RUSSIAN JOURNAL OF EARTH SCIENCES VOL. 10, ES4001, doi:10.2205/2007ES000218, 2008

3. Data Applied

[27]  Geodynamic zonation of the Atlantic Ocean lithosphere encloses deep-water areas, the Mid-Atlantic Ridge area, passive continental margins, and offshore area (Figure 1). The analysis does not include arc and backarc zones of the Caribbean and Scotia, seas different in geodynamics in environments characteristic of the entire Atlantic. Thus, the class of phenomena studied does not contain collision zones. It is the 82o N at the transition of the Mid-Atlantic Ridge zone to Gakkel Ridge. To the south, it is confined to Bouvet triple junction where a drastic change in the structural pattern of most geophysical anomalies takes place farther south. Analysis area is shown on Figure 1. The brief description of chosen parameters for each 1otimes1o cell multivariant vector will be presented.

3.1. Bottom Topography

[28]  Bottom topography in the first and the major parameter describing the top of the earth's crust and the lithosphere (see Figure 1). It was inferred from ETOPO5 (1995), lowpass filtered frequency filtration and recalculated to one degree cell. Its shape gathers together the effect of many processes: magmatism, ocean floor deformation, sedimentation, etc. In our classification it is assigned to group 1 of parameters describing the object geometry being so a direct measure of the required characteristics. Qualitatively, the bottom topography assumes the result of the crustal block movement under the effect of contacting forces (parameter of group 3). An accurate observation over the movement similar to ground-based GPS measurements for the ocean floor are absent, off processing of the bottom topography data indirectly, though not precisely, is accounting these movements.

3.2. The Thickness of Sedimentary Cover

2007ES000218-fig02
Figure 2
[29]  The thickness of sedimentary cover in the Atlantic ocean was inferred from data of Laske and Masters [1997] and is shown in Figure 2. These authors collected information on the sedimentary cover (averaging thickness by 30 arc minutes grid to make corrections to tomographic model. In the present paper the data are arranged to fit its detailed pattern. There are several reasons of sediment thickness usage. The main is that the ocean periphery being a zone of intense sediments deposition delivered from the continents exhibits unbalanced isostatic state between the crustal blocks and viscous mantle substrate caused by the higher load on the latter. This results in processes striving to restore the balance by response vertical movements trying to bring the medium back into isostatic equilibrium and decrease the aquiered disturbance. Another reason why the sedimentary cover should be accounted into estimates is a great difference in density between bottom and crystalline basement. This surface describes properties of group l and is responsible for object geometry - in this case a stratified conformity of the Earth's crust with the upper mantle.

3.3. Tomography Inferred From Surface Love Waves

2007ES000218-fig03
Figure 3
[30]  This parameter is based on the data provided by Larson et al. (http://www.seismology.harvard.edu, 1999) and shown on Figure 3. In fact, it is a "surrogate'' used to describe object geometry reflecting not directly but indirectly behavior of an efficient base of the lithosphere layer (depth to the lithosphere for the entire Atlantic has not been measured). It shows the dependence of surface wave phase velocities (waves propagating in an efficient surface layer) on the layer thickness. The greater the thickness, the less is wave velocity and, in contrast, velocity increases as thickness decreases. Therefore, phase velocity deviation from average value implies relative variations of the efficient surface layer. They are proportional to the required parameter, i.e. a depth to the lithosphere base. The 35 s period wave (the shortest of published) model was used for calculations, in this case penetrations of displacements along the wave front is not deep and approximately corresponds to the lithospheric layer. Long period waves involve deeper layers into wave motion. Figure 3 clearly shows that continental areas and those with low velocities in regions of intense magmatism can be easily distinguished from the oceanic zones underlain by thin and high velocity lithosphere. It also shows that mountain edifices with deep roots have adequate minima complying with geometry of the surface layer base, it implies that the selected parameter can be used as a "surrogate'' in description of the lithosphere base geometry. However, a peculiar linear anomaly zone extending along the azimuth at about 30o N and not orthogonally crossing the Mid-Atlantic Ridge near the equator can be discernible in the Atlantic.

3.4. Bouguer Anomaly

2007ES000218-fig04
Figure 4
[31]  Bouguer anomaly was calculated from EGM97 data [Hwang et al., l997] and from bottom topography data (ETOPO5, 1993) for the average density of the oceanic crust at 2.8 g cm-3 and the land density at 2.67 g cm-3, and 166 km in integration radius, shown in Figure 4. EGM97 matrix originally was on the 2-minute grid, therefore this and ETOPO5 matrix were smoothed prior to calculations to the 10-minute grid followed by averaging of the result to 1 arc degree cell. Gravity anomalies of EGM97 are free-air anomalies. In offshore area it means that about 80% of anomalous field variability is proportional to the most distinct density boundary - to the bottom topography inferred from echo soundings. Calculation of Bouguer anomaly, i.e. "addition'' of the crustal density masses into the water layer (which is less than an average value) that will eliminate the effect of bottom topography on the anomalous field. Hence, variability of residual field will show mainly a depth difference in density at the crust-mantle boundary along with lateral density heterogeneities in the crust and mantle. In deep water basins they can be small or fairly important where serpentization of the upper mantle rocks take place but the absence of deep seismic sounding does not allow to tell reliably their effect from variation in a depth to the crust base. However, lateral heterogeneities in the MAR area can be great and occupy extensive zones of hundreds of kilometers (e.g. Azores and Iceland plume areas). They show heated zones where the lithosphere has magma chambers and areas exhibiting high partial melting. These zones are marked by intense magmatism, and accordingly greater crustal thickness, the latter being the load on the viscous mantle substrate, increases a depth of the M-discontinuity. The above means that Bouguer anomalies are proportional to a depth of the crust-mantle boundary and the less the anomaly value, the greater is the depth. The cases when it is reasonable to introduce a correction into the anomalous field for thermal effect should be accounted for heat flow data or another parameter reflecting the heated state, for example, tomography inferred from S-waves. However, the parameters being independently used in our geodynamic analysis (see below), calculation of Bouguer anomalies accounted for thermal correction makes no sense. Thus Bouguer anomalies belong to the first group parameters defining the lithospheric layer geometry or the inner boundary to distinguish the "dense stages'' of the crust and mantle from the variation pattern of masses along the lithosphere. The features of Bouguer anomaly mentioned are the combination of contribution whose reliable separating seems difficult. Bouguer anomalies are "surrogate'' parameters for description of geometry and are true for description of mass variation.

3.5. Isostatic Anomalies

2007ES000218-fig05
Figure 5
[32]  The authors calculated isostatic anomalies by means of the Bouguer anomaly data and topography (ETOPO5) for the average density of the oceanic crust, continent density, and the mantle density of 2.8 g cm -3, 2.67 g cm -3, and 3.3 g cm -3, respectively, with radius of 166 km integrating by the Airy model and the surface reduction depth of 33 km (Figure 5). The long-wave components of above 900 km were eliminated from the anomalous field because they reflect sublithospheric heterogeneities and their effect obscures the processes operating in the upper shell of the Earth. On elimination of the anomalous field variability related to the upper boundary of the crustal masses and determination of Bouguer anomaly, the estimate of isostatic anomalies controls the hypothetic field variability caused by the change of the compensation surface topography due to the difference in thickness of the crustal blocks drifting upon the viscous mantle surface. The authors proceed from the fact that in case of isostatic equilibrium the position of the compensational surface and topography can be expressed by a simple equation:

H=T+hast(s c-s w)/(s m+s c)

where H is a depth of the compensational surface, T - level of reduction, s c - crustal density, s w - water density, s m - mantle density; this will allow us to calculate correction for Bouguer anomaly. They eliminate the effect of a hypothetical surface topography obtained as in the case of bottom topography. The residual field represents isostatic anomalies when their positive values imply an excess of masses above the compensation surface unlike the negative values pointing to their deficiency. The excess of masses might well result in submergence of the crustal block in a given site, while emergence together with the mantle part owes it to deficiency. If the action (e.g., thrust) is not completed then we'll get both the excess of masses (positive isostatic anomalies) and positive vertical movements of the crust. The interpretation of isostatic anomalies being ambiguous, its resolution calls for further investigation of the general tectonics of the region. In terms of geodynamics this parameter concerns directly the variation of crust density properties intensity of energy release in the crust, and generation of stresses (modulus of isostasy gradient) caused by the transition from disturbed state into equilibrium. This parameter is also a "surrogate'' in description of the resultant vertical movement of the crustal blocks subjected to energy release. The isostatic anomaly field presents the above properties as, in fact, an indivisible combination.

3.6. Heat Flow

2007ES000218-fig6b
Figure 6
[33]  Heat flow is inferred from the data by Pollack et al. [1991], Podgornykh and Khytorskoy [1997], see Figure 6. Figure 6b shows it is not equally studied in the Atlantic but we have to use it in our estimates as it has been determined by the authors mentioned. Grid was calculated for each cell degree of the region to use scattered and irregular cloud of values by means of "kriging'' technique, followed by reduction of its high frequency component to the level of other parameters to minimize the effect of irregular density of measurements. The obtained map (Figure 6a) differs in data available for the polar areas and for basins. Besides, there is an area on the Mid-Atlantic Ridge (Azores archipelago) also poorly studied. Heat flow being a true parameter of group 2 reflecting energy release, must be included into calculations. It is clear that irregular pattern of knowledge will make this parameter a tool for reliable classification in areas with a dense network of observations, and the opposite will take place in areas with low density of measurements, in such a way the latter will affect the results. In this case, it is the best we have at our disposal.

3.7. Tomography by S-Waves

2007ES000218-fig07
Figure 7
[34]  Tomography by S-waves was inferred from the data by Grand et al. [1997], Becker and Boschi [2001] and is shown in Figure 7. The uppermost segment of NGRAND model from 0 to 100 km, calculated by its authors for 2o by 2o blocks and represented by spherical harmonics of the 31 order. Matrix of tomographic values was recalculated for 1otimes1o grid. These values show the variation of S-wave propagation velocity from the average within the layer (in %). This parameter responds to heated zones exhibiting high partial melting. It clearly points to the presence of plumes commonly accompanied by magmatism and to zones of mid-oceanic ridges. These zones are characterized by negative, inferred from tomography, values: -3.5% and less because warmed up and viscous medium decreases seismic velocities.

[35]  So, this parameter is an almost indivisible combination of effects of energy release (heated state) and medium geometry (zone of prolific magma production and greater thickness of the crust). This parameter is "surrogate'' for both groups reflecting indirectly and not directly properties of the groups.

3.8. Tomography by P-Waves

2007ES000218-fig08
Figure 8
[36]  Tomography by P-waves is inferred from data by Van der Hilst et al. [1997], Becker and Boschi [2001] (see Figure 8). The uppermost section HWE97p from 0 to 100  km calculated by the above authors for 2otimes2o blocks and represented by spherical harmonics to the 31 order was used to study geodynamics of the lithosphere. Matrix for the tomographic part of values was recalculatad to 1otimes1o grid. Like in the case of S-waves, P-waves should be accounted for thermal state of the subsurface. So, tomography by these waves must be similar to that by S-waves. However, in practice such is not the case. According to Becker and Boschi [2001] the S- and P-models correlate better toward the middle part of the mantle (above l000 km) implying a similarity in cases responsible for variability of parameters. The behavior of the S- and P-models differs greatly in the mantle from that on the surface. If pattern by S-model is easily to explain, then distribution of values by P-model should be accounted for sources that caused velocity variation. The authors consider the presence of the stressed condition of the lithosphere and (or) related fracturing system as a possible cause. The system is responsible for a peculiar pattern of highs and lows distribution on the map (Figure 8). Low velocities are seen to be concentrated along the collision zones of the Earth whereas high velocities occur in the rear part. These zones are marked by large-scale fractures, their disposition in plan is not aligned with direction of forces generating collision area. These fractures might well "slow down'' the velocity of the P-waves. The parameter points to a stressed state of the medium and relates systems of faults with relieving of stress along them. Thus, it combines parameters of group 2 and group 3 showing both energy release in the medium and the resultant action of forces. The absence of an accepted regional geodynamic interpretation of this parameter makes its discussion in context of remaining parameters more stimulating.

3.9. Total Seismic Moment

[37]  This parameter is used for calculation of the total energy released during earthquakes. It was a global query (ANSS, February 2004, http://quake.geo.berkeley.edu/anss/) for events with a Richter magnitude of above 4.5 for a layer of 0 to 100 km. The approach published by Boldyrev [1998] was used in our estimates. Summation of released energy for events within a degree cell was calculated from the formula

M=(10(17.1+1.3cdot ( Mag-5)))/10+13 [Jcdot 10+13],

2007ES000218-fig09
Figure 9
where M is total moment, Mag - magnitude on the Richter scale. The estimate of total moment was followed by calculation of density moment for sq. km accounted for changes in area of call degree at high latitudes. The resultant value shown on the map (see Figure 9) is [J  km-2]cdot 10+13. This parameter is marked by an extremely irregular distribution in the study area. Besides, no more than 5% of the entire seismic energy of the planet is released along the Mid-Atlantic Ridge. Therefore, variably of the parameter is mainly out of the region unlike other parameters having within the region values close to minimal and maximal. A given magnitude of scale not imposing limits on record of seismic events along the distance, the entire region is regularly crated by values of the parameter, however, for most of the area it equals zero. Seismic moment belongs to group 2 showing energy release.

3.10. Lithospheric Component of the Earth's Magnetic Field

2007ES000218-fig10
Figure 10
[38]  This parameter was found by processing of satellite CHAMP data [Maus et al., 2002]. At the orbit altitude of about 450 km and trajectory crossing the earth poles this satellite made possible to register the lithospheric component of the anomalous magnetic field or the entire area of the Earth. The authors constructed magnetic anomaly map of full vector, vertical component, and gradient modulus of the full vector. Our study uses the last parameter and Figure 10 shows the map for recalculation to the altitude of 100 km. This altitude is approximately equal to the thickness of the lithosphere and provides a proper averaging. The gradient modulus of full vector has an advantage because the field lacks alternating signs due to changing direction of the magnetized field, i.e. the cause related to the features of the lithosphere which greatly simplifies an interpretation. This parameter is proportional to concentration of magnetically active minerals in the lithosphere and points to reasons responsible for variability of their concentration. Among them are: a depth of the Curie isotherm, the presence of serpentinization zones, zones of intense magmatism differing in composition from that of the adjacent territories, etc. In other words, this parameter is a "surrogate'' as concerns properties of group 2, namely, in energy release, and partly those of group 1 - geometry of deep-seated boundaries.


RJES

Citation: Sokolov, S. Yu., N. S. Sokolov, and L. V. Dmitriev (2008), Geodynamic zonation of the Atlantic Ocean lithosphere: Application of cluster analysis procedure and zoning inferred from geophysical data, Russ. J. Earth Sci., 10, ES4001, doi:10.2205/2007ES000218.

Copyright 2008 by the Russian Journal of Earth Sciences

Powered by TeXWeb (Win32, v.2.0).