Impact of climate extremes on hydrological ecosystem services in a heavily humanized Mediterranean basin

Ecological Indicators 37 (2014) 199–209

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Impact of climate extremes on hydrological ecosystem services in a heavily humanized Mediterranean basin ˜ a , D. Ennaanay b,c , H. Tallis c , S. Sabater a,d M. Terrado a,∗ , V. Acuna a

Catalan Institute for Water Research, Emili Grahit 101, Scientific and Technological Park of the University of Girona, 17003 Girona, Spain Riverside Technology Inc., 2950 E Harmony Road, Suite 390, Fort Collins, CO 80528, USA The Natural Capital Project, 371 Serra Mall, Stanford University, Stanford, CA 94305-5020, USA d Institute of Aquatic Ecology, University of Girona, 17071 Girona, Spain b

c

a r t i c l e

i n f o

Article history: Received 27 February 2012 Received in revised form 21 December 2012 Accepted 17 January 2013 Keywords: Ecosystem services Semi-arid basin Climate change Water scarcity Water quality

a b s t r a c t Climate change projections in the Mediterranean region are associated with more frequent extreme climate conditions, which could alter water availability and impact the delivery of ecosystem services. We assess the change in the delivery of three hydrological ecosystem services, one provisioning (water), and two regulating (water purification and erosion control), in the heavily humanized Llobregat River basin (Catalonia, NE Spain) in recently observed extreme wet and dry years. Results indicate that impacts on the delivery of services were especially important in dry years. The main sources of water supply were located in the northern part of the basin and they were the most affected by annual rainfall reduction. Drinking water and hydropower production were highly threatened in dry years, when benefits were almost 100% reduced with respect to the benefits obtained in normal years. The regulating service water purification provided higher benefits in dry years, when water quality was more likely to be compromised due to a decreased dilution capacity. Water purification benefits in normal years increased 127% in dry years. According to our results, no benefit was provided by water purification in wet years. Collectively, our findings emphasize that hydrological ecosystem services in semi-arid basins which are subject to chronic human pressure are very sensitive to the climate conditions of extreme years. We also find a spatial decoupling among areas of service supply and areas where the service is demanded. Management efforts in Mediterranean basins should consider both of these aspects. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Humans receive an array of benefits from the natural environment in the form of goods and services (Daily, 1997). Many ecosystem services are derived from freshwater and are commonly referred to as hydrological ecosystem services. These benefits are often regulated by terrestrial ecosystems and include provisioning services such as water supply for drinking, power production, industrial use and irrigation (Brauman et al., 2007), as well as regulating services such as water purification and erosion control (de Groot et al., 2010). The provision of hydrological ecosystem services is strongly dependent on watershed characteristics. Topography, land use/land cover (LULC), and climate have governing roles on the delivery of services (Brauman et al., 2007). Climate is one of the

∗ Corresponding author at: Catalan Institute for Water Research (ICRA), Emili Grahit 101, Parc Científic i Tecnològic de la Universitat de Girona, 17003 Girona, Spain. Tel.: +34 972 18 33 80; fax: +34 972 18 32 48. E-mail address: [email protected] (M. Terrado). 1470-160X/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ecolind.2013.01.016

major shaping factors in semi-arid basins, which present larger extremes than more humid areas. In semi-arid basins, climate is characterized by: (1) low annual precipitation but high intensity storms with significant spatial variability, (2) high potential evaporation, (3) low annual runoff with short-term high volume runoff, and (4) runoff losses in ephemeral channels (Branson et al., 1981). Both average conditions for normal and extreme years are important indicators of the hydrological regime in semi-arid regions, and therefore, both need to be considered when assessing services at the basin scale. Climate change is expected to intensify the hydrological cycle in semi-arid areas through the global increase in temperature, the concentration of rainfall in shorter periods of the year, and the more extended droughts (Hisdal et al., 2001). The Mediterranean region has been globally identified as one of the most vulnerable to global change (Schröter et al., 2005). Different potential impacts are projected for the region, including increased temperatures and reduced vegetation. Associated human impacts through changes in ecosystem services could include drinking water shortages, increased risk of forest fires, shifts in the distribution of species, and agricultural losses among others (Schröter et al., 2005).

200

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Mapping of ecosystem services has been a major topic at the regional to global scale (Eigenbrod et al., 2010), and often has been based on proxy-based maps (Costanza et al., 1997; Chan et al., 2006; Nelson et al., 2009). Several studies have assessed the effects of human changes on LULC, but less have focused on the impact of climate conditions in extreme years (Schröter et al., 2005; Metzger et al., 2008). There are some examples of the effects of extreme wet and dry years on water availability in areas of high water demand (Booker, 1995; Guo et al., 2000; Delpla et al., 2009). However, these effects are poorly understood in heavily humanized systems. Even in watersheds receiving large human pressures we can identify areas that deserve a higher protection since they are naturally providing ecosystem services that otherwise would need to be obtained artificially. Understanding how average climate conditions in extreme years affect these areas is essential to sustain a particular level of services’ delivery in the context of climate change. However, it is important to keep in mind that future extreme years for the Mediterranean area under global change could look nothing like the recent extremes assessed here. The ability of ecosystems to mediate hydrologic response to climate extremes is unclear yet, potentially important, and likely not linearly related to the delivery of hydrological services under average climate conditions in normal years. In this paper we assess the effects of recently observed extreme wet and dry years on the delivery of hydrological ecosystem services in a Mediterranean basin. Any disturbance to the hydrological regime is expected to impact annual water availability as well as nutrient and sediment dynamics in the basin. The Llobregat basin

(Catalonia, NE Spain), an area of 4950 km2 , is typical of semi-arid conditions and constitutes an example of a highly populated, highly impacted and severely exploited area in the Mediterranean region. We apply a spatially explicit modeling tool to evaluate the delivery of three ecosystem services, one provisioning (water), and two regulating (erosion control and water purification), in the basin. These are essential services in semi-arid areas, where water scarcity can constrain water-reliant activities. In the Llobregat basin, water scarcity is exacerbated by its extractive use for industry, human consumption, and agriculture. These activities contribute to the degradation of water quality (Sabater et al., 1987; Terrado et al., 2009; López-Doval et al., 2010). Erosion is also a major concern in many semi-arid areas worldwide, and is expected to increase under increasing land-use changes and flood frequency. In this study, we are interested in determining which parts of the basin would be the most impacted by the effect of average climate conditions in extreme years. We expect some of the assessed benefits to largely decrease under water scarcity, in particular those related to water provisioning. 2. Materials and methods 2.1. Study site The Llobregat basin (Catalonia, NE Spain) covers an area of 4950 km2 (Fig. 1). The river, which is 156.5 km long, has its headwaters in the Pyrenees mountains and flows southward into the Mediterranean Sea near the city of Barcelona. It is the main water

Fig. 1. Overview map and land use land cover of the Llobregat basin.

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

201

Fig. 2. Conceptual approach followed to link ecosystem services and human wellbeing.

source for Barcelona and its metropolitan area, with a population of more than 3 million people. Climate in the Llobregat basin is Mediterranean with strong seasonal fluctuations in temperature and rainfall. Annual rainfall varies substantially within the basin from more than 1000 mm in the Pyrenees to less than 600 mm near the coast. Three large reservoirs are located in the upper part of the basin: La Baells (115 × 106 m3 ), Sant Ponc¸ (24 × 106 m3 ), and La Llosa del Cavall (80 × 106 m3 ). A drinking water treatment plant (WTP) is located close to the outlet (Fig. 1). 2.2. Selected hydrological services Services were estimated after the conceptual framework of Haines-Young and Potschin (2010), which describes the pathway from ecosystem structures and processes to human wellbeing (Fig. 2). In the mentioned approach, “service” represents the supply of all possible benefits, while “benefit” involves the use of a service by humans even though no change in actual wellbeing is projected. Only certain benefits were assessed for each service, in the terms defined in the box “biophysical value” (Fig. 2). Water available for drinking and hydropower production were the benefits assessed for the water provisioning service. Drinking water constitutes the most important annual consumptive demand of water resources in the Llobregat basin (65%), followed by industry (25%), agriculture (8%) and livestock (2%) (Catalan Water Agency, 2002). Water purification is a regulating service provided by ecosystems through the retention of pollutants, ultimately preventing them from reaching the water course. The benefit assessed for water purification was higher water quality, defined here in terms of total nitrogen (TN) and total phosphorus (TP) concentration. Erosion control is a regulating service provided by ecosystems through sediment and soil retention. The assessed benefit for erosion control was avoided sedimentation in reservoirs, as sedimentation can affect reservoir water capacity and functioning for hydropower generation. Certainly, the selected services represent a part of all the hydrological services delivered in the basin, and the selected benefits only represent the service partially. However, we are not aiming to quantify ecosystem services in absolute terms, but only to illustrate the effect that extreme years can have on three of the provided services as they are defined here. Services were calculated at average conditions for normal and extreme years observed in the recent past. Average conditions of rainfall and temperature for normal years in the Llobregat basin were obtained from Ninyerola et al. (2000) for the period 1951–2000. Both parameters were obtained using an empirical and statistical procedure. Data from meteorological stations were used to build and validate the model, which was based on a multiple regression analysis and its corresponding validation. Extreme dry and wet annual conditions were obtained from a study covering

30 years, from 1971 to 2000 (Llebot, 2010). Dry and wet conditions were calculated by averaging rainfall and evapotranspiration data of the 5 driest and wettest years of these series. The impact of extreme years on the provision of ecosystem services was then assessed by comparing their provision in normal years to the provision in extreme years. 2.3. Modeling approach Two types of freshwater related tools could be useful in this kind of assessment (Vigerstol and Aukema, 2011): hydrologic tools and ecosystem service tools. While the former provide a higher degree of detail and mostly focus on ecosystem service drivers, the latter provide a more general picture of ecosystem services and are more accessible to non-experts. Here we used an ecosystem service tool, InVEST or Integrated Valuation of Ecosystem Services and Tradeoffs (Kareiva et al., 2011; Tallis et al., 2011), to assess the impact of recent extreme dry and wet years on the selected hydrological services in the Llobregat basin. InVEST is a spatially explicit tool consisting of a suite of models that use biophysical and economic data and relationships to estimate biophysical levels and economic values of ecosystem services. The model runs in a gridded map at an annual average time step, and results can be reported in either biophysical or monetary terms, depending on the needs and the availability of data. Information requirements and outputs (rasters and alphanumeric data at the annual scale) for each of the modeled services are given in Table 1. Two kinds of outputs were obtained: rasters, providing information per cell, and tables, containing information at the basin or sub-basin level. These last values were the ones compared to real data in the calibration and correspond to information contained in Table 1. InVEST user’s guide can be consulted for further detail (Tallis et al., 2011). The conceptual approach followed in this work constitutes one of the possible ways of assessing the benefits provided by the selected services, but it is not the only one. Other benefits apart from hydroelectricity and drinking water can be obtained from water provisioning, such as water for industry, agriculture and livestock. The amount of water provisioned by each cell in the landscape was obtained by calculating the net hydrological balance (Service; Table 1). No surface-groundwater interactions or sub-annual temporal dimension of water supply were considered in the model. For drinking water, only the fraction of water available for drinking purposes at every cell was considered. It was calculated as the remaining water fraction after removal of the demand for other consumptive uses (Catalan Water Agency, 2002) and the regulated environmental flow allocated at the outlet of the basin (Catalan Government, 2006). Water returns from non-consumptive uses were added to the model results as the difference between water demand and consumption. Available water was then

202

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Table 1 Data requirements and outputs for the selected ecosystem services. ES

Water provisioning – drinking water

Step

Service

Benefit

Data requirements DEM (m) Land use/land cover (LULC) Effective soil depth (mm) Average annual rainfall (mm) Average annual reference evapotranspiration (mm) Plant available water content (fraction [0,1]) Maximum root depth (mm) Evapotranspiration coefficient Zhang coefficient [0,10] Consumptive use by LULC (m3 y−1 ) Watershed above the point of interest

Service Water provisioning – hydropower production

Water purification – NP retention

Benefit

Service

Benefit

Service Erosion control

Benefit

Same as for drinking water Calibration coefficient Turbine efficiency (%) Reservoir fraction for hydropower (m3 y−1 ) Average annual head (m) DEM (m) LULC Soil depth (mm) Water yield (mm, output from water provisioning) Export coefficient (g ha−1 y−1 ) Nutrient filtration efficiency (%) Allowed level of nutrient pollution (kg y−1 ) Watershed above point of interest DEM (m) LULC Rainfall erosivity (R) (MJ mm ha−1 h−1 y−1 ) Soil erodibility (K) (Mg ha h ha−1 MJ−1 mm−1 ) Crop factor (C) Management practice factor (P) Sediment retention efficiency (%) Slope threshold Reservoir dead volume (m3 ) Watershed above point of interest

compared to the real water demand, since only the actual use of water was considered a benefit (Benefit; Table 1). The same water balance was performed for hydropower production, this time subtracting the amount of water devoted to any consumptive use upstream of reservoirs, since it was considered non-available for energy generation. The whole amount of produced energy was considered to be used by humans, and hence constituted a benefit (Benefit; Table 1). Although more than 100 small hydraulic plants exist in the Llobregat basin, the lack of information about diversion concessions forced us to use only power stations located in reservoir systems. Parameters modified during the calibration process were the Zhang coefficient and the plant evapotranspiration coefficient (Table 1). The Zhang coefficient corresponds to a constant representing the seasonal rainfall distribution in the studied area. The evapotranspiration coefficient was the most relevant parameter during calibration. It is related to the characteristics of vegetation, it is specific for each land use, and it is used to calculate the potential evapotranspiration that controls the annual average water yield in the basin. Seven gauging stations located along the mainstem of the Llobregat River were used for calibration of the water provisioning model. Raw data from the gauging stations consisted of daily water discharge (Catalan Water Agency, 2002). Daily values were aggregated in an annual average in order to make them comparable to InVEST results, obtained at the annual time step (Table 2). The period 1950–2000 was considered for calculation of the annual water discharge, although it was also dependent on data availability at each station.

Process

Output

Calculates cell level yield as difference between rainfall and evapotranspiration

Annual average water yield (mm y−1 )

Subtracts water consumed for other uses and identifies points of extraction

Annual average water yield available for drinking purposes (mm y−1 )

as for drinking

as for drinking

Estimates power generated by water available for hydropower

Energy production (kWh y−1 )

Calculates nutrient export and retention

Nutrient export (kg y−1 ) Nutrient retention (kg y−1 )

Subtracts retention equal to amount of maximum allowed level and identifies treatment facility locations

Nutrient retention for water quality (kg y−1 )

Calculates sediment export and retention at each cell using USLE and routing

Annual average erosion (kg y−1 ) Annual average sediment retention (kg y−1 )

Subtracts sediment equal to dead volume

Annual average sediment retention to reservoirs (kg y−1 )

The benefits of water purification were assessed for TN and TP. Although these single measures ignore many other sources of water pollution, in this work we considered them representative of the water purification service. However, the service could also be assessed for other kinds of non-point source emissions if data on loading and filtration rates of the pollutants of interest were available. The amount of nutrients exported from each cell was estimated from export coefficients, which corresponded to annual averages of nutrient fluxes, whereas the retained amount of nutrients by each cell was a function of the retention coefficients associated with each LULC type. The model estimates how much nutrient from a given cell reaches a stream (i.e. is not retained by vegetation along the flowpath), how much each cell retains (Service;

Table 2 Observed values used for model calibration and predicted values for mean, dry and wet conditions. Parameter

Observed

Water supply (×106 m3 y−1 ) TN export (Mg y−1 ) TP export (Mg y−1 ) Sediment export – 1 (Gg y−1 ) Sediment export – 2 (Gg y−1 )

606 6000a 420 200b 602–1418c

Predicted Mean

Dry

606 5998 422 150 1535

121 5312 385 66 719

Wet 1586 6384 448 246 2623

Source: Data sources are Ludwig et al. (2009) (a), Catari et al. (2009) (b), and Liquete et al. (2009) (c).

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Table 1), and the total amount of nutrient reaching the point(s) of interest (e.g. a water treatment plant). In general terms, it is expected that a higher proportion of nutrients from cells located closer to the river network reach the stream, since they have less chances to be retained by vegetation. In this case study, the point of interest was a drinking water treatment plant near the outlet of the basin. Drinking water quality standards (drinking WQS) established by the European Drinking Water Directive 80/778/EEC (European Council, 1980) determine the allowed loads of N and P in drinking water. Allowed loads were therefore used to calculate the nutrient loads exceeding the drinking WQS that would reach the treatment plant if retention by vegetation was not provided (Production; Fig. 2). We ‘discounted’ the drinking WQS from nutrient production considering that, strictly concerning drinking water, nutrient retention below the drinking WQS is not accounted as human benefit (Benefit; Table 1) (Fig. 2). We consider human benefit as the investment that otherwise would be necessary to maintain nutrient loads in water below the established drinking WQS. The term “useful retention” in Fig. 2 refers to the retention exceeding the maximum allowed loads. However, even if the substance is below the drinking WQS, a further reduction can still be useful. We are aware of the limitations of using drinking WQS, which are the result of a policy process combining health risk together with economic aspects. For this reason, we also considered water standards for the ‘good ecological status’ (ecological WQS; European Water Framework Directive, European Council, 2000) to calculate the nutrient loads exceeding ecological standards. The parameter adjusted during the calibration process was the nutrient filtration efficiency for each LULC type (Table 1). This parameter describes the capacity of vegetation to retain TN and TP as a percentage of the amount of nutrient flowing into a cell from upslope. Note that nutrient filtration efficiency was assumed to be constant when working with annual averages. However, at the event scale, this parameter would depend on the load, meaning that with higher loads the filtration efficiency of vegetation would be much lower than with events of smaller size. TN and TP models were calibrated using nutrient data from a water quality station at the outlet of the basin operated by the Catalan Water Agency. Monthly data concentrations of nitrate and phosphate were available for some years (1995–2000) and were used to estimate annual nutrient loads. The FLUX software (US Army Corps of Engineers) was used, which establishes a relationship between nutrient concentration and daily flow. The method flow weighted average under stratification was used for the estimation of annual nutrient loads. Nitrate and phosphate concentrations were subsequently transformed to TN and TP using the ratios: N-NO3 /TN = 0.75 and P-PO4 /TP = 0.048 (Ludwig et al., 2009). Annual loads estimated for TN coincided with the value reported in the latter study (Table 2). Erosion control was calculated with an InVEST model that uses an adjusted version of the Universal Soil Loss Equation (USLE) (Wischmeier and Smith, 1978). Each cell contribution to the load of suspended sediment in the river was estimated by first calculating erosion with USLE, and then routing eroded soil down the flowpath and allowing retention by present vegetation. The amount of soil retained on each cell was estimated as a percentage, depending on the ability of vegetation on the cell to stop erosion on the cell itself and the ability to retain sediment in surface water running on from upstream cells (Service; Table 1). Only sheet-wash erosion was included in the model. The exclusion of linear erosion (rill-interrill, gully or stream bank erosion) was due to model limitations. The benefits of erosion control were only calculated upstream of reservoirs, since we were assessing avoided reservoir sedimentation. However, sediment retention by the remaining landscape can report other benefits such as prevention of losses in soil fertility, since more fertilizers need to be applied in agricultural lands with high erosion, or decrease in the cost of water treatment. In this

203

Fig. 3. Observed versus predicted annual water yield at 7 different gauging stations along the mainstem of the Llobregat River, and effect of correcting the predicted discharge values by the returns from non-consumptive uses.

case, the landscape upstream of a reservoir was assumed to have a maximum soil export allowance which is equal to the reservoir’s dead volume or volume that, when filled with sediment, impacts the reservoir’s function. The difference between total sediment production and maximum soil export allowance gives the load of sediment exceeding the reservoir’s dead volume that would reach the reservoir if there was no retention by vegetation (Fig. 2). Again, this “discounting” is used because we assume that any retention of sediment when fluxes are lower than those that would fill the dead volume does not provide a reduction in dredge costs, infrastructure maintenance or production potential (Benefit; Table 1). One of the parameters of the erosion control model was the rainfall erosivity factor. Erosivity was calculated from the rainfall series of the Catalan Meteorological Service (Catari and Gallart, 2010). As is the only factor in the model susceptible to change under different climate conditions, it was used for simulation of dry and wet extremes. Calibration of erosion control was done at La Baells headwater reservoir using temporal changes in the reservoir’s bathymetry (Catari et al., 2009). The parameters adjusted during the calibration process were the slope threshold and the sediment retention efficiency (Table 1), both impacting the amount of sediment reaching the point of interest. The slope threshold is a slope value describing landscape characteristics such as slope management practices. It corresponds to the slope where practices stop or switch to terracing or stabilization. The sediment retention efficiency is the capacity of vegetation to retain sediment as a percentage of the amount of sediment flowing into a cell from upslope, and is specific for every LULC. Results of the erosion control model were then validated with data reported by Liquete et al. (2009) at the outlet of the basin (Table 2). 3. Results 3.1. Model calibration Model fit for water supply was good (Table 2) and the obtained Nash–Sutcliffe model efficiency index (Nash and Sutcliffe, 1970) was 0.995. In the initial calibration of the water supply (black dots in Fig. 3), water demands in the model were considered 100% consumptive, but this estimate underpredicted by 20% the amount of water at the river mouth. The model fit expectedly improved when return flows from waste water treatment plants (WWTPs)

204

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Fig. 4. Average annual water provisioning in the Llobregat basin under normal conditions (map and legend on the left) and changes under dry and wet climate conditions (two maps and legend on the right).

and industry were added to the whole calculation (white dots in Fig. 3). With this adjustment, prediction errors were always lower than 8% when considering water returns from non-fully consumptive uses. The model fit for exported TN and TP at the outlet of the basin was also good (Table 2). However, the model underpredicted exported loads of sediment to La Baells reservoir (1; Table 2) and overpredicted exported loads at the outlet (2; Table 2).

3.2. Ecosystem services and impact of climate extremes The average annual water volume reaching the outlet of the basin in normal years decreased 80% in recently observed dry years, while it was 160% higher in wet years (Table 2). Higher values of water provisioning occurred in high-altitude areas, characterized by high water yield production and low demand (Fig. 4). These locations were the ones experiencing the highest decrease during dry years, providing between 75% and 100% less water than in normal years. The central and southern parts of the basin experienced lower changes during drought conditions. Even in some parts, the provision in normal years increased 800% during wet years (Fig. 4). Decrease in uplands water provisioning in wet years was attributed to differences in rainfall patterns related to the complex topography of the studied area (Martín-Vide et al., 2010). Due to this heterogeneity, average conditions for the whole basin did not necessarily reflect average conditions in each of its parts. Hence the interest of using a spatially explicit model. Despite the differential change between uplands and lowlands, the former are still the water source in both wet and dry years (Fig. 4). Water purification results need to be interpreted considering both nutrient retention and nutrient export. The sum of exported and retained loads equals the nutrient production in the basin. The retention magnitude depends not only on the characteristics of a particular landscape unit but also on the loads entering from upstream areas to that landscape unit. The nutrient export to the stream or point of interest depends on the retention capacity of vegetation downslope. As such, landscape areas located closer to the river network presented the highest values of nutrient export to the stream (Fig. 5a). This was especially the case at the lower river course, where export from parcels closer to the river had fewer opportunities to be retained before reaching the point of interest. Nutrient retention did not follow the same pattern as that observed for nutrient export. It was higher in flat areas contiguous to steep slopes and presented moderate values close to the river network (Fig. 5b). Patterns were similar for both TN and TP. The spatial

patterns identified for sediment export and retention were similar to patterns observed for nutrients. The areas neighboring the river channel had the highest sediment export values (Fig. 5c). Sediment retention was lower at the upper part of the basin because of the shorter flowpaths and the lower sediment loads coming from upstream sources (Fig. 5d). Effects of extremely wet or dry years on the annual exported nutrient loads were low and closely related to changes in the water yield. At the basin scale, TN and TP loads reaching the outlet in normal years were estimated to be 11% and 9% lower respectively in dry years. These loads increased by 6% in wet years (Table 2). Response to extreme years varied significantly within the basin. Nutrient export occurring in normal years in the central and southern areas considerably increased in wet years (from 50% to 200%). The northern areas experienced an important decrease in nutrient export under dry conditions. In dry years, annual sediment export at the basin scale was 53% lower than in normal years while in wet years it was 70% higher (Table 2). As in the case of nutrients, the effects of extremely wet or dry years on soil erosion were not homogeneous throughout the basin. South-western areas were identified as the most vulnerable to Hortonian runoff (described by the rainfall erosivity factor). Erosion in mean conditions was estimated to increase from 75% to 100% in wet years and, as a result, the value of sediment retention also increased.

3.3. Benefits of ecosystem services and impact of climate extremes Under average normal conditions, the Llobregat basin supplied 603 × 106 m3 y−1 of water that could be used for drinking purposes after appropriate treatment, while the real drinking water demand in the basin is around 300 × 106 m3 y−1 (Catalan Water Agency, 2002). As here supply exceeds demand, benefits correspond to the volume of demanded water. Drinking water demand remains approximately constant in the basin unless exceptional water restrictions are enforced during prolonged droughts. The highest drinking water provision was in areas at the northern part of the basin and upstream of urban areas (Fig. 6a). Water provisioning for hydropower production in the basin was quantified at 304 × 106 m3 y−1 , allowing the production of up to 65,270 MWh y−1 of energy. This value corresponds to the potential energy generation in the case all available water was used for hydropower generation. However, power stations are not working continuously and the actual amount of produced electricity would be certainly lower (a 30% of the reported value if 2500 h of annual functioning

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

205

Fig. 5. Landscape distribution of ecosystem processes describing the erosion control and water purification services in the Llobregat basin: average annual sediment export (a), average annual sediment retention (b), average annual total nitrogen export (c), and average annual total nitrogen retention (d).

were considered) (Catalan Water Agency, 2007). Areas located at higher altitudes contributed the most to hydropower generation, with >1000 kWh ha−1 y−1 (Fig. 6b). Areas upstream of La Llosa del Cavall reservoir received larger values for hydropower production per landscape unit than those upstream of La Baells. These results reflected the double use of water generated within the territory

upstream of La Llosa for energy production, which is afterwards directed to Sant Ponc¸ reservoir. Water purification was quantified as 6168 Mg of nutrient retention. In normal years, the benefit obtained from this service was exclusively derived from the retention of TN, since the established drinking WQS was only exceeded by nitrogen loads reaching the river but not by phosphorus loads (Table 3).

Table 3 Quantification of the water purification service in the Llobregat basin under mean and extreme climate conditions according to total nutrient load (export + retention) and maximum load allowed fulfilling drinking water quality standards (a) and ecological standards established by the European Water Framework Directive (b). Climate conditions

Nutrient

Total load (Mg y−1 ) b

Drinking WQS (Mg y−1 )

Service (Mg y−1 )a

Ecological WQS (Mg y−1 )

Service (Mg y−1 )b

Mean

TN TP

16,810 1115b

10,642 3215

6168 0

2128 321

14,682 793

Dry

TN TP

15,054a 1020b

1596 482

13,458 538

319 48

14,735 972

Wet

TN TP

18,943b 1232

23,957 7238

0 0

4791 724

14,151 508

a b

Nutrient export exceeded the drinking WQS. Nutrient export exceeded the ecological WQS.

206

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Fig. 6. Landscape allocation of annual hydrological ecosystem services provided by the Llobregat basin under normal conditions: water provisioning for drinking use (a), water provisioning for hydropower production (b), water purification for drinking water (c) and erosion control for reservoir maintenance (d).

The annual retention amounted to 15,475 Mg of nutrients using the ecological WQS instead of the drinking WQS. In this case, both TN and TP retention contributed to the benefit obtained by the water purification service, since ecological standards were exceeded by both nutrients (Table 3). Quantification of erosion control gave an annual retention of 980 Gg of sediment contributing to avoid reservoir sedimentation. As with energy generation, only the landscape upstream of the reservoirs was assessed for benefits in terms of avoided reservoir sedimentation. Again, the landscape upstream La Llosa del Cavall reservoir contributed to avoid sedimentation in two adjacent reservoirs (Fig. 6d). Drinking water provided in normal years was almost 100% lower in dry years and 150% higher in wet years. The provision of drinking water attained 2.4 × 106 m3 y−1 in dry years. In this case, demand was not fulfilled and benefits were considered to equal the total water provision (Fig. 7). In wet years, drinking water provision was 1488 × 10 m3 y−1 . In spite of the increase in reference to normal years, obtained benefits were kept equal to 300 × 106 m3 y−1 , since

water demand in the basin was considered quite constant (Fig. 7). Hydropower production attained in normal years was 98% lower in dry years. In this case, water provisioning allowed the generation of 988 MWh y−1 of energy. Energy produced in normal years slightly increased in wet years, when 70,133 MWh y−1 were produced (Fig. 7). Extreme years also affected water purification. The amount of service obtained in normal years was 127% higher in dry years, when 13,996 Mg y−1 of nutrients were retained. Retention of both TN and TP was more beneficial in dry years because nutrient emission exceeded the drinking WQS (Table 3). However, the drinking WQS thresholds were not exceeded by either nutrient in wet years (because of dilution) and therefore, following our conceptual framework, the obtained benefit was zero (Fig. 7). Benefits from erosion control for reservoir maintenance obtained in normal years were 58% lower in dry years, when sediment retention was 410 Gg y−1 . Benefits were 65% higher in wet conditions, when the amount of retained sediment was 1622 Gg y−1 (Fig. 7). In the Llobregat basin, dry years were characterized by less aggressive rains, meaning lower erosion and consequently lower retention, whereas wet years were characterized by more intense rains that increased both erosion and retention. 4. Discussion The vulnerability of ecosystem services to climate extremes was assessed with the aim of identifying the magnitude of change of key hydrological ecosystem services in recent extreme dry and wet years. Not surprisingly, higher levels of provisioning services were delivered under average and wet climate conditions, while the regulating service water purification contributed most to human wellbeing in dry conditions. 4.1. Magnitude and distribution of hydrological benefits

Fig. 7. Annual change in the provision of hydrological ecosystem services in the Llobregat basin under normal and extreme (dry and wet) climate conditions.

Water provisioning was not homogeneously distributed throughout the Llobregat basin. Mountainous areas located in the Pyrenees contributed the most to the provision of freshwater in downstream areas (Fig. 6a). These areas functioned as regional

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

“water towers” (Viviroli et al., 2007). The average amount of water provisioning in the Llobregat basin was 19,500 m3 ha−1 y−1 , in the range of global values reported elsewhere (Naidoo et al., 2008). Water returns from urban and industrial areas have been identified to be important in many arid and semi-arid basins worldwide (Horn et al., 2008). In the case of the Llobregat basin, water returns can represent up to 40% of the discharge at the river mouth, and correspond approximately to 60% of the demand. In normal years, the average amount of water available to the drinking water supply was 750 m3 ha−1 y−1 . The water provisioning service was characterized by a flow connecting the supply region (mountainous headwaters) with the area of highest demand (the Barcelona Metropolitan area). Hydrological services such as water provisioning often show this regional spatial component, with a clear differentiation between source and sink areas (Brauman et al., 2007). With respect to the basin’s potential hydropower production (813 kWh ha−1 y−1 in mean conditions, corresponding to 0.21 kWh m−3 ), a similar spatial pattern was also observed. This service was again supplied by areas upstream of the demand areas, in this case the reservoirs themselves. Moreover, the existence of two reservoirs located consecutively in the landscape meant that the water supplied to that region provided much higher benefits in terms of energy production (Fig. 6b). However, this statement reflects only a narrow point of view. Indeed, any detailed evaluation of the potential benefits of reservoirs should also consider their impacts on other ecosystem services such as gene pool protection (de Groot et al., 2010). The average value of drinking water purification in the Llobregat basin was equivalent to 12 kg ha−1 y−1 of nutrient retention. In normal years, this corresponded exclusively to the retention of TN. The estimated average nutrient loads produced by the landscape of the basin were 34 kg ha−1 y−1 of TN and 2.3 kg ha−1 y−1 of TP, of which a total of 60–65% were retained before reaching the water treatment plant. These loads were similar to values reported in different European watersheds (Grizzetti et al., 2005). The legally defined maximum allowable nutrient levels for drinking water (50 mg/L of nitrate and 5 mg/L of phosphate in P2 O5 form) played an important role in this estimation. Concentrations transformed to TN and TP loads (Ludwig et al., 2009) were sensitive to different flow conditions. Under average conditions, the actual TN and TP loads reaching the river network at the drinking water treatment plant (Export; Fig. 2) were below the drinking WQS. This meant that nitrogen and phosphorus pollution was not problematic in the Llobregat River with respect to the current drinking WQS (Ludwig et al., 2009). However, in the case of nitrogen, the estimated load production of TN exceeded the drinking WQS (Table 3). This meant that without the nitrogen retention provided by the landscape, water treatment technologies would be required in order to attain the desirable drinking water quality standards. As a result, TN retention in the Llobregat basin provided the benefit of higher water quality. As TP production was below the maximum allowable level, phosphorus retention was considered to provide no benefit for drinking water purposes (Table 3). Despite not being problematic in terms of drinking water quality standards, environmental concentrations of nitrogen and phosphorus could be sufficient to trigger eutrophication, biodiversity loss and ultimately other water quality problems (Sabater et al., 1987; Ricart et al., 2010). To account for these issues, we also used the ecological WQS established by the European Water Framework Directive (10 mg/L of NO3 − and 0.5 mg/L of P2 O5 ) to assess the water purification service. Ecological WQS were more restrictive than drinking WQS and, as a result, both TN and TP loads exported to the water treatment plant already exceeded ecological standards in normal years. This meant that all the TN and TP retention in the basin, corresponding to 15,475 Mg y−1 , contributed to a higher standard of ecological water quality (Table 3). However, the production and therefore the service values calculated in Table 3 were probably underestimated due to model

207

restrictions. The water purification model was primarily developed for watersheds dominated by saturation excess runoff hydrology, and may be less applicable to locations like the Llobregat basin where infiltration runoff occurs as a result of flash rainfall events. Despite these limitations, areas closer to the river network, especially those in the southern part of the basin, were the ones with the highest nutrient export values (Fig. 6c). Management efforts aimed at decreasing nutrient inputs should focus on these regions. In a disturbed river basin like the Llobregat, services are valuable in certain places, and identifying these key areas is essential for management interventions to achieve their maximum potential effect. The average value of erosion control in the basin was equivalent to 12 Mg ha−1 y−1 of sediment retention. Estimated average soil erosion was 22 Mg ha−1 y−1 , similar to values observed in various regional studies (Barrow, 1991; Gunatilake and Gopalakrishnan, 1999; Lee et al., 2011). Approximately 85% of the sediment produced in the basin was retained by vegetation. Sediment loads reaching La Baells reservoir were underestimated compared to predictions by Catari et al. (2009) (1; Table 2). This was likely due to the exclusion of linear erosion in the model definition, which is associated with relatively strong rainfall events characteristic of semi-arid systems. Conversely, the annual amount of sediment predicted to reach the outlet was overestimated compared to estimations obtained by Liquete et al. (2009) (2; Table 2). InVEST likely overestimates export because it does not currently account for the trapping of sediment in reservoirs. Despite performing a retention function, areas downstream of reservoirs provided no reservoir maintenance benefits by definition (Fig. 6d). However, these downstream areas do play an important role in maintaining other socioeconomic benefits such as water quality and soil fertility through erosion control. The analysis presented in this work relied on calculations of average precipitation, temperature and water yield in the basin. The hydrological services we studied are meaningful at the annual scale because water management is carried out at this scale. In the case of erosion control, the service is relevant at even longer time scales. Average conditions were considered good descriptors of a, multiannual distribution (calculating the sum of annual values, distributions become approximately normally distributed due to the central limit theorem) even in the case of a semi-arid climate where extremes affect the system’s behavior.

4.2. Climate extremes and hydrological ecosystem services Water scarcity is already an issue in the Mediterranean region and will probably become more acute with climate change and expanding human development (Sabater and Tockner, 2010). Ecosystem vulnerability in these areas will likely become more evident and could result in large reductions of ecosystem service supply (Schröter et al., 2005). Data from the past 50 years were used to assess the effects of annual climate extremes in this work. However, they may not reflect future extremes under climate change. Due to recent extremes in climate conditions, the delivery of hydrological ecosystem services in the Llobregat basin varied significantly through time. The water towers, providing water to their immediate surroundings as well as areas located downstream, were the most affected by annual rainfall reduction. Water provisioning for drinking water use was the most threatened service in dry years. Indeed, taking into account the current annual consumption rate of drinking water in the Llobregat basin, it would clearly be difficult to cope with a period of prolonged drought. Water provisioning for hydropower as well as energy production itself were altered in dry years (Fig. 7), reflecting the predicted reduction in hydropower potential in southern and south-eastern Europe under dry conditions (Lehner et al., 2005).

208

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209

Nutrient retention increased in wet years and was related to higher nutrient transport. Aside from diffuse nutrient pollution, the Llobregat River, like many other Mediterranean areas, is characterized by important point emissions from WWTPs (Marti et al., 2004). These emissions were also included in the model and considered constant under different climate conditions. Point emissions cannot be retained by the landscape and exert a direct impact on water quality. Therefore, the water purification service provided a greater net benefit in dry years. During these low flow years, water quality was more likely to be compromised (Fig. 7) due to a decreased dilution capacity. Furthermore, higher nutrient concentrations are associated with higher treatment costs during seasonal low flow events (Yongguan et al., 2001). Effects observed for dry years also suggest that water purification benefits could be important at the seasonal scale, specifically during the driest season of the year, although intra-annual variation in services was not explored in this study. The assessment of the water purification service in ecological terms (ecological WQS from the WFD) revealed some variation in the benefits obtained under different climate conditions. Benefits obtained in normal years were much higher using the ecological WQS (2.5 times the benefits for drinking WQS; Table 3), but they only increased 1.5% in dry years and decreased 5% in wet years. The benefits of erosion control were higher in wet years (Fig. 7), when a larger amount of sediment would reach reservoirs if the sediment retention function was not supplied by the landscape. Predicted increases in rainfall intensity (Martín-Vide, 2004) together with a decline in vegetation cover characteristic of the Mediterranean region could enhance soil erosion, making the erosion control service even more valuable under future climate change. 5. Conclusions The delivery of ecosystem services in the Llobregat River basin was strongly dependent on climate conditions. Results indicate that, to adapt to climate change, either proactive or reactive management should be envisaged in Mediterranean basins characterized by semi-arid conditions, chronic human impact, and intensive water use. While provisioning services were identified as more relevant in normal and wet years, the regulating service water purification provided higher benefits in dry years, when threats to water quality were increased because of a decreased dilution capacity. Protection of water towers in semi-arid regions and areas expected to experience dramatic changes is shown as essential to ensure water provisioning in dry years, when the service is more at risk. However, the protection of water resources is not sufficient if consumption rates continue or increase in the future. Interventions should be planned to enhance the provision of regulating services. In particular, these interventions should focus on areas neighboring the river network, where benefits per surface area are estimated to be the highest. Aside from the land phase, the aquatic phase also plays an important role in the provision of regulating services. This suggests the inclusion of in-stream processes in order to gain modeling accuracy. Overall, this study highlights that in semi-arid basins under chronic human impact, hydrological ecosystem services are very sensitive to climate extremes, and that service supply and demand areas are usually spatially and temporally decoupled. Both aspects are relevant and need to be considered in basin management in semi-arid regions. Acknowledgements This research was supported by the Spanish Ministry of Science and Innovation through the SCARCE Consolider-Ingenio 2010 CSD2009-00065 project, by a Marie Curie European Reintegration

Grant within the 7th European Community Framework Programme, and by the Gordon and Betty Moore Foundation. The authors would like to thank Rebekah Kipp for her help during the revision process.

References Barrow, C.J., 1991. Land Degradation: Development and Breakdown of Terrestrial Environments. Cambridge University Press, Cambridge. Booker, J.F., 1995. Hydrologic and economic impacts of drought under alternative policy responses. Water Resour. Bull. 31, 889–906. Branson, F.A., Gifford, G.F., Renard, K.G., Hadley, R.F., 1981. Rangeland Hydrology. Kendall/Hunt Publishing Co., Iowa. Brauman, K.A., Daily, G.C., Duarte, T.K., Mooney, H.A., 2007. The nature and value of ecosystem services: an overview highlighting hydrologic services. Annu. Rev. Environ. Resour. 32, 67–98. Catalan Government, 2006. Resolution MAH/2465/2006 through which it is approved the sectorial planning of environmental flows in the internal basins of Catalonia. Official Journal of the Catalan Government. Catalan Water Agency, 2002. Description and prediction study about water demand in the internal basins of Catalonia and the Ebro Catalan basins. Department of Environment, Catalan Government. Catalan Water Agency, 2007. Assessment of the implementation of the sectorial plan for environmental flows from the internal basins of Catalonia to the hydroelectric sector. Department of Environment and Housing, Catalan Government. Catari, G., Gallart, F., 2010. Rainfall erosivity in the upper Llobregat basin, SE Pyrenees. Pirineos 165, 55–67. Catari, G., Gallart, F., Josa, R., 2009. Quantifiying soil erosion with GIS-based RUSLE in La Baells reservoir (Llobregat River basin), SE Pyrenees, Spain. Geophys. Res. Abstr., 11. Chan, K.M.A., Shaw, M.R., Cameron, D.R., Underwood, E.C., Daily, G.C., 2006. Conservation planning for ecosystem services. PLoS Biol. 4, e379. Costanza, R., d’Arge, R., de Groot, R.S., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O’Neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., van den Belt, M., 1997. The value of the world’s ecosystem services and natural capital. Nature 387, 253–260. Daily, G.C., 1997. Nature’s Services: Societal Dependence on Natural Ecosystems. Island Press, Washington, DC. de Groot, R.S., Alkemade, R., Braat, L., Hein, L., Willemen, L., 2010. Challenges in integrating the concept of ecosystem services and values in landscape planning, management and decision making. Ecol. Complex. 7, 260–272. Delpla, I., Jung, A.-V., Baures, E., Clement, M., Thomas, O., 2009. Impacts of climate change on surface water quality in relation to drinking water production. Environ. Int. 35, 1225–1233. Eigenbrod, F., Armsworth, P.R., Anderson, B.J., Heinemeyer, A., Gillings, S., Roy, D.B., Thomas, C.D., Gaston, K.J., 2010. The impact of proxy-based methods on mapping the distribution of ecosystem services. J. Appl. Ecol. 47, 377–385. European Council, 1980. Council Directive 80/778/EEC of 15 July 1980 relating to the quality of water intended for human consumption as amended by Council Directives 81/858/EEC and 91/692/EEC. European Council, 2000. Water Framework Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for community action in the field of water policy. Official Journal of the European Communities. Grizzetti, B., Bouraoui, F., de Marsily, G., Bidoglio, G., 2005. A statistical method for source apportionment of riverine nitrogen loads. J. Hydrol. 304, 302–315. Gunatilake, H.M., Gopalakrishnan, C., 1999. The economics of reservoir sedimentation: a case study of Mahaweli reservoirs in Sri Lanka. Int. J. Water Resour. Dev. 15, 511–526. Guo, Z., Xiao, X., Li, D., 2000. An assessment of ecosystem services: water flow regulation and hydroelectric power production. Ecol. Appl. 10, 925–936. Haines-Young, R., Potschin, M., 2010. The links between biodiversity ecosystem services and human wellbeing. In: Raffaelli, D., Frid, C. (Eds.), Ecosystem Ecology: A New Synthesis. BES Ecological Review Series. CUP, Cambridge. Hisdal, H., Stahl, K., Tallaksen, L.M., Demuth, S., 2001. Have streamflow droughts in Europe become more severe or frequent? Int. J. Climatol. 21, 317–333. Horn, M.A., Moore, R.B., Hayes, L., Flanagan, S.M., 2008. Methods for and Estimates of 2003 and Projected Water Use in the Seacoast Region. US Geological Survey Scientific Investigations, Southeastern New Hampshire. Kareiva, P., Tallis, H., Ricketts, T.H., Daily, G.C., Polasky, S., 2011. Natural Capital: Theory and Practice of Mapping Ecosystem Services. Oxford University Press, New York. Lee, Y., Yoon, T., Shah, F.A., 2011. Economics of integrated watershed management in the presence of a dam. Water Resour. Res. 47, http://dx.doi.org/10.1029/2010WR009172, W10509. Lehner, B., Czisch, G., Vassolo, S., 2005. The impact of global change on the hydropower potential of Europe: a model-based analysis. Energy Policy 33, 839–855. Liquete, C., Canals, M., Ludwig, W., Arnau, P., 2009. Sediment discharge of the rivers of Catalonia, NE Spain, and the influence of human impacts. J. Hydrol. 366, 76–88. Llebot, J.E., 2010. Second Report on Climate Change in Catalonia. Institute for Catalan Studies and Catalan Government, Barcelona.

M. Terrado et al. / Ecological Indicators 37 (2014) 199–209 López-Doval, J.C., Grossschartner, M., Höss, S., Orendt, C., Traunspurger, W., Wolfram, ˜ I., 2010. Invertebrate communities in soft sediments along a pollution G., Munoz, gradient in a Mediterranean river (Llobregat, NE Spain). Limnetica 29, 311–322. Ludwig, W., Dumont, E., Meybeck, M., Heussner, S., 2009. River discharges of water and nutrients to the Mediterranean and Black Sea: major drivers for ecosystem changes during past and future decades? Prog. Oceanogr. 80, 199–217. Marti, E., Aumatell, J., Godé, L., Poch, M., Sabater, F., 2004. Nutrient retention efficiency in streams receiving inputs from wastewater treatment plants. J. Environ. Qual. 33, 285–293. Martín-Vide, J., 2004. Spatial distribution of a daily precipitation concentration index in peninsular Spain. Int. J. Climatol. 24, 959–971. Martín-Vide, J., Brunet, M., Prohom, M., Rius, A., 2010. The climate in Catalonia. Present and recent tendencies. In: Llebot, J.E. (Ed.), Second Report on Climate Change in Catalonia. Institute of Catalan Studies and Catalan Government, Barcelona. Metzger, M.J., Schröter, D., Leemans, R., Cramer, W., 2008. A spatially explicit and quantitative vulnerability assessment of ecosystem service change in Europe. Reg. Environ. Change 8, 91–107. Naidoo, R., Balmford, A., Costanza, R., Fisher, B., Green, R.E., Lehner, B., Malcolm, T.R., Ricketts, T.H., 2008. Global mapping of ecosystem services and conservation priorities. Proc. Natl. Acad. Sci. U.S.A. 105, 9495–9500. Nash, J.E., Sutcliffe, J.V., 1970. River flow forecasting through conceptual models. Part I. A discussion of principles. J. Hydrol. 10, 282–290. Nelson, E., Mendoza, G., Regetz, J., Polasky, S., Tallis, H., Cameron, D.R., Chan, K.M., Daily, G.C., Goldstein, J., Kareiva, P.M., Lonsdorf, E., Naidoo, R., Ricketts, T.H., Shaw, M.R., 2009. Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Front. Ecol. Environ. 7, 4–11. Ninyerola, M., Pons, X., Roure, J.M., 2000. A methodological approach of climatological modelling of air temperature and precipitation through GIS techniques. Int. J. Climatol. 20, 1823–1841. Ricart, M., Guasch, H., Barceló, D., Brix, R., Conceic¸ao, M.H., Geiszinger, A., de Alda, ˜ I., Postigo, C., Romaní, A.M., Villagrasa, M., M.J.L., López-Doval, J.C., Munoz, Sabater, S., 2010. Primary and complex stressors in polluted Mediterranean rivers: pesticide effects on biological communities. J. Hydrol. 383, 52–61.

209

Sabater, S., Tockner, K., 2010. Effects of hydrologic alterations on the ecological quality of river ecosystems. In: Sabater, S., Barceló, D. (Eds.), Water Scarcity in the Mediterranean: Perspectives Under Global Change. The Handbook of Environmental Chemistry. Springer Verlag, Berlin Heidelberg, pp. 15–39. Sabater, S., Sabater, F., Tomas, X., 1987. Water quality and diatom communities in two Catalan rivers (NE Spain). Water Res. 21, 901–911. Schröter, D., Cramer, W., Leemans, R., Prentice, I.C., Araújo, M.B., Arnell, N., Bondeau, W., Bugmann, A., Carter, H., Gracia, T.T., de la Vega-Leinert, C.A., Erhard, A.C., Ewert, M., Glendining, F., House, M., Kankaanpaa, J.I., Klein, S., Lavorel, R.J.T., Lindner, S., Metzger, M., Meyer, M.J., Mitchell, J., Reginster, T.D., Rounsevell, I., Sabaté, M., Sitch, S., Smith, S., Smith, B., Smith, Jo., Sykes, P., Thonicke, M.T., Thuiller, K., Tuck, W., Zaehle, G., Zierl, S.B., 2005. Ecosystem service supply and vulnerability to global change in Europe. Science 310, 1333–1337. Tallis, H., Ricketts, T., Guerry, A.D., Wood, S.A., Sharp, R., Nelson, E., Ennaanay, D., Wolny, S., Olwero, N., Vigerstol, K., Pennington, D., Mendoza, G., Aukema, J., Foster, J., Forrest, J., Cameron, D., Arkema, K., Lonsdorf, E., Kennedy, C., Verutes, G., Kim, C.K., Guannel, G., Papenfus, M., Toft, J., Marsik, M., Bernhardt, J., 2011. InVEST 2.2.2 User’s Guide. The Natural Capital Project, Stanford. Terrado, M., Lavigne, M.-P., Tremblay, S., Duchesne, S., Villeneuve, J.-P., Rousseau, A.N., Barceló, D., Tauler, R., 2009. Distribution and assessment of surface water contamination by application of chemometric and deterministic models. J. Hydrol. 369, 416–426. Vigerstol, K.L., Aukema, J.E., 2011. A comparison of tools for modeling freshwater ecosystem services. J. Environ. Manage. 92, 2403–2409. Viviroli, D., Dürr, H.H., Messerli, B., Meybeck, M., Weingartner, R., 2007. Mountains of the world, water towers for humanity: typology, mapping, and global significance. Water Resour. Res. 43, http://dx.doi.org/10.1029/2006WR005653, W07447. Wischmeier, W.H., Smith, D.D., 1978. Predicting Rainfall Erosion Losses – A Guide to Conservation Planning, vol. 537. U.S. Department of Agriculture, Washington D.C. Yongguan, C., Seip, H.M., Vennemo, H., 2001. The environmental cost of water pollution in Chongqing, China. Environ. Dev. Econ. 6, 313–333.