Journal of Arid Environments 75 (2011) 1302e1309 Contents lists available at ScienceDirect Journal of Arid Environments journal homepage: www.elsevier.com/locate/jaridenv Climatic change and rainfall patterns: Effects on semi-arid plant communities of the Iberian Southeast J.D. Miranda*, C. Armas, F.M. Padilla, F.I. Pugnaire Estación Experimental de Zonas Áridas, Consejo Superior de Investigaciones Científicas, Carretera de Sacramento s/n, La Cañada de San Urbano, E-04120 Almería, Spain1 article info abstract Article history: The structure and functioning of semi-arid ecosystems are strongly influenced by precipitation patterns. Received 16 September 2010 Water availability in such environments is highly pulsed, and discrete rainfall events interspersed with Received in revised form drought periods are important components of the annual water supply. Plant communities do not only 2 January 2011 respond to rainfall quantity, but also to variations in time, so that relatively small changes in rainfall Accepted 11 April 2011 frequency (i.e., pulsed inputs) may have strong effects on communities. Within the Mediterranean basin, Available online 25 May 2011 climate change models forecast a decrease in mean annual precipitation and more extreme events (i.e., less rainy days and longer drought periods between events), along with seasonal changes. However, little Keywords: Drought is known on the consequences of these future precipitation changes on plant communities, especially in Mediterranean ecosystems semi-arid environments. Here, we summarize the few experiments that have manipulated rainfall Precipitation patterns in arid and semi-arid areas worldwide, and introduce the first results of a pioneer, long-term Rainfall manipulations rainfall exclusion in the semi-arid southeast region of the Iberian Peninsula. The experiment is not Resilience only manipulating the amount of rainfall, but also its frequency and seasonal distribution in a grassland- Vegetation response shrubland in the Tabernas desert (Almería, Spain). This work monitored the effect of precipitation changes on different ecosystem processes for five years, at the species and community level, concluding that this plant community (as other communities studied in the same area) exhibited great resilience to changes in rainfall availability, likely caused by plant adaptation to large intra- and inter-annual precipitation variability. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction important consequences on processes such as nutrient cycling, plant growth, or population and community dynamics (Weltzin Changes in climate and land-use induced by human activities et al., 2003). affect productivity, water regime and biodiversity, and result in Available evidence shows that the structure and function of changes in the structure and functioning of terrestrial ecosystems terrestrial ecosystems are vulnerable to hydrological modifications (Fay et al., 2000; Rustad and Norby, 2002). These alterations would even in the absence of changes in mean annual precipitation (IPCC, probably intensify in the near future (Sala et al., 1999). There is an 2007; Knapp et al., 2002; Mearns et al., 1997). Therefore, the extensive literature analyzing the potential effects of increased analysis of climate change should not rely exclusively on scenarios greenhouse gas concentrations in the atmosphere on ecosystem based on climatic averages (Katz and Brown, 1992) but incorporate function and services. However, an almost unexplored but impor- seasonal changes in amount and frequency of rain events (Mearns tant characteristic of global change is the alteration of rainfall et al., 1997) in order to accurately predict ecosystem responses to patterns, with probable increases in rainfall variability (IPCC, 2007). future precipitation regimes. General Circulation Models predict higher frequencies of extreme Mediterranean ecosystems are especially sensitive to climate precipitation events, fewer rainy days, and longer drought periods change because they have large topographic complexity with (IPCC, 2007; Rind et al., 1989). These alterations could have disparate land uses as well as marked water availability gradients (Lavorel et al., 1998). Mediterranean shrubland ecosystems are quite vulnerable to rising temperatures and reduced water avail- * Corresponding author. Present address: Estación Experimental del Zaidín, ability (Usodomenech et al., 1995) and the expected increase in Consejo Superior de Investigaciones Científicas, C/Profesor Albareda 1, E-18008 aridity and temperatures in regions already dry, such as the þ þ Granada, Spain. Tel.: 34 958 181 600; fax: 34 958 129 600. Mediterranean basin, will likely have serious consequences (IPCC, E-mail addresses: [email protected], [email protected] (J.D. Miranda). 2007). Rainfall predictions, however, are extremely complex at 1 Tel.: þ34 950 281 045; fax: þ34 950 277 100. the regional level. Unlike the increase in CO2, which occurs in 0140-1963/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaridenv.2011.04.022 J.D. Miranda et al. / Journal of Arid Environments 75 (2011) 1302e1309 1303 a geographically uniform manner, climate change is strongly provide control over the timing of drought but more elaborate dependent on local conditions (IPCC, 2007). shelters can include irrigation systems able to control the daily, In the Mediterranean basin rainfall has declined by 20% on weekly, and seasonal timing and the extent of wet and dry periods, average during the 20th century, although not uniformly in all independently of natural rainfall patterns (Arkin et al., 1976; Miller locations (Esteban-Parra et al., 1998; Piñol et al., 1998). In the SE of et al., 1991; Ries and Zachmeier, 1985; Svejcar et al., 1999). Finally, the Iberian Peninsula it is expected that 21st century conditions the use of rainout shelters also allows for the selection of a partic- will be significantly drier than the period 1961e1990 (Esteban- ular experimental design and location, with the potential of long- Parra et al., 1998). Several studies predict an increase in winter term manipulations (Fay et al., 2000). precipitation primarily in the form of high-intensity events (Frei Fay et al. (2000) and, recently, Miranda and Pugnaire (2009) et al., 1998) which, together with decreased spring and autumn reviewed the different designs of fixed and mobile rainout shel- rainfalls, can have important consequences for ecosystem function. ters. Although still few, some of these shelters have been set up in For instance, it may speed up desertification processes (Rodrigo, dry and arid environments. We have included a short description 2002) and reduce ecosystem productivity. Seasonal changes in and the main results of the most important long-term experiments rainfall can affect population dynamics of certain species that set up in these environments on Table 1. One of the earliest designs might become locally extinct (Pugnaire et al., 2004). It can also have was developed by Yahdjian and Sala (2002) in a semi-arid shrub deep impact on planteplant and planteanimal interactions leading steppe in Patagonia, Argentina. Using a fixed, partial shelter made to changes in the structure and composition of whole communities of polycarbonate bands folded in V-form, these shelters reduced (Suttle et al., 2007). For example, seed germination in autumn could the amount of rain by 30, 55 and 80% on small plots (200 Â 188 cm). be dramatically reduced so that recruitment of some species will be These authors also compared plot responses to artificial watering threatened and may disappear in the mid term (Miranda et al., that resembled dry and wet years, to find that recovery after 2009a). Overall, changes may imply an alteration of the regenera- watering was lower in plots that had experienced drought the tion niche affecting population dynamics; they may also imply previous year, which showed a link between time to recover and a loss of species richness as only species with broad environmental drought intensity (Yahdjian and Sala, 2006). tolerance will be able to respond to new ecological situations In the tallgrass prairie of Kansas (USA), a rainfall manipulation (Cramer et al., 2001; Jump and Peñuelas, 2005). experiment that altered both mean rainfall and seasonal distribu- Most climate change studies have been conducted in temperate tion of precipitation has been running since 1998 (Knapp et al., ecosystems in the northern hemisphere and the Arctic, but 2002) in one the first experiments that manipulated precipitation responses may be very different in warmer and drier environments frequency (Table 1). Results to date have shown that changes in the (Llorens et al., 2003). To our knowledge, very few experiments have temporal distribution of rainfall had a negative effect on plant addressed changing seasonal rainfall patterns, and none in the productivity, photosynthetic rates, and soil CO2 flux whereas Mediterranean basin. Thus, our purpose in this manuscript is: 1) to reducing precipitation by 30% had smaller effects (Fay et al., 2000, review the most important long-term experiments that have 2002, 2003). manipulated rainfall patterns in arid and semi-arid environments In an arid sagebush steppe in Oregon (USA), 4 treatments or worldwide; and, 2) to enshrine the only experiment set in a semi- distributions (control, natural distribution, winter and spring arid Mediterranean environment, where the amount of precipita- distribution) were compared (Svejcar et al., 1999).