3D Geophysical-petrological modelling of the lithosphere: how can GOCE data help us assessing the geothermal potential of Ireland? J. Fullea(1), Z. Martinec(2), M.R. Muller(2), A.G. Jones(2) 1. Institute of Geosciences (IGEO) CSIC-UCM Madrid, Spain 2. DIAS, Geophys. Section, 5 Merrion Square, Dublin, Ireland Structure of the presentation Summary LitMod (geophysical observables+mantle petrology) IRETHERM: lithospheric structure and regional thermal field in Ireland GOCE gravity gradients: new “directional “ constraints for the lithosphere?? LitMod3D (http://eps.mq.edu.au/~jafonso/Software1.htm) # LitMod3D is a 3D computational code that allows to model all relevant physical properties within a robust and thermodynamically self-consistent framework (Afonso et al., 2008, Fullea et al., 2009, G-cubed). LitMod3D LitMod : Thermodynamic modelling and bulk properties All thermophysical (e.g. density and seismic velocities) properties depend ultimately on T, P, and Composition. dG = V dP - S dT + Σi µi dXi + DdE In the mantle, stable mineral assemblages are computed by Gibbs free energy minimization either within the system CaO-FeO-MgO-Al2O3-SiO2 (CFMAS) or Na2O-CaO-FeO-MgO-Al2O3-SiO2 (NCFMAS) [Connolly, 2005]. Each mantle body is therefore characterized by a specific major-element composition (in wt.%), which translates into specific bulkrock properties. All necessary files containing thermodynamic information can be generated either with the freely available software Perple_X [www.perplex.ethz.ch, Connolly, 2005] or by using a simple interface provided with LitMod3D. LitMod3D LitMod3D : Thermal field Cartesian finite differences grid Mantle thermal conductivity dependent on T, P and C IRETHERM (http://www.iretherm.ie/index.html) Academic-government-industry collaborative project between DIAS, UCD, UCC, GSI, GSNI, and SLR Consulting targeting Ireland's geothermal energy potential through integrated modelling of new and existing geophysical and geological data. 4 4 1 1 2 2 INSET 3 5 7 6 6 7 8 8 9 10 3 5 9 10 TEMP (°C) 250 100 60 40 20 Left: regional heat-flow density contours, from ~40 mW/m2 (blue) to ~80 mW/m2 (red). Middle and right: potential survey areas to investigate eight “type” geothermal targets. Middle: overlain on the surface geology of Ireland (courtesy GSNI and GSI). Right: overlain on modelled temperatures at 2,500 m depth Inset: Lithospheric structure: P-wave tomography & S-receiver functions From O’Donnell et al. et al. 2011 From Wawerzinek et al. 2008 Indication of S-N lithospheric thinning but resolution is poor, particularly in the North! LAB depth from S-receiver functions (Landes et al., 2007) Geophysical observables: elevation and potential fields Geoid anomaly (n>10) EGM 2008 Bouguer anomaly (land + satellite data) FA anomaly (Smith & Sandwell 97) Elevation (ETOPO2 V9.1) Lithospheric models 3) N-S thinning model-composition From seismic refraction+1D inv. of geoid & elevation 1) Flat model Average mantle compositions, M&S95 refers to McDonough and Sun (1995) 2) N-S thinning model Lithospheric models: predicted SHF and P-wave tomography Calculated surface heat flow model 1 Calculated surface heat flow model 2 Flat and N-S thinning lithospheric models predict quite different SHF and Vp patterns 1) Flat model 2) N-S thinning model Synthetic seismic tomography, P-wave anomaly (average Vp model velocity subtracted at each depth) Lithospheric models: residuals Lithospheric models compatible with “traditional” potential field + topography data 1) Flat model 2) N-S thinning model GOCE data Model GOCO02S (satellite only, GOCE, GRACE and CHAMP, and SLR data), 8 months GOCE data (Nov 2009-July 2010) Goiginger H., et al. (2011) Coordinates convention: X East Y North (Cartesian reference frame in the forward models) Gravity gradients at z= 250 km computed by spherical harmonic synthesis code (Z. Martinec). Grid resolution: 15’x15’ (30 km X 30 km approx.). GOCE modelling: gravity gradients Gravity gradient residuals (synthetic -GOCE) for models 1, 2 and 3 Std. dev. of 30-50 mE in the different components Gravity gradient residuals for models 1, 2 and 3 differ in <1mE ! At the satellite height (250 km) the differences in the gravity gradients between the alternative lithospheric models are small Synthetic gravity gradient residuals at z= 250 km. Grid resolution: 9 km X 11 km approx.). GOCE modelling: gravity gradients Exploring different datum's : “raw” downward continuation of GOCE data GOCE data (diagonal components) z=20 km Not too close to the Earth’s surface different frequency content! Synthetic data (diagonal components) z=20 km Gravity gradients: ‘flat’ vs ‘N-S thinning’ models At “intermediate depths” the frequency content of synthetic and GOCE gravity gradients are comparable. 1) Flat model gravity gradient residuals at z= 100 km Statistics of the residuals (diagonal + zy component) Std. dev. The N-S thinning. model shows lower residuals than the flat model. The differences between the models are ~ tens of mE. 2) N-S thinning model Gravity gradients: ‘N-S thinning-compositional’ vs ‘N-S thinning’ models The N-S thinning model shows lower residuals than the N-S thinning-compositional model . The differences between the models are ~ tens of mE. 3) N-S thinning-composition model gravity gradient residuals at z= 100 km Statistics of the residuals (diagonal + zy component) Std. dev. All gravity grad components improve except Uxx 2) N-S thinning model Summary: imaging the Irish lithosphere • IRETHERM → Regional component in the thermal field in Ireland? • N-S thinning vs flat models → equivalent from « traditional» potential field + topography data modelling • GOCE gravity gradients→ at intermediate heights (z=100 km) the N-S thinn. model shows over all lower residuals than the other models (Uzz, Uyy, Uzy), the differences in std. dev. are ~tens of mE (resolution of GOCE data ~ 0.4 mE, R. Pail dixit). Potential for lithospheric modelling (even “medium” wavelengths ~ few hundreds of km, and moderate LAB topography variations )? LitMod3D LitMod3D : Potential fields ∆g FTP Gγ r = (ρ ) = G ρ − xy ln (r + z x 2 ρ (z) = ρ ∆N Mass distribution Right rectangular prisms in a certesian coordinate system FTP (ρ ) = + y 0 2 x ln 0 + z )− (y + r )+ y ln ( x + r )− z arctan x xy arctan + 2 zr 2 2 z arctan 2 xy zr y yz + 2 xr x2 y2 x1 y1 z2 + z1 2 arctan xz yr x2 y2 z2 x1 y1 z1 2 + γz Gρ0 g Gγ + 3g xy ln ( z + r ) + yz ln ( x + r ) + xz ln( y + r ) − P ( ) ( ) x2 y2 z2 x1 y1 z1 y x 2 xy xyr + y 2 + 3 z 2 ln( x + r ) + x + 3 z 2 ln( y + r ) − z 3 arctan 2 2 zr x2 y2 z2 x1 y1 z1 LitMod3D (http://eps.mq.edu.au/~jafonso/Software1.htm) LitMod3D : Forward modelling Objective: 3D interactive modelling of the lithosphere and uppermost mantle
© Copyright 2026 Paperzz