The effects of soil moisture variability on the vegetation pattern in Mediterranean abandoned fields (Southern Spain)

Catena 85 (2011) 1–11

Contents lists available at ScienceDirect

Catena j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c a t e n a

The effects of soil moisture variability on the vegetation pattern in Mediterranean abandoned fields (Southern Spain) J.D. Ruiz-Sinoga ⁎, J.F. Martínez-Murillo, M.A. Gabarrón-Galeote, R. García-Marín Department of Geography, University of Málaga, Campus of Teatinos s/n. 29071, Málaga, Spain

a r t i c l e

i n f o

Article history: Received 26 September 2009 Received in revised form 2 October 2010 Accepted 26 November 2010 Keywords: Abandoned field Soil moisture Vegetation pattern Resilience Hydrological connectivity Mediterranean

a b s t r a c t The demographic pressure decrease in Southern Spanish Mediterranean mountainous areas in the mid twentieth century led to the abandonment of agriculture and rupture in the geo-ecosystem balance which had existed until then. Since then, different phases of recovery have been put into action to return the landscape to its earlier natural condition according to climate and soil degradation state after the abandon. In Mediterranean climatic conditions (between subhumid and semi-arid regimes), degraded soil recovery has followed different tendencies rendering the landscape in heterogenic and complex one. This heterogeneity has manifested in the vegetation pattern of abandoned fields. In this paper, we analyze the state of three abandoned fields situated under different Mediterranean climatic conditions from the recovery point of view by means of monitoring the effects of spatial and temporal variability of soil moisture in the vegetation pattern over a period of two years sampling (Nov. 2002–Nov. 2004). The results showed that: i) more annual rainfall volume did not guarantee success in the biological recovery of the system due to the influence of other factors such as the degradation state of the soil post-abandonment or the steep slope gradient; ii) soil moisture variability tended to play a more important role in affecting vegetal cover in semi-arid conditions; and iii) in dry climatic conditions the system demonstrated greater signs of recovery (greater biodiversity). © 2010 Elsevier B.V. All rights reserved.

1. Introduction Some Mediterranean mountainous areas close to the coast are characterized by intensive human activity which has resulted in the destruction of large areas of natural vegetal cover and the cultivation of land not potentially for farming which had led to soil degradation processes (López-Bermúdez, 1993). Kosmas et al. (2000) pointed out that since the end of the 19th century land abandonment took place in marginal areas because of the agricultural crisis which constituted a real change in soil use at larger scale. These abandoned fields have experienced a process of vegetal re-colonization with different degrees of success, depending on climate and soil degradation. Where the resilience of the geo-ecosystem was low, recovery was weak or nil and processes of erosion appeared. In the rest of the Mediterranean areas, due to the fact that there was no continuity in the geo-ecosystem, a mosaic-like landscape of abandoned fields developed (Bergkamp, 1998) where three types of areas could coexist: i) those which had barely recovered after degradation because of agricultural abandonment, with low vegetal cover and likely intense desertification processes (water erosion processes and soil degradation); ii) those which recovered with a certain degree of success, with an increase in vegetal cover due to the wetter climatic

⁎ Corresponding author. Tel.: +34 952 131710; fax: +34 952 131700. E-mail address: [email protected] (J.D. Ruiz-Sinoga). 0341-8162/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.catena.2010.11.004

conditions and, in many cases, what could be defined as quasi-natural areas given that after more than fifty years of abandonment they look like natural areas rather than abandoned fields; and iii) those that are in an intermediate state. Both abandoned fields and Mediterranean semi-arid rangeland show a mosaic vegetation pattern in two phases (Arnau-Rosalén et al., 2008; Lesschen et al., 2009): vegetated and bare soil areas. This type of spatial pattern directly affects to water erosion by defining sink and source runoff areas during rainfall events (Bergkamp, 1998; Calvo et al., 2003; Cammeraat and Imeson, 1999; Martínez-Murillo and Ruiz-Sinoga, 2007) causing hydrological connectivity processes between the different parts of the hillslope and the rest of the drainage system of the catchment (Puigdefábregas et al., 1999). The vegetation pattern is modified according to the slope gradient and rainfalls, with the vegetal cover percentage diminishing as the climate becomes more arid and the abiotic factors exerting greater control than biotic ones in the runoff generation processes (Katra et al., 2007; Lavee et al., 1998; Rietkerk et al., 2002; Shnerb et al., 2003). When soil degradation conditions become apparent, even in very rainy areas, abiotic factors may control the hydrodynamics of the hillslope where water erosion may be extreme (Boix et al., 1995). In these circumstances, the influence of hydrodynamics on the vegetation pattern is even greater if we take temporal variability into account, because vegetation depends almost exclusively on rainfall from October to May in a Mediterranean climate. This dependency on soil moisture is even more evident on those hillslopes with a geological metamorphic and impermeable

2

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

substratum (e.g. phyllites and schists) with scarce infiltration capacity where the island fertility theory has been clearly observed in semi-arid environments (Puigdefábregas et al., 1999). This study was carried out in three field sites which had been abandoned fifty years ago at least with metamorphic soil and located in different Mediterranean climatic conditions with the aim of establishing the effects of soil moisture variability on the vegetation pattern. The objectives were next: i) to characterize the relationship between vegetation pattern and soil moisture from the point of view of spatial and temporal variability; ii) to determine the hydrological and erosive consequences; and iii) to establish the level of resilience of the three sites and their proximity to their natural previous conditions with respect to their climate regimes. 2. Study area The field sites have been selected according to the climatic gradient approach in which topography, geology and land use must be similar in order to compare them (Imeson and Lavee, 1998). They were located in the Cordillera Bética Litoral (Southern Spain) where a climatic gradient was registered (Fig. 1). Table 1 shows the geo-ecological features of the abandoned fields. All of them were characterized by a similar slope gradient and exposure, lithology, vegetal cover and land use but with different climatic conditions: Colmenar (CO) subhumid, Almogía (AL) dry-Mediterranean and Berja (BE) semi-arid. All of them were cultivated and abandoned at least 50 years ago, when shrub recolonization began with varying degrees of success. CO was located in the Mountains of Málaga and characterised by steep hillslopes with a Palaeozoic metamorphic geological substratum (Malaguide Complex with slates, phyllites and schists) where water erosion predominated. The 80-m-long south facing hillslope had a

convex topographic profile and 51% maximum slope gradient. According to FAO (2006), associations of Regosols and haplic Leptosols, including chromic Cambisol near the top appeared on the hillslope. Soil depth did not usually exceed 25 cm especially at the bottom (where there were rock outcrops). Rock fragment cover of soil was above 50% and the soil texture was sandy clay loam. AL was also located in the Mountains of Málaga and its surroundings shared the same geographical features as CO. The 78-m-long south-facing hillslope was characterised by a rectilinear–convex topographic profile with a maximum slope gradient equal to 32%. Soil characteristics were similar to those in CO in terms of type, gravel content and texture. BE was situated in the Sierra of Contraviesa. The 48-m-long south-facing BE hillslope had a rectilinear–concave topographic profile, a maximum slope gradient of 30%, Palaeozoic metamorphic geological substratum (Alpujarride Complex, mica-schists). Some erosion morphologies of concentrated overland flow were observed (rills). Soils were very limited because of erosion by water: haplic Leptosol with Regosol at the bottom of the hillslope, loam–sandy texture, abundant gravel and thickness not exceeding 20 cm. In general, soil properties were not so degraded as it could be expected when climatic conditions became more arid. Only this tendency was observed in the organic carbon content and aggregate stability which diminished from CO to BE. 3. Methods 3.1. Analysis of vegetation pattern After selecting the three field sites, we defined three open plots located at the top, middle and bottom section of each hillslope with a dimension of 5 × 5 m. The plot vegetation cover was described

Fig. 1. Location and general view of the field sites.

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

3

Table 1 Field site eco-geomorphological features. Field site

Colmenar

Almogia

Berja

Climate Rainfall (mm year−1) Slope gradient (%) Slope length (m) Geology Vegetal cover (%) Soil bulk density (g cm−1) Gravel content (%) Sand (%) Silt (%) Clay (%) Soil texture Organic matter (%) SOC (tn ha−1) Aggregate stability (%) Ksat (cm h−1)

Sub-humid Mediterranean 668.1 38.5 102 Phyllites 49.7 1.37 56.0 51.5 24.8 23.7 Sandy clay loam 4.5 17.6 73.2 10.9

Dry-Mediterranean 513.4 28.6 75 Phyllites 47.8 1.18 60.0 26.6 35.5 23.7 Loam 5.0 16.1 75.3 10.9

Semiarid Mediterranean 335.0 21.0 65 Mica-schists 50.7 1.35 50.0 52.3 32.9 13.6 Sandy loam 2.4 8.9 62.0 7.6

SOC: soil organic carbon; Ksat: saturated hydraulic conductivity.

according to the degree of surface coverage by the aerial system of the vegetation (Lacoste and Salanon, 1978). Vegetation pattern was mapped following a similar methodology to that applied by RuizSinoga et al. (2003), Calvo-Cases et al. (2005), Arnau-Rosalén et al. (2008) and Ruiz-Sinoga and Martínez-Murillo (2009). We placed a digital camera at the top of a 5-m-tall metal tower and took aerial photos for each plot. Photos covered an area of 2.5 × 3.0 m and hence we had to take several to cover the plots completely. The final vegetation pattern maps only showed the permanent vegetation cover (shrub in black colour) because we were only interested in the spatial and temporal changes of this type of vegetation. We mapped the plot's vegetation pattern every six months with a total of five maps for every hillslope section: Nov-2002, May-2003, Nov-2003, May-2004 and Nov-2004 (at the end of dry season with the minimum vegetal cover and at the end of wet season with the maximum). The photos were geo-referenced and interpreted using ArcGIS 9.2. The ratio vegetated area/non-vegetated area was calculated from the vegetal cover data. Vegetated area referred to those areas of hillslopes covered by permanent vegetation (trees and shrubs) and nonvegetated surface to the sum of the occupied area by bare soil, rock fragment cover and rock outcrops. 3.2. Measurement of soil moisture We measured the soil moisture (SM) of the top 15 cm of soil every two weeks for two years (Oct-2002 to Nov-2004) using a TDRTektronix 1502C with a probe inserted vertically into the soil surface. SM was always measured in the same place: three sample points located in the open plots of each field site. In total, 84 visits to each field site were made to measure SM. The water content was then calculated by means of the Time Domain Reflectrometry (TDR) method based on the equation of Topp et al. (1980). In order to analyze the temporal variability of SM, we employed three statistical techniques: standard deviation (Sdt), variation coefficient (VC) and the parametric test of relative differencing (δij) proposed by Vachaud et al. (1985). This test allows plotting the data with a view to highlighting the differences between sampling locations in terms of constancy in temporal stability. The relative difference (δij) is calculated from: δij =

Δij Sj

ð1Þ

where Δij = Sij −Sij

ð2Þ

and

Sij =

1 N ∑ S : N i = 1 ij

ð3Þ

Sji being the soil moisture content (cm3 cm−3) at sample point i on day j, and N the sampling locations. Thus, the mean relative difference for each location is defined as δj =

1 m ∑ δ m j = 1 ij

ð4Þ

where m is the number of sampling days. 3.3. Analyses of soil properties Disturbed and undisturbed soil samples were collected in the three field sites to analyze the next soil properties. Gravel content was measured by sieved. Texture was assessed by sieving using the method of Robinson (1922). BD was determined by the core method (Blake and Hartge, 1986). Aggregate stability was determined by analysing the stability of 4 mm aggregates with wet sieving (Kemper and Rosenau, 1986). The undisturbed samples in 100 cm3 cylinders enabled the measurement of Ksat using a constant-load permeameter and Darcy's Law (Klute, 1965). OM was analyzed using the AFNOR method of carbon measurement (AFNOR, Association française de normalisation, 1987). Organic carbon (kg ha− 1) stored in the first 10 cm of soil (SOC) was calculated for each field site as follows (5): OC = ð%SOC = 100Þ × soil mass

ð5Þ

where soil mass = depth (m) × BD (Mg m− 3) × 10,000 m2 ha− 1 × 1000 kg Mg− 1. 3.4. Statistical analysis The differences in the measured vegetation cover and SM variables were quantified based on the analysis of variance (ANOVA). Differences or patterns have been regarded as significant when the null hypotheses (no difference, no pattern) were rejected because probability (p) was lower than 0.05; p-valueb 0.01 was also used in some cases. F-test was used to determine if there was an F-distribution or a continuous probability distribution and the null hypothesis was found to be true. Relationships between variables were quantified using Pearson's correlation coefficient. Also statistic linear relationships were performed to model the variability of vegetation cover based on SM.

4

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

These procedures were carried out with the windows-based SPSS software version 16. 4. Results 4.1. Variability of soil moisture Table 2 displays SM figures, directly linked to rainfalls (Fig. 2). Despite being the wettest place, CO did not register the highest figure of SM, although in general terms SM was greater in CO and AL with respect to BE, as was also the case with SM differences. Even though abundant rainfalls were measured during the two year sampling period, the SM average did not rise above 0.10 cm3 cm−3 owing to the strong influence of the low values of SM during the summer in the three field sites. SM diminished rapidly when there was a lack of rain. The extremely dry season (CO and AL: four months; and BE: five to six months) rendered an edaphic drought which lasted until the end of September in CO and AL and until October in BE. This long dry season evened out the three hillslopes from the hydrological point of view by placing them all at wilting point. All these inter-annual SM changes were significant according to the F-test for p N 0.05 (Table 3). The spatial variability of SM intra-hillslopes showed important differences in the three field sites analyzed (Fig. 3). In CO and AL there was a greater disparity between the levels of SM in the hillslope sections, while in BE spatial uniformity was greater (Table 2). Both in CO and in AL the existence of a steeper slope in the area nearest the thalweg meant lower levels of SM. The same relationship with the topography but inversely, was observed on the BE hillslope where the lowest level was recorded at the top (slope gradient equal to 28%). The SM relationship with the slope was clearly defined statistically by Pearson's correlation coefficient: CO, r = −0.68; AL, r = −0.56; BE, r = −0.53; p b 0.05. Although the extreme figures were similar in the three field sites, the spatial variability of SM decreased as rainfall was reduced: both the Std and the VC were lower in BE. The same tendency was established by the parametric δij test. Grayson and Western (1998) considered more stable when the mean relative difference approached Table 2 Summary of the statistic test values of the soil moisture at the field sites. Units of soil moisture in cm3 cm−3.

Total

Top hillslope

Mid hillslope

Bottom hillslope

Mean dij Std VC Max Min Mean Std VC Max Min Mean Std VC Max Min Mean Std VC Max Min

Colmenar

Almogía

Berja

0.07 −0.18 0.05 0.71 0.27 0.001 0.07 0.04 0.60 0.24 0.001 0.07 0.06 0.80 0.27 0.001 0.05 0.05 0.83 0.20 0.001

0.07 −0.10 0.05 0.64 0.26 0.001 0.06 0.04 0.67 0.24 0.001 0.08 0.05 0.70 0.26 0.001 0.06 0.04 0.68 0.20 0.001

0.04 −0.01 0.03 0.68 0.22 0.001 0.04 0.03 0.69 0.21 0.001 0.05 0.03 0.76 0.20 0.001 0.04 0.04 0.82 0.22 0.001

Total: relationship between the mean ratio and the mean soil moisture for the whole hillslope. δij: relative differencing index. Std: standard deviation. VC: variation coefficient. Max: higher measured value of soil moisture. Min: lower measured value of soil moisture.

Fig. 2. Mean soil moisture variability and monthly rainfall at the abandoned fields since Oct-2002 to Nov-2004. Colmenar (top); Almogía (med); Berja (bottom).

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11 Table 3 Summary of F-test results. Significance leves (p-values): significantly different at p b 0.05* and p b 0.01**. Field site

CO AL BE

Soil moisture

Vegetal cover

Ratio veg./non-veg.

T.

S.

T.

S.

T.

S.

0.94* 1.30* 0.61*

– – –

1.61* 2.71* 4.14*

– 3.91** 4.06**

2.03* 2.45* 4.66**

– 4.54** 2.46*

T: temporal variability of the variable during the sampling period. S: spatial variability of the variable along the hillslope.

zero. Martínez-Fernández and Ceballos (2005) observed that the most stable areas from the point of view of SM were the most arid ones. Our findings were similar: the lowest figure of δij corresponded to BE. The major difference among the rainfall depth of wet and dry seasons, the higher variability of SM. During the dry season, the hydrological state of the soil was similar in the three field sites, though the soil was under the wilting point for a longer period in the driest field site. Fig. 4 shows the temporal variation of the δij test in the three field sites. Low rainfalls in BE complicated the generalization of high SM figures and the differentiation of hydrological behaviour along the hillslope, so that the results of the δij test were always close to zero. However, the most extreme figures of the test were usually obtained in CO where it was normal to record in the same day and in different areas, very low SM figures which moved away from the average hillslope figures. This indicated a certain complexity in its hydrological functioning. In AL the results of the test were at times very high indicating a strong SM variability on days when a large area of the hillslope presented a very low average. So greater SM variability on the hillslopes with greatest rainfall constituted a spatial and temporal limiting factor of growth and maintaining vegetation. Its functioning and spatial distribution would be connected to this SM variability, with a succession of vegetated and non-vegetated areas. In BE a priori there was also a limiting factor but in the opposite direction: the scarcity of water in a large area of the hillslope which limited the vegetation pattern all along it. 4.2. Changes in vegetation pattern The spatial distribution of the permanent vegetation in the three field sites was dominated by a vegetation pattern in a two phase mosaic (patches with vegetation and patches without vegetation) (Fig. 5). The mean vegetal cover was similar in the three field sites during the sampling period despite the different climatic conditions (vegetation cover: CO = 49.7%, AL = 47.8%, and BE = 50.7%). However, we observed differences between them. CO and AL generally showed a decrease in VC downslope: the highest slope was a limiting factor of the soil, which was very affected by water erosion, especially in CO where rock outcrops were observed. BE showed an opposite tendency. Statistically, for p N 0.01 (Table 3), VC changed significantly along the hillslope in the field sites with less rain (AL and BE), while in CO vegetation cover diminished downslope although statistically these spatial changes were not found to be significant. This spatial pattern could have as a hydrological consequence the appearance of flow disconnection processes if we assume the areas with vegetation as sink areas. The evolution of vegetation cover during the set period of time was indeed significant in the three field sites according to the F-test (Table 3), and especially in CO and AL, because the type of vegetation characteristic of these field sites required a greater amount of water to subsist, and therefore died during the dry season. This was the case for Cistus monspeliensis of which vegetal cover reduced in 75% given it is a chamaephyte vegetal species. However, in the driest field site, (BE), the vegetation type (xerophytic) is better adapted to a prolonged lack of water, and we observed less significant changes between the wet and dry seasons.

5

The ratio veg/non-veg was calculated for each date of vegetation pattern sampling (Table 4). Arnau-Rosalén et al. (2008) obtained similar ratio figures for similar climatic conditions but in calcareous hillslopes. Generally, in all the field sites, this ratio diminished at the end of the dry season given that the area of bare soil increased significantly after the summer because of the edaphic drought. In the same way, the ratio diminished lower down the slope in CO where vegetation cover decreased owing to the diminution of SM there, and in AL it tended to be inferior in the middle area where the edaphic conditions were poorer to retain water and thus for vegetation growth: more gravelly soils and sandy and greater bulk density. However, in BE this occurred at the top of the hillslope, with a steeper gradient slope and shallow soils with abundant rock fragments on the surface. The intra-hillslope differences increased at the end of the wet season as there was greater vegetal cover especially where the slope was less steep and the soil was able to infiltrate water allowing for a greater volume of retention: the top hillslope of CO and AL and the middle and bottom of BE. The vegetation pattern was composed of Mediterranean vegetal species, adapted to each climatic area. The presence or absence of a particular type of vegetal species could be used as an indicator of the level of resilience of the system in each studied area following that described by other researchers (López-Bermúdez et al., 1998; Ruiz Flaño, 1993). Table 5 shows the vegetal species observed in the sampling plots and some of the natural vegetal species typical of the climax situation in agreement with Ruiz-de la Torre (1991a,b), the existence of these would indicate a more evolved degree of vegetal succession in a bio-geographical region, as CO is at level 3, AL at level 4 and BE at level three in a scale of 1 to 6, according to the six levels of vegetation succession defined by Ruiz-de la Torre (1991a,b). Hence AL was the field site where it seemed that natural recovery conditions were best, showing greater biodiversity, and with even woodland substratum with Quercus suber sprotus. In BE biodiversity was relatively high, although the species Anthyllis cytisoides, a species which indicates recovery (Ruiz-de la Torre, 1991a), was missing. Lastly, CO was poorer in vegetal species as numerous scrubland and woodland species were lacking which indicated that it was far from its climactic condition. In fact, the environmental conditions of CO showed signs of degradation, based on reduced vegetal cover (VC), shallow soil (b20 cm) and rock outcrops on the lower part of the slope with a steep gradient (N50%). 5. Discussion The direct influence of SM on the development and survival of vegetation was evident. Kosmas et al. (2000) pointed out that climate and soil degradation are key factors in the recovery of abandoned fields. Such soil degradation has a direct effect given that it influences greatly the soil water holding capacity among other properties (Lavee et al., 1998). Depending on this soil degradation state, the geo-system will have a certain level of resilience and vegetal cover will have greater or lesser success in the re-colonization of such abandoned fields (López-Bermúdez et al., 1998; Ruiz Flaño, 1993). The vegetated/non-vegetated area showed a significant correlation with the mean SM of the six months previous to the date of the respective photographs of the vegetation pattern (mean soil moisture of dry and wet seasons) (Fig. 6). The relationship between both variables was lineal in CO and AL with Pearson's correlation coefficients being significant (Table 6) and regression for p N 0.05. However, this relationship was of logarithmic type in the semi-arid field site: the steepest slope factor indicated greater effectiveness in vegetation consumption of available water in the soil as there was a slight increase in SM. As this type of vegetation is more adapted to a lack of water when there is available, vegetation cover increased despite scarce rainfall. The temporal changes in the ratio were most

6

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

Fig. 3. Temporal variability of soil moisture from the sampled hillslope sections at the abandoned fields. CO: Colmenar; AL: Almogía; BE: Berja.

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

7

Fig. 4. Temporal variability of the relative differencing index at the field sites. CO: Colmenar; AL: Almogía; BE: Berja.

Fig. 5. Plot vegetation patterns at every sampled hillslope section. Black colour corresponds to shrub and white colour to bare soil. The number beside every map means the vegetal cover percentage.

8

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

Table 4 Temporal and spatial variability of the vegetated/non-vegetated ratio at the field sites. Field site

Hillslope section

Nov-02

May-03

Nov-03

May-04

Nov-04

Colmenar

Mean Top Medium Bottom Mean Top Medium Bottom Mean Top Medium Bottom

0.84 1.50 0.89 0.43 0.91 1.22 0.75 0.82 1.07 0.72 1.38 1.22

1.54 2.13 2.57 0.72 1.11 1.50 0.89 1.04 1.70 0.92 2.70 2.13

0.70 1.38 0.67 0.35 0.81 1.22 0.64 0.67 0.89 0.61 1.08 1.04

1.26 2.23 1.50 0.61 1.19 1.86 0.72 1.27 1.04 0.67 1.22 1.38

0.80 1.50 0.82 0.39 0.67 0.89 0.75 0.43 0.69 0.47 1.00 0.67

Almogia

Berja

Veg./non-veg. ratio equal to 1 means a vegetal cover of 50%.

pronounced in CO because its vegetation cover suffered most the lack of water in the summer. BE showed a strong variation the first year of sampling while the second year its evolution was more uniform, with figures remaining lower (drier year). From the intra-hillslope point of view, in both CO and BE the middle area showed more sudden spatial changes. This section was a transition section between the most and least degraded areas in both hillslopes. This occurred in both the high and low section of AL. All these changes were significant according to the F-test in the three field sites both spatially and temporally, corroborating the control which the areas of vegetation exert as sink areas in SM spatial pattern on Mediterranean hillslopes as has been observed in other studies (Bromley et al., 1997; Francis et al., 1986; Katra et al., 2007; Rodriguez-Iturbe et al., 1999; Sarah, 2002; Yair and Danin, 1980). This vegetation pattern in two phase mosaics is very common in abandoned fields re-colonized by scrubland and in dry and semi-arid Mediterranean areas (Arnau-Rosalén et al., 2008; Lavee et al., 1998). Likewise, there were also consequences from the hydrological point of view given that they defined spatially the sink and source areas during the runoff generation process and control hydrological connectivity between the different parts of a hillslope (Bergkamp, 1998; Calvo et al., 2003; Cammeraat, 2004; Cammeraat and Imeson, 1999). The SM δij test was correlated with the vegetal cover difference between one season and the next to establish the influence of soil moisture variability on the seasonal changes of vegetation pattern. We observed the correlation coefficient increased as the climate became drier (CO, r = −0.39; AL, r = 0.79; BE, r = 0.88; p-values

N0.05). The SM modifications in BE whose δij test figures were close to zero were significant than on the other hillslopes. Throughout the year SM underwent fewer changes in BE where the figures were of less than 10% almost all year, but these small changes were not an impediment to the vegetation responding rapidly and increasing soil cover. This was also indicated by the logarithmic relationship between the ratio veg/non-veg and the SM mean of the six months previous. On the contrary, in AL and above all in CO, the lower SM stability resulted in more evident changes in vegetation cover as was obtained earlier with the F-test. A lineal equation was drawn up from the points resulting from the graphic representation of the SM mean of each season and the vegetal cover of both the whole hillslope and of the sections, to obtain the theoretical vegetation cover according to the SM mean. Later we recalculated the vegetation cover figures expected according to the SM mean of each season and we compared them with those observed (Fig. 7). In CO and BE the mean square errors were greater than in AL, with a greater uncertainty, so that the seasonal changes in vegetal cover could not be reliably predicted from SM alone. In CO, other factors influenced the process: i) degraded soil with scarce development and limited infiltration capacity and water holding capacity (the year 2003–04 was more rainy with higher intensities, but with lower SM), and ii) the steep slope which limited this infiltration and favored water erosion. Both had negative effects by limiting vegetation cover growth, despite the fact that the increased rainfall could positively affect the resilience of the system. In AL the key factors were also degraded soil and a lower amount of water per year, but in contrast to CO, the slope gradient was less steep, thus reducing the conditions which favour runoff generation and therefore water erosion. So, although vegetation cover was not very high, there were sufficient signs of recovery from the point of view of the type of vegetal species on the hillslope. In BE the limiting factors were the same: i) a limited quantity of available water for the system during the year because of the scarcity of rain, to which the plants had to adapt, and ii) very shallow soil with scarce amounts of silt and clay (which also limited soil water holding capacity), a lack of structure and abundant surface rock fragments arranged like slates which limited water infiltration processes. However, this high sand content enhanced that the characteristic soil moisture content at wilting point were very low (b0.05 cm3 cm−3), which allowed the plants to benefit from an amount of available water when there was a very low percentage of SM during a large part of the dry season (Ruiz-Sinoga and Martínez-Murillo, 2009).

Table 5 Observed and non-observed vegetal species at the abandoned fields. Non-observed vegetal species corresponds to those that were presented nearby natural areas in the surrounding. Field site

CO

AL

BE

Observed vegetal species

Cistus monspeliensis Genista umbellata Lavandula stoechas Phlomis purpurea Helychrisum stoechas Ulex parviflorus

Genista spartiodes Thymus zygis Thymelea hirsuta Stipa tenacissima

Non-observed vegetal species

Quercus suber or ilex Cistus albidus Cistus ladanifer Quercus coccifera Eleagnus angustifolia Phillyrea angustifolia Crataegus monogyna Pistacia lenticus

Cistus monspeliensis Genista umbellata Lavandula stoechas Phlomis purpurea Helychrisum stoechas Ulex parviflorus Quercus coccifera Teucrium fruticans Cistus albidus Retama sphaerocarpa Chamaerops humilis Quercus suber Quercus ilex Calicotome villosa

Ulex parviflorus Anthyllis cytisoides

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

9

Table 6 Results of Pearson's correlation coefficients versus soil moisture (SM) and vegetated/ non-vegetated ratio (Ratio) and versus soil moisture and vegetal cover (VC). Significance level: p b 0.05. Field site

Hillslope section

Ratio/SM

VC/SM

Colmenar

Total Top Medium Bottom Total Top Medium Bottom Total Top Medium Bottom

1.00 0.93 0.95 0.97 0.89 0.71 0.71 0.79 0.83 0.85 0.72 0.86

0.99 0.94 0.99 0.96 0.87 0.71 0.71 0.78 0.87 0.85 0.78 0.87

Almogia

Berja

erosive potential (López-Bermúdez and Albadalejo, 1990), detaching fine soil particles and connecting different sections of the hillslope in a process already observed in other studies (Cammeraat, 2002, 2004; Davenport et al., 1998; Puigdefábregas et al., 1999; Yair and Kossovsky, 2002; Yair and Lavee, 1985). As pointed out by Hewlett and Hibbert (1967) and Solé-Benet et al. (1997), the slope gradient plays a crucial role in this hydrological process, even more so taking into account that in metamorphic areas with slates, phyllites and schists the slope usually increases downslope because of the drainage network dissection. On the other hand, the erosion of the spring rains (which can also be intense in the Mediterranean) would be inferior because the sink areas would occupy a maximum extension at the end of the wet season when vegetation cover occupies a greater surface area. In fact, Martínez-Murillo and Ruiz-Sinoga (2007) checked this seasonable effect on water erosion linked to changes in vegetation cover and SM in similar eco-geomorphological conditions. Likewise, Ruiz-Sinoga and Martínez-Murillo (2009) also corroborated seasonal changes in the soil–water–plant system in similar environmental conditions. 6. Conclusions

i) The spatial and temporal variability of soil moisture has repercussions in the vegetation pattern of the analyzed abandoned fields. This relationship was more evident when the climate was wetter: areas with different soil moisture content were

Fig. 6. Relationships between soil moisture and ratio veg/non-veg in the abandoned fields. Total: relationship between the mean ratio and the mean soil moisture for the whole hillslope. Med: medium section of the hillslope; Bot: bottom section of the hillslope. R2: value of the regression coefficient.

In short, after more than fifty years' abandonment, the geoecosystem has evolved subject to climatic, soil and topographic factors. The intra and inter-annual climatic variability caused SM spatial-temporal dynamics which in turn affected vegetation cover, and all those modifications in the vegetation pattern had hydrological and erosive consequences. Given that the lowest vegetation cover was registered at the end of the dry season, the soil was in a more vulnerable state just before the autumn rains: reduced vegetation cover faced with intense rainfall and soil limited by water infiltration favored the generation of runoff. And with a greater surface of bare soil, this runoff could run a greater distance downslope, increasing its

Fig. 7. Predicted versus measured vegetation cover for every field site.

10

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11

differentiated along the hillslopes according to their different states of degradation and especially their slope gradient, which favored greater vegetal cover at hilltop (less slope gradient) in both subhumid and dry climatic regimes. But when climate became more aridity, SM spatial pattern was more uniform, originating less significant changes in the vegetation which was better adapted to the lack of water and its cover was modified by only a slight increase in soil water content. The worst adaptation of the dominant vegetation type in CO and AL to water stress became evident at the end of the dry season because we obtained lower ratios veg./non-veg. ii) Intra-hillslope and inter-annual variability of the vegetation pattern a priori, resulted in a spatial and temporal functioning of sink and source areas of runoff. The three abandoned fields analyzed, each in different climatic conditions, were characterized by an increase of the areas with no vegetation as a consequence of the long rainfall and edaphic drought. This meant a larger surface of the hillslope vulnerable to the impact of water erosion, favored by the higher slope gradient. During the wet season, vegetal cover increased on the hillslopes and thus, the potential surface of sink areas. Also, as this dynamic pattern was in patches, hydrological disconnections could appear between the sections of the hillslope triggering one complex hydrological functioning in time and space at hillslope scale. iii) The abandoned field under dry Mediterranean climatic conditions appeared to be in a more advanced state of recovery considering the vegetal species diversity, and was therefore the field site closest to quasi-natural conditions. In addition, it could be very close to the climactic situation bearing in mind the limitations of a climate variable and poor in rainfall and the state of soil degradation. Lastly, the wettest field site was the one which was the furthest away from its original natural state. Acknowledgments The authors are grateful to the Agencia Andaluza del Agua and the III Plan Andaluz de Investigación, both organisms that belong to the Regional Government of Andalucía, for their financial support, without which this study would not have been possible.

References AFNOR (Association française de normalisation), 1987. Qualité de sols, methods d'analyse. AFNOR, Paris, France. Arnau-Rosalén, E., Calvo-Cases, A., Boix-Fayos, C., Lavee, H., Sarah, P., 2008. Analysis of soil surface component patterns affecting runoff generation. An example of methods applied to Mediterranean hillslopes in Alicante (Spain). Geomorphology 101, 595–606. Bergkamp, G., 1998. Hydrological influences on the resilience of Quercus spp. dominated geoecosystems in central Spain. Geomorphology 23, 101–126. Blake, G.R., Hartge, K.H., 1986. Bulk density. In: Klute, A. (Ed.), Methods of Soil Analysis, Part 1, 2nd Edition, Agronomy Monograph, Vol. 9. American Society of Agronomy, Madison, WI, pp. 363–375. Boix, C., Soriano, M.M., Tiemessen, I.R., Calvo, A., Imeson, A.C., 1995. Properties and erosional response of soils in a degraded ecosystem in Crete (Greece). Environment Monitoring and Assessment 37, 79–92. Bromley, J., Brouwer, J., Barker, A.P., Gaze, S.R., Valentine, C., 1997. The role of surface water redistribution in an area of patterned vegetation in a semi-arid environment, South-West Niger. Journal of Hydrology 198, 1–29. Calvo, A., Boix, C., Imeson, A.C., 2003. Runoff generation, sediment movement and soil water behaviour on calcareous (limestone) slopes of some Mediterranean environments in Southeast Spain. Geomorphology 50, 269–291. Calvo-Cases, A., Boix-Fayos, C., Arnau-Roselen, E., 2005. Pattern and thresholds of runoff generation and sediment transport on some Mediterranean hillslopes. In: García, C., Batalla, R. (Eds.), Catchment dynamics and river processes: Mediterranean and other climate regions. Elsevier, Amsterdam, pp. 31–51.

Cammeraat, L.H., 2002. A review of two strongly contrasting geomorphological systems within the context of scale. Earth Surface Processes and Landforms 27, 1201–1222. Cammeraat, L.H., 2004. Scale dependent thresholds in hydrological and erosion response of a semi-arid catchment in southeast Spain. Agriculture Ecosystems & Environment 104, 317–332. Cammeraat, L.H., Imeson, A.C., 1999. The evolution and significance of soil vegetation patterns following land abandonment and fire in Spain. Catena 37, 107–127. Davenport, D.W., Breshears, D.D., Wilcox, B.P., Allen, C.D., 1998. Viewpoint: sustainability of pinon–juniper ecosystems — a unifying perspective of soil erosion thresholds. Journal of Range Management 51, 231–240. FAO, 2006. Base referencial mundial del recurso suelo. Informes sobre recursos mundiales de suelos 84, FAO, Roma, 91 p. Francis, C.F., Thornes, J.B., Romero-Diaz, A., Lopez-Bermudez, F., Fisher, G.C., 1986. Topographic control of soil moisture, vegetation cover and land degradation in a moisture stressed Mediterranean environment. Catena 13, 211–225. Grayson, R.B., Western, A.W., 1998. Towards areal estimation of soil water content from point measurements: time and space stability of mean response. J. Hidrol. 207, 68–82. Hewlett, J.D., Hibbert, A.R., 1967. Factors affecting the response of small watersheds to precipitation in humid areas. In: Sopper, W.E., Lull, H.W. (Eds.), Forest Hydrology. Pergamon, New York, pp. 275–290. Imeson, A.C., Lavee, H., 1998. Soil erosion and climate change: the transect approach and the influence of scale. Geomorphology 23, 319–337. Katra, I., Blumberg, D.G., Lavee, H., Sarah, P., 2007. Topsoil moisture patterns on arid hillsides — Micro-scale mapping by thermal infrared images. Journal of Hydrology 334, 359–367. Kemper, W.D., Rosenau, R.C., 1986. Aggregate stability and size distribution, In: Klute, A. (Ed.), Methods of Soil Analysis. Part I. Physical and Mineralogical Methods, 2nd edition. American Society of Agronomy-Soil Science Society of America, Madison, pp. 425–442. Klute, A., 1965. Laboratory measurement of hydraulic conductivity of saturated soil. In: Black, C.A. (Ed.), Methods of soil analysis. Am. Soc. of Agron, Madison, pp. 210–221. Kosmas, C., Danalatos, N.G., Gerontidis, S., 2000. The effect of land parameters on vegetation performance and degree of erosion under Mediterranean conditions. Catena 40, 3–17. Lacoste, A., Salanon, R., 1978. Biogeografía. Oikos-Tau editorial, Barcelona. Lavee, H., Imeson, A.C., Sarah, P., 1998. The impact of climate change on geomorphology and desertification along a Mediterranean arid transect. Land Degradation and Development 9, 407–422. Lesschen, J.P., Schoorl, J.M., Cammeraat, L.H., 2009. Modelling runoff and erosion for a semi-arid catchment using a multi-scale approach based on hydrological connectivity. Geomorphology 109, 174–183. López-Bermúdez, F., 1993. Reflexiones sobre la degradación de los suelos y su gestión sostenible en la Cuenca Mediterránea. In: Instituto de Estudios Almerienses (editor), Desertificación y uso del suelo en la Cuenca Mediterránea, Paralelo 37, Vol. 16, Almería, pp. 211–218. López-Bermúdez, F., Albadalejo, J., 1990. Factores ambientales de degradación del suelo. In: López-Bermúdez, F., Albadalejo, J., Díaz, E. (Eds.), Degradación y regeneración del suelo en condiciones ambientales mediterráneas. CSIC, Murcia, pp. 15–46. López-Bermúdez, F., Romero-Díaz, A., Martínez-Fernández, J., 1998. Vegetation and soil erosion under a semiarid Mediterranean climate: a case study from Murcia (Spain). Geomorphology 24, 51–58. Martínez-Fernández, J., Ceballos, A., 2005. Mean soil moisture estimation using temporal stability analysis. Journal of Hydrology 312, 28–38. Martínez-Murillo, J.F., Ruiz-Sinoga, J.D., 2007. Seasonal changes in the hydrological and erosional response of a hillslope under dry-Mediterranean climatic conditions (Montes de Málaga, South of Spain). Geomorphology 88, 69–83. Puigdefábregas, J., Sole, A., Gutiérrez, L., del Barrio, G., Boer, M., 1999. Scales and processes of water and sediment redistribution in drylands: results from Rambla Honda field site in Southeast Spain. Earth Science Reviews 48, 39–70. Rietkerk, M., Boerlijst, M.C., van Langevelde, F., HilleRisLambers, R., van de Koppel, J., Kumar, L., Prins, H.H.T., de Roos, A.M., 2002. Self-organization of vegetation in arid ecosystems. The American Naturalist 160, 524–530. Robinson, G.W., 1922. A new method for mechanical analysis of soil and other dispersion. Journal of Agriculture Science 12, 306–321. Rodriguez-Iturbe, I., D'Odorico, P., Porporato, A., Ridolfi, L., 1999. On the spatial and temporal links between vegetation, climate and soil moisture. Water Resources Research 12, 3709–3722. Ruiz Flaño, P., 1993. Procesos de erosion en campos abandonados del Pirineo. El ejemplo del Valle de Aísa. Geoforma ediciones, Logroño. Ruiz-de la Torre, J., 1991a. Mapa forestal de España 1:200.000: Granada-Málaga. Ministerio de Agricultura, Pesca y Alimentación, Madrid. Ruiz-de la Torre, J., 1991b. Mapa forestal de España 1:200.000: Almería. Ministerio de Agricultura, Pesca y Alimentación, Madrid. Ruiz-Sinoga, J.D., Martínez-Murillo, J.F., 2009. Eco-geomorphological system response variability to the 2004–06 drought along a climatic gradient of the Littoral Betic Range (southern Spain). Geomorphology 109, 351–362. Ruiz-Sinoga, J.D., Martínez-Murillo, J.F., Gallegos-Reina, A., Delgado-Peña, J.J., LucasSantamaría, B., Romero-Lopera, A., Noguera-Robles, M.J., Márquez-Carrero, J., 2003. Variabilidad de los procesos de generación de escorrentía en laderas bajo condiciones mediterráneas. Baetica 25, 279–311. Sarah, P., 2002. Spatial patterns of soil moisture as affected by shrubs, in different climatic conditions. Environmental Monitoring and Assessment 73, 237–251. Shnerb, N.M., Sarah, P., Lavee, H., Solomon, S., 2003. Reactive glass and vegetation patterns. Physical Review Letters 90, 0381011–0381014.

J.D. Ruiz-Sinoga et al. / Catena 85 (2011) 1–11 Solé-Benet, A., Calvo, A., Cerdà, A., Lazaro, R., Pini, R., Barbero, J., 1997. Influences of micro-relief patterns and plant cover on runoff related processes in badlands from Tabernas (SE Spain). Catena 31, 23–28. Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resources Research 68, 574–582. Vachaud, G., Passerat de Silans, A., Balabanis, P., Vauclin, M., 1985. Temporal stability of spatially measured soil water probability density function. Soil Science Society of America Journal 49, 822–828.

11

Yair, A., Danin, A., 1980. Spatial variation in vegetation as related to the soil moisture regime over an arid limestone hillside, northern Negev, Israel. Oecologia 47, 83–88. Yair, A., Kossovsky, A., 2002. Climate and surface properties: hydrological response of small and semi-arid watersheds. Geomorphology 42, 43–57. Yair, A., Lavee, H., 1985. Runoff generation in arid and semi-arid zones. In: Anderson, M.G., Burt, T.P. (Eds.), Hydrological Forecasting. JohnWiley and Sons Ltd., London, pp. 183–220.