Introduction

Water resource management in the Mediterranean is a fine balancing act that pits the regions’ socio-economic growth against negative externalities of its water use in terms of both present and future sustainability (Iglesias et al. 2007). In Greece, an EU member state, the Water Framework Directive (WFD) forms the basis for the management and planning of its water resources through an integrated river basin approach (WFD 2000). The ‘polycentric’ nature of governance that the WFD seeks to establish via the delegation of regulatory authority of river basin management to sub-national units within member states has been a challenge to effectively implement in southern European nations (Rauschmayer et al. 2009). In particular, the islands of the Aegean archipelago have found it difficult to support and facilitate the adoption of an Integrated Water Basin Management Plan (IWBMP). There have been numerous challenges arising from the different site specific administrative, geographical, physical, and socio-economic factors (EC 2015, p. 101; Demetropoulou et al. 2010) that resulted in the River Basin Management Plan (RBMP) for the Aegean Islands River Basin District (GR14) to be only recently completed, in September 2015 (SSW 2015). It has been five years since Kanakoudis and Tsitsifli (2010) raised the question whether Greece was falling behind schedule in implementing the WFD and nearly twelve years since its transposition in the Greek national legislation.

In the Aegean islands, an issue of fundamental concern is the manifestation of the impacts of climate change on supply and availability of freshwater resources. Projections by the fifth iteration of the Intergovernmental Panel on Climate Change (IPCC) point towards a ‘likely to very likely’ decrease in precipitation over the Mediterranean by 30–45%, especially if temperature rises by 1.4 °C (Jimenez et al. 2014; Bates et al. 2008). IPCC further projects that over the coming years, meteorological droughts in the region will increase in intensity and frequency (Jimenez et al. 2014). Apart from natural and climatic factors, social, economic, and behavioral features also interact with the hydrogeological system to determine the net stress. Groundwater predominantly contributes to the water supply and consumption in Lesvos (Kaldellis and Kondili 2007). However, increasing demand for water in irrigation and industrial application has led to the displacement of previously sustainable practices in water consumption (Gikas and Tchobanoglous 2009). Furthermore, as Gikas and Angelakis (2009) point out, although the potential for reclaiming water from wastewater exists in the Aegean, current water resource programs have failed to realize sustainability in their management by not integrating them into practice. In the past few months, the Aegean and the island of Lesvos have been used as a popular gateway for migrant crossing into the European Union creating additional pressure on its resources (Trubeta 2015). An essential issue in addition to that of water supply lies in the threat posed by the water quality deterioration through heavy metal concentration and saltwater intrusion, issues of concern in the Aegean as shown by recent studies (Karagiannis and Soldatos 2007; Kontis and Gaganis 2012; Symeonakis et al. 2014; Zkeri et al. 2015).

To conceptualize a water system it is imperative that its inherent inter-linkages with socio-economic, cultural, and institutional factors be acknowledged. Only a comprehensive effort that encaptures all issues surrounding water resources simultaneously can truly reflect the overall picture of the hydrogeological system. Liu et al. (2011) demonstrated an optimized approach which integrated potable and non-potable water systems for the Islands of Syros and Paros–Antiparos. Further, holism in such management approaches can be attained by the application of vulnerability assessment as a tool as it allows simultaneous consideration of socio-cultural and political dimensions of water resources in addition to natural, climatic, and environmental factors. Any assessment of vulnerabilities of a system is therefore by nature-, case-, and context-specific. To the best of our knowledge, no literature exists that documents a vulnerability-centric approach for the Aegean islands. This study thus considers vulnerability assessment as a tool to quantify such issues through a case study of the Lesvos Island. The IPCC’s definition of vulnerability is used but in the broader context of any change to system dynamics as ‘the degree to which a (water) system is susceptible to, and unable to cope with, adverse effects of (climate) change’ (Döll 2009). Vulnerability (V) of a system can be expressed as a reflective function of exposure (E) and sensitivity (S) as one aspect with adaptability or adaptive capacity (AC) as the other (Eq. 1).

$$\it V=f(E+S-{AC})$$
(1)

The endpoint approach to vulnerability assessment (Sullivan and Meigh 2005; Hamouda et al. 2009), a quantitative methodology to evaluate vulnerability through the development of a numerical expression (a composite index), was applied as this better addresses system characteristics (S, AC). This analysis ensures that the system is not gauged in isolation and accounts for both the traditional approach of assessing water resources as natural systems and the stress they encounter through anthropogenic forcing factors. The limitation to this approach lies in its failure to quantify risks or estimate the probability of manifestation of changes in system dynamics (Pielke and Guenni 1999); however, this will be addressed in a future study.

Through the application of an endpoint approach for the geographical area of Lesvos Island, this study seeks to provide insights into various conspicuous and inconspicuous issues surrounding water resources in the Aegean. To shift the focus towards resilience and identify approaches to ensure adequate preparedness in the system, we also present a methodological approach that (quantitatively) integrates the calculated vulnerability index with adaptive management. The principal motivation for the methodological approach detailed in this study is to allow policy makers to better identify strategies to address vulnerabilities of the system through an iterative, self–learning approach, and provide a pathway for resolution of complex water management–decision problems.

Methodology

It has been suggested that the use of indicators (indicator–based frameworks) and its synopsis could lead to the distortion of facts when used incorrectly or when the data itself are not applicable to the given situation (Komnenic et al. 2009). However, uncertainty in our predictions of environmental changes and their accompanying impacts, especially in identifying and/or establishing causality between the social and ecological systems have hampered decision-making processes (Jones 2000; Barnett 2001); this too, at a time when there is widespread recognition that adaptation mechanisms must be envisioned and put in place (Berrang-Ford et al. 2011; Field 2012). In this scenario, studies have also argued to emphasize the efficacy of utilizing proxy indicators in vulnerability profiling given the ease with which they can be communicated to stakeholders and decision-makers (Villa and McLeod 2002; Vincent 2004; Adger and Vincent 2005).

Hence, in the present work, indicators have been used as proxies for profiling the vulnerability of Lesvos and its hydrogeological system to arrive at a composite index. For indicator development, the well-known Driver-Pressure-State-Impact-Response (DPSIR) framework was employed as it allows identification of cause-effect relationships. Literature was surveyed to screen previous attempts at quantifying water vulnerability as well as previously developed indices to select relevant indicators therein (see, Lane et al. 1999; Pielke and Guenni 1999; Owens et al. 2001; Villa and McLeod 2002; Molle and Mollinga 2003; Vincent 2004; Sullivan and Meigh 2005; Rijsberman 2006; Döll 2009; Hamouda et al. 2009; Brown and Matlock 2011; Stathatou et al. 2016). The methodological framework for vulnerability assessment proposed by Stathatou et al. (2016) was adopted and modified to suit the purposes of the present study. Relevance was considered in order to reflect the applicability of the chosen indicators to the geographical context as well as their time applicability since many have been revised or modified over time. Theoretical comprehension and statistical interrelationships between the causes and the effects were used to inform this process along with conceptual understanding of the system dynamics as shown earlier by Adger et al. (2004). However, data availability and reliability constrained the choice of indicators. Careful consideration was given to avoid any overlaps in indicators as this would result in overestimation of its contribution; for instance, coefficient of variation in rainfall was chosen as this is more representative of the Lesvian hydrology under climate change scenarios rather than variation in surface water flow that was used by Hamouda et al. (2011) in their study on the Nile river basin. Consequently, a comprehensive although non-exhaustive set of 25 indicators was identified for the studied area. The indicators were then grouped and assigned to reflect their contribution to water vulnerability (Table 1). To measure the extent of this contribution, each indicator was gauged against thresholds that were derived from literature sourced from journal publications, reports of international organizations (for ex., United Nations, World Resources Institute), World Development Indicators (World Bank 2015), and/or national averages (Hellenic Statistical Authority 2010). In case of discrepancy between national and international thresholds for the indicators, the latter was favored and consistently used throughout the study.

Table 1 Indicators, thresholds, weights and contribution to vulnerability
Full size table

It is acknowledged that the usable water resources in a region or a river basin are always less than the potential water resources (Iglesias et al. 2007). Hence, water budget estimation was carried out to analyze the water input–output for the system through the law of conservation of mass. The Lesvian hydrogeology exhibits the characteristics of a closed system, allowing the estimations of average annual flows to be determined by the general water balance equation (Eq. 2); where, P is the annual precipitation, ET is total evapotranspiration, R is the surface run off, and I is the water infiltration

$$ P\ = ET +\ R\ +\ I$$
(2)

Study area

This study focuses on Lesvos Island, located in the northern Aegean Sea with an area of 1626 km2 and population of 86,436 inhabitants. Lesvos receives on an average, nearly 1 billion m3 of freshwater annually of which only 13% infiltrates and recharges its underground aquifer network. The remaining 62% of rainfall is lost as evapotranspiration and a further 25% ends up as surface runoff (Karavitis and Kerkides 2002). There are currently three reservoirs in Lesvos for storing surface water in Kalloni, Mythimna, and Chalandra Eresos with capacities 540,000, 560,000, and 2,450,000 m3, respectively. The water from the reservoirs is used for irrigation; however it is underutilized due to infrastructure limitations. Consistent with the majority of Mediterranean streams, all streams in the island are characterized by seasonal flow that only exists for a period of 5–6 months (from November to April). Subsequently, only 143 million m3 of water undergoes infiltration (Fig. 1). Moreover, the quantity of infiltration of water varies with the local scale precipitation features (intensity, duration, frequency of appearance) as well as the physiographic and geologic features of the area (extent, slope, ground cover, soil permeability, underground layers capacity).

Fig. 1
figure 1

Illustration for the water budget (million m3) and hydrographic network of Lesvos Island, Greece

Full size image

The high geodiversity of the Island significantly influences the hydrogeological characteristics of its aquifers and especially their hydrochemical properties (Kanellopoulos et al. 2015); local and regional geology have been known to determine groundwater geochemistry and vegetation (Kanellopoulos and Argyraki 2013). For instance, a study by Zkeri et al. (2015) found that the geological environment on Lesvos Island influenced the natural occurrence and mobilization of Arsenic in the groundwater. Lesvos is primarily covered with Neogene volcanic rocks formed during late Oligocene epoch, middle Miocene calc-alkaline rocks, and shoshonitic volcanic rocks (Kontis and Gaganis 2012). On the basis of its lithology, the following different types of aquifers can be found on the Island: alluvial, karstic, ophiolithic rock, Neocene formation, and volcanic rock.

Demand for freshwater primarily stems from the agriculture and household sectors. Irrigation accounts for nearly three quarters of the total water consumption on the Island despite half of the utilized agricultural area cultivated with olive groves that require zero to minimal irrigation (Lavee et al. 1990). As is typical in Mediterranean areas with low land diversity and productivity, animal husbandry of sheep and goat is an integral part of livelihood in the north western part of Lesvos. As illustrated by Kizos et al. (2013), this area is home to an intensified animal production system driven by the market and hence is a significant consumer of groundwater. Domestic consumption follows irrigation in demand for water; Mytilene, the capital and home to nearly half of the total population, accounts for 40% of the total household water use. However, Mythimna (or Molyvos), Lesvos’s tourist center, records the highest per capita water consumption. While water demand from industries is insignificant when compared to the other two sectors, minimal or no treatment of industrial wastewater represents a significant threat to groundwater quality (Halvadakis 2000).

Radar charts

The plotting of multivariate data on radial axes has been used extensively in all spheres of science. These plots employ a circular graphing method where a series of equiangular spokes/rays project from a central point radially outward with each spoke representing a different variable (Budsaba et al. 2000; Sun et al. 2010). The length of a spoke is proportional to the magnitude of the corresponding variable relative to the maximum value across the whole dataset for that variable (Saary 2008). For the purpose of the current work, radar charts were plotted to graphically represent the multivariate relationship between the indicators and their respective thresholds in two dimensions. The advantage of the radar chart in a vulnerability assessment lies in its concurrent visualization of all indicators for each functional category viz., E, S, and AC. Moreover, as suggested by Hamouda et al. (2009), the aggregation of vulnerability indicators into a composite index could lead to a potential loss of information. Although the initiative to do this has been taken in the present study, plotting of radar diagrams was also performed to counter this possibility. To ensure consistency in data representation, the method detailed by Hamouda (2006) was followed. Some indicators were displayed as absolute values while some were inverted to ensure uniformity in vulnerability trend. Following the standardization of all indicators, a cut-off value of 3 was chosen to avoid chart distortion and all threshold values are constant at 1.

Parametric normalization

The estimation of the vulnerability index (VI) entails aggregation of all the indicators into a single composite index. The prerequisite to perform aggregation is the reduction of dimensionality while retaining the variation in datasets (OECD 2005). To do this, all indicator values (x) were transformed into dimensionless parameters (X) by expressing them as a ratio of their respective thresholds (t). Moreover, since each indicator has to be brought to a common scale for cross-comparison, unity-based normalization (or feature-scaling) was performed to standardize the features of the data and/or the range of independent variables to a scale ranging from 0 to 1 (Y). Subsequently, the decision problem of identifying indicators (as proxies for issues affecting the water system) of high concern to the Island was decomposed into sub-problems by employing Analytical Hierarchical Process (AHP). AHP reduces the complexity in decision making by allowing methodical comparison between each element (indicator) in a pairwise fashion through the use of collected/calculated data and/or value judgements (Su et al. 2003). This makes AHP a unique process as it captures both subjective and objective aspects of the process (Saaty 2008). A Pairwise Comparison Matrix (PCM) was developed that calculated the Perron Eigen vectors for each indicator, i.e., the weight (w i) assigned to each indicator (Table 1). Since pairwise comparisons of a large set of criteria are usually prone to inconsistencies, a consistency ratio (CR) was calculated for with matrix outputs accepted subject to the condition that CR < 0.1.

Vulnerability index

The normalized value for each indicator was combined with its respective weight to arrive at the final score (Z). This process was performed independently for E, S, and AC. In addition, adaptive capacity was further divided into three sub-indices: natural and physical capacity; socio-economic and cultural capacity; and legal and institutional capacity with AHP performed for each independently. The total adaptive capacity was computed as the equally weighted average of the three sub-indices. The overall vulnerability index was computed using Eq. 3. The equation is further expanded by inserting the indicators for each functional unit as in Eq. 4 (See Table 1). The index provides the estimated value for a given year ranging from zero (not vulnerable) to one (highly vulnerable) and depicts the severity of the stress experienced by the system under consideration.

$$VI=E+S-AC$$
(3)
$${\text{VI = }}\left\{ {{\text{CV + PCWA + PCWC + WEI + RWS + WQI + WWD}}} \right\}{\text{ + }}\left\{ {{\text{PD + PG + AGIrr}}} \right\} - \left\{ \begin{gathered} {\text{VCC + DLoss + IRREff + RWU + EAP + GRDP + PBPL + HDI + LBAg + }} \hfill \\ {\text{IEC + S }}\& {\text{ E + GovEx + WWQStd + R }}\& {\text{ DEx + GovEff}} \hfill \\ \end{gathered} \right\}$$
(4)

Results

Exposure

Exposure refers to the nature and extent to which the water system experiences perturbation and is characterized by the magnitude, duration, and frequency of such changes (Adger 2006). As seen through Fig. 2, the indicators chosen to describe the exposure of the system reveal high vulnerability despite good supply of water as annual rainfall the island receives in relation to the demand stemming from various sectors. Likewise, this endowment exhibits low reliability as seen through the temporal variation in rainfall (0.215) which is at risk of exceedance of its threshold (0.3). Spatial variability further exacerbates this as mean rainfall in the western portion of the island (400 mm year−1) is significantly lesser than that in the eastern one (700 mm year−1). In particular, per capita water availability is well below 1700 m3 year−1, a threshold widely used by world relief organizations. IPCC’s projection of a likely decrease in annual precipitation in the Aegean in conjunction with increase in drought frequency could potentially cascade into reduced forest productivity and increased forest fires in the Island (Jimenez et al. 2014).

Fig. 2
figure 2

Indicators for exposure and water stress in Lesvos, Greece

Full size image

Similarly, non-climatic forcing factors also contribute towards high exposure of the system. Per capita water consumption (65.7 m3 capita−1 year−1) far exceeds its threshold which is also reflected in the high degree of exploitation of limited abstract-able water resources. This has translated into a gradual deterioration of water quality which is at medium risk of exposure (WQI = 64.11; t = 49). The local council-run water board (DEYA) is responsible for the treatment of domestic and some of the industrial wastewater but the amount of discharge is insignificant and well within the threshold. Although many industries in Lesvos have independent facilities for waste treatment, most, including many olive oil processing mills discharge their wastewater without adequate or no treatment into the Aegean Sea (Kapellakis et al. 2008). Regulations on the type and quantity of chemicals released to water bodies exist, though the data on their use are absent and at best, vague estimations can be reached (T. Kizos, Department of Geography, University of the Aegean, personal communication, June 17, 2015) and have not been included in the present analysis.

Sensitivity

Sensitivity can be described as the extent to which a water system is adversely (or beneficially) affected by perturbations resulting in its direct or indirect change. It is geared towards relating stress on resources and human vulnerability to changes in the systems governing those resources. Figure 3 illustrates the indicators investigated to analyze the Lesvian hydrogeological system’s sensitivity. Examination of the radar diagram indicated that population growth does not factor in contributing to vulnerability with population having stabilized over the last ten years to around 90,000 people on the island. However, potential localized stress on water (both, temporal and spatial) has increased in the recent months as the number of migrant crossings into the EU through Lesvos has increased significantly; for instance, during March–April 2015, migrants accounted for nearly 10% of the total population on the island (Trubeta 2015). Current levels of population density (53 inh km−2) pose considerable pressure on water resources and risk exceedance of the threshold (55 inh km−2); despite being within the region of acceptable stress, its relative proximity to the threshold points to potential resource deficit in the future. Significant vulnerability stems from irrigation-dependent cultivated area which is very close to the EU average for the same signifying high water dependency and demand (FAO 2015). Since most of the island’s population is involved in agriculture or agriculture-related activities, the system’s sensitivity to sustain human livelihoods in addition to ensuring its ecological sustainability is high. Trickle down effects of water stress from irrigation are likely to have impacts that cascade down to the ecosystem and affect its ability to continue providing services that the island’s communities depend on.

Fig. 3
figure 3

Comparison between sensitivity indicators and their thresholds for Lesvos, Greece

Full size image

Adaptive capacity

Adaptive capacity is indicative of a system’s ability to evolve when dealing with unpredictable human–ecosystem interaction and exogenous risk while trying to accommodate a wider range of variability in order to cope with them (Adger 2006). Adaptability of a system is a function of natural and physical structures, socio-economic fabric, institutional, and governance aspects at local, regional, and national levels with each representing different risks and impacts associated with the system (Hill 2013). The pressures on water resources described earlier are further exacerbated due to poor adaptive capacity as seen through Figs. 4, 5 and 6.

Fig. 4
figure 4

Lesvos natural and physical capacity indicators of water resource vulnerability

Full size image
Fig. 5
figure 5

Lesvos socio-economic & cultural capacity indicators of water resource vulnerability

Full size image
Fig. 6
figure 6

Lesvos legal and institutional capacity indicators of water resource vulnerability

Full size image

Poor natural capacity is observed with anthropogenic-influenced land-use change in the past 25 years documented to be twice of its threshold (Fig. 4). Gatsis et al. (2006) attributed this phenomenon to rapid urbanization, deforestation (cultivar expansion into natural vegetation), and overgrazing. A similar pattern is observed in the system’s physical capacity on account of high losses in the distribution network as well as irrigation water use efficiency. Losses in distribution stem from inefficiencies both at source and in pipelines for transportation.

Minimal to zero compliance with environmental regulations in the Lesvian industries is observed in Fig. 5 with most utilizing either independently operated wastewater treatment facilities with no regulatory oversight or discharging wastewater without prior treatment, a common practice in olive oil processing mills (Kapellakis et al. 2008). The lack of human resources and expertise in water management and research is also evident. Besides, the Gross Regional Domestic Product exceeds the threshold and is indicative of higher demand for water.

Legal and institutional frameworks allow regulation, maintenance, planning, development, and provisioning of water services. Enhancing this capacity through targeted government expenditure is crucial in making the system resilient to change. Analysis of these capacities revealed that water quality and wastewater quality discharge standards exist in compliance with the EU norms and the WFD (Fig. 6). However, government expenditure and promotion of water research and management have been weak, possibly in light of the new austerity and fiscal measures imposed on the country. Greece has consistently trailed other EU member states when it comes to investing in R&D programs (Belegri-Roboli and Michaelides 2006).

Vulnerability index

Through the application of AHP, weights were assigned to each indicator (see Table 1) and Eq. 5 depict the importance and contribution of each towards the overall water vulnerability.

$$\begin{aligned} {\text{VI}} & {\text{ = 0}}{\text{.33 }} \times {\text{ }}\left( {{\text{0}}{\text{.1169 }} \times {\text{ CV + 0}}{\text{.1612 }} \times {\text{ PCWA + 0}}{\text{.2125 }} \times {\text{ PCWC + 0}}{\text{.2851 }} \times {\text{ WEI + 0}}{\text{.0225 }} \times {\text{ RWS + 0}}{\text{.1852 }} \times {\text{ WQI + 0}}{\text{.0166 }} \times {\text{ WWD}}} \right){\text{ + 0}}{\text{.33 }} \times {\text{ }}\left( {{\text{0}}{\text{.3158 }} \times {\text{ PD + 0}}{\text{.2791 }} \times {\text{ PG + 0}}{\text{.4051 }} \times {\text{ AGIrr}}} \right) - {\text{0}}{\text{.11 }} \times {\text{ }}\left( { - {\text{0}}{\text{.1585 }} \times {\text{ VCC - 0}}{\text{.2354 }} \times {\text{ DLoss}} - {\text{0}}{\text{.6061 }} \times {\text{ IRREff}}} \right) \\ - {\text{0}}{\text{.11 }} \times \left( {{\text{0}}{\text{.0318 }} \times {\text{ EAP}} - {\text{0}}{\text{.0318 }} \times {\text{ GRDP + 0}}{\text{.0807 }} \times {\text{ PBPL + 0}}{\text{.2825 }} \times {\text{ HDI + 0}}{\text{.0888 }} \times {\text{ LBAg - 0}}{\text{.2406 }} \times {\text{ IEC}} - {\text{0}}{\text{.2438 }} \times {\text{ S }}\& {\text{ E}}} \right) \\ - {\text{0}}{\text{.11 }} \times {\text{ }}\left( {{\text{0}}{\text{.3296 }} \times {\text{ GovEx + 0}}{\text{.0895 }} \times {\text{ WWQStd + 0}}{\text{.0895 }} \times {\text{ WQStd}} - {\text{0}}{\text{.1539 }} \times {\text{ R }}\& {\text{ DEx + 0}}{\text{.4716 }} \times {\text{ GovEff}}} \right) \\ \end{aligned}$$
(5)

For the Lesvian water system, the composite index was calculated to be 0.69, demonstrating high vulnerability. Delving into each function constituting the index, exposure was found to contribute the most towards the index. Therein, water consumption, quality, and exploitation were identified as issues of highest concern and the same is reflected in their contribution to the overall vulnerability (Table 1). By inspecting the system’s sensitivity, it is evident that substantial alterations in vulnerability can occur even through small changes in agricultural practices, land-use change, and settlement density. It was also concluded that adaptive capacity of the island displayed poor ability to respond to disturbances in water dynamics.

Coupling vulnerability assessments with adaptive management

The inherent uncertainties in predicting the dynamics of a water system emanate not only from its ecological dimension but also through its anthropogenic forcing factors. Socio-ecological issues of systems can only be addressed through the improved understanding of behavioral and governance aspects such as the decision-making processes, mechanisms for regulatory oversight and approaches to enforces rules therein (Dutra et al. 2015).

However, it is not uncommon for the public and governmental institutions to approach third-party commissioned proposals or plans with skepticism given the lack of ownership therein and/or a lack of understanding of the nature of the nature of the problem or solutions (Park and Brorson 2005; O’Dwyer and Owen 2007). Consultancy groups come with “ready-made answers” or “blueprint prescriptions” that often lack effectiveness, implementability, and sustainability (Copestake and Williams 2013). The corollary of applying best practices to most if not all circumstances is what Matt et al. (2013) describe as ‘isomorphic mimicry’; the adoption of the on-paper reforms that camouflage the evident lack of demonstrability of the adopted response mechanism in improving developmental conditions.

In September 2015, the Aegean Islands River basin District (GR14) completed its River Basin Management Plan (RBMP) in line with the WFD (SSW 2015). Similar to other River Basin Districts (RBDs) in Greece, third-party consultants have been involved in this process (EC 2015), however, a “primary consultant,” the Ministry of Environment, Energy and Climate Change—Special Secretariat for Water (SSW)Footnote 1 existed to provide coordination, technical assistance, and guidelines. A special feature of this RBD is the existence of many smaller independent units; the islands that represent areas with distinctly different hydrological and socio–economic characteristics. This particularity imposes an additional level of complexity and the need for developing effective smaller scale, island-specific RBMPs, and water management policies rather than one (unavoidably more general) management plan at the RBD scale.

On the backdrop of this, it is thus imperative to rethink our approach to the management of local resources. What is evident is that there is no dearth of strategies to counter or adapt to changes that affect water resources, rather, viable approaches to adopt and enact the same. Systems characterized by complexities require a combination of several policy enforcements simultaneously as it is difficult to provide one ‘best’ policy that addresses the issues surrounding the resource (Westgate et al. 2013). One approach that utilizes the results of a vulnerability assessment (and its composite index) to minimize the risks to the system could be through adaptive management. Adaptive management refers to a systematic approach for improving resources management by learning from management outcomes (Pahl-Wostl 2007), in other words, “learning to manage by managing to learn.” Such an approach could potentially be inclusive allowing stakeholders to prioritize issues they wish to address and influence the decisions that affect the way they consume, supply, and distribute water. More importantly, not only would it allow dissemination of information, it could also improve water resource management through its inherent iterative process (Allen et al. 2011). Public engagement in monitoring programs extends the ownership of actions/programs that reduce vulnerability (Hahn et al. 2006). Considering these factors and observations, the following recommendations are put forward to allow improved water resources management:

  1. i.

    The commissioning of a comprehensive vulnerability assessment allows identification of issues of high priority (and hence, concern) through their contribution to the systems’ vulnerability. However, during the assessment, assigning weight factors to indicators must involve public participation to reflect their risk perception. This follows the corollary that the stakeholder’s and public’s demand for clean water must also be accompanied by their willingness to aid in the preparation of assessments and responses that manage their resources.

  2. ii.

    For every identified issue (through proxy indicators), a list of potential response measures and strategies must be proposed to counter them. This must be refined through a participatory approach that engages the local population, decision-makers, and on-site scientific community to source context- and location-specific solutions. This would also draw on the traditional ecological knowledge of communities and allow any recommendations put forward to be tailored to suit local conditions.

  3. iii.

    The selection of responses, best practices and strategies must acknowledge and identify constraints such as the availability of financial capital, availability of local resources, motivation of political entities involved along with their electoral priorities (given their mandate stems from it), potential legislative hurdles, and policy overlaps.

  4. iv.

    Workshops must be held to identify the preference of stakeholders and the public on various proposed measures to reduce water vulnerability; ideally, consensus must be reached for every vulnerability function (E, S, and AC) as it is only through a simultaneous approach that addresses every function that synergistic effects be created across the system.

  5. v.

    It is necessary to adopt targets for every preferred indicator as this creates a timeline in addition to motivating and holding accountable the actors involved for achieving the same. Moreover, consideration must be given to the timing of implementing these measures. For instance, reforms that require significant political capital must be implemented parallel to the term of the entities holding office.

  6. vi.

    In order to comprehend the change in a system’s vulnerability following the implementation of preferred response measures, it is essential for continued monitoring and evaluation to analyze the variations in the composite index as well as the sub-indices over time.

These recommendations have been illustrated and summarized in Fig. 7. The figure describes the sequential steps involved in an indicator-based vulnerability assessment and the coupled adaptive management approach which is seen as an iterative process. It is acknowledged that each issue (E n, S n, and ACn) could be addressed through the adoption of several potential responses (R i to R n ). Further, some responses could simultaneously address more than one issue; for instance, in the hypothetical example shown in Fig. 7, R1 is seen as the response to E1, S2, AC1, and AC2. The process of selecting the response measures as well as the number of issues that needs to be addressed is seen as a combined preferential function of local stakeholders, the public, and the entity that carries out the assessment while considering the constraints (C 1 to C n ) for implementing the same. Implementation is preceded by the setting of targets (T 1 to T n ) to allow quantification of the magnitude of change and the time frame to achieve them. It is in monitoring and evaluation that realization of the iterative adaptive management approach transpires; it allows change in the system’s vulnerability to be studied over time to depict any modifications as a result of enactment of the preferred measures. If our attempts to reduce vulnerability fail to achieve the agreed-upon targets, the process of identifying the next preferred response is carried out iteratively (and thus is an adapted response). Integrating adaptive management with vulnerability assessment could however have obstacles one of which is institutional. The premise of having meaningful public participation lies with the ability of the functioning governance structures to accommodate the process of adaptive management. Without this, the shift in power from decision-makers to the stakeholders would remain a tokenistic gesture (Carter and Howe 2006).

Fig. 7
figure 7

Conceptualization of coupled vulnerability assessment and adaptive management approach

Full size image

Conclusion

Given the documented failures of traditional methods to assess freshwater systems, this study utilized vulnerability assessment as a quantitative tool to characterize water resources on Lesvos Island, Greece as a case study. Analysis indicated that water resources in Lesvos when conceptualized through an indicator-based framework are under significant stress from both natural and anthropogenic factors. A composite index was compiled using a set of 25 indicators and calculated to be 0.69, indicating high vulnerability in the system to perturbations. Further, the prospect of coupling vulnerability assessments with adaptive management to address issues related to water systems was discussed. It was evident during the present work that few studies exist that quantify (or predict) the change that the adoption of potential management strategies can have over the vulnerability index. Such quantification would allow better solutions to management decision problems especially for issues where several strategies seem plausible.