Linking regional sources and pathways for submarine groundwater discharge at a reef by electrical resistivity tomography, 222Rn, and salinity measurements



[1] Submarine groundwater discharge (SGD) is an important component of the hydrologic cycle connecting terrestrial to marine environments. SGD in fringing reefs and its consequences on biogeochemistry and ecology remain mostly unexplored. The 222Rn activity and salinity of seawater indicate a substantial groundwater contribution throughout most of the 20 km2 studied tropical reef in Pangasinan, Philippines. Over 30 km of electrical resistivity profiles with a penetration depth of 12.5 m shows widespread zones within the reef that are much more resistive than porous reef rocks or sediment saturated with typical seawater. Some discrete resistive areas are located close to where seawater has 222Rn peaks and where geologic lineaments are likely located suggesting that these are preferential pathways for fresher groundwater discharging to the reef. SGD at the site could be a major ecological factor connecting the reef to the subsurface environment which in turn may lead to connections to land.

1. Introduction

[2] Submarine groundwater discharge (SGD) connects terrestrial environments to coastal and offshore environments and is recognized as a key component of the hydrologic and biogeochemical cycles [Burnett et al., 2003; Michael et al., 2005; Moore, 1996]. It provides pathways for fluids and solutes whose fluxes could be larger than from surface water input or could be the sole pathway for materials from land where there is little to no surface drainage. However, very few of the previous SGD studies have been done on reefs and the few that did were conducted in environments abutting the shore. These studies show that SGD delivers substantial amounts of nutrients to reefs thereby affecting reef ecology and biogeochemistry [Paytan et al., 2006; Street et al., 2008]. Since SGD is potentially a critical natural component of coral reefs [e.g., D'Elia et al., 1981; Lapointe, 1997], its role needs to be fully understood in light of the gloomy prospect for coral reefs under combined natural and human-induced global change [Bellwood et al., 2004; Carpenter et al., 2008].

[3] Despite comprehensive efforts, one gap in SGD science that has been challenging to fill is a lack of observations which link offshore information from isotopic or chemical tracer measurements [Cable et al., 1996a; Moore, 1996], seepage measurements [Michael et al., 2003; Moore et al., 2002], ground-based [Moore et al., 2002] and remote thermal measurements [Johnson et al., 2008] to subsurface environments in offshore areas, coasts and further inland. However, there have been some recent comprehensive studies [Martin et al., 2007; Breier et al., 2009]. In some cases numerical models have filled this gap where processes at larger spatial scales and longer time scales can be linked to local and time-constrained coastal observations [Michael et al., 2005]. In addition to numerical modeling, large-scale electrical resistivity (ER) imaging can fill this gap. While there are excellent examples of its application in SGD studies [Breier et al., 2005; Swarzenski et al., 2006], we present a regional yet high-resolution application in a carbonate reef fringing an island with no obvious surface drainage. We collected more than 30 km of intersecting ER lines over a 10 km2 area over several days leading to a 3D ER map for the reef. The ER surveys are complemented by simultaneous 222Rn and salinity measurements of seawater and some groundwater samples from the island.

2. Study Site and Methods

[4] The reef we studied fringes Santiago Island at the northern tip of the Bolinao Peninsula in Pangasinan province, Philippines (Figure 1); Pangasinan is on the western side of the major island of Luzon abutting the South China Sea. The 20 km2 reef flat has an average water depth of 2 m and contains several ecosystems including one of the most extensive seagrass meadows in the country and a giant clam nursery. It is also an important fish and invertebrate refugia for the fishing communities in Pangasinan and neighbor provinces.

Figure 1.

False-color multispectral image of Santiago Island, its vicinity and key features, including the studied fringing reef. The interval for contour lines is 10 m. The bubbles show 222Rn activity from measurements in March 2009 and the tracks for continuous resistivity profiling conducted in December 2008. Orange dashed lines are inferred geologic lineaments based on interpretation of topographic and bathymetric data. The white box indicates approximate boundary for Figure 2. The inset map shows the Philippines with the location of the study site at the tip of Bolinao Peninsula.

[5] Santiago Island is separated from the main Bolinao Peninsula by the 30 m deep and 0.5 km wide Guiguiwanen channel. The island population is about fourteen thousand with agriculture and fishing as the locals' primary source of livelihood. The islanders rely on groundwater wells for daily domestic use except for cooking and drinking; these wells tap the lithologic contact between the Pleistocene coralline limestone and Plio-Pleistocene marl at around sea level. Some of the island dwellers have deeper (6 m) drilled wells used as source for cooking and drinking water.

[6] Similar to most of northwestern Luzon (the mainland), Bolinao experiences low rainfall conditions during the northeast monsoon period from November to March and high precipitation during the southwest monsoon from April to October. High rainfall is usually brought in by typhoons which occur mostly during the southwest monsoon period.

[7] Several geologic lineaments cut across Santiago Island and the Bolinao Peninsula (Figure 1). The north to north-northeast and north-northwest trending lineaments are consistent with the orientation of the East Zambales and Western Boundary faults and are likely associated with faults and fractures related to the active tectonic history of the area [Rimando and Knueffer, 2006]; we postulate that these could act as conduits for fluid flow.

2.1. Measurements of 222Rn

[8] SGD can be quantified using several isotopes in the U-Th decay series particularly Ra and Rn isotopes which are decay products of aquifer and sediment bound material and are typically enriched in groundwater or pore water [Swarzenski, 2007; Dulaiova et al., 2008]. We analyzed 222Rn in bottom waters along the reef following the approach described in Burnett and Dulaiova [2003] using the RAD-7 portable radon–in-air monitor (Durridge Co. Inc.) and also ran sediment equilibration experiments for several weeks to determine 222Rn diffusive flux from the sediment similar to Cable et al. [1996b]). In-situ measurements (Figure 1) were done in June 2008 which is within the wet season and again in December 2008 and March 2009 which are within the dry season. Pore-water samples from sites in the reef and groundwater samples from the island were also analyzed for 222Rn.

2.2. Continuous Resistivity Profiling (CRP)

[9] Electrical resistivity (ER) was continuously profiled using a towed SuperSting R8 marine ER system (Advanced Geosciences, Inc., AGI) similar to Breier et al. [2005]. A cable with eleven electrodes, spaced 5 m apart, was towed behind a boat. The SuperSting R8 simultaneously collects information from a differential GPS and a sonar transducer, and continuously controls current injection and potential field mapping using a translating dipole-dipole array. The CRP survey was conducted over three days in December 2008 for several hours a day while maintaining a speed of ∼ 1 knot. The apparent resistivity data was inverted piecewise using the CRP module of AGI EarthImager after manually segmenting continuous profiles into straight and computationally manageable sections. Each segment had 65% overlap with its neighboring segments. The resistivity of the water column was not held constant and not constrained by independent measurements during the inversion.

2.3. Conductivity-Temperature-Depth (CTD) Surveys

[10] A Seabird™ CTD-submersible pump flow-through system was used to directly profile electrical conductivity and temperature along the reef flat in July 2009. Water depth at which seawater was sampled was recorded using a depth logger connected to the opening of the submersible pump while position was measured with a GPS.

3. Results

3.1. Radon

[11] The 222Rn in samples from wells from Santiago Island ranged from 17 to 1674 dpm/L (Figure 1) which is significantly higher than the 0.07-3.9 dpm/L range measured from the reef waters (Figures 1 and 2). Sediment porewater 222Rn activity in selected sites within the reef flat ranges from 2 to 36 dpm/L, for both December 2008 and March 2009 surveys, which reflects mixing of the seawater with groundwater end members.

Figure 2.

Three-dimensional distribution of electrical resistivity inside the reef and the overlying water column (top fence diagram), in-situ 222Rn activity of bottom waters measured at the same period as the ER surveys (filled squares; the filled circles are measurements in March 2009), and bathymetry of the reef; the view is facing south. The color scale is for both ER and 222Rn activity. The circles representing ER values correspond to 5% of the actual grid nodes in the numerical model used in the inversion. The white lines in the ER profiles indicate bathymetry. The projected locations of the ER profiles are indicated by white lines on the bathymetric surface. The base bathymetric map is 4.88 × 3.30 km. The dotted yellow boxes shows location of resistivity plots from CTD casts shown in Figure 3.

[12] The 222Rn diffusive flux from the sediment determined through sediment equilibration experiments ranged between 6-26 dpm/m2-d. These values are at least an order of magnitude smaller than the estimates of 120-140 dpm/m2-d reported by Cable et al. [1996a, 1996b] which they described as an insignificant amount. Unfortunately, we cannot conduct a mass balance calculation [e.g., Cable et al., 1996a], due to lack of other information. While we do not have detailed information about currents in the reef for the study period, the diurnal tidal range was from 0.6 to 1.1 m during the time of the December period of the study indicating that the reef is well-flushed since the average depth is ∼ 2 m. Since there is no surface drainage on the north side of Santiago Island and since 222Rn was not detected beyond the reef crest in the open ocean, the 222Rn inventory in the reef is very likely due to SGD. Although our estimated 222Rn diffusive flux from sediment is much smaller than Cable et al. [1996a, 1996b], our total bottom water 222Rn concentrations are similar to those at their sites (∼1–2 dpm/L but sometimes as high as 4 dpm/L) where they estimated SGD to have a magnitude of several cm/d over hundreds of square kilometers.

[13] The 222Rn was generally high but variable throughout the reef for both study periods. In December, the highest 222Rn concentrations were found in areas adjacent to the northern shores of Santiago Island (Figure 2). However, high 222Rn activities were observed in the central reef area between Santiago Island and Silaki Island, a small island within the reef at its northern portion. The lowest values tend to be near the reef crest. Low 222Rn activity was observed in a few spots near the coast of Santiago Island, particularly northeast of a small peninsula on the northeastern part of the island.

3.2. Electrical Resistivity

[14] Based on typical resistivity-porosity relationships, i.e., Archie's Law, for carbonates [Jackson et al., 2002] and assuming a low resistivity of seawater (0.1–0.2 Ωm, but this may be as high as 1 Ωm due to salinity and temperature changes), the bulk resistivity of seawater-saturated unconsolidated carbonates is expected to fall between 0.4–0.5 Ωm, certainly no more than 1 Ωm. Since the reef consists mostly of unconsolidated sediment and/or very porous reef material, areas which are more resistive than 0.5 Ωm (or up to 1 Ωm to be more conservative) can be interpreted as having interstitial water or groundwater that is less saline than seawater. Of course, there is some uncertainty in this since porosity may vary.

[15] Since the resistivity of the water column was not constrained in the inversion, any differences in water resistivity may be exaggerated and areas of mixing in the water column may appear more pronounced. Note that the threshold resistivity of 0.5–1.0 Ωm, anything above which can be interpreted as having porewater fresher than seawater, corresponds to green in Figure 2; typical seawater should be blue in Figure 2. Most of the inverted ER sections result in a resistivity ∼ 0.2 Ωm for the water column. The closeness of the inverted ER for the water column to that typical of seawater supports the accuracy of the unconstrained inversion. Nevertheless, there are large variations in resistivity from the CRP surveys of the water. These are found in the same areas where bottom water 222Rn activity is high; this is best illustrated close to the northern perimeter of Santiago Island (Figure 2). At this same area there are relatively resistive zones within the reef. Moreover, the broad area of the reef between Santiago Island and Silaki Island which has high 222Rn activity is also associated with a highly resistive portion inside the reef. In fact, we interpret a lineament originating from Santiago Island, going through this area and connecting with Silaki Island (Figure 1). In general, a broad portion within the reef has a resistivity > 0.5 Ωm.

[16] Towards the eastern side of the reef, there are points of low 222Rn activity enveloped by areas with high 222Rn activity. This is obvious at Malilnep channel and just northeast of the small peninsula and on the northern part of Santiago Island. Electrical resistivity within the reef is very high just east of the small peninsula and below Malilnep channel (Figure 2). The eastern side of the peninsula corresponds to a lineament that is thought to be continuous with the Malilnep channel (Figure 1). At the east side of Malilnep channel, the resistive subsurface zones (yellow to red colors) reach the sediment-water interface. According to other researchers who regularly dive in this area, springs are found there. However, these areas- Malilnep channel and just east of the small peninsula- have mixed signals from 222Rn activity. This is consistent with the electrical resistivity data suggesting an area where seawater and other less saline water mix. The water in these areas both have high resistivity near the surface (yellow to red color in Figure 2) but lower resistivity for the bottom section (blue to green color in Figure 2) from where samples analyzed for 222Rn are pumped from.

3.3. Salinity Profiles

[17] Continuous CTD profiling in selected sites showed spatial and vertical variation in seawater resistivity indicative of mixing of different waters. At the three sites we conducted CTD casts (Figures 2 and 3), the average resistivity of the majority of the water column is ∼ 0.18 Ωm. In fact, in a few areas at Malilnep channel, where both the ER and the 222Rn data indicate mixing of saline and fresher water, the CTD measured a resistivity as high as 0.23 Ωm in some spots which is basically fresh water.

Figure 3.

Resistivity profiles from CTD casts in July 2009. Locations are indicated by dotted yellow boxes in Figure 2. (a) From south of Silaki Island, (b) from north of Silaki Island, and (c) from Malilnep channel. All views are facing south.

4. Summary, Conclusions and Implications

[18] We present results of simultaneous regional synoptic 222Rn measurements of bottom waters, three-dimensional continuous electrical resistivity profiling, and CTD casts conducted at a fringing reef. Groundwater samples from wells in the islands have very high 222Rn activity whereas sediment equilibration based diffusive fluxes of 222Rn from the sediment to the water column are extremely small. Since there is no discernible terrestrial surface water input, the abundance of 222Rn distribution across the entire reef and generally low salinity seawater suggests a substantial contribution from fresher-than-seawater, 222Rn-rich, groundwater. The electrical resistivity profiles shows the presence of zones within the reef whose resistivity is more than that for typical carbonate sediment saturated with seawater suggesting that the water in the pore spaces is fresher (more resistive) than seawater. The locations of these resistive subsurface zones correspond to areas with higher 222Rn activity. In some cases, areas with 222Rn high activity also correspond to portions of the water column which are electrically resistive suggesting that submarine groundwater discharge is affecting the salinity of the water column. These resistive subsurface zones are generally present throughout the reef but punctuated in some areas where fractures are thought to exist based on topographic and bathymetric data. The salinity of seawater measured by the CTD is lower than expected for the region, and in the same areas where 222Rn of water and ER in the reef is high, there are pronounced localized drops in salinity. These areas are interpreted as sources and pathways for submarine groundwater discharge which supplies fresher 222Rn-rich water to the reef. Submarine groundwater discharge is diffuse and prevalent throughout the reef but is more pronounced in areas that appear to have faults. These potentially represent a key pathway for nutrients and could be affecting reef ecology and biogeochemistry. In fact, a recent biological study in the reef identified the abundance near Malilnep channel of blue coral (Heliopora coerulea), a species thought to be tolerant of low salinity and nutrient-rich conditions. Moreover, corals transplanted and nursed as part of reef restoration studies died in certain areas and the researchers partly blamed freshwater seepage [Shaish et al., 2010]; these areas are in fact those with higher resistivity and therefore prone to submarine groundwater discharge. Additionally, nutrient analysis of waters from the reef indicates high nitrate concentrations in the vicinity of the Malilnep channel. Integrative studies and management of coral reefs in similar settings should therefore closely consider the effects of hydrogeologic processes.


[19] We thank the reviewers of this paper for their constructive comments. This study was made possible by a grant from Global Ecology Fund under the Coral Reef Targeted Research project. Field work by MBC was supported by the Philippine Department of Science and Technology-Balik Scientist Program and while he was hosted by the University of the Philippines-National Institute of Geological Sciences. We thank Brad Carr of Advanced Geosciences, Inc. for technical support with the ER surveying and analysis.