The Khanom Core Complex in Peninsular Thailand is a part of the crystalline basement of Sundaland and plays a key role in our understanding of the evolution of Thailand and SE Asia. The complex comprises ortho- and paragneisses, schists, meta-volcanics, subordinate calcsilicate rocks, and postkinematic granitoids. New petrochronological data reveal that the sedimentation and metamorphism of the paragneiss precursors (Haad Nai Phlao complex, Khao Yoi paragneisses) occurred in the Late Cambrian at the latest. A syn- to postsedimentary andesitic intrusion/extrusion in the Haad Nai Phlao complex at 495 ± 10 Ma defines a minimum age for the former event(s). In the Early Ordovician (477 ± 7 Ma), the Haad Nai Phlao complex and the Khao Yoi paragneisses were intruded by the Khao Dat Fa granite. During the Indosinian orogenic events, the Laem Thong Yang (211 ± 2 Ma) and Haad Nai Phlao (210 ± 2 Ma) granitoid plutons were intruded. Immediately afterward (ca. 208–205 Ma), the first metamorphic overprinting of the Laem Thong Yang granite and the Haad Nai Phlao complex including the Khao Dat Fa granite occurred. A second metamorphic overprinting of all lithological units and the contemporaneous intrusion of the Khao Pret granite followed in the Late Cretaceous and Early Paleogene (ca. 80–68 Ma). The tectonic formation of the core complex took place in the Eocene (<42 Ma), followed by exhumation and regional cooling below ca. 450°C and the latest cooling to ca. 120°C in the Miocene (ca. 20 Ma). The evolutionary data show that the Khanom Core Complex is part of Sibumasu, and its Late Cretaceous-Neogene cooling pattern and exhumation history can be directly related to the northward drift of India.

Thailand is located on the geological entity known as Sundaland, which consists of Gondwana-derived continental terranes that accreted over time to build the present-day mainland of Indochina [1, 2]. Two main continental terranes can be distinguished, Sibumasu in the west and Indochina in the east, along with an interjacent arc, called Sukhothai. Both terranes are crucial to understand the geological evolution of Gondwana, the various Tethys oceanic domains, Sundaland, and Southeast Asia. However, there is no agreement on the nature and exact locations of their boundaries, the characteristics of the basement evolution, or the tectonic models of their amalgamation (References [2-5] and references therein). This problem is accentuated by the scarcity of crystalline basement exposures. The available basement data are limited to three regions of exposure in northern and southeast Thailand and on the Thai peninsula.

The first description of crystalline basement rocks in Thailand was published by Heim and Hirschi [6]. These rocks are typically high- to medium-temperature low-pressure metamorphic and intermediate to acidic plutonic rocks [1, 7, 8]. Often, they are overlain by fossiliferous Phanerozoic sediments [1, 9]. Consequently, the first to assign a Precambrian age to the gneisses was Buravas [9]. In contrast to this interpretation, early radiometric dating in the 1990s of gneisses belonging to this so-called “Precambrian basement” showed that these were more likely formed during metamorphic events in the Middle-Late Triassic, the Late Cretaceous, and the Paleogene (e.g., References [10-12] and references herein). Based on these ages, a model for the tectonic evolution of Thailand was derived, wherein the collision between the Sibumasu terrane and Indochina in the Middle-Late Triassic and the collision of the Indian plate with the Asian plate in the Paleogene are key moments [1, 7]. However, this model was established on the basis of data almost exclusively from Northern Thailand. The continuation toward the Thai peninsula of crystalline basement rocks thought to belong to the Sibumasu terrane seems evident, but only few data are available to support this [5, 8, 13-15].

On Peninsular Thailand, the Khanom Core Complex (KCC) [15, 16] forms a significant part of the crystalline basement and is thus potentially key to our understanding of the plate tectonic evolution of Thailand and SE Asia. The KCC is located approximately 70 km east of Surat Thani on the western coast of the Gulf of Thailand just south of Khanom (Figures 1 and 2).

Traditionally, it is taken as a part of the Khao Luang Batholith of Southern Thailand [15-17]. It comprises orthogneisses, meta-volcanics, paragneisses and schists, subordinate calcsilicate rocks, and postkinematic granitoid intrusives. Currently, only four studies are available that provide data on specific parts of the KCC [12, 15, 18, 19], and no comprehensive model for the evolution of the KCC in the frame of the crustal evolution of Thailand is available.

The aim of our study is to combine petrological, geochemical, and geochronological data to potentially contribute to the assumption that the basement rocks found in Peninsular Thailand have a comparable igneous, metamorphic, and cooling evolution as their northern counterparts. For this reason, in situ SIMS U-Pb zircon, in situ (U+Th)-total Pb EMP monazite dating, and mica 40Ar/39Ar dating were applied to the main lithological units of the KCC, and the results combined with the existing data set.

1.1. The Geological Evolution of Thailand

The Paleozoic evolution of Sundaland is dominated by the successive dispersion, amalgamation, and accretion of Gondwana-derived continental terranes. Three tectonic phases which opened and subsequently consumed three Tethyan oceanic domains, the Paleo-Tethys (Devonian-Triassic), the Meso-Tethys (Early Permian-Late Cretaceous), and the Neo-Tethys (Late Triassic-Late Cretaceous), have been distinguished [2]. The Late Carboniferous to Triassic subduction of the Paleo-Tethys, resulting in the Late Triassic collision of Sibumasu with the Indochina-Cathaysia composite terrane, is known as the Indosinian Orogeny [2, 3, 20]. Between these two continental blocks, the Sukhothai Zone (synonyms being Lincang terrane, Sukhothai terrane, Chantaburi terrane, and Central and Eastern belts of the Malay Peninsula) and the Inthanon Zone (synonyms being Changning-Menglian suture, Chiang Mai-Inthanon suture, and Bentong-Raub suture) can be distinguished [2]. The Sukhothai Zone is interpreted as a volcanic arc developed on rifted fragments of the Indochina terrane that was probably separated in the Carboniferous by a back-arc basin now represented by the Nan-Uttaradit suture [2]. The Inthanon Zone is assumed to be an accretionary complex in which Paleo-Tethyan pelagic cherts, basaltic seamounts with carbonate caps, and minor amounts of siliciclastics have been thrust as nappes along a low-angle fault over sedimentary rocks of Sibumasu [2, 3, 12, 21]. Whereas the exact delineations of these individual tectonic zones is still under debate, their general alignment is presumed to be N-S [1].

The Triassic collision between the Sibumasu and Indochina terranes is the most influential tectonic episode of Sundaland in Thailand [3, 7, 8, 22, 23]. This collision ended by the termination of the Palaeo-Tethys subduction and Late Triassic plutonism related to collisional crustal thickening.

After the Triassic Indosinian Orogeny, a time of tectonic quiescence occurred from the Early Jurassic to the Cretaceous. During this time, shallow marine sediments were deposited in the western part of Thailand while, in the eastern part, continental sedimentation dominated [1]. This interval ended with the collision of India with Eurasia and the subduction of the Meso-Tethys beneath Sundaland, which reached its maximum activity in the Late Cretaceous and was accompanied by granite intrusions and intensified deformation [2, 3, 19, 24, 25]. In Thailand, the India-Eurasia collision led to well-documented lateral extrusion along major strike-slip faults [1, 26]. Four prominent strike-slip fault systems, including a number of splays, play an important part in this Cenozoic lateral extrusion of SE Asia (Figure 1 inset). These are the NW-SE striking sinistral Mae Ping and Three Pagodas shear zones in the west [8, 27, 28] and the NE-SW striking dextral Ranong and Khlong Marui shear zones in the south [13, 29]. A second dominant Cenozoic structural feature of the geology of Thailand is the formation of pull-apart basins. For some authors, the link between the strike-slip fault zones and these basins is obvious [1], while Morley et al. [30] postulated that the strike-slip fault zones were active prior to the formation of the basin, and consequently, another mechanism must be responsible for basin formation. They suggested that subduction-rollback was the major process involved in the formation of these basins, and strike-slip faulting played only a minor part. In addition, subduction-rollback could have played an important role in the genesis of the metamorphic core complexes in Thailand (Doi Inthanon, Khanom, Klaeng [30]). Data from Peninsular Thailand shed light on the transition from an orogenic late-stage situation into the development of continental rifting systems [18]. In this model, the evolution of rifting is controlled by the presence and migration of a free edge (path of India) and the existence of a strong backbone (Malay Peninsula) separating two basin geometries.

Thermochronological data from the metamorphic core complexes suggest the same succession of events (HT metamorphic event, postkinematic intrusion, exhumation, and rifting) but with the ages of the postkinematic intrusions and the subsequent cooling below ca. 400°C systematically getting younger northward. This age migration is thought to be linked to the northward motion of the Indian Plate [19].

1.2. The Geology of the KCC

The ca. 225 km2 KCC is situated in the northern part of Changwat Nakhon Si Thammarat, southern Thailand, just south of the town of Khanom (Figures 1 and 2).

The KCC was first described in detail by Kosuwan and Charusiri [16, 31]. Typically, five lithological units are distinguished (Figure 2) [12, 15, 16, 18, 32]:

  1. Haad Nai Phlao (HNP) gneiss — sillimanite–muscovite–biotite gneisses with a subordinate mineral assemblage of biotite + sillimanite + garnet and partly amphibole-bearing, occurring in alternating layers with marked differences in grain size. The finer-grained layers are equigranular, and the coarse-grained layers are more porphyroclastic. The contact between the two varieties is sharp and partly discordant. The gneisses also host calc-silicate enclaves. According to previous investigations, the gneisses are interpreted to be meta-quartz arenites to meta-arkoses.

  2. Khao Yoi gneiss — a rock assemblage of muscovite ± garnet gneisses and schists intercalated with quartzites, calc-silicates, and marbles. These gneisses have been previously deemed to be of sedimentary origin.

  3. Laem Thong Yang (LTY) gneiss — a porphyroclastic and periodically tourmaline-bearing biotite gneiss with augen-texture which is locally cut by a finer-grained biotite gneiss. Due to its texture, the gneiss was interpreted to be an orthogneiss.

  4. Khao Dat Fa granite — a garnet-bearing two-mica orthogneiss. Importantly, the garnet is Mn-rich in euhedral cores and overgrown by Ca-richer, irregular rims. Plagioclase shows concentric, euhedral oscillatory zoning and remnant textures suggestive of a cumulative origin.

  5. Khao Pret granite — structurally the youngest unit, a fine-grained biotite granite cross-cutting the former lithological units. A contact metamorphic aureole is developed in the Khao Yoi gneisses at the western contact zone.

Metamorphic conditions in the KCC reached from upper greenschist- to upper amphibolite-facies [16, 33]. The predominant regional foliation trends NW-SE dipping 50°–60° NE [16, 18]. Field evidence reveals at least three tectonic events, the superposition of which is reflected in a structural style characterized by multidirectional foliation and fault patterns (Figures 1 and 2).

The contact of the KCC with the surrounding Paleozoic sedimentary units is covered by Quaternary deposits and dense vegetation. In so far as decipherable, the KCC is bounded by detachment faults in the W and E. In the N and S, it is located between two major strike-slip fault systems, namely the Khanom fault in the N and Sichon fault in the S (Figures 1 and 2) [34]. The detachments faults are thought to be rotated, having been initially high-angle normal faults related to the early collapse of the continental crust at the onset of extension during the Eocene [30].

1.3. Published Age Data from the KCC

According to Kosuwan [16], the relative ages and most of the contact relations between the five major lithological units of the KCC are enigmatic. Only the Khao Pret granite can definitely be identified as the youngest lithological unit of the KCC. Published age data are presented in online supplementary Table S1.

Following the traditional view, the KCC is of Lower Paleozoic or even Precambrian age [16]. The first attempts to date the KCC rocks were made by Hansen & Wemmer [12]. Multigrain zircon U-Pb analyses were conducted on an amphibole-bearing HNP (HNP-a) gneiss from Khao Phra. The data yield a badly defined upper intercept age of 452–84/+93 Ma (Late Ordovician) and a lower intercept age of 45–18/+12 Ma (Lutetian, Eocene). While the Eocene age was interpreted as a strong thermal pulse or high-temperature metamorphism without significant deformation, the meaning of the Ordovician age remained uncertain.

A following study by Hansen et al. [32] involving in situ zircon U-Pb SIMS dating on the same sample gave a slightly older uppermost Paleocene age of 56.45 ± 0.72 Ma. The Ordovician age, however, could not be reproduced.

Further geochronological work was carried out by Kawakami et al. [15] using LA-ICP-MS U-Pb zircon and CHIME monazite dating. CHIME ages obtained of the LTY gneiss are 263 ± 36 Ma (monazite cores) and 66 ± 9 Ma (monazite rims). As the CHIME regressions show negative intercepts, the ages are deemed to be unreliable as is the monazite CHIME age of 72 ± 13 Ma found for an HNP gneiss sample. U-Pb zircon dating yielded a weighted average 206Pb/238U age of 477 ± 7 Ma for the Khao Dat Fa granite gneiss. The authors interpret this Ordovician age as an intrusion age and take this as evidence that some of the basement rocks of Sibumasu were formed during the Pan-African Orogeny (sic!). A monazite core CHIME age of 205 ± 5 Ma is interpreted as representing Indosinian metamorphism in the Triassic. A CHIME monazite rim age of 36 ± 24 Ma for the same sample is also reported, but no definite interpretation is given. The Late Cretaceous concordia age (67.5 ± 1.3 Ma) for the Khao Pret granite is interpreted as timing zircon crystallization from the granitic melt, the intrusion of which caused the sillimanite-grade contact metamorphism in the hosting metapelites and gneisses. Sautter et al. [19] presented zircon (ZFT) and apatite (AFT) fission track ages of the HNP gneiss (ZFT: 33.5 ± 1.4 Ma; AFT: 24.6 ± 1.1 Ma) and the Khao Pret granite (ZFT: 32.9 ± 1.6 Ma; AFT: 17.9 ± 1.1 Ma). In addition, modeling of the thermal history of confined fission track lengths on the sample from the HNP gneiss revealed fast cooling from 36 to 26 Ma (23°C/Ma), interpreted as being related to rapid tectonically controlled exhumation, followed by lower to moderate cooling from 27 to 22 Ma (12°C/Ma). Later, cooling at rates lower than 2°C/Ma is considered to be solely related to erosional unroofing.

The Khao Yoi schists have not been dated yet.

Full details of the analytical methods employed by us are given in the online Supplementary Material 1.

2.1. Whole-Rock Geochemistry

Whole-rock analyses were made by X-ray fluorescence for major and large ion lithophile trace elements and by inductively-coupled plasma mass spectrometry for high field strength and rare earth elements (REE) at the Bureau Veritas (Canada). REEs are normalized to the C1 chondrite values of McDonough and Sun [35].

2.2. Zircon Separation and Cathodoluminescence Imaging

Zircon was separated from three whole rock samples (TH1608, TH1611, and TH1612), each 10–15 kg, using conventional methods. Clear, crack-, and inclusion-free zircon grains were handpicked under a binocular microscope and mounted on standard 1-inch epoxy resin discs. The chemical zonation of the zircon crystals was revealed by CL imaging using a FET Philips 30 electron microscope, an accelerating voltage of 15 kV, and a beam current of 20 nA at the Institute of Earth Sciences, University of Silesia in Katowice, Poland. The most suitable locations of the spots for analysis were selected, and after the isotopic analyses, the zircon grains were again inspected in order to define the precise spot location with respect to the internal microstructures.

2.3. Zircon SIMS U-Pb Dating

In situ U-Th-Pb SIMS measurements were conducted using a Cameca IMS-1280 HR SIMS at the Institute of Geology and Geophysics, Chinese Academy of Sciences (Beijing, China) during two sessions. A hybrid dynamic multicollector U-Pb dating technique was used taking advantage of both the static multicollector mode and peak-hopping monocollector mode and using oxygen flooding. For the final age calculations, the offline version of IsoplotR was used [36]. On Tera–Wasserburg concordia plots, the spot’s uncertainty ellipses are at the two-sigma level and color coded according to the respective Th/U ratio.

2.4. Monazite EMP (Th+U)-Total Pb Dating and Mineral Chemistry

EMP monazite (Th+U)-total Pb dating was performed using a CAMECA SX-100 microprobe at the Department of Electron Microanalysis at the State Geological Institute of Dionýz Štúr in Bratislava, Slovakia. (Th+U)-total Pb spot and CHIME ages were calculated using the mathematical expressions of Suzuki et al. [37].

The monazite mineral chemistry was determined using the Cameca SX Five FE EMPA at the Department of Lithospheric Research, University of Vienna, Austria. The concentrations are reported either as weight-% or weight-ppm. Spots used for age calculation and for chemical characterization, especially for the monazite chemistry determinations, were set side to side whenever possible.

2.5. Mica40Ar/39Ar Dating

Mica 40Ar/39Ar geochronology was conducted at the Queen’s University Ar-Ar geochronology laboratory (Kingston, Ontario, Canada). Micaceous mineral separates were identified and purified by hand picking. Each separate contained >10 to ~40 mica crystals with a size of ca. 1–2 mm. Final ages were calculated using the online version of IsoplotR [36].

3.1. Investigated Samples

For the sake of consistency with published descriptions, the investigated samples are classified following the naming scheme and petrographical mapping of Kosuwan [16]. Investigated samples are:

  1. LTY gneiss (TH1608, KN30: 09°5'19.8"N/099°54'36.12"E; TH1610: 09°5'26.1"N/099°53'38.1"E; KN71: 09°6'31.98"N/099°54'5.90"E).

  2. HNP-a gneiss (TH1611: 09°8'22.7"N/099°52'33.4"E).

  3. Fine-grained HNP gneiss (KN19: 09°8'22.7"N/099°52'33.4"E).

  4. Coarse-grained HNP (HNP-cg) gneiss (TH1612, KN68: 09°8'21.1"N/099°52'35.9"E; KN18: 09°8'11.25"N/099°50'24.04"E; KN54: 09°7'56.40"N/099°50'26.06"E).

  5. Tong Yi leucocratic gneiss (TH1613: 09°6'49.9"N/099°53'55.3"E).

  6. Khao Pret granite (KN21: 09°6'43.28"N/099°52'15.79"E).

3.2. Petrography

Field photographs of the different lithologies are given in Figure 3 and thin-section photographs in Figure 4.

3.2.1. LTY Gneiss (Samples TH1608, TH1610, KN30, and KN71)

The LTY is a generally coarse-grained, partly slightly anatectic augen-gneiss occurring in the south-eastern part of the KCC at Thong Yi beach, Hin Lat waterfall, Khao Phlai Dam, and Phlai Dam beach (Figure 2). The average mineral assemblage of the LTY is K-feldspar (40–50 vol%, Or = 90), quartz (25–30 vol%), plagioclase (15–20 vol%, An = 15–30), and biotite (5–10 vol%; Figures 3(a), 3(e)). The porphyritic texture is dominated by K-feldspar and sparse plagioclase augen, both of which are up to a few centimeters in size. The matrix consists of fine-grained quartz, plagioclase, and biotite, where the orientation of biotite grains defines the foliation. Apart from some big K-feldspar clasts and biotite, mineral grains are anhedral. K-feldspars often show Karlsbad and microcline twinning, perthitic exsolutions, and remnants of the former anatectic melt along the crystal rims (Figure 4(b)). Variously sized inclusions of quartz, plagioclase, biotite, apatite, monazite, and zircon occur within the K-feldspars. Plagioclase is characterized by polysynthetic twinning, relict magmatic zonation, and, locally, intense sericite alteration. At the contact between K-feldspar and plagioclase, myrmekite replacing K-feldspars is sometimes present. Quartz grains show dynamic recrystallization, grain boundary migration, and, in places, a weak shape preferred orientation. Biotite, mostly aligned along the feldspar augen, forms characteristic clusters and shows kink-bands and pleochroic halos around zircon and monazite inclusions. Minor constituents are white mica, chlorite, epidote, zoisite, sericite, ilmenite, and in places tourmaline. Accessories are zircon, monazite, and xenotime, which often form a characteristic paragenesis, apatite, and Fe-oxides. Most of the zircon and monazite are associated with biotite clusters.

3.2.2. HNP Gneiss

Outcrops of the HNP can be found in the central- and more north-eastern part of the KCC around Nai Phlao beach and Samet Chun waterfall (Figure 2). It is composed of two varieties, namely coarse- and fine-grained biotite-gneisses. The structural, textural, and age relations between the two varieties are uncertain. On the one hand, they occur in alternating successions. Occasionally, a gradual transition between the two gneisses can be observed. On the other hand, some contacts are evidently discordant (Figure 3(d)). The alternating layers usually have a thickness from a few centimeters to several decimeters. Single layers of some meters were also observed. Furthermore, intercalations of fine-grained, strongly laminated white-greenish calcsilicates occur.

3.2.3. HNP-a Gneiss (Samples TH1611 and KN19)

The mineral assemblage of the HNP-a is quartz (20–30 vol%), plagioclase (20–30 vol%, An = 30–40), K-feldspar (15–20 vol%, Or = 95), biotite (5–10 vol%), hastingsitic amphibole (5–10 vol%), white mica (1–5 vol%), and clinozoisite (<5 vol%). Quartz shows a strong undulous extinction and abundant dynamic recrystallization. Plagioclase is typically polysynthetic twinned and sometimes shows kink-bands and undulous extinction. Weak sericitic alteration is also observed. Minor constituents are sericite, allanite, titanite, chlorite, zircon (Figures 3(b), 3(d)) and rare sillimanite and garnet. It is worth noting that, in contrast to the other samples, no monazite (and xenotime) could be found, and the abundant zircon is of comparably small grain size (<50 µm).

Calcsilicate lenses can be found intercalated with the HNP-a. These are aligned concordantly, decimeters in length, and show a ca. centimeter thick compositional layering that varies systematically from center to rim. While the center predominantly consists of diopsidic pyroxene and minor anorthite-rich plagioclase, the rim consists of quartz, albite-rich plagioclase, K-feldspar, and minor hornblende, titanite, and biotite. Quartz is dynamically recrystallized showing bulging and recrystallized subgrains. Plagioclase shows undulous extinction, weak sericitic alteration, and polysynthetic, sometimes kinked twin lamellae. Myrmekite can be observed where K-feldspar is in contact with plagioclase. Interestingly, titanite is euhedral and can reach a few millimeters in size. The other constituents are characteristically far smaller in grain size (50–500 µm) and anhedral. Accessories in the calcsilicates are apatite and zircon.

3.2.4. HNP-Cg Gneiss (Samples TH1612, KN18, KN54, and KN68)

The HNP-cg consists of K-feldspar (35–45 vol%, Or = 85–92), quartz (25–30 vol%), plagioclase (15–20 vol%, An = 19–25), biotite (5–10 vol%), and white mica (<5%). Minor constituents are apatite, sericite, chlorite, zoisite, and ilmenite. Accessories are zircon, monazite, and xenotime (Figures 3(c) and 4(e)). Thus, it is comparable in composition to the LTY with, on average, slightly more abundant biotite and white mica. Although the K-feldspar blasts are generally smaller and less euhedral than in the LTY, a porphyritic texture is noticeable. Minerals are anhedral, apart from the mica grains which define the foliation. Feldspars often show sericitization, and plagioclase sometimes shows polysynthetic twinning (Figure 4). Quartz is characterized by dynamic recrystallization. Karlsbad twinning of K-feldspars is not as dominant as in LTY. Biotite shows the typical brownish-greenish pleochroism and pleochroic halos around zircon and monazite grains. Monazite and zircon are commonly associated with biotite clusters.

3.2.5. TY-Lg Orthogneiss (Sample TH1613)

A leucocratic and nearly biotite-free orthogneiss (Figures 3(e) and 3(f)), not described by Kosuwan [16], was found at Thong Yi beach. Its composition is K-feldspar (30–40 vol%), quartz (25–30 vol%), plagioclase (20–25 vol%), and white mica (5–10 vol%). Biotite is extremely rare. K-feldspars often show Karlsbad twinning and sometimes relicts of normal magmatic zoning. Quartz and plagioclase myrmekites as well as quartz and white mica symplectites are abundant. Accessory minerals are apatite, monazite, zircon, and xenotime. Notably, monazite, zircon, and xenotime are rare compared with other lithologies. This rock only shows one stage of metamorphic overprinting, in contrast to the other gneisses [33]. The TY-lg clearly crosscuts the LTY (Figure 3(e)). The contact between the two gneisses is definitely magmatic, as is shown by the abundant occurrence of LTY enclaves and assimilated Kfsp-megacrysts and igneous apophyses of the TY-lg in the LTY (Figures 3(e) and 3(f)).

3.3. Whole Rock Geochemistry

The analytical data is given in online supplementary Table S2. In the Streckeisen QAP diagram, the samples from the LTY (TH1608 and TH1610) HNP-cg (TH1612 and TH1613) plot in the granite field, whereas the HNP-a sample (TH1611) shows a less felsic composition on the border of quartz monzonite and quartz monzo-diorite (Figure 5(a)).

All samples, with the exception of TH1611, which is metaluminous, show a peraluminous geochemical character (Figure 5(c)). In the tectonic discrimination diagram [38], TH1613 plots in the middle of the volcanic arc and syncollisional granite field, whereas the remaining samples plot on the border of this field with the within-plate granite field (Figure 5(d)).

Samples TH1608, TH1610, TH1611, and TH1612 contain on average ca. 170 weight-ppm REEs and have virtually identical REE patterns (Figure 5(b)) with an average (La/Lu)N = 9.17, moderately steep LREE-patterns (LaN = 180) and (La/Nd)N = 0.74. HREE patterns are flat with (Gd/Lu)N = 1.77. Eu anomalies are around 0.25. TH1611 is comparably LREE enriched, MREE depleted, and has the least prominent Eu anomaly. TH1613 in contrast is strongly depleted in all REEs (ΣREE = 88 weight-ppm) with a flat pattern (La/Lu)N = 3.84, (La/Nd)N = 0.05, (Gd/Lu)N = 0.34, and large Eu anomaly (Eu/Eu* = 0.06).

3.4. Monazite Chemistry

Monazite mineral chemistry analyses were made directly in thin sections from one LTY (TH1608, 35 points) and one HNP-cg (TH1612, 19 points). The data are presented in online supplementary Table S3 and S4. The monazite crystals contain on average 29 wt% P2O5, ca. 55 wt% LREE2O3, and ca. 4 wt% HREE2O3. Trace elements are ca. 2 wt%, and UO2 + ThO2 is ca. 9 wt%. All data points are plotted in a small cluster within the LREE2O3 - HREE2O3 - Y2O3 and LREE2O3 - HREE2O3+Y2O3 - ThO2+UO2 ternary diagrams (Figure 6) in the pure monazite apex.

They show a pseudolinear trend toward more HREE-Y-rich, that is, higher xenotime solid solution amounts and actinides-richer compositions (Figure 6(a)). No immediate trends of the chemical groups with spot locations are evident. Also, Eu anomalies do not show any significant inter- or intragroup correlation.

3.5. Monazite (Th+U)-Total Pb EMP Age Data

The monazite (Th + U)-total Pb EMP age data are given in online supplementary Table S3 and presented in Figures 7 and 8).

3.5.1. LTY Gneiss (TH1608)

Analyzed monazite crystals are between 20 and 200 µm in size, roundish to slightly elongated, and show strongly resorbed cores and rims. The internal zonation is in parts oscillatory and concentric with darker core domains and brighter rims, in parts patchy without systematic arrangement (Figure 7).

Twenty spots on eight crystals were analyzed (online supplementary Table S3). U contents range from 0.093 wt% to 0.636 wt% with an average of 0.344 wt%. Th contents are 1.23–12.26 wt% with an average of 7.65 wt%. Pb ranges from 0.010 wt% to 0.142 wt% with an average of 0.061 wt%. The resulting apparent spot ages range from 26.7 ± 15.0 to 209.5 ± 15.2 Ma (Figure 8(b)), and all ages fulfill the 0.9 < (Ca + Si)/(Th + U + Pb) < 1.1 criterion. Spot ages show a bimodal distribution with three spots defining a weighted mean age of 207.7 ± 15.2 Ma, whereas fifteen spots show a weighted mean age of 77.9 ± 4.2 Ma. The remaining two spots are < 51 Ma. The corresponding forced CHIME ages are 213.8 ± 7.2 and 79.9 ± 3.7 Ma, respectively (Figure 8(a)). Within the associated uncertainties, the CHIME ages are identical to the weighted mean spot ages. No correlation between spot ages, mineral REE and trace element contents, and spot location is evident (Figure 7).

3.5.2. HNP-cg Gneiss (TH1612)

Monazite crystals are between 20 and 100 µm in size, roundish to slightly elongated, and show, in parts, slightly resorbed rims. The internal zonation is in parts oscillatory and concentric and in parts patchy without systematic arrangement. Crystal cores tend to be brighter in the BSE emission (Figure 7).

Twenty-two spots on seven crystals were analyzed (online supplementary Table S3). U contents range from 0.069 wt% to 1.227 wt% with an average of 0.669 wt%. Th contents are 1.12–12.53 wt% with an average of 8.01 wt%. Pb ranges from 0.009 wt% to 0.085 wt% with an average of 0.039 wt%. Apparent spot ages range from 60.6 ± 11.8 to 219.5 ± 61.7 Ma (Figure 8(d)). All spot ages fulfill the 0.9 < (Ca + Si)/(Th + U + Pb) < 1.1 criterion. Spot TH1612C-1_mnz7/1 is aberrant in showing the lowest Y content (0.49 wt%) in combination with the highest spot age (219.5 ± 61.7 Ma). The remaining twenty-one spots show spot ages <140 Ma. These show a bimodal Y content distribution with a positive correlation of Y with U contents. Low Y spots (eleven spots, 1.23 ± 0.15 wt% Y) show an average spot age of 99.4 ± 28.0 Ma. Five spots out of these eleven spots define an age plateau at 90.9 ± 11.8 Ma (Figure 8(d)). The CHIME age is 71.2 ± 8.2 Ma with a high initial Pb content (Figure 8(c)). High Y spots (ten spots, 1.81 ± 0.08 wt% Y) show an average spot age of 65.6 ± 7.2 Ma (Figure 8(d)) and a CHIME age of 67.9 ± 7.2 Ma (Figure 8(c)). Within the associated uncertainties, the CHIME ages are identical to the mean spot ages. No correlation between spot ages, mineral REE and trace element contents, and spot location is evident (Figure 7).

3.6. Zircon U-Pb SIMS Age Data

The zircon U-Pb SIMS data are given in online supplementary Table S5 and presented in Figures 9 and 10.

3.6.1. LTY Gneiss (TH1608)

Zircon crystals from the LTY are commonly euhedral and show long prismatic habits (length ca. 100–600 µm) with aspect ratios between 3:1 and 5:1. CL images reveal well-discernible concentric and oscillatory zonations. Frequently, a bright CL-zoned core is surrounded by a CL dark layer, which is in turn surrounded by medium bright CL zonation (Figure 9).

Twenty-four spots on twelve crystals were analyzed (online supplementary Table S2). U concentrations range from 59 to 4373 wt-ppm, Th from 1.0 to 20.2 wt-ppm, and Pb from 3.2 to 158.0 wt-ppm. Th/U values range from 0.002 to 0.023 with a mean of 0.011 ± 0.007 (2 SD). All spots are concordant within the associated uncertainties (Figure 10(a)) and yield 206Pb/238U dates ranging from 199.2 ± 3.1 to 1011.1 ± 14.2 Ma. Twenty-one out of twenty-two spots form a cluster at 238U/206Pb > 25 that defines a concordia age of 210.1 ± 1.3 Ma (Figure 10(b)). The remaining two spots (036b, 077a) form singular occurrences at 238U/206Pb < 25. The cluster at 210.1 Ma can be separated into two subclusters based on the spot Th/U ratio. The low Th/U cluster is at Th/U = 0.005 ± 0.004 (2 SD, twelve spots), and the high Th/U cluster is at Th/U = 0.017 ± 0.005 (2 SD, ten spots). The two clusters do not overlap at the two-sigma confidence level, the division being at Th/U = 0.012. The spots with Th/U > 0.012 give a concordia age of 207.5 ± 2.0 Ma (Figure 10(c)), whereas those with Th/U < 0.010 give a concordia age of 212.2 ± 1.7 Ma (Figure 10(d)).

3.6.2. HNP-cg Gneiss (TH1612)

Zircon crystals range in length between 50 and 300 µm with a mostly euhedral shape. They show aspect ratios between 1:1 and 5:1. A complex oscillatory zonation is commonly well established and typically shows a relatively CL bright core with a darker mantle and a medium bright outer zonation (Figure 9). Otherwise, CL dark and bright patchy domains are present.

Twelve spots on four crystals were analyzed (online supplementary Table S2). U concentrations range from 177 to 1477 wt-ppm, Th from 1.3 to 7.4 wt-ppm, and Pb from 6.2 to 51.4 wt-ppm. Th/U values range from 0.002 to 0.018 with a mean of 0.007 ± 0.005 (2 SD). All spots are concordant within the associated uncertainties (n = 12; Figure 10(h)) and yield 206Pb/238U dates ranging from 202.5 ± 3.0 to 215.5 ± 3.2 Ma. Eleven spots form a cluster that defines a concordia age of 209.7 ± 2.1 Ma (Figure 10(h)). The remaining spot (112c) is omitted because of the large uncertainty on 207Pb/206Pb.

3.6.3. HNP-a Gneiss (TH1611)

Zircon crystals are ca. 50–200 µm in length with an aspect ratio of 1:1 or 1:2 and show variable internal structures (Figure 9). In some, a CL patchy core with an oscillatory zoned rim can be seen. Smaller equant zircon crystals tend to be more patchy, while longer prismatic zircon shows more pronounced oscillatory zonation.

Seventeen spots on ten crystals were analyzed (online supplementary Table S5). U concentrations range from 36 to 6742 wt-ppm, Th from 0.6 to 30.1 wt-ppm, and Pb from 9.3 to 244.4 wt-ppm. Th/U values range from 0.000 to 0.046 with a mean of 0.022 ± 0.016 (2 SD). All spots are concordant within the associated uncertainties (n = 17; Figure 10(f)) and yield 206Pb/238U dates ranging from 52.5 ± 1.0 to 497.3 ± 7.8 Ma. Three spots (020a, 025a, and 025b) from CL bright oscillatory-zoned mantle domains of prismatic crystals define a concordia age of 494.5 ± 9.6 Ma (Figure 10(g)). A second cluster is formed by three spots (011a, 023a, and 023b) at 52.8 ± 1.0 Ma (Figure 10(h)) from CL gray homogeneous zircon domains and comparably small stubby crystals. Five spots from homogeneous CL dark gray domains (018a, 019a, 019b, 027a, and 027b) are grouped together, although they do not form a statistically valid cluster. They show a weighted mean date of 63.9 ± 2.6 Ma. The remaining six spots from nonspecific to strongly altered zircon domains with no characteristic CL emission provide dates intermediate to ca. 400 and ca. 65 Ma.

3.7. Mica40Ar/39Ar Data

The mica 40Ar/39Ar data are given in online supplementary Table S6. A total of seventeen hand-picket biotite multigrain fractions (0.30-0.25 mm) from six samples and six hand-picket muscovite multigrain fractions (0.30-0.25 mm) from five samples were analyzed.

3.7.1. LTY Gneiss (KN30 and KN71)

The biotite (KN71) total fusion 40Ar/39Ar age is 35.3 ± 2.3 Ma (40Ar/36Ar = 1117 ± 281). Two muscovite fractions (KN30-2 and KN71) have 40Ar/39Ar ages of 40.2 ± 2.7 (40Ar/36Ar = 819 ± 44) and 40.2 ± 2.3 Ma (40Ar/36Ar = 620 ± 35), respectively. The weighted mean age is 40.2 ± 2.4 Ma.

3.7.2. HNP-cg Gneiss (KN18, KN54, and KN68)

The KN18-3 single fraction total fusion 40Ar/39Ar age is 36.5 ± 1.7 Ma (40Ar/36Ar = 1363). Three biotite fractions of sample KN54 show total fusion 40Ar/39Ar ages between 34.0 ± 4.2 (KN54-1a; 40Ar/36Ar = 219) and 38.1 ± 2.3 Ma (KN54-1b; 40Ar/36Ar = 1877). The weighted mean average 40Ar/39Ar age of the three fractions is 36.4 ± 3.5 Ma. Two biotite fractions of sample KN68 have total fusion 40Ar/39Ar ages of 38.1 ± 10.0 (KN68-1a; 40Ar/36Ar = 2112) and 43.0 ± 8.3 Ma (KN68-1b; 40Ar/36Ar = 3957). The weighted mean average 39Ar/40Ar age is 41.1 ± 12.8 Ma. The overall weighted mean 39Ar/40Ar age of all six biotite fractions is 36.6 ± 2.3 Ma.

Two muscovite fractions (KN18-2 and KN68) have 40Ar/39Ar ages of 41.9 ± 3.7 (40Ar/36Ar = 2745) and 41.7 ± 7.7 Ma (40Ar/36Ar = 3640), respectively. The weighted mean age is 41.9 ± 6.5 Ma.

3.7.3. Fine-Grained HNP Gneiss (KN19)

The biotite sample KN19-1 total fusion 39Ar/40Ar age is 23.8 ± 8.8 Ma (40Ar/36Ar = 246 ± 119).

3.7.4. Khao Pret Granite (KN21)

Five biotite fractions show total fusion ages between 29.6 ± 12.4 (KN21-6b; 40Ar/36Ar = 497 ± 411) and 46.2 ± 7.5 Ma (KN21-1b; 40Ar/36Ar = 1325 ± 991), respectively, with no systematic relation between 40Ar/36Ar and 39Ar/40Ar. The weighted mean average 39Ar/40Ar age of all five biotite fractions is 40.7 ± 1.4. With sample KN21-1, a five-step heating experiment was made. Individual step 39Ar/40Ar ages range from 38.2 ± 2.5 to 41.9 ± 2.3 Ma. With the exception of the first low-temperature step, 40Ar/36Ar ratios are ≥ 1994 ± 1612. The weighted mean 39Ar/40Ar age of the stepwise heating experiment is 40.7 ± 1.4 Ma. The total weighted mean average 39Ar/40Ar age of all ten biotite fractions is 40.7 ± 1.3.

Two muscovite fractions (KN21-1 and KN21-6) have 39Ar/40Ar ages of 39.8 ± 3.7 (40Ar/36Ar = 782 ± 51) and 34.3 ± 3.8 Ma (40Ar/36Ar = 363 ± 38), respectively. The weighted mean age is 37.1 ± 5.2 Ma.

4.1. LTY Gneiss

Petrographically, the LTY is a homogeneous coarse-grained biotite augen-gneiss that can be interpreted as a metamorphic igneous body. Compared with the geochemical data presented by Kosuwan [16], our sample TH1608 is relatively SiO2 poor, whereas sample TH1610 from a more internal part of the LTY has an intermediate SiO2 content and is CaO richer. The REEs of both samples are identical and not characteristically different from the samples of the HNP varieties, with moderately enriched LREEs, flat MREE-HREE patterns, and a small negative Eu anomaly. Accordingly, the LTY is a former granite with a peraluminous chemistry and a possible volcanic arc to syncollision granite signature.

The interpretation is supported by the zircon U-Pb age data and the corresponding zircon zonation patterns. All the LTY zircon crystals show a well-developed oscillatory chemical zonation which is characteristic for growth from a melt. Furthermore, cores are rare. The SIMS spots do not show any correlation between apparent ages and position within the zircon crystals. Thus, we interpret the cluster of twenty-one (out of twenty-four) spots defining a Triassic concordia age of 210.1 ± 1.3 Ma as reflecting zircon growth in the crystallizing LTY protolith. Based on the Th/U ratios, the cluster can be divided into two subclusters: one with a concordia age of 211.2 ± 1.7 Ma (Th/U < 0.01) and a second of 207.5 ± 2.0 Ma (Th/U > 0.01). These two concordia ages are not identical within the associated uncertainties and might represent two zircon growth events during the formation of the LTY protolith. Alternatively, the older age (ca. 211 Ma) might be dating igneous zircon growth during LTY protolith formation and the younger one (ca. 208 Ma) a metamorphic-hydrothermal overprinting event shortly after the LTY protolith formation. Zircon zonation and the Th/U systematics do not directly favor the latter interpretation, but the abundance of literature data [1, 3, 8, 17] providing evidence for an Indosinian (Middle to Late Triassic) metamorphic event in Thai basement rocks should not be ignored in this case. We thus favor the interpretation of an igneous formation of the LTY at ca. 211 Ma and a first metamorphic overprint at ca. 208 Ma. The two erratic SIMS spots with apparent ages >250 Ma are from zircon core domains and do not add any additional information.

The TH1608 monazite crystals show anhedral habits, the indication of a skeletal structure, and patchy zoning—all features normally ascribed to metamorphic monazite and monazite consumption and recrystallization [39, 40]. Three monazite (U + Th)-total Pb ages (out of twenty) giving the weighted mean spot age of 207.7 ± 15.2 Ma, also point to an Indosinian event. The CHIME age of these three spots is 213.8 ± 7.2 Ma, identical within the uncertainties with the mean spot age. These monazite ages are identical within the uncertainties with the zircon concordia ages and are most reasonably interpreted as representing growth ages. However, the younger zircon age cluster at 207.5 ± 2.0 Ma, which we tentatively interpret as the age of a metamorphic event overprinting the somewhat older igneous LTY protolith (211 Ma), is within the uncertainty of the age provided by the three Triassic monazite spots (207.7 ± 15.2 Ma). Alternatively, the Triassic monazite ages can also be interpreted as dating a metamorphic overprinting event, in as much as the monazites do not show any specific mineral chemistry that would distinguish these older domains from the majority of the far younger metamorphic ages. But in view of the associated uncertainties, the problematic zircon Th/U values, and the relative scarcity of the monazite ages, any final attribution has to remain open. By far, the dominant age group found in the TH1608 monazite shows Cretaceous ages (ca. 95–67 Ma). Fifteen spots form a cluster and define a weighted mean age of 77.9 ± 4.2 Ma. The corresponding CHIME age is identical within uncertainties at 79.9 ± 3.7 Ma. In view of the monazite habit and zonation pattern, this Late Cretaceous age of ca.78 Ma is interpreted as dating a metamorphic overprinting event leading to monazite breakdown and recrystallization. The remaining two spots ages <51 Ma appear to have little geological significance. A monazite CHIME date of 66 ± 9 Ma, interpreted as being somewhat unreliable due to a negative initial Pb content [15], supports our interpretation of the presence of a Late Cretaceous metamorphism in the LTY.

The mica 40Ar/39Ar ages are interpreted as dating the cooling of the LTY after the Late Cretaceous metamorphic event dated by some of the monazites. In the absence of rigorous petrochronological data defining the closure of the 40K/40Ar system in the investigated muscovite and biotite, we interpret the muscovite 40Ar/39Ar age of 40.2 ± 3.4 Ma as cooling of the LTY sample below ca. 450°C, whereas the biotite 40Ar/39Ar age of 35.3 ± 2.3 Ma defines the cooling below ca. 300°C in the Eocene. This indicates an Eocene cooling gradient of ca. 31°C/My for this temperature interval.

4.2. HNP Gneiss

The HNP sensu lato is far more heterogeneous in lithology and fabrics than the LTY; strong lithological layering and intercalations comprising a number of different rock types are common features. Simplifying, the HNP can be taken as being composed of three different varieties, namely coarse-grained gneisses, fine-grained biotite-gneisses, and amphibole-bearing biotite-gneisses. Importantly, weakly strained contacts between the varieties are, in places, discordant, possibly suggesting the existence of former igneous contacts. From our fieldwork, we conclude that the amphibole-bearing biotite-gneiss variety makes up about 10% of the total HNP, whereas the HNP-cg makes up at least 50%. The remaining ca. 40% is attributed to the fine-grained HNP.

4.3. HNP-cg Gneiss

The HNP-cg (TH1612) with its characteristic porphyritic texture is interpreted as a meta-igneous rock and is closely comparable to the LTY. The peraluminous chemistry and a possible volcanic arc to syncollision granite signature support this conclusion. Importantly, sample TH1612 is identical in composition to sample TH1610 from the internal LTY.

The zircon CL zonation patterns and SIMS U-Pb ages support this conclusion also. All investigated zircon crystals exhibit a marked oscillatory zonation. All SIMS spots were located in such domains. Eleven spots (out of twelve) are concordant within the associated uncertainties and form a cluster that defines a concordia age of 209.7 ± 2.1 Ma. This Late Triassic age is identical to the igneous zircon growth age proposed for the LTY (211.2 ± 1.7 Ma).

The twenty-one (U+Th)-total Pb monazite spot ages show a bimodal Y content distribution with a positive correlation of Y with U contents. Low Y spots (eleven spots, 1.23 ± 0.15 wt% Y) give a badly defined weighted mean spot date of 99.4 ± 28.0 Ma. The corresponding CHIME date is 71.2 ± 8.2 Ma but shows a high initial Pb content (ca. 100 wt-ppm Pb), raising some doubts as to the usefulness of the CHIME and spot dates. The high Y spots (ten spots, 1.81 ± 0.08 wt% Y), in contrast, define a weighted mean spot age of 65.6 ± 7.2 Ma and a CHIME age of 67.9 ± 7.2 Ma. All data are identical within the assigned analytical uncertainties. We interpret the ca. 66 Ma age to represent either monazite growth or recrystallization during a fluid-dominated metamorphic event. On the other hand, if the bimodal Y distribution of the monazites is taken to reflect two different fluid-dominated events overprinting the HNP-cg sample, we can postulate that one event took place at ca. 72 Ma (low Y) and the second at ca. 66 Ma (high Y).

The one spot (mnz7/1) giving an exceptionally high age of 219.5 ± 61.7 Ma is also aberrant in showing the lowest Y content (0.49 wt%) and, thus, the purest monazite composition of all analyzed domains. Though it only supports a very imprecise potential age for initial monazite growth, the spot age is not incompatible with the igneous zircon age of 209.7 ± 2.1 Ma.

The mica 40Ar/39Ar ages are interpreted as dating the cooling of the HNP-cg. We interpret the muscovite 40Ar/39Ar age of 41.9 ± 6.5 Ma as marking the cooling of the HNP-cg samples through the ca. 450°C isotherm and the biotite 40Ar/39Ar age of 36.6 ± 2.3 Ma cooling through ca. 300°C. This results in an Eocene cooling gradient of ca. 28°C/My for this temperature interval.

4.4. Fine-grained HNP Gneiss

The fine-grained HNP, in showing far more lithological variance than the HNP-cg, very characteristic calcsilicate enclaves, and more biotite-rich and more basic intercalations, is clearly a paragneiss-sequence [18, 31, 33]. No petrochronological data are available except for the comparably young biotite 40Ar/39Ar cooling age of 23.8 ± 8.8 Ma.

4.5. HNP-a Gneiss

The HNP-a occurs in cm to m thick layers within the fine-grained HNP biotite-gneisses. Apart from the specific mineralogy, its whole-rock geochemistry also indicates that it must have a different origin than the gneisses discussed earlier. It is comparably SiO2 poor and metaluminous and shows only a small Eu anomaly. In view of the close and intimate association with the paragneisses and calcsilicates of the “normal” fine-grained HNP, it is best interpreted as a meta-latitic to meta-andesitic lava or tuff. This is directly supported by the variety of zircon crystals in the sample. The zircon crystals provide a vast range of dates from ca. 497 to 53 Ma without showing any dominant age cluster(s). Nevertheless, geologically significant conclusions can be drawn. Three spots (out of seventeen) from CL-bright oscillatory-zoned mantle domains in three prismatic crystals define a concordia age of 494.5 ± 9.6 Ma. This Late Cambrian (Furogian) age is interpreted as directly dating the igneous formation, possibly the extrusion of the precursor of the HNP-a. A similar but imprecise zircon U-Pb upper intercept age of 452 −84/+93 Ma was reported by Hansen & Wemmer [12]. Due to the large uncertainties, the age was not considered significant. Our age implies that the time of sedimentation of the fine-grained HNP precursor is Late Cambrian or older. As the sedimentation age of the fine-grained HNP is otherwise unknown, this is noteworthy.

A second-age cluster is formed by another three spots at 52.8 ± 1.0 Ma from CL gray homogeneous zircon domains and small stubby crystals. This age is best interpreted as reflecting a pervasive fluid event leading to an incomplete recrystallization of the igneous zircon. A U-Pb SIMS lower intercept age of 56.5 ± 0.7 Ma from another HNP-a from the Khao Phra Mountain in the north-central part of the KCC reported by Hansen et al. [32] and ascribed to Paleocene metamorphism is identical to our age and supports our interpretation.

Five spots from homogeneous CL dark gray domains are grouped together, although they do not form a statistically valid cluster. They provide a weighted mean date of 63.9 ± 2.6 Ma. The remaining six spots are from nonspecific to strongly altered zircon domains with no characteristic CL emission that provide dates between ca. 400 and ca. 65 Ma. This behavior is best explained by the nature of the rock. Due to its comparable limited thicknesses and its intimate association with the biotite gneisses and calcsilicates of the fine-grained HNP, any penetrating fluid easily passing through the rock might provoke severe zircon recrystallization and alteration not matched in the far more massive LTY and HNP-cg.

Based on petrographical similarities, Hansen et al. [32] and Neugschwentner [33] stated that the HNP-a might be an equivalent of the Khao Dat Fa granite. But, according to our age data, it is conceivable that the Khao Dat Fa granite intruded into the HNP-a, though there is no field evidence to support or disprove this to date.

4.6. TY-Lg Gneiss

The leucocratic Thong Yi gneiss was not identified as an independent lithodeme by former investigations. It is definitely intrusive into the LTY as it displays discordant magmatic contacts, abundant occurrences of LTY enclaves, assimilated LTY Kfsp-megacrysts, and apophyses into the LTY. The LTY enclaves are polymetamorphic, whereas the leucocratic gneiss is only monometamorphic. Thus, it is younger than the first Indosinian metamorphic overprinting of the LTY dated at 207.5 ± 2 Ma. At the moment, no additional information is available.

4.7. Implications for the Geology of the KCC

Initially, the lithological naming by Kosuwan [16] as maintained by Kawakami et al. [15], Hansen et al. [32], and Sautter et al. [19] was chosen by us as a basis for the petrographical description and possible genetic subdivision of the various gneisses encountered in the KCC. However, our field and analytical work make it evident that the subdivision published earlier cannot be upheld; the new petrochronology data do not support them.

Consequently, to better reflect the new analytical data, their interpretation, and modern naming conventions, we propose to rename the rock association in the KCC as presented below. The proposal is based mainly on a lithological profile along the coast from HNP beach in the north to LTY beach in the south and along the river beds of Hin Lad and Samet Chunt in the more central part of the KCC:

  • HNP Complex (a lithodemic unit, formerly HNP) consisting of a pre-Indosinian meta-volcano-sedimentary sequence and a granitic intrusion into the former. The HNP Complex is formed by the following lithodemes (in stratigraphical order):

    • HNP paragneiss (formerly called fine-grained HNP) — a pre-Late Cambrian to Late Cambrian paragneiss formation. According to the whole rock geochemistry, the protolith had a quartz-arenitic to arkosic composition.

    • Khao Phra meta-andesite (formerly called HNP-a) — a ca. 495 ± 10 Ma (Late Cambrian, Furogian) andesitic (to latitic) meta-volcanic amphibole-bearing gneiss intimately associated with the HNP paragneiss lithodeme.

    • HNP orthogneiss (formerly called HNP-cg) — a Late Triassic (209.7 ± 2.1 Ma) former granitic intrusion into the HNP paragneiss. The meta-granite might be an equivalent to the LTY orthogneiss.

  • Khao Yoi paragneiss — pre-Indosinian muscovite-garnet paragneisses and schists intercalated with quartzites, calcsilicates, and marbles of unknown sedimentation and overprinting ages.

  • Khao Dat Fa orthogneiss (formerly Khao Dat Fa granite or Khao Dat Fa granite gneiss)—a ca. 477 ± 7 Ma old homogeneous granitic meta-pluton intruded into the HNP and Khao Yoi paragneisses.

  • LTY orthogneiss — a ca. 211 ± 2 Ma old, homogeneous granitic meta-pluton intruded into the HNP paragneiss. This might be an equivalent to the HNP orthogneiss. This igneous event is accompanied by the formation of numerous pegmatites and aplitic dikes crosscutting the gneisses in a number of generations.

  • TY-lg orthogneiss — a post-211 Ma leucocratic meta-granite intruded into the LTY orthogneiss and possibly into the HNP paragneiss. The intrusion age is most likely Late Cretaceous.

  • Khao Pret granite — a Late Cretaceous (ca. 68 Ma) pluton, intruded into all former lithodemic units.

All lithodemic units except the TY-lg orthogneiss and the Khao Pret granite are polymetamorphic with metamorphic grades ranging from higher greenschist-facies lithologies (Khao Yoi paragneiss) to sillimanite-bearing upper amphibolite-facies rocks. Metamorphic ages range from pre-Indosinian (probably pre-Late Cambrian) to Indosinian and Late Cretaceous.

The cooling from amphibolite-facies conditions and initial exhumation of the complex must postdate the Khao Pret granite intrusion (<68 Ma). The muscovite 40Ar/39Ar ages suggest a regional cooling of the KCC below ca. 450°C in the Eocene (ca. 42 Ma). The Khao Pret granite potentially shows a younger muscovite 40Ar/39Ar cooling age (37.1 ± 5.2 Ma), suggesting that the granite might have acted as a latent hot spot during the cooling and exhumation of the KCC. Cooling to ca. 300°C (biotite 40Ar/39Ar ages) was reached some 5 My later. The final cooling steps were ca. 33 Ma for the ZFT ages and ca. 20 Ma for the AFT ages [19]. As above, the Khao Pret granite seems to have somewhat younger cooling ages. The number of KCC cooling ages is not sufficient to decipher any variation in the cooling patterns within the complex. We thus cannot decide whether the KCC was exhumed as one rigid body or as individual crustal blocks with differential exhumation and cooling histories.

4.8. Regional Geological Implications

In the following, the above scenario is combined with the evolution of other basement occurrences in Thailand. A compilation of age data relevant to the discussion of the KCC is presented in Figure 11.

The sedimentation age(s) of the HNP and Khao Yoi paragneiss protoliths remain uncertain but must be contemporaneous or older than the Late Cambrian igneous zircon age of 495 ± 10 Ma of the HNP meta-andesite. The igneous protolith age of a granitic orthogneiss from Khao Tao Ma in the Hub-Kapong to Pran Buri area has been dated at 502 ± 8 [41]. This is identical to our age of the KCC meta-andesite. This consistency is further evidence for the existence of the Late Cambrian crust in Peninsular Thailand. Our conclusions also show, that, in parts, the sedimentary protoliths of the various paragneisses encountered in the basement areas must be of a Late Cambrian or even older age, and the Khao Dat Fa granite was intruded into the HNP Complex at 477 ± 7 Ma. According to Lin et al. [14], the chemical similarity and spatial continuity of the Khao Tao Ma granitic orthogneiss with other pre-Neotethys marginal Eurasian and Sibumasu granitoids indicate a paleogeographic association in a similar magmatic arc-related regime along the Gondwana India-Australia margin. The ca. 495 Ma HNP meta-andesite also shows volcanic arc-related geochemistry and thus fits perfectly into this paleogeographical association. The Khao Dat Fa granite has a U-Pb zircon age of 477 ± 7 Ma and thus has to be seen as belonging to this Lower Paleozoic crust as well.

Geochronological studies on granitic rocks interpreted as belonging to the Sibumasu terrane show Indosinian ages ranging from the Middle to the Late Triassic [3, 8, 10, 15, 23, 42-44]. Whether all of these ages reflect only igneous activity or also metamorphism is open to discussion. Some suggest protolith emplacement [42, 43], whereas others suggest metamorphism [10]. Furthermore, Hansen et al. [32] state that the regional distribution of zircon ages implies that all samples supposed to belong to the Sibumasu terrain show the presence of at least two metamorphic events, namely, in the Late Triassic during the Indosinian Orogeny and in the Cretaceous possibly related to the northward movement of India. They do not show any signs of an earlier, pre-Triassic event. In contrast, samples from the Sukhothai arc do show evidence of Middle Ordovician metamorphism. However, age data published since 2014 do not support this picture. The investigation of different zircon generations across the Thai crystalline basement allowed to develop a model for the crustal evolution of the Sibumasu terrane in which inherited zircon cores suggest the presence of Neo-Proterozoic to Paleozoic precursor rocks [45]. In this respect, the KCC shows an identical pre-Indosinian evolution to that hypothesized for the basement outcrops north of the Thai Peninsula. The model also promotes a two-step Mesozoic granitic igneous evolution. The first step may be related to anatectic melting of the pre-Indosinian crust during the Middle Triassic (ca. 242–235 Ma), and the second to postintrusive cooling during the Late Triassic (ca. 220–213 Ma). However, in contrast to what is postulated for more northerly regions, no Middle Triassic igneous or metamorphic activity is recorded in the KCC to date. We postulate that for the gneisses of the KCC only one Triassic tectonothermal event around 212–205 Ma occurred during the late collisional stage of the Indosinian Orogeny. No information on the post-Indosinian to Late Cretaceous evolution of the KCC is available.

The Cenozoic-Paleogene evolution can be summarized as follows. The monazite ages (ca. 80–60 Ma) from the HNP and LTY orthogneisses and zircon rims from the HNP meta-andesite reflect partial recrystallization due to pervasive fluid-overprinting during the Cretaceous tectonothermal event. This event is well documented throughout Sibumasu. A temporal fixpoint in the KCC evolution is the intrusion of the Khao Pret granite at 67.5 ± 1.3 Ma. The granite being postkinematic, the age must also likely define a minimum age for the thermal maximum (ca. 670°C) of the regional upper amphibolite-facies metamorphism in the area [33]. Thus, cooling from high-T conditions and initial exhumation of the KCC must be younger than ca. 68 Ma. Cooling to below ca. 450°C was reached in the Eocene at ca. 42 Ma and to ca. 300°C some 5 My later. The final cooling steps are recorded at ca. 33 Ma by the ZFT ages (ca. 200°C) and at ca. 20 Ma by the AFT ages (ca. 120°C). The Khao Pret granite shows a somewhat delayed cooling, suggesting that it may have acted as a hot-spot during this cooling and exhumation event.

Our zircon and monazite ages and the mica cooling ages fit well into the regional cooling pattern and exhumation model claimed for other basement outcrops in northern Malaysia, Thailand, and Burma [19]. The model suggests a similar succession of tectonothermal and sedimentary events throughout west Sundaland that were driven by the gradual northward migration of India from the Cretaceous to the Neogene. The model shows a more or less linear relation of thermochronological ages with geographic latitude. In the north Malayan Stong complex, the igneous and metamorphic evolution began in the Middle Cretaceous (ca. 90 Ma), whereas, in the Burmese Mogok Metamorphic Belt and Gaoligong, it did not begin before ca. 25 Ma. The corresponding ages from the KCC (ca. 68 Ma) fit into this model. The cooling and exhumation patterns of western Sundaland show that the Cretaceous to Neogene cooling rates were relatively low (ca. 10°K/My; 650°C–100°C in ca. 50 My) in the south, whereas in the north in Burma the same cooling rates were very high (ca. 50°K/My; 600°C–100°C in ca. 10 My). Again, the thermochronological ages from the KCC fit well into this model.

Applying the new lithodemic division, the KCC (Changwat Nakhon Si Thammarat, Thailand) shows the following evolution:

  1. Sedimentation and/or metamorphism of the HNP paragneiss and Khao Yoi paragneiss precursor rocks in the Late Cambrian (Furogian, >495 ± 10 Ma) at the latest.

  2. HNP meta-andesite protolith syn- to postsedimentary extrusion into the HNP complex precursor (meta-?)sediments in the Late Cambrian (Furogian, 495 ± 10 Ma).

  3. Khao Dat Fa orthogneiss protolith intrusion into the HNP and the Khao Yoi paragneisses (Early Ordovician, 477 ± 7 Ma).

  4. LTY orthogneiss protolith intrusion (211 ± 2 Ma) and HNP orthogneiss protolith intrusion (210 ± 2 Ma) into the HNP paragneiss during the Indosinian Orogeny in the Norian (Late Triassic).

  5. Subsequent first metamorphic overprinting of the LTY granite and the HNP complex including the Khao Dat Fa granite also in the Late Triassic (ca. 208–205 Ma).

  6. Metamorphic overprinting of all lithodemic units and contemporaneous intrusion of the Khao Pret granite in the Late Cretaceous and Early Paleogene (ca. 80–68 Ma).

  7. Formation of the KCC in the Eocene (<42 Ma), exhumation, and regional cooling below ca. 450°C in the Eocene and latest cooling to ca. 120°C in the Miocene (ca. 20 Ma).

The evolutionary data show that the KCC is part of Sibumasu, and its Late Cretaceous-Neogene cooling pattern and exhumation history can be directly related to the northward drift of India.

Jiranan Homnan, Michael Oester, Markus Palzer, and David Heuser are thanked for their help in the field and discussions. Franz Biedermann (University of Vienna, Austria) is thanked for help in sample preparation. Padhraig Kennan (University College Dublin, Ireland) is thanked for substantially refining the English. Two anonymous reviewers and journal editor Chuan-Lin Zhang are thanked for their help in improving the original manuscript. UK was funded by the ASEAN European Academic University Network (ASEA-UNINET). The research activities of JB were co-financed by funds granted under the Research Excellence Initiative of the University of Silesia in Katowice (ZFIN11651022). QL acknowledges funding by the National Nature Sciences Foundation of China (Grant no. 41673059). The research is also funded by the Thailand Science research and Innovation Fund Chulalongkorn University (DIS66230008) granted to PK and PC.

The authors declare that there is no conflict of interest regarding the publication of this paper.

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).

Supplementary data