Water Resour Manage (2016) 30:3861–3878 DOI 10.1007/s11269-016-1393-8
Evaluating the Relative Importance of Groundwater Recharge Sources in a Subtropical Alluvial Plain Using Tracer-Based Ternary End Member Mixing Analysis (EMMA) Tsung-Ren Peng 1 & Chun-Chun Huang 1 & Jui-Er Chen 2 & Wen-Jun Zhan 1 & Li-Wei Chiang 3 & Liang-Cheng Chang 4
Received: 16 March 2016 / Accepted: 6 June 2016 / Published online: 17 June 2016 # Springer Science+Business Media Dordrecht 2016
Abstract In Taiwan’s humid climate, proximal fan groundwater (PFG) is mainly sourced from local precipitation (LP), mountain front recharge (MFR), and mountain block recharge (MBR). This study evaluates the relative importance of the above sources’ respective contributions to the PFG of the Langyang alluvial plain (LAP), northeastern Taiwan. To this end, we first identify stable isotopic characteristics of these target waters and evaluate the hydrological relations among them. Further, we employ ternary end member mixing analysis (EMMA) based on δ18O and electrical conductivity to semi-quantitatively calculate contributing fractions and amounts of water for respective LP, MFR, and MBR end members. EMMA results indicate that the respective contribution fractions of LP, MFR, and MBR to PFG at the LAP are approximately 28, 60, and 12 %, respectively. Further, we employ the obtained contribution fractions to understand the corresponding water amounts of each end-member contributed to PFG. In total, 325 × 106 m3 of water recharges PFG annually; of which, 226 × 106 m3/yr. is from MFR, 76 × 106 m3/yr. from LP, and 23 × 106 m3/yr. from MBR. MFR is clearly the greatest source of water at the LAP and local water resource management and protection authorities should concentrate their energies on this important contributor to groundwater. To keep these results in context, limitations to the EMMA approach are evaluated in the text.
* Tsung-Ren Peng
[email protected]
1
Department of Soil and Environmental Sciences, National Chung Hsing University, Taichung 40227, Taiwan
2
Central Geological Survey, Ministry of Economic Affairs, New Taipei 23568, Taiwan
3
Green Energy and Environment Research Laboratories, Industrial Technology Research Institute, Hsinchu 31040, Taiwan
4
Department of Civil Engineering, National Chiao-Tung University, Hsinchu 30010, Taiwan
3862
T.-R. Peng et al.
Keywords Mountain front recharge (MFR) . Mountain block recharge (MBR) . End member mixing analysis (EMMA) . Groundwater recharge . Taiwan
1 Introduction Groundwater is one of humanity’s vital water resources. Evaluating recharge sources and available volumes of groundwater is therefore very important to proper water-resource management (Viviroli et al. 2011; Mas-Pla et al. 2013; Ebrahimi et al. 2016). Conventionally, mountains have been described as the water towers of the world for their contributions to runoff surface and groundwater in adjacent alluvial plains. The process by which mountains plentifully supply groundwater to adjacent alluvial plains is known as mountain system recharge (MSR) (e.g., Wilson and Guan 2004; Ajami et al. 2011). Understanding the linkage between mountain water sources and basin aquifers is essential for local water resource management (Smerdon et al. 2009; Beniston and Stoffel 2014). Generally, MSR in arid and semi-arid regions consists of two main components: mountain front recharge (MFR) and mountain block recharge (MBR) (e,g.,Wilson and Guan 2004; Ajami et al. 2011). MFR describes the process by which mountain stream runoff infiltrates alluvial fan aquifers via streambed sediments and MBR describes how precipitation infiltrates and percolates through mountain bedrock recharging alluvial fan aquifers via movement of deep groundwater. However, in humid climates, local precipitation (LP) is an important contributor to basin groundwater and needs to be included in groundwater recharge studies (Kao et al. 2012; Liu and Yamanaka 2012; Peng et al. 2014). Although MBR is an important issue in groundwater studies, reliable measurement of MBR’s contribution to basin aquifers is difficult as MBR is not easy to observe (Kambhammettu et al. 2011; Ping et al. 2014). Various physical, chemical, and numerical methods have been used to study MBR’s processes and contributions (Wilson and Guan 2004; Ajami et al. 2011). Each approach reveals a specific aspect of the MBR process with all study methods relying on mutual fundamental assumptions relating to meteoric, hydrologic, or geological parameters. These assumptions create uncertainties in MBR evaluations. In this study ternary end member mixing analysis (EMMA) is used based on chemical and isotopic tracers to partition the various contributions from LP, MFR, and MBR to groundwater in an alluvial plain. Using ternary tracer-based EMMA, Liu and Yamanaka (2012) and Peng et al. (2014) revealed spatial variations in the contribution fractions of LP, MFR, and MBR to the groundwaters of plains adjacent to mountain ranges. This case study is of Langyang alluvial plain (LAP), northeastern Taiwan (Fig. 1). Taiwan is an island located in the Western Pacific. It is 35,873 km2 in area with a subtropical to tropical climate. Estimated groundwater storage in montane and hill regions of elevation higher than 100 m above sea level (a.s.l.) covering 63 % of the island is approximately 12.5–16.5 × 109 m3. This volume is much greater than the estimated 4.5–5.8 × 109 m3 of groundwater stored in alluvial plains of elevation below 100 m a.s.l. (Water Resources Agency 2003), and plentiful mountain groundwater is thought to be stored in rock-fracture zones and contribute greatly to groundwater stored in adjacent alluvial plains areas. Consequently, this study evaluates the importance of mountain catchment water including MFR and MBR to proximal-fan groundwater through tracer-based ternary EMMA. First, stable isotopic characteristics of target waters are determined and then the hydrological relations among LP, MFR, and MBR’s contributions are evaluated. Next, EMMA is used to
Evaluating Relative Importance of Groundwater Sources Using EMMA
3863
Fig. 1 a Map showing the Langyang alluvial plain (LAP) and indicating collection sites of precipitation (LP), stream water (MFR), and groundwaters including mountain groundwater (MBR) and groundwater in the proximal-fan region (PFG). Additionally, two profiles are used to illustrate the hydrogeological characteristics of the LAP (Central Geological Survey 2013). b Map showing the flow nets in April and September, 2013, respectively, and indicating sub-regions LAF, NAF, and SAF in the LAP
semi-quantitatively calculate the respective contribution fractions and volumes of water attributable to end members: LP, MFR, and MBR.
2 Conceptual EMMA Model Using tracer-based EMMA to partition various potential sources contributing to a mixture is a popular approach in environmental studies (Clark and Fritz 1997; Bosley et al. 2002; Dor et al. 2011; Liu et al. 2011). In this study, proximal fan groundwater (PFG) is assumed mainly sourced from local precipitation (LP), mountain front recharge (MFR), and mountain block recharge (MBR). The ternary EMMA in terms of the δ value and elemental concentration (C) used to partition LP, MFR, and MBR contributions to PFG is expressed as follows: δPFG QPFG ¼ δLP QLP þ δMFR QMFR þ δMBR QMBR
ð1aÞ
C PFG QPFG ¼ C LP QLP þ C MFR QMFR þ C MBR QMBR
ð1bÞ
QPFG ¼ QLP þ QMFR þ QMBR
ð1cÞ
In Eq. (1), Q stands for water quantity. The subscript PFG represents groundwater in the proximal-fan aquifer; LP, MFR, and MBR denote waters from the respective contributions of LP, MFR, and MBR to PFG.
3864
T.-R. Peng et al.
Equation (1) can be rewritten in terms of fraction values and expressed as follows (Genereux 1998; Phillips and Gregg 2001): δPFG ¼ δLP RLP þ δMFR RMFR þ δMBR RMBR
ð2aÞ
C PFG ¼ C LP RLP þ C MFR RMFR þ C MBR RMBR
ð2bÞ
1 ¼ RLP þ RMFR þ RMBR
ð2cÞ
That is, the contribution fraction of each end-member is denoted by R: . RLP ¼ QLP QPFG
ð3aÞ
. RMFR ¼ QMFR QPFG
ð3bÞ
. RMBR ¼ QMBR QPFG
ð3cÞ
The solutions of RLP, RMFR, and RMBR derived from Eq. (2) are well known in the literature and expressed as follows (Maurya et al. 2011; Vanderzalm et al. 2011): h . . i RLP ¼ ðδMBR –δPFG Þ ðδMBR –δMFR Þ–ðC MBR –C PFG Þ ðC MBR –C MFR Þ ð4aÞ . . i .h ðC LP –C MBR Þ ðC MBR –C MFR Þ–ðδLP –δMBR Þ ðδMBR –δMFR Þ h . . i RMFR ¼ ðδMBR –δPFG Þ ðδMBR –δLP Þ–ðC MBR –C PFG Þ ðC MBR –C LP Þ . . i .h ðC MFR –C MBR Þ ðC MBR –C LP Þ–ðδMFR –δMBR Þ ðδMBR –δLP Þ
ð4bÞ
RMBR ¼ 1–RLP –RMFR
ð4cÞ
The EMMA applied is based on the following ideas (Barthold et al. 2011; Liu and Yamanaka 2012): (1) the product (PFG) results from well mixed reactants (LP, MFR, and MBR); (2) the mixing is a linear and complete process; (3) the substances employed as tracers are conservative; and (4) the source reactants have distinct concentrations in respect to various reactants. In this study, δ18O rather than δ2H is employed as the δ item in the studied EMMA because analytical precision is better for δ18O than δ2H (0.2 vs. 0.8 ‰, see Section 2.3). Less analytical precision leads to greater uncertainty in subsequent calculations due to error propagation (Berendsen 2011). Although Cl− is a suitable ion-tracer for the EMMA approach (Maurya et al. 2011; Liu and Yamanaka 2012), we use electronic conductivity (EC) as it can give more reasonable results than Cl− for C in EMMA. This is because preliminary checking led us to suspect that Cl− concentrations may be affected by anthropogenic pollution in the study region, meaning it is not well conserved. Moreover, many reports (e.g., Clark and Fritz 1997; Barth et al. 2006; Peng et al. 2014) indicate EC can be employed as a tracer to determine water sources, migration, and water mixing. Further, based on the fraction values obtained by Eq. (4), each water-amount (Q) item in Eq. (1) can be derived based on Eq. (3) if one of the Q items is known. In this study, QLP is
Evaluating Relative Importance of Groundwater Sources Using EMMA
3865
considered first as there are sufficient parameters for its derivation. QLP is derived from local climatic and hydrogeological measurements (Irmak et al. 2002; Snyder et al. 2005): QLP ¼ A PE
ð5aÞ
PE ¼ I C ðP–0:7 E A Þ
ð5bÞ
EA ¼ E K p
ð5cÞ
In Eq. (5), A is area of the study region; PE, effective precipitation; P, observed precipitation; IC, infiltration coefficient; EA, actual evaporation; E, pan evaporation; Kp, pan coefficient. According to Eq. (3), once QLP is obtained, QPFG can be derived based on Eq. (3a). Subsequently, QMFR and QMBR can also be derived based on Eq. (3b-c).
3 Materials and Methods 3.1 Study Area Langyang alluvial plain (LAP) is located in the northeastern Taiwan (Fig. 1a). Climatically, annual mean rainfall of LAP is approximately 2840 mm. Percentage wise, 38 % falls during summer (June to September), and 29 % in winter (November to February) (Central Weather Bureau 1981–2010). In geography, the region is bound by the Xueshan Range, Central Range, and Pacific Ocean to the northwest, southwest, and east, respectively. The Xueshan Range and Central Range are separated by the Langyang Stream. Geologically, the Central Range comprises alternating argillite, phyllite, and slate, and the Xueshan Range is lithologically characterized by argillite and metasandstone (Central Geological Survey 1995). Geomorphologically, the LAP is a confluent fan mainly formed by fluvial deposits of the Langyang Stream (Fig. 1a). Besides fluvial deposits from the Langyang Stream, the LAP receives fluvial sediments from catchment tributaries of the Xueshan Range and Central Range in the northern and southern parts of the LAP (Fig. 1a), respectively (Cheng 2010). Additionally, according to the flow net produced by equipotential lines (Fig. 1b), this study divides the proximal fans of the LAP into three sub-regions: the Langyang alluvial fan (LAF), northerncatchment alluvial fan (NAF), and southern-catchment alluvial fan (SAF) (Fig. 1b). The respective areas of the LAF, NAF, and SAF are 120, 86, and 61 km2 and average catchment elevations are approximately 894, 364, and 274 m a.s.l. (Table 1). Hydrogeologically, the LAP is composed of Quaternary deposits with texture specified by alternating coarse sand/gravel and fine silt/clay (Fig. 1a; Central Geological Survey 2013). In this study, the border that divides the proximal- and mid-fan is based on the matrix profiles indicated in Fig. 1a and the isoline of apparent resistivity at 100 Ω-m detected by geoelectrical resistivity (Fig. 1b; Central Geological Survey 2013). In the proximal-fan region, resistivity is about 100 Ω-m; additionally, according to the texture profiles of the LAP (Fig. 1a), no significant impermeable (silt/clay) layer presents in the proximal-fan region. Therefore, the proximal-fan region of the studied LAP is a phreatic (unconfined) aquifer, any surface water such as precipitation or nearby stream water recharging the shallow groundwater can easily percolate into the deep aquifer. By comparison, resistivity is lower than 100 Ω-m in the midfan region and hydrogeologically the area is a semi-confined aquifer characterized by silt/clay
Altitude (m.a.s.l.)
(μS/cm)
(‰)
(‰)
(‰)
(‰)
δ2H
δ18O
δ2H
δ18O EC
July 2013
March 2013
(μS/cm)
EC
214
137 280
185
LS-2
LS-3 LS-4
LS-5
−46.8
−46.2
−46.7 −36.0
−31.4
−8.0
−7.7
−7.9 −6.7
−6.3 273
544 376
508
274
−6.3
−8.5 −6.6
−8.3
−7.8
−31.6
−52.7 −33.2
−52.3
−46.4
282
548 296
553
241
−5.7
−7.2 −6.3
−7.8
−7.9
(‰)
δ18O
476
NS-3
−34.3
−34.2
−6.2
−6.2 108
131 −6.3
−6.2 −34.7
−34.3 104
126
−6.1
−6.5
35
SS-2
Mean
12
SS-1
−26.9
−24.9
−5.7
−5.5 206
166 −5.6
−5.4 −23.4
−26.0
236
132
−4.9
−5.3
Mean Southern-catchment alluvial fan (SAF) (with catchment area of 61 km2 and average elevation of 274 m a.s.l.)
175
NS-2
−24.7
−26.1
−33.8
−35.6
−37.7
−31.0
−41.6 −34.1
−46.5
−46.3
(‰)
δ2H
November 2013
Northern-catchment alluvial fan (NAF) (with catchment area of 86 km2 and average elevation of 364 m a.s.l.) NS-1 195 −6.4 −36.0 164 −6.5 −35.5 158 −6.5
Mean
295
LS-1
Langyang alluvial fan (LAF) (with catchment area of 120 km2 and average elevation of 894 m a.s.l.)
Stream sampling site
313
77
90
94
114
190
337 294
427
203
(μS/cm)
EC
−24.4 ± 0.8 −25.3 ± 1.2
−5.4 ± 0.4 −5.4 ± 0.3
−26.3 ± 0.5
−35.1 ± 1.2
−6.3 ± 0.2 −5.5 ± 0.2
−34.2 ± 0.5
−34.8 ± 0.7 −6.2 ± 0.1
−6.3 ± 0.2
−36.4 ± 1.1
−41.5 ± 7.8 −6.5 ± 0.0
−31.3 ± 0.3 −7.3 ± 0.9
−47.0 ± 5.6 −34.4 ± 1.5
−48.3 ± 3.4
−46.5 ± 0.2
(‰)
δ2H
−6.1 ± 0.3
−7.9 ± 0.7 −6.5 ± 0.2
−7.9 ± 0.3
−7.9 ± 0.1
(‰)
δ18O
Mean
188 ± 83
252 ± 55
125 ± 45
121 ± 26
101 ± 9
117 ± 20
145 ± 27
356 ± 128
248 ± 51
476 ± 121 322 ± 47
496 ± 64
239 ± 36
(μS/cm)
EC
Table 1 Background of respective proximal-fan regions, stable isotopic composition, and EC concentration of stream water (MFR water) in respective proximal-fan regions
3866 T.-R. Peng et al.
Evaluating Relative Importance of Groundwater Sources Using EMMA
3867
deposits. Additionally, the infiltration coefficient is estimated at 0.15 for the LAF and 0.10 for both the NAF and SAF (Water Resources Agency 2003).
3.2 Water Samples Water samples including precipitation, stream water, and groundwater of the LAF, NAF, and SAF regions were collected for isotopic determination. The sampling locations are indicated in Fig. 1 and described as follows: (1) Local precipitation (LP): The LP samples in the LAP were collected at one site on each rainy day from January 2013 to December 2014. (2) Stream water in the mountain front region (Mountain Front Recharge (MFR)): The MFR samples were collected at 5 sites, LS-1 to LS-5, in the LAF, 3 sites, NS-1 to NS-3, in the NAF, and 2 sites, SS-1 to SS-2, in the SAF (Table 1). (3) Mountain groundwater recharging proximal-fan groundwater (Mountain Block Recharge (MBR)): These waters include samples from northern mountain groundwater (MGN) and southern mountain groundwater (MGS). The MGN data are seasonal measurements from 9 wells in the piedmont region along the Xueshan Range, and the MGS samples were seasonally collected at a hot spring pouring out in a thermal field of the Central Range. Both necessary isotopic and EC data of the MGN and MGS applied in this study are taken from unpublished data in our laboratory (see Appendix). (4) Proximal-fan groundwater (PFG): Groundwater samples in the proximal fan were collected from 14 wells belonging to 10 stations. Geographically, stations LG-1 to LG-3 are in the LAF; NG-1 to NG-4 in the NAF, and SG-1 to SG-3 in the SAF (Table 2).
3.3 Analyses of Stable Hydrogen and Oxygen Isotopes All collected water samples were stored in well-sealed plastic bottles of 100 mL in size until isotopic analyses could be made. Water samples were analyzed with a Liquid-Water Isotope Analyzer (LWIA) produced by Los Gatos Research, USA. The LWIA is based on off-axis integrated cavity output spectroscopy (OA-ICOS) lasers and connected with a CTC LC-PAL liquid auto-sampler (LAS) to measure both 2H/1H and 18O/16O ratios directly and concurrently for water molecules (H2O) from a single run (Lis et al. 2008; Peng et al. 2014). All the 2H/1H and 18O/16O results are reported in δ-notation (‰) relative to the international Vienna Standard Mean Ocean Water (VSMOW). In this study, the long-term analytical precisions (1 sigma) of isotope analyses for laboratory water standards are ±0.8 ‰ for δ2H and ±0.2 ‰ for δ18O.
4 Results 4.1 Local Precipitation (LP) On average, the year-round precipitation-weighted mean δ values of LP samples are −4.5 ± 2.7 ‰ and −21.5 ± 26.7 ‰ for δ18O and δ2H, respectively. As shown in Fig. 2a, the δ values of precipitation vary widely and exhibit distinct seasonal variations with relatively high δ values in winter and low values in summer. These phenomena are due to monsoons effect (Peng et al. 2010). The precipitation-weighted mean δ values of winter months (November–February) are −3.3 ± 0.4 ‰ and −2.8 ± 5.0 ‰ for δ18O and δ2H, respectively, and
3868
T.-R. Peng et al.
Table 2 Stable isotopic composition, EC concentrations of proximal-fan groundwater (PFG), and fraction results indicative of relative contributions among LP, MFR, and MBR to the PFG of respective proximal-fan regions Well*/PFG of respective proximal fan
Screen depth (m)
April 2013
September 2013
Mean
δ18O δ2H
δ18O δ2H
δ18O δ2H
EC
(‰)
(μS/ (‰) cm)
(‰)
(‰)
EC
(μS/ (‰) (‰) cm)
Contribution fractions EC
RLP RMFR RMBR
(μS/ (%) (%) cm)
(%)
Langyang alluvial fan (LAF) LG-1
26–44
−7.3 −43.0 444
−7.2 −41.5 514
−7.3 −42.2 479
0
96
4
LG-2
7–49
−6.5 −36.5 177
−6.4 −36.0 176
−6.4 −36.3 177
31
69
0
LG-3S
38–62
−6.3 −34.6 256
−6.7 −38.0 295
−6.5 −36.3 276
27
73
0
LG-3D
136–160 −6.6 −36.3 412
−6.7 −36.6 432
−6.7 −36.5 422
19
77
4
−6.7 −37.8 338
19
79
2
Mean Northern-catchment alluvial fan (NAF) NG-1 41–50 −6.2 −32.6 311 NG-2S 31–90 −5.8 −30.5 369
−6.0 −34.0 377
−6.1 −33.3 344
44
26
30
−5.8 −31.6 385
−5.8 −31.0 377
65
0
35
NG-2D
138–156 −6.0 −32.4 222
−6.1 −32.0 231
−6.1 −32.2 227
30
54
15
NG-3
12–14
−6.5 −36.0 204
−6.5 −36.0 204
3
87
10
NG-4
7–19
−5.9 −32.6 255
−6.1 −34.2 263
34
46
20
−6.1 −33.1 283
35
43
22
−5.4 −26.4 283 −5.2 −26.1 296 −5.3 −25.1 391
0 29
97 67
3 5
−6.2 −35.9 270
Mean Southern-catchment alluvial fan (SAF) SG-1S SG-1D SG-2S
28–49 68–86 13–37
−5.5 −24.9 279 −5.1 −25.6 291 −5.2 −24.7 384
15
78
7
SG-2D
60–78
−5.3 −24.0 1274 −5.3 −23.4 1298 −5.3 −23.7 1286 13
53
34
SG-3
6–27
−4.7 −20.7 327
Mean
−5.4 −27.8 287 −5.2 −26.6 300 −5.3 −25.5 398 −4.5 −21.1 345
−4.6 −20.9 336
91
0
9
−5.1 −24.4 518
30
58
12
* Each station is generally constructed with one well except for LG-3, NG-2, SG-1, and SG-2, which are equipped with two wells of different screen depths. Note, based on well-screen depths, shallow stationgroundwaters are labeled BS″ while deep-station groundwaters are labeled BD^
those of summer months (June–September) are −7.7 ± 2.0 ‰ and −53.9 ± 17.5 ‰ for δ18O and δ2H, respectively. Moreover, those mean δ values for the transition months between summer and winter are −2.5 ± 1.3 ‰ and −7.8 ± 14.0 ‰ for δ18O and δ2H, respectively. Additionally, the year-round precipitation-weighted local meteoric water line (LMWL) is δ2H = (9.6 ± 0.5) δ18O + (21.9 ± 2.5) (Fig. 2).
4.2 Stream Water in the Mountain Front Region (MFR Water) δ values of stream waters for respective sampling months and yearly means are listed in Table 1. All δ data of the studied stream waters distribute along with LMWL of the LAP as per Fig. 2b, indicating that all the MFR waters originate from precipitation and their δ values do not indicate water–rock interactions and extensive water evaporation. Obviously, δ values of MFR samples in the LAF are more depleted in 18O and 2H isotopes than those of the NAF and SAF (Table 1). For example, the yearly mean
Evaluating Relative Importance of Groundwater Sources Using EMMA 100
(a)
LMWL Summer months Winter months Transition months
80 60 40 20 0
δ 2H ( ‰ )
Fig. 2 Plot of δ2H vs. δ18O of water samples of this study. a Seasonal precipitation samples. The year-round local meteoric water line (LMWL) is also indicated in the figure. b Water samples of MFR (including LS, NS, SS), MBR (including MGN, MSN), and PFG (including LG, NG, SG)
3869
-20 -40 -60 -80 -100 -120 -140 -20
-15
-10
-5
0
5
δ 18O (‰ ) 60 40 X
δ 2H ( ‰ )
20 0 -20
(b)
LMWL SMOW Summer precipitation Winter precipitation LS NS SS MGN MGS LG NG SG
X
-40 -60 -80 -13
-11
-9
-7
-5
-3
-1
1
δ 18O ( ‰)
δ18O value of MFR samples in the LAF is −7.3 ± 0.9 ‰, and those mean values for MFR samples of NAF and SAF are −6.3 ± 0.2 and −5.4 ± 0.3 ‰, respectively. The discrepancy in stream water δ values from river to river reflects the difference in altitude of their respective sources. As indicated in Table 1, average elevation is approximately 894 m a.s.l. for catchments of the LAF, and 364 and 274 m a.s.l. for catchments of the NAF and SAF, respectively. Besides spatial variation, seasonal differences in δ values are trivial for stream water in each proximal-fan region (Table 1). For example, in the LAF, the mean δ18O value is −7.5 ‰ (range from −8.5 to −6.3 ‰) for summer samples, and −7.0 ‰ (range from −7.9 to −5.7 ‰) for winter samples. The range of δ18O values of summer stream water overlaps the range in values of winter stream water.
4.3 Mountainous Groundwater Recharging Proximal-Fan Groundwater (MBR Water) The MBR of this study includes MGS and MGN. Their isotopic and EC data applied are taken from unpublished data of our laboratory and listed in the Appendix. The mean δ values of
3870
T.-R. Peng et al.
MGN are −8.3 ‰ for δ18O (ranging from −8.6 to −8.0 ‰) and −48.9 ‰ (ranging from −55.6 to −40.6 ‰) for δ2H. On the other hand, the mean δ values of MGS are −5.4 ‰ for δ18O (ranging from −6.0 to −5.0 ‰) and −44.2 ‰ (ranging from −47.9 to −34.8 ‰) for δ2H. Mean δ data of MGN waters locate along with the LMWL of the LAP (Fig. 2b), indicating that the studied MGN waters originate from precipitation and their δ values do not indicate water–rock interactions nor affects due to seawater. By comparison, mean δ data of the MGS waters deviate from the LMWL of the LAP (Fig. 2b), implying that the studied MGS waters experienced a certain degree of water–rock interaction or were affected by thermal water (Panichi and Gonfiantini 1981; Liu et al. 1990).
4.4 Proximal-Fan Groundwater (PFG) The mean δ values of the PFG samples in the LAF are −6.7 ‰ for δ18O and −37.8 ‰ for δ2H (Table 2). Those isotopic values in the NAF are −6.1 ‰ for δ18O and −33.4 ‰ for δ2H; and in the SAF, they are −5.1 ‰ for δ18O and −24.4 ‰ for δ2H. In general, seasonal differences in δ values of each well are minor; the δ differences are smaller than 0.4 ‰ for δ18O and 3.4 ‰ for δ2H (Table 2). Furthermore, the differences in δ values between shallow and deep groundwater are minor. For example, in the LAF, the δ18O values of LG-3S is −6.5 ‰, similar to −6.7 ‰ of LG-3D. In the NAF, the δ18O values of NG-2S is −5.8 ‰, similar to −6.1 ‰ of NG-2D, and in the SAF the δ18O values of SG-1S is −5.4 ‰, similar to −5.2 ‰ of SG-1D, and the δ18O values of SG2S is −5.3 ‰, identical to −5.3 ‰ of SG-2D (Table 2). The small differences in δ values between seasons and depths indicate that the studied PFG is well-mixed water from its various sources.
4.5 EC Concentration of the Studied Water Averaged EC concentrations of the various waters of this study are indicated in Fig. 3. On the whole, the LP samples have lower observable EC concentration (approximately 37 μS/cm) than the MFR, MBR, and PFG waters. The MBR waters including MGS and MGN exhibit higher EC concentrations than any other studied waters. The mean EC value is approximately 3435 μS/cm for the MGS water and 971 μS/cm for the MGN water (Fig. 3 and see Appendix). High EC values for the MBR waters imply that they experienced a greater degree of water– rock interaction than other waters did (Appelo and Postma 1996; Barth et al. 2006). This reaction is also evidenced by the δ data of the MGS waters deviating from the LMWL (Fig. 2b). In general, EC concentration is lower in the MFR water than in PFG water (Fig. 3). In the LAF region, the mean EC concentrations between the MFR and PFG are similar, approximately 356 vs. 338 μS/cm (Fig. 3a and Tables 1–2). This similarity implies that high hydrological connectivity exists between MFR and PFG in the LAF. In the NAF region, the mean EC concentration of MFR is approximately 121 μS/cm (Fig. 3b and Table 1), which is smaller than the 283 μS/cm of the PFG (Table 2); further, in the SAF region, the mean EC concentration of MFR is approximately 188 μS/cm, smaller than the 518 μS/cm of the PFG (Fig. 3c and Tables 1–2).
Evaluating Relative Importance of Groundwater Sources Using EMMA 4000
(%)
(a) LAF
MBR(-5.4, 3435)
19 LP
3000
EC (μS/cm)
Fig. 3 Plot of δ18O vs. EC of the PFG in respective proximal-fan regions of LAF (a), NAF (b), and SAF (c). The plots also illustrate the relative contributions of three end-source waters LS, MFR, and MBR for each PFG in respective proximal-fan regions. The δ18O value (left) and EC concentration (right) of each end-member are indicated
3871
2000
79 MFR
1000
0 1200
EC (μS/cm)
1000
(-4.5, 37) LP
(-7.3, 356) MFR -9
-8
-7
-6
-5
-4
2
MBR
(%)
(b) NAF
35 LP
MBR (-8.3, 971)
800
43 MFR
600 400 200 0 4000
(-4.5, 37) 22 MBR LP
(-6.3, 121) MFR -9
-8
(c) SAF
-7
-6
-5
-4
(%)
MBR (-5.4, 3435)
30 LP
EC (μS/cm)
3000
2000
58 MFR
1000
0
(-5.4, 188) MFR -7
-6
12 MBR
LP(-4.5, 37) -5
-4
δ 18O ( ‰ )
-3
-2
5 Discussion 5.1 Verifying Sources of Groundwater in the Proximal-Fan Region Local precipitation (LP) is the most likely source of PFG. According to Fig. 1a, no impermeable layer is evident below the ground surface in the proximal region; therefore, precipitation or surface water can easily infiltrate and percolate into the aquifer. The contribution of LP to PFG can be evidenced by groundwater-level records (Fig. 4). For example, in the LAF region, groundwater levels of all studied wells vary significantly with the amount of precipitation
3872
T.-R. Peng et al.
1000
65 60
1500 55
Monthly rainfall Level of LG-1
Groundwater level (m a.s.l.)
16
2000 2010
2011
2012
2013
(b) NAF
0
14
500
12 1000 10 1500
8
2000 2010
2011
2012
2013
35
500
30
1000
25
1500
Monthly rainfall Level of LG-3S Level of LG-3D
2000 2010
2011
2012
2013
2014
10
Monthly rainfall Level of NG-1
6 2009
0
20 2009
2014
Precipitation (mm)
50 2009
40
8
500
6 1000 4 1500
2 0 2009
2014
0
Precipitation (mm)
500 70
Groundwater level (m a.s.l.)
Groundwater level (m a.s.l.)
75
Precipitation (mm) Groundwater level (m a.s.l.)
0
Precipitation (mm)
(a) LAF 80
Monthly rainfall Level of NG-2S Level of NG-2D
2000 2010
2011
2012
2013
2014
(c) SAF
500
4 1000 2 1500
0 Monthly rainfall Level of SG-3
-2 2009
2000 2010
2011
2012
2013
2014
Precipitation (mm)
6
10
0
8
500
6 1000 4 1500
2 0 2009
Precipitation (mm)
0
Groundwater level (m a.s.l.)
Groundwater level (m a.s.l.)
8
Monthly rainfall Level of SG-1S Level of SG-1D
2000 2010
2011
2012
2013
2014
Fig. 4 Hydrographs with monthly records of precipitation and groundwater tables for selected well stations in respective proximal-fan regions. a Wells LG-1 and LG-3 in the LAF; b Wells NG-1 and NG-2 in the NAF; and c Wells SG-1 and SG-3 in the SAF
(Fig. 4a). The fluctuating relationship between rainfall and groundwater table of LG-1 indicates that precipitation is an obvious contributor to groundwater. This precipitationinduced fluctuation also happens in the deep groundwater as evidenced by water table records of LG-3. Similarly, such precipitation-induced fluctuations also occurred at well stations in the NAF and SAF regions (Panels b–c in Fig. 4). Furthermore, groundwater level hydrographs show a constant baseline during low or no rain periods (Fig. 4). This evidence suggests that besides the LP, MFR and/or MBR are important contributors to groundwater of the proximal-fan region. MFR indicates infiltration of mountain stream runoff in alluvial fan aquifers. According to groundwater-level records (Fig. 4), water tables of the PFGs range from 26 to 80 m.a.s.l. in the LAF, 2–12 m.a.s.l. in the NAF, and 0–6 m.a.s.l. in the SAF. By comparison, the elevations of floor-of-stream water sampling sites are about 222, 282, and 24 m.a.s.l. in the LAF, NAF, and SAF, respectively (Table 1). Comparisons of elevations between stream-floor and groundwater
Evaluating Relative Importance of Groundwater Sources Using EMMA
3873
tables demonstrate that stream water may recharge the PFG (Banks et al. 2011; Brunner et al. 2011). Although MBR is difficult to observe (Wilson and Guan 2004), the plot of mean δ18O values vs. EC suggests MBR water is a contributor to PFG (Fig. 3). In Fig. 3a, data points for all PFG samples in the LAF are plotted within a triangle with end members of LP, MFR, and MBR. A similar distribution giving PFG data located within the three-end-member triangle also occurs for the NAF and SAF regions (Panels b–c in Fig. 3). This distribution feature suggests that besides LP and MFR water, water sources that recharge the PFG include MBR water. The δ18O and EC values of end members for LAF, NAF, and SAF regions are indicated in Fig. 3. In short, the annual mean δ18O and EC values are commonly used for end member LP of LAF, NAF, and SAF regions. The δ18O and EC values of MGS are used to represent the end member MBR in the LAF and NAF regions and those values of MGN are used to represent the end member MBR in the NAF. The δ18O and EC values of MFR in respective regions are the mean values of stream waters for the corresponding regions.
5.2 Relative Importance of Respective Groundwater Sources in an Alluvial Aquifer The relative contributions to PFG among the three end sources of LP, MFR, and MBR can be semi-quantitatively estimated using ternary EMMA as in Eq. (4). Based on the δ18O and EC values of the related three end members in the proximal fan (Fig. 3), the semi-quantitative fractions of LP, MFR, and MBR contributing to PFG are derived and listed in Table 2. Briefly, the mean RLP values are approximately 19 %, 35 %, and 30 % for LAF, NAF, and SAF, respectively; the mean RMFR values, approximately 79 %, 43 %, and 58 % for LAF, NAF, and SAF, respectively; and the mean RMBR values, approximately 2 %, 22 %, and 12 % for LAF, NAF, and SAF, respectively. Obviously, MFR water is the most important contributor to PFG in each proximal-fan region, especially in the LAF region. The notable infiltration of MFR into PFG is most likely due to no impermeable layer being evident below the ground surface in the proximal region (Fig. 1b). The RMFR of LAF is approximately 80 %, greater than about 50 % for SAF/NAF region. This difference in RMFR can be attributed to the catchment size of LAF being larger than that of the SAF/NAF region (Table 1). The larger the catchment size is, the greater the discharge provided. However, in the LAF region the significance of MFR may restrict the contributions of LP and MBR, causing RLP or RMBR to be smaller than in SAF or NAF. Further, based on the contribution fractions obtained, water volumes of each end member to PFG can be derived if QLP in Eq. (3) is obtained. The QLP can be derived according to Eq. (5). The values of required parameters in Eq. (5) for respective LAF, SAF, and NAF regions are listed in Table 3. In light of the QLP and contribution-fraction values in Eq. (3), water volumes of QMFR and QMBR for MFR- and MBR-recharge contributions to groundwater in the proximal fan at LAF, SAF, and NAF are further derived and listed in Table 3. Overall, 325 × 106 m3/yr. of water is recharged into groundwater in the proximal fan of LAP (Table 3). The 325 × 106 m3/yr. estimated by this study is of the same order as estimates ranging from 237 to 304 × 106 m3/yr. derived by water balance or numerical approaches of previous studies (Central Geological Survey 2013). Of the total recharge (325 × 106 m3/yr), 226 × 106 m3/yr. is from MFR, 76 × 106 from LP, and 23 × 106 m3/yr. from MBR (Table 3).
2838
SAF
893
893 893
Area (A)
61
120 86
(km2)
0.10
0.15**** 0.10
Infiltration coefficient (Ic)
0.80
0.80***** 0.80
Pan coefficient (Kp)
30
19 35
(%)
RLP RMFR
58
79 43
(%)
RMBR
12
2 22
(%)
QLP**
14
42 20
(106 m3/yr)
***** A pan coefficient (Kp) value of 0.8 is set for each region LAF, NAF, and SAF (Chang 1995)
**** An infiltration coefficient (IC) value of 0.15 is set for LAF, and 0.10 is set for both NAF and SAF (Water Resources Agency 2003)
*** QPFG = QLP /RLP ; QMFR = QPFG × RMFR; QMBR = QPFG × RMBR (Refer to Eq. (3))
** QLP = A× IC × [P – 0.7 (E× Kp)] (Refer to Eq. (5))
* Long-term yearly mean precipitation (P) and evaporation (E) data from Central Weather Bureau (1981–2010)
2838 2838
Evaporation* (E)
(mm/yr)
Precipitation* (P)
(mm/yr)
LAF NAF
Regions
28
173 25
(106 m3/yr)
QMFR***
6
4 13
(106 m3/yr)
QMBR***
48
219 58
(106 m3/yr)
QPFG***
Table 3 The water amounts of respective end-sources of LP, MFR, and MBR contributed to the PFG of respective proximal-fan regions, including the required parameter values that are used to estimate
3874 T.-R. Peng et al.
Evaluating Relative Importance of Groundwater Sources Using EMMA
3875
Despite the encouraging results, it is recognized that this approach has its limitations and it is still possible to improve this tracer-based evaluation. First, the calculation fractions of RLP, RMFR, and RMBR derived by our tracer model are estimated by a ternary EMMA in terms of δ18O and EC values (Eq. 2). Because the error in the derived fraction is hard to precisely estimate due to complicated statistical error propagation (Berendsen 2011; Maurya et al. 2011; Peng et al. 2011), the derived contribution fractions of this study are deemed semi-quantitative estimates. Calculating the precise uncertainty of derived fractions is difficult to accomplish. Next, more reliable estimation of recharge volumes while using Eq. 5 depends on the precision of the parameters measured. Determination of parameters relies more on educated guesswork than proven factual information. For example, it is difficult to precisely define the border of each proximal-fan region (Central Geological Survey 2013; Bowen et al. 2014). In this study, according to Eq. 5, the error in recharge volume estimations on an annual scale induced by deviations in area estimates of proximal-fan regions is approximately 1 × 106 m3/ km2. The errors for other parameters are: 1 × 106 m3/10 mm for precipitation; −0.6 × 106 m3/ 10 mm for evaporation; 20 × 106 m3/ 0.01 for infiltration coefficient; and −7 × 106 m3/0.1 for pan coefficient. All those limitations described above raise the degree of error in our tracer model.
6 Conclusion This study used ternary EMMA based on δ18O and EC concentrations to evaluate the relative importance of respective contributions of LP, MFR, and MBR to PFG of the LAP, northeastern Taiwan. The semi-quantitative results derived from EMMA indicate that the average contribution fractions of LP, MFR, and MBR to PFG are approximately 28, 60, and 12 %, respectively. These fractional results clearly indicate that mountain catchment water, including MFR and MBR, contributes most to groundwater storage in the adjacent alluvial basin, and most of this recharge water is from MFR. The finding is very significant in that it provides a good understanding of the hydrological environment of the studied LAP. In terms of local hydrological implications, local water resource management and protection should concentrate on the MFR; that is, surface water in the proximal-fan. However, tracer-based EMMA has its limitations. First, errors in contribution fractions of LP, MFR, and MBR derived by the used EMMA in terms of δ18O and EC values are difficult to precisely estimate due to complicated statistical error propagation. Thus, derived contribution fractions are only deemed semi-quantitative estimates. Next, a more reliable estimation of recharge volumes requires more precise estimation of parameters. The above limitations raise the degree of error in our tracer model. The results of this study provide new insight into how mountain catchment water recharges adjacent basin groundwater and benefits water-resource management in the studied region. The results and experience gathered from this study will be of great interest to similar future studies elsewhere. Acknowledgments The authors are very grateful to anonymous reviewers and the Associate Editor for their constructive comments, which greatly improved our manuscript. Work on this paper is divided into two parts. That focusing on meteoric water is an achievement attributable to assistance from the National Science Council, Taiwan (NSC 101-2116-M-005-001 and NSC 102-2116-M-005-001) and that regarding groundwater is attributable to assistance from the Central Geological Survey, Ministry of Economic Affairs, Taiwan (B10249).
3876
T.-R. Peng et al.
Appendix Table listing stable isotopic composition and EC concentration of mountain groundwaters (MBR water) represented by MGN and MGS waters of this study MBR water
MGN
MGS
Sample site
Well-1 Well-2 Well-3 Well-4 Well-5 Well-1 Well-2 Well-3 Well-4 Well-5 Well-6 Well-7 Well-8 Well-1 Well-2 Well-3 Well-4 Well-5 Well-6 Well-7 Well-8 Well-1 Well-2 Well-3 Well-4 Well-5 Well-6 Well-7 Well-8 Well-9 Mean Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Spring-1 Mean
δ18O
δ2H
EC
(‰)
(‰)
(μS/cm)
−8.5 −8.4 −8.4 −8.5 −8.2 −8.6 −8.5 −8.4 −8.3 −8.2 −8.1 −8.3 −8.3 −8.2 −8.2 −8.4 −8.5 −8.3 −8.3 −8.2 −8.2 −8.0 −8.3 −8.2 −8.2 −8.0 −8.3 −8.0 −8.2 −8.2 −8.3 −6.0 −5.4 −5.8 −5.4 −5.2 −5.1 −5.3 −5.3 −5.3 −5.4 −5.0 −5.6 −5.4
−50.9 −50.1 −51.8 −50.0 −48.7 −49.4 −51.4 −49.8 −49.0 −49.5 −50.2 −47.5 −48.3 −49.4 −54.0 −49.1 −42.7 −55.6 −49.9 −49.1 −52.8 −43.1 −48.5 −43.1 −50.1 −40.6 −44.3 −46.9 −50.1 −49.8 −48.9 −40.0 −47.9 −46.5 −45.7 −45.8 −42.1 −41.7 −46.4 −34.8 −47.6 −46.6 −45.8 −44.2
962 948 952 943 928 951 1005 922 953 975 973 917
882 939 964 993 947 976 924 973 1354 971
3710 3160 3435
Sampling date
August 2006 August 2006 August 2006 August 2006 August 2006 November 2006 November 2006 November 2006 November 2006 November 2006 November 2006 November 2006 November 2006 March 2007 March 2007 March 2007 March 2007 March 2007 March 2007 March 2007 March 2007 July 2007 July 2007 July 2007 July 2007 July 2007 July 2007 July 2007 July 2007 July 2013 December 1991 March 1992 July 1992 November 1992 February 1993 May 1993 August 1993 November 1993 January 1994 February 1994 July 2013 November 2013
Evaluating Relative Importance of Groundwater Sources Using EMMA
3877
References Ajami HA, Troch PA, Maddock IIIT, Meixner T, Eastoe C (2011) Quantifying mountain block recharge by means of catchment-scale storage-discharge relationships. Water Resour Res 47:W04504. doi:10.1029/ 2010WR009598 Appelo CAJ, Postma D (1996) Geochemistry, groundwater and pollution. Balkema, Rotterdam, pp. 142–174 Banks EW, Simmons CT, Love AJ, Shand P (2011) Assessing spatial and temporal connectivity between surface water and groundwater in a regional catchment: implication for regional scale water quantity and quality. J Hydrol 404:93–102. doi:10.1016/j.jhydrol.2011.04.017 Barth JAC, Stichler W, Bergemann M, Reincke H (2006) Can conductivity and stable isotope tracers determine water sources during flooding? An example from the Elbe River in 2002. Int J River Basin Manag 4(2):77– 83. doi:10.1080/15715124.2006.9635278 Barthold FK, Tyralla C, Schneider K, Vache KB, Frede H, Breuer L (2011) How many tracers do we need for end member mixing analysis (EMMA)? A sensitivity analysis. Water Resour Res 47:W08519. doi:10.1029/ 2011WR010604 Beniston M, Stoffel M (2014) Assessing the impacts of climatic change on mountain water resources. Sc Total Environ 493:1129–1137. doi:10.1016/j.scitotenv.2013.11.122 Berendsen HJC (2011) Errors: classification and propagation. In: A Student’s guide to data and error analysis. Cambridge University Press, New York, pp. 18–26. doi:10.1017/CBO9780511921247.004 Bosley KL, Witting DA, Chambers R, Wainright SC (2002) Estimating turnover rates of carbon and nitrogen in recently metamorphosed winter flounder (Pseudopleuronectes americanus) with stable isotopes. Mar Ecol Prog 236:233–240. doi:10.3354/meps236233 Bowen EE, Hamada Y, O’Connor BL (2014) Mapping mountain front recharge areas in arid watersheds based on a digital elevation model and land cover types. J Water Resour Protec 6:756–771. doi:10.4236/jwarp.2014. 68072 Brunner P, Cook PG, Simmons CT (2011) Disconnected surface water and groundwater: from theory to practice. Ground Water 49:460–467. doi:10.1111/j.1745-6584.2010.00752.x Central Geological Survey (1995) Explanatory text of the geologic map of Sanshin (1:50,000). Ministry of Economic Affairs, Taipei, Taiwan Central Geological Survey (2013) Hydrogeological investigation and groundwater recharge model simulations: recharge areas delineation and groundwater resources assessment (1/4). Ministry of Economic Affairs, Taipei, Taiwan Central Weather Bureau (1981–2010) Monthly Precipitation – Ilan. http://www.cwb.gov.tw/V7/climate/ monthlyMean/Taiwan precp.htm ; accessed on 1 June 2015. Chang CC (1995) Groundwater study of the Ilan area. PhD Thesis of National Taiwan Normal University, Taipei, Taiwan, pp. 259. Cheng YC (2010) The paleoenvironment analysis of the Ilan plain. Master’s Thesis of National Taiwan Ocean University, Keelung, Taiwan, p. 75 Clark ID, Fritz P (1997) Tracing the hydrological cycle. In: Environmental isotopes in hydrogeology. CRC Press, Florida, pp. 35–60 Dor N, Syafalni S, Abustan I, Rahman MTA, Nazri MAA, Mostafa R, Mejus L (2011) Verification of surfacegroundwater connectivity in an irrigation canal using geophysical, water balance and stable isotope approaches. Water Resour Manage 25:2837–2853. doi:10.1007/s11269-011-9841-y Ebrahimi H, Ghazavi R, Karimi H (2016) Estimation of groundwater recharge from the rainfall and irrigation in an arid environment using inverse modeling approach and RS. Water Resour Manage 30:1939–1951. doi:10. 1007/s11269-016-1216-6 Genereux D (1998) Quantifying uncertainty in tracer-based hydrograph separations. Water Resour Res 34(4): 915–919. doi:10.1029/98WR00010 Irmak S, Haman DZ, Jones JW (2002) Evaluation of class a pan coefficients for estimating reference evapotranspiration in humid location. J Irrig Drain Eng 128(3):249–253. doi:10.1061/(ASCE)0733-9437(2002) 128:3(153) Kambhammettu BVNP, King JP, Allena P (2011) Evaluation of mountain-front recharge estimation techniques for southern New Mexico basins. Int J Water Res Environ Eng 3(3):66–72 Kao YH, Liu CW, Wang SW, Lee CH (2012) Estimating mountain block recharge to downstream alluvial aquifers from standard methods. J Hydrol 426–427:93–102. doi:10.1016/j.jhydrol.2012.01.016 Lis G, Wassenaar LI, Hendry MJ (2008) High-precision laser spectroscopy D/H and 18O/16O measurements of microliter natural water samples. Anal Chem 80:287–293. doi:10.1021/ac701716q Liu Y, Yamanaka T (2012) Tracing groundwater recharge sources in a mountain–plain transitional area using stable isotopes and hydrochemistry. J Hydrol 464-465:116–126. doi:10.1016/j.jhydrol.2012.06.053
3878
T.-R. Peng et al.
Liu KK, Yui TF, Shieh YN, Chiang SC, Chen LH, Hu JY (1990) Hydrogen and oxygen isotopic compositions of meteoric and thermal waters from the Chingshui geothermal area, northeastern Taiwan. Proc Geol Soc China 33(2):143–165 Liu Y, Leng X, Deng Z, Wang L, Zhang L, Liu S, An S (2011) Effects of watershed vegetation on tributary water yields during the wet season in the Heishui Valley, China. Water Resour Manage 25:1449–1464. doi:10. 1007/s11269-010-9754-1 Mas-Pla J, Menció A, Marsiñach A (2013) Basement groundwater as a complementary resource for overexploited stream-connected alluvial aquifers. Water Resour Manage 27:293–308. doi:10.1007/s11269-012-0186-y Maurya AS, Shah M, Deshpande RD, Bhardwaj RM, Prasad A, Gupta SK (2011) Hydrograph separation and precipitation source identification using stable water isotopes and conductivity: river ganga at Himalayan foothills. Hydrol Proc 25:1521–1530. doi:10.1002/hyp.7912 Panichi C, Gonfiantini R (1981) Geothermal waters. In: Gat J R, Gonfiantini R (eds) Stable isotope hydrology: deuterium and oxygen–18 in the water cycle. International Atomic Energy Agency Technical Report 210, Vienna, pp. 241–271. Peng TR, Wang CH, Huang CC, Fei LY, Chen CTA, Hwong JL (2010) Stable isotopic characteristic of Taiwan’s precipitation: a case study of western Pacific monsoon region. Earth Planet Sc Lett 289:357–366. doi:10. 1016/j.epsl.2009.11.024 Peng TR, Liu KK, Wang CH, Chuang KS (2011) A water isotope approach to assessing moisture recycling in the island-based precipitation of Taiwan: a case study in the western Pacific. Water Resour Res 47:W08507. doi: 10.1029/2010WR009890 Peng TR, Lu WC, Chen KY, Zhan WJ, Liu TK (2014) Groundwater-recharge connectivity between a hills-andplains’ area of western Taiwan using water isotopes and electrical conductivity. J Hydrol 517:226–235. doi: 10.1016/j.jhydrol.2014.05.010 Phillips DL, Gregg JW (2001) Uncertainty in source partitioning using stable isotopes. Oecologia 127:171–179. doi:10.1007/s004420000578 Ping J, Nichol C, Wei X (2014) Quantification of groundwater recharge using the chloride mass balance method in a semi-arid mountain terrain, south interior British Columbia, Canada. J Chem Pharm Res 6(1):383–388 Smerdon BD, Allen DM, Grasby SE, Berg MA (2009) An approach for predicting groundwater recharge in mountainous watersheds. J Hydrol 365:156–172. doi:10.1016/j.jhydrol.2008.11.023 Snyder RL, Orang M, Matyac S, Grismer ME, M ASCE (2005) Simplified estimation of reference evapotranspiration from pan evaporation data in California. J Irrig Drain Eng 131 (3): 249–253. doi:10.1061/(ASCE) 0733-9437(2005)131:3(249). Vanderzalm JL, Jeuken BM, Wischusen JDH, Pavelic P, Le Gal La Salle C, Knapton A, Dillon PJ (2011) Recharge sources and hydrogeochemical evolution of groundwater in alluvial basins in arid Central Australia. J Hydrol 397:71–82. doi:10.1016/j.jhydrol.2010.11.035 Viviroli D, Archer DR, Buytaert W, Fowler HJ, Greenwood GB, Hamlet AF, Huang Y, Koboltschnig G, Litaor MI, L’opez-Moreno JI, Lorentz S, Schädler B, Schreier H, Schwaiger K, Vuille M, Woods R (2011) Climate change and mountain water resources: overview and recommendations for research, management and policy. Hydrol Earth Syst Sc 15:471–504. doi:10.5194/hess-15-471-2011 Water Resources Agency (2003) Groundwater resources maps of Taiwan. Ministry of Economic Affairs, Taipei, Taiwan Wilson JL, Guan H (2004) Mountain block hydrology and mountain front recharge. In: Hogan JF, Phillips FM, Scanlon BR (eds) Groundwater recharge in a desert environment: the southwestern United States. American Geophysical Union, Washington, D. C.. doi:10.1029/009WSA08