Physicochemical characteristics of groundwater

The results of the physical and chemical analyses are summarized in Table 3. Fig. 8

plots groundwater depth versus target water quality parameters and the relations between

parameters. Fig. 8a and the ORP value in Table 3 indicate that the groundwater in the

proximal fan tended to be highly oxidative, whereas the mid-fan and the distal fan were

under reductive state. Low ORP and low DO in the distal fan indicate that the groundwater

(h)

(i) (j)

(g)

was anaerobic or under comparatively reduced conditions (Kao et al., 2011; Kurosawa et

al., 2008).

Fig. 8a, 8b, 8c, 8d, and 8f show that the concentrations of DO, TOC, NO3−, SO42−,

and Fe decreased with depth in the proximal fan, whereas DO, NO3−

, and SO42−

elevated in the depth of from 25 to 150 m and from 250 to 275 m. Kao et al. (2011)

suggested that the recharge of meteoric water and extensive pumping of groundwater for

agricultural requirement led to the entry of atmospheric O2 into unconfined granular

aquifers of the proximal fan, changing the reductive groundwater to be oxidizing state, or

even complex redox condition. Fe and As concentrations remained low in aforementioned

depth, which may be attributed to the co-precipitation or adsorption of As on Fe

oxyhydroxides (Pierce and Moore, 1982; Peterson and Carpenter, 1983).

In the mid-fan, Fe, As, and SO42− increased near surface of upper 50 m (Fig. 8c, 8f,

and 8g), implying that the reductive dissolution of As-containing Fe oxyhydroxides

resulted in the release of As into the groundwater, and meanwhile the oxidation of pyrite

might also cause the As release in the complex redox environment. High concentrations

of NH4+ in the upper 50 m (Fig. 8e) suggest that the N sources may be NH4+ fertilizers,

manure, sewage water, or other transformation products (e.g., DNRA from NO3− sources),

which were evidenced by N and O isotope composition. TOC, SO42−

, NH4+, Fe, and As

decreased with depth, suggesting that when sulfate-reducing bacteria reduced SO42− and

the oxidation of iron occurred, the production of sulfide minerals (e.g., FeS) and ferric

oxides subsequently precipitated As, adsorbed or co-precipitated aqueous As, lowering

As concentration in groundwater (Akai et al., 2004; Kirk et al., 2004; Swartz et al., 2004).

Similar situation occurred in the distal fan; however, TOC concentrations increased

with depth, forming two separate trends (Fig. 8b). High TOC in the upper zone of the

aquifer may be the result of the application of manure or the presence of septic waste and

livestock, and the deeper zone may be attributed to local geologic sediments.

Fig. 8h indicates that there are two distinct trends of EC versus the concentrations of

SO42− in the mid-fan and the distal fan. The upper trend in the distal fan may be related

to seawater infiltration or intrusion. Wang et al. (2007) and Kao et al. (2011) reported the

occurrence of salinization in the southwestern coastal area of Taiwan and the average EC

value of salinized groundwater was greater than that of non-salinized groundwater.

Seawater infiltration and intrusion into the coastal area of the Choushui River alluvial fan

was verified by ¹⁸O mass balance calculation, resulting in elevation of EC and SO42− in

the distal fan. Also, the average concentration of Cl⁻ in the distal fan (1790.86 mg/L) was

greater than that in the proximal fan (9.10 mg/L) and the mid-fan (62.46 mg/L). The lower

trend in the mid-fan may be related to evaporation or evapotranspiration (responsible for

elevation of EC) and pyrite oxidation (responsible for elevation of SO42−), or

paleo-marine environment (responsible for elevation of both). Kao et al. (2011) suggested that

the mid-fan of the Choushui River alluvial fan was under the paleo-marine environment,

and this may contribute to the elevation of EC and SO42−.

Moreover, Fig. 8d clearly shows that high NO3− concentrations were found in the

oxidative proximal fan, whereas low NO3− concentrations were observed in the reductive

distal fan. On the contrary, the As concentrations increased from the proximal to the distal

fan (Fig. 8g). Fig. 8i shows the plot of As versus NH4+ concentrations. The relationship

between As and NH4+ is not clear. Due to a sequence of complex redox processes, the

concentration correlation between As and NH4+ may not be confirmed without any further

evidence. Fig. 8j indicates a positive correlative relationship between As concentrations

and TOC, and as TOC increases, As increases. Redox conditions are controlling As

solubility. Because both nitrate and ferric iron are electron acceptors, the bacteria utilizing

them are essentially competing against one another. Denitrification is more

thermodynamically favored and thus its occurrence may actually be consuming electron

donors (i.e., organic matter) more rapidly and making them less available for iron

reducing bacteria. The organic matter may be a more important control on redox

conditions and thus As concentrations, as other studies have shown (Liu et al., 2003;

Anawar et al., 2013). Many studies have reported the positive correlation between As and

TOC because TOC implies the organic substrates for the requirement of reaction of As

release from Fe oxyhydroxides into groundwater (Farooq et al., 2010; Hsu et al., 2010;

Anawar et al., 2013; Pi et al., 2015; Lu et al., 2016). Notably, a spatial positive correlation

is evident by As, NH4+, and TOC in Fig. 9. Elevated concentrations of As, NH4+, and

TOC were mostly observed in the mid to distal fan of the Choushui River alluvial fan,

which was opposite to the trend in NO3− concentrations. The samples of MT-1, PT-1, and

CP-1 contained relatively high TOC and NH4+ concentrations as well as high As

concentrations (Table 3). These sampling wells are located in the distal fan (Fig. 4b),

where the groundwater is in a reductive state. CH-1 and IW-1 contained high NH4+

concentrations and relatively low As concentrations, which may be explained by the

intrusion of manure, septic waste, or organic waste, and subsequent mineralization

without triggering the release of As (Kurosawa et al., 2008).

Table 4 shows the As species analysis results. As⁵⁺ concentrations in four sampling

wells (KY-1, CS-2, CK-1, SH-1) of the proximal fan were greater than As³⁺

concentrations, caused by the relatively oxidizing condition of groundwater. As³⁺

concentrations in all wells of the mid-fan and the distal fan were distinctly greater than

As⁵⁺ concentrations because the groundwater was in reducing state. The concentrations

of Fe²⁺ (observed more in the reductive groundwater) and Fe³⁺ (observed more in the

oxidative groundwater) show similar behavior.

Table 5 presents a correlation analysis between the physical and chemical parameters.

The correlation of the NH4+ concentration with NO3−, ORP, and DO was statistically

significant. Specifically, NH4+ concentrations showed significant moderate-to-high

negative correlations with NO3−, ORP, and DO, with the correlation coefficients being

−0.511, −0.752, and −0.568, respectively. This suggests that the NH4+ concentration

increased in reductive conditions, thereby causing the NO3− concentration to decrease

through denitrification or dissimilatory NO3− reduction to NH4+ (DNRA).

As concentrations showed significant moderate-to-high negative correlations with

NO3−, the ORP, and DO, with correlation coefficients of −0.514, −0.777, and −0.658,

respectively. This suggests that the release of As from sediments may increase when the

groundwater becomes more reductive, and denitrification or DNRA may occur

simultaneously with As release. The lack of significant correlations between As and

SO42− also suggests that in the reductive environment of the groundwater, complex sulfur

disproportionation geochemical processes control the As concentration in the

groundwater (Kao et al., 2011). Kirk et al. (2004) show that sulfate reduction can be an

important limit on As in solution; i.e., the production of sulfide removes As from solution.

Sulfate and As thus tend to have a "mutually exclusive" relationship, and our data in this

study generally show this relationship.

Fig. 9. Plots of spatial distribution of As, NH4+, and TOC for the 46 groundwater samples.

As > 0.5 mg/L As > 0.01 mg/L~

NH4+> 15 mg/L NH₄⁺> 1 mg/L

~

TOC > 2 mg/L TOC > 0.5 mg/L~

Table 4. Fe and As species analysis results for the 46 groundwater samples collected from the Choushui River alluvial fan in September 2015.

Region Well

Table 5. Correlations between the physical and chemical analysis results for the 46 groundwater samples of the Choushui River alluvial fan.

Correlation coefficient 1.000 0.005 0.068 0.089 -0.085 -0.498**

Significance . 0.977 0.696 0.613 0.627 0.002

Sample number 36 36 35 35 35 35

TOC (mg/L)

Correlation coefficient 1.000 0.657** -0.569** -0.423** 0.347*

Significance . 0.000 0.000 0.004 0.020

Sample number 46 45 45 45 45

EC (μmho/cm)

Correlation coefficient 1.000 -0.538** -0.234 0.358*

Significance . 0.000 0.121 0.016

Sample number 45 45 45 45

ORP (mV)

Correlation coefficient 1.000 0.775** -0.733**

Significance . 0.000 0.000

Sample number 45 45 45

DO (mg/L)

Correlation coefficient 1.000 -0.540**

Significance . 0.000

Spearman's rank correlation is used to calculate the correlation coefficient; **: p value < 0.01; *: p value < 0.05