1 VARIABILIDAD CLIMATICA Y ECOSISTEMASCLAUDIO MENENDEZ, ANDREA CARRIL, PEDRO FLOMBAUM, ANNA SORENSSON CIMA/CONICET-UBA, DCAO/FCEN, UMI IFAECI/CNRS IANIGLA, NOV. 2011

2 ANDES: FUERTES GRADIENTES EN PARÁMETROS CLIMÁTICOS

3

4 Cfa: templado, sin estación seca y verano caliente

5

6 BWh: árido-desierto caliente BWk: árido-desierto fríoBSh: árido- estepa caliente ET: polar-tundra Csa: templado, verano seco y caliente Cwa: templado, con invierno seco y verano caliente Csb: templado, verano seco y cálido Cfa: templado, sin estación seca y verano caliente BSk: árido- estepa fría GCMs y RCMs: resolución demasiado gruesa para representar detalles topográficos (y otros forzantes relacionados con características del terreno y de la vegetación) La simulación de la precipitación en zonas montañosas es poco confiable Los modos naturales de variabilidad climática (p.e. ENSO, SAM) pueden modular los patrones de precipitación en diferentes escalas temporales (los modelos acoplados globales tienen dificultades para capturar estos mecanismos) Existen estudios que muestran que modelos de alta resolución pueden simular los patrones de mesoescala observados (pero son muy caros en modo climático) Cfb: templado, sin estación seca y con verano cálido BWk: árido-desierto frio

7 f (BIODIVERSIDAD, CLIMA, …)PROCESOS ECOLOGICOS = f (BIODIVERSIDAD, CLIMA, …)

8 f (BIODIVERSIDAD, CLIMA, …)PROCESOS ECOLOGICOS = f (BIODIVERSIDAD, CLIMA, …) Colaborador externo: Osvaldo Sala - Arizona State University

9 BIODIVERSIDAD Y ESTABILIDADHIPOTESIS #1: BIODIVERSIDAD α ESTABILIDAD i.e. ecosistemas con alta diversidad son más estables que los ecosistemas con baja diversidad (Elton 1958, Tilman1996) PPNA alta biodiversidad baja biodiversidad tiempo

10 BIODIVERSIDAD Y ESTABILIDADHIPOTESIS #1: BIODIVERSIDAD α ESTABILIDAD i.e. ecosistemas con alta diversidad son más estables que los ecosistemas con baja diversidad (Elton 1958, Tilman1996) ~ PORTFOLIO EFFECT PPNA alta biodiversidad baja biodiversidad tiempo

11 BIODIVERSIDAD Y ESTABILIDADHIPOTESIS #2: PPNA α BIODIVERSIDAD i.e. tasa de funcionamiento de un ecosistema es proporcional al número de especies net primary production (NPP) La biota es responsable de una fracción importante del intercambio de materia y energía en los ecosistemas. Ante escenarios de reducción de la biodiversidad, es necesario comprender cómo se verá afectado el funcionamiento de los ecosistemas. Las hipótesis que vinculan a la biodiversidad con el funcionamiento de los ecosistemas proponen que tanto la tasa de funcionamiento como la estabilidad de los ecosistemas aumentan con la biodiversidad. Varios experimentos exploraron la relación entre la riqueza de especies de plantas y la productividad primaria neta, que es la principal entrada de energía a los ecosistemas. Estos experimentos muestran un claro patrón de aumento en la productividad con el número de especies como resultado de el uso complementario de los recursos, efectos positivos entre especies, y la presencia de especies más productivas (efecto de muestro). La cantidad de experimentos que vinculan la biodiversidad con la estabilidad del ecosistema es menor, y si bien sugieren que la biodiversidad amortigua la variabilidad ambiental la evidencia es todavía inconclusa.

12 BIODIVERSIDAD Y ESTABILIDADHIPOTESIS #2: PPNA α BIODIVERSIDAD i.e. tasa de funcionamiento de un ecosistema es proporcional al número de especies Es posible estimar series de NPP a partir de la dendrocronologia ? net primary production (NPP) La biota es responsable de una fracción importante del intercambio de materia y energía en los ecosistemas. Ante escenarios de reducción de la biodiversidad, es necesario comprender cómo se verá afectado el funcionamiento de los ecosistemas. Las hipótesis que vinculan a la biodiversidad con el funcionamiento de los ecosistemas proponen que tanto la tasa de funcionamiento como la estabilidad de los ecosistemas aumentan con la biodiversidad. Varios experimentos exploraron la relación entre la riqueza de especies de plantas y la productividad primaria neta, que es la principal entrada de energía a los ecosistemas. Estos experimentos muestran un claro patrón de aumento en la productividad con el número de especies como resultado de el uso complementario de los recursos, efectos positivos entre especies, y la presencia de especies más productivas (efecto de muestro). La cantidad de experimentos que vinculan la biodiversidad con la estabilidad del ecosistema es menor, y si bien sugieren que la biodiversidad amortigua la variabilidad ambiental la evidencia es todavía inconclusa.

13 BIODIVERSIDAD Y ESTABILIDAD: DISEÑO DEL EXPERIMENTOSeleccionar sitios correspondientes a diferentes tipos de clima; En cada sitio debe haber un gradiente de diversidad de especies de árboles; Estimación de la productividad a partir de anillos de los árboles; Estimación de la variabilidad climática a partir de reanálisis y observaciones

14 DOES TREE DIVERSITY BUFFERS CLIMATE VARIABILITY?High tree diversity Low tree diversity San Martín Junín Mendoza

15 MODELADO DINAMICO DE LA VEGETACIONY DEL CLIMA Entre otros varios parametros, el modelo calcula NPP y tiene en cuenta hasta 7 especies vegetales distintas en cada punto de grilla.

16 MODELADO DINAMICO DE LA VEGETACIONY DEL CLIMA Entre otros varios parametros, el modelo calcula NPP y tiene en cuenta hasta 7 especies vegetales distintas en cada punto de grilla. Colaborador externo: Patrick Samuelsson - Rossby Centre, Swedish Meteorological and Hydrological Institute

17 RCA3 COUPLED TO THE DYNAMIC VEGETATION MODEL LPJ-GUESS:RCA-GUESS El plan de investigación de Anna propone investigar la interacción clima-vegetación en Sudamérica empleando un RCM de última generación al cual se le acoplará un nuevo esquema de vegetación dinámica.

18 RCA3 COUPLED TO THE DYNAMIC VEGETATION MODEL LPJ-GUESS:RCA-GUESS Tareas: Simular la vegetación del pasado reciente Forzar LPJ-GUESS con climatologías observacionales  i) off line forzado por CRU, ii) acoplado con RCA forzado por reanálisis Caracterizar las interacciones entre la vegetación y la atmósfera (análisis estadísticos) Evaluar el rol de las interacciones vegetación-clima en un contexto de cambio climático  Simular el período con y sin vegetación dinámica

19 ** necesitamos datos de NPP para evaluar el modelo**RCA3 COUPLED TO THE DYNAMIC VEGETATION MODEL LPJ-GUESS: RCA-GUESS Tareas: Simular la vegetación del pasado reciente Forzar LPJ-GUESS con climatologías observacionales  i) off line forzado por CRU, ii) acoplado con RCA forzado por reanálisis ** necesitamos datos de NPP para evaluar el modelo** Caracterizar las interacciones entre la vegetación y la atmósfera (análisis estadísticos) Evaluar el rol de las interacciones vegetación-clima en un contexto de cambio climático  Simular el período con y sin vegetación dinámica

20 Se grafican las áreas en las que cambian los biomasEJEMPLO DE CAMBIOS EN LA VEGETACION SIMULADOS POR UN MODELO DINAMICO DE VEGETACION (UNA VERSION DE LPJ) Metodología Se hace corresponder a cada celda una de las 18 categorías de vegetación (bioma) Se fuerza el modelo de vegetación con un AOGCM y se comparan clima “actual” y “futuro” Se grafican las áreas en las que cambian los biomas Our simulation showed that by 2045, the most extensive changes are likely to occur in the parts of the world that are already dry—such as central Asia—and along the existing boundaries of major vegetation types (Fig 1A). The model predicts increases in the amount of vegetation and woody cover in many dry areas because an increase in atmospheric CO2 will allow plants to survive increased levels of drought. Furthermore, the boundaries of major vegetation types will shift along both moisture and temperature gradients. Thomas et al., 2008

21 EJEMPLO DE CAMBIOS EN LA VEGETACION SIMULADOS POR UN MODELO DINAMICO DE VEGETACION (UNA VERSION DE LPJ) Our simulation showed that by 2045, the most extensive changes are likely to occur in the parts of the world that are already dry—such as central Asia—and along the existing boundaries of major vegetation types (Fig 1A). The model predicts increases in the amount of vegetation and woody cover in many dry areas because an increase in atmospheric CO2 will allow plants to survive increased levels of drought. Furthermore, the boundaries of major vegetation types will shift along both moisture and temperature gradients. CAMBIOS EN LAI Thomas et al., 2008

22 VARIABILIDAD CLIMATICA

23 PRECIP. Y TEMP. EN SANTIAGO DE CHILEThe position of the jet axis coincides with zone of maximum meridional temperature gradient, and therefore is a proxy for the preferred path, or storm track, of synoptic-scale disturbances (Trenberth, 1991). Although subtle, there is an intensification and equatorward shift of the storm track and similar changes in the strength and position of the subtropical anticyclones during the austral winter (JJA; Physick, 1981), thus leading to the rainy season in central Chile and Argentina (30–40°S).

24 VECTOR VIENTO E ISOTACAS EN 200 hPaPRECIP. Y TEMP. EN SANTIAGO DE CHILE The position of the jet axis coincides with zone of maximum meridional temperature gradient, and therefore is a proxy for the preferred path, or storm track, of synoptic-scale disturbances (Trenberth, 1991). Although subtle, there is an intensification and equatorward shift of the storm track and similar changes in the strength and position of the subtropical anticyclones during the austral winter (JJA; Physick, 1981), thus leading to the rainy season in central Chile and Argentina (30–40°S).

25 PRECIP. Y TEMP. EN MENDOZA

26 PRECIP. Y TEMP. EN MENDOZAPRECIPITACION Y LINEAS DE CORRIENTE EN 200 hPa PRECIP. Y TEMP. EN MENDOZA

27 REGRESION ENTRE INDICE SAM Y PRESION EN SUPERFICIE ANOMALIAS 500 hPa, JULIO 1987 (EL NIÑO) Changes in the global atmospheric circulation induced by anomalies in the ocean-atmosphere system in the tropical Pacific are at the origin of many ENSO-related climate anomalies in subtropical and extratropical areas around the globe (that is, teleconnection patterns). In South America, these signals are significant along the subtropical western border (central Chile) and over the southeast portion of the continent. Regarding the links between SST anomalies in the tropical Pacific and rainfall, a warm-wet / cold-dry relationship over central Chile and the subtropical Andes during the wet season (May–September) has been documented in several studies (Quinn and Neal, 1983; Kiladis and Diaz, 1989; Aceituno and Garreaud, 1995). A detailed analysis of this ENSO signal revealed that the tendency for positive (negative) rainfall anomalies during El Niño (La Niña) is significant in the 33°S to 36°S latitudinal band during the austral winter, while from 36°S to 39°S the same signal is best defined during the spring (Montecinos et al., 2000). The tendency to above average wintertime rainfall during El Niño years is consistent with a relatively higher frequency of blocking anticyclones to the west of the Antarctic Peninsula, and the subsequent northward displacement of the South Pacific storm track (Rutllant and Fuenzalida, 1991). As the blocks in the southeast Pacific remain stationary for 5 to 15 days (Sinclair, 1996; Renwick, 1998), the mid-latitude storm track shifts equatorward, producing stormy conditions in central Chile and dry, cold conditions at the southern tip of the continent (Rutllant and Fuenzalida, 1991). Rutllant y Fuenzalida, 1991 Goosse et al., 2010

28 Annalisa Cherchi y Andrea CarrilANOMALIAS DE SST Z* (200 hPa) It has been proposed that the summertime circulation anomalies over South America (fig. 3.5) are part of large scale wave train emerging from the South Pacific, referred to as the Pacific South American (PSA) modes. During wintertime, the PSA modes dominate the intraseasonal variability in the SH (Ghil and Mo, 1991; Mo and Higgins, 1997). Mo and Higgins (1997) suggest that tropical convection over the western Pacific serves as a catalyst in the development of standing PSA modes, therefore connecting intraseasonal variability in the remote tropics with intraseasonal variability in subtropical South America. The PSA modes have also been associated with the onset of blocking anticyclones in the south Pacific to the west of the Antarctic Peninsula (Renwick and Revell, 1999), and whose subsequent maintenance arises from a complex interaction between the mean flow and the transient disturbances (Marques and Rao, 1999). The tropospheric-deep anticyclonic anomalies (barotropic structure) tend to split the mid-latitude zonal flow into equatorward and poleward branches. As the blocks in the southeast Pacific remain stationary for 5 to 15 days (Sinclair, 1996; Renwick, 1998), the mid-latitude storm track shifts equatorward, producing stormy conditions in central Chile and dry, cold conditions at the southern tip of the continent (Rutllant and Fuenzalida, 1991). Annalisa Cherchi y Andrea Carril

29 Volumen, complejidad y disponibilidad de datos climáticos de diferentes tipos y provenientes de diferentes fuentes se está incrementando rápidamente: - CMIP5 / IPCC AR5 - Nuevos reanálisis multidecádicos (p.e. 20CR, Compo et al., 2011) - CLARIS LPB (nuevas simulaciones regionales, ) - Nuevas simulaciones “de cambio climático” (p.e. IPSL-CM5A-LR con escenario RCP4.5) - Experimentos idealizados con AGCMs con “alta” resolución (p.e. ECHAM4/CMCC) - Nuevos datos observacionales (p.e. Observatorio Nacional de la Degradación de Tierras y Desertificación) El volumen, complejidad y disponibilidad de datos climáticos de diferentes tipos y provenientes de diferentes fuentes se está incrementando rápidamente (Overpeck et al., 2011). Este fenómeno se aceleró durante el corriente año (2011) y se espera que muy particularmente durante 2012, inducido por la proximidad del próximo informe de IPCC (AR5) y por nuevos proyectos o programas que pondrán a disposición de la comunidad científica nuevas bases de datos observacionales y provenientes de simulaciones con modelos numéricos. Algunos ejemplos relevantes incluyen una nueva fase del proyecto de intercomparación de modelos climáticos acoplados globales (CMIP5, nuevos reanálisis multidecádicos tales como 20CR (Compo et al., 2011), y nuevas simulaciones (cubriendo todo o buena parte del período ) con modelos regionales de alta resolución desarrolladas en el contexto de CORDEX (http://wcrp.ipsl.jussieu.fr/SF_RCD_CORDEX.html), programa en el cual se incluye a Sudamérica como uno de los dominios de integración para los modelos. Además, y de particular relevancia para nuestra región, el proyecto CLARIS LPB (http://www.claris-eu.org/, del que participan todos los integrantes de esta propuesta de investigación PIP 2012) está en su fase de finalización y, por lo tanto, se dispondrá de la base de datos observacional y de modelos que actualmente se está organizando y que incluye un ensamble de simulaciones de cambio climático (con períodos correspondientes al clima actual y futuro) realizadas con modelos climáticos regionales (RCM, según su sigla en inglés). En consecuencia, un desafío para la comunidad climatológica y meteorológica es aprovechar esta oportunidad que se presenta de disponer de un gran volumen de nueva y mejor información acerca de la evolución climática en un período que incluye las últimas décadas y que se proyecta hasta el 2100.

30 GRACIAS

31

32 OTROS SLIDES

33 VARIABILIDAD INTERANUAL DE PRESIONERA 40 Atlas

34

35

36

37 The uncertainty is larger over parts of Bolivia and Brazil where the number of models projecting a wetter climate is similar to the number of models projecting a drier climate. Regions of considerable uncertainty are also found over parts of central and western Argentina, which is a region located between robust drying (Southern Andes) and moistening (Rio de la Plata) areas (different models place the boundaries between drying and moistening regions differently). However, even in the regions where relatively large consensus is reached for the response, the fact that most models are not able to reproduce the regional precipitation patterns in their control experiment with sufficient accuracy contributes to enhancement of the uncertainty. This highlights that the robustness of the large-scale response is only a necessary but not sufficient condition for its trustworthiness.

38 Fig. 4. Schematic of the land water balance (left) and land energy balance (right) for a given surface soil layer. dS/dt refers to the change in water content within the layer (soil moisture, surface water, snow; depending on the depth of the layer, this may include ground water changes), while dH/dt refers to the change of energy within the same layer. SWnet refers to the net shortwave radiation (SWin-SWout) and LWnet refers to the net longwave radiation (LWin-LWout). Note that H2O and CO2 refer to atmospheric water vapour and atmospheric CO2 and their role as greenhouse gases. For simplicity other greenhouse gases are not indicated on the .gure. For explanations of the other abbreviations, see Section 3.

39 CMIP3 ensemble annual mean biases (IPCC AR4)PRECIP.: Mean of the 21 models minus observations (CMAP) Surf.Air TEMP.: Mean of the 21 models minus observations (HadCRUT2v)

40 Inferring supply and demand limitation of ETET is limited by atmospheric demand ET is limited by soil moisture supply Classifying the regions with either T (demand limitation) or P (supply limitation) Inferring supply and demand limitation of ET Distinguishing land-ET response due to atmospheric demand from that due to terrestrial moisture-supply limitation is a classic ecohydrological problem. ET responds to changing atmospheric demand, for example to changing temperature, if there is sufficient moisture supply. In contrast, if the soils are too dry, ET becomes restricted by soil moisture. We infer the primary limitation of ET, atmospheric demand or moisture supply using correlations between annual means of ET and annual means of temperature and precipitation . Since temperature, radiation, and vapour pressure deficit are strongly correlated, temperature can be used as a proxy for atmospheric demand. Supplementary Figure 2. Inferred supply and demand limitation of ET. Panels (a) and (b) show Pearson’s correlation coefficient of ET and temperature, and ET and precipitation respectively. Correlations that are not significant (p>0.1) are displayed in gray. Panel (c) provides a picture of the globe displaying where demand or supply limitation dominates the interannual ET behaviour. It is based on a superimposition of the correlation maps shown in (a) and (b) and classifying the regions according to the larger correlation coefficient with either temperature (demand limitation) or precipitation (supply limitation). Jung et al., 2010

41 Inferring supply and demand limitation of ETET is limited by soil moisture (soils are relatively dry) ET responds to changing atmospheric demand (if there is sufficient moisture supply) Jung et al., 2010

42 ENSAMBLE W ENSAMBLE S-W Correlaciones positivas –aqui en amarillo/rojo - indican que SM condiciona la ET

43 Andes  topografía compleja Fuertes gradientes en parámetros climáticos (p.e. temp. y precip.) GCMs y RCMs: resolución demasiado gruesa para representar detalles topográficos (y otros forzantes relacionados con características del terreno y de la vegetación) La simulación de la precipitación en zonas montañosas es poco confiable Los modos naturales de variabilidad climática (p.e. ENSO, SAM) pueden modular los patrones de precipitación en diferentes escalas temporales (los modelos acoplados globales tienen dificultades para capturar estos mecanismos) Existen estudios que muestran que modelos de alta resolución pueden simular los patrones de mesoescala observados (pero son muy caros en modo climático) Nieve+Hielo  clave para ciclo hidrológico en zonas montañosas Doble problema en los Andes: calentamiento y disminución de la precipitación Consecuencias p.e. sobre: Escurrimiento (runoff), caudal de ríos  timing / volumen Vegetación: puede ser sensible a cambios en el ciclo anual de diferentes factores (T, P, nieve, runoff) Although mountains diff er considerably from one region to another, one common feature is the complexity of their topography. Related characteristics include rapid and systematic changes in climatic parameters, in particular temperature and precipitation, over very short distances (Becker and Bugmann, 1997); greatly enhanced direct runoff and erosion; systematic variation of other climatic (e.g., radiation) and environmental (e.g., soil types) factors. In some mountain regions, it has been shown that temperature trends and anomalies have an elevation dependence (Giorgi et al., 1997), a feature that is not, however, systematically observed in all upland areas (e.g., Vuille and Bradley, 2000, for the Andes). Few model simulations have attempted to directly address issues related specifi cally to future climatic change in mountain regions, primarily because the current spatial resolution of GCMs and even RCMs is generally too crude to adequately represent the topographic detail of most mountain regions and other climate-relevant features such as land cover that are important determinants in modulating climate in the mountains (Beniston et al., 2003). High-resolution RCM simulations (5-km and 1-km grid scales) are used for specifi c investigations of processes such as surface runoff , infi ltration, evaporation and extreme events such as precipitation (Weisman et al., 1997; Walser and Schär, 2004; Kanada et al., 2005; Yasunaga et al., 2006) and damaging wind storms (Goyette et al., 2003), but these simulations are too costly to operate in a ‘climate mode’. Because of the highly complex terrain, empirical and statistical downscaling techniques have often been seen as a very valuable tool to generate climate change information for mountainous regions (e.g., Benestad, 2005; Hanssen-Bauer et al., 2005). Projections of changes in precipitation patterns in mountains are unreliable in most GCMs because the controls of topography on precipitation are not adequately represented. In addition, it is now recognised that the superimposed eff ects of natural modes of climatic variability such as ENSO or the NAO can perturb mean precipitation patterns on time scales ranging from seasons to decades (Beniston and Jungo, 2001). Even though there has been progress in reproducing some of these mechanisms in coupled oceanatmosphere models (Osborn et al., 1999), defi ciencies remain and prevent a good simulation of these large-scale modes of variability (see also Section 8.4). However, several studies indicate that the higher resolution of RCMs and GCMs can represent observed mesoscale patterns of the precipitation climate that are not resolved in coarse-resolution GCMs (Frei et al., 2003; Kanada et al., 2005; Schmidli et al., 2006; Yasunaga et al., 2006). Snow and ice are, for many mountain ranges, a key component of the hydrological cycle, and the seasonal character and amount of runoff is closely linked to cryospheric processes. In temperate mountain regions, the snowpack is often close to its melting point, so that it may respond rapidly to minor changes in temperature. As warming increases in the future, regions where snowfall is the current norm will increasingly experience precipitation in the form of rain (e.g., Leung et al., 2004). For every degree celsius increase in temperature, the snow line will on average rise by about 150 m. Although the snow line is diffi cult to determine in the fi eld, it is established that at lower elevations the snow line is very likely to rise by more than this simple average estimate (e.g., Martin et al., 1994; Vincent, 2002; Gerbaux et al., 2005; see also Section 4.2). Beniston et al. (2003) show that for a 4°C shift in mean winter temperatures in the European Alps, as projected by recent RCM simulations for climatic change in Europe under the A2 emissions scenario, snow duration is likely to be reduced by 50% at altitudes near 2,000 m and by 95% at levels below 1,000 m. Where some models predict an increase in winter precipitation, this increase does not compensate for the eff ect of changing temperature. Similar reductions in snow cover that will aff ect other mountain regions of the world will have a number of implications, in particular for early seasonal runoff (e.g., Beniston, 2003), and the triggering of the annual cycle of mountain vegetation (Cayan et al., 2001; Keller et al., 2005). Because mountains are the source region for over 50% of the globe’s rivers, the impacts of climatic change on mountain hydrology not only aff ect the mountains themselves but also populated lowland regions that depend on mountain water resources for domestic, agricultural, energy and industrial supply. Water resources for populated lowland regions are infl uenced by mountain climates and vegetation; shifts in intra-annual precipitation regimes could lead to critical water amounts resulting in greater fl ood or drought episodes (e.g., Barnett et al., 2005; Graham et al., 2007).

44 Efectos del calentamiento: En algunas regiones: nevadas  lluvias Zonas montañosas templadas  Criósfera próxima a su melting point  Criosfera sensible a pequeños cambios de temperatura Efectos del calentamiento: En algunas regiones: nevadas  lluvias Línea de nieve: a mayor altura (~150 m / 1C) Menor duración de la nieve, especialmente en niveles bajos Although mountains diff er considerably from one region to another, one common feature is the complexity of their topography. Related characteristics include rapid and systematic changes in climatic parameters, in particular temperature and precipitation, over very short distances (Becker and Bugmann, 1997); greatly enhanced direct runoff and erosion; systematic variation of other climatic (e.g., radiation) and environmental (e.g., soil types) factors. In some mountain regions, it has been shown that temperature trends and anomalies have an elevation dependence (Giorgi et al., 1997), a feature that is not, however, systematically observed in all upland areas (e.g., Vuille and Bradley, 2000, for the Andes). Few model simulations have attempted to directly address issues related specifi cally to future climatic change in mountain regions, primarily because the current spatial resolution of GCMs and even RCMs is generally too crude to adequately represent the topographic detail of most mountain regions and other climate-relevant features such as land cover that are important determinants in modulating climate in the mountains (Beniston et al., 2003). High-resolution RCM simulations (5-km and 1-km grid scales) are used for specifi c investigations of processes such as surface runoff , infi ltration, evaporation and extreme events such as precipitation (Weisman et al., 1997; Walser and Schär, 2004; Kanada et al., 2005; Yasunaga et al., 2006) and damaging wind storms (Goyette et al., 2003), but these simulations are too costly to operate in a ‘climate mode’. Because of the highly complex terrain, empirical and statistical downscaling techniques have often been seen as a very valuable tool to generate climate change information for mountainous regions (e.g., Benestad, 2005; Hanssen-Bauer et al., 2005). Projections of changes in precipitation patterns in mountains are unreliable in most GCMs because the controls of topography on precipitation are not adequately represented. In addition, it is now recognised that the superimposed eff ects of natural modes of climatic variability such as ENSO or the NAO can perturb mean precipitation patterns on time scales ranging from seasons to decades (Beniston and Jungo, 2001). Even though there has been progress in reproducing some of these mechanisms in coupled oceanatmosphere models (Osborn et al., 1999), defi ciencies remain and prevent a good simulation of these large-scale modes of variability (see also Section 8.4). However, several studies indicate that the higher resolution of RCMs and GCMs can represent observed mesoscale patterns of the precipitation climate that are not resolved in coarse-resolution GCMs (Frei et al., 2003; Kanada et al., 2005; Schmidli et al., 2006; Yasunaga et al., 2006). Snow and ice are, for many mountain ranges, a key component of the hydrological cycle, and the seasonal character and amount of runoff is closely linked to cryospheric processes. In temperate mountain regions, the snowpack is often close to its melting point, so that it may respond rapidly to minor changes in temperature. As warming increases in the future, regions where snowfall is the current norm will increasingly experience precipitation in the form of rain (e.g., Leung et al., 2004). For every degree celsius increase in temperature, the snow line will on average rise by about 150 m. Although the snow line is diffi cult to determine in the fi eld, it is established that at lower elevations the snow line is very likely to rise by more than this simple average estimate (e.g., Martin et al., 1994; Vincent, 2002; Gerbaux et al., 2005; see also Section 4.2). Beniston et al. (2003) show that for a 4°C shift in mean winter temperatures in the European Alps, as projected by recent RCM simulations for climatic change in Europe under the A2 emissions scenario, snow duration is likely to be reduced by 50% at altitudes near 2,000 m and by 95% at levels below 1,000 m. Where some models predict an increase in winter precipitation, this increase does not compensate for the eff ect of changing temperature. Similar reductions in snow cover that will aff ect other mountain regions of the world will have a number of implications, in particular for early seasonal runoff (e.g., Beniston, 2003), and the triggering of the annual cycle of mountain vegetation (Cayan et al., 2001; Keller et al., 2005). Because mountains are the source region for over 50% of the globe’s rivers, the impacts of climatic change on mountain hydrology not only aff ect the mountains themselves but also populated lowland regions that depend on mountain water resources for domestic, agricultural, energy and industrial supply. Water resources for populated lowland regions are infl uenced by mountain climates and vegetation; shifts in intra-annual precipitation regimes could lead to critical water amounts resulting in greater fl ood or drought episodes (e.g., Barnett et al., 2005; Graham et al., 2007).

45

46