Precipitation dynamics in southern Spain: trends and cycles



The temporal variability of precipitation in southern Spain was analysed to detect trends and cycles. The objectives of this study were (1) to determine whether there has been a decrease in precipitation in southern Spain and (2) to evaluate if any decrease is related to temporal cycles or is a well-defined trend. The study was based on analysis of data series from nine monitoring stations in the Andalusia area of the Spanish Mediterranean. The Pettit test and temporal variability analyses were carried out using the temporal databases (1960–2006) from the selected stations. The results indicate that (1) there were significant differences in precipitation throughout the study period based on the location of the stations; (2) a negative trend in precipitation was observed for the east coast and inland stations, and an increasing trend in precipitation was observed for stations on the west coast; and (3) a general decreasing trend in seasonal precipitation, especially in winter, autumn, and spring, was observed for all stations except those on the west coast. Copyright © 2010 Royal Meteorological Society

1. Introduction

Meteorological series of precipitation and temperature provide useful information about climate behaviour and its repercussions for natural systems. Conclusions from interpretation of such information can be applied to various social and economic spheres to support decision making related to agricultural production, regulation of irrigation, water resources for urban and industrial use, energy planning, tourism and commercial activities, and protection of the environment.

Concern about the future of water resources in the context of climate change has aroused great interest in the study of precipitation trends. In the last 25 years, studies of temporal variability and trends in precipitation have proliferated at both global and regional scales (Díaz et al., 1989; Amanatidis et al., 1993; Hulme, 1995; Brunetti et al., 2001; Norrant and Douguédroit, 2005; López-Bustins et al., 2008a; Mehta and Yang, 2008). Many studies in Spain have analysed temporal variability in precipitation. In the western Mediterranean area, Moreno-García and Martín-Vide (1986) noted a decrease in precipitation in some places, whereas Quereda et al. (2000) reported an increase in precipitation in the Catalonia and Castellón areas, based on analysis of 11 data series. Similarly, in a study of the longest precipitation series for the Balearic Islands, Guijarro (2001) noted varying trends, with an obvious northeast-southeast gradient, despite the relatively small spatial scale of the study area involved. If confirmed, this gradient could indicate changes in the patterns of atmospheric circulation in the western Mediterranean (Wheeler and Martín-Vide, 1992; Galán et al., 1999; González-Hidalgo et al., 2001; Saladié et al., 2002; Esteban-Parra et al., 2003; Rodrigo and Trigo, 2007; De Luis et al., 2009; González-Hidalgo et al., 2009, 2010). Furthermore, changes in precipitation gradients could have important implications for the Mediterranean ecogeomorphological system and for soil-water-plant relationships where degradation and erosion processes are enhanced by increasingly arid conditions (Lavee et al., 1998; Ruiz-Sinoga and Martínez-Murillo, 2009; Ruiz-Sinoga and Romero-Díaz, 2010).

Drought periods at the end of the 20th and the beginning of the 21st centuries (especially the dry periods 1978–1982, 1992–1995, and 2005–2007; García Marín, 2009) and increasing demand for water resources resulting from population growth and economic development have posed research challenges at a time when climate change hypotheses indicate a decline of 20-25% in precipitation for the middle of the 21st century (IPCC, 2007). Indeed, a decrease in winter precipitation (especially during March) has been observed at the beginning of spring and in summer, resulting in a decrease in mean annual precipitation for the Iberian Peninsula (Rodrigo et al., 2000; Aguilar et al., 2006; Rodrigo and Trigo, 2007; Del Río et al., 2010).

At the regional scale, most models agree on the likely consequences of continued atmospheric warming. A predicted initial effect is a reduction in the temperature difference between the poles and the equator, with subtropical high pressure systems and the paths of mid-latitude low-pressure systems extending to higher latitudes. This will cause an increase in winter precipitation in regions between 50° and 70° latitude. However, between 10° and 50° latitude there will be a decrease in water availability as a consequence of increasing evaporation, especially during equinoctial periods (Royer and Mahouf, 1992).

According to the most recent IPCC report, the global temperature could increase 1.5–4.5 °C, largely as a consequence of anthropogenic climatic change and appreciable changes in rainfall patterns. However, in the inland regions of the Iberian Peninsula the temperature may increase by 6 °C by the end of the 21st century (De Castro et al., 2007; Giorgi and Lionello, 2008).

Such a change could have major climatic and biogeographical impacts in the western Mediterranean basin, which is on the meridional border of the temperate zone (35–45°N). Consequently, throughout the 21st century there may be a significant readaptation of the biogeographical system, driven by an increase in evapotranspiration and a decrease in precipitation. According to the current models (including HadAM3H, ARPEGE, REMO, RACMO, and PROMES), this will be caused by a progressive increase in the distance from the zone where polar fronts are generated. For example, the European PRUDENCE and ENSEMBLES models have modelled the future climate of Europe, including the Iberian Peninsula (Hewitt, 2005; Christensen et al., 2007). Integrated within these models, the regional climate model PROMES predicts greater variability in the mean seasonal precipitation than the climate features being the simulated figures of precipitations generally lesser than the real climatic ones for the southern zone of the Iberian Peninsula (De Castro et al., 2005). In a study of the Iberian Peninsula for the period 1960–1990, comparison of simulated data from the regional PROMES model with actual data from rain gauges indicated that this model reliably represents the various climatic patterns of Spain; the degree of certainty is acceptable within the regionalization of climatic change predicted at the global scale by the HadCM3 model (De Castro et al., 2005). In summary, the seasonal rainfall predictions for the middle of the 21st century in the least favourable scenario of the PROMES model are as follows

  • (1)Winter: an increase in precipitation of more than 10 mm in the northeast quadrant of the peninsula, decrease of more than 10 mm in the southern quadrant and the Mediterranean areas of the peninsula, and no appreciable changes in the remaining part of the territory.
  • (2)Spring: a decrease in precipitation of more than 20 mm over most of the Iberian Peninsula and no appreciable change for the islands.
  • (3)Summer: decrease in precipitation of more than 40 mm in the north of Galicia, the Cantabrian coast, the Pyrenees, and the northeast of the Iberian Peninsula, and decreases of 10–40 mm in the remaining part of the territory with the exception of the Canary Islands, where no appreciable change is expected.
  • (4)Autumn: an increase in precipitation of more than 10 mm in the northeast of Spain, a decrease of more than 20 mm in the southwestern half, and no significant change over the remaining part of the territory.

In view of these predictions, it is of interest to establish whether clear precipitation trends already exist in the southern Spanish Mediterranean. This is confirmed by the forecasts issued at regional levels by the various climate change models noted above. The objectives of this study were to (1) analyse the temporal patterns of precipitation during the last half of the 20th century in the southern Spanish Mediterranean region, (2) establish whether there has been a decrease in precipitation, and (3) assess whether any decrease is related to temporal cycles or is a well-defined trend.

2. Study area

At a European scale, the southern Spanish Mediterranean has a Mediterranean climate, but a regionalization involving coastal, inland, and mountainous climates is also evident (Pita, 2003):

The Mediterranean subtropical climate affects most of the Andalusian Mediterranean coast. This climate is characterized by mild winter temperatures because of three basic factors: the moderating influence of the sea, the south facing slope of the coast (which traps solar radiation), and protection against northerly winds by the Betic Mountains. Precipitation is very variable and irregular and generally decreases in a west-east gradient (Figure 1) from the Strait of Gibraltar (>1500 mm/year) to the subdesert on the coast of Almería (<250 mm/year) (Capel, 1987).

Figure 1.

Location of the study areas and average annual rainfall in the southern Spanish Mediterranean basin

The Mediterranean subdesert climate is characteristic of the entire southeastern sector of the Iberian Peninsula (the province of Almería). This area has annual precipitation typically less than 250 mm and less than 150 mm in some areas close to Cabo de Gata. Precipitation occurs in a small number of torrential rainfall periods, which exacerbates erosion and soil degradation processes because there is little vegetation cover, which also contributes to the high level of evapotranspiration that occurs under the high temperature and insolation conditions (Castillo, 1989).

The inland and mountainous climate affects the highest altitude areas of the region, mainly the Sierra Nevada and surrounding high mountains; more precipitation occurs at lower altitudes approaching the Strait of Gibraltar. In general, the effect of altitude is reflected in reduced temperatures and increased rainfall, although these factors are very much determined by the topographic conditions and the surrounding relief (the mean annual precipitation exceeds 1000 mm/year). Both temperature and rainfall distribution patterns remain inalterable, and the summer drought is as pronounced as it is in the remaining part of the region because orographic effects do not overcome the marked subsidence of air resulting from the presence of high subtropical pressure. The combination of summer droughts and extreme winter temperatures makes conditions extremely difficult for plants, animals, and human occupation (Pita, 2003).

2.1. Descriptive analysis of the study monitoring stations

In addition to central figures and dispersion averages for the seasonal and annual series, various patterns were evident (Table I).

  • 1.The stations close to the southern Spanish Mediterranean coast had significantly lower mean precipitation than the other stations.
  • 2.Spring precipitation was more irregular, indicating relative temporal instability, and the variability in spring precipitation was greatest. The variability of equinoctial precipitation, especially in spring, may affect both the vegetation active period and dry farming production, which is regulated by water resources. The crops most affected in the study region are cereals (barley and wheat) and fruit trees (e.g. almonds). The latter do not need watering because final maturation of the fruit is directly related to spring rain.
  • 3.Analysis of average seasonal precipitation showed that among the stations there was a relatively uniform spatial pattern common to the southern Spanish Mediterranean, with most precipitation occurring in autumn, less in winter and spring, and little in summer.
Table I. Average annual rainfall, average number of rainy days, standard deviation, coefficient of variation, and the proportion of the seasonal rain (%) for the monitored stations
Average rainfall65.041.514.673.4194.5
Average rainy days11.68.32.712.935.5
Standard deviation38.630.321.838.075.8
Coefficient variation0.
Proportion of rains33.521.67.537.4100
Average rainfall135.960.020.4145.0361.3
Average rainy days14.38.62.814.139.8
Standard deviation84.255.523.996.0158.7
Coefficient variation0.
Proportion of rains37.616.65.640.1100
Average rainfall164.172.521.8218.7477.1
Average rainy days18.111.12.719.351.2
Standard deviation80.146.232.9116.7164.3
Coefficient variation0.
Proportion of rains35.015.34.745.1100
Average rainfall276.193.832.5320.6723.0
Average rainy days21.511.62.921.357.3
Standard deviation146.656.952.0184.9310.6
Coefficient variation0.
Proportion of rains38.213.04.544.3100
San Roque
Average rainfall319.2109.325.1384.6838.2
Average rainy days22.85.92.423.254.3
Standard deviation170.562.140.8202.8315.6
Coefficient variation0.
Proportion of rains38.
El Higueral
Average rainfall109.581.539.7120.0350.8
Average rainy days12.49.33.412.037.1
Standard deviation63.461.339.288.3131.9
Coefficient variation0.
Proportion of rains31.223.211.334.2100
Sierra Nevada
Average rainfall219.3123.241.2267.7651.4
Average rainy days16.113.14.516.350.0
Standard deviation129.581.141.9137.5218.3
Coefficient variation0.
Proportion of rains34.318.86.340.6100
Average rainfall163.087.333.9201.9486.1
Average rainy days20.513.03.320.457.2
Standard deviation78.446.236.6107.2165.5
Coefficient variation0.
Proportion of rains34.118.16.940.9100
Average rainfall414.3170.441.0501.01126.7
Average rainy days23.213.72.823.763.4
Standard deviation211.198.852.5267.9418.0
Coefficient variation0.
Proportion of rains36.815.13.644.5100

A summer water deficit affects all parts of the region almost uniformly, and summer precipitation rarely reaches 8% of the annual figure (Table I). This temporal pattern is a consequence of frequent displacement of the jet stream towards the latitude of the study area and that substantial activity results from the associated low-pressure systems from October to December. Most spring precipitation originates from depressions that integrate strong surface convection with marked instability at high altitudes. The low pressures are particularly well developed in the study region because of the presence of intensively heated soil surfaces and the mountainous relief, which enhance the ascent of air. A feature common to the entire region is that precipitation occurs on a limited number of days each year (usually < 25%), and less than 10-15% of days in the southeast (Table I).

3. Database and methods

Meteorological stations in Spain have recorded precipitation from the middle of the 20th century. However, despite the existence of numerous stations for monitoring precipitation behaviour and evolution, the data quality is questionable at times, and serious gaps can occur in temporal series, preventing comparisons with other series.

For this study, we used the daily temporal series from nine rain-gauge stations, five located along the southern Mediterranean coast and another four inland (Table II). These were selected as representative of the precipitation in the surrounding areas, according to Ramos-Calzado et al. (2008). The data were supplied by the Andalusian Water Agency (Government of Andalusia).

Table II. Basic characteristics of the rain-gauge stations
NameCo-ordinates UTMAltitude a.s.l. (m)Number of yearsInitial year–final year
San Roque4009846-285647109461962–2007
El Higueral4131968-544293886461962–2007
Sierra Nevada4091190-4836811319461962–2007

The homogeneity of the temporal series was tested using the Anclim software (Štepánek, 2001), according to the guidelines of the Andalusian Agency of Water. The standard normal homogeneity test (Alexandersson and Moberg, 1997) and the EP test (Easterling and Peterson, 1995) were applied to the data. No reference series were used because of a lack of objective criteria to ensure that they were homogeneous over the experimental period. The Pettit test was also applied; this nonparametric test, developed by Wijngaard et al. (2003), is based on the series of interval {ri:i = 1, …, n}, which is defined as the data position in the series sorted from the lowest to the highest value. Data of the same value are assigned to the same interval, corresponding to the arithmetic average of the intervals for those elements. The series homogeneity is considered for the null hypothesis and is calculated using Equation (1):

equation image(1)

In the case of a discontinuity in the K-esima position, the final value presents an extreme one near that position. The significance level and the eigenvalues were tabulated and depending on the distribution size, the next value is calculated with Equation (2):

equation image(2)

The temporal variability of precipitation was also analysed. The polygonal line should be replaced by another extendable one which correctly fit to establish or deduce an evolutionary prognosis of precipitation for the years following the last year in the temporal series, i.e. after 2007. Therefore, it is advisable to substitute fictitious points that are in a recognized distribution line (e.g. a straight line or parabola) for the actual precipitation data. The residual distance to the respective actual figures should be as close as possible. In this study, the precipitation trend was calculated using the linear regression trend method and moving averages within a 5-year range. A new series was created that enabled detection of a secular trend in the series studied (Walford, 1995; Brightman, 1999). The nonparametric Mann–Kendall (Z-test; 95% reliability) was used to evaluate the statistical significance of the trends obtained.

4. Results

4.1. Evolution and trends of annual rainfall

The Pettit test revealed optimum homogeneity in the database for each of the stations. Table III shows the annual trend and its effect. Significant differences were associated with the locations of the stations.

  • 1.For east coast stations (Almería and Motril), there was a negative trend in rainfall, although it was not very significant. The Málaga station, located centrally in the southern Spanish Mediterranean, could also be included in this group.
  • 2.For west coast stations (Marbella and San Roque), there were highly significant positive trends, with increases in annual precipitation of 27 and 22 mm/year, respectively.
  • 3.For inland stations (Sierra Nevada, Antequera, and Gaucín), there were significant negative trends, with a large reduction in precipitation.
Table III. Trends in annual rainfall based on the Mann–Kendall test
ObservatoriesPeriodnTest ZAnnual rainfall variation (mm/year)
  • Level of significance:

  • *

    p < 0.05;

  • **

    p < 0.01.

Almería1962–200746− 0.89− 0.55
Motril1965–200743− 0.08− 0.21
Málaga1962–200746− 0.59− 0.57
San Roque1962–2007460.362.22
El Higueral1962–200746− 0.82− 1.52
Sierra Nevada1962–200746− 1.97*− 3.33
Antequera1962–200746− 2.12*− 3.12
Gaucín1962–200746− 1.67**− 4.67

The precipitation trends were less evident for the southeastern Mediterranean stations (Almería, Motril, and El Higueral) relative to the those in the west (Marbella, San Roque, and Gaucín), independent of their location (inland or coastal). The significance of these trends was assessed using the Mann–Kendall test, which indicated that only the inland stations (Sierra Nevada, Antequera, and Gaucín) had significant negative trends (Table III). The only exception was the El Higueral observatory, where the negative trend was less significant.

Figure 2 shows the annual precipitation, the trend line, and the moving 5-year precipitation average for softening the great annual variability and detecting probable wet and dry cycles during the study period. One station from each of the study areas is provided as an example in Figure 2.

Figure 2.

Annual rainfall, temporal evolution (moving 5-year average), and trends in precipitation at study stations: (A) Almería, (B) San Roque, and (3) Sierra Nevada

Although substantial interannual variability is the most obvious feature, remarkable differences were found in the number and duration of wet and dry cycles and were related to geographical location. The Almería station (east coast) had fewer wet cycles (four) than the rest. The dry cycles were longer, but the wet cycles were more intense: the dry cycles lasted for up to 10 years (1978–1987), but the wet cycles lasted for approximately 4 years.

Towards the western coastal stations (Marbella and San Roque), the wet cycles tended to last as long as the dry cycles, which were less intense (i.e. a positive pluviometric trend). The results for the inland stations (El Higueral, Antequera, Gaucín, and Sierra Nevada) were much more significant. The number of wet cycles was similar to that of the southwest Mediterranean coast, but because the amount of precipitation was less there was a decreasing trend in annual precipitation (Figure 2).

Following the analysis of the annual precipitation trend we assessed the seasonal evolution of precipitation, as this is readily understood by society (García-Marín and Calvo, 2008). With respect to cumulative precipitation, autumn and spring are the two most important environmental and agronomic periods for dry farming, provided that rainfall is distributed throughout each season. A negative trend in precipitation and prolonged dry periods during spring and autumn, which are key periods for vegetation development, could increase the physical perception of the phenomenon. This analysis also contributes to the clarification of the temporal evolution of annual precipitation in southern Spain during the second half of the last century.

4.2. Evolution and trends of seasonal rainfall

Analysis of the seasonal precipitation trends and their possible causes facilitates understanding of the temporal variability patterns that occurred in southern Spain during the second half of the 20th century. Table IV and Figure 3 show the seasonal trends, their statistical significance, and the distribution of precipitation.

Figure 3.

(A) Evolution of winter rainfall at the Sierra Nevada and Málaga stations. Winter rainfall in the southern Spanish Mediterranean is declining. (B) The trends in spring rainfall are diverse. In the Mediterranean southeast (Almería station), the trends are negative, whereas in the western zone (Marbella station), the trends are positive or increasing. (C) In autumn, the trends are decreasing for inland rainfall (Antequera station) and the Mediterranean southeast. The southwest stations (Marbella) show positive trends. This figure is available in colour online at

Table IV. Trends in seasonal rainfall based on the Mann–Kendall test
ObservatoriesPeriodTest ZSeasonal rainfall variation (mm/year)
  • Level of significance:

  • *

    p < 0.05;

  • **

    p < 0.01.

Almería1962–2007− 0.99− 1.500.21− 0.30− 0.34− 0.400.30− 0.11
Motril1965–2007− 1.17−− 1.72− 0.440.281.67
Málaga1962–2007− 1.74**0.411.86**− 0.03− 0.700.170.15− 0.19
Marbella1962–2007− 0.841.481.95**1.95**− 1.220.670.722.56
San Roque1962–2007− 0.27− 0.262.16*1.14− 0.79− 0.230.452.79
El Higueral1962–2007− 0.61− 0.820.02− 0.79− 0.54− 0.220.22− 0.98
Sierra Nevada1962–2007− 1.85**− 1.001.45− 0.68− 2.02− 0.700.47− 1.08
Antequera1962–2007− 2.76**− 0.570.96− 1.48− 1.51− 0.230.11− 1.49
Gaucín1962–2007− 2.84**0.801.78**− 0.53− 5.830.960.59− 0.39

In winter, there was a generalized trend of decreasing precipitation, with highly significant negative trends for the inland stations (Sierra Nevada, Antequera, and Gaucín) and a significant trend for the Málaga station on the coast (Table IV).

During the equinox periods, the trends were variable, but with some similarities depending on the study area. Two patterns of spring precipitation were observed: (1) a nonsignificant decrease for the more easterly stations (coastal and inland) and (2) an increase for the southwest stations (with the exception of the San Roque station), although these trends were not very strong.

During autumn in the study region, when a change in precipitation can have marked impacts on natural ecosystems and crops following the long dry summer, three patterns were observed: (1) a negative trend in precipitation for inland stations (El Higueral, Sierra Nevada, Antequera, and Gaucín) and the most easterly station (Almería); (2) a positive, but less significant trends for the stations located on the central-eastern coast (Motril and Málaga); and (3) a positive, but much more significant trends for the western stations (Marbella and San Roque).

In summer, the trend was positive for all stations, which can be explained by the greater incidence of convective precipitation, although showers and storms were generally weak and infrequent. According to the Mann–Kendall test, the trends were statistically significant for the southwestern observatories.

5. Discussion

According to Sumner et al. (2003), in the southern Spanish Mediterranean there has been a marked reduction in moisture coming eastwards from the Atlantic Ocean, from where most of the rain in winter and spring is usually derived. Consequently, total annual precipitation will decline. However, according to the cited authors, the number of days with fog and reduced precipitation will increase in the western coastal areas.

Quereda et al. (2000) noted that the temporal evolution of precipitation is strongly correlated to solar activity: wet cycles develop during major sunspot periods and vice versa. However, this correlation is not perfect. For example, solar activity was marked in 1980 but precipitation was minimal, although this may have been a consequence of the effects of the Mount Saint Helens (1980) and Chinchón Volcano (1982) volcanic eruptions, the aerosols of which could block solar radiation with consequences in its greater precipitation potential (Quereda et al., 2000).

Positive correlations between solar radiation and wet/dry cycles were also evident for the high rainfall years 1989–1990 and during an intense drought that occurred in the period 1994–1995. From 1995 to 2004, a new period of maximum solar activity was correlated with a wet cycle, whereas from 2004 to 2007, a period of drought corresponded to a decrease in the number of sunspots.

On the southwestern Mediterranean coast, a negative correlation between the NAO index and precipitation from October to March was reported by Martín-Vide and Fernández (2001). The correlation was lower during April and May. In contrast, the summer months had no NAO. September is a transition between summer and the clearly significant period of NAO sign.

In the positive phase of the NAO index, high subtropical pressure is accentuated, which causes a change in the path of the winter low-pressure systems towards the southwest Spanish coast, reducing precipitation. The NAO index has shown a predominantly positive trend since the 1970s, resulting in reduced winter precipitation.

Other studies have indicated similar precipitation trends in the Iberian Peninsula. According to López-Bustins et al. (2008b), there has been trend of decreasing precipitation in the western half of the southern Spanish Mediterranean, especially in winter. As well as analysing precipitation in relation to the NAO index, they noted that the decrease in precipitation is related to a synoptic situation typical of the mid-winter period, with a strong anticyclone centred in central Europe (which causes atmospheric stability over the central and western Iberian Peninsula) and northeast circulation in the eastern Mediterranean. The frequency of this pattern increased by 2.71 days/decade in the mid-winter months throughout the period 1959–2000. This resulted in an increase in atmospheric pressure in the central European plain during winter, which is an outcome corroborated by various studies (Stefanicki et al., 1998; Maugeri et al., 2004; Trigo, 2006). This pattern of atmospheric circulation also typically corresponds to a positive NAO phase, which explains the reduced precipitation.

Sumner et al. (2003) related the observed increase in precipitation during summer to an increase in storms at the end of summer, associated with an accentuated impact of humid air masses coming from the Mediterranean Sea.

Comparison of the results of the aforementioned regional climate models with the trend analysis revealed several similarities and some notable differences. An important similarity was in forecasts of the evolution of winter precipitation, with the forecast reduction in winter precipitation from this study being greater than the forecast by the regional models (a reduction of more than 10 mm for the last third of the 21st century). Spring trends in precipitation varied according to the location of the stations (Table IV). The trend of declining precipitation found for the most easterly stations coincided with the findings of the models, but the forecasts for the western stations (Marbella and Gaucín) did not. The summer precipitation trends found in this study are completely opposite to those predicted by the models, but for autumn precipitation the forecasts of the models coincide with the trends from inland stations in this study (El Higueral, Sierra Nevada, Antequera, and Gaucín).

According to Guijarro (2002), the increased spatial-temporal variability in seasonal precipitation and the low statistical significances occasionally found (as a result of the high variability in rainfall) can lead to underestimates of the importance of these precipitation trends. Nevertheless, if they continue in the areas studied there will be major changes in the distribution of water resources, which are critical for economic activities (tourism and agriculture, in particular).

6. Conclusions

Concerns about the scarcity of water resources and its implications in the southern Spanish Mediterranean highlight the urgent need for critical analysis of the temporal evolution of precipitation in the area, with the aim of establishing whether a progressive reduction is occurring. The results show marked contrasts in the southern Spanish Mediterranean related to the geographical location of the stations in the study:

  • 1.In the southeastern coast a decrease in precipitation was observed (Almería, Motril, and Málaga stations), but the trend was not statistically significant.
  • 2.In the southwestern coast (Marbella and San Roque stations) a trend of increasing precipitation was observed. This is in contrast to the reported decrease in precipitation over the southern Spanish Mediterranean and the trends from current climatic change models. These differences may reflect complex cyclical variations in Mediterranean precipitation.
  • 3.In the inland area of Andalusia (El Higueral, Sierra Nevada, Antequera, and Gaucín stations) statistically significant trends of decreasing precipitation were evident, with the trend in winter and autumn dominating the decrease. In this area, there was a strong negative correlation between the NAO index and precipitation from October to May. In the last 30 years, the NAO index has been predominantly positive, which may explain this reduction in rainfall.

The contrasting results could have considerable implications for society and the economy, and the decrease in inland precipitation will cause greater problems than similar changes on the coast. This is because the reservoirs controlling and storing water resources are primarily located in the inland part of the study region, and a decrease in precipitation in this area could reduce water reserves. These reservoirs supply most of the population, which is primarily located along the coast and also the irrigation area of the region. Studies of trends in precipitation and its evolution provide useful tools for managers and administrators of water resources, especially in areas where periods of drought can result in major economic losses, and social and political conflict.