Variability in the water footprint of arable crop production across European regions

water Article Variability in the Water Footprint of Arable Crop Production across European Regions Anne Gobin 1,*, Kurt Christian Kersebaum 2, Josef Eitzinger 3, Miroslav Trnka 4,5, Petr Hlavinka 4,5, Jozef Takáˇc 6, Joop Kroes 7, Domenico Ventrella 8, Anna Dalla Marta 9, Johannes Deelstra 10, Branislava Lali´c 11, Pavol Nejedlik 12, Simone Orlandini 9, Pirjo Peltonen-Sainio 13, Ari Rajala 13, Triin Saue 14, Levent ¸Saylan 15, Ruzica Striˇcevic 16, Višnja Vuˇceti´c 17 and Christos Zoumides 18 1Flemish Institute for Technological Research (VITO), 2400 Mol, Belgium 2Leibniz Centre for Agricultural Landscape Research (ZALF), 15374 Müncheberg, Germany; [email protected] 3Institute of Meteorology, University of Natural Resources and Life Sciences, 1180 Vienna, Austria; [email protected] 4Global Change Research Institute, The Czech Academy of Sciences, 61300 Brno, Czech Republic; [email protected] (M.T.); [email protected] (P.H.) 5Department of Agrosystems and Bioclimatology, Mendel University in Brno, 61300 Brno, Czech Republic 6National Agricultural and Food Centre, Soil Science and Conservation Research Institute, Bratislava 82713, Slovakia; [email protected] 7 Wageningen Environmental Research (Alterra), 6700 AA Wageningen, The Netherlands; [email protected] 8Consiglio per la Ricerca in Agricoltura e L’analisi Dell’economia Agraria, Unità di Ricerca per i Sistemi Colturali degli Ambienti Caldo-Aridi, 70125 Bari, Italy; [email protected] 9Department of Agrifood Production and Environmental Sciences (DISPAA), University of Florence, 50155 Florence, Italy; [email protected] (A.D.M.); [email protected] (S.O.) 10 Norwegian Institute of Bioeconomy Research (NIBIO), 1431 Ås, Norway; [email protected] 11 Faculty of Agriculture, University of Novi Sad, Novi Sad 21000, Serbia; [email protected] 12 Earth Science Institute of Slovak Academy of Science, Bratislava 84005, Slovakia; [email protected] 13 Natural Resources Institute Finland (Luke), 31600 Jokioinen, Finland; [email protected] (P.P.-S.); [email protected] (A.R.) 14 Estonian Crop Research Institute, Tallinn Technical University, Jõgeva 48309, Estonia; [email protected] 15 Department of Meteorology, Faculty of Aeronautics and Astronautics, Istanbul Technical University, 34469 Istanbul, Turkey; [email protected] 16 Faculty of Agriculture, University of Belgrade, Zemun-Belgrade 11080, Serbia; sr[email protected] 17 Meteorological and Hydrological Service, Zagreb 10000, Croatia; [email protected] 18 Energy, Environment & Water Research Center, The Cyprus Institute, Nicosia 2121, Cyprus; [email protected] *Correspondence: [email protected]; Tel.: +32-14-336775 Academic Editor: Ashok K. Chapagain Received: 24 October 2016; Accepted: 31 January 2017; Published: 8 February 2017 Abstract: Crop growth and yield are affected by water use during the season: the green water footprint (WF) accounts for rain water, the blue WF for irrigation and the grey WF for diluting agri-chemicals. We calibrated crop yield for FAO’s water balance model “Aquacrop” at field level. We collected weather, soil and crop inputs for 45 locations for the period 1992–2012. Calibrated model runs were conducted for wheat, barley, grain maize, oilseed rape, potato and sugar beet. The WF of cereals could be up to 20 times larger than the WF of tuber and root crops; the largest share was attributed to the green WF. The green and blue WF compared favourably with global benchmark values (R 2 = 0.64–0.80; d = 0.91–0.95). The variability in the WF of arable crops across different regions in Europe is mainly due to variability in crop yield ( cv = 45%) and to a lesser extent to variability in crop water use ( cv = 21%). The WF variability between countries ( cv = 14%) is lower than the variability between seasons ( cv = 22%) and between crops ( cv = 46%). Though modelled yields Water 2017,9, 93; doi:10.3390/w9020093 www.mdpi.com/journal/water Water 2017,9, 93 2 of 22 increased up to 50% under sprinkler irrigation, the water footprint still increased between 1% and 25%. Confronted with drainage and runoff, the grey WF tended to overestimate the contribution of nitrogen to the surface and groundwater. The results showed that the water footprint provides a measurable indicator that may support European water governance. Keywords: water footprint; arable crops; cereals; Europe; crop water use; yield 1. Introduction The water footprint (WF) concept has created awareness of sustainable water use following a global assessment of national production, consumption and international trade [ 1 ]. Traditional water consumption statistics have been given for different sectors, such as domestic, agricultural and industrial water use, but these show little about how much water is actually used. The water footprint provides a way to compare water use of regions, sectors, commodities and nations. Leading work in understanding water availability and risk has come from the food industries through the analysis of water quantities that companies use throughout their supply chain. With water being inherently local, the water footprint calculations highlight the risks of local exploitations that could potentially disrupt both business operations and the surrounding community. Water is a precious commodity, certainly in drought-prone regions and at times of drought in any part of the world. The economic cost of drought has been enormous. In 2003, combined drought and heat waves led to 30% reduction in primary productivity [ 2 ], and an estimated 13 billion € loss in European agricultural production [ 3 ]. With water shortages already threatening growth, the future of Europe’s agriculture will be tied closely to water availability. In addition climate models project that southern Europe will face increased drought and central Europe prolonged dry spells [ 4 , 5 ] frequently combined with heat waves [ 6 ]. The rising population, coupled with increasing demands by the agriculture and energy industries presents an interdependent relationship often referred to as the water–food–energy nexus; the demand for water will likely outweigh supply by 2050 unless changes in food and energy preferences are implemented [ 7 ]. While access to water has been recognized as a basic human right, the increasingly high demand for water resources should be valued according to its supply. The WF is closely linked to the concept of virtual water, which is the volume needed to produce a commodity or service. Importing virtual water can be perceived as a partial solution to problems of water scarcity, particularly in dry regions [ 8 ]. National, regional and global water and food security can be improved when water-intensive commodities are traded from places where they are economically viable to places they are not. Food import offers an alternative to reduce pressure on domestic water resources and enables more productive water use as expressed by the WF of food [ 9 ]. Other research has taken a life cycle assessment (LCA) approach to evaluate the water footprint of products, processes and organisations as initiated by [ 10 ]. Subsequently, an ISO 14046 standard was set to specify the principles, requirements and guidelines [ 11 ]. The ISO standard may introduce complexity by creating water footprints for each environmental impact, e.g., for water availability, scarcity, eutrophication and eco-toxicity, across the life cycle of a product which is beyond the crop water footprint that this research focuses on. The WF of crops forms the basis for WF estimations of crop products and derived commodities [ 12 ]. In terms of water volumes used, the crop WF estimations consider three major sources of water, i.e., water from rain (green WF), irrigation (blue WF) and water for diluting chemicals (grey WF) [ 13 ]. In a comparison of different irrigation and water conservation methods for four locations [14], it was concluded that a combination of drip irrigation and synthetic mulching allowed for the largest reduction in the WF of maize, potato and tomato. The inter-annual variability of the crop WF highlighted inter alia the importance of increased yields for 22 crops for the period 1978–2008 in Water 2017,9, 93 3 of 22 China [ 15 ]. Understanding the variability is a prerequisite to making projections of good water governance under different scenarios of global change. Our study contributes to understanding the variability of the WF across regions, soils and annual weather conditions in Europe. We hypothesize that the variability in the water footprint of arable crops across different regions in Europe is mainly due to variability in crop yield and to a lesser extent to variability in crop water use. Therefore, the objectives of this study were to quantify the variability in water used to grow arable crops across different regions in Europe; to estimate their yield variability; to establish the variability in the WF of these different crops; and, to compare the results with benchmark values from global model estimates as in [ 16 ]. Understanding the sources of variability in the WF is important to elucidate water consumption patterns in relation to crop production, which in turn enables more efficient water management and agricultural water governance within the framework of a water–food–energy nexus. 2. Materials and Methods 2.1. Data We collected temperature, rainfall, wind speed, solar radiation and relative humidity data from 41 meteorological stations across different regions in Europe for the period 1992–2012 (Table 1; Figure 1). Reference evapotranspiration was calculated using the modified Penman-Monteith approach [ 17 ]. The climatological diagrams of temperature, precipitation and evapotranspiration for these locations demonstrate a wide variation in weather conditions (Figure 2) and soils (Appendix A). The dominant soil type(s) for 45 locations were described in terms of texture; chemical composition; volumetric water content at saturation, field capacity and wilting point of different soil horizons up to 1.5 m or to an impervious layer. With the exception of polder regions, groundwater was absent and water leaching from the root zone was discharged as drainage. In each location major arable crops were selected for calculating the water footprint (Table 1). Table 1. Meteorological stations and crops per region (Location see Figure 1). Country Region Meteo Stations 1Major Crops 2 AT Marchfeld Gross Enzersdorf,Fuchsenbigl WHB, BAR, MAZ, SBT BE Flanders Koksijde,Gent, Ukkel, Peer WHB, BAR, MAZ, SBT, POT CY Country Nicossia, Pafos, Larnaca WHD, POT, BAR, MAZ CZ Eastern Czech Domaninek,Lednice,Verovany WHB, BAR, MAZ, RAP DE-1 Märk. Oderland Muncheberg, Manschnow WHB, BAR, SBT, RAP, POT, MAZ DE-2 North-East Lower Saxony Braunschweig WHB, BAR, SBT EE Country Kuusiku, Tartu, Tallinn, Võru, Pärnu, Väike-Maarja, Kuressaare WHB, BAR, POT, RAP FI-1 Häme Jokioinen BAR, WHB, BAR, POT, RAP FI-2 South Finland Mikkeli,Ylistaro,Laukaa,Piikio BAR, WHB, BAR, POT, RAP HR Koprivnica-Križevci Križevci MAZ IT-1 Foggia Foggia WHD, SBT IT-2 Val d’Orcia Radicofani WHB, WHD, BAR NO South Eastern Norway Søråsjordet BAR NL Flevoland Lelystad WHB, POT, SBT, MAZ PL Mazovia D ˛abrowice WHB, BAR, POT, SBT, RAP SK Danube Lowland Bratislava-letisko, Hurbanovo, Nitra, Jaslovske Bohunice WHB, BAR, MAZ SR Vojvodina Rimski Sancevi WHD, MAZ, SBT, POT TR Thrace Edirne, Kırklareli, Tekirda˘g WHD, WHB, BAR, MAZ Notes: 1 In bold are meteorological stations located in the vicinity of experimental fields; 2 BAR is barley (Hordeum vulgare L.); MAZ is maize (Zea mays L.); POT is potato (Solanum tuberosum L.); SBT is sugar beet (Beta vulgaris L.); RAP is oilseed rape (Brassica napus L.); WHB is common wheat (Triticum aestivum L.); and, WHD is durum wheat (Triticum turgidum L.). Water 2017,9, 93 4 of 22 Water2017,9,934of22   Figure1.LocationofdifferentmeteorologicalstationsacrossEurope. 2.2.CropWaterUse FAO’s“Aquacrop”modelversion5.0[18]wasusedtocalculatethecropwaterfootprint.The growthmoduleisevapotranspirationdriven,wherecroptranspiration(T)isconvertedtobiomass throughawaterproductivityparameter[19,20].Theevaporativepoweroftheatmosphere(ET0)is convertedtoactualevapotranspiration(ET)andseparatedintonon‐productivewaterfluxes,i.e.,soil evaporation(E),andproductivewaterfluxes,i.e.,croptranspiration(T).Soilmoistureconditions determineEfromthesoilsurfacenotcoveredbycanopy[19,20].Cropcanopyexpandsfrom seedlingtomaturityasdeterminedbyaccumulatedgrowingdegreedays. Cropcalendarandgrowthcharacteristicswerecollectedforthemajorarablecropsineach location(Table1).Thecropgrowthparametersweresetusingexperimentalfielddatacollectedfor eachregion(AppendixA,[21]).Forregionswithoutexperimentalfielddataavailable,cropgrowth parameterswerederivedfromfarmers’fields’data. Allweather,soilandcropinputdata(Figure2;AppendixA)wereinsertedintothemodel.The model’sphenologicalmodulewasruningrowingdegreedaystocapturecropgrowthdynamics duringthegrowingseason.Rainfedmodelrunsforthedifferentlocationswerefollowedby sprinklerirrigationruns,at80%fieldcapacity,andaccordingtolocalfarmpractices.Therefore, regionswherenoirrigationwasreportedwereexcludedfromtheirrigationmodelruns.  Figure 1. Location of different meteorological stations across Europe. 2.2. Crop Water Use FAO’s “Aquacrop” model version 5.0 [ 18 ] was used to calculate the crop water footprint. The growth module is evapotranspiration driven, where crop transpiration (T) is converted to biomass through a water productivity parameter [ 19 , 20 ]. The evaporative power of the atmosphere (ET0) is converted to actual evapotranspiration (ET) and separated into non-productive water fluxes, i.e., soil evaporation (E), and productive water fluxes, i.e., crop transpiration (T). Soil moisture conditions determine Efrom the soil surface not covered by canopy [ 19 , 20 ]. Crop canopy expands from seedling to maturity as determined by accumulated growing degree days. Crop calendar and growth characteristics were collected for the major arable crops in each location (Table 1). The crop growth parameters were set using experimental field data collected for each region (Appendix A, [ 21 ]). For regions without experimental field data available, crop growth parameters were derived from farmers’ fields’ data. All weather, soil and crop input data (Figure 2; Appendix A) were inserted into the model. The model’s phenological module was run in growing degree days to capture crop growth dynamics during the growing season. Rainfed model runs for the different locations were followed by sprinkler irrigation runs, at 80% field capacity, and according to local farm practices. Therefore, regions where no irrigation was reported were excluded from the irrigation model runs. Water 2017,9, 93 5 of 22 Water2017,9,935of22  NorthernEurope(EE,FI,NO)  Søråsjordet,Norway Jokioinen,Finland WesternEurope(BE,DE,NL)  Lelystad,TheNetherlands(NL) Gent,Belgium(BE) Figure 2. Cont. Water 2017,9, 93 6 of 22 Water2017,9,936of22  CentralEurope(AT,CZ,DE,SK)  GrossEnzersdorf,Austria(AT) Hubanovo,Slovakia(SK) SouthEurope(CY,HR,IT,SR,TR)  Foggia,Italy(IT) Tekirdağ,Turkey(TR) Figure2.ClimatologicaldiagramsfordifferentmeteorologicalstationsalongabroadtransectinEuropefortheperiod1992–2012.Pisprecipitation(mm);ET0isreference evapotranspiration(mm);Tmeanisaveragetemperature(°C).Atwolettercodereferstothecountries.  Figure 2. Climatological diagrams for different meteorological stations along a broad transect in Europe for the period 1992–2012. Pis precipitation (mm); ET0 is reference evapotranspiration (mm); Tmean is average temperature (◦C). A two letter code refers to the countries. Water 2017,9, 93 7 of 22 2.3. Water Footprint Calculations Irrigated agriculture receives water from irrigation (blue water) and from precipitation (green water), while rainfed agriculture only receives green water. Green water is originated by precipitation and is the soil water held in the unsaturated zone available to plants, while blue water refers to the manageable water in rivers, lakes, wetlands and aquifers [ 22 ]. The green WF and blue WF reflect the rainfed and irrigated crop water use per harvested crop with calculation methods established by [ 13 ]. The grey WF accounts for water used to dilute nutrient pollution to meet ambient water quality standards; for reasons of comparison we focused on nitrogen pollution [16]. WFgreen =10·∑lgp d=1ETd,green Y(1) WFblue =10·∑lgp d=1ETd,blue Y(2) WFgrey =[[∝·AR]/[cmax −cnat]] Y(3) where ETd is the daily evapotranspiration in mm · day −1 , accumulated over the length of the growing period (lgp, in days), under rainfed (green) and irrigated (blue) conditions. The factor 10 converts water depths from millimetres into water volumes per land surface (m 3· ha −1 ). The nominator reflects crop water use in m3·ha−1, whereas the denominator (Y) is crop yield in Mg·ha−1. The green water evapotranspiration under irrigated conditions was estimated as the total evapotranspiration simulated in a scenario without irrigation. The blue water evapotranspiration equalled the total evapotranspiration simulated in the scenario with irrigation minus the simulated green water evapotranspiration. For the grey WF, we assumed that the nitrogen fraction ( α ) that reached free flowing water bodies through leaching or runoff equalled 10% of the application rate (AR in kg · ha −1· year −1 ). Fertilizer application rates were reduced significantly in the European Member States following the introduction of the Nitrates Directive in 1991 and the Water Framework Directive in 2000. Reporting mechanisms are in place so that nitrogen application rates and derived gross nitrogen balances are available from Eurostat for the period 1992–2012 [ 23 ]. Fertilizer consumption rates are available per hectare of arable land in the World Bank database [ 24 ]. We assumed drinking water standards for water quality with a difference between maximum acceptable and natural background concentration (cmax −cnat) of 10 mg·L−1[16]. 2.4. Yield Statistics Yield is an important component of the WF. Yields, area and production of wheat, barley, grain maize, potato, sugar beet and oilseed rape differed distinctly across the different regions in Europe, as shown for 2012 regional statistical yields (Figure 3). The harvested production of cereals in 2012–2015 in the EU-28 was estimated at one ninth of global cereals production; wheat (44%–47%), maize (21%–22%) and barley (19%–20%) account for a high share [ 25 ]. Despite a European-wide system of production quota, sugar beet remains the most important root crop for north-western Europe. Potato production is more widely spread across the different European Member States, as reflected by the presence of yield data in different regions (Figure 3). Oilseed rape, the main oilseed crop across Europe, showed an upward trend in production during the last decade due to its use for bioenergy purposes [ 25 ]. Regional statistical yields were compared with modelled yields assuming a humidity of 14% for cereals, 80% for root crops and 9% for oilseed rape [25]. Water 2017,9, 93 8 of 22 Water2017,9,938of22    Figure3.Yields(Mg∙ha −1 )formajorarablecropsacrosstheEuropeanregionsfortheyear2012basedonregionalstatistics.Atwolettercodereferstothecountrythatthe regionbelongsto.  Figure 3. Yields (Mg · ha −1 ) for major arable crops across the European regions for the year 2012 based on regional statistics. A two letter code refers to the country that the region belongs to. Water 2017,9, 93 9 of 22 2.5. Statistical Analysis The statistical analysis was done in R using the core functionalities [ 26 ] and the hydroGOF package [ 27 ]. Common statistical measures were used to describe the datasets. The coefficient of variation (cv), i.e., the ratio of the standard deviation to the mean expressed in %, was used to compare the spread of variables. The Pearson correlation coefficient (r 2 ) was used as a measure of strength of an association between two variables. Statistical metrics to describe the agreement between modelled and statistical yields and between our and benchmark WFs were the mean average error (MAE), the root mean square error (RMSE) and the index of agreement (d) [ 27 ]. The regression lines on the graphs and the associated coefficient of determination (R 2 ) were provided as a measure of how well the statistical yields or the benchmark WFs were approximated by our modelled results. 3. Results The water footprint (WF) of arable crops across different regions in Europe showed a large variability. We presented this large variability in relation to the different components that comprised the water footprint: evaporation and transpiration; biomass and yield; and, the green, blue and grey WF. Since these components were intrinsically linked to the water balance, a general comparison was made of the major water balance input and output. 3.1. Water Balance The water balance was driven by reference evapotranspiration, calculated from solar radiation, wind speed, temperature and relative humidity using the modified Penman–Monteith equation [ 17 ]. In all studied regions (Table 1), the reference evapotranspiration was higher than the precipitation accumulated over the growing season of spring sown crops (Figure 4). For autumn sown crops this difference was less pronounced. In northern and western European regions cumulative precipitation was higher than cumulative evapotranspiration during the growing season for the period 1992–2012. Simulated sub-surface drainage was in all cases higher than simulated surface runoff, but this difference was not always significant (Figure 5). A surplus on the water balance led to higher runoff and drainage during the growing season, and vice versa for a deficit. Due to higher precipitation during winter a surplus occurred during the growing season of autumn sown crops (Figure 5). Water2017,9,939of22  2.5.StatisticalAnalysis ThestatisticalanalysiswasdoneinRusingthecorefunctionalities[26]andthehydroGOF package[27].Commonstatisticalmeasureswereusedtodescribethedatasets.Thecoefficientof variation(cv),i.e.,theratioofthestandarddeviationtothemeanexpressedin%,wasusedtocompare thespreadofvariables.ThePearsoncorrelationcoefficient(r2)wasusedasameasureofstrengthofan associationbetweentwovariables.Statisticalmetricstodescribetheagreementbetweenmodelledand statisticalyieldsandbetweenourandbenchmarkWFswerethemeanaverageerror(MAE),theroot meansquareerror(RMSE)andtheindexofagreement(d)[27].Theregressionlinesonthegraphs andtheassociatedcoefficientofdetermination(R2)wereprovidedasameasureofhowwellthe statisticalyieldsorthebenchmarkWFswereapproximatedbyourmodelledresults. 3.Results Thewaterfootprint(WF)ofarablecropsacrossdifferentregionsinEuropeshowedalarge variability.Wepresentedthislargevariabilityinrelationtothedifferentcomponentsthatcomprised thewaterfootprint:evaporationandtranspiration;biomassandyield;and,thegreen,blueandgrey WF.Sincethesecomponentswereintrinsicallylinkedtothewaterbalance,ageneralcomparison wasmadeofthemajorwaterbalanceinputandoutput. 3.1.WaterBalance Thewaterbalancewasdrivenbyreferenceevapotranspiration,calculatedfromsolarradiation, windspeed,temperatureandrelativehumidityusingthemodifiedPenman–Monteithequation[17]. Inallstudiedregions(Table1),thereferenceevapotranspirationwashigherthantheprecipitation accumulatedoverthegrowingseasonofspringsowncrops(Figure4).Forautumnsowncropsthis differencewaslesspronounced.InnorthernandwesternEuropeanregionscumulativeprecipitation washigherthancumulativeevapotranspirationduringthegrowingseasonfortheperiod1992–2012. Simulatedsub‐surfacedrainagewasinallcaseshigherthansimulatedsurfacerunoff,butthis differencewasnotalwayssignificant(Figure5).Asurplusonthewaterbalanceledtohigherrunoff anddrainageduringthegrowingseason,andviceversaforadeficit.Duetohigherprecipitation duringwinterasurplusoccurredduringthegrowingseasonofautumnsowncrops(Figure5).  Figure4.Precipitation(Pinmm)andreferenceevapotranspiration(ET0inmm)duringthegrowing seasonofautumnandspringsowncropsacrosstheEuropeanregionsfortheperiod1992–2012.A twolettercodereferstothecountrythattheregionbelongsto. Figure 4. Precipitation (Pin mm) and reference evapotranspiration (ET0 in mm) during the growing season of autumn and spring sown crops across the European regions for the period 1992–2012. A two letter code refers to the country that the region belongs to. Water 2017,9, 93 16 of 22 Water2017,9,9316of22   Figure13.Greywaterfootprint(inm3∙Mg−1)formodelledandstatisticalarableyieldsfortheperiod 1992–2012andforfourdifferentnitrogenapplicationrates.Notethedifferencesinscalebetween crops.Atwolettercodereferstothecountrythattheregionbelongsto. 4.Discussion Weusedthe“Aquacrop”modeltoestimatecropgrowthandevapotranspirationunderboth rainfedandirrigatedconditions.Thismodelhasbeendevelopedtosimulateyieldresponsetowater underwater‐limitedconditions[18–20].Regions,cropsandsoilsthataresensitivetodryspellsand droughtprovideforawater‐limitedenvironment.Reviewsofmodelbehaviourshowmixedresults withrespecttowateruseefficienciesandyields[28].Anintercomparisonofeightmodelsshowed between13%and19%uncertaintyintheestimationofevapotranspiration(“Aquacrop”:cv=15%), andbetween13%and34%fortranspiration(“Aquacrop”:cv=24%)forwheat[21]. ThegreenWFshowedthelargestvariabilityforcereals(cv=51%–52%),closelyfollowedby potato(cv=48%);thelowestvariabilitywasforoilseedrape(cv=36%)andsugarbeet(cv=28%) (Table3).Forallarablecrops,theyieldismorevariable(   =45%;cvinTable2)thanthecrop evapotranspiration(   =21%;cvinTable3).Thisclearlydemonstratestheimportanceofyieldsand theirvariabilityforthewaterfootprint.Similartothefindingsof[29],rootandtubercropshave muchlowerWFsascomparedtocerealsandoilseedcrops(Table3),owingtoacombinedeffectof higheryieldsandhighermoisturecontentsatharvest.Cerealsandoilseedcropshaveamuch smallerharvestablefractionofthetotalbiomassproducedpersurfacearea,andthereforehave largerwaterfootprints. BetweenthedifferentregionsinEurope,highyieldingwesternEuropeanregionshaveWFs thatcanbeuptosixtimeslowerthantheWFsofregionsinnorthernorsouthernEurope (Figures9,11and13).Athreefoldincreasecanoccurbetweenseasons,certainlyinregionswith variableyields.Ananalysisofvariance(ANOVA)demonstratedcleareffectsofcrops,countries, seasonsandtheirfactorialinteractionsonthegreenwaterfootprint(p<0.001).Thelargestvariability betweenseasonsisforthesouthern(CY,TR)andnortherncountries(EE,FI,NO)andforWHD, BARandRAP.Thevariabilitybetweencountries(   =14%;range:7%–26%)islowerthanthe variabilitybetweenseasons(   =22%;range:10%–50%)andthevariabilitybetweencrops (   =46%;range:29%–52%). Table3.Greenwaterfootprint(WF)(WFginm3∙Mg−1)andevapotranspiration(ETinmm)forthe majorarablecropsinEuropefortheperiod1992–2012.ForcropabbreviationsseeFigure3. CropWFg∙mWFg∙sWFg∙cv ET∙mET∙sET∙cv RAP18576613640510325 WHB1108580524598719 WHD1414720513756217 BAR901458513378224 MAZ590304523737320 POT15775483326419 SBT6719293518625 Notes:Wheremdenotesmean,sstandarddeviationandcvcoefficientofvariation(%). Figure 13. Grey water footprint (in m 3· Mg −1 ) for modelled and statistical arable yields for the period 1992–2012 and for four different nitrogen application rates. Note the differences in scale between crops. A two letter code refers to the country that the region belongs to. 4. Discussion We used the “Aquacrop” model to estimate crop growth and evapotranspiration under both rainfed and irrigated conditions. This model has been developed to simulate yield response to water under water-limited conditions [ 18 – 20 ]. Regions, crops and soils that are sensitive to dry spells and drought provide for a water-limited environment. Reviews of model behaviour show mixed results with respect to water use efficiencies and yields [ 28 ]. An intercomparison of eight models showed between 13% and 19% uncertainty in the estimation of evapotranspiration (“Aquacrop”: cv = 15%), and between 13% and 34% for transpiration (“Aquacrop”: cv = 24%) for wheat [21]. The green WF showed the largest variability for cereals (cv = 51%–52%), closely followed by potato (cv = 48%); the lowest variability was for oilseed rape (cv = 36%) and sugar beet (cv = 28%) (Table 3). For all arable crops, the yield is more variable ( cv = 45%; cv in Table 2) than the crop evapotranspiration ( cv = 21%; cv in Table 3). This clearly demonstrates the importance of yields and their variability for the water footprint. Similar to the findings of [ 29 ], root and tuber crops have much lower WFs as compared to cereals and oilseed crops (Table 3), owing to a combined effect of higher yields and higher moisture contents at harvest. Cereals and oilseed crops have a much smaller harvestable fraction of the total biomass produced per surface area, and therefore have larger water footprints. Table 3. Green water footprint (WF) (WFg in m 3· Mg −1 ) and evapotranspiration (ET in mm) for the major arable crops in Europe for the period 1992–2012. For crop abbreviations see Figure 3. Crop WFg·m WFg·s WFg·cv ET·m ET·s ET·cv RAP 1857 661 36 405 103 25 WHB 1108 580 52 459 87 19 WHD 1414 720 51 375 62 17 BAR 901 458 51 337 82 24 MAZ 590 304 52 373 73 20 POT 157 75 48 332 64 19 SBT 67 19 29 351 86 25 Notes: Where m denotes mean, s standard deviation and cv coefficient of variation (%). Between the different regions in Europe, high yielding western European regions have WFs that can be up to six times lower than the WFs of regions in northern or southern Europe (Figures 9,11 and 13) . A threefold increase can occur between seasons, certainly in regions with variable yields. An analysis of variance (ANOVA) demonstrated clear effects of crops, countries, seasons and their factorial interactions on the green water footprint (p< 0.001). The largest variability between seasons is for the southern (CY, TR) and northern countries (EE, FI, NO) and for WHD, BAR and RAP. The variability between countries ( cv = 14%; range: 7%–26%) is lower than the variability between seasons ( cv = 22%; range: 10%–50%) and the variability between crops ( cv = 46%; range: 29%–52%). Water 2017,9, 93 17 of 22 The breakdown of the crop water footprints in different components enabled a better understanding of the different contributing factors involved. A comparison between our calculations for the European regions and water footprint benchmarks for crop production provided by [ 16 ], revealed a good agreement for the green WF (R 2 = 0.80; d = 0.95), a reasonably good agreement for the blue WF (R 2 = 0.64; d = 0.91) and a lower agreement for the grey WF (R 2 = 0.25; d = 0.73), where R 2 is the coefficient of determination and d is the index of agreement [ 27 ]. Overall the best fit was obtained for the green WF (Figure 14). All WF were highly influenced by yield so that only well calibrated models able to model yield can be successfully deployed to estimate the WF. Yield variability determined the WF variability. Water2017,9,9317of22  Thebreakdownofthecropwaterfootprintsindifferentcomponentsenabledabetter understandingofthedifferentcontributingfactorsinvolved.Acomparisonbetweenour calculationsfortheEuropeanregionsandwaterfootprintbenchmarksforcropproductionprovided by[16],revealedagoodagreementforthegreenWF(R2=0.80;d=0.95),areasonablygood agreementfortheblueWF(R2=0.64;d=0.91)andaloweragreementforthegrey WF(R2=0.25;d=0.73),whereR2isthecoefficientofdeterminationanddistheindexofagreement [27].OverallthebestfitwasobtainedforthegreenWF(Figure14).AllWFwerehighlyinfluenced byyieldsothatonlywellcalibratedmodelsabletomodelyieldcanbesuccessfullydeployedto estimatetheWF.YieldvariabilitydeterminedtheWFvariability.  Figure14.Comparisonofgreen,blueandgreywaterfootprints(WF)ofthisstudywithbenchmark WF[16]inm3∙Mg−1forarablecrops.Thebluelinerepresentstheidentityline. Cropgrowthandproductionaremostlyaffectedbythedistributionofgreenwaterduringthe growingseason.Bluewaterdirectlyinfluencestheyieldprovidedwaterisavailableforirrigation: wemodelledyieldincreasesofupto50%whichhighlightthebenefitsofirrigation.Waterscarcity, exacerbatedfurtherbyclimatechange,isanissueofmajorconcerninaridandsemi‐aridregions withseriousimpactsonfoodsecurity,sustainabilityandeconomy.Thefactthatthemajorityof availablewaterisconsumedbyagriculturalactivities,particularlyinaridandsemi‐aridcountries, underpinstheneedformonitoringandreducingwaterconsumptionpatternsinagriculturalareas. OurmodelledyieldincreasesdidnotresultinlowercropWFsundersprinklerirrigation;drip irrigationmayresultinlowerWFsascalculatedby[14].TheWFofagriculturalcropsallowsfor decisionmakingandbettermanagementofthewaterpotential. AppliedtoagriculturalproductionthegreyWFistheamountoffreshwaterrequiredforthe assimilationofanypollutant,incasunitrogenrunoffduetoagriculturalcropproduction.Nitrogen applicationratesdifferconsiderablybetweenregions,andregulationsareinplacetolimittheinput forexampleinnitrogenvulnerablezonesornatureconservationareas[30]. Otherimportantsourcesofvariationarecroptype,farmingsystem,soiltype,andthe rainfall‐runoffregime.Whenrunoffanddrainageweretakenintoaccount(Figure5),thegreyWF showedalotmorevariabilitybetweentheyearsandcouldbeaslowaszeroduringsomeyearsfor summercrops.Inaddition,thepollutionandhencegreywaterisattributedtoasinglecropthereby neglectingtheroleofacropintherotation.ForexampleoilseedrapehadahighgreyWF,despite thecrop’scapacitytodepletenitrogenfromthepreviouscropbeforewinterandthereforereduceN leaching.Acontraryexampleisthehighleachingriskofbaresoilduringwinterpriortosugarbeet. Theseeffectswerenotincorporatedintheapplicationsofthegreywaterfootprintofcrops[13,16]. Recentapplicationsconcentratedonanitrogenbalancetobudgetuptakeandlosses,andarrivedat higherestimatesofnitrogen‐relatedwaterpollutioninriverbasinsowingtodifferencesin computationalmethods[31].ThereforethegreyWFshouldbecomparedwithcautionbetween studiesandagriculturalsystems. Figure 14. Comparison of green, blue and grey water footprints (WF) of this study with benchmark WF [16] in m3·Mg−1for arable crops. The blue line represents the identity line. Crop growth and production are mostly affected by the distribution of green water during the growing season. Blue water directly influences the yield provided water is available for irrigation: we modelled yield increases of up to 50% which highlight the benefits of irrigation. Water scarcity, exacerbated further by climate change, is an issue of major concern in arid and semi-arid regions with serious impacts on food security, sustainability and economy. The fact that the majority of available water is consumed by agricultural activities, particularly in arid and semi-arid countries, underpins the need for monitoring and reducing water consumption patterns in agricultural areas. Our modelled yield increases did not result in lower crop WFs under sprinkler irrigation; drip irrigation may result in lower WFs as calculated by [ 14 ]. The WF of agricultural crops allows for decision making and better management of the water potential. Applied to agricultural production the grey WF is the amount of freshwater required for the assimilation of any pollutant, in casu nitrogen runoff due to agricultural crop production. Nitrogen application rates differ considerably between regions, and regulations are in place to limit the input for example in nitrogen vulnerable zones or nature conservation areas [30]. Other important sources of variation are crop type, farming system, soil type, and the rainfall-runoff regime. When runoff and drainage were taken into account (Figure 5), the grey WF showed a lot more variability between the years and could be as low as zero during some years for summer crops. In addition, the pollution and hence grey water is attributed to a single crop thereby neglecting the role of a crop in the rotation. For example oilseed rape had a high grey WF, despite the crop’s capacity to deplete nitrogen from the previous crop before winter and therefore reduce N leaching. A contrary example is the high leaching risk of bare soil during winter prior to sugar beet. These effects were not incorporated in the applications of the grey water footprint of crops [ 13 , 16 ]. Recent applications concentrated on a nitrogen balance to budget uptake and losses, and arrived at higher estimates of nitrogen-related water pollution in river basins owing to differences Water 2017,9, 93 18 of 22 in computational methods [ 31 ]. Therefore the grey WF should be compared with caution between studies and agricultural systems. Practices to reduce the WF of crop production start with awareness of the WF of different crop management systems. A transition to less water-demanding crops with higher water productivities or higher water use efficiencies offers opportunities to optimize plant water use. Soil and water conservation techniques and water saving irrigation methods, e.g., drip irrigation and deficit irrigation, could further reduce water demands [ 14 ]. Advanced techniques lower the crop water demand, but cannot markedly decrease the WF; achieving more stable and higher yields, however, can. The high dependency on yield warrants strategies to increase agricultural productivity which is accomplished through breeding programs and/or through optimizing resources use during the crop growth season. The grey WF is partly regularized through the Water Framework and Nitrates Directives with designated nitrate vulnerable zones and limitations on nitrogen and phosphorus applications [ 30 ]. Ensuring that water quality is minimally affected offers good perspectives for nutrient smart precision farming. Overall the water footprint and its assessment process helps establish a greater awareness of water consumption patterns among different stakeholders involved. 5. Conclusions We calculated the green and blue water footprint with FAO’s “Aquacrop” model, and the grey water footprint on the basis of nitrogen application rates for six major arable crops in 45 locations across Europe for the period 1992–2012. The WF of cereals is larger than the WF of tuber and root crops owing mainly to the difference in yield and moisture content at harvest between these crop types. Since yield has a larger variability than crop water use, yield estimates are of paramount importance to the crop WF. The WF for wheat, for example, can be up to five or six times larger in northern and southern Europe as compared to high yielding western European regions. The WF variability between crops was larger than the variability between seasons and in turn larger than the variability between countries. Yield increases under sprinkler irrigation were not high enough to reduce the water footprint. Water saving irrigation and soil conservation techniques, however, may result in WF reductions. The green and blue WF, but not the grey WF, compared favourably with internationally available benchmark values. Confronted with drainage and runoff, the grey WF tended to overestimate the contribution of nitrogen to the surface and groundwater. Other agro-hydrological methods to calculate the grey WF resulted in even larger values which points to caution when comparing different studies. The large variability between crops, regions and seasons; and between yields and water use as major components of the WF highlights the importance of crop yield variability. The water footprint is a measurable indicator that may support European water governance. Acknowledgments: The authors acknowledge funding from COST ES1106, Belgian Science Policy contract number SD/RI/03A, JPI FACCE MACSUR (D.M. 24064/7303/15), 2812ERA 147 (from BLE Germany), the Czech project LD13030, the Ministry of Education, Youth and Sports of the Czech Republic within the National Sustainability Program I (NPU I), grant number LO1415, and contract nr III43007 of the Ministry of Education and Science of the Republic of Serbia. The authors thank their national meteorological services and institutes for providing weather data. Support for data collection is acknowledged from Sabina Thaler for Austrian data, Eva Pohanková for Czech data and Toprak Aslan for Turkish data. Author Contributions: Anne Gobin, Kurt Christian Kersebaum, Josef Eitzinger and Mirek Trnka conceived and designed the experiments of calculating the water footprint. “Aquacrop” model runs were performed by Anne Gobin. All authors contributed to data collection and interpretation of the results. Conflicts of Interest: The authors declare no conflict of interest. Appendix A The dominant soil type(s) and meteorological stations of each region are provided in Table A1, together with references to relevant datasets. Crop characteristics used for calibration are provided in Table A2. Water 2017,9, 93 19 of 22 Table A1. Soil hydrological properties of the topsoil for different locations across Europe. CTRY Location Soil Type Texture 1FC (%) WP (%) Pore Space References AT Gross Enzersdorf Chernozem Silt Loam 35 21 43 [32] AT Gross Enzersdorf Parachernozem Sandy Loam 28 8 39 AT Gross Enzersdorf Fluvisol Clay Loam 35 22 42 AT Fuchsenbigl Calcaric Chernozem Silt Loam 38 23 53 BE Koksijde Calcaric & Gleyic Fluvisol Marine Clay 39 23 50 [33–35] BE Gent Albeluvisol Sandy Loam 22 10 47 BE Peer Podzol Loamy Sand 16 8 46 BE Ukkel Luvisol Silty Loam 34 12 49 CY Larnaca Chromic Vertisol Clay 42 28 50 [36,37] CY Nicosia Vertic-Chromic Luvisol Clay Loam 38 24 46 CY Pafos Eutric Fluvisol Loam 32 18 46 CZ Domaninek Dystric Cambisol Loam 30 15 47 [38] CZ Lednice Chernozem Silt Loam 35 16 49 CZ Verovany Chernozem Silt Loam 33 14 47 DE Manschnow Fluvic Gleysol Clay loam 39 15 46 [39] DE Manschnow Cambisol Sandy Loam 31 9 40 [39] DE Manschnow Podzol Sandy Loam 14 5 42 [39] DE Müncheberg Eutric Cambisol Loamy sand 26 11 36 [40,41] DE Braunschweig Luvisol Sandy Loam 24 6 46 [41] EE Kuusiku Calcic Luvisol Silt Loam 28 7 40 [42] EE Väike-Maarja Calcaric Cambisol Sandy Loam 28 8 45 EE Tartu Mollic Cambisol Loam 30 9 48 EE Võru Stagnic Luvisol Loamy Sand 20 6 42 EE Tallinn Haplic Albeluvisol Sand 16 3 44 EE Kuressaare Gleysol Clay 35 22 50 EE Pärnu Gleysol Clay loam 32 20 48 FI Jokioinen Haplic Umbrisol Silt loam 35 21 45 [43] FI Mikkeli Mollic Cambisol Sandy Loam 28 7 42 FI Ylistaro Verti-Gleyic Cambisol Silt Loam 35 15 48 FI Laukaa Eutric Regosol Silty Clay 46 25 55 FI Piikiö Vertic Cambisol Clay Loam 36 22 48 HR Križevci Gleyic Luvisol Silt loam 36 12 41 [44] IT Foggia Alluvial vertisol Clay Loam 42 24 55 [45] IT Radicofani Vertic Cambisol Silty Clay 42 27 51 [46,47] NL Lelystad Gleyic Fluvisol Marine Clay 36 16 45 [48] NO Søråsjordet Gleyic Podzoluvisol Silt Loam 37 20 50 [49] PL D ˛abrowice Podzol Loamy Sand 23 17 40 SK Jasl.Bohunice Chernozem Silty Loam 34 14 44 [50] SK Nitra Luvisol Clay Loam 36 17 44 SK Bratislava Fluvisol Sandy Loam 32 12 44 SK Hurbanovo Phaeozem Clay Loam 35 18 44 SR Rimski Sancevi Chernozem Loam 34 17 51 [51] TR Kirklareli Cambisol Sandy Clay Loam 35 17 42 [52] TR Tekirda˘g Fluvic Cambisol Sandy Clay Loam 39 28 46 [53] TR Edirne Cambisol Clay Loam 37 23 41 [53] Notes: 1Soil texture is classified according to the USDA nomenclature. Table A2. Planting (P) and harvesting (H) dates of arable crops in the different regions. Region * WHB/WHD BAR RAP MAZ POT SBT P; H P; H P; H P; H P; H P; H AT 12/10; 30/7 25/3; 30/6 7/5; 26/9 16/4; 5/9 12/4; 18/8 BE 15/10; !/8 15/10; 15/7 15/9; 15/7 1/5; 30/9 10/4; 30/9 10/4; 15/10 CY 15/11; 30/5 15/11; 4/5 15/1; 24/5 CZ 3/10; 30/7 30/3; 25/7 28/8; 20/7 30/4; 15/9 DE-1 2/10; 30/7 20/9; 15/7 28/8; 24/7 16/5; 15/9 DE-2 25/10; 30/7 25/9; 25/6 15/4; 30/9 EE 30/8; 10/8 25/4; 3/8 1/5; 15/9 5/5; 10/9 FI-1 30/8; 20/8 15/5; 20/8 15/5; 10/9 15/5; 10/9 FI-2 10/9; 15/8 10/5; 15/8 10/5; 1/9 15/5; 5/9 HR 29/4; 3/10 IT-1 15/11; 20/6 22/3; 18/8 IT-2 15/11; 15/7 15/11; 10/7 Water 2017,9, 93 20 of 22 Table A2. Cont. Region * WHB/WHD BAR RAP MAZ POT SBT P; H P; H P; H P; H P; H P; H NL 20/10; 30/7 5/9; 17/7 30/4; 15/10 25/4; 20/9 10/4; 15/10 NO 25/4; 15/8 PL 20/9; 14/7 24/4; 16/7 28/8; 17/8 15/4; 30/9 15/4; 20/9 SK 7/10; 20/7 24/3; 10/7 20/4; 3/9 15/4; 15/9 SR 15/11; 10/7 20/4; 30/9 30/3; 10/7 30/3; 15/10 TR 15/11; 30/6 30/9; 15/6 9/4; 20/8 15/3; 20/8 Notes: * For the name of the region see Table A1. References 1. Hoekstra, A.Y.; Mekonnen, M.M. The water footprint of humanity. Proc. Natl. Acad. Sci. USA 2012 ,109, 3232–3237. [CrossRef] [PubMed] 2. Ciais, P.; Reichstein, M.; Viovy, N.; Granier, A.; Ogée, J.; Allard, V.; Aubinet, M.; Buchmann, N.; Bernhofer, C.; Carrara, A.; et al. Europe-wide reduction in primary productivity caused by the heat and drought in 2003. Nature 2005,437, 529–533. [CrossRef] [PubMed] 3. COPA-COGECA. COPA-COGECA. Assessment of the impact of the heat wave and drought of the summer of 2003 on agriculture and forestry. In Fact Sheets of the Committee of Agricultural Organisations in the European Union and the General Committee for Agricultural Cooperation in the European Union; COPA-COGECA: Brussels, Belgium, 2003. 4. Dai, A. Increasing drought under global warming in observations and models. Nat. Clim. Chang. 2013 ,3, 52–58. [CrossRef] 5. Trnka, M.; Olesen, J.E.; Kersebaum, K.C.; Skjelvag, A.O.; Eitzinger, J.; Seguin, B.; Peltonen Sainio, P.; Rotter, R.; Iglesias, A.; Orlandini, S.; et al. Agroclimatic conditions in Europe under climate change. Glob. Chang. Biol. 2011,17, 2298–2318. [CrossRef] 6. Trnka, M.; Hlavinka, P.; Semenov, M.A. Adaptation options for wheat in Europe will be limited by increased adverse weather events under climate change. J. R. Soc. Interface 2015,12, 20150721. [CrossRef] [PubMed] 7. Damerau, K.; Patt, A.G.; van Vliet, O.P. Water saving potentials and possible trade-offs for future food and energy supply. Glob. Environ. Chang. 2016,39, 15–25. [CrossRef] 8. Allan, J.A. Virtual water: A strategic resource, global solutions to regional deficits. Ground Water 1998 ,36, 545–546. [CrossRef] 9. Zoumides, C.; Bruggeman, A.; Hadjikakou, M.; Zachariadis, T. Policy-relevant indicators for semi-arid nations: The water footprint of crop production and supply utilization of Cyprus. Ecol. Indic. 2014 ,43, 205–214. [CrossRef] 10. Ridoutt, B.G.; Pfister, S. A revised approach to water footprinting to make transparent the impacts of consumption and production on global freshwater scarcity. Glob. Environ. Chang. 2010 ,20, 113–120. [CrossRef] 11. International Organization for Standardization. ISO 14046: 2014: Environmental Management: Water Footprint—Principles, Requirements and Guidelines; International Organization for Standardization: Geneva, Switzerland, 2014. 12. Chenoweth, J.; Hadjikakou, M.; Zoumides, C. Quantifying the human impact on water resources: A critical review of the water footprint concept. Hydrol. Earth Syst. Sci. 2014,18, 2325–2342. [CrossRef] 13. Mekonnen, M.M.; Hoekstra, A.Y. The green, blue and grey water footprint of crops and derived crop products. Hydrol. Earth Syst. Sci. 2011,15, 1577–1600. [CrossRef] 14. Chukalla, A.D.; Krol, M.S.; Hoekstra, A.Y. Green and blue water footprint reduction in irrigated agriculture: Effect of irrigation techniques, irrigation strategies and mulching. Hydrol. Earth Syst. Sci. 2015 ,19, 4877–4891. [CrossRef] 15. Zhuo, L.; Mekonnen, M.M.; Hoekstra, A.Y. The effect of inter-annual variability of consumption, production, trade and climate on crop-related green and blue water footprints and inter-regional virtual water trade: A study for China (1978–2008). Water Res. 2016,94, 73–85. [CrossRef] [PubMed] Water 2017,9, 93 21 of 22 16. Mekonnen, M.M.; Hoekstra, A.Y. Water footprint benchmarks for crop production: A first global assessment. Ecol. Indic. 2014,46, 214–223. [CrossRef] 17. Allen, R.G.; Pereira, L.S.; Raes, D.; Smith, M. Crop Evapotranspiration—Guidelines for Computing Crop Water Requirements; FAO Irrigation and Drainage Paper; Food and Agriculture Organization: Rome, Italy, 1998; p. D05109. 18. FAO Crop-Model to Simulate Yield Response to Water. Available online: http://www.fao.org/nr/water/ aquacrop.html (accessed on 23 August 2016). 19. Steduto, P.; Hsiao, T.C.; Raes, D.; Fereres, E. AquaCrop—The FAO crop model to simulate yield response to water: I. Concepts and underlying principles. Agron. J. 2009,101, 426–437. [CrossRef] 20. Raes, D.; Steduto, P.; Hsiao, T.C.; Fereres, E. Aquacrop the FAO crop model to simulate yield response to water: II. Main algorithms and software description. Agron. J. 2009,101, 438–447. [CrossRef] 21. Kersebaum, K.C.; Kroes, J.; Gobin, A.; Takác, J.; Hlavinka, P.; Trnka, M.; Ventrella, D.; Giglio, L.; Ferrise, R.; Moriondo, M.; et al. Assessing the uncertainty of model based water footprint estimation using an ensemble of crop growth models on winter wheat. Water 2016,8, 571. [CrossRef] 22. Rockström, J.; Falkenmark, M.; Karlberg, L.; Hoff, H.; Rost, S.; Gerten, D. Future water availability for global food production: The potential of green water for increasing resilience to global change. Water Res. 2009 ,45, W00A12. [CrossRef] 23. Eurostat. Gross Nutrient Balance. Available online: http://ec.europa.eu/eurostat/cache/metadata/EN/ aei_pr_gnb_esms.htm (accessed on 23 August 2016). 24. World Bank, Fertilizer Consumption Rates per Hectare of Arable Land. Available online: http:// databank.worldbank.org/data/reports.aspx?source=2&series=AG.CON.FERT.ZS&country (accessed on 23 August 2016). 25. Eurostat. Handbook for Annual Crop Statistics. European Commission-Eurostat, Directorate E Sectoral and Regional Statistics, Unit E1 Agriculture and Fisheries. Available online: http://ec.europa.eu/eurostat/ cache/metadata/Annexes/apro_acs_esms_an1.pdf (accessed on 23 August 2016). 26. R Core Group. The R Project for Statistical Computing. Available online: https://www.r-project.org/ (accessed on 20 November 2016). 27. Zambrano-Bigiarini, M. hydroGOF: Goodness-of-Fit Functions for Comparison of Simulated and Observed Hydrological Time Series. 2011. R Package Version 0.3-2. Available online: http://CRAN.R-project.org/ package=hydroGOF (accessed on 23 August 2016). 28. Evett, S.R.; Tolk, J.A. Introduction: Can Water Use Efficiency Be Modeled Well Enough to Impact Crop Management? Agron. J. 2009,101, 423–425. [CrossRef] 29. Mekonnen, M.M.; Hoekstra, A.Y. The Green, Blue and Grey Water Footprint of Crops and Derived Crop Products. Value of Water Research Report Series No. 47; UNESCO-IHE: Delft, The Netherlands, 2010. Available online: http://www.waterfootprint.org/Reports/Report47-WaterFootprintCrops-Vol1.pdf (accessed on 23 August 2016). 30. Velthof, G.L.; Lesschen, J.P.; Webb, J.; Pietrzak, S.; Miatkowski, Z.; Pinto, M.; Kros, J.; Oenema, O. The impact of the Nitrates Directive on nitrogen emissions from agriculture in the EU-27 during 2000–2008. Sci. Total Environ. 2014,468, 1225–1233. [CrossRef] [PubMed] 31. Mekonnen, M.M.; Hoekstra, A.Y. Global gray water footprint and water pollution levels related to anthropogenic nitrogen loads to fresh water. Environ. Sci. Technol. 2015 ,49, 12860–12868. [CrossRef] [PubMed] 32. Eitzinger, J.; Trnka, M.; Semerádová, D.; Thaler, S.; Svobodová, E.; Hlavinka, P.; Siska, B.; Takáˇc, J.; Malatinská, L.; Nováková, M.; et al. Regional climate change impacts on agricultural crop production in Central and Eastern Europe–hotspots, regional differences and common trends. J. Agric. Sci. 2013 ,151, 787–812. [CrossRef] 33. Gobin, A. Impact of heat and drought stress on arable crop production in Belgium. Natl. Hazards Earth Syst. Sci. 2012,12, 1911–1922. [CrossRef] 34. Gobin, A. Modelling climate impacts on arable yields in Belgium. Clim. Res. 2010,44, 55–68. [CrossRef] 35. Gobin, A. The water footprint of Belgian arable crops. Ital. J. Agric. Meteorol. 2015,4, 91–97. 36. Zoumides, C.; Bruggeman, A.; Zachariadis, T.; Pashiardis, S. Quantifying the poorly known role of groundwater in agriculture: The case of Cyprus. Water Resour. Manag. 2013,27, 2501–2514. [CrossRef] Water 2017,9, 93 22 of 22 37. Camera, C.; Zomeni, Z.; Noller, J.; Zissimos, A.; Christoforou, I.; Bruggeman, A. A high resolution map of soil types and physical properties for Cyprus: A digital soil mapping optimization. Geoderma 2017 ,285, 35–49. [CrossRef] 38. Hlavinka, P.; Trnka, M.; Kersebaum, K.; Cermak, P.; Pohankova, E.; Ors_ag, M.; Pokorný, E.; Fischer, M.; Brtnický, M.; Zalud, Z. Modelling of yields and soil nitrogen dynamics for crop rotations by HERMES under different climate and soil conditions in the Czech Republic. J. Agric. Sci. 2014,152, 188–204. [CrossRef] 39. Kersebaum, K.C.; Nendel, C. Site-specific impacts of climate change on wheat production across regions of Germany using different CO2response functions. Eur. J. Agron. 2014,52, 22–32. [CrossRef] 40. Mirschel, W.; Barkusky, D.; Hufnagel, J.; Kersebaum, K.C.; Nendel, C.; Laacke, L.; Luzi, K.; Rosner, G. Coherent multi-variable field data set of an intensive cropping system for agro-ecosystem modelling from Müncheberg, Germany. Open Data J. Agric. Res. 2016,2, 1–10. [CrossRef] 41. Kollas, C.; Kersebaum, K.C.; Nendel, C.; Manevski, K.; Müller, C.; Palosuo, T.; Armas-Herrera, C.M.; Beaudoin, N.; Bindi, M.; Charfeddine, M.; et al. Crop rotation modelling–a European model intercomparison. Eur. J. Agron. 2015,70, 98–111. [CrossRef] 42. Kadaja, J.; Saue, T. Effects of different irrigation and drainage regimes on yield and water productivity of two potato varieties under Estonian temperate climate. Agric. Water Manag. 2016,165, 61–71. [CrossRef] 43. Peltonen-Sainio, P.; Jauhiainen, L.; Hakala, K. Crop responses to temperature and precipitation according to long-term multi-location trials at high-latitude conditions. J. Agric. Sci. 2011,149, 49–62. [CrossRef] 44. Vuˇceti´c, V. Modelling of maize production in Croatia: Present and future climate. J. Agric. Sci. 2011 ,149, 145–157. [CrossRef] [PubMed] 45. Ventrella, D.; Stellacci, A.M.; Castrignanò, A.; Charfeddine, M.; Castellini, M. Effects of crop residue management on winter durum wheat productivity in a long term experiment in Southern Italy. Eur. J. Agron. 2016,77, 188–198. [CrossRef] 46. Dalla Marta, A.; Orlando, F.; Mancinii, M.; Guasconi, F.; Motha, R.; Qu, J.; Orlandini, S. A simplified index for an early estimation of durum wheat yield in Tuscany (Central Italy). Field Crops Res. 2015 ,170, 1–6. [CrossRef] 47. Guasconi, F.; Dalla Marta, A.; Grifono, D.; Mancini, M.; Orlando, F.; Orlandini, S. Influence of climate on durum wheat production and use of remote sensing and weather data to predict quality and quantity of harvests. Ital. J. Agrometeorol. 2011,3, 21–28. 48. Van Bakel, P.J.T.; Massop, H.T.L.; Kroes, J.G.; Hoogewoud, J.; Pastoors, R.; Kroon, T. Actualisatie Hydrologie voor STONE 2.3. In Aanpassing Randvoorwaarden en Parameters, Koppeling Tussen NAGROM en SWAP, en Plausibiliteitstoets; WOt-Rapport 57; Wettelijke Onderzoekstaken Natuur & Milieu (MNP), Alterra: Wageningen, The Netherlands, 2008. 49. Deelstra, J.; Kværnø, S.H.; Skjevdal, R.; Vandsemb, S.; Eggestad, H.O.; Ludvigsen, G.H. A General Description of Skuterud Catchment; Jordforsk (Now NIBIO) Report No. 61/05; Bioforsk: Ås, Norway, 2005. 50. Takáˇc, J.; Skalský, R.; Morávek, A.; Klikušovská, Z.; Bezák, P.; Bárdyová, M. Spatial Patterns of Agricultural Drought Events in Danube Lowland in the 1961–2013 Period. In Proceedings of the International Scientific Conference towards Climatic Services, Nitra, Slovakia, 15–18 September 2015. 51. Striˇcevi´c, R.; Cosi´c, M.; Djurovi´c, N.; Peji´c, B.; Maksimovi´c, L. Assessment of the FAO “Aquacrop” model in the simulation of rainfed and supplementally-irrigated maize, sugar beet and sunflower. Agric. Water Manag. 2011,98, 1615–1621. 52. ¸Saylan, L.; Çalda˘g, B.; Bakano˘gulları, F. Investigation of Potential Effects of Climate Change on Crop Growth by Crop Growth Simulation Models; TUBITAK Project No.: 1080567; TUBITAK: Istanbul, Turkey, 2012; p. 258. (In Turkish) 53. Çakir, R. Istranca (Yıldız) Da˘gı Güneyinde Yer Alan Vertisol Ordosu Topraklarının Toprak Taksonomisine Göre Belirlenmesi, Toprak ve Su Mühendisli˘gi Yönünden ˙ Irdelenmesi. Ph.D. Thesis, Trakya Üniversitesi, Edirne, Turkey, 1997. © 2017 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).