperature and carbon uptake are reasonably well simulated when compared to observed trends. Land carbon fluxes show much more variation between models than ocean carbon fluxes, and recent land fluxes appear to be slightly underestimated. It is possible that recent modelled climate trends or climate–carbon feedbacks are overestimated resulting in too much land carbon loss or that carbon uptake due to CO2 and/or nitrogen fertilization is underestimated. Several one thousand year long, idealized, 2 × and 4 × CO2 experiments are used to quantify standard model characteristics, including transient and equilibrium climate sensitivities, and climate–carbon feedbacks. The values from EMICs generally fall within the range given by general circulation models. Seven additional historical simulations, each including a single specified forcing, are used to assess the contributions of different climate forcings to the overall climate and carbon cycle response. The response of surface air temperature is the linear sum of the individual forcings, while the carbon cycle response shows a non-linear interaction between land-use change and CO2 forcings for some models. Finally, the preindustrial portions of the last millennium simulations are used to assess historical model carbon-climate feedbacks. Given the specified forcing, there is a tendency for the EMICs to underestimate the drop in surface air temperature and CO2 between the Medieval Climate Anomaly and the Little Ice Age estimated from palaeoclimate reconstructions. This in turn could be a result of unforced variability within the climate system, uncertainty in the reconstructions of temperature and CO2, errors in the reconstructions of forcing used to drive the models, or the incomplete representation of certain processes within the models. Given the forcing datasets used in this study, the models calculate significant land-use emissions over the pre-industrial period. This implies that land-use emissions might need to be taken into account, when making estimates of climate–carbon feedbacks from palaeoclimate reconstructions.

The dependence of radiative forcing and feedback on evolving patterns of surface temperature change in climate models

Experiments with CO2 instantaneously quadrupled and then held constant are used to show that the relationship between the global-mean net heat input to the climate system and the global-mean surface-air-temperature change is nonlinear in Coupled Model Intercomparison Project phase 5 (CMIP5) Atmosphere-Ocean General Circulation Models (AOGCMs). The nonlinearity is shown to arise from a change in strength of climate feedbacks driven by an evolving pattern of surface warming. In 23 out of the 27 AOGCMs examined the climate feedback parameter becomes significantly (95% confidence) less negative – i.e. the effective climate sensitivity increases – as time passes. Cloud feedback parameters show the largest changes. In the AOGCM-mean approximately 60% of the change in feedback parameter comes from the topics (30N-30S). An important region involved is the tropical Pacific where the surface warming intensifies in the east after a few decades. The dependence of climate feedbacks on an evolving pattern of surface warming is confirmed using the HadGEM2 and HadCM3 atmosphere GCMs (AGCMs). With monthly evolving sea-surface-temperatures and sea-ice prescribed from its AOGCM counterpart each AGCM reproduces the time-varying feedbacks, but when a fixed pattern of warming is prescribed the radiative response is linear with global temperature change or nearly so. We also demonstrate that the regression and fixed-SST methods for evaluating effective radiative forcing are in principle different, because rapid SST adjustment when CO2 is changed can produce a pattern of surface temperature change with zero global mean but non-zero change in net radiation at the top of the atmosphere (~ -0.5 Wm-2 in HadCM3).

A general-method for tracking analysis and its application to meteorological data

Recent interest in the validation of general circulation models (GCMs) has been devoted to objective methods. A small number of authors have used the direct synoptic identification of phenomena together with a statistical analysis to perform the objective comparison between various datasets. This paper describes a general method for performing the synoptic identification of phenomena that can be used for an objective analysis of atmospheric, or oceanographic, datasets obtained from numerical models and remote sensing. Methods usually associated with image processing have been used to segment the scene and to identify suitable feature points to represent the phenomena of interest. This is performed for each time level. A technique from dynamic scene analysis is then used to link the feature points to form trajectories. The method is fully automatic and should be applicable to a wide range of geophysical fields. An example will be shown of results obtained from this method using data obtained from a run of the Universities Global Atmospheric Modelling Project GCM.

Convectively coupled equatorial waves in high resolution Hadley centre climate models

Current global atmospheric models fail to simulate well organised tropical phenomena in which convection interacts with dynamics and physics. A new methodology to identify convectively coupled equatorial waves, developed by NCAS-Climate, has been applied to output from the two latest models of the Met Office/Hadley Centre which have fundamental differences in dynamical formulation. Variability, horizontal and vertical structures, and propagation characteristics of tropical convection and equatorial waves, along with their coupled behaviour in the models are examined and evaluated against a previous comprehensive study of observations. It is shown that, in general, the models perform well for equatorial waves coupled with off-equatorial convection. However they perform poorly for waves coupled with equatorial convection. The vertical structure of the simulated wave is not conducive to energy conversion/growth and does not support the correct physical-dynamical coupling that occurs in the real world. The following figure shows an example of the Kelvin wave coupled with equatorial convection. It shows that the models fail to simulate a key feature of convectively coupled Kelvin wave in observations, namely near surface anomalous equatorial zonal winds together with intensified equatorial convection and westerly winds in phase with the convection. The models are also not able to capture the observed vertical tilt structure and the vertical propagation of the Kelvin wave into the lower stratosphere as well as the secondary peak in the mid-troposphere, particularly in HadAM3. These results can be used to provide a test-bed for experimentation to improve the coupling of physics and dynamics in climate and weather models.

North Atlantic Oscillation response to transient greenhouse gas forcing and the impact on European winter climate: a CMIP2 multi-model assessment

This study investigates the response of wintertime North Atlantic Oscillation (NAO) to increasing concentrations of atmospheric carbon dioxide (CO2) as simulated by 18 global coupled general circulation models that participated in phase 2 of the Coupled Model Intercomparison Project (CMIP2). NAO has been assessed in control and transient 80-year simulations produced by each model under constant forcing, and 1% per year increasing concentrations of CO2, respectively. Although generally able to simulate the main features of NAO, the majority of models overestimate the observed mean wintertime NAO index of 8 hPa by 5-10 hPa. Furthermore, none of the models, in either the control or perturbed simulations, are able to reproduce decadal trends as strong as that seen in the observed NAO index from 1970-1995. Of the 15 models able to simulate the NAO pressure dipole, 13 predict a positive increase in NAO with increasing CO2 concentrations. The magnitude of the response is generally small and highly model-dependent, which leads to large uncertainty in multi-model estimates such as the median estimate of 0.0061 +/- 0.0036 hPa per %CO2. Although an increase of 0.61 hPa in NAO for a doubling in CO2 represents only a relatively small shift of 0.18 standard deviations in the probability distribution of winter mean NAO, this can cause large relative increases in the probabilities of extreme values of NAO associated with damaging impacts. Despite the large differences in NAO responses, the models robustly predict similar statistically significant changes in winter mean temperature (warmer over most of Europe) and precipitation (an increase over Northern Europe). Although these changes present a pattern similar to that expected due to an increase in the NAO index, linear regression is used to show that the response is much greater than can be attributed to small increases in NAO. NAO trends are not the key contributor to model-predicted climate change in wintertime mean temperature and precipitation over Europe and the Mediterranean region. However, the models' inability to capture the observed decadal variability in NAO might also signify a major deficiency in their ability to simulate the NAO-related responses to climate change.

Comparison between observations and models of the Mozambique Channel transport: Seasonal cycle and eddy frequencies

A time series of the observed transport through an array of moorings across the Mozambique Channel is compared with that of six model runs with ocean general circulation models. In the observations, the seasonal cycle cannot be distinguished from red noise, while this cycle is dominant in the transport of the numerical models. It is found, however, that the seasonal cycles of the observations and numerical models are similar in strength and phase. These cycles have an amplitude of 5 Sv and a maximum in September, and can be explained by the yearly variation of the wind forcing. The seasonal cycle in the models is dominant because the spectral density at other frequencies is underrepresented. Main deviations from the observations are found at depths shallower than 1500 m and in the 5/y–6/y frequency range. Nevertheless, the structure of eddies in the models is close to the observed eddy structure. The discrepancy is found to be related to the formation mechanism and the formation position of the eddies. In the observations, eddies are frequently formed from an overshooting current near the mooring section, as proposed by Ridderinkhof and de Ruijter (2003) and Harlander et al. (2009). This causes an alternation of events at the mooring section, varying between a strong southward current, and the formation and passing of an eddy. This results in a large variation of transport in the frequency range of 5/y–6/y. In the models, the eddies are formed further north and propagate through the section. No alternation similar to the observations is observed, resulting in a more constant transport.

Constraining a North Atlantic Ocean general circulation model with chlorofluorocarbon observations

Measurements of anthropogenic tracers such as chlorofluorocarbons and tritium must be quantitatively combined with ocean general circulation models as a component of systematic model development. The authors have developed and tested an inverse method, using a Green's function, to constrain general circulation models with transient tracer data. Using this method chlorofluorocarbon-11 and -12 (CFC-11 and -12) observations are combined with a North Atlantic configuration of the Miami Isopycnic Coordinate Ocean Model with 4/3 degrees resolution. Systematic differences can be seen between the observed CFC concentrations and prior CFC fields simulated by the model. These differences are reduced by the inversion, which determines the optimal gas transfer across the air-sea interface, accounting for uncertainties in the tracer observations. After including the effects of unresolved variability in the CFC fields, the model is found to be inconsistent with the observations because the model/data misfit slightly exceeds the error estimates. By excluding observations in waters ventilated north of the Greenland-Scotland ridge (sigma (0) < 27.82 kg m(-3); shallower than about 2000 m), the fit is improved, indicating that the Nordic overflows are poorly represented in the model. Some systematic differences in the model/data residuals remain and are related, in part, to excessively deep model ventilation near Rockall and deficient ventilation in the main thermocline of the eastern subtropical gyre. Nevertheless, there do not appear to be gross errors in the basin-scale model circulation. Analysis of the CFC inventory using the constrained model suggests that the North Atlantic Ocean shallower than about 2000 m was near 20% saturated in the mid-1990s. Overall, this basin is a sink to 22% of the total atmosphere-to-ocean CFC-11 flux-twice the global average value. The average water mass formation rates over the CFC transient are 7.0 and 6.0 Sv (Sv = 10(6) m(3) s(-1)) for subtropical mode water and subpolar mode water, respectively.

Statistical downscaling of river flows

An extensive statistical ‘downscaling’ study is done to relate large-scale climate information from a general circulation model (GCM) to local-scale river flows in SW France for 51 gauging stations ranging from nival (snow-dominated) to pluvial (rainfall-dominated) river-systems. This study helps to select the appropriate statistical method at a given spatial and temporal scale to downscale hydrology for future climate change impact assessment of hydrological resources. The four proposed statistical downscaling models use large-scale predictors (derived from climate model outputs or reanalysis data) that characterize precipitation and evaporation processes in the hydrological cycle to estimate summary flow statistics. The four statistical models used are generalized linear (GLM) and additive (GAM) models, aggregated boosted trees (ABT) and multi-layer perceptron neural networks (ANN). These four models were each applied at two different spatial scales, namely at that of a single flow-gauging station (local downscaling) and that of a group of flow-gauging stations having the same hydrological behaviour (regional downscaling). For each statistical model and each spatial resolution, three temporal resolutions were considered, namely the daily mean flows, the summary statistics of fortnightly flows and a daily ‘integrated approach’. The results show that flow sensitivity to atmospheric factors is significantly different between nival and pluvial hydrological systems which are mainly influenced, respectively, by shortwave solar radiations and atmospheric temperature. The non-linear models (i.e. GAM, ABT and ANN) performed better than the linear GLM when simulating fortnightly flow percentiles. The aggregated boosted trees method showed higher and less variable R2 values to downscale the hydrological variability in both nival and pluvial regimes. Based on GCM cnrm-cm3 and scenarios A2 and A1B, future relative changes of fortnightly median flows were projected based on the regional downscaling approach. The results suggest a global decrease of flow in both pluvial and nival regimes, especially in spring, summer and autumn, whatever the considered scenario. The discussion considers the performance of each statistical method for downscaling flow at different spatial and temporal scales as well as the relationship between atmospheric processes and flow variability.

Alternatives to linear analysis of energy balance data from lactating dairy cows

The current energy requirements system used in the United Kingdom for lactating dairy cows utilizes key parameters such as metabolizable energy intake (MEI) at maintenance (MEm), the efficiency of utilization of MEI for 1) maintenance, 2) milk production (k(l)), 3) growth (k(g)), and the efficiency of utilization of body stores for milk production (k(t)). Traditionally, these have been determined using linear regression methods to analyze energy balance data from calorimetry experiments. Many studies have highlighted a number of concerns over current energy feeding systems particularly in relation to these key parameters, and the linear models used for analyzing. Therefore, a database containing 652 dairy cow observations was assembled from calorimetry studies in the United Kingdom. Five functions for analyzing energy balance data were considered: straight line, two diminishing returns functions, (the Mitscherlich and the rectangular hyperbola), and two sigmoidal functions (the logistic and the Gompertz). Meta-analysis of the data was conducted to estimate k(g) and k(t). Values of 0.83 to 0.86 and 0.66 to 0.69 were obtained for k(g) and k(t) using all the functions (with standard errors of 0.028 and 0.027), respectively, which were considerably different from previous reports of 0.60 to 0.75 for k(g) and 0.82 to 0.84 for k(t). Using the estimated values of k(g) and k(t), the data were corrected to allow for body tissue changes. Based on the definition of k(l) as the derivative of the ratio of milk energy derived from MEI to MEI directed towards milk production, MEm and k(l) were determined. Meta-analysis of the pooled data showed that the average k(l) ranged from 0.50 to 0.58 and MEm ranged between 0.34 and 0.64 MJ/kg of BW0.75 per day. Although the constrained Mitscherlich fitted the data as good as the straight line, more observations at high energy intakes (above 2.4 MJ/kg of BW0.75 per day) are required to determine conclusively whether milk energy is related to MEI linearly or not.

A likelihood ratio appropach to family-based association studies with covariates

We introduce a procedure for association based analysis of nuclear families that allows for dichotomous and more general measurements of phenotype and inclusion of covariate information. Standard generalized linear models are used to relate phenotype and its predictors. Our test procedure, based on the likelihood ratio, unifies the estimation of all parameters through the likelihood itself and yields maximum likelihood estimates of the genetic relative risk and interaction parameters. Our method has advantages in modelling the covariate and gene-covariate interaction terms over recently proposed conditional score tests that include covariate information via a two-stage modelling approach. We apply our method in a study of human systemic lupus erythematosus and the C-reactive protein that includes sex as a covariate.

Describing long-term trends in precipitation using generalized additive models

With the current concern over climate change, descriptions of how rainfall patterns are changing over time can be useful. Observations of daily rainfall data over the last few decades provide information on these trends. Generalized linear models are typically used to model patterns in the occurrence and intensity of rainfall. These models describe rainfall patterns for an average year but are more limited when describing long-term trends, particularly when these are potentially non-linear. Generalized additive models (GAMS) provide a framework for modelling non-linear relationships by fitting smooth functions to the data. This paper describes how GAMS can extend the flexibility of models to describe seasonal patterns and long-term trends in the occurrence and intensity of daily rainfall using data from Mauritius from 1962 to 2001. Smoothed estimates from the models provide useful graphical descriptions of changing rainfall patterns over the last 40 years at this location. GAMS are particularly helpful when exploring non-linear relationships in the data. Care is needed to ensure the choice of smooth functions is appropriate for the data and modelling objectives. (c) 2008 Elsevier B.V. All rights reserved.

Neural and wavelet network models for financial distress classification

This work analyzes the use of linear discriminant models, multi-layer perceptron neural networks and wavelet networks for corporate financial distress prediction. Although simple and easy to interpret, linear models require statistical assumptions that may be unrealistic. Neural networks are able to discriminate patterns that are not linearly separable, but the large number of parameters involved in a neural model often causes generalization problems. Wavelet networks are classification models that implement nonlinear discriminant surfaces as the superposition of dilated and translated versions of a single "mother wavelet" function. In this paper, an algorithm is proposed to select dilation and translation parameters that yield a wavelet network classifier with good parsimony characteristics. The models are compared in a case study involving failed and continuing British firms in the period 1997-2000. Problems associated with over-parameterized neural networks are illustrated and the Optimal Brain Damage pruning technique is employed to obtain a parsimonious neural model. The results, supported by a re-sampling study, show that both neural and wavelet networks may be a valid alternative to classical linear discriminant models.

Model selection approaches for non-linear system identification: a review

The identification of non-linear systems using only observed finite datasets has become a mature research area over the last two decades. A class of linear-in-the-parameter models with universal approximation capabilities have been intensively studied and widely used due to the availability of many linear-learning algorithms and their inherent convergence conditions. This article presents a systematic overview of basic research on model selection approaches for linear-in-the-parameter models. One of the fundamental problems in non-linear system identification is to find the minimal model with the best model generalisation performance from observational data only. The important concepts in achieving good model generalisation used in various non-linear system-identification algorithms are first reviewed, including Bayesian parameter regularisation and models selective criteria based on the cross validation and experimental design. A significant advance in machine learning has been the development of the support vector machine as a means for identifying kernel models based on the structural risk minimisation principle. The developments on the convex optimisation-based model construction algorithms including the support vector regression algorithms are outlined. Input selection algorithms and on-line system identification algorithms are also included in this review. Finally, some industrial applications of non-linear models are discussed.

Emulating AOGCM results using simple climate models

Three simple climate models (SCMs) are calibrated using simulations from atmosphere ocean general circulation models (AOGCMs). In addition to using two conventional SCMs, results from a third simpler model developed specifically for this study are obtained. An easy to implement and comprehensive iterative procedure is applied that optimises the SCM emulation of global-mean surface temperature and total ocean heat content, and, if available in the SCM, of surface temperature over land, over the ocean and in both hemispheres, and of the global-mean ocean temperature profile. The method gives best-fit estimates as well as uncertainty intervals for the different SCM parameters. For the calibration, AOGCM simulations with two different types of forcing scenarios are used: pulse forcing simulations performed with 2 AOGCMs and gradually changing forcing simulations from 15 AOGCMs obtained within the framework of the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. The method is found to work well. For all possible combinations of SCMs and AOGCMs the emulation of AOGCM results could be improved. The obtained SCM parameters depend both on the AOGCM data and the type of forcing scenario. SCMs with a poor representation of the atmosphere thermal inertia are better able to emulate AOGCM results from gradually changing forcing than from pulse forcing simulations. Correct simultaneous emulation of both atmospheric temperatures and the ocean temperature profile by the SCMs strongly depends on the representation of the temperature gradient between the atmosphere and the mixed layer. Introducing climate sensitivities that are dependent on the forcing mechanism in the SCMs allows the emulation of AOGCM responses to carbon dioxide and solar insolation forcings equally well. Also, some SCM parameters are found to be very insensitive to the fitting, and the reduction of their uncertainty through the fitting procedure is only marginal, while other parameters change considerably. The very simple SCM is found to reproduce the AOGCM results as well as the other two comparably more sophisticated SCMs.

Identifying and quantifying nonconservative energy production/destruction terms in hydrostatic Boussinesq primitive equation models

This paper seeks to illustrate the point that physical inconsistencies between thermodynamics and dynamics usually introduce nonconservative production/destruction terms in the local total energy balance equation in numerical ocean general circulation models (OGCMs). Such terms potentially give rise to undesirable forces and/or diabatic terms in the momentum and thermodynamic equations, respectively, which could explain some of the observed errors in simulated ocean currents and water masses. In this paper, a theoretical framework is developed to provide a practical method to determine such nonconservative terms, which is illustrated in the context of a relatively simple form of the hydrostatic Boussinesq primitive equation used in early versions of OGCMs, for which at least four main potential sources of energy nonconservation are identified; they arise from: (1) the “hanging” kinetic energy dissipation term; (2) assuming potential or conservative temperature to be a conservative quantity; (3) the interaction of the Boussinesq approximation with the parameterizations of turbulent mixing of temperature and salinity; (4) some adiabatic compressibility effects due to the Boussinesq approximation. In practice, OGCMs also possess spurious numerical energy sources and sinks, but they are not explicitly addressed here. Apart from (1), the identified nonconservative energy sources/sinks are not sign definite, allowing for possible widespread cancellation when integrated globally. Locally, however, these terms may be of the same order of magnitude as actual energy conversion terms thought to occur in the oceans. Although the actual impact of these nonconservative energy terms on the overall accuracy and physical realism of the oceans is difficult to ascertain, an important issue is whether they could impact on transient simulations, and on the transition toward different circulation regimes associated with a significant reorganization of the different energy reservoirs. Some possible solutions for improvement are examined. It is thus found that the term (2) can be substantially reduced by at least one order of magnitude by using conservative temperature instead of potential temperature. Using the anelastic approximation, however, which was initially thought as a possible way to greatly improve the accuracy of the energy budget, would only marginally reduce the term (4) with no impact on the terms (1), (2) and (3).