New Heat-Flow Observations in a Hotspot Swell: the Reunion-Mascarene Plateau
P. Chiozzi1, M. Verdoya1, C. Pasqua1,2
1University of Genoa, Depart. of Earth, Environment and Life Sciences, Genoa, Italy 2ELC – Electroconsult, Milan, Italy • Terrestrial heat-flow data, generally derived from sea-bottom measurements, can be important to validate theories on the origin of hotspot swells (e.g. Crough, 1978; Courtney and White, 1986; Watson and McKenzie, 1991; Sleep, 1994). • Strikingly, hotspots are generally characterised by the lack of high heat-flow values. This is often argument against reheating of the oceanic lithosphere as a mechanismforming swells (e.g. Von Herzen et al., 1989). • However, it was recently argued that the thermal signature of hotspots can be widely obscured by fluid circulation (e.g. Harris and McNutt, 2007). • Other analyses conclude that hydrothermal flow may redistribute heat only near the swell axes and that the small heat-flow anomalies indicate that the mechanisms producing hotspots do not significantly perturb the thermal state of the lithosphere (e.g. Stein and Von Herzen, 2007). • Closely spaced heat-flow determinations (1-2 km or less) can be useful to discriminate environments in which heat is transferred by conduction or advection
• The heat-flow data analysis also may benefit from seismic profiles providing details of sediment interfaces; the occurrence of fluid flow can be better understood
• So far, these requirements are satisfied in only for two hot spot swells: the Hawaii-Emperor chain and the Reunion-Mascarene Plateau
Focus on the Reunion-Mascarene Plateau hotspot area (western Indian Ocean) Review the available sea-bottom heat-flow determinations Present results of the first heat-flow observations on Mauritius Island, SW end of the Reunion-Mascarene Plateau The Reunion-Mascarene Plateau is an aseismic ridge which is part of the Chagos-Laccadive Ridge emerging south of the Deccan Traps flood basalt province in western India and continuing south- westward to Reunion Island, in the Indian Ocean; volcanism spans in age from 64 to 0 Ma.
It includes the Seychelles, Saya de Malha and Nazareth banks, extending north of the Mascarene archipelago Mauritius Island, the second largest island of the archipelago which also includes Reunion and Rodriguez.
A wide topographic swell is associated with this plateau.
Bathymetric map by Amante and Eakins (2009) and white boxes shown ages of the chain (Tiwari et al., 2007). Marine heat-flow data • Several heat flow determinations have been made in the investigated area since 1960’s. Early measurements were randomly carried out in the vicinity of the Mascarene Plateau. Data are contained in the heat-flow database of the IHFC. • We selected from the database only measurements ranked as of higher quality. To remove the effects of fluid circulation, we carefully filtered heat flow measurements (following Hasterok et al., 2011): - sediment thickness > 400 m - distance from seamounts > 60 km
• Another HF dataset is available from Bonneville et al. (1997). They carried out high-resolution heat-flow determinations along seismic reflection profiles in a zone not directly affected by hot spot activity since at least 15 Ma.
• We processed data exhibiting successful probe penetrations (> 2.5 m) and further refined data selection by rejecting measurement sites with poor information on both thermal gradients and in situ thermal conductivity. Data correction • Sediment perturbation: a simple model of constant sedimentation rate (Von Herzen and Uyeda, 1963).
• The deposition rate was inferred from the sediment thickness maps (Divins (2007) and seafloor age data (Müller et al., 2008). The sediment thickness along the heat-flow profile of Bonneville et al. (1997) was estimated by converting travel-times into sediment thickness with the relationship by Tucholke et al. (1982).
• A thermal diffusivity of 0.3 mm2 s-1, with no compaction or associated fluid expulsion (Hasterok et al., 2011).
• Average value 66 11 mW m-2
• Range 37-88 mW m-2
Distribution of heat-flow data (full circle) at sea in the Reunion-Mascarene Plateau hotspot from Von Herzen and Langseth (1965), Birch and Halunen (1966), Sclater (1966), Anderson et al. (1977) and Bonneville et al. (1997). Relief model is by Amante and Eakins (2009) and map of seafloor age (in Ma) by Müller et al. (2008). Comparison with reference HF model
• In order to interpret ocean heat-flow data, it is fundamental the choice of a reference model of conductive heat transfer through the lithosphere. • We adopted the model recently proposed by Hasterok (2013) which relies on the recalibration of the classical cooling-plate model by removing the effects of hydrothermal circulation and sedimentation from the global heat-flow database.
q 506 .7 t For t t 48 .1 Ma
q 53 106 e 0 .034607 t For t t 48 .1 Ma 1.5
1.3 Median values (squares) grouped by age for 2.5 Ma bins of relative heat flow (ratio between heat-flow data 1.0 and the values predicted by the reference model) versus seafloor age. Bars: interquartile ranges. Gray step line: values used for the plate model by Hasterok, 0.8 Observed/predicted heat flow heat Observed/predicted LSDA-11 LSDA-15 2013,. Q1 and Q3: first and third quartiles
Median Q1 Q3 Heat flow data
0.5 50 55 60 65 70 75 80 85 Age (Ma)
A deviation from the reference cooling-plate model of the heat-flow data of the RM plateau would be expected, for they are located on the hotspot swell
However, the heat-flow measurements are generally consistent with the model, with the exception of two values of, which present a relative heat flow lower than that expected, probably due to superficial perturbations
In conclusion, marine data do not put into evidence any significant heat-flow anomaly on the hotspot swell
However, this neither proves nor disproves the usefulness and meaning of heat-flow over hot spots; in spite of the careful data filtering, a possible bias caused by fluid flow in the marine sediments cannot be excluded Heat flow observations in Mauritius Island • So far, the only available thermal information for the Mascarene archipelago consisted of temperature measurements in exploration holes on the Reunion Island (Demange et al., 1989; Sanjuan et al., 2000) the only island of the Mascarene with active volcanism. Unfortunately, no detail on temperature data accuracy, thermal conductivity and consequently on heat flow is available • The recent drilling of a geothermal exploration borehole in the Mauritius Island gave an opportunity for heat flow observations on land • Mauritius is the second youngest island of the mantle plume track. Volcanism has developed during three phases, dated 7.8−5.5 Ma, 3.5−1.9 Ma, and 0.7−0.03 Ma, which are termed Older, Intermediate, and Younger Series, respectively (Paul et al., 2007; Moore et al., 2011 and references therein)
20° 00' S Grand Baie 1 2 3 Port Loui 4
BH1226
20° 15' S
Maherourg Location of the heat flow site (BH1226) and geologic map of Mauritius. Volcanic Series: (1) Older, (2) Intermediate, (3) Younger; (4) caldera limit 57° 20' E 57° 40' E •The temperature profile in the hole section between 35 and 180 m shows distortions ascribable to water strikes.
•Minor inflows occur at about 50 and 110 m, whereas a prominent perturbation of the thermal profile is visible at 150-200 m.
•Water strikes implies advection and thus the temperature gradient inferred in this section does not mirror the deep thermal regime.
•The deepest part of the thermal profile shows instead a conductive thermal regime and the average thermal gradient is 43 mK m-1.
Temperature (dots), thermal gradient versus depth and stratigraphy of the borehole BH1226. (1) Older, (2) Intermediate, (3) Younger volcanic Series. Bold line marks the lateritic levels. Dashed line indicates the least square linear regression Thermal conductivity measurements on water-saturated core samples recovered from the hole by means of a transient divided bar apparatus (Pasquale et al., 2015)
Average thermal conductivity a of samples sorted according to the volcanic cycle. V= vesicular; M= massive. s.d. = standard deviation.
There is no substantial difference between the Younger and Older volcanic series rocks thermal conductivity. In lava flows, the main controlling factor seems the rock structure, being thermal conductivity lower in vesicular lavas (on average 1.10 W m-1 K-1) and larger in massive rocks (1.52 W m-1 K-1), whereas tuff exhibits intermediate values (1.25 W m-1 K-1). By combining the thermal conductivity with the temperature gradient calculated with 5 m step from 215 m depth to the hole bottom we obtain a heat flow value of 6118 mW m-2
On-shore observations seem to confirm the low HF of the marine data and therefore absence of a thermal anomaly on the Reunion-Mascarene Plateau
This again suggests that under the hotspot area the lithosphere thermal conditions are close to the reference plate model Plot of effective elastic thickness (dots) vs. age of the lithosphere at time of loading. Data: 1 Reunion-Mauritius, 2 Nazareth Bank, 3 Saya de Malha Bank, 4 Seychelles (Tiwari et al., 2007); 5 Mauritius, 6 Reunion (Boneville et al., 1988); 7 Mauritius-Nazareth Bank (Crosby and McKenzie 2009); 8 Reunion, 9 Mauritius (Calmant et al., 1990). Isotherms are calculated with the cooling plate model by Hasterok (2013) Residual topography
The topography predicted using the depth-age model by Hasterok (2013)
d ( t ) d o 414 .5 t t 17 .4 Ma
d ( t ) d o 3109 2520 exp ( 0 .034607 t ) t 17 . 4 Ma
was subtracted from the global relief model ETOPO 1 after correction for sediment loading performed using method proposed by Sykes (1996) and the sediment thickness data
compiled by Divins (2004). The ridge depth do is 2514 m in accordance to the seafloor-age global dataset of Crosby and McKenzie (2009)
5000
4000 Profile A-A'
) Profile B-B' Profile C-C' 3000 The occurrence of a topographic swell and volcanic 2000 activity along the hotspot area argue for a thermal 1000 anomaly in the mantle Bathymetric anomaly (m 0 0 5 10 15 20 Distance (°) Thermal model
We used a simple conductive model (Crosby and McKenzie, 2009), applicable for a plate moving slowly with respect to the plume (13 mm yr-1; Gripp and Gordon 2002), to evaluate the possible lithosphere reheating due to a plume Lower boundary condition: plume head temperature
Potential temperature (°C) 1200 1300 1400 1500 0
Reunion 1 3 0.2
Plate thickness 2 1 = 90 km 2 = 60 km 3 0.4 3 = 30 km Melt production rate (km /yr) (km rate production Melt 0.6 Active plume decompression beneath intact lithosphere
Mantle temperature of a plume lying beneath intact lithosphere as a function of melt production for different lithospheric thicknesses (after White and McKenzie, 1995). The melt rate of the Reunion-Mauritius hotspot is 0.022 km3 yr-1 (Crosby and McKenzie (2009). A lithosphere thickness of 85 km was used as inferred from the cooling plate model for a seafloor age of 80 Ma
Tp = 1480 °C Thermal model
• The temperature at the base of the plate is permanently raised from about 1370 to 1500 °C
T(z) prior to plume impingement
T(z) at t = 0
T(z) at t = 10 Ma
T(z) steady state temperature after perturbation • For age spanning from 64 to 0 Ma the maximum heat flow anomaly is < 5 mW m-2
• The model leads us to believe that only about < 25% of elevation has a thermal origin, i.e. due to lithosphere reheating Conclusions The careful review of the HF marine data lead us to suggest that no significant thermal anomaly exists in the RM hotspot area
This is in contrast with previous studies which pointed to at least a small HF increase. This is due to more selective criteria in filtering the HF dataset but also the use of new reference models for ocean floor age and heat flow
New thermal data (Mauritius) corroborate the result that the thermal anomaly in the hotspot area is absent or negligible, both offshore and onshore
This seems to suggest that the “low” (normal), marine HF is not due to fluid circulation in sediments
Large residual topography anomaly, which points to a positively buoyant, large-scale, mantle upwelling, does not seem accounted for by thermal isostasy
If plate is fast, no thermal effect is expected. In the RM hotspot, despite the relatively low absolute velocity of the plate, thermal calculation argues for a negligible effect both in HF and topography even for plume head temperature in excess of 150-200 °C
Dynamic support of the RM swell. Moreover, the southward decrease of both the swell size and the magma volume could be evidence of a decrease of the plume activity over time, consistently with recently proposed modeling of mantle convection