water

Article Spatio-Temporal Patterns of the 2010–2015 Extreme Hydrological Drought across the Central , Argentina

Juan Antonio Rivera 1,2,* ID , Olga C. Penalba 3,4, Ricardo Villalba 1 and Diego C. Araneo 1

1 Instituto Argentino de Nivología, Glaciología y Ciencias Ambientales (IANIGLA), CCT-Mendoza/Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Av. Ruiz Leal s/n, 5500 Mendoza, Argentina; [email protected] (R.V.); [email protected] (D.C.A.) 2 Facultad de Ciencias Veterinarias y Ambientales, Universidad Juan Agustín Maza, Av. Acceso Este, Lateral Sur 2245, 5521 Mendoza, Argentina 3 Departamento de Ciencias de la Atmósfera y los Océanos, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Intendente Güiraldes 2160, C1428EGA Buenos Aires, Argentina; [email protected] 4 Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Godoy Cruz 2290, C1425FQB Buenos Aires, Argentina * Correspondence: [email protected]; Tel.: +54-261-428-6010

Academic Editors: Paulo Barbosa and Jürgen Vogt Received: 1 June 2017; Accepted: 23 August 2017; Published: 30 August 2017

Abstract: During the period 2010–2015, the semi-arid Central Andes in Argentina (CAA) experienced one of the most severe and long-lasting hydrological droughts on record. Since the snowmelt is the most important source of water, the reduced snowfall over the mountains propagated the drought signal through the streamflows in the adjacent foothills east of the Andes ranges. Motivated by the widespread impacts on the socio-economic activities in the , this study aims to characterize the recent hydrological drought in terms of streamflow deficits. Based on streamflow data from 20 basins, we used the standardized streamflow index (SSI) to characterize hydrological droughts during the period 1971–2016. We found that the regional extent of the 2010–2015 hydrological drought was limited to the basins located north of 38◦ S, with mean duration of 67 months and maximum drought severity exhibiting a heterogeneous pattern in terms of spatial distribution and time of occurrence. The drought event reached extreme conditions in 14 of the 15 basins in the CAA, being record-breaking drought in six of the basins. This condition was likely driven by a cooling in the tropical Pacific Ocean resembling La Niña conditions, which generated a decrease in snowfall over the Andes due to suppressed frontal activity.

Keywords: hydrological drought; semi-arid region; streamflow; Central Andes; drought; hydroclimatic variability; water resources; standardized streamflow index; Argentina; snowmelt

1. Introduction Given the recent changes in the frequency, duration and intensity of droughts, a comprehensive understanding of water scarcity is needed at different temporal and spatial scales. This requirement is pressing given the marked increase in demand of water for agriculture, energy production, industry and human consumption. Droughts occur in virtually all climates, but have larger impacts in arid and semi-arid of the world where droughts are characterized by long duration (up to several years) and high intensity [1]. The scarcity of precipitation over an extended period define the meteorological

Water 2017, 9, 652; doi:10.3390/w9090652 www.mdpi.com/journal/water Water 2017, 9, 652 2 of 18 drought, whereas its consequences on the hydrological cycle, such as abnormally low streamflows, low levels in lakes and reservoirs, and deeper groundwaters, is known as hydrological drought [2]. The documented precipitation decrease in the subtropics during the last 3–4 decades has favored the occurrence of persistent large-scale droughts [3]. Frequent drought events over the subtropics have been mainly attributed to the recent poleward expansion and intensification of the descending branch of the Hadley Circulation [4]. Indeed, widespread droughts have affected recently several regions in the Southern Hemisphere. In , an uninterrupted rainfall decline was recorded in Central from 2010 to date [5,6], while the semi-arid northeastern Brazil has experienced since 2010 the longest and most intense drought in decades [7], with more than 10 million people affected and large losses on rainfed agriculture [8]. A period of unprecedented rainfall shortage, called the “Millennium Drought”, has also been registered in [9], at the time that 2015 was the year with the lowest national annual rainfalls since records started in 1904 in South Africa [10]. The arid Central Andes of Argentina (CAA, 31◦ S–37◦ S; 67◦ W–71◦ W) has also been affected by persistent severe droughts. The Water Administration Department from the Mendoza province (Departamento General de Irrigación) established the hydrological emergency in 2010. This hydrological status for the region continues today (May 2017). The Hydraulic Department from the San Juan province (Departamento de Hidráulica) reported the onset of the hydrological drought in 2009, when streamflow reduction amounted to 40%. Dams and reservoirs faced a marked reduction in water storages with severe consequences for intensive agriculture, the dominant regional activity only possible through irrigation [11]. Being the major wine producer region in Argentina, the agro-industrial activities in CAA depend largely on grape production [12]. Impacts of the 2010–2015 drought also affected the hydropower generation [13] and international tourism [14]. The amount of penalties in relation to water misuse increased in quantity and cost [15,16], in an attempt to optimize the limited water resources by the water administration agencies. To face the increasing severity of drought across the CAA, large-scale grape producers gradually set the new grape production areas at upper elevations by the Andes foothills, looking for lower temperatures, proximity to water sources, better water quality, and less environmental pollution [11]. These geographical changes in areas of grape production have intensified the upstream-downstream water conflicts between producers within basins. At regional scale, conflicts between provincial administrations have also been exacerbated due to the recent dominant drought conditions prevailing in most Central Andes basins during the last decade [17,18]. Climate change will exacerbate the spatial and temporal patterns of meteorological and hydrological droughts, including changes in seasonal distribution of water stress, thereby producing more frequent and more intense droughts with longer duration [1]. In this sense, the knowledge of processes causing hydrological drought and its spatial variability is essential for a sustainable management of water resources [19]. Basins in the CAA have a marked snowmelt-driven hydrological regime, with a runoff peak in the warm season and low flows during winter months [20]. Above normal streamflows occur during El Niño years; however, during La Niña events, the occurrence of low snowfall—i.e., potential hydrological drought conditions—is not straightforward [21]. The hydrological cycle in the CAA is likely to be modified under future climate scenarios with a long-term decrease in streamflow linked to reduced snow accumulation and an early peak in streamflows due to anticipated warmer springs [22,23]. Under future projections, it is crucial to deepen our knowledge on hydrological drought across the region and particularly the associated large impacts. The objective of this study is to characterize the recent extreme hydrological drought (2010–2015) along the CAA in terms of its onset, spatial extension, mean duration and maximum severity. The assessment of hydrological droughts will be performed through the analysis of streamflow data, a key variable to identify drought events with reference to some specific threshold levels [24]. A comparison of this long-lasting event with previous droughts recorded during the last 46 years will be performed in terms of drought characteristics, its links with snow accumulation and the related atmospheric and oceanic drivers. To further identify the spatial extension of this hydrological drought, we also included in the assessment some of the major basins from North (NP, 37◦ S–41◦ S; Water 2017, 9, 652 3 of 18

68◦ W–72◦ W), totalizing 20 rivers located between 31◦ S and 41◦ S. Recommendations based on both the need of an adequate streamflow monitoring system and the role of thresholds for drought declaration purposes are also discussed.

2. Materials and Methods

2.1. Data Monthly streamflow records from 20 stations located between 31◦ S and 41◦ S (Figure1) were retrieved from the hydrological database belonging to the Water Resources Agency of Argentina (http://bdhi.hidricosargentina.gov.ar/). Our records included the major rivers from the CAA and NP regions. These stations were selected based on the quality of the data, the spatial representativeness of the stations over the study area and the length of the runoff series. Raw streamflow data were previously subjected to quality control and gap filling routines, as described in [25,26]. Data infilling was minimal since the selected stations have less than 5% missing data. Linear regressions between the selected series and data from neighboring gauge stations were used to complete the data gaps. Following these procedures, the 1971–2016 common period between records was considered to characterize the hydrological droughts, although some records have data until the year 2015. Given that streamflow variability over the CAA is modulated by snowmelt [21,27], snow records, expressed as snow water equivalent between 1989 and 2015, from six stations distributed between 29◦ S and 37◦ S were also included in our analysis (Figure1). Selected stations are summarized in Table1. It is worth mentioning that, between the headwaters and the location of the stream gauges, there are no major reservoir or dam constructions that can alter the natural flow and, therefore, affect the estimation of streamflow drought occurrences.

Table 1. Geographical information on selected stations. IDs with an asterisk (*) correspond to snow course.

Altitude Mean Streamflow ID River Station Name Lat (◦ S) Lon (◦ W) (m a.s.l.) (m3/s) 1205 de los Patos Alvarez Condarco 31.92 69.70 1923 20.6 1211 San Juan KM 101 31.25 69.18 1310 53.4 1403 Atuel La Angostura 35.10 68.87 1302 37.5 1407 Cuevas Punta de Vacas 32.87 69.77 2406 7.2 1413 Mendoza Guido 32.92 69.24 1408 49.6 1415 Salado Cañada Ancha 35.20 69.78 1680 11.1 1419 Tunuyán Valle de Uco 33.78 69.27 1199 28.9 1420 Punta de Vacas 32.88 69.76 2450 24.9 1421 Vacas Punta de Vacas 32.85 69.76 2400 4.7 1423 Diamante La Jaula 34.66 69.31 1500 33.8 1425 Poti Malal Gendarmería 35.87 69.95 1485 7.6 1426 Pincheira Pincheira 35.51 69.80 1750 5.4 1427 Grande La Gotera 35.87 69.89 1400 108.2 2001 Barrancas Barrancas 36.80 69.89 950 38.3 2002 Colorado Buta Ranquil 37.08 69.75 850 151.3 2004 Neuquen Paso de Indios 38.53 69.41 498 295.5 2005 Chimehuin Naciente 39.79 71.21 875 62.6 2010 Agrio Bajada del Agrio 38.37 70.03 660 79.1 2021 Cuyín Manzano Cuyín Manzano 40.77 71.18 675 9.9 2040 Quilquihue Junín de los Andes 40.05 71.10 750 31.2 9250 * Jachal Cerro Negro 29.89 69.56 4172 - 9145 * Mendoza Toscas 33.16 69.89 3000 - 9131 * Diamante Laguna Diamante 34.20 69.70 3301 - 9104 * Atuel Laguna Atuel 34.51 70.05 3423 - 9121 * Grande Valle Hermoso 35.14 70.20 2253 - 9120 * Grande Paso Pehuenches 35.98 70.39 2555 - Water 2017, 9, 652 4 of 18 Water 2017, 9, 652 4 of 18

Figure 1. Spatial distribution of the hydrometeorological stations used in this study. The main rivers Figure 1. Spatial distribution of the hydrometeorological stations used in this study. The main rivers and the topography of the Central Andes in Argentina are also indicated. and the topography of the Central Andes in Argentina are also indicated. 2.2. Hydroclimatic Features 2.2. Hydroclimatic Features The Andes strongly affect the regional precipitation patterns through interactions with the Thecontinental Andes stronglyatmospheric affect circulation the regional and precipitation the incursion patterns of moist through air masses interactions from the with Pacific the continentalOcean atmospheric[25]. With circulation a mean elevation and the of incursion 3500 m north of moist of 35° air S, masses the Andes from act the as Pacific a permanent Ocean [barrier25]. With to the a mean elevationhumid of 3500air masses m north from of the 35◦ mid-latiS, the Andestude South act as Pacific a permanent Ocean. barrierAt high to elevations the humid in airthe massesAndes, fromthe the climate shows a Mediterranean regime with a marked precipitation peak during the cold months mid-latitude South . At high elevations in the Andes, the climate shows a Mediterranean (April to October) and little precipitation during the warm summer season (November to March regime with a marked precipitation peak during the cold months (April to October) and little precipitation [20]). The precipitation pattern drives the CAA hydrological cycle, exemplified in Figure 2 for the duringMendoza the warm River. summer The annual season streamflow (November cycle to was March defined [20]). as The the precipitationperiod between pattern July of drives year t theandCAA hydrologicalJune of year cycle, t + exemplified1. A pronounced in Figure streamflow2 for the peak Mendoza during the River. spring-summer The annual season streamflow is associated cycle was definedwith as snowmelt the period and between glacier July melting of year duet andto warmer June of temperatures. year t + 1. A pronouncedAt the Mendoza streamflow River, the peak mean during the spring-summermonthly streamflow season for is associatedJanuary exceeds with snowmelt100 m3/s. andIn contrast, glacier meltingthe low dueflow to season warmer is observed temperatures. At theduring Mendoza the cold River, winter the meanmonths, monthly with mo streamflownthly streamflow for January values exceeds close to 10020 m m3/s3/s. (Figure In contrast, 2). Due theto low flow seasonthe strong is observed rain shadow during effect, the cold climate winter east months, of the with Andes monthly is arid streamflow to semi-arid, values with close annual to 20 m 3/s (Figureprecipitation2). Due to thetotals strong ranging rain shadowbetween effect,100 and climate 400 eastmm. ofThese the Andes totals isare arid mostly to semi-arid, associated with with annual summer rainfalls favored by moist air masses from the Amazon and Atlantic basins [28]. precipitation totals ranging between 100 and 400 mm. These totals are mostly associated with summer South of 35° S, the mean elevation of the Andes decreases to about 1500 m. Along these ranges rainfallsand favoredadjacent by foothills, moist air rainfall masses is fromabundant the Amazon(reduced) and during Atlantic winter basins (summer) [28]. in response to the ◦ Southnorthward of 35 (southward)S, the mean shift elevation of the ofSouth the AndesEastern decreases Pacific High to about off the 1500 Chilean m. Along cost [26]. these Annual ranges and adjacentprecipitations foothills, rainfallrange from is abundant over 1200 (reduced)mm at the duringhigh elevations winter (summer)of the continental in response divide to to the less northward than (southward)300 mm east shift of of the the Andes. South The Eastern annual Pacificcycle of Highthe rivers off south the Chilean of 35° S costshows [26 two]. Annualannual maxima, precipitations as rangeexemplified from over for 1200 the mm Agrio at the River high in elevationsFigure 2. One of the streamflow continental peak divide is associated to less thanwith 300the mmwinter east of the Andes.rainfalls The (June–July, annual cycle >100 ofm the3/s) riversand the south second of 35with◦ S showsthe snowmelt two annual at higher maxima, mountains as exemplified during for the Agriospring–early River insummer Figure (October–November,2. One streamflow peak~120 m is3/s). associated The annual with streamflow the winter cycle rainfalls for the (June–July, rivers >100south m3/s) of and38° S the goes second from April with of the year snowmelt t to March atof higheryear t + 1. mountains during spring–early summer

(October–November, ~120 m3/s). The annual streamflow cycle for the rivers south of 38◦ S goes from April of year t to March of year t + 1. Water 2017, 9, 652 5 of 18 Water 2017, 9, 652 5 of 18

Figure 2. Hydrological cycles from two regional representative basins: the Mendoza River for the Figure 2. Hydrological cycles from two regional representative basins: the Mendoza River for the Central Andes of Argentina (CAA) and the Agrio River for North Patagonia (NP). Central Andes of Argentina (CAA) and the Agrio River for North Patagonia (NP). 2.3. Methods 2.3. Methods Hydrological drought was assessed in terms of streamflow variations, using the standardized Hydrologicalstreamflow index drought (SSI, [29]), was a assessed widely used in terms index ofconceived streamflow as an variations, extension from using the the standardized standardized streamflowprecipitation index index (SSI, [29(SPI,]), a[30]) widely to depict used indexhydrologic conceived aspects as of an droughts. extension In fromthis sense, the standardized the SSI precipitationquantifies index the number (SPI, [30 of]) standard to depict deviations hydrologic that aspects the streamflow of droughts. deviates In this from sense, the theclimatological SSI quantifies the numbermean of ofa location, standard by deviationstransforming that monthly the streamflowstreamflows into deviates z-scores from [31]. the For climatologicalthe calculation of mean the of SSI, streamflow series were divided in 12 monthly series of 46 years. Each series was fitted to a a location, by transforming monthly streamflows into z-scores [31]. For the calculation of the SSI, lognormal probability density function, i.e., the distribution that better fits to streamflow records streamflow series were divided in 12 monthly series of 46 years. Each series was fitted to a lognormal across the study area [32]. The 12 probability density functions were transformed to the standard probabilitynormal density distribution function, with i.e.,mean the 0 distribution and variance that 1 to better finally fits obtain to streamflow the SSI time records series. across A detailed the study area [32description]. The 12 of probability the calculation density steps functions for deriving were th transformede SSI is provided to the by standardVicente-Serrano normal et distributional. [29]. with meanThis index 0 and was variance obtained 1 to through finally obtainthe SCI thepackage SSI time for R series. [33], a A commonly detaileddescription used package of to the calculate calculation steps forstandardized deriving drought the SSI isindices. provided In this by study, Vicente-Serrano the SSI was calculated et al. [29]. on This time index scale wasof 3 months obtained (SSI3), through the SCIfollowing package previous for R [33 work], a commonly from [34,35]. used Nevertheless, package to the calculate index can standardized be obtained droughton many indices.time scales In this study,(i.e., the accumulating SSI was calculated streamflow on time monthly scale da ofta 3over months 1, 3, 6, (SSI3), 12 or 24 following months). previous work from [34,35]. Nevertheless,Following the index the assessment can be obtained described on in many [30], timethe hy scalesdrological (i.e., drought accumulating is defined streamflow as a period monthly in which the SSI is continuously negative and the index reaches a value of −1.0 or less. The drought data over 1, 3, 6, 12 or 24 months). begins when the SSI first falls below zero and ends when a positive value of SSI appears, following a Following the assessment described in [30], the hydrological drought is defined as a period in value of −1.0 or less. Thus, streamflow departures from average conditions need to exceed one whichstandard the SSI is deviation. continuously This negativedefinition and facilitates the index clear reaches identifications a value ofof −the1.0 onset or less. and The the droughtend of the begins whenhydrological the SSI first drought falls below event zero and anddetermination ends when of aot positiveher common value used of statistics SSI appears, such followingas the drought a value of −1.0duration, or less. its Thus, magnitude streamflow and severity. departures Three fromdrought average categories conditions were used need for hydrological to exceed one drought standard deviation.assessment, This definition based on facilitatesthe categories clear established identifications for the of theSPI onset to define and themeteorological end of the hydrologicaldrought droughtconditions event and[36]. determinationTable 2 shows the of SSI other categories, common which used are statistics in line with such recent as the research drought based duration, on the its magnitudeSSI [31,32,37,38] and severity. and will Three help drought to compare categories the results were obtained used forconsidering hydrological meteorological drought drought assessment, basedassessment on the categories based on established the SPI over for SSA the (e.g., SPI to[39]). define meteorological drought conditions [36]. Table2 shows the SSI categories, which are in line with recent research based on the SSI [31,32,37,38] and will help to compare the results obtained considering meteorological drought assessment based on the SPI over SSA (e.g., [39]). Water 2017, 9, 652 6 of 18 Water 2017, 9, 652 6 of 18

Table 2.2. StandardizedStandardized streamflowstreamflow indexindex (SSI)(SSI) categories.categories.

CategoryCategory Index Value Excess ≥1.00 Excess ≥1.00 NormalNormal −−0.990.99 to 0.99 ModerateModerate drought drought −−1.49 to −1.001.00 SevereSevere drought drought −−1.99 to −1.501.50 ExtremeExtreme drought drought ≤−≤−2.002.00

3.3. Results

3.1.3.1. The 2010–20152010–2015 HydrologicalHydrological DroughtDrought onon aa RecentRecent HistoricalHistorical ContextContext (1971–2016)(1971–2016) InIn orderorder toto putput thethe 2010–20152010–2015 hydrologicalhydrological droughtdrought inin aa historicalhistorical context,context, wewe analyze,analyze, overover thethe lastlast 4646 years, the variations variations in in the the SSI3 SSI3 for for the the Me Mendozandoza and and Agrio Agrio Rivers Rivers as as representative representative records records of ofCAA CAA and and NP, NP, respectively. respectively. Figure Figure 33 showsshows the the vari variationsations in monthly streamflowstreamflow andand SSI3SSI3 forfor thethe MendozaMendoza RiverRiver fromfrom JulyJuly 19711971 toto JuneJune 2016.2016. Based on FigureFigure3 3,, we we observed observed that that above-average above-average streamflowsstreamflows areare mainlymainly recordedrecorded inin thethe periodsperiods 1977–19951977–1995 andand 2006–2010,2006–2010, whereaswhereas below-averagebelow-average streamflowsstreamflows werewere registered registered in in 1971–1977, 1971–1977, 1995–2001 1995–2001 and and 2010–2016. 2010–2016. Wet Wet and and dry periodsdry periods are not are only not basedonly based on streamflow on streamflow peaks duringpeaks during the warm the season,warm season, but also but along also the along annual the low annual flows. low For flows. instance, For aninstance, increase an(decrease) increase (decrease) in the streamflow in the streamflow during the during low-flow the low-flow season is season observed is observed together together with an increasewith an (decrease)increase (decrease) in the streamflow in the streamflow peaks in summer peaks wetin summer (dry) periods. wet (dry) Regarding periods. the Regarding hydrological the droughthydrological for CAA drought in the for recent CAA period, in the we recent noted peri thatod, severity we noted reached that theseverity extreme reached category the betweenextreme Decembercategory between 2010 and December January 2011, 2010 being and theJanuary lowest 2011 SSI3, being value onthe record. lowestEven SSI3 whenvaluedrought on record. duration Even seemswhen drought to be comparable duration toseems the eventto be betweencomparable 1974 to and the 1977,event the between number 1974 of monthsand 1977, with the SSI3 number values of lowermonths than with−1.0 SSI3 is remarkablevalues lower larger than in − the1.0 mostis remarkable recent drought. larger When in the comparing most recent the drought. Mendoza When River withcomparing the most the rivers Mendoza in the River CAA, with similar the patternsmost rivers regarding in the CAA, severity, similar duration patterns and timingregarding of the severity, recent hydrologicalduration and drought timing emergeof the recent (not shown), hydrological consistent drought with theemerge homogeneous (not shown), behavior consistent of streamflow with the variationshomogeneous across behavior the CAA of [ 21streamflow,40], and particularly variations inacross line withthe CAA the regionalization [21,40], and particularly performed byin [line31] basedwith the on regionalization the SSI for the period performed 1961–2006. by [31] based on the SSI for the period 1961–2006.

FigureFigure 3.3. TemporalTemporal evolutionevolution ofof monthlymonthly streamflowstreamflow (red(red line)line) andand monthlymonthly SSI3SSI3 (black(black line)line) forfor thethe MendozaMendoza riverriver (1413(1413 station)station) overover thethe periodperiod JulyJuly 1971–June1971–June 2016.2016.

StreamflowStreamflow andand SSI3SSI3 variationsvariations acrossacross NP,NP, basedbased onon thethe recordsrecords fromfrom AgrioAgrio RiverRiver areare depicted inin FigureFigure4 4.. BothBoth monthlymonthly streamflowstreamflow andand SSI3SSI3 recordsrecords showshow aa moremore “noisy”“noisy” patternpattern in in comparison comparison withwith thethe MendozaMendoza RiverRiver (Figure(Figure3 ).3). ThisThis resultresult isis consistentconsistent with with Caragunis Caragunis et et al. al. [ [41],41], whowho showedshowed thatthat thethe contributioncontribution ofof lowlow frequencyfrequency variationsvariations toto thethe streamflowstreamflow variabilityvariability accountsaccounts forfor 40%40% andand ~15%~15% ofof the the total total variance variance across across the the CAA CAA and and NP, respectively.NP, respectively. Main Main wet periods wet periods are observed are observed during theduring period the 1979–1983,period 1979–1983, whereas whereas hydrological hydrological droughts droughts are identified are identified in 1996, 1999 in 1996, and 20131999 (Figureand 20134). (Figure 4). Comparing with records from CAA, there are several differences in timing, duration and occurrence of the maximum severity during the recent hydrological drought. Several drought events

Water 2017, 9, 652 7 of 18

ComparingWater 2017 with, 9, 652 records from CAA, there are several differences in timing, duration and occurrence7 of 18 of the maximum severity during the recent hydrological drought. Several drought events were recorded since 2007,were with recorded the maximum since 2007, severities with th betweene maximum 2012 severities and 2014 between (Figure 20124). Thisand 2014 latter (Figure hydrological 4). This latter drought hydrological drought event shows the lowest SSI3 for the Agrio River record, and is comparable in event shows the lowest SSI3 for the Agrio River record, and is comparable in duration and magnitude duration and magnitude with the droughts of 1996/1997 and 1998/1999. Analyzing the spatial with thepattern droughts of the of recent 1996/1997 hydrological and 1998/1999. drought Analyzingacross NP, thewe spatialnoted patternthat most of thebasins recent experienced hydrological droughtextreme across hydrological NP, we noted droughts that most between basins 2012 experienced and 2013, with extreme durations hydrological ranging droughtsfrom seven between to over 2012 and 2013,24 months. with durations The extreme ranging hydrological from seven drought to in over 1998 24/1999 months. shows The the extremelowest SSI3 hydrological value on record drought in in 1998/1999three showsof the five the basins lowest south SSI3 of value 38° S. on record in three of the five basins south of 38◦ S.

FigureFigure 4. Temporal 4. Temporal evolution evolution of monthlyof monthly streamflow streamflow (red(red line) and and monthly monthly SSI3 SSI3 (black (black line) line) for the for the AgrioAgrio river river (2010 (2010 station) station) over over the the period period July July 1971–June 1971–June 2016.

Visual inspection of the SSI3 series for each of the selected rivers indicate that, in the basins of VisualNP, the inspection timing of the of maximum the SSI3 series severity for of each the 2010–2015 of the selected hydrologic riversal drought, indicate its that, duration, in the onset basins of NP, theand timing demise of were the not maximum consistent severity with those of therecord 2010–2015ed for the hydrological drought across drought, the CAA its (Figures duration, 3 and onset and demise4). The werespatial not representativeness consistent with thoseof the recorded2010–2015 for hydrological the drought drought across is the limited CAA (Figuresto the basins3 and 4). The spatiallocated representativeness north of 38° S. Nevertheless, of the 2010–2015 the intensif hydrologicalication of hydrological drought is limiteddroughtto over the NP basins during located north2012/2013 of 38◦ S. Nevertheless,is associated with the the intensification same physical of mechanisms hydrological that drought sustained over the NP drought during conditions 2012/2013 is associatedover the with CAA. the Details same physicalon this aspect mechanisms are provided that in sustained next sections. the drought conditions over the CAA. The percentage of stations within each SSI3 category is useful to quantify the spatial extension Details on this aspect are provided in next sections. of drought events, even when it is limited by the lack of information in the immediate surroundings. TheNevertheless, percentage streamflow of stations integrates within the each hydrologic SSI3 categoryal processes is useful over to a quantifyriver basin the and spatial the selected extension of droughtstations events, belong even to similar when and it is spatially limited bycontiguous the lack ofbasins information over the instudy the area, immediate i.e. providing surroundings. an Nevertheless,acceptable streamflow estimation of integrates the proportion the hydrologicalof area under drought processes conditions. over a riverThe temporal basin and evolution the selected of stationsthis belongindex, represented to similar andin Figure spatially 5, was contiguous obtained for basins each month over theduring study the area,period i.e. 1971–2016 providing by an acceptablecalculating estimation the percentage of the proportion of stations ofwith area SSI3 under below drought the selected conditions. thresholds The (see temporal Table 2). evolution The of thispercentage index, represented of stations in in the Figure moderate5, was drought obtained category for includes each month those duringstations showing the period severe 1971–2016 and by calculatingextreme drought the percentage conditions. ofLikewise, stations the with percentage SSI3 below of stations the selectedin the severe thresholds category (see includes Table 2). The percentagethose stations of stationsshowing inextreme the moderate drought. We drought used th categorye percentage includes of stations those instead stations of showingthe number severe of stations given the length differences in the records between CAA and NP regions, particularly and extreme drought conditions. Likewise, the percentage of stations in the severe category includes after 2015. In general, regional hydrological droughts are observed during 1976/1977, 1991/1992, those1996/1997, stations showing 1998/2000 extreme and 2010–2016, drought. with We more used than the percentage50% of the ofstudy stations area insteadunder hydrological of the number of stationsdrought given conditions the length at different differences severity in levels the records(Figure 5). between During December CAA and 1996, NP regions,January 1997 particularly and after 2015.between In July general, and October regional 2015, hydrological all the analyzed droughts rivers are show observed hydrological during droughts 1976/1977, at different 1991/1992, 1996/1997,severity 1998/2000 levels. The and2010–2015 2010–2016, hydrological with moredrought than is substantially 50% of the longer study than area any under other hydrological drought droughtevent conditions in the last at46 differentyears. On severityaverage, levelsbetween (Figure April 52010). During and December December 2015 1996, (69 months), January 63% 1997 of and betweenthe Julyrivers and along October the 2015,study all area the analyzedwere affected rivers by show moderate hydrological to extreme droughts streamflow at different drought severity levels.conditions. The 2010–2015 In terms hydrological of severity, however, drought the is substantiallydroughts of 1996/1997 longer and than 1999/2000 any other affected drought a larger event in the lastpercentage 46 years. of On stations average, under between extreme Aprildry co 2010nditions, and although December with 2015 shorter (69 durations. months), 63% of the rivers along the study area were affected by moderate to extreme streamflow drought conditions. In terms of severity, however, the droughts of 1996/1997 and 1999/2000 affected a larger percentage of stations under extreme dry conditions, although with shorter durations. Water 2017, 9, 652 8 of 18 Water 2017, 9, 652 8 of 18

Water 2017, 9, 652 8 of 18

FigureFigure 5.5. Temporal evolution of the percentage of stations in the Central Andes of Argentina and Temporal evolution of the percentage of stations in the Central Andes of Argentina and NorthNorth PatagoniaPatagonia with with hydrological hydrological droughts droughts during during 1971–2016. 1971–2016. Yellow, Ye orange,llow, orange, and red and colors red indicate colors Figure 5. Temporal evolution of the percentage of stations in the Central Andes of Argentina and moderate,indicate moderate, severe, and severe, extreme anddrought extreme conditions,drought conditions, respectively. respectively. North Patagonia with hydrological droughts during 1971–2016. Yellow, orange, and red colors In termsindicate of moderate, severity severe, and and spatial extrem extension,e drought conditions, three of respectively.the most important hydrological droughts In terms of severity and spatial extension, three of the most important hydrological droughts over the last 46 years were recorded in 1996/1997, 1999/2000 and 2010–2015 (Figure 5). To compare over theIn last terms 46 years of severity were and recorded spatial in extension, 1996/1997, three 1999/2000 of the most and important 2010–2015 hydrological (Figure5 ).droughts To compare the spatial extension of these droughts during the month showing the maximum intensity, Figure 6 the spatialover the extension last 46 years of thesewere recorded droughts in during1996/1997, the 1999/2000 month showing and 2010–2015 the maximum (Figure 5). intensity, To compare Figure 6 displays the regional patterns of the hydrological drought categories for these three events. As displaysthe spatial the regional extension patterns of these of droughts the hydrological during th droughte month showing categories the for maximum these three intensity, events. Figure As expected 6 expecteddisplays from the Figureregional 5, patternsthe worst of conditions the hydrological in terms drought of severity categories and extensionfor these three are registered events. As during from Figure5, the worst conditions in terms of severity and extension are registered during December Decemberexpected 1996, from withFigure all 5, the worstrecords conditions under droughtin terms ofconditions, severity and 18 extension reaching are the registered severe categoryduring and 1996, with all the records under drought conditions, 18 reaching the severe category and 14 under 14 Decemberunder extreme 1996, with hydrological all the records drought under droughtconditions. conditions, The 18spatial reaching extension the severe of category the hydrological and extreme hydrological drought conditions. The spatial extension of the hydrological drought during drought14 under during extreme January hydrological 1999 seems drought to be conditions. limited to Thethe stationsspatial extension south of of33° the S, hydrologicalalthough the two January 1999 seems to be limited to the stations south of 33◦ S, although the two stations with normal stationsdrought with during normal January conditions 1999 seems have to negativebe limited SSI3 to the almost stations reachi southng of the33° S,moderate although category.the two In a conditionsstations havewith negativenormal conditions SSI3 almost have reaching negative theSSI3 moderate almost reachi category.ng the In moderate a regional category. perspective, In a [31] regional perspective, [31] showed that this drought event extended farther south, affecting the showedregional that perspective, this drought [31] event showed extended that this farther drought south, event affecting extended thefarther basins south, in Centralaffecting Patagonia the basins in Central Patagonia reaching 45° S. During August 2015, only 13 stations located across the reachingbasins 45 in◦ CentralS. During Patagonia August reaching 2015, only45° S. 13 During stations August located 2015, across only 13 the stations CAA havelocated available across the records. CAA have available records. All rivers were under moderate hydrological drought, eight and five of All riversCAA have were available under moderaterecords. All hydrological rivers were under drought, moderate eight hydrological and five of themdrought, reaching eight and the five severe of and them reaching the severe and extreme drought categories, respectively. Hydrological drought extremethem drought reaching categories, the severe respectively. and extreme Hydrological drought categories, drought respectively. severity exhibits Hydrological a more heterogeneousdrought severityseverity exhibits exhibits a amore more heterogeneous heterogeneous pattern,pattern, attrib attributeduted by byRivera Rivera et al. et [25] al. to[25] geomorphological to geomorphological pattern, attributed by Rivera et al. [25] to geomorphological factors within each basin. However, factorsfactors within within each each basin. basin. HoweHowever, additionaladditional studies studies are are needed needed to properlyto properly account account for the for the additional studies are needed to properly account for the difference between basins. differencedifference between between basins. basins.

FigureFigure 6. Spatial 6. Spatial distribution distribution patterns patterns forfor the the hydrological hydrological droughts droughts according according to drought to drought categories categories during:during: December December 1996 1996 (A (A);); January January 1999 ( (BB);); and and August August 2015 2015 (C). (C No). Nodata data are available are available for the for NP the NP gauges during August 2015. gaugesFigure during6. Spatial August distribution 2015. patterns for the hydrological droughts according to drought categories during: December 1996 (A); January 1999 (B); and August 2015 (C). No data are available for the NP gauges during August 2015.

Water 2017, 9, 652 9 of 18 Water 2017, 9, 652 9 of 18

3.2.3.2. Spatial and TemporalTemporal PatternsPatterns ofof thethe 2010–20152010–2015 HydrologicalHydrologicalDrought Drought ForFor assessingassessing the the spatial spatial and and temporal temporal patterns pattern of thes of 2010–2015 the 2010–2015 hydrological hydrological drought, drought, 15 stations 15 fromstations the from CAA the were CAA analyzed. were analyzed. SSI3s from SSI3s July 2009from to July June 2009 2016 to are June shown 2016 in are Figure shown7. In in most Figure gauges, 7. In themost onset gauges, of the the hydrological onset of the drought hydrological started drough during thet started second during half of the 2010. second SSI3s half rapidly of 2010. fall below SSI3s −rapidly1.0, with fall extreme below − droughts1.0, with inextreme the Mendoza, droughts Tupungato in the Mendoza, and Vacas Tupungato Rivers during and Vacas the summer Rivers ofduring 2011 andthe summer severe drought of 2011 conditionsand severe indrought the rest conditions of the rivers in the in the rest CAA. of the Across rivers thein the CAA CAA. the 2010–2015Across the hydrologicalCAA the 2010–2015 drought hydrological is a continuous, drought long-lasting is a contin event.uous, Only long-lasting three stations event. recorded Only three SSI3 positivestations valuesrecorded in OctoberSSI3 positive 2012 values (Diamante in October River), 2012 October-November (Diamante River), 2013 October- (GrandeNovember River) and 2013 November (Grande 2013River) (Barrancas and November River). 2013 Three (Barrancas peaks in droughtRiver). Three intensity peaks are in evident: drought the intensity first between are evident: 2010 and the 2012, first thebetween second 2010 between and 2012, 2014 the and second 2015 and between the last 2014 between and 2015 2015 and and the 2016. last Stations between reaching 2015 and extreme 2016. hydrologicalStations reaching droughts extreme during hydrological the first intensity droughts peak du arering located the first between intensity 32◦ Speak and are 34◦ locatedS, whereas between those related32° S and to the34° secondS, whereas intensity those peakrelated between to the second 34◦ S and intensity 37◦ S. Apeak more between heterogeneous 34° S and spatial 37° S. patternA more isheterogeneous observed for spatial the 2015–2016 pattern intensityis observed peak, for asthe represented 2015–2016 intensity in Figure peak,6. The as hydrological represented droughtin Figure demise6. The hydrological is recorded between drought thedemise end ofis yearrecorded 2015 be andtween the beginningthe end of ofyear 2016, 2015 although and the some beginning stations of continued2016, although under some hydrological stations drought continued conditions under (i.e.,hydrological Tunuyán, drought Vacas, Poti condit Malalions and (i.e., Pincheira, Tunuyán, see TableVacas,1 forPoti more Malal details). and Pincheira, see Table 1 for more details).

Figure 7. SSI3 evolution from July 2009 to June 2016 for the 15 stations located across the Central Figure 7. SSI3 evolution from July 2009 to June 2016 for the 15 stations located across the Central Andes Andes of Argentina. of Argentina.

The spatial distribution of the hydrological drought severity at different stages during the 2010– 2015The drought spatial event distribution is shown in of Figure the hydrological 8 for the mo droughtnths of January, severity May at different and September stages in during the years the 2010–20152011 to 2015. drought As previously event is shownrecorded in Figurefor the8 temporalfor the months variability of January, in SSI3, May differences and September in drought in thespatial years patterns 2011 to arise 2015. from As previouslythe comparison recorded between for thethe temporalCAA and variability NP rivers. in Less SSI3, intense differences drought in droughtseverities spatial are observed patterns over arise the from NP thebasins, comparison except for between the period the between CAA and September NP rivers. 2012 Less and intense May drought2013. As severities shown in are Figure observed 7, the over hydrological the NP basins, drought except develops for the periodand intensifies between quickly, September with 2012 severe and Maydrought 2013. conditions As shown in in Figureall CAA7, thebasins hydrological during Ja droughtnuary 2011. develops The hydrological and intensifies drought quickly, event with severeacross droughtthe CAA conditions decreases in in all severity CAA basins since during January January 2012, 2011.with Thefew hydrologicalstations showing drought moderate event across drought the CAAconditions decreases in September in severity 2012 since and January January 2012, 2013. with This few stationsspatial showingpattern contrasts moderate with drought the increase conditions in inseverity September over 2012NP, suggesting and January that 2013. the Thisintensificatio spatial patternn of drought contrasts could with be the related increase to reduced in severity rainfalls over ◦ NP,south suggesting of 38° S. thatOver the the intensification period 2010–2015, of drought higher could drought be related severity to reducedwas recorded rainfalls in southrivers oflocated 38 S. ◦ Overbetween the period35° S and 2010–2015, 36° S. Water higher demand drought for severity irrigati wason in recorded the CAA in riversis larger located during between spring-summer 35 S and ◦ 36months;S. Water therefore, demand when for irrigationcomparing in the the hydrological CAA is larger drought during categories spring-summer during months; January therefore,it can be whenseen that comparing the large the number hydrological of stations drought under categoriesdrought was during observed January during it can 2011 be and seen 2015. that This the result large numberis in line ofwith stations drought under intensity drought peaks was in observed Figure 7. during 2011 and 2015. This result is in line with drought intensity peaks in Figure7.

Water 2017, 9, 652 10 of 18

Water 2017, 9, 652 10 of 18

FigureFigure 8. 8.Spatial Spatialdistribution distribution ofof hydrologicalhydrological droughts by by categories categories during during January, January, May May and and SeptemberSeptember from from 2011 2011 to to 2015. 2015.

Table 3 summarizes the main hydrological drought characteristics over the period 2010–2016. ExceptTable for3 summarizes the San Juan theRiver, main the hydrologicalremaining gauges drought show characteristics that the onset of over the hydrological the period 2010–2016. drought Exceptoccurred for the between San Juan June River, and October the remaining 2010. The gauges drought show ended that during the onset the first of the pa hydrologicalrt of 2016, although drought occurredfour rivers between still remain June and under October drought 2010. conditions The drought at the ended end of during the records. the first The part mean of 2016,hydrological although fourdrought rivers duration still remain was under67 months, drought being conditions one of the at longest the end dry of periods the records. in the TheCAA mean rivers. hydrological Six rivers droughtregistered duration the lowest was 67SSI3 months, values being(i.e., maximum one of the drought longest severity) dry periods of the in entire the CAA record rivers. (1971–2016). Six rivers registeredA large theheterogeneity lowest SSI3 is valuesobserved (i.e., in maximum the date droughtof the maximum severity) hydrological of the entire drought record (1971–2016). severity, A largeconsistent heterogeneity with the findings is observed on meteorologic in the date ofal the drought maximum conditions hydrological reported drought by [42,43]. severity, consistent with the findings on meteorological drought conditions reported by [42,43].

Water 2017, 9, 652 11 of 18

Water 2017, 9, 652 Table 3. Characteristics of the 2010–2015 hydrological drought. 11 of 18

ID River Onset Demise Duration (Months) Maximum Severity 1205 de los Patos Table August 3. Characteristics 2010 January of the 2016 2010–2015 hydrological 65 drought.−2.20 (July 2015) 1211 San Juan November 2009 January 2016 74 −2.11 (August 2015) 1403ID Atuel River October Onset 2010 June Demise 2016 Duration 68 (Months)−2.01 Maximum (May Severity2014) * 1407 Cuevas September 2010 February 2016 65 −2.83 (October 2013) * 1205 de los Patos August 2010 January 2016 65 −2.20 (July 2015) 14131211 Mendoza San Juan JulyNovember 2010 2009 FebruaryJanuary 2016 2016 67 74 −2.28−2.11 (January (August 2011) 2015) * 14151403 Salado Atuel July October 2010 2010 May June 2016 2016 70 68 −2.17−2.01 (October (May 2014) 2013) * 14191407 Tunuyán Cuevas OctoberSeptember 2010 2010 February - 2016 69 65 −2.17−2.83 (October (October 2015) 2013) * 14201413 Tupungato Mendoza October July 2010 2010 March February 2016 2016 65 67 −2.40−2.28 (September (January 2011)2011) * * 1415 Salado July 2010 May 2016 70 −2.17 (October 2013) 14211419 Vacas Tunuy án March October 2010 2010 - - 76 69 −2.44−2.17 (March (October 2011) 2015) * 14231420 Diamante Tupungato November October 2010 2010 February March 2016 23 + 39 65 *** −2.04−2.40 (September (September 2014) 2011) * 14251421 Poti Malal Vacas July March 2010 2010 - - 72 76 −1.952.44 (June (March 2011) 2011) * 14261423 Pincheira Diamante JulyNovember 2010 2010 February - 2016 23 72 + 39 *** −2.72−2.04 (November (September 2015) 2014) * 1425 Poti Malal July 2010 - 72 −1.95 (June 2011) 1427 Grande June 2010 June 2015 ** 40 + 19 *** −3.09 (May 2014) 1426 Pincheira July 2010 - 72 −2.72 (November 2015) * 20011427 Barrancas Grande July June 2010 2010 June June 2015 2015 ** ** 40 40+ 19 + 19 *** *** −1.88−3.09 (May (May 2014) 2014) 20022001 Colorado Barrancas June July 2010 2010 June June 2016 2015 ** 40 72 + 19 *** −2.45−1.88 (June (May 2015) 2014) Notes:2002 * indicates Colorado that the maximum June 2010 severity during June 2016 the recent drought 72 was the lowest−2.45 SSI3 (June value 2015) over the periodNotes: 1971–2016. * indicates ** indicates that the maximumthe last date severity of the during streamfl the recentow records. drought *** was indicates the lowest that SSI3 two value drought over the events period were 1971–2016. ** indicates the last date of the streamflow records. *** indicates that two drought events were recorded recordedbetween between 2010 and2010 2015, and although 2015, although the period the with period positive with SSI3 positive values was SSI3 shorter values than was 3 months, shorter suggesting than 3 months, that suggestingthe two that drought the two events drou couldght beevents merge could in a single be merg drought.e in a single drought.

3.3.3.3. Drivers Drivers of of the the Hydrological Hydrological Drought Drought SinceSince hydrologicalhydrological drought drought conditions conditions over over the CAA the respondCAA respond to lower thanto lower average than accumulation average accumulationof snow over of the snow Andes over [21 the,44], Andes we explore [21,44], the we relationship explore the between relationship drought between and the drought mean and regional the meansnow regional water equivalent snow water anomaly equivalent (SWEA) anomaly over the (SWEA) period 1989–2015.over the period The annual1989–2015. (July The to June annual across (July the toCAA) June percentageacross the CAA) of stations percentage with hydrological of stations droughtwith hydrological conditions drought highlights conditions the strong highlights link between the stronglarge negative link between (positive) large anomalies negative (positive) in SWEA andanomalies the occurrences in SWEA of and widespread the occurrences hydrological of widespread droughts hydrological(excesses; r = droughts 0.67, p < 0.01;(excesses; Figure r 9=). 0.67, The p largest < 0.01; negative Figure 9). anomalies The largest inSWEA negative were anomalies registered in inSWEA 1996 wereand 1998,registered leading in 1996 to severe and 1998, to extreme leading hydrological to severe to drought extreme conditions hydrological over drought most of conditions the CAA basins over most(Figures of the5 and CAA6). Abasins dry pattern (Figures was 5 alsoand observed6). A dry overpattern NP was in those also years observed (see Figuresover NP4 and in those5), but years the (seesmall Figures contribution 4 and 5), of but rainfall, the small instead contribution of snow, likelyof rainfall, is the instead main driver. of snow, The likely 2010–2015 is the main drought driver. was Theentirely 2010–2015 consistent drought with sixwas years entirely in a rowconsistent showing wi negativeth six years SWEA in anomaliesa row showing (Figure negative9) and a mean SWEA of anomalies68% of the (Figure basins affected9) and a bymean drought of 68% between of the 2010/2011basins affected and 2015/2016.by drought Positive between SWEA 2010/2011 leads and to a 2015/2016.lower percentage Positive of SWEA stations leads showing to a lower droughts percentage (Figure of9). stations showing droughts (Figure 9).

FigureFigure 9. 9. TemporalTemporal evolution evolution of of mean mean regional regional sn snowow water water equivalent equivalent anomalies anomalies (SWEA; (SWEA; 1989–2015) 1989–2015) andand mean mean annual annual (July (July to to June) June) percentage percentage of of st stationsations showing showing hydrologic hydrologicalal droughts (1981/1982 to to 2015/2016).2015/2016).

Water 2017, 9, 652 12 of 18 Water 20172017,, 99,, 652652 12 of 18 Previous research shows that La Niña events are related to lower than average snow Previous research shows that La Niña events are related to lower than average snow accumulationPrevious researchover the shows CAA that[21]. La The Niña sea events surface are temperature related to lower (SST) than anomaly average field snow for accumulation the months accumulation over the CAA [21]. The sea surface temperature (SST) anomaly field for the months overwith theat least CAA 70% [21]. of The the sea gauges surface recording temperature hydrological (SST) anomaly droughts field foris shown the months in Figure with at10 least for the 70% period of the with at least 70% of the gauges recording hydrological droughts is shown in Figure 10 for the period gauges1989–2014. recording A clear hydrological La Niña pattern droughts is evident is shown over in Figurethe tropical 10 for Pacific the period Ocean, 1989–2014. with cold A clearSST over La Niña the 1989–2014. A clear La Niña pattern is evident over the tropical Pacific Ocean, with cold SST over the patternequatorial is evident ocean. overThis thepattern tropical emerges Pacific after Ocean, discar withding cold the SST year over 2015, the which equatorial recorded ocean. a Thisvery patternstrong equatorial ocean. This pattern emerges after discarding the year 2015, which recorded a very strong emergesEl Niño afterevent. discarding Warm SSTs the over year 2015,the subtropical which recorded Pacific a veryOcean strong east ElofNiño Australia event. are Warm produced SSTs over by the El Niño event. Warm SSTs over the subtropical Pacific Ocean east of Australia are produced by the subtropicaladvection of Pacific warm Ocean water east from of Australiathe tropics are toward produced the by subtropics the advection (Figure of warm 10). Based water fromon the the Oceanic tropics advection of warm water from the tropics toward the subtropics (Figure 10). Based on the Oceanic towardNiño Index the subtropics (see [45] for (Figure details), 10). we Based noted on thethat Oceanic the years Niño 2010/2011 Index (see and [45 2011/2012] for details), were we classified noted that as Niño Index (see [45] for details), we noted that the years 2010/2011 and 2011/2012 were classified as theLa yearsNiña 2010/2011years, while and 2012/2013 2011/2012 and were 2013/2014 classified asalso La Niñacharacterized years, while by 2012/2013cold SSTs and over 2013/2014 the tropical also La Niña years, while 2012/2013 and 2013/2014 also characterized by cold SSTs over the tropical characterizedPacific Ocean bynot cold reached SSTs over La Niña the tropical threshold. Pacific Including Ocean not the reached year 2015, La Niña anomalies threshold. over Including the tropical the Pacific Ocean not reached La Niña threshold. Including the year 2015, anomalies over the tropical yearPacific 2015, Ocean anomalies still remain over the below tropical zero, Pacific whereas Ocean the still warm remain region below zero,in the whereas subtropics the warm show region similar in Pacific Ocean still remain below zero, whereas the warm region in the subtropics show similar theanomalous subtropics values show (not similar shown). anomalous values (not shown). anomalous values (not shown).

Figure 10. Composite of SST (Extended Reconstructed Sea Surface Temperature v3b, [46]) anomalies Figure 10. Composite of SST (Extended Reconstructed SeaSea Surface Temperature v3b, [[46])46]) anomalies (1989–2014) during the months showing more than 70% of the stations with hydrological drought (1989–2014)(1989–2014) during the months showing more than 70% ofof thethe stationsstations withwith hydrologicalhydrological droughtdrought conditions. Color areas are significant at the 95% significance level. conditions. Color areas are significantsignificant atat thethe 95%95% significancesignificance level.level. Figure 11 shows the composite anomaly for the 500 hPa geopotential height (Z500) during the Figure 11 shows the compositecomposite anomaly for the 500 hPahPa geopotentialgeopotential height (Z500) during the same months used for developing the SST anomaly composite. Figure 11 displays a strengthening same months used forfor developingdeveloping thethe SSTSST anomalyanomaly composite.composite. Figure 11 displaysdisplays aa strengtheningstrengthening and southward latitudinal location of the semi-permanent south Pacific subtropical anticyclone. This and southward southward latitudinal latitudinal location location of ofthe the semi-perma semi-permanentnent south south Pacific Pacific subtropical subtropical anticyclone. anticyclone. This circulation pattern is associated with a decrease of the westerly zonal winds at subtropical latitudes Thiscirculation circulation pattern pattern is associated is associated with a withdecrease a decrease of the westerly of the westerly zonal winds zonal atwinds subtropical at subtropical latitudes and the decrease of the frontal activity over the study area. latitudesand the decrease and the decreaseof the frontal of the activity frontal over activity the study over the area. study area.

Figure 11. Composite of Z500 (NCEP/NCAR Reanalysis, [47]) anomalies (1989–2014) during the Figure 11.11. Composite of Z500Z500 (NCEP/NCAR(NCEP/NCAR Reanalys Reanalysis,is, [47]) [47]) anomalies (1989–2014) during thethe months showing more than 70% of the stations with hydrological drought conditions. Color areas are months showing more than 70% of the stations with hydrological drought conditions. Color areas are significant at the 95% significance level. significantsignificant atat thethe 95%95% significancesignificance level.level.

Water 2017, 9, 652 13 of 18

4. Discussion The period of persistent below-average snow accumulation over the Central Andes between the years 2009 and 2015 was remarkably anomalous over the last 27 years and significantly reduced the snow contribution to discharge for the major rivers along the adjacent Chilean and Argentinean territories. These rivers sustain the agricultural oases that fed almost 2.5 million inhabitants along the CAA adjacent lands [25]. Therefore, understanding the hydrological drought variability and forcings, as well as the associated impacts to human and ecosystems is a key aspect for water management over the semi-arid CAA. In this sense, the term “snow drought” is gaining traction in the scientific community [48,49], a condition that refers to the combination of general drought and reduced snow storage. Snow drought is not a new hazard across the CAA, we just need to look at the snowpack records from the years 1968 or 1996 to verify that is a recurrent phenomenon commonly linked to meteorological drought conditions over central Chile (see [50] for example). Nevertheless, climate change and climate variability on interannual and interdecadal time scales added new facets to snow droughts, which motivated the definition of dry snow drought—accounting for the lack of precipitation—and warm snow drought—condition where temperatures prevent precipitation from accumulating on the landscape as a snowpack [48]. Further studies are needed to clarify its contribution to the hydrological drought across the CAA. Moreover, several other factors need to be addressed to fully understand the hydrological cycle in the region, as for example the glacier melt contribution during drought periods, the groundwater dynamics or the losses of snow due to sublimation processes. Ground observations are limited across the CAA, highlighting the need to use remote sensing products or modeled hydrometeorological data to address these challenges. Streamflow data were used as a key variable that synthetize the hydrological processes to define hydrological droughts. Even when this kind of drought involves several components of the hydrological cycle, such as snow, groundwater and reservoir and lake levels, streamflow variations were able to depict the main hydrological drought periods across the CAA. Using the SSI [51], found a significant link between streamflow from the main rivers across the CAA and Llancanelo lake areal averages, the major water body in the region. A decrease of 41.6% in the surface of the lake was observed between October and November 2010 [51], in line with the sharp decrease in the SSI3 across the CAA (Figure7). Moreover, the years 2010–2013 are the longest period of sustained small lake size in the period 1984–2013 [51]. The results from our study support the homogeneous behavior of streamflow and hydrological droughts, considering the differences observed between the spatial and temporal variations of the NP basins during the 2010–2015 drought period. The use of the standardized indices for drought characterization is increasing steadily, in line with the needs of monitoring of different hydrological variables across several climatic regions. In Argentina, several agencies as the National Weather Service and the National Institute of Agricultural Technology use the SPI for meteorological drought monitoring and assessment. In this sense, it is expected that the outcomes from the SSI can be easily interpreted by meteorological and hydrological agencies, water managers and scientists focused on drought research and monitoring [31]. One of the main advantages of using the SSI is its flexibility when contrasting hydrological regimes and flow characteristics are involved [52], allowing the comparison among different regions and basins regardless of streamflow magnitudes [53]. We believe that the SSI has the potential to improve hydrological drought monitoring across the CAA and NP and might be integrated with the actual streamflow monitoring, which is based on the percentage of normal streamflow. The effect of multidecadal climate variability limits the applicability of the SSI, as shown by [35]. Nevertheless, this study attempted to limit this negative effect by selecting only 46 years of data. In this sense, water agencies can declare hydrological drought conditions with a minimum of 30 years of data, a period of records that can allow a good density of gauges across the CAA. For hydrological drought declaration, thresholds are a fundamental piece. In this work, the streamflow drought categories (Table1) were selected based on the thresholds typically used to define meteorological droughts based on the SPI. This operational definition can diverge from the conceptual or political definitions of hydrological drought [35]. For example, categories as mild Water 2017, 9, 652 14 of 18 drought can be used to have a better picture than the wide “normal” category, which covers 68% of the SSI variability. This limitation was observed in Figure6, in which the rivers of San Juan province were under normal conditions but the SSI3 value was close to −1.0. Long-term trends in streamflows across the CAA are highly dependent on the period considered for the analysis, as shown by [21,54,55] among others. Recent negative trends in snow cover extent during 1979–2014 were mainly attributed to changes between 3000 and 5000 m a.s.l. [56], and can be responsible for the recent trends in streamflow across the CAA. Further studies are needed to properly quantify the role of the Pacific Decadal Oscillation (PDO) in the 2010–2015 hydrological drought, considering longer streamflow records to adequately represent the interdecadal variability. A strong non-linearity in the behavior of long-term trends was observed considering the seasonal number of days with low flows [25], linked to the decadal behavior of the PDO. The cold (warm) phase of the PDO is linked to below (above)—average winter snowpacks and annual discharges [25,57]. Nevertheless, Boisier et al. [5] have shown that the PDO accounts for half of the recent precipitation trends recorded in the Central Andes. This negative trend in precipitation reconciles with model results only if the anthropogenic forcing and the observed SST changes are accounted for jointly. When considering future changes in the hydroclimatology of the Central Andes, the region is likely to become drier and warmer in response to anthropogenic climate change [22,23,58–60], with an increase in meteorological drought frequency and severity [42]. These projected changes will likely change the characteristics of future hydrological drought across the CAA, an assessment that needs to be performed to assist adaptation strategies.

5. Conclusions The 2010–2015 hydrological drought across the semi-arid CAA and NP regions was assessed in terms of streamflow variations through the SSI3. The main characteristics of this dry episode—spatial distribution, temporal evolution, and timing of severity—indicate that its spatial extension was limited to the basins north of 38◦ S. This result prevents concluding that the recent hydrological drought affected in a similar way the CAA and NP. The northern extension is limited due to the data availability until 31◦ S, but an assessment of the drought event affecting Central Chile [6] showed similar drought conditions as north as 28◦ S. In this sense, the core region was the CAA, where snow accumulation plays a relevant role in the streamflow amount and timing. In a regional perspective, 68% of the CAA basins were continuously affected by hydrological drought from moderate to extreme severity between 2010 and 2016. The mean duration of hydrological drought was 67 months—i.e., five years and seven months—standing out as the longest drought period of the record (1971–2016). Moreover, the temporal consistency of the drought event is comparable on the long-term context with the drought of 1967/1968 to 1971/1972 [61]. Taking into account population growth, impacts of the recent hydrological drought could be stronger than the recorded during the period 1967/1968 to 1971/1972. We need to point out that four of the analyzed rivers continued under hydrological drought until the end of our records (June 2016), but with SSI3 values close to zero. Six of the rivers recorded the lowest SSI3 value during the 2010–2015 drought, with 14 out of 15 stations reaching the extreme category. The role of snow accumulation was assessed through the SWEA, identifying that the regional hydrological droughts were associated to lower than normal snow accumulation over the Central Andes. This feature is likely to be linked to La Niña pattern, with cold SSTs over the tropical Pacific Ocean and positive anomalies in the subtropical Pacific east of Australia. This oceanic forcing generated a rainfall shortage that contributed to almost 100 days with low flows in the basins of NP between 2012 and 2013 [26], a factor that was observed through severe hydrological drought conditions mainly between September 2012 and May 2013. Nevertheless, even when El Niño-Southern Oscillation seems to be one of the main forcings on hydrological drought variability over the CAA and NP, a lack of spatio-temporal consistency in the 2010–2015 drought was observed, indicating that other factors could be playing a relevant role in the spatial extension of hydrological drought and its severity. Further studies are needed to clarify this issue. Water 2017, 9, 652 15 of 18

Acknowledgments: This work was carried out with the aid of the following projects: UBACyT 20020130100263BA from the University of Buenos Aires (UBA) and PIP0137 from the National Research Council of Argentina (CONICET). We thank the Subsecretaría de Recursos Hídricos de Argentina for providing the snow and streamflow records used in the study. Author Contributions: J.A.R. came up with the concept for the manuscript, conducted the analysis and wrote the manuscript with the inputs from all co-authors (O.C.P., R.V. and D.C.A.). Conflicts of Interest: The authors declare no conflict of interest.

References

1. Schwabe, K.; Albiac, J.; Andreu, J.; Ayers, J.; Caiola, N.; Hayman, P.; Ibanez, C. Summaries and Considerations. In Drought in Arid and Semi-Arid Regions; Schwabe, K., Albiac, J., Connor, J.D., Hassan, R.M., Meza-González, L., Eds.; Springer: Dordrecht, The Netherlands, 2013; pp. 471–507. 2. Tallaksen, L.M.; Van Lanen, H.A.J. Hydrological Drought: Processes and Estimation Methods for Streamflow and Groundwater; Elsevier: Amsterdam, The Netherlands, 2004. 3. Dai, A. Drought under global warming: A review. WIREs Clim. Chang. 2011, 2, 45–65. [CrossRef] 4. Cai, W.; Cowan, T.; Thatcher, M. Rainfall reductions over Southern Hemisphere semi-arid regions: The role of subtropical dry zone expansion. Sci. Rep. 2012, 2, 702. [CrossRef][PubMed] 5. Boisier, J.P.; Rondanelli, R.; Garreaud, R.D.; Muñoz, F. Anthropogenic and natural contributions to the Southeast Pacific precipitation decline and recent megadrought in central Chile. Geophys. Res. Lett. 2016, 43, 413–421. [CrossRef] 6. Garreaud, R.; Alvarez-Garreton, C.; Barichivich, J.; Boisier, J.P.; Christie, D.; Galleguillos, M.; LeQuesne, C.; McPhee, J.; Zambrano-Bigiarini, M. The 2010–2015 mega drought in Central Chile: Impacts on regional hydroclimate and vegetation. Hydrol. Earth Syst. Sci. Discuss. 2017.[CrossRef] 7. Marengo, J.A.; Alvala, R.S.; Moraes, O.L.; Brito, S.; Cunha, A.P. The 2010–2016 drought in the semiarid Northeast Brazil region: Monitoring, climatic features and socio-economic impacts. In Proceedings of the Multihazard Early Warning Conference, Cancún, México, 22–23 May 2017. 8. Paredes-Trejo, F.; Barbosa, H. Evaluation of the SMOS-Derived Soil Water Deficit Index as Agricultural Drought Index in Northeast of Brazil. Water 2017, 9, 377. [CrossRef] 9. Van Dijk, A.I.J.M.; Beck, H.E.; Crosbie, R.S.; De Jeu, R.A.M.; Liu, Y.Y.; Podger, G.M.; Timbal, B.; Viney, N.R. The Millennium drought in southeast Australia (2001–2009): Natural and human causes and implications for water resources, ecosystems, economy, and society. Water Resour. Res. 2013, 49, 1040–1057. [CrossRef] 10. Bureau for Food and Agricultural Policy. Policy Brief on the 2015/2016 Drought; BFAP: Cape Town, South Africa, 2016. 11. Montaña, E.; Diaz, H.P.; Hurlbert, M. Development, local livelihoods, and vulnerabilities to global environmental change in the South American Dry Andes. Reg. Environ. Chang. 2016, 16, 2215–2228. [CrossRef] 12. Agosta, E.A.; Cavagnaro, M. Variaciones interanuales de la precipitación de verano y el rendimiento del cultivo de la vid en Mendoza. Geoacta 2010, 35, 1–10. 13. Villatoro, G. Cortocircuitos por la creciente crisis hídrica del río Atuel. Diario Uno, 25 January 2015. Available online: http://www.diariouno.com.ar/pais/cortocircuitos-la-creciente-crisis-hidrica-del-rio- atuel-20150118-n43593.html (accessed on 26 May 2017). 14. Romanello, C. Por la sequía, el turismo en el dique Ullum es cada vez menor. Diario Los Andes, 19 March 2015. Available online: http://www.losandes.com.ar/noticia/por-la-sequia-el-turismo-en-el-dique-ullum-es- cada-vez-menor-838737 (accessed on 26 May 2017). 15. Diario Los Andes. En sólo 11 días se hicieron 470 multas por derroche de agua en Mendoza. Diario Los Andes, 13 November 2014. Available online: http://www.losandes.com.ar/article/en-solo-11-dias-se-hicieron-470- multas-por-derroche-de-agua-en-mendoza (accessed on 26 May 2017). 16. Diario Uno. Aumentó el valor de las multas por derrochar agua: Las más caras serán de $5.700. Diario Uno, 20 March 2017. Available online: http://www.diariouno.com.ar/mendoza/aumento-el-valor-las-multas- derrochar-agua-las-mas-caras-seran-5700--20170320-n1361153.html (accessed on 26 May 2017). 17. McIntyre, O. Environmental Protection of International Water Courses under International Law; Ashgate Publishing Limited: Hampshire, UK, 2007. Water 2017, 9, 652 16 of 18

18. Rojas, F.; Wagner, L. Conflicto por la apropiación del río Atuel entre Mendoza y La Pampa (Argentina). HALAC 2016, 6, 278–297. [CrossRef] 19. Nikbakht, J.; Tabari, H.; Hosseinzadeh Talaee, P. Streamflow drought severity analysis by percent of normal index (PNI) in northwest Iran. Theor. Appl. Climatol. 2013, 112, 565–573. [CrossRef] 20. Masiokas, M.H.; Christie, D.C.; Le Quesne, C.; Pitte, P.; Ruiz, L.; Villalba, R.; Luckman, B.H.; Berthier, E.; Nussbaumer, S.U.; González-Reyes, A.; et al. Reconstructing the annual mass balance of the Echaurren Norte glacier (Central Andes, 33.5 S) using local and regional hydroclimatic data. Cryosphere 2016, 10, 927–940. [CrossRef] 21. Masiokas, M.; Villalba, R.; Luckman, B.; Le Quesne, C.; Aravena, J.C. Snowpack variations in the Central Andes of Argentina and Chile, 1951–2005: Large-scale atmospheric influences and implications for water resources in the region. J. Clim. 2006, 19, 6334–6352. [CrossRef] 22. Boninsegna, J.; Villalba, R. Los escenarios de Cambio Climático y el impacto en los caudales. In Documento Sobre la Oferta Hídrica en los Oasis de Riego de Mendoza y San Juan en Escenarios de Cambio Climático; Secretaría de Ambiente y Desarrollo Sustentable: Buenos Aires, Argentina, 2006. 23. Villalba, R.; Boninsegna, J.A.; Masiokas, M.H.; Cara, L.; Salomón, M.; Pozzoli, J. Cambio climático y recursos hídricos. El caso de las tierras secas del oeste argentino. Ciencia Hoy 2016, 25, 48–55. 24. Jain, V.K.; Jain, M.K.; Pandey, R.P. Effect of the Length of the Streamflow Record on Truncation Level for Assessment of Streamflow Drought Characteristics. J. Hydrol. Eng. 2014, 19, 1361–1373. [CrossRef] 25. Rivera, J.A.; Araneo, D.C.; Penalba, O.C. Threshold level approach for streamflow droughts analysis in the Central Andes of Argentina: A climatological assessment. Hydrol. Sci. J. 2017.[CrossRef] 26. Rivera, J.A.; Araneo, D.C.; Penalba, O.C.; Villalba, R. Regional aspects of streamflow droughts in the Andean rivers of Patagonia, Argentina. Links with large-scale climatic oscillations. Hydrol. Res. 2017.[CrossRef] 27. Araneo, D.; Villalba, R. Variability in the annual cycle of the Río Atuel streamflows and its relationship with tropospheric circulation. Int. J. Climatol. 2015, 35, 2948–2967. [CrossRef] 28. Schwerdtfeger, W. The atmospheric circulation over Central and South America. In Climates of Central and South America; Schwerdtfeger, W., Ed.; Elsevier: New York, NY, USA, 1976; Volume 2, pp. 2–12. 29. Vicente-Serrano, S.M.; López-Moreno, J.I.; Beguería, S.; Lorenzo-Lacruz, J.; Azorin-Molina, C.; Morán-Tejada, E. Accurate computation of a streamflow drought index. J. Hydrol. Eng. 2012, 17, 318–332. [CrossRef] 30. McKee, T.B.; Doesken, N.J.; Kleist, J. The relationship of drought frequency and duration to time scales. In Proceedings of the Eight Conference on Applied Climatology, Anaheim, CA, USA, 17–23 January 1993; American Meteorological Society: Boston, MA, USA, 1993; pp. 179–184. 31. Rivera, J.A.; Penalba, O.C. Spatio-temporal assessment of streamflow droughts over Southern South America: 1961–2006. Theor. Appl. Climatol. 2017, 1–13. [CrossRef] 32. Rivera, J.A.; Penalba, O.C. Distribución de probabilidades de los caudales mensuales en las regiones de Cuyo y Patagonia (Argentina). Aplicación al monitoreo de sequías hidrológicas. Meteorológica 2017, in press. 33. Gudmundsson, L.; Stagge, J.H. SCI: Standardized Climate Indices such as SPI, SRI or SPEI; R Package Version 1.0-2; ETH: Zurich, Switzerland, 2016. 34. Shukla, S.; Wood, A.W. Use of a standardized runoff index for characterizing hydrologic drought. Geophys. Res. Lett. 2008, 35, L02405. [CrossRef] 35. Núñez, J.; Rivera, D.; Oyarzún, R.; Arumí, J.L. On the use of Standardized Drought Indices under decadal climate variability: Critical assessment and drought policy implications. J. Hydrol. 2014, 517, 458–470. [CrossRef] 36. Lloyd-Hughes, B.; Saunders, M.A. A drought climatology for . Int. J. Climatol. 2002, 22, 1571–1592. [CrossRef] 37. Barker, L.J.; Hannaford, J.; Chiverton, A.; Svensson, C. From meteorological to hydrological drought using standardised indicators. Hydrol. Earth Syst. Sci. 2016, 20, 2483–2505. [CrossRef] 38. Rangecroft, S.; Van Loon, A.F.; Maureira, H.; Verbist, K.; Hannah, D.M. Multi-method assessment of reservoir effects on hydrological droughts in an arid region. Earth Syst. Dyn. Discuss. 2016.[CrossRef] 39. Penalba, O.C.; Rivera, J.A. Precipitation response to El Niño/La Niña events in Southern South America—Emphasis in regional drought occurrences. Adv. Geosci. 2016, 42, 1–14. [CrossRef] Water 2017, 9, 652 17 of 18

40. Compagnucci, R.H.; Araneo, D.C. Identificación de áreas de homogeneidad estadística para los caudales de ríos andinos argentinos y su relación con la circulación atmosférica y la temperatura superficial del mar. Meteorológica 2005, 30, 41–53. 41. Caragunis, J.; Rivera, J.A.; Penalba, O. Variabilidad de baja frecuencia en los caudales de los ríos del centro-norte de Argentina y su contribución al desarrollo de sequías hidrológicas. In Proceedings of the XXVIII Reunión Científica de la Asociación Argentina de Geofísicos y Geodestas, La Plata, Buenos Aires, Argentina, 17–21 April 2017. 42. Penalba, O.C.; Rivera, J.A. Future changes in drought characteristics over Southern South America projected by a CMIP5 ensemble. Am. J. Clim. Chang. 2013, 2, 173–182. [CrossRef] 43. Rivera, J.A. Aspectos Climatológicos de las Sequías Meteorológicas en el Sur de Sudamérica. Análisis Regional y Proyecciones Futuras. Ph.D. Thesis, University of Buenos Aires, Buenos Aires, Argentina, 23 March 2014. 44. Rivera, J.A.; Araneo, D.C.; Penalba, O.C.; Villalba, R. Linking climate variations with the hydrological cycle over the semi-arid Central Andes of Argentina. Past, present and future, with emphasis on streamflow droughts. In Geophysical Research Abstracts 19, EGU2017-995, Proceedings of the European Geosciences Union 2017 General Assembly, Vienna, Austria, 23–28 April 2017; EGU: Vienna, Austria, 2017. 45. Historical El Niño/La Niña Episodes (1950–Present). Available online: http://www.cpc.ncep.noaa.gov/ products/analysis_monitoring/ensostuff/ensoyears.shtml (accessed on 26 May 2017). 46. Smith, T.M.; Reynolds, R.W.; Peterson, T.C.; Lawrimore, J. Improvements to NOAA’s Historical Merged Land-Ocean Surface Temperature Analysis (1880–2006). J. Clim. 2008, 21, 2283–2296. [CrossRef] 47. Kalnay, E.; Kanamitsu, M.; Kistler, R.; Collins, W.; Deaven, D.; Gandin, L.; Iredell, M.; Saha, S.; White, G.; Woollen, J.; et al. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol. Soc. 1996, 77, 437–470. [CrossRef] 48. Harpold, A.A.; Dettinger, M.; Rajagopal, S. Defining snow drought and why it matters. Eos 2017, 98. [CrossRef] 49. Snow Drought. Available online: http://europeandroughtcentre.com/2017/04/20/snow-drought/ (accessed on 26 May 2017). 50. Masiokas, M.H.; Villalba, R.; Christie, D.A.; Betman, E.; Luckman, B.H.; Le Quesne, C.; Prieto, M.R.; Mauget, S. Snowpack variations since AD 1150 in the Andes of Chile and Argentina (30◦–37◦ S) inferred from rainfall, tree-ring and documentary records. J. Geophys. Res. 2012, 117, D05112. [CrossRef] 51. Bianchi, L.; Rivera, J.; Rojas, F.; Britos Navarro, M.; Villalba, R. A regional water balance indicator inferred from satellite images of an Andean endorheic basin in central-western Argentina. Hydrol. Sci. J. 2017, 62, 533–545. [CrossRef] 52. Hong, X.; Guo, S.; Zhou, Y.; Xiong, L. Uncertainties in assessing hydrological drought using streamflow drought index for the upper Yangtze River basin. Stoch. Environ. Res. Risk Assess. 2015, 29, 1235–1247. [CrossRef] 53. Lorenzo-Lacruz, J.; Morán-Tejeda, E.; Vicente-Serrano, S.M.; López-Moreno, J.L. Streamflow droughts in the Iberian Peninsula between 1945 and 2005: Spatial and temporal patterns. Hydrol. Earth Syst. Sci. 2013, 17, 119–134. [CrossRef] 54. Carril, A.F.; Doyle, M.E.; Barros, V.R.; Núñez, M.N. Impacts of climate change on the oases of the Argentinean cordillera. Clim. Res. 1997, 9, 121–129. [CrossRef] 55. Lauro, C.; Vich, A.; Moreiras, S.M. Variabilidad del régimen fluvial en cuencas de la región de Cuyo. Geoacta 2016, 40, 28–51. 56. Mernild, S.H.; Liston, G.E.; Hiemstra, C.A.; Malmros, J.K.; Yde, G.C.; McPhee, J. The Andes Cordillera. Part I: Snow distribution, properties, and trends (1979–2014). Int. J. Climatol. 2017, 37, 1680–1698. [CrossRef] 57. Masiokas, M.H.; Villalba, R.; Luckman, B.H.; Mauget, S. Intra- to multidecadal variations of snowpack and streamflow records in the Andes of Chile and Argentina between 30◦ and 37◦ S. J. Hydrometeorol. 2010, 11, 822–831. [CrossRef] 58. Barnett, T.P.; Adam, J.C.; Lettenmaier, D.P. Potential impacts of a warming climate on water availability in snow-dominated regions. Nature 2005, 438, 303–309. [CrossRef][PubMed] 59. Vicuña, S.; Garreaud, R.D.; McPhee, J. Climate change impacts on the hydrology of a snowmelt driven basin in semiarid Chile. Clim. Chang. 2011, 105, 469–488. [CrossRef] Water 2017, 9, 652 18 of 18

60. Vicuña, S.; McPhee, J.; Garreaud, R.D. Agriculture Vulnerability to Climate Change in a Snowmelt-Driven Basin in Semiarid Chile. J. Water Resour. Plan. Manag. 2012, 138, 431–441. [CrossRef] 61. Rivera, J.A.; Araneo, D.C.; Penalba, O.C. Streamflow droughts over the central Andes of Argentina: Current crisis (2010–2014) in historical perspective. In Proceedings of the 11th International Conference on Southern Hemisphere, Santiago de Chile, Chile, 5–9 October 2015.

© 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/).