Đang tải... (xem toàn văn)
seawater intrusion under hunman activities and sea level rise , tác động của con người và nước biển dâng đến xâm nhập mặn được đánh giá tại yucatan Peninsula. Saltwater intrusion is a major hazard to coastal communities as it causes degradation of fresh water resources. The impact of rising sea level on the saltwater intrusion into coastal aquifers has been studied for decades, but how human activities affect the extent of saltwater intrusion is poorly understood. Human activities are known to influence groundwater availability indirectly by affecting precipitation patterns and directly by extracting groundwater and reducing recharge. In this paper the authors investigated the integrated impacts of human activities and rising sea level on aquifer recharge in Quintana Roo, Mexico, by incorporating anthropogenic impacts on groundwater recharge into an analytical saltwater intrusion model. The impact of human activities on groundwater extraction was firstly calculated; then, the resulting groundwater recharge was used in a Ghyben–Herzberg analytical model to determine the inland distance of saltwater intrusion.
Nat Hazards DOI 10.1007/s11069-016-2621-5 ORIGINAL PAPER The integrated impacts of human activities and rising sea level on the saltwater intrusion in the east coast of the Yucatan Peninsula, Mexico Yujun Deng1 • Caitlin Young2 • Xinyu Fu1 • Jie Song1 Zhong-Ren Peng1 • Received: February 2016 / Accepted: October 2016 Ó Springer Science+Business Media Dordrecht 2016 Abstract Saltwater intrusion is a major hazard to coastal communities as it causes degradation of fresh water resources The impact of rising sea level on the saltwater intrusion into coastal aquifers has been studied for decades, but how human activities affect the extent of saltwater intrusion is poorly understood Human activities are known to influence groundwater availability indirectly by affecting precipitation patterns and directly by extracting groundwater and reducing recharge In this paper the authors investigated the integrated impacts of human activities and rising sea level on aquifer recharge in Quintana Roo, Mexico, by incorporating anthropogenic impacts on groundwater recharge into an analytical saltwater intrusion model The impact of human activities on groundwater extraction was firstly calculated; then, the resulting groundwater recharge was used in a Ghyben–Herzberg analytical model to determine the inland distance of saltwater intrusion The analytical model tested six scenarios stemming from different combinations of human development patterns, hydrological settings, hydraulic conditions and rising sea level to obtain the range of possible inland movement of saltwater intrusion Our results indicate that the groundwater recharge will decrease to 32.6 mm year-1 if human activities increase by 50 % more With 1-m sea level rise, inland saltwater intrusion distance is estimated to be up to 150 and km under head-controlled and flux-controlled scenarios, respectively A sensitivity analysis of the model reveals that the large hydraulic conductivity of the Quintana Roo aquifer (0.26–68.8 m s-1) is the most important factor in determining saltwater intrusion distance Therefore, in this aquifer, the response to human activities is greatly exceeded by natural hydrogeological conditions Keywords Rising sea level Saltwater intrusion Human activities Coastal water resources & Zhong-Ren Peng zpeng@dcp.ufl.edu Department of Urban and Regional Planning, University of Florida, Gainesville, FL 32603, USA Department of Geological Sciences, University of Florida, Gainesville, FL 32603, USA 123 Nat Hazards Introduction Saltwater intrusion is the invasion of saltwater into fresh groundwater in coastal aquifers due to changes in hydraulic gradient at the coastline (Bruington 1972; Barlow and Reichard 2010) It is known to contaminate potable water wells and even jeopardize water supply in vulnerable communities that rely on sole source aquifers (Vengosh and Rosenthal 1994; Feseker 2007; Germain et al 2008; Barlow and Reichard 2010; Liu and Dai 2012; Cai et al 2015) Whereas some communities are more resilient to the saltwater intrusion due to local environmental characteristics such as abundant precipitation, low evapotranspiration or low hydraulic conductivity, all of which can mitigate the effects of saltwater intrusion (Feseker 2007; Qi and Qiu 2011) However, the local environment can be altered by human activities through changing land use, increasing the amount of impervious surface as well as increasing groundwater extraction, which in turn alter fresh water stores in the local aquifer (Nicholls and Cazenave 2010; Beck and Bernauer 2011; Goudie 2013) Thus, it is essential to consider the impacts of human activities to obtain an accurate estimation of saltwater intrusion in association with rising sea level The overexploitation of groundwater in Taiwan and coastal Tunisia both altered local groundwater hydrology, which allowed for extensive saltwater intrusion and led to eliminate many potable water wells (Willis and Finney 1988; Narayan et al 2003; El Ayni et al 2013) As a consequence, local residents may continue to consume water from salt-contaminated wells under clear health risks In the coastal communities of Bangladesh that experiences severe saltwater intrusion, people consume much more salt compared to people inland, thereby putting them at risk of hypertension (Rasheed et al 2014) In response to the potential hazards, the shrinkage of local fresh water reservoirs caused by saltwater intrusion necessitates an effective community planning, which requires an understanding of what drives saltwater intrusion in the local aquifer Rising sea level is recognized as one important factor increasing saltwater intrusion, but the extent of intrusion is also related to the aquifer composition and local hydraulic conductivity (Werner and Simmons 2009; Chang et al 2011; Guha and Panday 2012; Carretero et al 2013; Werner et al 2013) Current estimates of sea level rise predict rates will exceed 3.0 mm year-1, while this number could go even higher depending on the stability of the West Antarctic ice sheet (Church and White 2006; Rahmstorf 2007; Nicholls and Cazenave 2010; Church et al 2013) The fifth assessment report of Intergovernmental Panel on Climate Change (IPCC) points out that sea level will continue to rise more than m by 2100 in the worst case (Pachauri et al 2014) Thus, coastal communities will be faced with the intensified saltwater intrusion and potential water shortage with sea level rise (Werner and Simmons 2009; Chang et al 2011; Guha and Panday 2012; Carretero et al 2013) Nonetheless, it is possible to decrease the effects of saltwater intrusion by implementing sustainable water management policies Many pumping strategies have been proposed to manage saltwater intrusion, but these strategies are site specific (Willis and Finney 1988; Hallaji and Yazicigil 1996; Cheng et al 2000; Mantoglou 2003; Sreekanth and Datta 2013; Cai et al 2015) In addition to site-specific pumping strategy, both current and future development projections should be included in the water management plan to sustain local water resources (Alley et al 1999) Moreover, many researchers have pointed out that human activities impose compounding impacts on environment including alteration in temperature and precipitation, which results in changes in evapotranspiration (Santer et al 1996; Vitousek et al 1997; Tett et al 1999; Mitchell et al 2001; Thorne et al 2002; Hegerl et al 2004; Sun et al 2008; Min et al 2011) Therefore, an effective water management plan 123 Nat Hazards must start with an understanding of the synergy between local hydrogeology and human activities in order to determine where and how much human activities actually control the extent of saltwater intrusion The increases in human activities, e.g., population, vehicle and agricultural land, would contribute to the growth in water consumption and subsequently the groundwater extraction (Gleick 1996; Peters and Meybeck 2000; Voăroăsmarty et al 2000; Schnoor 2015; Mehdizadeh et al 2015) Groundwater recharge can be viewed as the net of precipitation minus evapotranspiration, overland flow and groundwater extraction Therefore, managing human activities might be the only approach to sustain groundwater recharge as precipitation, evaporation and hydraulic condition are inherent properties of the climate and aquifer (Carretero et al 2013) In this study we incorporate human activities into the groundwater recharge term of a Ghyben–Herzberg and the Dupuit-Forchheimer solution analytical hydrological model in order to determine how humans will alter the hydrological balance in a karst aquifer undergoing sea level rise Including human activities in estimating saltwater intrusion should provide better understanding that how human impacts will alter the hydrogeological balance during saltwater intrusion which will lead to more effective water management planning The authors employ a partial least square regression (PLSR) model to determine the impact of human activities on aquifer recharge and extraction (Q) PLSR was developed as an econometric tool, but it is also widely used in chemometrics, chemical engineering and monitoring industrial processes This method can construct a dependable mathematical model with adequate predictive power even when factors are many and highly collinear (Tobias 1995; Martens and Dardenne 1998; Helland 2006; Rosipal and Kraămer 2006; Jie et al 2007; Mevik and Wehrens 2007; Carrascal et al 2009; Abdi 2010) In this study the authors use PLSR to predict the impact of varying extent of human activities on groundwater extraction and groundwater recharge These values are then incorporated into a saltwater intrusion model to estimate the inland distance of saltwater intrusion under a scenario of up to 1-m sea level rise To illustrate the integrated impacts of human activities and rising sea level on the saltwater intrusion, six scenarios resulted from the combination of four human development patterns, three hydraulic conditions and up to m of sea level rise, were tested under two hydrological settings: (1) flux-controlled scenarios, where the groundwater discharge rate will stay still despite the rising sea level, and (2) head-controlled scenarios, where the inland aquifer head will not rise despite the rising sea level In any particular place, it is possible to have either flux-controlled or head-controlled scenario or the combination of them (Carretero et al 2013) Thus it is imperative to assess both models to provide information for water management In this study, we quantify the migration of toe of saltwater intrusion as a function of sea level rise, groundwater recharge, ground fresh water discharge, hydraulic conductivity and water table under sea level Study area The state of Quintana Roo is located on the east coast of the Yucatan Peninsula, Mexico The aquifer is comprised of a large carbonate platform with high hydraulic conductivity, with macroscale connections (conduits) between the fresh water aquifer and the coastal ocean The coastal zone has undergone rapid population and tourism growth from 1980 to 2012 According to the database of Instituto Nacional de Estadı´stica y Geografı´a (INEGI, http://www.inegi.org.mx/), the population was 220,000 in 1980 and increased to 1,300,000 123 Nat Hazards in 2012 (INEGI Information Databank; INEGI ‘‘Population and Housing Census 2010’’), simultaneously annual tourist influx increased from one million to nearly nine million (INEGI Information Databank) Therefore, growth in human activities has resulted in a heavily impacted coastal aquifer in Quintana Roo which is in need of a long-term water management plan This plan should account for the relationship of human activities with groundwater extraction and recharge Groundwater extraction in Quintana Roo has increased from 29.11 million cubic meters of water to 873 million cubic meters of water between 1980 and 2012 (Comisio´n Nacional del Agua) Quintana Roo has undergone increases in tourism for three decades, and projected annual tourism rate indicates accelerated rate of growth for the next decade Thus groundwater extraction, due to both expanding population and tourism, is likely to increase in the future Quintana Roo is comprised of a karst aquifer, which is a sole source aquifer for drinking water The minimum air temperature and the median air temperature both have increased 1–2 °C since 2004 (Comisio´n Nacional del Agua, ‘‘Servicio Meteorolo´gico Nacional’’ 2014) The average air temperature increased from 25.7 °C in 2004 to 27 °C in 2014, which leads to an increase in the evaporation rates The annual report of Comisio´n Nacional del Agua points out that the annual evaporation is larger than the annual precipitation Hence the aquifer recharge primarily occurs during the wet season when precipitation is greater than evaporation Average precipitation and evaporation in the wet season from 1980 to 2012 were 777.67 and 719.52 mm year-1, respectively Due to the hydraulic conductivity, overland flow is negligible as water rapidly infiltrates the highly fractured carbonate rocks (Comisio´n Nacional Del Agua [National Water Commission of Mexico] 2010, 2011, 2012, 2015) (Fig 1) Table gives the aquifer parameters used to estimate the migration of saltwater intrusion; these values were obtained from field experiments and derived from the literature A Nortek Vector Acoustic Doppler Velocimeter (ADV) was placed at an offshore spring vent Velocity components (u, v, w), temperature and pressure data were collected at 64 Hz in 10-min bursts every 30 Data beyond three standard deviations from the mean were considered spurious and were replaced by the average of the nearest neighbors (Monismith et al 2010) The conduit outflow data were multiplied by the area of the vent opening (0.62 m2) to estimate volumetric flow rate into and out of the mouth, q (m3 s-1) As the offshore spring cycled between discharge and backflow conditions, the average groundwater discharge rate was calculated by averaging the flow rate only during times of discharge over the two-week period of the field experiment (Fig 2) Hydraulic conductivity (K) for Quintana Roo is highly variable, with values ranging from 0.26 to 68.84 m s-1 (Bauer-Gottwein et al 2011) We used these maximum and minimum values to model the length of saltwater intrusion in addition to a K value calculated from data gathered during this study Two CTDs continuously recorded pressure measurements, one located at Pargos spring and the other at a Cenote located km inland The difference between tidal amplitude was used to calculate K, following a method suitable for karst aquifers (Martin et al 2012) Methodology 3.1 Partial least square regression As discussed, PLSR can construct predictive model even when predictors are highly correlated PLSR aims to extract the underlying or latent factors that account for most of 123 Nat Hazards Fig The state of Quintana Roo on the Yucatan Peninsula (inset) with study area outlined Enlarged study area shows Pargos, an offshore spring used to determine q0 and UNAM well, used to confirm values for hydraulic conductivity Table Aquifer parameters used in saltwater intrusion modeling Parameter Value Source Groundwater density ratio 40 Carretero et al (2013) Aquifer depth (m) 40 Gondwe et al (2010) Horizontal water discharge (q0) (m3 s-1) 0.48 Field data Hydraulic conductivity (m s-1) 68.8 Bauer-Gottwein et al (2011) Hydraulic conductivity (m s-1) 0.26 Bauer-Gottwein et al (2011) Hydraulic conductivity (m s-1) 0.66 Field data 123 Nat Hazards Fig Discharge volume of water from Pargos spring during March–April 2014 Positive values indicate times when the aquifer was discharging to the lagoon, while negative values indicate times when the lagoon was backflooding water into the aquifer Horizontal fresh groundwater flow to the sea per unit length of shoreline was calculated by averaging the positive discharge values and then multiplying this average by the average fresh fraction of the discharging water The average fresh fraction was determined from the average salinity of the discharging water the variation in the response although there are many predictors Figure indicates the schematic outline of PLSR, whose ultimate goal is to predict the response given by the predictors in the population To achieve this goal, the X score, t, and Y score, u, will be firstly extracted from the predictors (represented by matrix X) and responses (represented by matrix Y) in the sample Then scores t are used to predict u which will subsequently be Sample T Regression Extract Latent Variables U Extract Latent Variables Predictor Regression Response Population Predict Predictor Fig The schematic outline of PLSR 123 Response Nat Hazards used to predict the responses Generally, PLSR is a process of indirect modeling, constructing the prediction for responses given by the predictors To be more specific, PLSR iteratively obtained the components, called latent variables or principal components It starts with the singular value decomposition of the crossproduct matrix S ¼ X T Y, extracting information regarding the variation and correlation of X and Y The first left and right singular vectors, w and q, are the weights for X and Y, respectively To obtain X score and Y score: t ¼ Xw ¼ Ew ð1Þ u ¼ Yq ¼ Fq ð2Þ where E and F are initialized as X and Y The scores t and u are often normalized: t t ¼ p tT t 3ị u u ẳ p uT u 4ị p ẳ ET t 5ị q ẳ FT t ð6Þ To obtain the factor loadings: Finally, the information regarding this principal component must be subtracted from current data matrices E and F Enỵ1 ẳ En tpT 7ị Fnỵ1 ¼ Fn tqT ð8Þ Therefore, the next component must be estimated from the singular value decomposition of the cross-product matrix En?1Fn?1 The vectors w, t, p and q are stored in the corresponding matrices W, T, P and Q, respectively, after each iteration Then scores T can be used to calculate the regression coefficients, and later the normalized variables are converted back: 1 9ị R ẳ W PT W 1 B ¼ R T T T T T Y ẳ RQT 10ị The sum of squares of X explained by the principal components is calculated by PPT, and the proportion of explained variance is obtained by dividing the sum of explained variance by the corresponding sum of total variance The optimal subset of principal components is determined using cross-validation The authors developed a predictive model to reveal the impacts of human activities on groundwater extraction To be accurate, this study selected the predictors from the most common human activities in the following categories: industry, transportation, agriculture, livestock, economy, demographics and human waste Table shows the predictors selected to represent the human activities in different fields and the time variable, trend The variable, trend, was a dummy variable representing the effects of other factors 123 Nat Hazards Table Description of predictors Category Indicator Description Industry Volume of timber The timber is the primary industrial produce in Quintana Roo Automotive The increase in vehicles and road reveals the urban expansion and economic growth Transportation Motorcycle Length of road Agriculture Livestock Total area sown The increase in agricultural area indicates the growth in irrigation Cattle Cattle and poultry are the primary livestock in Quintana Roo Poultry Economy Tourists Tourism is the driving force of economic growth in Quintana Roo Lodging room Hotel Demographics Population Population growth increases the water demand Human waste Garbage Increasing garbage reflects the growth in human activities Time Trend Reduce the dynamics of data along time excluding human activities Then this model can to the best reduce the disturbance from other factors Due to the high collinearity, the authors applied PLSR to the predictors to extract the principal components, finding the largest variation in these predictors The resulting variation represented the hidden information behind the predictors, and this information should be able to explain the original data Afterward, these principal components extracted from predictors were used to construct a model with the principal components extracted from the response, groundwater extraction Subsequently, the regression between them later was converted back to the original predictors and responses To have a validly predictive model, the principal components should be able to explain 80 % or more variation in original data, and the cross-validation indices should be larger than 0.0975 as well 3.2 Saltwater intrusion length estimation In this study, both the Ghyben–Hertzberg relationship and the Dupuit–Forchheimer approximation were adopted to estimate the integrated impact of rising sea level and human activities on the saltwater intrusion for a homogeneous, isotropic, unconfined aquifer subjecting to a constant recharge and steady-state conditions The horizontal flux of fresh water at position (xi) was computed using the following equations (Falkland 1991): qi xi ị ẳ q0 Wxi ẳ K h ỵ ahÞðdh=dxÞ ð11Þ which was integrated and transferred into (Custodio and Bruggeman 1987) h2i ¼ 2q0 xi Wx2i K ð1 þ aÞ ð12Þ where xi (m) is the distance from inland to coastal shore, and hi (m) is the elevation of the water table at inland position xi (m), W (mm year-1) is the net recharge, q0 (m3 s-1) is the 123 Nat Hazards Fig Illustration of saltwater intrusion model variables used for the coastal aquifer in Quintana Roo, MX Groundwater recharge (W) varied according to human activities The new position of the saltwater toe (x) was modeled, and the movement of the salt water–fresh water interface was described as Xt Aquifer depth (z), horizontal movement of groundwater (q0) were determined from literature and field data The system exhibits multiple types of permeability leading to a range of hydraulic conductivity, from low values associated with matrix permeability to pipe-flow permeability associated with conduit flow horizontal fresh groundwater flow to the sea per unit length of shoreline, and K is the hydraulic conductivity (Table 1) The net recharge is calculated from the following equation: W ¼ precipitation evapotranspiration overland flow Q ð13Þ where precipitation and evapotranspiration stemmed from historical average value, overland flow is negligible given few surface water bodies are in Quintana Roo and Q is groundwater extraction which was calculated with PLSR model The ground water density ratio, a, is equal to qf/(qs - qf), where qf is the density of fresh groundwater and qs is the density of saltwater and a is assumed to be 40 In all the scenarios, the toe of the saltwater wedge is designated as xt (m), which is calculated as (Custodio and Bruggeman 1987): rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi q0 q20 K ð1 ỵ aịz20 xt ẳ 14ị W W2 Wa2 where z0 (m) is the depth of the aquifer below the mean sea level The rising sea level will change z0 (m) In saltwater intrusion predictions, the water table elevation (hi) is required and it can be determined by a simple steady-state mass balance and Darcy’s law to be: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi W hxi ị ẳ 15ị xi xtị q0 xi ỵ xT ị ỵ hT ỵ z0 Þ2 k where hT (m) is the water table height at the toe of sea water wedge, and xT (m) is the distance from the toe of salt water–fresh water interface to the inland 123 Nat Hazards In the flux-controlled scenarios, the horizontal discharge of fresh groundwater to the sea, q0 (m3 s-1), is constant despite the sea level changes However, the water table elevation, hi (m), will rise at the same rate of sea level rises to maintain a constant flow On the contrary, in the head-controlled scenarios the inland head is consistent regardless of the Table Modeling scenarios Scenarios Increase in human activities (%) Hydrological settings Hydraulic conductivity (m s-1) A 0–50 Flux controlled 68.8 0–1 B 0–50 Head controlled 68.8 0–1 C 0–50 Flux controlled 0.26 D 0–50 Head controlled 0.26 0–1 E 0–50 Flux controlled 0.66 0–1 F 0–50 Head controlled 0.66 0–1 Sea level rise (m) 0–1 Groundwater recharge was computed based on the increase in human activities which are stated in Sect 4.2 Then the results were used to estimate saltwater intrusion length in association with hydrological settings, parameter sets and sea level rise Table R-squared coefficient of groundwater extraction model t1 t2 t3 0.2319 0.0036 R-squared coefficients for each component Extraction model 0.7534 Table Explained variance of each variable in PLSR model Extraction model t1 t2 t3 Explained variance of X–Y by T 123 Automotive 0.8919 0.9953 0.9956 Motorcycle 0.7585 0.9945 0.9973 Tourists 0.9773 0.9833 0.9909 Lodging room 0.9723 0.9943 0.9973 Population 0.9904 0.9928 0.9939 Garbage 0.9670 0.9850 0.9857 Total area sown 0.7761 0.9931 0.9953 Length of road 0.7282 0.9282 0.9997 Stocks of cattle 0.8010 0.9673 0.9815 Trend 0.9779 0.9951 0.9951 Use 0.7534 0.9853 0.9889 Nat Hazards square value of 0.003571 at the third component Table shows the extent to which the information of each variable is captured by the components Almost all the variables were captured well by the first component, which was able to explain at least 73 % variation for both predictors and response Then the leftover information in the data was obtained by the second and third principal component, which extracted about 99 % information from original data in the end, enabling the authors to construct a validly predictive model The cross-validation results of this model at third component are shown in Table The h indices were computed using equation, Q2h ¼ PRESS RSSh1 , where RSSh-1 was the squared sum of residuals using the th-1 component, and PRESSh was the prediction error sum of squares using the th component A component th was considered to be significant if Q2h was Fig Comparison between origin and prediction in water consumption model Impacts of Human Activties on Groundwater 45.00 40.00 35.00 30.00 25.00 20.00 15.00 10.00 5.00 0.00 Zero Development (0%) Low Development (10%) Groundwater Recharge W (mm/year) Median Development (30%) High Development (50%) Groundwater Abstraction Q (mm/year) Fig Impacts of human activities for different development patterns ranging from to 50 % increase in human activities 123 Nat Hazards 30% Percent of q0 is fresh water a 1000 ΔT (m) 800 600 400 200 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.042 0.04 0.038 Water Recharge (m/year) 70% Percent of q0 is fresh water b 500 ΔT (m) 400 300 200 100 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 100% Percent of q0 is fresh water c 300 ΔT (m) 250 200 150 100 50 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) Fig Saltwater intrusion length for scenario A: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c 100 % of q0 is fresh water 123 Nat Hazards Fig Saltwater intrusion length for scenario B: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c c 100 % of q0 is fresh water In response to the decreasing water recharge and other parameters in scenario B, the qmin ranged from 60,329 to 53,852 m2 s-1 greater than or equal to LimQ2 = 0.0975 In other words, the model had significant predictive power at third component Table shows the regression coefficients of variables selected in the model Except for agriculture and livestock, the other human activities had positive coefficients, which collectively indicated that groundwater extraction will increase with the growth in human activities Specifically, the growth in transportation, tourism and demographics all contributed to the growth in groundwater extraction The transportation was more than a proxy to the local economy, the increase in transportation facilitated the delivery of bottled water, which has become the primary source of drinking water for the majority Mexicans (Green and MacQuarrie 2014) Though the per capita bottled water consumption in Mexico was already the greatest in the world, 69.8 gallons per capita in 2014, it is very likely that the number will continue increasing due to the deficient infrastructure of public supply system (International Bottled Water Association n.d.) Accordingly, the transportation will also develop to satisfy the demand of delivering the bottled water, which is extracted from more pristine aquifers away from population centers Figure shows the comparison between original data and prediction for groundwater extraction The goodness of fit for this model was acceptable given the prediction fitted the original data very well and the value of root-mean-square error (RMSE) was 25.46 The small value of RMSE indicated the small differences between original data and prediction More than that, the figure also revealed the trend that the groundwater extraction increased very quickly in recent years 4.2 Scenario analysis The critical variables tested in the scenario analysis include groundwater recharge (W), hydraulic conductivity and flux- versus head-controlled aquifer systems Groundwater recharge (W) was determined by using the PLSR model to determine how the growth in human activities will increase the amount of groundwater extraction (Q) and alter recharge due to changes in human activities, according to Eq 13 The results of this analysis are shown in Fig Starting with the current groundwater recharge and Q values of 40.1 and 17.3 mm year-1, respectively, we see marked decreases in both values with increasing human activities The maximum amount of increase calculated, 50 % more than current activities, will decrease the annual recharge to 32.6 mm year-1 and simultaneously increase Q to 25.5 mm year-1 The resulting W used in scenario analyses ranges from 0.032 m year-1 (0 % development) to 0.042 m year-1 (50 % development) Three values of hydraulic conductivity were used to explore the importance of human activities on saltwater intrusion The hydraulic conductivities were selected based on their representation in the literature and field-determined values The hydraulic conductivities used were 68.8 m s-1 for scenario A and B, 0.26 m s-1 for scenario C and D and 0.66 m s-1 for scenario E and F (Table 3) The maximum value was 68.8 m s-1, calculated by Bauer-Gottwein in 2011 Field-derived data were used to calculate the hydraulic conductivity between an inland Cenote and Pargos spring, which was found to be 0.66 m s-1 Finally, a minimal value 0.26 m s-1 was chosen as calculated by BauerGottwein 123 Nat Hazards 30% Percent of q0 is fresh water a 104 15 ΔT (m) 10 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 70% Percent of q0 is fresh water b 104 ΔT (m) 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 100% Percent of q0 is fresh water c 104 ΔT (m) 1 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 123 Nat Hazards Fig Saltwater intrusion length for scenario C: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c c 100 % of q0 is fresh water The maximum groundwater discharge (q0) measured was 0.48 m3 s-1 in study area (Fig 2) Three cases representing 30, 70 and 100 % fresh water discharge during offshore discharge were tested in each scenario The recorded salinity values for Pargos spring ranged between 22 and 35 Therefore a 30 % fresh water discharge scenario is closest to current conditions at this offshore spring Other offshore springs sampled as part of a larger study in Quintana Roo ranged in salinity from 14 to 27, so the scenarios encompass the range of observed fresh water fractions in discharging offshore springs (Null et al 2014) In flux-controlled scenarios, the groundwater discharge would not change despite the sea level rises and the head in aquifers would rise at the same rate as sea level rises In contrast, for head-controlled scenarios, the hydraulic gradient will shallow and groundwater discharge will decrease For head-controlled scenarios the minimal groundwater discharge in study area was calculated according to Eq 16 r WK ỵ aịz20 16ị qmin ẳ a2 where W is recharge, K is hydraulic conductivity, z0 is depth of aquifer, and a is groundwater density ratio as described above The groundwater discharge would at most decrease to qmin in head-controlled scenarios when sea level rises Flux- and head-controlled scenarios were tested using three different hydraulic conductivity values Assuming a hydraulic conductivity of 68.8 m s-1 and a 30 % fresh discharge (q0), the maximum inland penetration of saltwater is 1000 and 15,000 m under head- and flux-controlled scenarios, respectively (Figs 7, 8) In both cases an increase in human activities, and the consequent decrease in W, caused the saltwater to penetrate further inland Under a flux-controlled scenario (Fig 7) saltwater intrusion length would increase with the rising sea level no matter how much fresh water was discharged from aquifer to ocean However, as the fraction of fresh water in the discharging water decreased, the more saltwater intrusion penetrated inland Saltwater penetrated 1000 m inland under a 30 % fresh water scenario, but this distance decreased to 300 m when 100 % of discharge was fresh water The qmin ranged from 60,329 to 53,852 m2 s-1 here (Fig 8) Scenarios C and D use a hydraulic conductivity 0.26 m s-1, as which is a value observed in other studies of Yucatan karst aquifer The lower hydraulic conductivity resulted in less landward penetration of the saltwater toe (Figs 9, 10) In a flux-controlled scenario the maximum inland penetration of saltwater is m, which is significantly less than the maximum observed in scenario A, which reached a maximum of 1000 m inland Under head-controlled conditions when the discharge is 30 % fresh water the saltwater intrusion length is 700 m inland, and qmin ranges 3707–3515 m2 s-1 The final scenarios, E and F, utilized a hydraulic conductivity of 0.66 m s-1, which was determined from field data gathered in this study Flux-controlled scenario finds that when discharging groundwater (q0) is 30 % fresh saltwater will penetrate 10 m inland, but this value decreases to m inland when discharging groundwater is 100 % fresh (Fig 11) Under head-controlled conditions, scenario F, saltwater will penetrate 500–1600 m inland, depending on the fresh water composition of the discharging groundwater (Fig 12) In this scenario qmin ranges 5907–5207 m2 s-1 123 Nat Hazards 123 Nat Hazards Fig 10 Saltwater intrusion length for scenario D: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c c 100 % of q0 is fresh water In response to the decreasing water recharge and other parameters in scenario D, the qmin ranged from 3707 to 3515 m2 s-1 Discussion Results from this study found that human activities impacted both the amount of groundwater extraction (Q) and the amount of aquifer recharge A 50 % increase in human activities over today’s current level would result in a decrease of *10 mm year-1 in groundwater recharge and an equivalent increase in groundwater usage (Fig 6) The PLSR modeling revealed that transportation has a large impact on the amount of groundwater usage and recharge (Table 5) This information can be utilized by local governments to effectively manage water resources and develop strategies to offset local water usage as population expands Mahesha and Lakshmikant (2014) hypothesized that one effective way to minimize the saltwater intrusion is by identifying the optimal location and rate of fresh water pumping Our results confirmed that the rate of fresh water pumping is critical in determining the extent of saltwater intrusion Although human activities impact the extent of saltwater intrusion in Quintana Roo, the magnitude of their impact was small compared to how hydrogeological conditions impact saltwater intrusion For example, in scenario F, which utilized field-derived hydraulic conductivity value of 0.66 m s-1, differences in water recharge attributed to human impacts alter the inland salt migration distance by less than 100 m (Fig 12) In comparison, the fraction of fresh water discharging from the aquifer altered the inland distance of saltwater intrusion by more than km (Fig 12) Thus, the fraction of discharging fresh water played a key role in determining the length of saltwater intrusion Despite this, there are little data regarding the range of discharging water composition in Quintana Roo Measured salinity values in the Puerto Morelos lagoon area in this study range from 4.2 in groundwater discharging from the beach to 27.8 in water discharging from offshore springs (Null et al 2014) Therefore, uncertainty regarding the composition of discharging groundwater makes determining the exact extent of saltwater intrusion challenging Another critical factor affecting saltwater intrusion length is the hydraulic conductivity in Quintana Roo The saltwater intrusion length decreased with less hydraulic conductivity for both flow-controlled and head-controlled settings In the extreme case, when the hydraulic conductivity was 68.84 m s-1, the saltwater intrusion was able to penetrate 150 km, which would contaminate about 1860 billion cubic fresh water However, the loss would dramatically decrease to about 2640 million cubic fresh water if the hydraulic conductivity reduced to 0.26 m s-1 The dynamics of hydraulic conductivity in Quintana Roo makes the aquifer vulnerable to the sea level rise, whose impacts on saltwater intrusion in Quintana Roo were larger than in other well-studied coastal aquifers, where rising sea level was found to be the least significant factor in determining saltwater intrusion (Green and MacQuarrie 2014) Though the hydraulic conductivity and sea level rise to a great extent determined how saltwater intrusion affects local aquifers in Quintana Roo, the varied hydrological settings also played an important role As shown in Fig the aquifer is comprised of both matrix and conduit permeability Therefore it is likely that our predictions of saltwater intrusion are accurate at different scales Conduits allow for the rapid flow of water exchange between the coast and the aquifer A hydraulic conductivity of 68.8 m s-1 is reasonable in 123 Nat Hazards 123 Nat Hazards Fig 11 Saltwater intrusion length for scenario E: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c c 100 % of q0 is fresh water a conduit setting Thus within conduits it is likely that saltwater intrusion will extend 15 to 20 km, as results from scenario B suggest (Fig 8) The Quintana Roo aquifer also contains low permeability matrix rock, as modeled using a conductivity of 0.26 m s-1 Results from head-controlled modeling of this hydraulic conductivity (scenario F) indicated that saltwater penetration would reach a maximum of 1.5 km, an order of magnitude lower than penetration within conduits (Fig 12) Thus, the aquifer will likely have multiple salt water–fresh water interfaces that affect potable water quality for residents living in the coastal zone Compared with head-controlled settings, in which the saltwater intrusion length increased exponentially, the flux-controlled settings were much less vulnerable given the saltwater intrusion could penetrate much less than in head-controlled settings (Werner and Simmons 2009) However, the places where the flux-controlled setting dominates can be turned into head-controlled setting once the groundwater extraction exceeds its recharge, due to lowering of the water table (Carretero et al 2013) In order to maintain sufficient recharge for a flux-controlled system, water management agencies will need to actively control human activities, thereby preventing human alteration of the hydrological balance Currently, sufficient precipitation in Quintana Roo sustains groundwater recharge, whereas the karstic islands in Bahamas experience water scarcity although with abundant precipitation (Jones 2014; World Bank 2015) Presently evaporation rates not outstrip precipitation during certain portions of the year, but this could change due to changes in air temperature associated with climate change (Roderick and Farquhar 2002) A combination of increased air temperature associated with increased evaporation would decrease groundwater recharge in Quintana Roo, moving the system toward water scarcity as observed on small karst islands A perturbation in the fresh water–salt water balance may take years to re-establish During this time local inhabitants and tourism operations will need to invest in alternative methods to obtain fresh water These methods such as shipping or piping water from inland areas and desalinization will increase human activities, thereby creating a negative feedback loop that will continually deplete coastal fresh water resources Poor water resources management has caused many health-related problems in developing countries, and public involvement is essential to avoid such a result (Falkenmark and Widstrand 1992; Dungumaro and Madulu 2003) Our results indicate a clear and urgent need to understand the role of anthropogenic effects in saltwater intrusion due to the varied effects in different natural hydrogeological conditions, which necessitates the local study of anthropogenic effects In this paper the authors focused on how human activities alter water quantity, but water quality is equally important in karst aquifers with high hydraulic conductivity Moreover, a density-dependent numeric model would more accurately represent this hydrologically complex system, providing a more refined estimation of saltwater intrusion Our study demonstrates an increase in human activities over the next decade, which will lead to an increase in anthropogenic impacts on the aquifer The fresh water lens of the aquifer in Quintana Roo are already heavily contaminated by human sewage as evidenced by high levels of nitrogen and fecal coliform bacteria (Herna´ndez-Terrones et al 2011) Anthropogenic nutrient loading impacts seagrass meadows around offshore spring discharge points by introducing phosphorous after rain events and increasing nitrogen availability in coastal hotel zones (Carruthers et al 2005) A profound understanding regarding the impacts of human 123 Nat Hazards 30% Percent of q0 is fresh water a 10 ΔT (m) 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 70% Percent of q0 is fresh water b ΔT (m) 1 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 100% Percent of q0 is fresh water c ΔT (m) 2.5 1.5 0.5 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 123 Nat Hazards 30% Percent of q0 is fresh water a 2000 ΔT (m) 1500 1000 500 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 70% Percent of q0 is fresh water b 800 ΔT (m) 600 400 200 0.5 Sea Level Rise (m) 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) 100% Percent of q0 is fresh water c 600 500 ΔT (m) 400 300 200 100 0.5 Sea Level Rise (m) 123 0.032 0.034 0.036 0.038 0.04 0.042 Water Recharge (m/year) Nat Hazards b Fig 12 Saltwater intrusion length for scenario F: a 30 % of q0 is fresh water; b 70 % of q0 is fresh water; c 100 % of q0 is fresh water In response to the decreasing water recharge and other parameters in scenario F, the qmin ranged from 5907 to 5207 m2 s-1 activities on saltwater intrusion is indispensable for developing an efficient water use management given only human activities are manageable Conclusions The impact of sea level rise on the saltwater intrusion in karst aquifers was substantially different between flux-controlled and head-controlled scenarios Flux-controlled scenarios reduced the extent of saltwater intrusion substantially, independent of the hydraulic conductivity and composition of discharging groundwater In Quintana Roo hydraulic conductivity was the most significant factor in determining the extent of saltwater intrusion As aquifer permeability ranged from low permeable matrix rock to pipe-flow permeability of conduits, the extent of saltwater intrusion would not be uniform Instead, saltwater intrusion may extend for kilometers within conduits but only hundreds of meters in matrix rock Human activities alter the amount of recharge to the aquifer, which can alleviate the saltwater intrusion, but this effect is minor in comparison with the importance of discharging water composition and flux- versus head-controlled system Instead, the importance of human activities lies in the amount of aquifer recharge, which controls whether the aquifer is head or flux controlled Increasing human activities will lower aquifer recharge, thereby reducing inland hydraulic head and switching the system from flux to head control This switch dramatically increases the modeled extent of saltwater intrusion Future integrated modeling of hydraulic conditions and human activities should include water quality impacts in addition to water quantity Furthermore, the potential anthropogenic effects on hydraulic conditions, such as aquifer pumping and injection of wastewater, must be evaluated in order to fully understand the total impact of human activities on water resources These findings are applicable to other karst aquifers worldwide as they have similar hydraulic conditions, particularly high hydraulic conductivities and conduit flow connections to the coastal ocean Acknowledgments This study is funded by NSF Project 12-594 [Coastal SEES (Track 1): Planning for hydrological and ecological impacts of sea level rise on sustainability of coastal water resources] The authors appreciate the help of Professor Jonathan B Martin, Professor Andrew V Ogram and Professor Arnoldo Valle-Levinson from University of Florida, Daniel Miret and Dr Ismael Tapia Marin˜o from the Center for Research and Advanced Studies of the National Polytechnic Institute and other colleagues in this project References Abdi H (2010) Partial least squares regression and projection on latent structure regression (PLS regression) Wiley Interdiscip Rev Comput Stat 2(1):97–106 Alley WM, Reilly TE, Franke OL (1999) Sustainability of ground-water resources, vol 1186 US Department of the Interior, US Geological Survey Barlow PM, Reichard EG (2010) Saltwater intrusion in coastal regions of North America Hydrogeol J 18(1):247–260 Bauer-Gottwein P, Gondwe BR, Charvet G, Marı´n LE, Rebolledo-Vieyra M, Merediz-Alonso G (2011) Review: the Yucata´n Peninsula karst aquifer, Mexico Hydrogeol J 19(3):507–524 123 Nat Hazards Beck L, Bernauer T (2011) How will combined changes in water demand and climate affect water availability in the Zambezi river basin? Glob Environ Change 21(3):1061–1072 Bruington AE (1972) Saltwater intrusion into aquifers1 JAWRA J Am Water Resour Assoc 8(1):150–160 doi:10.1111/j.1752-1688.2012.00659.x Cai J, Taute T, Schneider M (2015) Recommendations of controlling saltwater intrusion in an inland aquifer for drinking-water supply at a certain waterworks site in Berlin (Germany) Water Resour Manag 29(7):2221–2232 Carrascal LM, Galva´n I, Gordo O (2009) Partial least squares regression as an alternative to current regression methods used in ecology Oikos 118(5):681–690 Carretero S, Rapaglia J, Bokuniewicz H, Kruse E (2013) Impact of sea-level rise on saltwater intrusion length into the coastal aquifer, Partido de La Costa, Argentina Cont Shelf Res 61:62–70 Carruthers TJB, Van Tussenbroek BI, Dennison WC (2005) Influence of submarine springs and wastewater on nutrient dynamics of Caribbean seagrass meadows Estuar Coast Shelf Sci 64(2):191–199 Chang SW, Clement TP, Simpson MJ, Lee KK (2011) Does sea-level rise have an impact on saltwater intrusion? Adv Water Resour 34(10):1283–1291 Cheng AD, Halhal D, Naji A, Ouazar D (2000) Pumping optimization in saltwater-intruded coastal aquifers Water Resour Res 36(8):2155–2165 Church JA, White NJ (2006) A 20th century acceleration in global sea-level rise Geophys Res Lett 33:L01602 doi:10.1029/2005GL024826 Church JA, Clark PU, Cazenave A, Gregory JM, Jevrejeva S, Levermann A et al (2013) Sea level change In Stocker TF et al (eds) Climate change 2013: the physical science basis Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change Cambridge University Press, Cambridge, pp 1137–1216 Comisio´n Nacional Del Agua [National Water Commission of Mexico] (2010) Statistics on water in Mexico, 2010 Edition National Water Commission (CONAGUA) Web 13 Aug 2015 Comisio´n Nacional Del Agua [National Water Commission of Mexico] (2011) Statistics on Water in Mexico, 2011 Edition National Water Commission (CONAGUA) Web 13 Aug 2015 Comisio´n Nacional Del Agua [National Water Commission of Mexico] (2012) Atlas del agua en Me´xico 2012 [Atlas of water in Mexico 2012] National Water Commission (CONAGUA) Web 13 Aug 2015 Comisio´n Nacional Del Agua [National Water Commission of Mexico] (2014) Atlas del agua en Me´xico 2013 [Atlas of water in Mexico 2013] National Water Commission (CONAGUA) Web 13 Aug 2015 Comisio´n Nacional Del Agua [National Water Commission of Mexico] (2015) Atlas del agua en Me´xico 2014 [Atlas of water in Mexico 2014] National Water Commission (CONAGUA) Web 13 Aug 2015 Custodio E, Bruggeman G (1987) Groundwater problems in coastal areas: a contribution to the International Hydrological Programme UNESCO, Paris Dungumaro EW, Madulu NF (2003) Public participation in integrated water resources management: the case of Tanzania Physics Chem Earth Parts A/B/C 28(20):1009–1014 El Ayni F, Manoli E, Cherif S, Jrad A, Assimacopoulos D, Trabelsi-Ayadi M (2013) Deterioration of a Tunisian coastal aquifer due to agricultural activities and possible approaches for better water management Water Environ J 27(3):348–361 Falkenmark M, Widstrand C (1992) Population and water resources: a delicate balance Popul Bull 47(3):1–36 Falkland A (1991) Hydrology and water resources of small islands: a practical guide UNESCO Press, Paris Feseker T (2007) Numerical studies on saltwater intrusion in a coastal aquifer in northwestern Germany Hydrogeol J 15(2):267–279 Germain DJS, Cohen DK, Frederick JJ (2008) A retrospective look at the water resource management policies in Nassau County, Long Island, New York J Am Water Resour As 44(5):1337–1346 Gleick PH (1996) Basic water requirements for human activities: meeting basic needs Water Int 21(2):83–92 Gondwe BRN, Lerer S, Stisen S, Marı´n L, Rebolledo-Vieyra M, Merediz-Alonso G et al (2010) Hydrogeology of the south-eastern Yucatan peninsula: new insights from water level measurements, geochemistry, geophysics and remote sensing J Hydrol 389(1–2):1–17 Goudie AS (2013) The human impact on the natural environment: past, present, and future Wiley, New York Green NR, MacQuarrie KTB (2014) An evaluation of the relative importance of the effects of climate change and groundwater extraction on seawater intrusion in coastal aquifers in Atlantic Canada Hydrogeol J 22(3):609–623 Guha H, Panday S (2012) Impact of sea level rise on groundwater salinity in a coastal community of South Florida1 JAWRA J Am Water Resour As 48(3):510–529 123 Nat Hazards Hallaji K, Yazicigil H (1996) Optimal management of a coastal aquifer in southern Turkey J Water Resour Plan Manag 122(4):233–244 Hegerl GC, Zwiers FW, Stott PA, Kharin VV (2004) Detectability of anthropogenic changes in annual temperature and precipitation extremes J Clim 17(19):3683–3700 Helland I (2006) Partial least squares regression Encyclopedia of statistical sciences Herna´ndez-Terrones L, Rebolledo-Vieyra M, Merino-Ibarra M, Soto M, Le-Cossec A, Monroy-Rı´os E (2011) Groundwater pollution in a karstic region (NE Yucatan): baseline nutrient content and flux to coastal ecosystems Water Air Soil Pollut 218(1–4):517–528 INEGI Information Databank, INEGI (n.d) Instituto Nacional De Estadı´stica Y Geografı´a, Mexico Web 12 Oct 2014 International Bottled Water Association (n.d.) Per capita consumption of bottled water worldwide in 2009 and 2014, by leading countries (in gallons) InStatista—The Statistics Portal Retrieved 26 June 2016, from http://www.statista.com/statistics/183388/per-capita-consumption-of-bottled-water-worldwidein-2009/ Jie Y, Jun F, Dexiu H, Jianwen Y (2007) Application of partial least-squares regression to safety monitoring of water conservancy projects Trans Chin Soc Agric Eng 2007 doi:10.3969/j.issn.1002-6819.2007.3 028 Jones JAA (2014) Water sustainability: a global perspective Routledge, London Liu CCK, Dai JJ (2012) Seawater intrusion and sustainable yield of basal aquifers1 JAWRA J Am Water Resour Assoc 48:861–870 doi:10.1111/j.1752-1688.2012.00659.x Mahesha A, Lakshmikant P (2014) Saltwater intrusion in coastal aquifers subjected to freshwater pumping J Hydrol Eng 19(2):448–456 Mantoglou A (2003) Pumping management of coastal aquifers using analytical models of saltwater intrusion Water Resour Res 39(12):1335 doi:10.1029/2002WR001891 Martens HA, Dardenne P (1998) Validation and verification of regression in small data sets Chemometr Intell Lab Syst 44(1):99–121 Martin JB, Gulley J, Spellman P (2012) Tidal pumping of water between Bahamian blue holes, aquifers, and the ocean J Hydrol 416:28–38 Mehdizadeh SS, Vafaie F, Abolghasemi H (2015) Assessment of sharp-interface approach for saltwater intrusion prediction in an unconfined coastal aquifer exposed to pumping Environ Earth Sci 73(12):8345–8355 Mevik BH, Wehrens R (2007) The pls package: principal component and partial least squares regression in R J Stat Softw 18(2):1–24 Min SK, Zhang X, Zwiers FW, Hegerl GC (2011) Human contribution to more-intense precipitation extremes Nature 470(7334):378–381 Mitchell JFB, Karoly DJ, Hegerl GC, Zwiers FW, Allen MR, Marengo J (2001) Detection of climate change and attribution of causes In: Houghton JT et al (eds) Climate change (2001): the scientific basis Cambridge University Press, Cambridge, pp 638–695 Monismith SG, Davis KA, Shellenbarger GG, Hench JL, Nidzieko NJ, Santoro AE, Reidenbach MA, Rosman JH, Holtzman R, Martens CS, Lindquist NL, Southwell MW, Geninf A (2010) Flow effects on benthic grazing on phytoplankton by a Caribbean reef Limnol Oceanogr 55(5):1881–1892 Narayan KA, Schleeberger C, Charlesworth PB, Bistrow KL (2003) Effects of groundwater pumping on saltwater intrusion in the lower Burdekin Delta, North Queensland In: MODSIM 2003 International Congress on modelling and simulation, vol 2, pp 212–217 Nicholls RJ, Cazenave A (2010) Sea-level rise and its impact on coastal zones Science 328(5985):1517–1520 Null KA, Knee KL, Crook ED, de Sieyes NR, Rebolledo-Vieyra M, Herna´ndez-Terrones L, Paytan A (2014) Composition and fluxes of submarine groundwater along the Caribbean coast of the Yucatan Peninsula Cont Shelf Res 77:38–50 Pachauri RK, Allen MR, Barros VR, Broome J, Cramer W, Christ R, et al (2014) Climate change 2014: synthesis report Contribution of Working Groups I, II and III to the fifth assessment report of the Intergovernmental Panel on Climate Change Peters NE, Meybeck M (2000) Water quality degradation effects on freshwater availability: impacts of human activities Water Int 25(2):185–193 Population and Housing Census 2010 (n.d.) Retrieved September 2014, from http://www.inegi.org.mx/est/ contenidos/proyectos/ccpv/cpv2010/Default.aspx Qi SZ, Qiu QL (2011) Environmental hazard from saltwater intrusion in the Laizhou Gulf, Shandong Province of China Nat Hazards 56(3):563–566 Rahmstorf S (2007) A semi-empirical approach to projecting future sea-level rise Science 315(5810):368–370 123 Nat Hazards Rasheed S, Jahan S, Sharmin T, Hoque S, Khanam MA, Land MA et al (2014) How much salt adults consume in climate vulnerable coastal Bangladesh? BMC Public Health 14(1):584 Roderick ML, Farquhar GD (2002) The cause of decreased pan evaporation over the past 50 years Science 298(5597):1410–1411 Rosipal R, Kraămer N (2006) Overview and recent advances in partial least squares In: Saunders C, Grobelnik M, Gunn S, Shawe-Taylor J (eds) Subspace, latent structure and feature selection Springer, Berlin, pp 34–51 Santer BD et al (1996) A search for human influences on the thermal structure in the atmosphere Nature 382:39–46 Schnoor JL (2015) Water unsustainability Daedalus 144(3):48–58 Servicio Meteorolo´gico Nacional (n.d.) Retrieved October 2014, from http://smn.cna.gob.mx/index php?option=com_content&view=article&id=12&Itemid=77 Sreekanth J, Datta B (2013) Design of an optimal compliance monitoring network and feedback information for adaptive management of saltwater intrusion in coastal aquifers J Water Resour Plan Manag 140(10):04014026 Sun G, McNulty SG, Moore Myers JA, Cohen EC (2008) Impacts of multiple stresses on water demand and supply across the southeastern United States1 JAWRA J Am Water Resour As 44(6):1441–1457 Tett SFB, Stott PA, Allen MR, Ingram WJ, Mitchell JFB (1999) Causes of twentieth century temperature change Nature 399:569–572 Thorne PW et al (2002) Assessing the robustness of zonal mean climate change detection Geophys Res Lett 29:1920 doi:10.1029/2002GL015717 Tobias RD (1995) An introduction to partial least squares regression In: Proceedings of the twenty annual SAS users group international, Orlando, FL, pp 2–5 Vengosh A, Rosenthal E (1994) Saline groundwater in Israel: its bearing on the water crisis in the country J Hydrol 156(1):389–430 Vitousek PM, Mooney HA, Lubchenco J, Melillo JM (1997) Human domination of Earth’s ecosystems Science 277(5325):494–499 Voăroăsmarty CJ, Green P, Salisbury J, Lammers RB (2000) Global water resources: vulnerability from climate change and population growth Science 289(5477):284–288 Werner AD, Simmons CT (2009) Impact of sea-level rise on sea water intrusion in coastal aquifers Groundwater 47(2):197–204 Werner AD, Bakker M, Post VE, Vandenbohede A, Lu C, Ataie-Ashtiani B et al (2013) Seawater intrusion processes, investigation and management: recent advances and future challenges Adv Water Resour 51:3–26 Willis R, Finney BA (1988) Planning model for optimal control of saltwater intrusion J Water Resour Plan Manag 114(2):163–178 World Bank (2015) World databank [Climate Change Knowledge Portal for Development Practitioners and Policy Makers] Retrieved from, http://sdwebx.worldbank.org/climateportal/index.cfm?page=country_ historical_climate&ThisRegion=Latin%20America&ThisCCode=BHS 123