A straight lineament a-a’ (Fig. 2) limits the northwestern foot of the Yuehkuang Mountains (maximum height is 649 m at Mt. Yuehkuang), mostly underlain by late Miocene sedimentary rocks and striking NNE-SSW on the north and NE-SW on the south. Several lineaments parallel to the lineament a-a’ are present on the southern part of the Yuehkuang Mountains (Fig. 2), but they also seem to be structural relief due to lithologic control, because no deformed terraces or alluvial lowland are observed.
We found one positive evidence for the deformation of terrace on the northeastern extension of this lineament a-a’, where at least three steps of fluvial terraces (TI to TIII in descending order) are preserved (Fig. 3). TI (202 m above sea level, TII (150 m) and TIII group (lower than 120 m). Most terraces gently slope down westward toward the main river (Tzuhsien River, Fig. 2). However, TI at Kuanyin Mt slopes eastwards from 202 m at the western margin to 180 m at the eastern margin (Fig. 3, profile a-a’). Such an eastward dip of the terrace is abnormal,
Figure 3. Terrace and fault map of western part of the Yuehkuang Mountains . Dashed line at the base of the terrace scarp is estimated location of the Chishan Fault as a buried reverse fault. Profile a-a’ across Terrace I shows eastward back tilting, and possible location of a buried Chishan Fault.
considering the origin of the terrace, which is located east side of the south-flowing
Tzuhsien River. We interpret that this eastward dip represents a back tilting resulted from a buried reverse fault with east-dipping fault plane (part of the Chishan Fault).
The height of supposed fault scarp on the western margin of TI is ca. 100 m, but probably smaller than this value, taking the back tilting and down cutting by the river into account. Even so, such a high scarp still suggests repeated activities of faulting.
The profile of TII at Tsukeng (location in shown in Fig. 3）is rather horizontal, except for the inner edge, which is covered by talus deposits. Such a nearly horizontal slope of TII may show some effect of back tilting, but it is difficult to confirm because of too much artificial modification. For the lower TIII group, we could not observe the evidence of deformation.
It is certain that this part of the Chishan Fault has activated after the formation of TI, and possibly after that of Terrace II during the late Quaternary. However, since we have no age control for terraces, it is difficult to estimate slip rate and establish progressive deformation.
On the southwestern end of lineament a-a’ (Fig. 2), three hills (1,2,3) make a series of cols between these hills and the main part of the Yuehkuang Mountains. It may reflect the location of the Chishan Fault, but no direct evidence of recent deformation is found. Thus, we conclude that this part of lineament a–a’ follows the Chishan Fault, but the present morphology is a structural relief, reflecting the non-resistant weak shattered zone on the fault.
The southern margin of Yuehkuang Mountains may be also structurally controlled, with some submerged morphologies. No evidence of active deformation is found (Fig. 2).
II-2. Chishan Fault and Lineaments in the Chungliao Mountains and Lingkou Mountains
There are several subparallel straight lineaments, nearly parallel to the trend of the mountains, in the Chungliao and Lingkou Mountains areas (A-J in Fig.4). Dotted zones A, B and C represent the relatively wide (less than 500 m) summit of the both mountains, where Miocene (is it correct?) sandstone layers are exposed. The top of Chungliao Mountains (the highest point is 421 m at Mt. Chungliao) is a flat summit making a remarkable topographic high, corresponds to the hanging wall of the Chishan Fault, which is represented by lineament G. Similarly, C corresponds to the top of the Lingkou Mountains (maximum height is 261 m at Kaoshiung Mt.). Top of the Lingkou Mountains is prominent, but no fault origin was proposed. It seems that the ridge coincides with the presence of hard sandstone and represents structural relief. A series of short, sharp and NEE-SWW striking ridges (D) are also sandstone ridges, which alternate with straight linear depressions, corresponding to the presence
Figure 4. Distribution of lineaments in the Chungliao and Lingkou Mountains, interpretes from aerial photographs and superimposed on shaded relief map based on DEM. Dotted zones (A-C) are topographical high (corresponding to the presence of sandstone), and solid lines (D) are narrow sandstone ridges. Dashed lines (E-J) are lineaments, expressed as depression or valleys, mostly correspond to the presence of mudstone. G1 coincides with the know trace of the Chishan Fault.
Locs.1-4 are fault sites described in the text
of mudstone. Thus, ridges of D group and depression in between are clearly structural
tween
posures on this lineament G1. A very wide
Chis
s f
relief. The presence of sandstone contributes to the formation of high summit or ridges, although their scale is variable depending on the thickness of sandstone.
In contrast, lineaments E to J (Fig. 4) are valleys or depressions be sandstone ridges. Their length is also variable from ca. 10 km to only 1 km or less.
Among them, G1 is the longest, extending with nearly NE to SW direction. It slightly bends however, at both ends. This trace is surprisingly coincident with the known Chishan Fault trace, although we traced this lineament just connecting topographic low based on the aerial photo interpretation. Thus, lineament G1 is certainly a topographic expression of the Chishan Fault.
There are at least two confirmed fault ex
and extensively sheared fault zone cutting bedrock is exposed at Loc.1. Fault appeared in the tunnel at Loc. 2 (Sung et al., 2004). However, the following two fault exposures, discovered by Chen et al. (2005), are not located on this lineament G1. At Loc. 3 on linearment G3, Holocene alluvial deposits of 7189±160 yrs BP is thrusted up by early Pleistocene sediments (Chen et al, 2005), indicating the Holocene activity of the Chishan Fault. The upper part of the Holocene deposits with unknown age is not deformed at this locality, however. It is a question if Loc. 3 is actually located on G1, or on G3. If the former is the case, the Chishan Fault trace extremely bends southward and again changes its strike to NE to SW. If the latter is the case, a short lineament G3 is an active portion of the Chishan Fault during the Holocene. As the third case, it is possible that the G1 and G3 join together. We cannot decide which one is the probable interpretation at this stage. Certain thing is that the fault exposure of Loc. 3 is not located on the direct southwestward extension of G1. At Loc. 4 on the northern end of G2. Lingkou conglomerate is faulted, showing Quaternary faulting (Chen et al., 2005). Almost the same problem as Loc. 3 occurs at Loc. 4, where the fault exposure is not located on the main part of G1, but on short lineament of G2.
Lineament G1 exactly coincides with the geologically identified location of the han Fault, but at its both ends, the fault exposures may be located not exactly on lineament G1, but on lineament G2 and G3, slightly away from old fault trace (G1). Then the Chishan Fault may not be a imple and single trace, but is composed o multiple traces, making a fault zone. Also, we cannot identify the deformed terrace or alluvial lowland on the lineaments shown in Fig. 4 as the key for the recent faulting. Because of intensive erosion, the Chishan Fault is expressed as only depression along the shattered zone, without clear morphologic evidence of recent faulting. In this sense, lineament G1 may be still tectonically controlled structural relief. Two lineaments on the northwest of lineament E and F are also just structural relief.
of the Kaoping River e south of this figure. Fl
Figure 5. Te ea of west
. Terrace distribution is limited in this area. Landslide removed terraces on exural scarp is identified in the areas shown in Fig. 9 and 10.
rrace and fault map on shaded relief map based on DEM in the northern ar
th
Figure 6. Terrace and fault map on shaded relief map based on DEM in the southern area of west of the Kaoping River. T2 terrace is widely distributed, in addition to the local presence of T1 and
ca. 6 km long, and parallel to A. It seems to be also structural relief. Straight lineament I can be divided into three, I1, I2 and I3. They may be also structural origin. But I2 may express tectonic relief because it marks a topographic boundary between mountains to
T3.
Lineament H1 and H2, east of the sandstone topographic high (A) is
the east and
distinguish tectonic
III. T est bank of the Kaoping River.
Characteristic landforms of the west of the Kaoping River are the presence of three steps of fluvial terraces (T1, T2, and T3 (Figs. 5 and 6), and flexural scarp of the eastern margin of T2 and T3 at northern part (Fig. 5). We describe firstly terrace features and then their deformation.
III-1. Terraces on the west of the Kaoping River
hree steps of fluvial terraces, T1, T2 and T3 in descending order, are present along the west bank of the Kaoping River (Figs 5 and 6). Correlation of these terraces with those on the Yuehkuang Mountains is difficult because of lack of terrace continuity; hence we use different symbols for both areas. N-S trending terrace risers, often parallel to the Kaoping River, suggest that most terraces were formed by the south-flowing Kaoping River. These terraces are not exactly mapped in the previous studies.
1 is only found at west of Chuliao, southernmost of the study area at about 120 m (inner margin) to 90 m (outer margin) above sea level (Figs. 6 and 7) . Terrace riser, separating from T2, is ca. 20 m high and very clear (profile a-a’ in Fig. 7). T1 is underlain by fluvial gravels, composed of subrounded gravels with few matrixes, at least 13 m thick at Loc.5 (Fig. 7). This terrace can be traced northward for 8 km distance, as dissected narrow terrace remnants of 100-120 m above sea level (Figs. 6 and 7
T2 is the extensive terrace (ca.1 km wide at north and ca. 2.5 km wide at south Figs. 5, 6, and 7）and is developed over 25 km long starting southward from the Chishan Park to the north. Since the inner margin of T2 is N-S trending, parallel to the Kaoping River, terraces, especially T2, could be formed by the Kaoping River. This at there was a relatively stable time to permitting the formation of
glomerate metimes is sometimes difficult because of their similarities in facies. Hence the thick
alluvial lowland to the west. Further detailed fieldworks are required to relief and structural one.
erraces and fault on the w
T
T
).
suggests th
extensive T2, which is absent in southern part of Fig. 5 due to removal by intensive landslide. Distinction between terrace deposits and underlying Lingkou Con
so
ness of terrace deposits is often uncertain. One excellent exposure at south of the Chishan Park (Loc. 6 in Fig. 9) reveals that a clear unconformity between the eastward dipping (25°E) Plistoocene Lingkou Conglomerate and T2 deposits of ca. 5 m thick (Photo A), mainly composed of subrounded fluvial gravels.
Figure 7. Terrace map of the Chuliao area and profile a-a’ across T1 and T2. Eastern margin of T2 is limited by sharply defined erosional cliff, retreated from the original tectonic scarp.
Figure 8. Longitudinal profile of terraces，modern flood plain and the river floor along the Kaoping River, drawn from the photomap of 1:5,000 in scale. Remarkable difference etween the inner margin height and outer margin height of T2 is caused by flexural deformation on the north, and by presence of secondary deposits on the south.
The longitudinal height distribution of terraces is shown in Fig. 8. The inner
ery close to the base of flexural scarp. b-b’
profile is for T3, younger terrace than T2. Notice that the apparent amount of vertical offset of T3 is smaller than T2, suggesting progressive deformation.
margin of T2 ranges from 140 m at Chishan Park (Fig. 9) and 125 m at Shangcholi (Fig. 10), where the inner margin of T2 is sharply defined. Considerable height difference between the outer and inner margin at these two locations are due to strong eastward deformation of T2, as will be discussed in III-2. In the southern area, the inner margin height is higher than the upper stream, reaching 160 m at maximum, which is significantly higher than the height of Shangcholi, located further upstream.
In addition, the inner margin height is rather irregular, and T2 surface dips eastwards considerably, although its outer margin is limited by clear terrace riser from the alluvial lowland. These features suggest that the inner part of original T2 terrace surface, originally created by the Kaoping River, is covered by alluvial fan deposits of various scales, transported by eastward flowing tributaries. A dashed line shown in Fig. 8 is a possible boundary between the original terrace surface and secondary covered fan surface, judging from the different gradient of terrace surfaces and ranges from ca. 115 m to 60 m. The southward gradient of T2, represented by a dashed line, is 40 m/22 km, and larger than 15 m/22 km for the alluvial lowland (Fig. 8).
Figure 9. Detailed contour map of fluvial terraces and their deformation in the Chishan Park area (drown by 20 m grid DEM). Contour map and profiles represent the remarkable eastward warping scarp deforming T2 and T3. Smooth and convex profiles on the eastern margin is very distinctive.
A buried fault (Kaoping River Fault) is estimated v
Terraces lower than T2 are collectively called as T3 (group) and are fragmentary preserv
osits, although the longitudinal profile of T2 is steeper than the modern flood plain or ed at several sites (Figs. 5 and 6). One is at the Chishan Park, where a small T3 with ca.10 m lower height than T2 (Fig. 9) at its inner margin is present along the small east flowing stream, clearly indicating tributary origin. The other is in the southern part of near Chuliao, where relatively wide T3 is present at ca. 50 m in height, separated by ca. 30 m high terrace riser (Fig. 6). The correlation within T3 group is uncertain.
We have no age control from these terraces, underlain by fluvial gravel without prominent sand or clay lense for dating. It is also difficult to find out a direct relationship between the terrace formation and sea level changes, because the terraces of the study area do not continue to any of marine terraces. However, since this area is not so far from the coast (Fig. 1), it may possible to assume some of terraces correlate with sea level change. We tentatively assume that the most extensive and relatively well preserved T2 terrace as a possible correlative with high sea level of MIS 5e, judging from the following reasons: First, it is unlikely to say that T2 can continue to the lowest sea level of MIS 2, morphology of which should be very much steeper than the alluvial lowland and should be found beneath the Holocene alluvial dep
present river floor. Secondly, the formation of an extensive terrace requires some stable time as mentioned before, Then thirdly, the high sea level of MIS 5e usually leave its morphological trace as a prominent and well preserved marine terraces with presence of younger marine terraces, corresponding to MIS 5c or 5a when the uplift rate is high enough to emerge these terraces on land. Thus, as the first approximation, we assume that T2 possibly represent MIS 5e of ca. 125 ka, although we know that the degree of preservation of the terrace of this stage, is very much changeable depending on the original scale and nature of deposits in addition to the height etc. If the above mentioned correlation is the case, higher T1 can be correlated with MIS 7.
In any case, the estimation of formation age of terraces is very tentative, leaving important problems to be solved, supported by dating. Although we are uncertain about the age of terrace, it is quite obvious that T2 is clearly tectonically deformed and provides a positive evidence as the active fault for the Kaoping River Fault.
III-2. Identification of the active Kaoping River Fault
A straight scarp that limits the eastern (outer) margin of T2, and strikes nearly N-S direction, is very distinctive (Figs. 4, 5, and 6). This scarp starts from the Chishan Park at north and continues southward to the west of Pingtung and is ca. 30 km long.
Scarp profile is divided into two types.
Flexural scarp: The first type is typically seen at the Chishan Park (Fig. 9) and
Shangcholi (Fig. 10), and is characterized by the smooth and convex profile without any c
Kaoping River Fault after Shyu et al.(2005) for this fault. We did not f
Figur
lear knick point between the terrace surface and scarp (profiles in Figs. 9 and 10), photo. 2). Tilted gravel bed parallel to the scarp shows that the scarp is not erosional origin but former terrace surface and represents flexural or warping scarp, created by deformation on the hanging wall of subsurface fault. We consider that this fault is a reverse fault with east facing scarp and with westward dipping fault plane, probably high angle, judging from scarp profile and rather straight trace in the northern part.
We use the name of
ind positive evidence to prove the left-lateral offset, proposed by Shyu et al.
(2005) from our field observation. Flexural scarp of this pattern is often recognized on the reverse fault in Taiwan (e.g. Chen et al., 2003; Chang et al., 2004; Ota et al., 2004;
Lai et al., 2006 for the Chulungpu Fault and Ota et al., 2005 for Tunglo Fault). We assume that the base of the scarp represents the location of the fault, although the base is partly buried by Holocene deposits and no fault exposure was found yet. We examined some lineaments found on the alluvial lowland by the photo interpretation, but no evidence of Holocene activity was found in the field.
e 10. Detailed contour map of fluvial terraces and their deformation in the Shangchouli area (drawn by 20 m grid DEM) and flexural scarp deforming T2. Two profiles across T2 on the Shangchouli area aslso clearly show the eastward convex scarp by a buried Kaoping River fault.
The amount of vertical displacement is unable to be exactly determined, because of very gradual change from the original slope to deformed flexural scarp. The possible minimum amount of vertical offset of T2 ranges from 50 to 60 m (profiles a-a’ to d-d’ of Fig. 9, and profile a-a’ , b-b’ in Fig.10). At the Chishan Park, even a local T3 is also deformed, but the amount of vertical offset is about 30 m (profile b-b’
of Fig. 7B), smaller than that of T2, and indicates a progressive deformation since T2 formation.
Erosional scarp modified from flexural scarp: Intensive down cutting and
land-sliding in the mudstone area completely removed terrace remnants for 8 km between Shangchouli and Dapingting (Fig. 5), thus here we cannot identify the tectonic scarp. Southward from Dapingting again T2 is extensively preserved, but scarp morphology is different from Chishan Park and Shangchouli area. There is a sharp knick point between terrace surface and terrace riser, as illustrated in the profile shown in Fig. 11. This is the second type of the scarp, and we interpret that this steep scarp was originally formed as a flexural scarp same as those in the northern area, but now is modified by fluvial erosion of the Kaoping River, because this river flows just base of the cliff. Fig. 12A and B illustrates how profiles are different between the orthern part (tectonic scarp) and the southern part (erosional scarp, modifying riginal tectonic scarp) and how they were formed. Such different profiles as a no
composite result of tectonic and erosional processes are typically preserved in Sado Island, Japan (Ota et al., 1992; Fig. 12C and D). In this study area, the flexural scarp is completely removed, thus we might have difficulty to assume the presence of initial tectonic scarp, unless the preserved flexural scarp at the Chishan Park and Shanchouli.
In the case of Sado Island, main flexural scarp with secondary range-facing scarp is well preserved in the Kuninaka plain (Fig. 12 C). In contrast, the major tectonic scarp is eroded out by marine erosion on the open coast, however, still the preserved range-facing scarp indicates that the present sharp cliff is modified from the initial tectonic scarp (Fig.12 D). This gives us a good analogy of the interpretation for two types of scarp along the Kaoping River. Difference between A, B（Kaoping River fault）and C,D (Sado Island) can be explained by the different fault dip. In the case of Sado Island (C, D), range-facing scarplet and associated tectonic bulge on the hanging
In the case of Sado Island, main flexural scarp with secondary range-facing scarp is well preserved in the Kuninaka plain (Fig. 12 C). In contrast, the major tectonic scarp is eroded out by marine erosion on the open coast, however, still the preserved range-facing scarp indicates that the present sharp cliff is modified from the initial tectonic scarp (Fig.12 D). This gives us a good analogy of the interpretation for two types of scarp along the Kaoping River. Difference between A, B（Kaoping River fault）and C,D (Sado Island) can be explained by the different fault dip. In the case of Sado Island (C, D), range-facing scarplet and associated tectonic bulge on the hanging