A verification framework for interannual-to-decadal predictions experiments

Decadal predictions have a high profile in the climate science community and beyond, yet very little is known about their skill. Nor is there any agreed protocol for estimating their skill. This paper proposes a sound and coordinated framework for verification of decadal hindcast experiments. The framework is illustrated for decadal hindcasts tailored to meet the requirements and specifications of CMIP5 (Coupled Model Intercomparison Project phase 5). The chosen metrics address key questions about the information content in initialized decadal hindcasts. These questions are: (1) Do the initial conditions in the hindcasts lead to more accurate predictions of the climate, compared to un-initialized climate change projections? and (2) Is the prediction model’s ensemble spread an appropriate representation of forecast uncertainty on average? The first question is addressed through deterministic metrics that compare the initialized and uninitialized hindcasts. The second question is addressed through a probabilistic metric applied to the initialized hindcasts and comparing different ways to ascribe forecast uncertainty. Verification is advocated at smoothed regional scales that can illuminate broad areas of predictability, as well as at the grid scale, since many users of the decadal prediction experiments who feed the climate data into applications or decision models will use the data at grid scale, or downscale it to even higher resolution. An overall statement on skill of CMIP5 decadal hindcasts is not the aim of this paper. The results presented are only illustrative of the framework, which would enable such studies. However, broad conclusions that are beginning to emerge from the CMIP5 results include (1) Most predictability at the interannual-to-decadal scale, relative to climatological averages, comes from external forcing, particularly for temperature; (2) though moderate, additional skill is added by the initial conditions over what is imparted by external forcing alone; however, the impact of initialization may result in overall worse predictions in some regions than provided by uninitialized climate change projections; (3) limited hindcast records and the dearth of climate-quality observational data impede our ability to quantify expected skill as well as model biases; and (4) as is common to seasonal-to-interannual model predictions, the spread of the ensemble members is not necessarily a good representation of forecast uncertainty. The authors recommend that this framework be adopted to serve as a starting point to compare prediction quality across prediction systems. The framework can provide a baseline against which future improvements can be quantified. The framework also provides guidance on the use of these model predictions, which differ in fundamental ways from the climate change projections that much of the community has become familiar with, including adjustment of mean and conditional biases, and consideration of how to best approach forecast uncertainty.     

Angular and Frequency-Dependent Wave Velocity and Attenuation in Fractured Porous Media

Wave-induced fluid flow generates a dominant attenuation mechanism in porous media. It consists of energy loss due to P-wave conversion to Biot (diffusive) modes at mesoscopic-scale inhomogeneities. Fractured poroelastic media show significant attenuation and velocity dispersion due to this mechanism. The theory has first been developed for the symmetry axis of the equivalent transversely isotropic (TI) medium corresponding to a poroelastic medium containing planar fractures. In this work, we consider the theory for all propagation angles by obtaining the five complex and frequency-dependent stiffnesses of the equivalent TI medium as a function of frequency. We assume that the flow direction is perpendicular to the layering plane and is independent of the loading direction. As a consequence, the behaviour of the medium can be described by a single relaxation function. We first consider the limiting case of an open (highly permeable) fracture of negligible thickness. We then compute the associated wave velocities and quality factors as a function of the propagation direction (phase and ray angles) and frequency. The location of the relaxation peak depends on the distance between fractures (the mesoscopic distance), viscosity, permeability and fractures compliances. The flow induced by wave propagation affects the quasi-shear (qS) wave with levels of attenuation similar to those of the quasi-compressional (qP) wave. On the other hand, a general fracture can be modeled as a sequence of poroelastic layers, where one of the layers is very thin. Modeling fractures of different thickness filled with CO₂ embedded in a background medium saturated with a stiffer fluid also shows considerable attenuation and velocity dispersion. If the fracture and background frames are the same, the equivalent medium is isotropic, but strong wave anisotropy occurs in the case of a frameless and highly permeable fracture material, for instance a suspension of solid particles in the fluid.     

On the feasibility of using a lake water balance model to infer rainfall: an example from Lake Victoria

Abstract

A water balance model of Lake Victoria that can be used to assess rainfall from lake level, is derived. The model utilizes satellite estimates of rainfall directly over the lake. The model, initially derived and calibrated for the period 1956–1978, is reformulated here in such a way that all water balance terms except evaporation can be calculated from a combination of catchment rainfall and level of the lake. The reformulated model is validated and used to predict lake level fluctuations during the period 1931–1994. An error analysis is also performed. The model is then “inverted” to solve for mean rainfall conditions during various intervals of changing lake levels. For modern periods with known rainfall conditions, the error in model estimates is of the order of 1%.     

Programme of the Galactic meridian: Star counts and Galactic models 2. Distribution of absolute proper motions

A comparison of the observed distribution of absolute proper motions with a kinematical model of the Galaxy is presented. Proper motions with respect to galaxies were obtained for about 40 000 stars along the main Galactic meridian and in two fields near the North Galactic pole (programme MEGA). The Galaxy is considered as composed of the disk (main sequence and disk red giants), the thick disk and spheroid populations. For each subsystem, spatial velocity components and their dispersions were computed. The distribution of kinematical parameters were modelled for stars located in different directions of the Galaxy.     

Statistical analysis of the radon-222 potential of rocks in Virginia,U.S.A.

More than 3,200 indoor radon-222 (²²²Rn) measurements were made seasonally in an area of about 1,000 square kilometers of the Coastal Plain and Piedmont physiographic provinces in Virginia, U.S.A. Results of these measurements indicate that some geological units are associated, on the average, with twice as much indoor²²²Rn as other geological units, and that indoor²²²Rn varies seasonally. The Kruskal-Wallis test was used to test whether indoor²²²Rn concentrations for data gathered over the winter and summer seasons differ significantly by rock unit. The tests concluded that indoor²²²Rn concentrations for different rock units were not equal at the 5-percent significance level. The rocks associated with the highest median indoor²²²Rn concentration are specific rocks in the Mesozoic Culpeper basin, including shale and siltstone units with Jurassic diabase intrusives, and mica schists in the Piedmont physiographic province. The pre-Triassic Peters Creek Schist has the highest ranking in terms of indoor²²²Rn concentration. The rocks associated with the lowest indoor²²²Rn concentrations include coastal plain sediments, the Occoquan Granite, Falls Church Tonalite, Piney Branch Mafic and Ultramafic complex, and unnamed mafic and ultramafic inclusions, respectively. The rocks have been ranked according to observed²²²Rn concentration by transforming the average rank of indoor²²²Rn concentrations to z scores.     

Hydrothermally altered hyaloclastites at the Earth’s surface in the rift zone of Iceland: Problem of the biochemogenic accumulation of trace elements

Results of the mineral and chemical study of surface-altered hyaloclastites from the Theistareykir and Nämafjall geothermal zones, the North Iceland rift, are considered. The paper scrutinizes the composition of hydrothermally altered rocks mainly consisting of clay minerals related to the transformation of bedrock hyaloclastites. The clay minerals also make up slope talus or proluvial deltas around highlands. The major clay minerals in areas of the present-day fumarole activity are kaolinite and smectite, while the redeposited talus-proluvial sediments are mainly composed of smectites. The studied rocks contain framboids, globules, rods, and other structures with morphology and size similar to those of mineralized bacteria. The local concentration of Ag in proluvial deposits is attributed to the intense bacterial activity. It is supposed that elevated contents of some trace elements could be related to an intense microbiotic influence.     

Middle Paleozoic paleomagnetism of east Kazakhstan: post-Middle Devonian rotations in a large-scale orocline in the central Ural–Mongol belt

In order to test different hypotheses concerning the Paleozoic evolution of the Ural–Mongol belt (UMB) and the amalgamation of Eurasia, we studied Middle Devonian basalts from two localities (11 sites) and Lower Silurian volcanics, redbeds, and intra-formational conglomerates from three localities (20 sites) in the Chingiz Range of East Kazakhstan. The Devonian rocks prove to be heavily overprinted in the late Paleozoic, and a high-temperature, presumably primary, southerly, and down component is isolated at only four sites from a homoclinal section. Most Silurian redbeds are found to be remagnetized in the late Paleozoic; in contrast, a bipolar near-horizontal remanence, isolated from Silurian volcanics, is most probably primary as indicated by positive tilt and conglomerate tests. Analysis of paleomagnetic data from the Chingiz Range shows that southward-pointing directions in Ordovician, Silurian, and Devonian rocks are of normal polarity and hence indicate large-scale rotations after the Middle Devonian. The Chingiz paleomagnetic directions can be compared with Paleozoic data from the North Tien Shan and with the horseshoe-shaped distribution of subduction-related volcanic complexes in Kazakhstan. Both paleomagnetic and geological data support the idea that today's strongly curved volcanic belts of Kazakhstan are an orocline, deformed mostly before mid-Permian time. Despite the determination of nearly a dozen new Paleozoic paleopoles in this study and other recent publications by our team, significant temporal and spatial gaps remain in our knowledge of the paleomagnetic directions during the middle and late Paleozoic. However, the paleomagnetic results from the Chingiz Range and the North Tien Shan indicate that these areas show generally coherent motions with Siberia and Baltica, respectively.     

The postglacial history of three Picea species in New England, USA

Given the difficulty of separating the three Picea species—P. glauca, P. mariana, and P. rubens (white, black, and red spruce)—in the pollen record, little is known about their unique histories in eastern North America following deglaciation. Here we report the first use of a classification tree analysis (CART) to distinguish pollen grains of these species. It was successfully applied to fossil pollen from eight sites in Maine and one in Massachusetts. We focused on the late glacial/early Holocene (14,000 to 8000 cal yr B.P.) and the late Holocene (1400 cal yr B.P. to present)—the two key periods since deglaciation when Picea has been abundant in the region. The result shows a shift from a Picea forest of P. glauca and P. mariana in the late glacial to a forest of P. rubens and P. mariana in the late Holocene. The small number of P. rubens grains identified from the late glacial/early Holocene samples (<5%) suggests that that species was either absent or rare at most of the sites. The occurrence and distribution of the three species do not reveal any geographic or temporal trend during late glacial time, but the data suggest that they were distributed in local patches on the landscape. The results of this study indicate that the recent population expansion of Picea (1000 to 500 cal yr B.P.) was likely the first time since deglaciation that P. rubens was abundant in the region.     

P–Y Curves for Piles Under Seismic Lateral Loads

This paper describes an experimentally based procedure for building dynamic P–Y curves for clays from the Campeche sound in the Gulf of Mexico and for Mexico City clays. Cyclic triaxial and resonant column tests were used to derive hyperbolic stiffness–strain and damping–strain functions that depend on soil plasticity and on relative consistency. The dynamic P–Y curves incorporate these functions and also include hysteretical and geometrical damping characteristics of the pile–soil system; they can be used directly in dynamic interaction analyses of structure–pile–soil systems in the time or the frequency domains. Usual practice relies on pseudo static formulations and a simplified alternative to include P–Y curves in this kind of analysis is also proposed, which is essentially the same as the dynamic ones but does not include radiation or hysteretical damping.     

Suction-controlled experiments on Boom clay

The understanding of the thermo-hydro-mechanical behaviour of a clay barrier is needed for the prediction of its final in situ properties after the hydration and thermal transient in a radioactive waste repository.

As part of the CEC 1990–1994 R&D programme on radioactive waste management and storage, the CEA (Fr), CIEMAT (Sp), ENRESA (Sp), SCK · CEN (B), UPC (Sp) and UWCC (UK) have carried out a joint project on unsaturated clay behaviour (Volckaert et al., 1996). The aim of the study is to analyse and model the behaviour of a clay-based engineered barrier during its hydration phase under real repository conditions. The hydro-mechanical and thermo-hydraulic models developed in this project have been coupled to describe stress/strain behaviour, moisture migration and heat transfer. A thermo-hydraulic model has also been coupled to a geochemical code to describe the migration and formation of chemical species.

In this project, suction-controlled experiments have been performed on Boom clay (B), FoCa clay (Fr) and Almeria bentonite (Sp). The aim of these experiments is to test the validity of the interpretive model developed by Alonso and Gens (Alonso et al., 1990), and to build a database of unsaturated clay thermo-hydro-mechanical parameters. Such a database can then be used for validation exercises in which in situ experiments are simulated.

The Boom clay is a moderately swelling clay of Rupellian age. It is studied at the SCK · CEN in Belgium as a potential host rock for a radioactive waste repository. In this paper, suction-controlled experiments carried out on Boom clay by SCK · CEN are described. SCK · CEN has performed experiments to measure the relation between suction, water content and temperature and the relation between suction, stress and deformation. The applied suction-control techniques and experimental setups are detailed. The results of these experiments are discussed in the perspective of the model of Alonso and Gens. The influence of temperature on water uptake was rather small. The measured swelling-collapse behaviour can be explained by the Alonso and Gens model.