Responses of atoll freshwater lenses to storm-surge overwash in the Northern Cook Islands James P. Terry & Anthony C. Falkland Abstract A category 5 tropical cyclone swept a storm surge across remote Pukapuka Atoll in the Northern Cook Islands (South Pacific Ocean) in late February 2005. Groundwater salinity (specific conductance) observations are reported for the 2-year post-storm period, with the aim of investigating the effects of saltwater intrusion on thin freshwater lenses within the atoll islets. This is the first article to present field observations of such an event. Specific conductance at shallow depths increased dramatically from potable conditions (approximately 1,000 μS/ cm) to brackish levels unsuitable for drinking (up to 10,000 μS/cm) shortly after the cyclone. Subsequently, the freshwater lenses required 11 months to recover. Within the thickest aquifer, a well-defined saline plume formed at 6m depth, sandwiching a freshwater layer beneath it and the base of the lens. Plume dispersal proceeded only gradually, owing to its formation at the start of the SW Pacific regional dry season and the low tidal range on Pukapuka. Consequently, the remnant of the plume was still present 26 months after the saltwater incursion. An important finding was that the freshwater horizon preserved at depth maintained salinity levels below 1,800 μS/ cm (i.e. within usable limits) for at least 5 months after surface overwash. Keywords South Pacific . Island hydrology . Salt-water/ fresh-water relations . Overwash . Freshwater lens
Received: 6 April 2009 / Accepted: 5 October 2009 Published online: 27 November 2009 * Springer-Verlag 2009 J. P. Terry ()) Department of Geography, National University of Singapore, AS2, 1 Arts Link, Kent Ridge, Singapore 117570 e-mail:
[email protected] Tel.: +65-6516-7144 A. C. Falkland ()) Island Hydrology Services, 9 Tivey Place, Hughes, ACT 2605, Australia e-mail:
[email protected] Hydrogeology Journal (2010) 18: 749–759
Introduction and aims The importance of freshwater lenses on atolls One of the reasons that permanent settlement is so precarious on the remote coral atolls of the tropical Pacific Ocean is that these small low-lying islands have severely limited freshwater resources. Apart from rainwater, the only natural potable water supply on atolls, upon which the islanders are completely dependent, is in the form of thin ‘lenses’ of fresh groundwater. Any degradation of these fragile freshwater resources can have serious consequences for the atoll population. Such lenses are developed as unconfined aquifers within the coarse carbonate sands, gravels and underlying reef limestone that make up the island substrate. Freshwater lenses are formed by rainwater percolation into the substrate and they essentially ‘float’ above the seawater beneath. A wide transition zone of brackish water occurs at the base of the freshwater zone, and both flow and freshwater/saltwater mixing occurs mostly in a vertical rather than a horizontal direction (Underwood et al. 1992). Replenishment of freshwater lenses is entirely reliant on groundwater recharge from rainfall. Much of the native vegetation on low atoll islands, particularly coconut trees, has evolved to become phreatophytic, i.e. they use this shallow groundwater directly to supply most of their moisture demands (Falkland and Brunel 1993). The existence of freshwater lenses is absolutely critical for both sustaining human habitation and ecosystem maintenance on atoll islands, yet such lenses are uniquely fragile. Exploitable water resources may be as low as 20% of the mean annual recharge to the lenses (Anthony 1991) and lenses are also highly vulnerable to a range of human and natural disturbances (White et al. 2007a). Anthropogenic damage can often result from inappropriate application of pesticides or fertilizers to crops, or faecal contamination owing to a lack of adequate sanitation and effluent disposal (Dillon 1997). Saline intrusion due to upconing of seawater is another problem caused by excessive groundwater pumping (Diaz Arenas and Falkland 1991). In comparison, natural (and humanassisted, semi-natural) processes of lens degradation are mainly the consequence of three effects. These are as follows. Firstly, physical erosion of an island itself, for example during powerful storms or owing to long-term sea-level rise, can lead to a reduction in size of the DOI 10.1007/s10040-009-0544-x
750
freshwater aquifer contained within (Terry and Thaman 2008). Secondly, meteorological drought reduces recharge, for example owing to changes in precipitation patterns across the Pacific driven by the El Niño-Southern Oscillation (ENSO; White et al. 1999; van der Velde et al. 2006; White et al. 2007b). For extended droughts, lens recharge becomes insufficient to replace the freshwater lost through mixing with underlying seawater and outflow from the lens perimeter (White et al. 1999), and a depleted lens may take 1.5 years to recover (Bailey et al. 2009). Thirdly, lens salinisation may occur after overtopping of an island by either tsunami waves or storm surges (Spennemann 2006).
Storm-surge threats Of the above-mentioned threats, storm-surge impacts on the condition of atoll groundwater resources are one of the least-well monitored and poorly understood. A storm surge is a temporary rise in sea level that often accompanies tropical cyclones (Terry 2007). Surge is generated by the combination of low atmospheric pressure near the centre of the storm and strong winds piling seawater up against the shore. A cyclone does not need to make landfall to cause a storm surge but the closer a cyclone track moves towards land, the more likely it is for a high storm surge to be produced. As seawater is pushed up against a coastline, coral reefs that normally offer some protection become more deeply submerged than usual, allowing large wind-driven waves to sweep either partially or totally across an atoll. In an effort to understand better the effects of stormsurge overwash on atoll freshwater lenses, the aim of this study is to investigate the impact of Cyclone Percy on Pukapuka Atoll in the Northern Cook Islands. Percy tracked close to Pukapuka in February–March 2005 and was the strongest storm to strike in recent decades. Freshwater lens salinity increases were measured shortly after the cyclone struck and the subsequent timescale of the lens recovery was monitored. Implications of the results are discussed in the context of climate change and water resources vulnerability on atolls.
Study area Physical geography of Pukapuka Atoll The Cook Islands is an independent island nation in the tropical South Pacific, in free association with New Zealand. The archipelago comprises volcanic, limestone, atoll and mixed-geology islands, scattered widely across an oceanic expanse of 2.2 million km2. The capital, Avarua, is located on the largest island Rarotonga, in the Southern Cooks group. Remote Pukapuka (10.85°S, 165.83°W) in the Northern Cooks group (Fig. 1) is a coral atoll that lies approximately 1,150 km northwest of Rarotonga. Pukapuka’s atoll reef is triangular in plan view comprising three main coral islets (called motu in Polynesian languages), one formed at each apex of the triangle. These are named Wale, Hydrogeology Journal (2010) 18: 749–759
Motu Ko and Motu Kotawa. Wale and Motu Ko are separated by a string of tiny unvegetated sand cays along the eastern atoll reef. The islets are built of Holocene coarse coralline sands and reach just 2 m or so above sea level, which is typical of atoll islets. Pukapuka has a diurnal microtidal environment, with mean and spring tidal ranges of only 30.5 and 36.6 cm respectively; mean tide level is 45.7 cm above Mean Lower Low Water Datum (NOAA 2008). For comparison, larger mean and spring tidal ranges are measured at Suva, Fiji (116.4 and 130.2 cm); Nukualofa, Tonga (106.7 and 121.9 cm) and Apia, Samoa (79.3 and 97.5 cm). The thickness of the Holocene sediments, consisting of coral sands and other coral material, ranges from 7–12 m (Table 1), unconformably overlying foundations of older Pleistocene limestone (Gray et al. 1992; Hein et al. 1997). Average saturated hydraulic conductivity of the Holocene coral sands is estimated to be approximately 10 m/day. While no direct saturated hydraulic conductivity measurements were made on Pukapuka, drilled sediments were similar to those observed on other atoll islands, the conductivities of which were measured using borehole falling head tests, including on Tarawa and Kiritimati atolls, Kiribati (Falkland and Woodroffe 1997; White et al. 2007b) and South Keeling atoll in the Cocos (Keeling) Islands (Falkland 1994; Woodroffe and Falkland 1997). Typical saturated hydraulic conductivity of underlying Pleistocene limestone is considerably greater, e.g. >1,000 m/day (Oberdorfer et al. 1990), due to karstification and diagenetic change during past episodes of sea-level change (Underwood et al. 1992). This prevents the downward extension of any freshwater lens into this bedrock. The total land area of Pukapuka’s small islets is just 3.8 km2. This land supports a population of about 640 people, most of which lives in three villages on Wale islet. Since 1900, the population has, in general, been growing, so freshwater demand has increased over the long-term. Freshwater for consumption and domestic use is obtained from rainwater tanks and a limited number of shallow wells. Giant swamp taro (Cyrtosperma chamissonis) is the most important subsistence food to supplement coconuts and fish, and is planted in the naturally fresh or slightly brackish swamps in the centre of all three islets. The atoll has a tropical marine climate with a mean annual rainfall of 2,845 mm, of which 63% falls in the 6-month wet season from October to March. Based on hydrological studies on coral islands elsewhere in the Indian and North Pacific oceans, freshwater lens recharge may be less than 50% of annual precipitation due to high evapotranspiration losses (Hamlin and Anthony 1987; Falkland 1994), although a simple surface-water balance for Pukapuka indicates that a water surplus occurs, on average, for every month of the year. Only when monthly rainfall is less than about 120 mm, is it likely that a water deficit would occur (Falkland 2005). However, droughts (e.g. July–September 1976, with a total rainfall of only 97 mm) are periods when lens recharge is significantly reduced. In the wet season months, tropical cyclones develop over warm waters in the tropical South Pacific. On average, nine DOI 10.1007/s10040-009-0544-x
751
Fig. 1 Location and geography of Pukapuka Atoll, showing also sites of groundwater borehole installation and monitoring on the three islets
cyclones occur annually across the region (Terry 2007), but of these, only 4–6 storms per decade normally cross Cook Islands waters. Yet, in 2005 an unprecedented six cyclones affected the archipelago. Of these, Cyclone Percy passed closest to Pukapuka Atoll and was the only storm to cause overwash of the islets, as described in the following section.
Impact of Cyclone Percy Percy was the sixth tropical cyclone to form in the South Pacific Ocean in 2005 and attained category-5 intensity (most severe category) on the standard Saffir-Simpson scale. At 18:00 on 24 February the storm was officially named ‘Cyclone Percy’ near 8.5°S 178.4°W by the Hydrogeology Journal (2010) 18: 749–759
Regional Specialized Meteorological Centre (RSMC) based in Fiji (RSMC-Nadi 2005). Central pressure dropped to 925 hPa around 06:00 on 27 February. Percy then decelerated and at 21:00 passed within 15–30 km to the south of Pukapuka Atoll in the Northern Cook Islands (Fig. 2). At this point, the cyclone recommenced intensification with reorganised deep convection (Clarke 2005). Maximum sustained 10-min winds of 90 knots (167 km/h) and gusts up to 120 knots (222 km/h) were recorded on the atoll (NIWA 2005). For the atoll nation of Tokelau, the cyclone destruction was the worst since a similar storm in 1966. Waves swept across the low-lying islands from both the ocean and the lagoon sides of the atolls, clashing in the middle of the islands and swamping villages (OCHA 2005). Percy also DOI 10.1007/s10040-009-0544-x
752 Table 1 Details of monitoring boreholes and freshwater lens characteristics prior to Cyclone Percy. Details in column 7–14 are averaged from a set of ten measurements between February 2004 and January 2005. SD standard deviation Borehole No.
Island
Borehole 1 Wale Borehole 2 Borehole 3 Borehole 4 Borehole 5
GPS co-ordinates South West
10°51′ 6.2″ Wale 10°50′ 58.9″ Wale 10°50′ 55.4″ Motu Kotawa 10° 53′ 33.8″ Motu Ko 10°54′ 37.0″
165°50′ 15.4″ 165°50′ 21.8″ 165°50′ 21.7″ 165°52′ 13.2″ 165°50′ 5.1″
Water levelb (m) Deptha to Deptha to base of Pleistocenehole (m) Holocene Mean Max. Min. SD unconformity (m)
Deptha to base of lensc (m) Mean Max. Min. SD
Estimated lens thickness (m)
13.0
10.1
1.92
2.05
1.85 0.08 7.4
8.5
7.0
0.61 5.5
10.0
10.5
1.28
1.35
1.05 0.09 4.6
4.9
4.5
0.15 3.3
18.5
11.5
1.70
1.95
1.45 0.15 6.3
7.0
6.0
0.32 4.6
12.0
11.9
1.12
1.35
0.85 0.16 4.3
4.8
3.5
0.47 3.2
22.0
7.2
2.30
2.65
1.95 0.21 12.3
13.2
12.1 0.34 10.0
a Below ground level. Note: there is no tide recorder at Pukapuka from which to establish ground height or lens surface/base relative to mean sea level (msl). The installation of a temporary tide recorder at two sites proved unsuccessful due to problems of wave damage to data logging instruments and the relatively short period of observations. Hence, depths are reported relative to ground level rather than msl as the best alternative at each borehole site b Top of the freshwater lens, measured as the depth below ground c The base of the freshwater lens was taken at the 2,500-μS/cm specific-conductance level (see section Groundwater characteristics prior to disturbance). The 2,500 μS/cm SC level was measured from the salinity profiles shown in Fig. 6
Fig. 2 Track of Cyclone Percy through the Cook Islands in February–March 2005; date markers are for midnight UTC. Inset: Visible satellite image at 00:40 1 March UTC, after passing close to Pukapuka and Nassau a few hours earlier. Satellite base image courtesy of NASA Hydrogeology Journal (2010) 18: 749–759
DOI 10.1007/s10040-009-0544-x
753
struck the Northern Cook Islands very hard and a state of emergency was declared. Almost all houses on Pukapuka and nearby Nassau were destroyed or damaged, leaving many people homeless. Post-storm flood conditions were precarious for the local population as fresh food supplies were wiped out by saltwater incursion into the taro swamps (Fig. 3). Many rainwater tanks lost their roof catchments and the few freshwater wells were polluted by seawater, with consequent risks to public health. Coconuts could not be used as an emergency measure as most had been torn from trees and lay rotting on the ground. An urgent shipment of drinking water, organised by the Red Cross, had to be transported from American Samoa.
Methodology Overview of groundwater investigations An assessment of groundwater resources was carried out on the three islets of Pukapuka in February 2004, upon request of the Cook Islands government. This was done as part of a programme to investigate water resources and provide recommendations for water supply improvements in the populated outer islands of the Cook Islands (Falkland 2005). The assessment included surveys of groundwater level and salinity in wells, ponds and swamps, electromagnetic (EM) surveys and installation of monitoring boreholes.
The groundwater surveys enabled a general assessment of the locations of fresh and brackish groundwater on the islets. These surveys were conducted using a tape measure and hand-held specific conductance (SC) meter. In addition, groundwater level and SC were recorded onto data loggers installed at four pits dug for drilling operations and one well. The data was recorded over typically two days at each pit and over 13 days at the well. Figure 4 shows the groundwater level and salinity (SC) variations over approximately 40 h at a dug pit near borehole 1. The groundwater level (in relation to the base of the pit) shows the influence of sea-level tides, as well as the short-term drawdown caused by a pump. The maximum amplitude of the groundwater signal during the observation period is approximately 9 cm. While the sea level amplitude was not measured, it is known that the mean tidal range is 30.5 cm (refer to section Physical geography of Pukapuka Atoll). Assuming that mean tides applied during the observation period, the tidal efficiency (TE; groundwater-to-sea-level amplitude ratio), is approximately 0.30. This is a moderately high TE value for atoll groundwater indicating relatively high permeability of the coral sediments. From Fig. 4, there was very little variation of the groundwater salinity due to the sea level influence, with the SC values ranging from 580 to 605 µS/cm. Slightly higher SC variations are noted during pumping of the pit. The electromagnetic (EM) surveys were carried out to map the shape and horizontal extent of the lenses on each of the three islets. The EM data assisted with the siting of the monitoring boreholes at locations of interest within the lens boundaries.
Borehole installation and monitoring
Fig. 3 Central swamp a with healthy taro plantation (the principal subsistence crop) on Wale islet in early 2005 and b seawater stagnating in the same location several days after the storm-surge overwash caused by Cyclone Percy in early March 2005. Photo credits: A. Falkland and Douglas Clark Hydrogeology Journal (2010) 18: 749–759
Five multi-level monitoring boreholes (Fig. 5) were drilled and installed on the lenses. The boreholes enabled calibration of the EM survey data and also a means of ongoing monitoring of the groundwater, both level and water quality, at selected locations. Three of the boreholes were installed on Wale (boreholes 1–3), one on Motu Kotawa (borehole 4) and one on Motu Ko (borehole 5). Boreholes were 89 mm in diameter, drilled to depths of 10–22 m, and installed with monitoring tubes made from 8-mm nylon tubing to different depths. The base of each tube was fitted with a glass filter to prevent ingress of sand and the top of each tube was fitted with a snap coupling to enable a portable electric pump to extract samples for salinity (specific conductance) measurements. Boreholes were backfilled with gravel and bentonite layers to form hydraulically isolated zones around the end of each tube. No permanent casing was fitted inside the boreholes. Each borehole was fitted with a 50-mm PVC pipe to below the water table and protected with a cast-iron cover set flush with the ground surface. After borehole installation, a programme of groundwater salinity measurements commenced, carried out by local personnel who were trained in the use of the sampling pump and the hand-held SC meter. Monitoring DOI 10.1007/s10040-009-0544-x
754
Fig. 4 Groundwater level (m) and specific conductance (SC) variations at water pit dug for borehole 1
Fig. 5 Monitoring borehole design Hydrogeology Journal (2010) 18: 749–759
DOI 10.1007/s10040-009-0544-x
755
was conducted initially at approximately monthly intervals. Prior to Cyclone Percy, ten sets of groundwater level and SC measurements were made between February 2004 and January 2005. Following Cyclone Percy, the frequency of monitoring was less with only eight sets of measurements made between May 2005 and May 2007. No detailed measurements of either groundwater level or SC variations due to sea level movements were conducted in the monitoring tubes. Previous studies (e.g. Oberdorfer et al. 1990; Hunt 1997) have shown that TE increases with depth. Multi-level monitoring boreholes using PVC pipes on Diego Garcia atoll, Hunt (1997) showed TE’s of 0.04–0.35 near the water table and 0.95 at a depth of 20 m. Similar observations were not possible for the Pukapuka boreholes owing to the small diameter tubes used, but similar results would be expected. Observations of SC variations at three multi-level monitoring boreholes using nylon tubes on Home Island, South Keeling atoll in the Cocos (Keeling) Islands were conducted at 3-h intervals over a 48-h period (unpublished data, 1993). The results show that there are notable SC variations in each tube due to sea level movements but that these variations are not significant in terms of defining the freshwater lens thickness. No such detailed measurements were made for the Pukapuka boreholes but similar variations are likely to occur.
Results from the borehole monitoring Groundwater characteristics prior to disturbance The borehole monitoring data was used to determine the average depth to groundwater table and base of the freshwater lenses at each location, and thus provide an estimation of lens thickness on the three islets (Table 1). For the purpose of groundwater resource evaluation, the base of freshwater lenses was defined as SC equal to 2,500 µS/cm. This SC value is approximately equivalent to a chloride-ion concentration of 600 mg/L, based on measurements of the relationship between groundwater SC and chloride concentration on Pukapuka, Manihiki, Rakahanga and Penrhyn atolls in the Cook Islands (A.C. Falkland, Ecowise Environmental, unpublished data, 2005 and 2006). It is noted that this concentration exceeds the World Health Organization’s drinking water guideline value for taste of 250 mg/L (WHO 2006), equivalent to a SC value of about 1,400 µS/cm. The higher SC value is used here for the lower boundary of the freshwater lens because, on Pacific atolls, such water is routinely used for non-potable domestic purposes such as cooking and washing clothes. Besides, there is no health-based value for chloride in drinking water and 2,500 µS/cm is suggested as the upper limit for human use (equivalent to approximately 1,375 mg/L total soluble salts; George et al. 1996). At the approximate centre of each lens, freshwater thickness was estimated to be 5.5, 3.2 and 10.0 m thick on Wale, Motu Kotawa and Motu Ko, respectively, before the cyclone. Subsequent to the stormsurge overwash caused by Cyclone Percy in March 2005, Hydrogeology Journal (2010) 18: 749–759
borehole salinity profiles were determined from SC readings on eight occasions until May 2007.
Post-disturbance observations Figure 6 illustrates the pre-Cyclone Percy (i.e. baseline) conditions of the freshwater lenses (see Jan-05 profiles) and the post-cyclone temporal changes in groundwater salinity (SC) resulting from overwash on the three islets. The following observations are made from the graphs. In late January 2005, a month before Cyclone Percy struck, groundwater at the surface of the lenses (just below the watertable) was fresh (SC between 639 and 1,178 μS/cm). In May 2005, 2 months after the cyclone, measurements showed increased salinity, especially at borehole 5 on Motu Ko islet (10,050 μS/cm). By September 2005, 6 months post-event, the surface of the freshwater lenses had partially recovered, with water table salinity decreasing in boreholes 1 and 3–5. Monitoring in late January—March 2006 indicated that lens-surface salinity had improved to precyclone conditions, i.e. below 1,000 μS/cm. The exception is at borehole 2 where the freshwater lens surface did not fully recover for 26 months until May 2007. Deeper within the freshwater lenses, some interesting observations were made. On Moto Ko islet at borehole 5, the salinity profile for May 2005 shows that seawater incursion produced a strong saline plume (15,450 μS/cm) at about 6 m below the surface. However, at a depth of about 12 m, sandwiched beneath the saline plume and the base of the lens, is a thin horizon of fresh water (1,599 μS/ cm). Until September 2005, the salinity plume remained static but its SC decreased to 8,560 μS/cm at 6 m depth. Over the next 5 months, the plume migrated slowly downwards to about 9 m depth and continued to decrease in salinity. By May 2007 (26 months post-cyclone), salinity levels at 6 m depth recovered to pre-cyclone conditions, although below this, the remains of the saline slug were still visible (3,970 μS/cm at 9 m). The salinity of the thin horizon of fresh water preserved beneath the saline plume remained below 1,800 μS/cm for 5 months. Between 6 and 26 months post-cyclone, the trapped fresh layer retained its position, but increased in salinity before stabilising at a SC value of approximately 4,000 μS/cm at 12 m below the surface. A similar effect to that just described, albeit not so strongly pronounced, is also observed at borehole 3 on Wale islet (Fig. 6). There, a saline plume also formed at about 6 m, again trapping a fresher layer beneath it at about 7.5 m depth. In this case, the plume showed increasing salinity between May and September 2005 (2–6 months post-cyclone), after which the plume and the fresher layer beneath it both dissipated entirely.
Discussion and implications Freshwater lens damage and recovery In pre-cyclone conditions, the islets of Pukapuka Atoll had moderately thick fresh groundwater resources, especially DOI 10.1007/s10040-009-0544-x
756
Fig. 6 Salinity profiles in groundwater lenses on Wale, Motu Kotawa and Motu Ko islets. The Jan-05 profile is closely representative of average pre-disturbance conditions measured from borehole installation in February 2004 to early 2005. The initial effects of saline intrusion after the storm-surge overwash generated by Cyclone Percy are seen in the May-05 profiles; later profiles show the subsequent recovery of the freshwater lenses until the most recent data in May-07. Notes: Only the top part of the salinity profiles (to 15 m depth) are shown for clarity; the upper limit of the X-axes (SC equals 25,000 µS/cm) is approximately 50% seawater salinity; on the Y-axes, depth means depth below ground surface
Motu Ko. Within the boundaries of the freshwater lenses, the groundwater salinity at the water table (i.e. lens surface) was equivalent to chloride-ion values that are lower than aesthetic (taste) guideline limits for drinking water (WHO 2006). The groundwater measurements indicate that Cyclone Percy in February–March 2005 caused deterioration of the freshwater lenses on all three islets of Pukapuka. The storm surge that swept partially over the islets caused seawater to penetrate into the porous coralline substrate, contaminating the aquifers. After the storm surge had receded, saline intrusion from stagnant seawater that collected in the central depressions (taro swamps) was a continuing source of contamination Hydrogeology Journal (2010) 18: 749–759
(Fig. 3). While open wells can assist the rapid vertical transmission of seawater into the groundwater (Carlson et al. 2008), wells are few in number on the islets of Pukapuka, so this route is thought to have been less important. Immediately after the overwash caused by Cyclone Percy, the intruding saline water would have formed a layer of denser fluid above the lighter freshwater layer. Since this configuration is unstable, it may assist aquifer recovery by vertical water overturning (Villholth et al. 2008), in conjunction with the beneficial effects of recharge from rainwater. For all three islets on Pukapuka, saline contamination at the surface of the freshwater DOI 10.1007/s10040-009-0544-x
757
lenses lasted 11 months or longer. This is a relatively long period for groundwater recovery, but is similar to the recovery time observed in coastal aquifers of Louisiana in the US, which suffered contamination during hurricanes Katrina and Rita in late 2005 (Van Biersel et al. 2007). On Pukapuka, the freshwater lens recovery time is a function both of climatic conditions during the overwash event and the amount of groundwater recharge over subsequent months. Cyclone Percy produced torrential precipitation over Pukapuka, with 220 mm recorded on 28 February, coincident with the timing of the storm surge. It is very likely that this heavy rainfall mitigated the effects of the surge. The groundwater salinisation would have been more severe if the overwash had occurred in the absence of this heavy rainfall. Yet the event occurred near the end of the regional cyclone season (March), immediately prior to the dry season that lasts normally from April to September. Although daily rainfall data for 2005 on Pukapuka are incomplete, based on long-term records, it is seen that the wet season receives 63% of the annual rainfall, approximately 1,800 mm. Therefore, if the cyclone had struck the atoll earlier in the wet season, say November or December 2004, it is likely that the freshwater lenses would have recovered more quickly due to the flushing effect from the high recharge caused by large rainfalls (205 and 198 mm, respectively) in November and December 2004.
Saline plume and preservation of a trapped freshwater layer On Motu Kotawa the freshwater lens is relatively thin compared to the other islets (3.2 m at borehole 4). Consequently, the post-cyclone response was one of saline water mixing and dispersal within the main body of the lens. Evidence for this is the observed thinning of the freshwater zone and concurrent thickening of the transition zone at the base of the lens by 1–2 metres (Fig. 6). The latter characteristic is consistent with recent modelling experiments of storm overwash effects on the groundwater transition zone by Anderson and Lauer (2008). A comparable response is seen at borehole 1 on Wale near the edge of the central swamp. In contrast, Motu Ko islet had the most significant freshwater aquifer before disturbance, approximately 10 m thick. The response to saline intrusion on this islet was for the freshwater lens to become stratified, with the formation of a salinity profile ‘inversion’ at depth. A notable feature was a well-defined and persistent saline plume that developed within the lens. A saline plume also formed on Wale islet at boreholes 2 and 3, which are located reasonably close to the central swamp. In aquifers on sand islands along the coast of Sri Lanka, a similar plume-like feature was described after the devastating tsunami of December 2004 (Dharmagunawardhane and Vithanage 2008). Of particular interest here is that the saline plume on Motu Ko effectively preserved a layer of freshwater beneath it, which retained SC levels below 1,800 μS/cm Hydrogeology Journal (2010) 18: 749–759
for at least 5 months after the surface contamination. There is no geologic explanation for the trapped freshwater layer as the substrate at this depth is limestone that is heavily weathered and fractured. Accessing this water resource whilst protecting the fragile freshwater layer would be a difficult challenge, especially because vertical boreholes are not considered an appropriate method of groundwater abstraction on atolls compared with shallow horizontal slotted pipes (infiltration galleries), owing to the threat of saltwater upconing into the base of the lens. The trapped but buoyant layer of fresh water seems to have been an important influence retarding the slow rate of downwards migration of the saline plume, which was able to penetrate only 3 m (from 6 to 9 m depth) over a period of 9 months. An important control affecting the persistence of the freshwater layer is Pukapuka’s microtidal environment with an average tidal range of only 30.5 cm (NOAA 2008). This means that tidally driven vertical mixing would have been a relatively inefficient mechanism for dispersing the trapped freshwater layer on Motu Ko.
Future concerns For the small island nations of the South Pacific, it has already been recognised that the vital contribution of clean, potable freshwater will continue to take on an increasing importance for future public health and environmental and ecological stability (Depledge 1999; White et al. 2008). The potential effects of climate change and sea-level rise on existing groundwater resources are therefore a major concern for Pacific Island peoples, especially those living on low-lying atolls (Burns 2002; White et al. 2007a). This study has shown that wave overwash by cyclone-induced storm surges can produce aquifer contamination on coral islands that may last at least 10 months. A longer duration of groundwater salinisation (up to three years) was recorded on the barrier islands off the coast of North Carolina after overwash by Hurricane Emily in 1993, but there, sediment permeability is an order of magnitude lower than generally observed on atolls, so the effect of tidal oscillations as well as mixing by density differences in the transition zone is reduced (Anderson 2002; Anderson and Lauer 2008). At present, there is much debate on the possible effects of global warming, increasing sea-surface temperatures and elevated-CO2 concentrations on the incidence of tropical cyclones (Terry 2007). By and large, for the South Pacific region, empirical observations remain inconclusive, although some evidence is emerging that cyclones may be becoming more intense (Webster et al. 2005; Elsner et al. 2008). Nonetheless, one clear signal which can be identified is that El Niño episodes tend to spread out cyclone origin points farther to the east (Hastings 1990). This means that compared to normal years, El Niño events spawn more cyclones east of 180° (Chu 2004), across to central parts of the South Pacific that would otherwise be less at risk of severe storms. This finding is supported by a comprehensive examination by DOI 10.1007/s10040-009-0544-x
758
de Scally (2008) of an improved historical cyclone database (1820–2006) for the Cook Islands. His work noted that a far greater number of cyclones occurred during El Niño conditions (even weak ones) compared with neutral ENSO or La Niña conditions. Consequently, if current projections of stronger or more persistent El Niño-like conditions in a greenhouse-enhanced world become established over the next few decades, then the atolls of Tuvalu, Tokelau, the Northern Cook Islands and French Polynesia (i.e. near to and east of the dateline) will be under greater threat from cyclone-induced overwash and subsequent short–medium term contamination of their limited freshwater resources. On individual cycloneaffected atolls, storm-surge effects will therefore exacerbate concomitant pressures on fresh groundwater sustainability owing to sea-level rise, overexploitation and pollution. Such effects on water resources may seriously threaten the viability of communities that currently inhabit these remote islands. Acknowledgements The authors wish to acknowledge the efforts of the Pukapuka Island Secretary and the staff of the Pukapuka Island Administration for obtaining the groundwater data during the period of monitoring. Mrs Li Kheng Lee, cartographer at the Geography Dept. of the National University of Singapore is thanked for drafting Figs.1 and 5. Dr Kolja Rotzoll and another anonymous reviewer provided constructive suggestions that enabled many improvements to the original version of the manuscript.
References Anderson WP Jr (2002) Aquifer salinization from storm overwash. J Coast Res 18:413–420 Anderson WP Jr, Lauer RM (2008) The role of overwash in the evolution of mixing zone morphology within barrier islands. Hydrogeol J. doi:10.1007/s10040-008-0340z Anthony SS (1991) Case study 7: Majuro Atoll. In: Falkland A (ed) Hydrology and water resources of small islands: a practical guide. UNESCO, Paris, pp 368–374 Bailey RT, Jenson JW, Olsen AE (2009) Numerical modeling of atoll island hydrogeology. Ground Water 47:184–196 Burns WCG (2002) Pacific Island developing country water resources and climate change. In: Gleick PH, Burns WCG, Chalecki EL, Cohen M (eds) The world's water 2002–2003: the biennial report on freshwater resources. Island, Washington, DC, pp 113–131 Carlson DA, Van Biersel TP, Riley Milner L (2008) Storm-damaged saline-contaminated boreholes as a means of aquifer contamination. Ground Water 46:69–79 Chu P-S (2004) ENSO and tropical cyclone activity. In: Murnane RJ, Liu K-B (eds) Hurricanes and typhoons: past, present and future. Columbia University Press, New York, pp 297–332 Clarke S (2005) Tropical Cyclone Percy (TD-10F/TC-20P) 24 February–5 March. http://www.australiasevereweather.com/ cyclones/2005/summ0502.htm. Cited September 2008 Depledge D (1999) Water. In: Rapaport M (ed) The Pacific Islands: environment and society, 1st edn. Bess, Honolulu, HI, pp 66–73 de Scally FA (2008) Historical tropical cyclone activity and impacts in the Cook Islands. Pac Sci 62:443–459 Dharmagunawardhane HA, Vithanage M (2008) Status of a tsunami affected coastal aquifer along the east coast of Sri Lanka. In: Bhattacharya P, Ramanathan AL, Mukherjee AB, Bundschuh J, Chandrasekharam D, Keshari AK (eds) Groundwater for sustainable development: problems, perspectives and challenges. Taylor and Francis, London, pp 223–231 Hydrogeology Journal (2010) 18: 749–759
Diaz Arenas A, Falkland A (1991) Characteristics of small islands. In: Falkland A (ed) Hydrology and water resources of small islands: a practical guide. UNESCO, Paris, pp 1–9 Dillon P (1997) Groundwater pollution by sanitation on tropical islands. Technical documents in hydrology no. 6, IHP-V. UNESCO, Paris, 31 pp Elsner JB, Kossin JP, Jagger TH (2008) The increasing intensity of the strongest tropical cyclones. Nature 455:92–95 Falkland AC (1994) Climate, hydrology and water resources of the Cocos (Keeling) Islands. Atoll Res Bull 400:52 Falkland A (2005) Pukapuka, Cook Islands, report on water investigations, February 2004–February 2005. Report No. EHYD 2005/46, prepared on behalf of GHD Pty Ltd for the Australian International Development Assistance Agency and the Cook Islands Government, March 2005, Ecowise Environmental, Canberra, Australia Falkland AC, Brunel JP (1993) Review of hydrology and water resources of the humid tropical islands. In: Bonell M, Hufschmidt MM, Gladwell JS (eds) Hydrology and water management in the humid tropics. International Hydrology Series, Cambridge University Press, Cambridge, pp 135–163 Falkland A, Woodroffe CD (1997) Geology and hydrogeology of Tarawa and Christmas Island, Kiribati. In: Vacher HL, Quinn TM (eds) Geology and hydrogeology of carbonate islands. Developments in sedimentology, vol 54. Elsevier, Amsterdam, pp 577–610 George R, Weaver D, Terry J (1996) Environmental water quality: a guide to sampling and measurement. Department of Agriculture, Gov. of Western Australia, Perth, 38 pp Gray SC, Hein JR, Hausmann R, Radtke U (1992) Geochronology and subsurface stratigraphy of Pukapuka and Rakahanga atolls, Cook Islands: late Quaternary reef growth and sea-level history. Palaeogeogr Palaeoclim Palaeoecol 91:377–394 Hamlin SN, Anthony SS (1987) Groundwater resources of the Laura area, Majuro Atoll, Marshall Islands. US Geol Surv Water Resour Invest Rep. 87-4047, 69 pp Hastings PA (1990) Southern oscillation influences on tropical cyclone activity in the Australian/south-west Pacific region. Int J Climatol 10:291–298 Hein JR, Gray SC, Richmond BR (1997) Geology and hydrogeology of the Cook Islands. In: Vacher HL, Quinn TM (eds) Geology and hydrogeology of carbonate islands. Developments in sedimentology, vol 54, Elsevier, Amsterdam, pp 503–535 Hunt CD (1997) Hydrogeology of Diego Garcia. In: Vacher HL, Quinn TM (eds) Geology and hydrogeology of carbonate islands. Developments in sedimentology, vol 54, Elsevier, Amsterdam, pp 909–931 NIWA (2005) The island climate update, No. 56. May 2005. National Institute for Water and Atmospheric Research, Wellington, New Zealand, 6 pp NOAA (2008) Tidal station location and ranges. National Oceanic and Atmospheric Administration, Washington, DC. http://coops.nos.noaa.gov/tides06/tab2wc3.html. Cited October 2008 Oberdorfer JA, Hogan PJ, Buddemeier RW (1990) Atoll island hydrogeology: flow and freshwater occurrence in a tidally dominated system. J Hydrol 120:327–340 OCHA (United Nations Office for the Coordination of Humanitarian Affairs) (2005) Cook Islands and Tokelau: tropical Cyclone Percy, OCHA situation report No. 5, OCHA, New York. http://www.reliefweb.int/rw/RWB.NSF/db900SID/ SNAO-6B2KVA?OpenDocument. Cited September 2006 RSMC-Nadi (2005) Tropical cyclone summary 2004–2005 season. Regional Specialized Meteorological Centre Nadi–Tropical Cyclone Centre. Fiji Meteorological Service, Nadi, Fiji, 14 pp Spennemann DHR (2006) Freshwater lens, settlement patterns, resource use and connectivity in the Marshall Islands. Transforming Cultures eJournal 1(2):42–63 Terry JP (2007) Tropical cyclones, climatology and impacts in the South Pacific. Springer, New York, 210 pp Terry JP, Thaman RR (2008) Physical geography of Majuro and the Marshall Islands. In: Terry JP, Thomas FR (eds) The Marshall Islands: environment, history and society in the atolls. Faculty DOI 10.1007/s10040-009-0544-x
759 of Islands and Oceans, The University of the South Pacific, Suva, Fiji, pp 1–22 Underwood MR, Peterson FL, Voss CL (1992) Groundwater lens dynamics of atoll islands. Water Resour Res 28:2889–2902 Van Biersel TP, Carlson DA, Milner LR (2007) Impacts of hurricanes storm surges on the groundwater resources. Environ Geol 53:813–826 van der Velde M, Javaux M, Vanclooster M, Clothier BE (2006) El Niño-Southern oscillation determines the salinity of the freshwater lens under a coral atoll in the Pacific Ocean. Geophys Res Lett 33:L21403.1–L21403.5 Villholth KG, Amerasinghe P, Jeyakumar P (2008) Tsunami impacts on shallow groundwater and associated water supplies on the east coast of Sri Lanka. In: Bhattacharya P, Ramanathan AL, Mukherjee AB, Bundschuh J, Chandrasekharam D, Keshari AK (eds) Groundwater for sustainable development: problems, perspectives and challenges. Taylor and Francis, London, pp 211–222 Webster PJ, Holland GL, Curry JA, Chang H-R (2005) Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309:1844–1846
Hydrogeology Journal (2010) 18: 749–759
White I, Falkland A, Scott D (1999) Droughts in small coral islands: case study, South Tarawa, Kiribati. Technical documents in hydrology No. 26, IHP-V. UNESCO, Paris, 55 pp White I, Falkland A, Perez P, Dray A, Metutera T, Metai E, Overmars M (2007a) Challenges in freshwater management in low coral atolls. J Clean Prod 15:1522–1528 White I, Falkland A, Metutera T, Metai E, Overmars M, Perez P, Dray A (2007b) Climatic and human influences on groundwater in low atolls. Vadose Zone J 6:581–590 White I, Falkland A, Metutera T, Katatia M, Abete-Reema T, Overmars M, Perez P, Dray A (2008) Safe water for people in low, small island Pacific nations: the rural-urban dilemma. Development 51:282–287 WHO (2006) Guidelines for drinking water quality. First Addendum to the third edition, vol 1, Recommendations. World Health Organization, Geneva, 515 pp Woodroffe CD, Falkland A (1997) Geology and hydrogeology of the Cocos (Keeling) Islands. In: Vacher HL, Quinn TM (eds) Geology and hydrogeology of carbonate islands. Developments in sedimentology, vol 54. Elsevier, Amsterdam, pp 885–908
DOI 10.1007/s10040-009-0544-x
