needed to attain coverage, frequency, and species com- Riparian position comparable with that of undisturbed zones. Key words: Mediterranean semiarid streams, plant res- Restoration in toration, plant structure, species frequency. Summer-Dry Riverbeds of Southeastern Introduction iparian ecosystems are spatially and temporally 1,3 M. Jacoba Salinas R dynamic and are shaped by fluvial geomorphic 2 processes. Thus, there are physical and biological links José Guirado between terrestrial and aquatic environments (Gregory et al. 1991) and biotopes in which animals may seek ref- Abstract uge and food, while enriching the soil in detritus. Stud- ies in North America have shown that semiarid and An evaluation was made of the development of two arid riparian ecosystems support higher species rich- experimental plots where restoration of dominant ri- ness and densities of wildlife than do other nearby eco- parian plant species was conducted in December 1991 systems (Johnson & Simpson 1971; Carothers et al. 1974). along two semiarid Mediterranean summer-dry wa- Semiarid Mediterranean riparian vegetation is the only tercourses. An overall comparison was made of the type of arboreal/shrubby vegetation in these areas and vegetation structure, species cover, floral composition, differs markedly in phenology from the other commu- and species richness of the plots restored using vege- nities, thus giving the landscape distinctive features. It tation from nearby undisturbed plots along the same detains erosion materials (a serious problem in the watercourse. The monitoring was performed in Octo- semiarid Mediterranean region, because the hillsides ber 1993, October 1995, September 1997, and October have sparse vegetation cover), thus decreasing the 1999. In the restored zones previously rooted cuttings amount of solids in suspension in the watercourses and of the species most representative of these communi- improving the quality of the water (Pesson 1978; Man- ties were planted, using the undisturbed zones as veg- teiga 1992). It slows down the flow of torrential rains etation models. Climatological conditions (particu- and collects the material carried, reducing the effects larly the rainfall regime during the planting period) downstream. Furthermore, the highly developed root substantially favored the success of the planting es- systems reinforce the banks of the streams. tablishment. The results show that a simple planting In recent years several techniques have been devel- technique accompanied by monitoring during the oped to stabilize river and stream banks using arboreal first year is adequate to achieve success in establish- or shrubby vegetation, either alone or in combination ment of planting species. It is necessary to take pre- with inert materials (Glover & Ford 1990, unpublished cautions against herbivory of small of Chamae- data; Glennon & Ritz 1994, unpublished data; Leiser et rops humilis, Ficus carica, and Retama sphaerocarpa. al. 1994, unpublished data). The more the vegetation The planting itself causes some disturbance in the soil grows, the greater the protection it offers, provided the that may alter the species composition, giving an ad- water flow is not obstructed (López 1988a). All these vantage to ruderal species over others. More time is advantages, together with the considerable enhance- ment of the landscape that this vegetation affords under semiarid Mediterranean climatic conditions, justify con- sidering this type of vegetation as being of primary impor- 1Departamento de Biología Vegetal, Facultad de Ciencias, C/ tance. The maintenance and/or restoration of vegetation Severo Ochoa s/n, Universidad de Granada, Granada, Spain thus deserve to be given priority in land management 18071. projects. 2 Delegación Provincial de la Consejería de Medio Ambiente The watercourses chosen for this experiment are de la Junta de Andalucía, Centro Residencial Oliveros, Bloque among the most seriously altered by human activity in Singular, 2a planta, Almería, Spain 04071. this area of southern Spain. Original riverside commu- 3Address correspondence to M. J. Salinas, e-mail: maria@ goliat.ugr.es. nities have been gravely degenerated or destroyed by tree cutting, the introduction of exotic species, the di- © 2002 Society for Ecological Restoration version and channeling of water for agriculture, and the

DECEMBER 2002 Restoration Ecology Vol. 10 No. 4, pp. 695–702 695

Plant Restoration in Semiarid Summer-Dry Riverbeds use of river beds and shores for cultivation or even as roads (due to low or nonexistent water flow) (Salinas 1995; Salinas et al. 1999). At present the damaging hu- man activities have ceased to a large degree (primarily agriculture, due to low profitability), and therefore large-scale restoration projects are now feasible. In ad- dition, these courses are located in the interior of the maritime-terrestrial Natural Park Cabo de Gata-Níjar, an area protected since 1987 for its unique semiarid flora and fauna. One of the prime objectives of those Figure 1. Monthly distribution of rainfall in millimeters for managing this park is to preserve and restore degraded the period 1991–1999. The thicker line represents mean plant communities. monthly rainfall for the period 1941–1970. Data from meteoro- To examine the success of restoring riparian vegeta- logic station, Almería airport. tion in these environments, we monitored plots situated on the banks of two different watercourses where ripar- ian vegetation has been reintroduced, comparing their presenting both areas with well-conserved natural ri- development with undisturbed vegetation of other parian vegetation and areas where the natural vegeta- nearby zones with similar characteristics. tion was completely destroyed by human activity (Sali- nas 1995). Materials and Methods Sampling of the Vegetation Site Description Along the undisturbed stretches four plots of 120 3 m The study was conducted in the southeastern Iberian were randomly chosen on both banks of the two water- Peninsula, along the lower stretches of two small streams courses (Amoladeras and Agua rambla, two per site, with sporadic seasonal flow known as ramblas, located one per each bank). In July and October 1991 we mea- in the interior of the maritime-terrestrial Natural Park sured the percentage cover and the frequency of all the Cabo de Gata-Níjar (Almería, Spain). The streams begin phanerogamous plant species in each plot by the line- on the southern slope of the Sierra de Alhamilla (18 km intercept method (Bullock 1996). The intercept mea- from the ), located approximately be- surements were taken at heights of 1.5 m, 50 cm, and 10 tween 3640 and 37N and 215 and 220W, with alti- cm. Sampling was repeated in October 1993, October tude varying from 0 to 1387 m, and discharge into the 1995, September 1997, and October 1999. Mediterranean Sea. The flow is highly irregular, with the total absence of surface water over most of the Planting length of these streams and during most of the year. The mean annual water flow is about 33.1 hm3 (Martín- The plant species chosen for restoration were those oc- Vivaldi 1991). The climate is semiarid Mediterranean curring most frequently and with greatest cover in un- (Le Houèrou 1982), with a mean annual rainfall of 241 disturbed plots (Table 1). Planting was done in Decem- mm (average for 1941 to 1970, data from the Instituto ber 1991 at sites 1 and 2, where natural vegetation was Nacional de Meteorología, Almería airport, Fig. 1), al- completely destroyed. Plots were selected to be the though interannually it fluctuates greatly (ranging from same size as the undisturbed plots (120 3 m) and 65.9 mm in 1998 to 551.5 mm in 1989). Rain falls mainly were located in zones where the riparian vegetation in autumn and winter. Winters are mild and summers was completely destroyed on both banks of the water- hot; the median annual temperature is 18.3C (average courses (four plots in total, two per site). The number of from 1941 to 1970, data from the Instituto Nacional de individuals planted in each species is indicated in Table Meteorología, Almería airport). The most frequent geolog- 1. The procedure was as follows: An excavator was ical materials of the bedrocks are of Neogene-Quater- used to dig to a depth of 1 m on the banks to reach the nary origin consisting of sandstones and conglomerates soil layers that remain the wettest, particularly in sum- (Martín-Vivaldi 1991). Upland vegetation in the study mer. Afterward we made holes 0.5 m in depth and 1.5 area is dominated by shrubs and lowland vegetation by m apart. Each individual was planted bare root. The highly diverse shrubby communities, with scant cover- plots were monitored (wilt and possible elimination of age (Alcaraz et al. 1989; Peinado et al. 1992; Mota et al individuals by the visitors) for the first few months af- 1997). ter planting and were irrigated weekly during the first Two experimental sites were selected, site 1 (Amolad- summer. The plants came from cuttings rooted in a eras rambla) and site 2 (Agua rambla), characterized by nursery, with the exception of Retama sphaerocarpa

696 Restoration Ecology DECEMBER 2002

Plant Restoration in Semiarid Summer-Dry Riverbeds

Table 1. Dominant plant species for each undisturbed plot and number of planted individuals per 120 3–m plot at Amoladeras rambla (site 1) and Agua rambla (site 2).

Number of Planted Individuals Dominant Species in Undisturbed Plots Life Form Site 1 Site 2 Tamarix canariensis Willd. Halophyle phreatophyte shrub (saltcedar) 80 48 Nerium oleander L. Perennial shrub phreatophyte 42 28 Chamaerops humilis L. Palm shrub 17 18 Retama sphaerocarpa (L.) Boiss. Mediterranean shrub 16 14 Tamarix africana Poiret Halophyle phreatophyte shrub (saltcedar) 13 10 Ficus carica L. Deciduous tree 8 8

(broom), which were germinated from seed. An effort rennial grasses (Poaceae), very frequent in the sur- was made to plant after a rainy period, so that the sub- rounding shrublands (Stipa tenacissima and Lygeum strate would be saturated with water. During the first spartum) and other ruderals (Hyparrhenia hirta and few months there was no wilting or loss of planted indi- Piptatherum miliaceum). The presence of annuals, quite viduals. The sampling method and dates were similar scarce in these seminatural plant communities except to that described above for undisturbed plots. when there is sufficient rainfall to ensure their germina- tion and growth, proved highly variable among the dif- ferent years analyzed; ruderal species were the most Analysis frequent, outstanding examples being Calendula arvensis Statistical analyses were done using data sets from the (marigold), Lobularia maritima (Cruciferae), and Oxalis two sites studied, separately and jointly. To compare pes-caprae (Oxalidaceae). Among the lianas the most xe- the changes in the structure of the undisturbed plots romorphic species in the wettest garrigas dominated with the restored plots, we established five plant groups, Rubia peregrina (Rubiaceae) and Asparagus stipularis which in turn corresponded to six strata: trees, shrubs, (wild asparagus). Species richness proved greater at site cryptophytes and hemicryptophytes (sensu Raunkiaer 2 than at site 1. 1934), herbs, lianas, and annuals. Comparisons were The result of the reinstallation of Ficus carica was un- made between natural and restored plots for each pe- even. Although at site 1 it increased its cover over the riod analyzed by a t-test, following arcsine transforma- study period, at site 2 all the plants wilted between 1995 tion (Zar 1984), using cover and species frequency. Flo- and 1997. At both sites Tamarix canariensis, T. africana, ral similarity between natural and restored plots in the and Nerium oleander took hold satisfactorily, increasing different periods monitored at each site was compared their cover along the study periods (Tables 2 & 3). Al- by the Bray-Curtis similarity index (Mueller-Dombois though there was survival among some stalks of Retama & Ellenberg 1974), using species number. sphaerocarpa, the cover of this plant diminished over time at both sites. The failure was total in the reintro- Results duction of Chamaerops humilis (Palmae); between 1995 and 1997 all plants introduced at both sites wilted. The vegetation that developed in the seminatural plots Species richness increased progressively at both sites: studied was represented by a shrub stratum dominated at site 1 from 16 in 1991 to 31 in 1999 and at site 2 from by Tamarix canariensis (saltcedar) and Nerium oleander 22 to 31 over the same period, due to the spontaneous (oleander) (Tables 2 & 3). Trees are rare, with the excep- introduction of species. At site 1 notable among the tion of Ficus carica (fig tree), a small and relatively fre- spontaneous shrubs was Nicotiana glauca (Solanaceae) quent tree. The strata comprised of chamaephytes and and Atriplex halimus (barilla). Meanwhile, chamaephytes hemicryptophytes proved the most diverse. The most and hemicryptophytes included Artemisia campestris, frequent species were ruderals: Dittrichia viscosa (Com- Lavandula multifida (lavender), and Phagnalon saxatile positae), Ballota hirsuta (Labiatae), Artemisia campestris (Compositae). Among herbs Hyparrhenia hirta was the (Compositae), Artemisia barrelieri, and Anthyllis cyti- only spontaneous plant. Among lianas there were the soides (Leguminosae). This strata presented numerous two species present in the seminatural plots, appearing species with a highly restricted geographic distribution, in the last periods studied. In trees and annuals no such as Anthyllis terniflora and Helianthemum almeriense spontaneous species appeared. (Cistaceae). At site 2 Nicotiana glauca was the only spontaneous Conversely, the herbaceous strata of annuals and li- shrub. Notable representatives of chamaephytes and anas were very poor. The former were restricted to pe- hemicryptophytes were Eryngium campestre (Apiaceae),

DECEMBER 2002 Restoration Ecology 697

Plant Restoration in Semiarid Summer-Dry Riverbeds

Table 2. Species richness and percentage of cover of the species recorded in undisturbed (U) and restored (R) plots in the study years at site 1 (Amoladeras rambla).

Species Life Form U-91 U-93 U-95 U-97 U-99 R-91 R-93 R-95 R-97 R-99 Ficus carica L. T 1.5 1.9 2.3 1.9 2.6 1.3 1.2 1.6 2.8 2.8 Tamarix canariensis Willd. S 38.4 45.2 46.4 39.4 41.6 9.2 11.3 15.2 18.4 2.4 Nerium oleander L. S 24.2 29.3 24.2 22.5 28.3 7.3 9.2 11.1 14.3 16.5 Arundo donax L. S 7.4 8.1 1.4 8.2 6.6 4.1 5.2 3.8 4.3 5 Retama sphaerocarpa (L.) Boiss. S 9 8.6 7.9 9.3 7.5 1.1 0.7 0.5 0.5 0.3 Tamarix africana Poiret S 8.6 7.4 6.8 7.1 5.9 2.1 2.7 3 4.3 5.1 Chamaerops humilis L. S 5.7 4.2 4.8 3.5 4.7 0.3 0.3 0.2 Salsola oppositifolia Desf. S 3.2 2.8 1.9 1.5 2 Nicotiana glauca R. C. Graham S 2.1 2.4 1.9 1.5 1.3 0.1 0.6 0.6 0.7 Rhamnus lycioides L. S 0.8 0.5 0.7 0.6 0.4 Whitania frutescens (L.) Pauquy S 0.5 0.3 0.4 0.5 0.8 Atriplex halimus L. S 0.2 0.3 0.2 0.2 0.3 0.5 0.2 0.1 0.1 Lycium intrincatum Boiss. S 0.2 0.1 0.1 0.2 0.1 Artemisia barrelieri Besser C 1.2 1.1 1.36 1.1 1.5 0.15 0.3 0.4 0.5 0.5 Dittrichia viscosa (L.) W. Greuter C 1.1 1.3 1.5 1.2 1.4 2.1 2.5 2.5 2.4 2.5 Thymus hyemalis Lange C 1.1 0.9 0.5 0.1 0.1 Artemisia campestris L. C 1.5 0.9 1.2 1.3 1.1 0.1 0.1 0.2 Anthyllis cytisoides L. C 0.9 0.5 1.1 0.8 0.7 0.1 0.1 Anthyllis terniflora (Lag.) Pau C 0.5 0.3 0.75 0.9 0.4 0.1 Helianthemum almeriense Pau C 0.5 0.7 0.1 0.4 0.1 Thymelaea hirsuta (L.) Endl. C 0.4 0.2 0.75 0.6 0.7 0.1 Ballota hirsuta Bentham C 0.4 0.3 0.5 0.7 0.55 0.8 0.2 0.4 0.45 0.5 Lavandula multifida L. C 0.3 0.7 0.5 0.3 0.2 0.1 0.1 0.1 Ononis viscosa L. C 0.3 0.3 0.2 0.4 0.3 Phagnalon saxatile (L.) Cass. C 0.12 0.8 0.4 0.3 0.1 0.1 0.1 corymbosa L. subsp. corymbosa C 0.1 0.15 0.1 0.2 0.1 0.3 0.1 0.12 0.1 0.6 Eryngium campestre L. C 0.6 0.5 0.2 0.25 0.2 0.15 Ruta angustifolia Pers. C 0.5 0.2 0.1 0.1 Suaeda priunosa Lange C 0.3 0.4 0.5 0.3 0.1 Glaucium flavum Crantz C 0.3 0.1 0.1 0.1 Launaea arborescens (Batt.) Murb. C 0.2 0.3 0.4 0.7 0.2 0.1 0.1 Fagonia cretica L. C 0.2 0.1 Andryala ragusina L. C 0.2 0.3 0.1 0.1 0.5 0.1 Dorycnium pentaphyllum Scop. C 0.2 0.4 Stipa tenacissima L. H 2.4 2.9 1.7 3.7 3.2 Hyparrhenia hirta (L.) Stapf H 1 1.1 1 1.1 1 0.8 1 2.1 2.1 Piptatherum miliaceum (L.) Cosson H 0.4 0.3 0.6 0.3 0.3 0.1 0.3 0.4 0.3 0.4 Lygeum spartum L. H 0.1 0.1 0.1 Rubia peregrina L. subsp. peregrina L 0.5 0.4 0.7 0.8 0.9 0.1 Asparagus stipularis Forsk. L 0.3 0.3 0.1 0.2 0.1 0.2 0.1 Calendula arvensis L. A 0.7 0.87 0.9 0.2 0.6 0.7 0.6 0.4 Oxalis pes-caprae L. A 0.2 0.5 0.3 0.4 0.2 Lobularia maritima (L.) Desv. A 0.62 0.68 0.3 0.3 0.2 0.1 0.6 Species richness 40 37 36 37 37 16 19 21 25 31

Life forms: T, tree; S, shrub; C, chamaephyte or hemicryptophyte; H, herb; L, liana; A, annual.

Dorycnium pentaphyllum (Leguminosae), and Glaucium ond study period (1993) and maintaining highly irregu- flavum (Papaveraceae). There were no spontaneous spe- lar behavior in successive periods. At site 2 Artemisia cies among trees, herbs, or lianas. In annuals Calendula campestris, Dittrichia viscosa, and Artemisia barrelieri, rud- arvensis and Oxalis pes-caprae were spontaneous species. erals, were notable among the chamaephytes and hemi- Among the chamaephytes and hemicryptophytes ini- cryptophytes present initially in the restored plots, which tially present in the restored plots, at site 1, notable spe- increased their cover over the periods analyzed (Table cies included Artemisia barrelieri and Ballota hirsuta, both 3). The same was true of the herbs Piptatherum milia- ruderal, which increased their cover over the study pe- ceum and Hyparrhenia hirta. Lobularia maritima was the riods (Table 2). The same is true of the herb Piptatherum only annual present from the beginning, and its cover miliaceum. The three species of annuals present showed increased considerably during the third study period similar behavior, increasing their cover during the sec- (1995), remaining relatively stable afterward.

698 Restoration Ecology DECEMBER 2002 Plant Restoration in Semiarid Summer-Dry Riverbeds

Table 3. Species richness and percentage of cover of the species recorded in undisturbed (U) and restored (R) plots in the study years at site 2 (Agua rambla).

Species Life Form U-91 U-93 U-95 U-97 U-99 R-91 R-93 R-95 R-97 R-99 Ficus carica L. T 1.7 1.7 2.1 2.2 1.8 1.0 1.2 1 Tamarix canariensis Willd. S 28.4 25.4 3.5 32.2 25.8 8.3 1.3 12.8 13.6 15.6 Nerium oleander L. S 9.6 8.1 6.7 7.7 8.5 3.2 5.6 6.9 7.5 9.1 Retama sphaerocarpa (L.) Boiss. S 7.9 6.4 6.1 8.2 9.7 0.8 1.2 1.3 0.5 0.3 Tamarix africana Poiret S 3.8 5.1 4.8 7.8 5.8 1.3 2.5 3.2 5.6 6.2 Chamaerops humilis L. S 1.7 1.6 1.8 2.3 1.6 1.2 1.5 1.0 Arundo donax L. S 1.5 2.1 2.0 2.5 2.0 1.5 2.4 1.7 1.3 0.8 Ziziphus lotus (L.) Lam. S 1.2 1.1 1.3 2.8 3.6 Nicotiana glauca R. C. Graham S 0.8 1.1 0.6 2.6 1.7 0.1 0.3 0.5 0.4 Atriplex halimus L. S 0.2 0.2 0.6 0.3 0.2 0.5 0.5 0.5 0.1 0.1 Lycium intrincatum Boiss. S 0.1 0.1 0.1 0.2 0.1 Artemisia campestris L. C 1.6 1.3 1.7 2 1.7 0.1 0.4 0.5 0.6 0.6 Dittrichia viscosa (L.) W. Greuter C 0.9 1.1 1.3 0.7 1.1 0.8 0.3 0.4 0.3 0.5 Ballota hirsuta Bentham C 0.85 0.6 0.4 0.6 0.7 0.5 0.5 Anthyllis cytisoides L. C 0.5 0.7 0.8 1 0.9 0.1 Artemisia barrelieri Besser C 0.5 0.9 1.1 0.7 0.9 0.6 0.1 0.1 0.15 0.2 Carlina corymbosa L. subsp. corymbosa C 0.13 0.4 0.3 0.3 0.5 0.5 0.1 0.1 0.3 0.2 Dorycnium pentaphyllum Scop. C 0.7 0.1 0.1 0.1 0.1 0.3 0.5 0.6 Andryala ragusina L. C 0.5 0.3 0.5 0.9 0.5 0.1 Thymelaea hirsuta (L.) Endl. C 0.5 0.9 0.8 0.4 0.1 Helianthemum almeriense Pau C 0.5 0.5 0.3 0.1 0.1 Glaucium flavum Crantz C 0.4 0.2 0.3 0.1 0.1 0.2 0.1 0.1 Eryngium campestre L. C 0.3 0.6 0.3 0.1 0.1 0.8 0.5 Lavandula multifida L. C 0.2 0.3 0.5 0.6 0.3 0.1 Launaea arborescens (Batt.) Murb. C 0.2 0.5 0.1 0.4 0.2 0.1 Ruta angustifolia Pers. C 0.2 0.1 0.2 0.1 Anthyllis terniflora (Lag.) Pau C 0.1 0.2 0.15 0.3 0.1 Thymus hyemalis Lange C 0.1 0.3 0.1 0.2 0.1 Suaeda pruinosa Lange C 0.5 0.6 Hyparrhenia hirta (L.) Stapf H 2.4 1.8 1.9 1.5 2.7 0.1 0.1 0.1 0.2 0.2 Stipa tenacissima L. H 0.9 0.8 0.5 0.6 0.3 Piptatherum miliaceum (L.) Cosson H 0.8 0.9 1.1 1.3 1.9 0.1 0.1 0.3 0.3 0.5 Asparagus stipularis Forsk. L 0.2 0.1 0.5 Calendula arvensis L. A 0.9 0.5 0.4 0.1 0.2 0.6 Lobularia meritima (L.) Desv. A 0.56 0.65 0.84 0.11 0.5 0.1 0.3 0.6 0.7 0.6 Oxalis pes-caprae L. A 0.5 0.75 0.67 0.61 0.6 0.31 0.2 0.3 0.2 Species richness 47 47 45 46 46 22 27 30 28 31

Life forms: T, tree; S, shrub; C, chamaephyte or hemicryptophyte; H, herb; L, liana; A, annual.

Significant differences were found between the shrub Comparisons of species presence between undisturbed cover of the natural and restored plots in all years com- and restored plots were made using the mean of the pared (Fig. 2). In the herbs the same occurred. In the group Bray-Curtis index at two sites (Table 4) and showed a of chamaephytes and hemicryptophytes the differences ap- slight though progressive increase over the years, from peared only during the first 2 years studied (1991 and 1993). 0.40 in 1991 to 0.59 in 1999. Comparisons between un- In the remaining groups no differences appeared in cover. disturbed and restored plots, by life forms and by sites During the first year (1991) significant differences (Table 5), show a progressive increase in the similarity were found between natural and restored plots in the in shrubs, chamaephytes, and hemicryptophytes at both specific frequency of the trees, of the chamaephytes and sites. The herbaceous group shows a different trend at hemicryptophytes, and of the lianas (Fig. 3). In chamae- both sites (Tables 2 and 3). At site 1 the similarity in- phytes and hemicryptophytes significant differences also creases in the first half of the study period and after- appeared in 1993. Among the lianas these differences ward diminishes and stabilizes; at site 2 it increases pro- persisted throughout all the years analyzed. Where sig- gressively over the study period. The lianas could be nificant differences were found, the trees of the restored compared only at site 1, where it presented a slight in- plots registered a frequency greater than in the undis- crease in similarity in the two last periods. The annuals turbed plots, whereas in the other groups frequency show a trend to decrease in similarity at site 1; at site 2 proved greater in the undisturbed plots. the pattern is quite irregular, with no identifiable trend.

DECEMBER 2002 Restoration Ecology 699 Plant Restoration in Semiarid Summer-Dry Riverbeds

Figure 3. Plant frequency by life form in the undisturbed and restored plots during the different years studied, means of site Figure 2. Plant cover by life form in the undisturbed and re- 1 and site 2. *p 0.05; ***p 0.001. stored plots during the different years studied, means of site 1 and site 2. *p 0.05; ***p 0.001.

Discussion ulations of such herbivores, primarily rabbits, are very numerous, selectively consuming certain species. In ad- Among the reintroduced species Chamaerops humilis and dition, there was a strong attack on introduced plants of Ficus carica showed negative results. Problems concern- Retama sphaerocarpa. This appears to be the reason why ing these species appear to be linked to intense her- this species showed no increase in cover over the study bivory by small mammals. In the restored area the pop- period, despite its reintroduction being successful. The

700 Restoration Ecology DECEMBER 2002 Plant Restoration in Semiarid Summer-Dry Riverbeds

Table 4. Similarity comparisons using mean of Bray-Curtis teristic in common, their ruderal nature. The high fre- index for undisturbed (U) and restored (R) plots, using data quency of these species after planting could be ex- pooled from sites 1 and 2. plained by the heavy rainfall in 1996 and 1997, which Year R-91 R-93 R-95 R-97 R-99 heavily disturbed these communities. However, this phenomenon occurred both in the natural and restored U-91 0.40 U-93 0.47 plots. Therefore, mineralization was possibly greater, U-95 0.52 and a subsequent increase in available nitrates and am- U-97 0.55 monium in the soil was a consequence of the perturba- U-99 0.59 tion caused by the working of the soil for planting (Zink et al. 1995). This phenomenon has been observed in other disturbed arid communities (Grantz et al. 1998). species of Tamarix were not attacked by mammal herbi- Among the most frequent spontaneous shrubs was vores, probably because of the high salt content (partic- Nicotiana glauca. It was accidentally introduced from ularly sodium ions) in the foliar tissues of these plants. South America (Juscafresa 1975; López 1988b). This spe- Similarly, the presence of oleandrina and flavons in the cies invades river beds and banks of the most disturbed tissues of Nerium oleander, being highly toxic, discour- stretches of these courses. For the next study period we aged herbivory (López 1988b). might expect a progressive decline in ruderal species as That plantings were done after heavy rains appeared the community cover becomes denser. The response of to have notable influence on establishment. The follow- N. glauca is difficult to predict, but given its competitive ing year was rainy, particularly the first few months of strength against planted species, it may become a stable 1992. Nevertheless, this precipitation did not produce component of the restored community, and its elimina- flash floods (characteristic of these watercourses; Smith tion may then be considered. et al. 1998), which normally would have disturbed the The frequent presence of Arundo donax (reed) is due to plantings; therefore rooting of the plants was allowed. the introduction of this Asiatic species (Rivera & Obón In the following periods the years were rainy (1993– 1991) by humans to use in hedges and to stabilize river- 1994), encouraging growth of plantings. banks (Salinas 1995). This introduction appears to have This study did not reveal the origin of the spontane- occurred in the Mediterranean Basin at least by the first ous species, but most of the new species appeared to century A.D. (Font Quer 1981). At present, though its originate from the dispersion by plants near the study growth has not increased, it has naturalized and consti- area, because the seed bank is apparently not vital to ri- tutes one more element in these communities and thus parian sites subject to flowing water (Malanson 1993). is difficult to eliminate. As Aronson et al. (1993) defined, The spontaneous species and those present from the for ecosystems subjected to long periods of prolonged initial period that increased their cover have a charac- human disturbance, restoration, in the narrowest sense,

Table 5. Similarity comparisons by life forms and by sites using mean of Bray-Curtis index for undisturbed (U) and restored (R) plots in shrubs and chamaephytes and hemicryptophytes.

Site 1 (Amoladeras Rambla) Site 2 (Agua Rambla) R-91 R-93 R-95 R-97 R-99 R-91 R-93 R-95 R-97 R-99 n 13 n 10 Shrubs U-91 0.42 0.55 U-93 0.45 0.64 U-95 0.53 0.66 U-97 0.59 0.67 U-99 0.61 0.75 n 21 n 18 Chamaephytes and hemicryptophytes U-91 0.24 0.03 U-93 0.33 0.07 U-95 0.40 0.21 U-97 0.45 0.22 U-99 0.54 0.31

n, number of species.

DECEMBER 2002 Restoration Ecology 701 Plant Restoration in Semiarid Summer-Dry Riverbeds aims toward the complete return of a site to a preexist- Ecological census techniques: a handbook. Cambridge Uni- ing state in taxonomic terms, the re-establishment of all versity Press, Cambridge, United Kingdom. indigenous species and the extirpation of all exotic ones. Carothers, S. W., R. R. Johnson, and S. W. Aitchison. 1974. Popu- lation structure and social organization of the southwestern Recently, it was shown that ecosystem alteration may riparian birds. American Zoologist 14:97–108. commonly be caused by invasive non-indigenous spe- Font Quer, P. 1981. Plantas medicinales. Editorial Labor, Barce- cies (Gordon 1998). If restoring ecosystem functioning is lona, Spain. sought, both Nicotiana and Arundo must be eliminated. Gordon, D. R. 1998. Effects of invasive, non-indigenous plant spe- The restored communities are far from equaling the cies on ecosystem processes: lessons from Florida. Ecological Applications 8:975–989. percentage of cover, frequency, and species richness of un- Grantz, D. A., D. L. Vaughn, R. J. Farber, B. Kim, L. Ashbaugh, T. disturbed areas. Groups such as lianas have very scarce VanCuren, R. Campbell, D. Bainbridge, and T. Zink. 1998. representation compared with undisturbed plots. Non- Transplanting native plants to revegetate abandoned farm- ruderal species that were not restored did not appear lands in the western Mojave Desert. Journal of Environmen- spontaneously, except a few in the last years analyzed. The tal Quality 27:960–967. Gregory, S. V., F. J. Swanson, W. A. McKee, and K. W. Cummins. time required to reach a state approaching seminatural is 1991. An ecosystem perspective of riparian zones. BioScience greater than that needed to establish the dominant species. 41:540–551. Our results show that the successful restoration of ri- Johnson, R. R., and J. M. Simpson. 1971. Important birds from parian vegetation in semiarid zones is relatively easy, but Blue Point cottonwoods, Maricopa County, Arizona. Condor slow. The loss of species by unforeseen causes, such as 73:379–380. Juscafresa, B. 1975. Flora medicinal, tóxica, aromática y condi- herbivory, should be taken into account. Satisfactory mentaria. Editorial Aedos, Barcelona, Spain. short-term results are attainable, fundamentally with re- Le Houèrou, H. N. 1982. The arid bioclimates in the Mediterra- spect to species diversity and the establishment of nean isoclimatic zone. Ecologia Mediterranea 8:103–114. planted species. During the first few years, an increase López, F. 1988a. Corrección de torrentes y estabilización de cauces. was found in ruderal species and a weak development in Food and Agricultural Organization (FAO), Rome, . López, G. 1988b. La guía de Incafo de los árboles y arbustos de la non-ruderal ones. It is necessary to continue monitoring Península Ibérica. Editorial Incafo, Madrid, Spain. the restoration effort to determine whether the non-rud- Malanson, G. P. 1993. Riparian landscapes. Cambridge Univer- eral species establish themselves spontaneously in the sity Press, Cambridge, United Kingdom. community. If we seek to reach a high degree of restora- Manteiga, L. 1992. Conservación y gestión de los cursos fluviales tion, with the elimination of exotic flora, then Nicotiana en la España peninsular. Quercus 76:36–43. Martín-Vivaldi, M. E. 1991. Estudio hidrográfico de la “Cuenca Sur” glauca and Arundo donax should be eliminated, although de España. Servicio de publicaciones de la Universidad de this is debatable and technically difficult to achieve. Granada/Confederación Hidrográfica del Sur, Granada, Spain. Mota, J. F., J. Cabello, M. Cueto, F. Gómez, E. Giménez, and J. Peñas. 1997. Datos sobre la vegetación del sureste del Alm- Acknowledgments ería (Desierto de Tabernas, Karst en yesos de Sorbas y Cabo de Gata). Universidad de Almería, Almería, Spain. This research is part of the project “Study and Regenera- Mueller-Dombois, D., and H. Ellenberg. 1974. Aims and methods tion of Forest Systems and Resources in the Ramblas of of vegetation ecology. John Wiley & Sons, New York. Peinado, M., F. Alcaraz, and J. M. Martínez. 1992. Vegetation of the arid SE Iberian Peninsula” (CICYT, FOR91-0632), one southeastern Spain. Flora et Vegetatio Mundi Band X. J. of the aims of which is to further our knowledge of the re- Cramer, Berlin, Germany. lationships between the many factors involved in such en- Pesson, P. 1978. Ecología forestal. Editorial Omega, Madrid, Spain. vironments and the vegetation. We thank the Delegación Raunkiaer, C. 1934. The life forms of plants and statistical plant provincial de Almería de la Consejería de Medio Ambi- geography. Clarendon Press, Oxford. Rivera, D., and C. Obón. 1991. La guía de Incafo de las plantas ente for permitting us to monitor the plantings performed útiles y venenosas de la Península Ibérica y Baleares (ex- by this institution in the Parque Natural del Cabo de Gata- cluídas medicinales). INCAFO, Madrid, Spain. Níjar. We also thank David Nesbitt for his assistance with Salinas, M. J. 1995. Estudio y regeneración de las comunidades the English language version of the text. forestales riparias en el sureste semiárido peninsular. Servicio de publicaciones de la Universidad de Granada. Granada, Spain. Salinas, M. J., G. Blanca, and A. T. Romero. 1999. Análisis compara- LITERATURE CITED tivo de dos comunidades vegetales riparias de cuencas semi- áridas del Sureste Ibérico (España). Boletín de la Real Sociedad Alcaraz, F., T. E. Díaz, S. Rivas-Martínez, and P. Sánchez-Gómez. Española de Historia Natural (Sección Biología) 95:43–56. 1989. Datos sobre la vegetación del sureste de España: pro- Smith, S. D., D. A. Devitt, A. Sala, J. R. Cleverly, and D. E. Busch. vincia biogeográfica Murciano-Almeriense. Itinera Geobo- 1998. Water relations of riparian plants from warm desert re- tanica 2:5–133. gions. Wetlands 18:687–696. Aronson, J., C. Floret, E. Le Floc’h, C. Ovalle, and R. Pontanier. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall. Englewood 1993. Restoration and rehabilitation of degraded ecosystems Cliffs, New Jersey. in arid and semiarid lands. I. A view from the south. Resto- Zink, T. A., M. F. Allen, B. Heindl-Tenhunen, and E. B. Allen. ration Ecology 1:8–17. 1995. The effect of a disturbance corridor on an ecological re- Bullock, J. 1996. Plants. Pages 111–138 in W. J. Sutherland, editor. serve. Restoration Ecology 3:304–310.

702 Restoration Ecology DECEMBER 2002