Results of Post-Pumping Land Elevation Change

Chapter 2: Thirty-Year Land Elevation Change from Subsidence to Uplift

2.3 Analyses and Results

2.3.2 Results of Post-Pumping Land Elevation Change

Several interesting and important characteristics of the land surface level change (subsidence and uplift) in the Taipei Basin during the post-pumping time can be inferred from the contour maps. First, the whole basin still subsided after the stop of groundwater pumping, but with a decreasing trend from 40-70 mm/yr to 0-30 mm/yr , during 1975 to 1989, denoted here as Phase 1 (Fig. 2-5). Second, a large part of the land became to uplift since 1989, denoted as Phase 2 (Fig. 2-6). However, the surface elevation change was far from homogeneous. For instance, enhanced and persistent subsidence occurred in north-western margin of the basin near Guandu (for the place names mentioned, please refer to Fig. 2-2b) throughout the entire time span (1975-2003), and concentrated subsidence in western margin of the basin (Wuku, Shulin) and north-eastern basin (Dazhi-Sungshan) was also very persistent before 1996. Other locus of subsidence (e.g., Banchiao, Zhonghe, Jingmei) as well as irregular and short-lived subsidence or uplift were also observed.

A period-by-period account of land elevation change in the Taipei Basin is presented as followed. Basin-wide decreasing subsidence characterized all three periods in Phase 1. During 1975-1980 (Fig. 2-5a; 94 benchmark), subsidence was dominant and was mostly concentrated in the northwestern (Guandu, exceeded 60 mm/yr; and Wuku, maximum 75 mm/yr) and northeastern portions (Dazhi, nearly 70

mm/yr). In 1980-1985 (Fig. 2-5b; 76 benchmark), significant subsidence continuously occurred along the western margin of the basin (Guandu, reached 12 mm/yr; Wuku, over 10 mm/yr) and in the northeastern corner (Dazhi, more than 20 mm/yr), however, with smaller rates compared to the previous period. For 1985-1989 (Fig. 2-5c; 157 benchmark), a large part of areas in the basin were continuous to subside (in particular, the western margin), but some small areas began to show slight uplifting, especially in the central and southern parts, although some of the benchmark records showing extremely sharp uplift might be anomalies. In general, the subsidence rate continued to decrease, comparing to the two previous periods; nonetheless, localized subsidence was still persistent along the western margin of the basin (Guandu, attained 30 mm/yr;

northern Wuku, >10 mm/yr; Shulin, 10 mm/yr) and other patches in the basin (south Dazhi, maximum 15 mm/yr; and Jingmei, around 5 mm/yr).

Starting from about 1989 (Phase 2), uplift began to appear significantly in the Taipei basin and declined in magnitude in the later half of this Phase. During 1989 to 1994 (Fig. 2-6a; 111 benchmark), we observed that the surface subsidence almost ceased and a large portion of the Taipei Basin began to slight uplift, especially in the centre of the basin (central Taipei, 10 to 15 mm/yr; Banchiao, 10 to 15 mm/yr).

Nevertheless, slight subsidence still persisted in western margin of the basin (e.g., 0 to 5 mm/yr in Guandu and Wuku). In 1994-1996 (Fig. 2-6b; 146 benchmark), slight subsidence continuously occurred and was localized in western side of the basin (Guandu, as high as 14 mm/yr; Wuku, maximum 10 mm/yr; and Shulin to Banchiao, up to 6 mm/yr), and in southern Taipei (Zhonghe, reached 10 mm/yr). On the other hand, uplift seemed to begin to dominate the eastern half of the basin, but with a lesser magnitude. In 1996-2000 (Fig. 2-6c; 103 benchmark), the whole Taipei Basin was generally experiencing mild uplift, except for north-western margin of the basin (Guandu – Wuku, subsidence with a possible bull-eye reaching maximum over 20 mm/yr), and other local areas (Dazhi, Zhonghe, and Jingmei). During the most recent period, 2000-2003 (Fig. 2-6d; 94 benchmark), slight uplift prevailed in the Taipei ascribed to the compaction of the water-depleted aquifers, resulting from the residual effect of the rapidly declined piezometric head due to over-pumping of groundwater.

Fig. 2-5. Contour maps of observed land surface elevation change in Phase 1 (1975-1989) of post-pumping period. (a) 1975-1980. (b) 1980-1985. (c) 1985-1989. Negative rate values indicate subsidence, and vice versa. Contour interval is 5 mm/yr. Coordinates shown are of TWD 67 Transverse Mercator system. The whole basin revealed a general subsidence from 1975-1985, with some localized effects. See detailed discussion in text.

Fig. 2-6. Contour maps of observed land surface elevation change in Phase 2 (1989-2003) of post-pumping period. (a) 1989-1994. (b) 1994-1996. (c) 1996-2000. (d) 2000-2003. Annotations see Fig. 2-5. The basin showed a general uplift from 1989 to 2003; however, local subsidence occurred at several places.

See details in text.

While massive pumping was ceased, recharge of groundwater from two major rivers (Xindian and Dahan Rivers) to the aquifers has raised the water table for more than 30 m from the 1970s to 2003 (Fig. 2-3c). We anticipate that the recovery of the groundwater level would produce an elastic rebound for the aquifers, thus caused basin uplift since 1989 in Phase 2. However, the concentrated and rather persistent subsidence and the reduced uplift occurred along the western margin and northern part of the Taipei Basin, compared to the central part of the basin, suggesting under impacts of other factors.

Fig. 2-7. Evolution of ground elevation change rate in post-pumping period from 1975 to 2003. (a) Western Taipei as the Wuku area. (b) Central Taipei. Distribution of the benchmarks used for calculation is shown on Fig. 2-4.

2.4 Mechanisms of land elevation changes during post-pumping

The land elevation change of the Taipei basin could be a product of complexly interfering natural and artificial processes. The heterogeneous pattern of land vertical changes in the Taipei Basin certainly reflects the spatial and temporal context of different agents. In the spatial domain, these agents, according to where the processes take place, can be described and classified by their operation depth. Here we propose main mechanisms to explain the land elevation changes in the Taipei Basin, attributing effects of land elevation change of the Taipei basin to three major components: (1) surface soil compaction (the shallow component), (2) deformation of aquifers, including compaction and elastic rebound due to groundwater effect (the intermediate component), and (3) tectonic load (the deep and crustal-scale component).

2.4.1 Near surface soil compaction (the shallow component)

Local enhanced subsidence can be observed in several areas within the Taipei Basin in many periods we analysed. Subsidence caused by construction (loads of new

buildings) cannot be omitted, but such phenomenon might be extremely localized and short-lived. Otherwise, subsidence in several places, such as Dazhi, Sungshan, Jingmei, Banchiao and Shulin, occurred throughout many periods although not always present (Figs. 5 and 6). Since their close proximity to the three major river channels in the Taipei basin (i.e., Keelung, Dahan and Xindian Rivers), we intend to interpret that a large part of subsidence in these areas to be contributed by compaction of recent flood or overbank deposits (especially mud). Effect of in situ sediment compaction, often referred as natural subsidence (Waltham, 2002), therefore must be evaluated in order to further distinguish the role and intensity of natural and artificial factors played on the observed ground level change.

According to a recent in situ study in the Taipei Basin conducted by Lin et al.

(1999), compaction of the loose sediments occurred mainly in the clayey layers of the uppermost 50 m Holocene deposits (generally corresponding to the Sungshan Formation) in three monitored wells in the western part of the basin at Wuku and Banchiao, in a time span of three and a half years from March, 1994 to September, 1997. For older sediments underneath, the compaction and consolidation were observed to be rather slow and was ignorable regardless of the types of the deposits.

Based on their results, we estimated the averaged annual rate of compaction for 1-meter-thick clayey layers to be ranging from 0.14 to 0.28 mm/yr (Table 1) by dividing the cumulative compaction of the clay sections in the monitored wells by the monitoring interval of three and half years. Meanwhile we reconstructed an isopach map for clay thickness within the topmost deposits of the Sungshan Formation in the basin (Fig. 2-8a) based on 350 borehole records from Central Geological Survey (open source, available on http://210.69.81.69/geo/frame/gsb88.cfm). The proxy of present variation of shallow soil compaction rate in the Taipei Basin can be obtained by multiplying the deduced average annual compaction rate for 1-meter-thick clay (i.e., 0.2 ± 0.07mm/yr) with the shallow clay thickness isopach. We estimate contemporary shallow soil compaction to contribute 1-8 mm/yr of subsidence in the basin with most land ranging 2 to 5 mm/yr (Fig. 2-8b).

Natural subsidence rates of surface soil documented elsewhere in similar flood plain and deltaic environments fall in the same order of magnitude (Jelgersma, 1996), such as Ravenna (4-6 mm/yr, Teatini et al., 2005), Romagna (0-5 mm/yr, Gambolati et al., 1999), Venice (0.5-1.3 mm/yr, Gatto and Carbognin, 1981), and Po River Delta (3-5 mm/yr, Gambolati and Teatini, 1998) in Italy and coastal Louisiana, USA (shallow component 1.5-2.5 mm/yr, Dokka, 2006). In addition, the modelling results of Meckel et al. (2006) based on Holocene stratigraphy of lower Mississippi River Delta in southern Louisiana, which closely resembles the compacting shallow

sediments of the Sungshan formation in the Taipei Basin, is applied to evaluate our proxy. Taking the surface soil in the Taipei Basin (50-meter thick and deposited within the Holocene) into their cumulative distribution functions of compaction rate model, the 90th percentile compaction rate (figure 2 of Meckel et al., 2006) ranges from 1.5 to 4 mm/yr. This suggests that our estimate is reasonable and not likely to underestimate present soil compaction.

Comparison between the contour maps of the observed elevation change (Figs. 5 and 6) and the estimated shallow soil compaction (Fig. 2-8b) shows that concentrated subsidence are indeed in a close relation with thicker soil/clay layers, especially along the main river courses (Fig. 2-3a). For example, ultra thick clay with resultant peak soil compaction is likely the main source of continuous subsidence in the marshy Guandu area along the Danshui River in the north-western corner of the basin.

However, the Wuku area in western Taipei, on the contrary, is not subject to high soil compaction. As a consequence, persistent subsidence in the western margin of the Taipei Basin (Guandu, Wuku, and sometimes extending to Shulin) cannot be attributed totally to natural compaction especially during Phase 2, and requires other explanations to decipher its origin.

Fig. 2-8. (a) Isopach of the Holocene clayish sediments (topmost 50-m deposits) in Taipei basin, constructed from 350 borehole records from Central Geological Survey (borehole sites marked in cross). (b) Estimated compaction rate of clayey layers within the Holocene deposits (soil compaction). See text for reference.

(Veterans Affair Office) ~10-20 10 (18-8) 0.2857

2.4.2 Tectonic load (the deep and crust-scale component)

Tectonic subsidence from activities of the Shanchiao Fault is thought to affect ground surface elevation change since present extensional deformation across the fault was documented (Yu et al., 1999a). It’s expected to be most pronounced in the near-fault hanging wall region, that is, the western margin of the basin, where the basin basement is dramatically deepened and reaches maximum depth. The long-term averaged tectonic subsidence rate in western Taipei Basin thus can be estimated by stratigraphic offset across the fault. The onset of basin sediment deposition was estimated to be around 0.4 Ma when basin basement was lowered to near sea level (Chen et al., 1995; Wei et al., 1998; Teng et al., 2001) and since then approximately 700 meters of basin sediments were accumulated. Dividing the total thickness of late-Quaternary basin deposits (700 m) by the time span of sedimentation (0.4 Myr) we obtain the late Quaternary averaged tectonic subsidence rate to be 1.75 mm/yr since 0.4 Ma. Note that it may contain contributions from co-seismic slips (Huang et al., 2007), thus representing the upper limit of value for interseismic tectonic subsidence rate. In the central part of the basin tectonic subsidence may attain half of the rate in western Taipei, approximately 0.88 mm/yr. We consider the tectonic subsidence rates to be constant over the investigated time intervals since no major earthquakes were recorded in shallow crust of the Taipei area during 1975 to 2003.

More accurate assessment of the ongoing tectonic subsidence could be derived from levelling of benchmarks attached to deep boreholes (e.g. Dokka, 2006) and is in urgent need for earthquake hazard research and mitigation in the Taipei metropolis.

2.4.3 Deformation of aquifers (the intermediate component)

In the hydro-mechanical coupling scheme as illustrated by Waltham (2002), abstracting groundwater would reduce pore-water pressure in aquifers, usually as sand or conglomerate beds, which behave in a seemingly elastic manner (Karig and Hou,

1992). However, this would be followed by reduced pore-water pressure in usually intercalated clay and aquitard materials in order to regain hydraulic equilibrium. Clay compaction is largely an irreversible one-way process, and is the major source for severe pumping-induced land subsidence (Terzaghi, 1925; Holzer, 1984; Phien-wej et al., 2006). When groundwater is recharged into the starved aquifers to recover the pore pressure, expansion of aquifer sand and gravel layers occurs and contributes to uplift of ground surface as elastic rebound (Waltham, 2002). In detail of this hydro-mechanical coupling, during piezometric head rise, the recharged sections of the aquifer will release the formerly imposed compressive strain to show dilatation, producing elastic rebound. On the other hand, the starved sections of aquifer and aquitard will continue to show compression as the result of time-dependent consolidation behaviour caused by past piezometric drawdown, until groundwater table has risen to remove deficits of pore pressure in these sections. Ground level change reflected on groundwater recovery is the competition between subsidence from starved aquifer (plus aquitard) section and uplift from the recharged aquifer section. The evolution generally follows the pattern of curve B in Fig. 2-1. Similar mechanism was also invoked to explain land level fluctuation in Las Vegas controlled by seasonal water table variations (Amelung et al., 1999).

According to well monitor records in the Taipei Basin (Fig. 2-3c), groundwater table started to rise as soon as pumping ceased in 1975. Compaction of Aquifer 1 during massive groundwater extraction would arrest gradually, following the recharge of groundwater in the aquifer. Near the end of Phase 1 as piezometric head was largely recovered, deformation of aquifer would become neutral as that rebound had developed to surpass remaining compaction. The uplift due to elastic rebound of Aquifer 1 was not visible on land surface until 1989 at the beginning of Phase 2 (Fig.

2-6). We interpret that at this time the amount and rate of elastic rebound outpaced residual aquifer clay compaction and other subsidence effects including the compaction of surface soil and tectonic subsidence of normal faulting. Elastic rebound of recharged aquifer would then gradually decrease in amount and rate since the groundwater level stabilized from mid 1990s. Ground level change pattern of the Taipei Basin after cessation of massive pumping with natural groundwater recharge thus fits well with the scheme of curve B in Fig. 2-1. The evolution of the post-pumping elevation change is characterized by a relatively longer period of waning subsidence (Phase 1: 1975-1989) followed by a shorter period of slight uplift (Phase 2: 1989-2003), due to a combination of decreasing compaction and increasing refill in Aquifer 1. The time lag between the start of piezometric level rise (t_stop in Fig.

2-1) and the first occurrence of the observed uplift of land surface (t_rebound in Fig. 2-1) in Taipei basin is about 15 years, which is believed to be closely related to the

recharge rate of Aquifer 1. The contrast on amount between the non-revertible, plastic clay compaction and the elastic rebound of aquifer strata may be a viable explanation to the fact that the recent uplift (since 1989) in the Taipei basin has a much lesser magnitude than that of the severe subsidence during massive pumping as before the 1970s. Most surface uplift due to rebound was observed in the central to eastern portion of the basin, roughly corresponding to area of most intense anthropogenic subsidence during pumping (compare Fig. 2-6 and Fig. 2-3a).

Similar phenomenon following cessation of massive groundwater exploitation was documented in Venice, Italy by Gatto and Carbognin (1981) where a land rebound of 2 cm, 15% of subsidence due to groundwater withdraw, occurred about five years later than the onset of regional piezometric head recovery from natural recharge.

Artificial recharge of a severely depleted hydrocarbon reservoir in Long Beach, California, induced an observed rebound of 33.5 cm where precedent subsidence due to petroleum production reached 9 meters in maxima (Allen and Maguya, 1969). Both cases demonstrated that aquifer rebound is generally about one order less in magnitude than the preceding subsidence. This is in good accordance with the situation in the Taipei Basin, whose subsidence is of two meters and rebound is of 10-20 cm in general. Recent studies of geodetic records in regions including Ravenna, Italy (Teatini et al., 2005) and Shanghai, China (Chai et al., 2004) displayed a sharp change on subsidence rates showing a switch from severe to slight subsidence while groundwater utilization was controlled or stopped; however, no uplift due to aquifer rebound was observed in these above regions yet. Differences in groundwater exploitation history, piezometric level evolution, and local hydrogeological properties (e.g. aquifer compressibility, porosity, permeability) are major factors for the discrepancies documented in the localities cited above.

As mentioned above in section 2.2.2, most groundwater extraction in the Taipei basin was from Aquifer 1, which is generally located about 50 to 100 m underground within the Jingmei Formation. We intend to call the direct resultant effects from post-pumping compaction and rebound of aquifer strata due to natural recharge of groundwater as the “intermediate” component, compared with the compaction of the surface soil above the Aquifer 1 and the possible crust-scale subsidence due to the normal faulting of the Shanchiao Fault. This component is expected to have basin-wide impact since the Aquifer 1 is distributed throughout the entire basin.

2.4.4 Synthesis and discussion

The mechanisms of land elevation change in the Taipei Basin during the past 30-year post-pumping period are summarized in Fig. 2-9, showing the effect from three major depth-related components in the central and western parts of the basin.

Soil compaction (the shallow component) is estimated with averaged rates of 3.5 mm/yr and 2 mm/yr for Central Taipei and the western margin, respectively (see Fig.

2-8, section 2.4.1.). The tectonic load from the Shanchiao Fault (the deep and crustal-scale component) contributes subsidence of approximately 1.75 mm/yr for the western margin of the basin and 0.9 mm/yr for Central Taipei (discussed in section 2.4.2.). As for aquifer deformation (the intermediate component), due to combination of compaction and rebound of natural recharge, its deformation rate is estimated by subtracting the soil compaction rate and the tectonic subsidence rate from the land elevation change rate. We thus obtain the evolution of the aquifer deformation rate in western and central Taipei basin, as the dashed curved in Fig. 2-9.

Fig. 2-9. Mechanisms of three depth-related component responsible for land elevation change in Western Taipei (Wuku) and Central Taipei. The shallow component represents the surfacial soil compaction with the average rates estimated after Lin et al. (1999). The intermediate component represents the aquifer deformation consists of compaction and elastic rebound within Aquifer 1. The deep and crustal-scale component represents asymmetric tectonic subsidence due to the activity of the Shanchiao fault in the western margin of the basin. Calculation of each component is explained in the text.

As revealed by Fig. 2-9, among the three major mechanisms, the aquifer deformation appears to be the main driving source of ground elevation changes, that is, subsidence in Phase 1 and uplift in Phase 2. Furthermore, the observed prominent land uplift in Phase 2, which is interpreted to be mainly related to the aquifer rebound, was significantly less in western margin than in Central Taipei.

In order to illuminate the actual extent of effects from aquifer deformation (compaction vs. rebound) and tectonic load, we sum up the ground level change in the Taipei area during Phase 2 (1989-2003, Fig. 2-10a) then remove estimate of shallow soil compaction to create the map of residual cumulative ground level change (Fig.

2-10b). The major remaining factors affecting land elevation fluctuation during Phase 2 are (1) elastic rebound of recharged Aquifer 1 contributing uplift and (2) tectonic subsidence. Note that the residual compaction of Aquifer 1 had declined to a neglected minimal rate as groundwater table almost fully recovered and stabilized.

Most of the Taipei Basin, except the western margin, showed rather homogeneous uplift of 13-18 cm (spatial variation) during 1989-2003 (or 9-13 mm/yr in average), and is interpreted as a manifestation of elastic rebound of Aquifer 1 due to natural recharge. The western part of the basin, in contrast, exhibited less amount of uplift (1-9 cm, or 0-6 mm/yr in average throughout the area) which decreased toward the west and northwest (also exhibited in Fig. 2-7). The “differential uplift” was partially resulted from variations of elastic rebound of Aquifer 1 possibly related to changes in sedimentary facies (hence the thickness and composition of the aquifer strata) and

在文檔中台北都會區山腳活斷層之地震地質研究 (頁 29-0)