Hyperspectral Mineral Identification using SVM and SOM

Remote sensing techniques involving hyperspectral imagery have applications in a number of sciences that study some aspects of the surface of the planet. The analysis of hyperspectral images is complex because of the large amount of information involved and the noise within that data. Investigating images with regard to identify minerals, rocks, vegetation and other materials is an application of hyperspectral remote sensing in the earth sciences. This thesis evaluates the performance of two classification and clustering techniques on hyperspectral images for mineral identification. Support Vector Machines (SVM) and Self-Organizing Maps (SOM) are applied as classification and clustering techniques, respectively. Principal Component Analysis (PCA) is used to prepare the data to be analyzed. The purpose of using PCA is to reduce the amount of data that needs to be processed by identifying the most important components within the data. A well-studied dataset from Cuprite, Nevada and a dataset of more complex data from Baffin Island were used to assess the performance of these techniques. The main goal of this research study is to evaluate the advantage of training a classifier based on a small amount of data compared to an unsupervised method. Determining the effect of feature extraction on the accuracy of the clustering and classification method is another goal of this research. This thesis concludes that using PCA increases the learning accuracy, and especially so in classification. SVM classifies Cuprite data with a high precision and the SOM challenges SVM on datasets with high level of noise (like Baffin Island).

Mineralogical characterization of a highly-weathered soil by the Rietveld Method

The mineralogical characterization through mineral quantification of Brazilian soils by X-ray diffraction data using the Rietveld Method is not common. A mineralogical quantification of an Acric Ferralsol from the Ponta Grossa region, state of Paraná, Brazil, was carried out using this Method with X-Ray Diffraction data to verify if this method was suitable for mineral quantification of a highly-weathered soil. The A, AB and B3 horizons were fractioned to separate the different particle sizes: clay, silt, fine sand (by Stokes Law) and coarse sand fractions (by sieving), with the procedure free of chemical treatments. X-ray Fluorescence, Inductively Coupled Plasma Atomic Emission Spectrometry, Infrared Spectroscopy and Mössbauer Spectroscopy were used in order to assist the mineral identification and quantification. The Rietveld Method enabled the quantification of the present minerals. In a general way, the quantitative mineralogical characterization by the Rietveld Method revealed that quartz, gibbsite, rutile, hematite, goethite, kaolinite and halloysite were present in the clay and silt fractions of all horizons. The silt fractions of the deeper horizons were different from the more superficial ones due to the presence of large amounts of quartz. The fine and the coarse sand fractions are constituted mainly by quartz. Therefore, a mineralogical quantification of the finer fraction (clay and silt) by the Rietveld Method was successful.

Unmixing hyperspectral data: independent and dependent component analysis

The development of high spatial resolution airborne and spaceborne sensors has improved the capability of ground-based data collection in the fields of agriculture, geography, geology, mineral identification, detection [2, 3], and classification [4–8]. The signal read by the sensor from a given spatial element of resolution and at a given spectral band is a mixing of components originated by the constituent substances, termed endmembers, located at that element of resolution. This chapter addresses hyperspectral unmixing, which is the decomposition of the pixel spectra into a collection of constituent spectra, or spectral signatures, and their corresponding fractional abundances indicating the proportion of each endmember present in the pixel [9, 10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. The linear mixing model holds when the mixing scale is macroscopic [13]. The nonlinear model holds when the mixing scale is microscopic (i.e., intimate mixtures) [14, 15]. The linear model assumes negligible interaction among distinct endmembers [16, 17]. The nonlinear model assumes that incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [18]. Under the linear mixing model and assuming that the number of endmembers and their spectral signatures are known, hyperspectral unmixing is a linear problem, which can be addressed, for example, under the maximum likelihood setup [19], the constrained least-squares approach [20], the spectral signature matching [21], the spectral angle mapper [22], and the subspace projection methods [20, 23, 24]. Orthogonal subspace projection [23] reduces the data dimensionality, suppresses undesired spectral signatures, and detects the presence of a spectral signature of interest. The basic concept is to project each pixel onto a subspace that is orthogonal to the undesired signatures. As shown in Settle [19], the orthogonal subspace projection technique is equivalent to the maximum likelihood estimator. This projection technique was extended by three unconstrained least-squares approaches [24] (signature space orthogonal projection, oblique subspace projection, target signature space orthogonal projection). Other works using maximum a posteriori probability (MAP) framework [25] and projection pursuit [26, 27] have also been applied to hyperspectral data. In most cases the number of endmembers and their signatures are not known. Independent component analysis (ICA) is an unsupervised source separation process that has been applied with success to blind source separation, to feature extraction, and to unsupervised recognition [28, 29]. ICA consists in finding a linear decomposition of observed data yielding statistically independent components. Given that hyperspectral data are, in given circumstances, linear mixtures, ICA comes to mind as a possible tool to unmix this class of data. In fact, the application of ICA to hyperspectral data has been proposed in reference 30, where endmember signatures are treated as sources and the mixing matrix is composed by the abundance fractions, and in references 9, 25, and 31–38, where sources are the abundance fractions of each endmember. In the first approach, we face two problems: (1) The number of samples are limited to the number of channels and (2) the process of pixel selection, playing the role of mixed sources, is not straightforward. In the second approach, ICA is based on the assumption of mutually independent sources, which is not the case of hyperspectral data, since the sum of the abundance fractions is constant, implying dependence among abundances. This dependence compromises ICA applicability to hyperspectral images. In addition, hyperspectral data are immersed in noise, which degrades the ICA performance. IFA [39] was introduced as a method for recovering independent hidden sources from their observed noisy mixtures. IFA implements two steps. First, source densities and noise covariance are estimated from the observed data by maximum likelihood. Second, sources are reconstructed by an optimal nonlinear estimator. Although IFA is a well-suited technique to unmix independent sources under noisy observations, the dependence among abundance fractions in hyperspectral imagery compromises, as in the ICA case, the IFA performance. Considering the linear mixing model, hyperspectral observations are in a simplex whose vertices correspond to the endmembers. Several approaches [40–43] have exploited this geometric feature of hyperspectral mixtures [42]. Minimum volume transform (MVT) algorithm [43] determines the simplex of minimum volume containing the data. The MVT-type approaches are complex from the computational point of view. Usually, these algorithms first find the convex hull defined by the observed data and then fit a minimum volume simplex to it. Aiming at a lower computational complexity, some algorithms such as the vertex component analysis (VCA) [44], the pixel purity index (PPI) [42], and the N-FINDR [45] still find the minimum volume simplex containing the data cloud, but they assume the presence in the data of at least one pure pixel of each endmember. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. Hyperspectral sensors collects spatial images over many narrow contiguous bands, yielding large amounts of data. For this reason, very often, the processing of hyperspectral data, included unmixing, is preceded by a dimensionality reduction step to reduce computational complexity and to improve the signal-to-noise ratio (SNR). Principal component analysis (PCA) [46], maximum noise fraction (MNF) [47], and singular value decomposition (SVD) [48] are three well-known projection techniques widely used in remote sensing in general and in unmixing in particular. The newly introduced method [49] exploits the structure of hyperspectral mixtures, namely the fact that spectral vectors are nonnegative. The computational complexity associated with these techniques is an obstacle to real-time implementations. To overcome this problem, band selection [50] and non-statistical [51] algorithms have been introduced. This chapter addresses hyperspectral data source dependence and its impact on ICA and IFA performances. The study consider simulated and real data and is based on mutual information minimization. Hyperspectral observations are described by a generative model. This model takes into account the degradation mechanisms normally found in hyperspectral applications—namely, signature variability [52–54], abundance constraints, topography modulation, and system noise. The computation of mutual information is based on fitting mixtures of Gaussians (MOG) to data. The MOG parameters (number of components, means, covariances, and weights) are inferred using the minimum description length (MDL) based algorithm [55]. We study the behavior of the mutual information as a function of the unmixing matrix. The conclusion is that the unmixing matrix minimizing the mutual information might be very far from the true one. Nevertheless, some abundance fractions might be well separated, mainly in the presence of strong signature variability, a large number of endmembers, and high SNR. We end this chapter by sketching a new methodology to blindly unmix hyperspectral data, where abundance fractions are modeled as a mixture of Dirichlet sources. This model enforces positivity and constant sum sources (full additivity) constraints. The mixing matrix is inferred by an expectation-maximization (EM)-type algorithm. This approach is in the vein of references 39 and 56, replacing independent sources represented by MOG with mixture of Dirichlet sources. Compared with the geometric-based approaches, the advantage of this model is that there is no need to have pure pixels in the observations. The chapter is organized as follows. Section 6.2 presents a spectral radiance model and formulates the spectral unmixing as a linear problem accounting for abundance constraints, signature variability, topography modulation, and system noise. Section 6.3 presents a brief resume of ICA and IFA algorithms. Section 6.4 illustrates the performance of IFA and of some well-known ICA algorithms with experimental data. Section 6.5 studies the ICA and IFA limitations in unmixing hyperspectral data. Section 6.6 presents results of ICA based on real data. Section 6.7 describes the new blind unmixing scheme and some illustrative examples. Section 6.8 concludes with some remarks.

A turquesa de Itacupim, Pará

Na ilha de Itacupim, localizada na região costeira do nordeste do Pará, foram encontrados veios de fosfatos de alumínio contendo turquesa, além de quartzo e argilominerais. A ilha é sustentada por espesso perfil laterítico maturo desenvolvido sobre complexo alcalino-ultramáfico mineralizado em apatita. Os veios e vênulas são de espessura centimétrica, normalmente constituídos de wavellita fibro-radial, onde pode ser observada turquesa verde-azulada, em massas subesferolíticas, microcristalinas, intercrescidas com caulinita e oxi-hidróxidos de Mn, além de quartzo. A identificação mineral foi realizada por DRX, microscopia óptica, análises químicas de rocha total, MEV/SED. Os teores de CuO são inferiores aos das turquesas em geral, compensados por Fe₂O₃ e ZnO. Os subesferolitos de turquesa contêm inúmeras inclusões micrométricas de goyazita ou svanbergita. A ocorrência da turquesa, na forma de veios e vênulas, seu aspecto porcelanado e a conhecida relação desse mineral com ambiente hidrotermal sugerem que a turquesa de Itacupim também seja de origem hidrotermal, reforçada pela sua associação com wavellita, goyazita ou svanbergita, quartzo e argilominerais. Ela não foi encontrada no perfil laterítico. Seu aspecto compacto e sua cor esverdeada abrem perspectivas para seu uso como mineral de gema.

Sonic log prediction in carbonates

This work is conducted to study the complications associated with the sonic log prediction in carbonate logs and to investigate the possible solutions to accurately predict the sonic logs in Traverse Limestone. Well logs from fifty different wells were analyzed to define the mineralogy of the Traverse Limestone by using conventional 4-mineral and 3-mineral identification approaches. We modified the conventional 3-mineral identification approach (that completely neglects the gamma ray response) to correct the shale effects on the basis of gamma ray log before employing the 3-mineral identification. This modification helped to get the meaningful insight of the data when a plot was made between DGA (dry grain density) and UMA (Photoelectric Volumetric Cross-section) with the characteristic ternary diagram of the quartz, calcite and dolomite. The results were then compared with the 4-mineral identification approach. Contour maps of the average mineral fractions present in the Traverse Limestone were prepared to see the basin wide mineralogy of Traverse Limestone. In the second part, sonic response of Traverse Limestone was predicted in fifty randomly distributed wells. We used the modified time average equation that accounts for the shale effects on the basis of gamma ray log, and used it to predict the sonic behavior from density porosity and average porosity. To account for the secondary porosity of dolomite, we subtracted the dolomitic fraction of clean porosity from the total porosity. The pseudo-sonic logs were then compared with the measured sonic logs on the root mean square (RMS) basis. Addition of dolomite correction in modified time average equation improved the results of sonic prediction from neutron porosity and average porosity. The results demonstrated that sonic logs could be predicted in carbonate rocks with a root mean square error of about 4μsec/ft. We also attempted the use of individual mineral components for sonic log prediction but the ambiguities in mineral fractions and in the sonic properties of the minerals limited the accuracy of the results.

A PETROPHYSICAL EVALUATION OF THE TRENTON-BLACK RIVER FORMATION OF THE MICHIGAN BASIN

This work is conducted to study the geological and petrophysical features of the Trenton- Black River limestone formation. Log curves, crossplots and mineral identification methods using well-log data are used to determine the components and analyze changes in lithology. Thirty-five wells from the Michigan Basin are used to define the mineralogy of Trenton-Black River limestone. Using the different responses of a few log curves, especially gamma-ray, resistivity and neutron porosity, the formation tops for the Utica shale, the Trenton limestone, the Black River limestone and the Prairie du Chien sandstone are identified to confirm earlier authors’ work and provide a basis for my further work. From these, an isopach map showing the thickness of Trenton-Black River formation is created, indicating that its maximum thickness lies in the eastern basin and decreases gradually to the west. In order to obtain more detailed lithological information about the limestone formations at the thirty-five wells, (a) neutron-density and neutron-sonic crossplots, (b) mineral identification methods, including the M-N plot, MID plot, ϱmaa vs. Umaa MID plot, and the PEF plot, and (c) a modified mineral identification technique are applied to these wells. From this, compositions of the Trenton-Black River formation can be divided into three different rock types: pure limestone, partially dolomitized limestone, and shaly limestone. Maps showing the fraction of dolomite and shale indicate their geographic distribution, with dolomite present more in the western and southwestern basin, and shale more common in the north-central basin. Mineral identification is an independent check on the distribution found from other authors, who found similar distributions based on core descriptions. The Thomas Stieber method of analysis is best suited to sand-shale sequences, interpreting hree different distributions of shale within sand, including dispersed, laminated and structural. Since this method is commonly applied in clastic rocks, my work using the Thomas Stieber method is new, as an attempt to apply this technique, developed for clastics, to carbonate rocks. Based on the original assumption and equations with a corresponding change to the Trenton-Black River formation, feasibility of using the Thomas Stieber method in carbonates is tested. A graphical display of gamma-ray versus density porosity, using the properties of clean carbonate and pure shale, suggests the presence of laminated shale in fourteen wells in this study. Combined with Wilson’s study (2001), it is safe to conclude that when shale occurs in the Trenton-Black River formation, it tends to be laminated shale.

Climatic records of the late Holocene from sediment core GeoB9504-3 off Senegal

High-resolution climatic records of the late Holocene along the north-west African continental margin are scarce. Here we combine sediment grain size, elemental distribution and mineral assemblage data to trace dust and riverine sources at a shallow-marine sediment depocentre in the vicinity of the Senegal River mouth. The aim is to understand how these terrigenous components reflect climate variability during the late Holocene. Major element contents were measured and mineral identification was performed on three sub-fractions of our sediment core: (i) fluvial material <2 µm, (ii) aeolian material of 18-63 µm and (iii) a sub-fraction of dual-origin material of 2-18 µm. Results show that more than 80% of the total Al and Fe terrigenous bulk content is present in the fluviogenic fraction. In contrast, Ti, K and Si cannot be considered as proxies for one specific source off Senegal. The Al/Ca ratio, recording the continental river runoff, reveals two dry periods from 3010 to 2750 cal a BP and from 1900 to 1000 cal a BP, and two main humid periods from 2750 to 1900 cal a BP and from 1000 to 700 cal a BP. The match between (i) intervals of low river runoff inferred by low Al/Ca values, (ii) reduced river discharge inferred by integrated palynological data from offshore Senegal and (iii) periods of enhanced dune reactivation in Mali confirms this interpretation.

Identificación de minerales mediante difracción de rayos X utilizando bases de datos online de libre acceso

El artículo proporciona una metodología sencilla para la identificación de minerales mediante la técnica de difracción de rayos X de polvo utilizando bases de datos mineralógicos de libre acceso y online. Las bases de datos utilizadas son la base de datos mineralógicas webmineral y la base de datos de estructuras cristalinas de la American Mineralogist Crystal Structure Database, AMS. En el presente trabajo se han elaborado 3 actividades resueltas de estudios reales y en orden creciente de dificultad. Se ha pretendido hacer hincapié en puntos donde el profesor puede interactuar con el alumno y promover la capacidad de análisis, síntesis y razonamiento crítico del alumno ante un problema de investigación en geología. Finalmente se ha elaborado un Anexo donde se recogen recomendaciones para que el profesor desarrolle sus propias actividades.

Bone mineral density (BMD), micro-architecture estimation (TBS) and vertebral fracture assessment (VFA) extracted from a single DXA device in combination with clinical risk factors improve significantly the identification of women at high risk of fracture

Introduction: Osteoporosis (OP) is a systemic skeletal disease characterized by a low bone mineral density (BMD) and a micro-architectural (MA) deterioration. Clinical risk factors (CRF) are often used as a MA approximation. MA is yet evaluable in daily practice by the Trabecular Bone Score (TBS) measure. TBS is a novel grey-level texture measurement reflecting bone micro-architecture based on the use of experimental variograms of 2D projection images. TBS is very simple to obtain, by reanalyzing a lumbar DXA-scan. TBS has proven to have diagnosis and prognosis value, partially independent of CRF and BMD. The aim of the OsteoLaus cohort is to combine in daily practice the CRF and the information given by DXA (BMD, TBS and vertebral fracture assessment (VFA)) to better identify women at high fracture risk. Method: The OsteoLaus cohort (1400 women 50 to 80 years living in Lausanne, Switzerland) started in 2010. This study is derived from the cohort COLAUS who started in Lausanne in 2003. The main goals of COLAUS is to obtain information on the epidemiology and genetic determinants of cardiovascular risk in 6700 men and women. CRF for OP, bone ultrasound of the heel, lumbar spine and hip BMD, VFA by DXA and MA evaluation by TBS are recorded in OsteoLaus. Preliminary results are reported. Results: We included 631 women: mean age 67.4±6.7 y, BMI 26.1±4.6, mean lumbar spine BMD 0.943±0.168 (T-score -1.4 SD), TBS 1.271±0.103. As expected, correlation between BMD and site matched TBS is low (r2=0.16). Prevalence of VFx grade 2/3, major OP Fx and all OP Fx is 8.4%, 17.0% and 26.0% respectively. Age- and BMI-adjusted ORs (per SD decrease) are 1.8 (1.2- 2.5), 1.6 (1.2-2.1), 1.3 (1.1-1.6) for BMD for the different categories of fractures and 2.0 (1.4-3.0), 1.9 (1.4-2.5), 1.4 (1.1-1.7) for TBS respectively. Only 32 to 37% of women with OP Fx have a BMD < -2.5 SD or a TBS < 1.200. If we combine a BMD < -2.5 SD or a TBS < 1.200, 54 to 60% of women with an osteoporotic Fx are identified. Conclusion: As in the already published studies, these preliminary results confirm the partial independence between BMD and TBS. More importantly, a combination of TBS subsequent to BMD increases significantly the identification of women with prevalent OP Fx which would have been miss-classified by BMD alone. For the first time we are able to have complementary information about fracture (VFA), density (BMD), micro- and macro architecture (TBS & HAS) from a simple, low ionizing radiation and cheap device: DXA. Such complementary information is very useful for the patient in the daily practice and moreover will likely have an impact on cost effectiveness analysis.

Bone mineral density (BMD), micro-architecture estimation (TBS) and vertebral fracture assessment (VFA) extracted from a single DXA device in combination with clinical risk factors improve significantly the identification of women at high risk of fracture.

Osteoporosis (OP) is a systemic skeletal disease characterized by a low bone mineral density (BMD) and a micro-architectural (MA) deterioration. Clinical risk factors (CRF) are often used as a MA approximation. MA is yet evaluable in daily practice by the trabecular bone score (TBS) measure. TBS is very simple to obtain, by reanalyzing a lumbar DXA-scan. TBS has proven to have diagnosis and prognosis values, partially independent of CRF and BMD. The aim of the OsteoLaus cohort is to combine in daily practice the CRF and the information given by DXA (BMD, TBS and vertebral fracture assessment (VFA)) to better identify women at high fracture risk. The OsteoLaus cohort (1400 women 50 to 80 years living in Lausanne, Switzerland) started in 2010. This study is derived from the cohort COLAUS who started in Lausanne in 2003. The main goal of COLAUS is to obtain information on the epidemiology and genetic determinants of cardiovascular risk in 6700 men and women. CRF for OP, bone ultrasound of the heel, lumbar spine and hip BMD, VFA by DXA and MA evaluation by TBS are recorded in OsteoLaus. Preliminary results are reported. We included 631 women: mean age 67.4 ± 6.7 years, BMI 26.1 ± 4.6, mean lumbar spine BMD 0.943 ± 0.168 (T-score − 1.4 SD), and TBS 1.271 ± 0.103. As expected, correlation between BMD and site matched TBS is low (r2 = 0.16). Prevalence of VFx grade 2/3, major OP Fx and all OP Fx is 8.4%, 17.0% and 26.0% respectively. Age- and BMI-adjusted ORs (per SD decrease) are 1.8 (1.2-2.5), 1.6 (1.2-2.1), and 1.3 (1.1-1.6) for BMD for the different categories of fractures and 2.0 (1.4-3.0), 1.9 (1.4-2.5), and 1.4 (1.1-1.7) for TBS respectively. Only 32 to 37% of women with OP Fx have a BMD < − 2.5 SD or a TBS < 1.200. If we combine a BMD < − 2.5 SD or a TBS < 1.200, 54 to 60% of women with an osteoporotic Fx are identified. As in the already published studies, these preliminary results confirm the partial independence between BMD and TBS. More importantly, a combination of TBS subsequent to BMD increases significantly the identification of women with prevalent OP Fx which would have been misclassified by BMD alone. For the first time we are able to have complementary information about fracture (VFA), density (BMD), micro- and macro architecture (TBS and HAS) from a simple, low ionizing radiation and cheap device: DXA. Such complementary information is very useful for the patient in the daily practice and moreover will likely have an impact on cost effectiveness analysis.

Identification of oligomers in polyethyleneterephthalate bottles for mineral water and fruit juice - Development and validation of a high-performance liquid chromatographic method for the determination of first series cyclic trimer

Cyclic oligomers were identified in PET bottles used for mineral water and fruit juice using MS and H-1 and C-13 NMR: a first series cyclic trimer, a first series cyclic tetramer, a first series cyclic dimmer and a second series cyclic trimer. An analytical method to determine first series cyclic trimer in these bottles was developed and validated, using HPLC. The first series cyclic trimer levels were 316-462 mg/100 g of PET bottle. (c) 2005 Elsevier B.V. All rights reserved.

Palaeosurfaces mapping and associated supergene copper deposits identification as mineral exploration tool Itapeva and Ribeirão Branco region – Ribeira Valley, state of São Paulo, Brazil

Alguns exemplos de formação de minério supérgeno e sua interação com a morfogênese (paleosuperfícies) e resistência mineralógica são discutidas aqui. Registros geomorfológicos e mineralógicos na caracterização paleosuperfícies associam o intemperismo de minerais primários e sua relação com as concentrações de cobre e minério de ferro no sudeste do estado de São Paulo, Brasil. Distinguiram-se duas paleosuperfícies geradas por várias fases de intemperismo e controladas pela estrutura geológica. A primeira, mais antiga, é a paleosuperfície superior (900 - 1000 m de altitude), situado em Ribeirão Branco (Alto do Brancal), foram desenvolvidas em rochas silico-calcários. É formada por lateritas de ferro enriquecido por produtos secundários de cobre. O segundo nível de paleosuperfície é mais novo está localizado na região de Itapeva (Santa Blandina e Bairro do Sambra). Esta paleosuperfície é formada por percolação de cobre através da rocha alterada (saprolito). Outras características podem ser observadas como produtos neoformados em lateritas. Eles são classificados em dois tipos: a argila como produtos silico-cuprífero (com quantidades significativas de ferro) e de cobre minerais (crisocola, fixas nas vertentes). Essas feições reconheceram a presença de minérios de cobre e seu controle morfogénetico ajudando na exploração e prospecção de minérios supérgenos.

Bone mineral density (BMD) and vertebral trabecular bone score (TBS) for the identification of elderly women at high risk for fracture: the SEMOF cohort study

PURPOSE To determine the predictive value of the vertebral trabecular bone score (TBS) alone or in addition to bone mineral density (BMD) with regard to fracture risk. METHODS Retrospective analysis of the relative contribution of BMD [measured at the femoral neck (FN), total hip (TH), and lumbar spine (LS)] and TBS with regard to the risk of incident clinical fractures in a representative cohort of elderly post-menopausal women previously participating in the Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk study. RESULTS Complete datasets were available for 556 of 701 women (79 %). Mean age 76.1 years, LS BMD 0.863 g/cm(2), and TBS 1.195. LS BMD and LS TBS were moderately correlated (r (2) = 0.25). After a mean of 2.7 ± 0.8 years of follow-up, the incidence of fragility fractures was 9.4 %. Age- and BMI-adjusted hazard ratios per standard deviation decrease (95 % confidence intervals) were 1.58 (1.16-2.16), 1.77 (1.31-2.39), and 1.59 (1.21-2.09) for LS, FN, and TH BMD, respectively, and 2.01 (1.54-2.63) for TBS. Whereas 58 and 60 % of fragility fractures occurred in women with BMD T score ≤-2.5 and a TBS <1.150, respectively, combining these two thresholds identified 77 % of all women with an osteoporotic fracture. CONCLUSIONS Lumbar spine TBS alone or in combination with BMD predicted incident clinical fracture risk in a representative population-based sample of elderly post-menopausal women.