Silica Geothermometry Applications in the Taiwan Orogenic Belt

The silica concentrations from 130 hot spring measurements were calculated and used to indicate the heat flow distribution in the active Taiwan orgenic belt. The heat flow spatial distribution seems to be controlled by the Taiwan arc-continent collision zone tectonic development. Two abnormally high silica heat flow regions, 160 and 190 mW m-2, are identified in northeastern and southeastern Taiwan, respectively. In northeastern Taiwan, the Chingshui area, the data suggests that the anomalous heat flow distribution may be controlled mainly by hot fluids and faults. The highest heat flow in southeastern Taiwan may be generated by heat advection caused by a rapidly uplifting hot crust with a decreasing mature arc-continent collision zone exhumation rate. In addition, we estimated the exhumation rate in different tectonic zones in Southern Taiwan by applying fission track ages and silica heat flow values. The exhumation rates are 1.33, 1.72 3.87, and 0.51 1.74 mm yr-1 in the tectonic zones around Taiwan for advanced, initial arc-continent collision and accretionary wedge deformation, respectively.


INTRODUCTION
It is well-known that a surface heat flow map is a useful tool for constructing regionally thermal regimes and revealing lithosphere and tectonic evolution thermal structure, especially for an active collision orogen.In general, a conventional heat flow map is derived from measuring subsurface rock temperatures and thermal conductivity.This method is common, direct and precise.However, drilling is time consuming and extremely costly.Fortunately, the silica geothermometry provides an inexpensive and quick alternative in comparison.Even though it is not as precise as the former method, it could produce sufficient information for understanding regional surface heat flow.Several studies have successively applied this technique in the USA, Egypt, Mexico, Anatolia, Fujian, and Australia (Swanberg and Morgan 1980;Prol-Ledesma and Juarez 1986;Vugrinovich 1987;Wan et al. 1989;Ilkişik 1995;Pirlo 2002).
The island of Taiwan is located within an actively complex arc-continent collision.The Philippine Sea plate has been moving toward the WNW at about 80 mm yr -1 resulting in an ongoing mountain-building process (Yu et al. 1997).Due to the oblique collision, Taiwan exhibits various uplift rates (Liu 1982;Lan et al. 1990;Lundberg and Dorsey 1990;Liu et al. 2001;Lee et al. 2006) in all geological physiographic provinces of Taiwan Island from the north to the south, which include (1) the Coastal Range, the northern extend of Luzon arcs, consisting of sedimentary and volcanic rocks; (2) the Longitudinal Valley, the plate boundary between the Eurasian and Philippine Sea Plates; (3) the eastern Central Range, a pre-Tertiary metamorphic complex rock, containing the Yuli belt and western Tailuko belt; (4) the western Central Range composed of argillite, including the Backbone Range and the Hsuehshan Range belts; (5) the Western Foothills, the fold-and-thrust belt, consisting of clastic Oligocene-Pleistocene sedimentary rocks, and; (6) the Coastal Plain containing younger sediment deposits (Fig. 1) (Ho 1988).This oblique collision contributes to several unique general distribution patterns such as the age and metamorphic grade of Taiwan, increasing from the west to the east across Taiwan Island (Chen and Wang 1995) and the aseismic zone in the Central Range of Taiwan, which demonstrate the arc-continent collision zone complexity of Taiwan Island (Lin 2000).Several papers have been published that model the Taiwan arc-continent collision thermal structure and mountain-building mechanism (Barr and Dahlen 1989;Hwang and Wang 1993;Lin 2000;Song and Ma 2002;Zhou et al. 2003;Chi and Reed 2008), based on the conventional borehole measurements (Fig. 1a) (Lee and Cheng 1986).However, the sample distribution from previous studies is either restricted to certain areas or inadequately represented the area, especially for the Central Range of Taiwan.
The silica geothermometry (Swanberg and Morgan 1980), using hydro-geochemical data to estimate regional surface heat flow, is well suited to Taiwan due to its abundance of hot springs (Fig. 1b), which are distributed widely and easy to access for sampling purposes.In order to improve our understanding of the Taiwan mountain-building process in terms of thermal evolution, this study makes the first attempt at using the silica regional heat flow data to construct the possible thermal structures and controlling components in Taiwan.

MATERIAL AND METHOD
Enormous contributions have been made in silica temperature development which is based on the quartz solubility in water (Kennedy 1950;Bödvarsson 1960;Morey et al. 1962;Fournier and Rowe 1966;Mahon 1966).This geothermometer methodology has been proven the best method to calculate the subsurface temperature.There are four assumptions utilized in applying the silica geothermometer.They are (1) the equilibrium between water and rock is valid, (2) no other water body enters as water ascends to the surface, (3) no precipitation occurs as the water reaches the surface, and (4) there is an unlimited supply for silicon dioxide.All collected samples were filtered through 0.22-μm cellulose filters in the field and stored in high-density polyethylene (HDPE).The silica concentration of these springs was measured using ICP-AES (Type Jobin-Yvon ULTIMA2).
One hundred and thirty spring samples were collected from all over Taiwan, as shown in Fig. 1b.All samples were analyzed to construct a regional silica heat flow map involving the three necessary processes.The silica concentration of these springs was measured using ICP-AES (Type Jobin-Yvon ULTIMA2).The silica concentrations were greater than 6 ppm (Fig. 2) indicating that quartz reached equilibrium with spring water (Morey et al. 1962).The silica temperature of these springs was then estimated using silica geothermometry (Fournier and Rowe 1966).To obtain regional representatives of the silica temperatures in this study, we set up the following criteria.If springs were located within the intersected region of two spring circles within 10 km radius (Fig. 3a), these springs were noted as one group (Fig. 3b).On the contrary, if the springs belonged to different physiographic provinces, they would not be averaged or grouped into separated blocks.Last, heat flow data were calculated using a linear equation, T SiO2 = mq + T 0 , where T SiO2 is the silica geothermometry (°C), q is heat flow (mW m -2 ), T 0 is the long-term mean annual surface temperature, and m is the slope and designated as 0.68 (Swanberg and Morgan 1980).The slope m is a constant and corresponds to the minimum mean groundwater circulation depth at about 1.4 ± 0.5 km.Although these values are not fully confirmed, they are not As was mentioned above, this study followed the silica geothermometry method steps to calculate the silica heat flow value of Taiwan.The silica heat flow map produced in this study was made using the Kriging method.

RESULTS
The estimated silica heat flow values for different physiographic provinces are 78 ± 23 mW m -2 in Western Foothills; 117 ± 39 mW m -2 in west Central Range, 126 ± 49 mW m -2 in east Central Range; and 104 ± 13 mW m -2 in Coastal Range (Table 1).The new Taiwan surface heat flow map is shown in Fig. 4a and an AA' profile was made to exhibit the variations in North-South silica heat flow distribution.Two prominent peaks appear in the Chingshui area with value of 160 mW m -2 and in the Chihpen area with the value of 190 mW m -2 , respectively (Fig. 4b).An interesting distribution pattern in northeastern-southwestern profile of Taiwan as the silica heat flow distribution increases from the Ilan Plain (100 mW m -2 ) to the Chingshui areas (160 mW m -2 ), then decreases to the Lushan area (100 mW m -2 ).As the profile turns to the southeastern-northwestern direction the silica heat flow increases from the Lushan area (100 mW m -2 ) to the Chihpen area and reaches the peak (190 mW m -2 ).However, a lower silica heat flow area, about 80 mW m -2 was found in Eastern Taiwan around the Haulien area, which is the location of a bended subducted slab of the Philippine Sea Plate (Lee et al. 2008).Note: N = number of observations; T = silica temperature in °C; σ is standard deviation; q = silica heat flow in mW m -2 .

Comparison Conventional and Silica Heat Flow
Previous studies have shown some unjustified high heat flow values, greater than 300 mW m -2 distributed in Taiwan (Fig. 5), particularly, in the Central Mountain Range (Lee and Cheng 1986).The heat flow spatial distribution in this study was significantly better than those found for the Western Foothills and Coastal Plain (Fig. 1).The average silica and conventional heat flow values are 110 ± 41 and 125 ± 105 mW m -2 , respectively (Figs. 6a and b).Some conventional data were too high, more than 300 mW m -2 , to be reasonable in the Central Mountain Range.Similar results were found for silica with reasonable conventional heat flow in different physiographic provinces, as shown in Fig. 6.
According to the silica geothemometry method, the standard silica geo-temperature deviation value is usually less than 25°C to represent more accuracy and reliable silica  heat flow value in one block (Swanberg and Morgan 1980).
In this study, most standard deviation values for silica geotemperature were less than 25°C (Table 1).In addition, the silica results and most quality conventional heat flow values were comparable.For example, the silica and conventional heat flow values were 91 and 100 mW m -2 (Huang 1990), in the Kuantzuling area of the Western Foothills, respectively.The Jinlun area in the Central Mountain Range presented silica and conventional heat flow values of 170 and 140 mW m -2 (Gou 2008), respectively.The silica heat flow values are 160 mW m -2 and the conventional values are from 164 -192 mW m -2 in the Chingshui area of the Central Mountain Range (Huang et al. 1986).As a result this silica geothermometry method is adequate for revealing regional heat flow distributions.

The Silica Heat Flow Map Variation and its Interpretation
Two anomalies with higher silica heat flow are located at the Chingshui and Chihpen areas and can be identified in Fig. 4b.Their values are 1.5 -2 times higher than the average continental heat flows in the rest of the world.In general, the surface heat flows are contributed by basement, radiogenic elements, exhumation, burial, fault/fluid, magma, and tectonic components (Blackwell 1978;Bodell and Chapman 1982;Lachenbruch et al. 1985;Armstrong and Chapman 1998).Understanding these components provides information for the regionally thermal regimes needed that reveal the thermal structure.This study discusses what  unique components control the anomalously high heat flow in the Taiwan orogen.
In northeastern Taiwan our results show that a heat source exists beneath the Chingshui area of the Ilan Plain (Fig. 7a).This existing heat source is supported by other methods and results, the helium isotopic ratio of soil gases, seismic data, geomagnetic and magneto telluric data and borehole information (Yu and Tsai 1979;Yang et al. 2005;Lan et al. 2006;Tong et al. 2008).
Few researches have investigated the mechanism that generates this heat anomaly.In fact the Philippine Sea Plate subducts northward under the Eurasian Plate, producing the Ryukyu Trench-Okinawa Trough system.The Okinawa Trough is generally recognized as a back-arc basin in an arccontinent collision system.However, it is not just a simple back-arc basin but rather an embryonic rift zone based on major and trace elements from volcanic rocks in the Southernmost Part of the Okinawa Trough (SPOT), (Chung et al. 2000).It has been shown to extend southwestward from the Okinawa Trough into the Ilan Plain using geodetic monitoring and micro-earthquake results (Liu 1995;Lai et al. 2009).The SPOT extension has been shown to cause the normal faults and magmatic intrusion in the arc collapse/subduction region (Chiang 1976).Thus, the origin of abnormally high silica heat flow in the Chingshui area likely results from the southwestward propagation of hot fluids from the Okinawa Trough into the Ilan Plain (Fig. 7a).
The silica heat flow spatial distribution shows a decreasing distribution from the Chingshui to the Lushan areas southwestwardly and the Ilan Plain northeastwardly (Fig. 6b).Chang (1989) showed a similar radiogenic spatial distribution for the Hsuehshan Range and the Backbone Ridge belts suggesting that both of them should produce the same heat generation near the surface.The decreasing silica heat flow pattern is likely caused by a heat transfer from below.A boundary fault, the Lishan fault, and many subsidiary faults between the Backbone Ridge and the Hsuehshan Range belts have been recognized in this region (Fig. 7a) (Chiang 1976;Lee et al. 1997).A fault can act as a channel that assists heat/ fluid transfer.As a result, the Lishan and other faults in the Ilan Plain likely function as conduits that transfer heat and fluid southwestward and northeastward from Chingshui to the Lushan areas and the Ilan Plain (Fig. 7a).
An aseismic zone was recognized by Lin (2000) based on the seismicity distribution of the Central Mountain Range.With conventional heat flow data numerical modeling, this aseismic zone was attributed to a faster exhumation rate along the Southern Cross-Island Highway (Lin 2000).In southeastern Taiwan a different mechanism, exhumation, was proposed as the major contributor to anomalous surface heat flow.In addition, the upward extrusion of the Central Mountain Range has been proposed as the exhumation process explaining fanning structures with inverted metamorphism (Yui and Chu 2000).Our results showed that the ex-humation is most likely located at the Chihpen area, 55 km south of the Southern Cross-Island Highway.GPS analysis showed strong dilatation within the Chihpen area (Hsu et al. 2009) and the zircon fission track displays the fastest exhumation rate since 0.4 Ma in the same area.These data indicate that buoyant hot materials are exhuming upward beneath the Chihpen area now (Fig. 7b).
The oblique arc-continent collision in Taiwan propagates southwardly, generating different exhumation rates in various tectonic zones.The exhumation rates for Taiwan were calculated using the fission track data under the constant geothermal gradient values (30°C km -1 ) (Lee et al. 2006).The gradient temperature seems underestimated for a reasonable exhumation rate calculation in Southern Taiwan, leading to exhumation rate miscalculation.It seems logical to have more reliable geothermal gradients for estimating the exhumation rate in Southern Taiwan.We, therefore, combined the silica heat flow and fission track dating to estimate the thermal evolution of Southern Taiwan.This is because the heat decays with time and higher heat flow exists in active tectonic zones than in stable areas according to the Fourier's law of heat flow (Carslaw and Jaeger 1986).The silica heat flows are estimated between 120 -190 mW m -2 and increase from the advanced to initial arc-continent collision zones.In addition, the zircon and apatite fission track dating display partial annealing and total reset zones in Southern Taiwan (Fuller et al. 2006).According to the fission track partial and total reset zones, cooling (the tectonic zones of advance and initial arc-continent collision) and accretionary ages (accretionary wedge deformation zone) (Huang et al. 2000;Fuller et al. 2006) can be distinguished.The estimated geothermal gradient values are about 55, 71, and 39°C km -1 in the advanced, initial arc-continent collision, and accretionary wedge deformation tectonic zones using the silica heat flow, respectively.Furthermore, the fission track ages of apatite are 1.5, 0.4 -0.9, and 1.6 -5.5 Ma, and the corresponding silica geotemperature values are about 130, 160, and 100°C at depths of 1.9 km in the advanced, initial arc-continent collision and accretionary wedge deformation tectonic zones, respectively.We combined the geothermal gradient values and apatite fission track ages to estimate the exhumation rate of Southern Taiwan (Tsao 1996).We estimated the exhumation rate in different tectonic zones at about 1.33 mm yr -1 in the advanced arc-continent collision zone, 1.72 -3.87 mm yr -1 in the initial arc-continent collision zone and 0.51 -1.74 mm yr -1 in the accretionary wedge deformation zone, respectively (Fig. 8).These values are much lower than the previously published exhumation rates (Lee et al. 2006).In light of the discrepancy between this study and others, further discussion on the exhumation rate of Taiwan is needed.

CONCLUSIONS
This paper constructed a heat flow map of Taiwan and examined three major controlling factors (1) hot fluids, (2) faults, and (3) exhumation which dominate the surface heat flow in Taiwan.The Okinawa Trough opening produces a heat fluid source beneath the Chingshui area with this heat transferring into the Lushan area and the Ilan Plain along the Lishan and other faults.The thermal anomaly around the Chihpen area, however, results from rapid exhumation in the initial arc-continent collision zone.
Fig. 1.(a) Most of the drill wells are located in the Western Foothills.(b) The spring area distribution is nearly all located in the Central Mountain Range.The left corner is the Philippine Sea plate moving towards the WNW at about 80 mm yr -1 .

Fig. 2 .
Fig. 2. All of the spring concentrations are higher than 6 ppm in Taiwan.

Fig. 4 .
Fig. 4. (a)The silica heat flow contour map (this map does not include the Tatun Volcano Group) in three tectonic zones of Taiwan.Based on the geologic and geophysical characteristics, Taiwan can be divided into three tectonic zones including the arc collapse subduction, the advanced and the arc collapse collision zones(Huang et al. 2000).Triangles denote submarine volcanoes.(b) The regional silica heat flow distribution AA' profile is across Taiwan Island from north to south.In the arc collapse/subduction zone, the peak value is located at the Chingshui area and decreases northeastwardly to the Ilan Plain and southwestwardly to the Lushan area.In the arc-continent collision zone the highest value is located at the Chihpen area and declines to the north.

Fig. 5 .
Fig. 5.This shows that both heat flow values are consistent in the Kuantzuling area, the sedimentary terrain of the Western Foothills and the Jinlun and Chingshui areas, the metamorphic terrain of the Central Mountain Range.
Fig. 6.(a) This histogram presents that the silica heat flow values and their numbers in different geological terrain.(b) Showing the conventional heat flow data-set.Some heat flow values are higher than 300 mW m -2 in the Central Range.

Fig. 7 .
Fig. 7.The schematic map illustrates the abnormal silica heat flow distribution.(a) In the arc collapse/subduction of northern Taiwan, the heat source is beneath the Chingshui area, and the heat transmits to the Lushan area along the Lishan fault.The solid star indicates the hot fluids/heat source beneath the Chingshui area, and the dashed line shows the fault traces.(b) The silica heat flow distribution is consistent with the Southern Taiwan tectonic zones that indicate propagation toward southernmost Taiwan.The solid star shows that the fastest exhumation rate is located in the Chihpen area.The gray arrow is the propagation direction of the increasing silica heat flow and the blue arrow is the propagation direction of the decreasing silica heat flow.

Fig. 8 .
Fig. 8.The exhumation rate in different tectonic zone of the Southern Taiwan.

Table 1 .
Springs of Taiwan and silica heat flow data.