Thèse-EliseBuisson-Chapitre2

Elise Buisson

Description

Chapter 2 Vegetation dynamics in La Crau Influence of former cultivation on the unique Mediterranean steppe of France and consequences for conservation management. 1* 2 1 3Christine Römermann , Thierry Dutoit , Peter Poschlod and Elise Buisson 1 Institute of Botany, University of Regensburg Universitätsstrasse 31 D-93051 Regensburg, Germany 2 UMR INRA-UAPV 406, Écologie des Invertébrés Site Agroparc 84914 Avignon France 3 Institute of Mediterranean Ecology and Paleoecology, UMR/CNRS 6116 University P. Cézanne, FST Saint Jérôme, case 462 13397 Marseille Cedex 20 France Article published in Biological Conservation in Jan. 2005. Abstract In Europe, the actual landscape has been mainly influenced by human activities. Agricultural intensification led to a considerable habitat loss and fragmentation, especially for dry semi-natural grasslands. This current study investigates the impact of former melon and cereal cultivation (cultivation period: 1950–1987) on the semi-natural vegetation of the Crau, representing the last xeric Mediterranean steppe in France. Today, the ex-cultivated melon and cereal fields are characterised by different vegetation compositions, species richness and evenness compared to the undisturbed steppe community. Also the abiotic conditions (N, P, K, pH, soil granule fractions) have been changed by former cultivation practices. The rather transient seed bank of the steppe was depleted during the cultivation periods ...

Informations

• Publié par: Miphing
• Nombre de lectures: 23
• Langue: English

Extrait

Chapter 2

Vegetation dynamics in La Crau Influence of former cultivation on the unique Mediterranean steppe of France and consequences for conservation management.   Christine Römermann 1* , Thierry Dutoit 2 , Peter Poschlod 1 and Elise Buisson 3   1 Institute of Botany, University of Regensburg Universitätsstrasse 31 D-93051 Regensburg, Germany 2 UMR INRA-UAPV 406, Écologie des Invertébrés Site Agroparc 84914 Avignon France 3 Institute of Mediterranean Ecology and Paleoecology, UMR/CNRS 6116 University P. Cézanne, FST Saint Jérôme, case 462 13397 Marseille Cedex 20 France Article published in Biological Conservation in Jan. 2005 . Abstract In Europe, the actual landscape has been mainly influenced by human activities. Agricultural intensification led to a considerable habitat loss and fragmentation, especially for dry semi-natural grasslands. This current study investigates the impact of former melon and cereal cultivation (cultivation period: 19501987) on the semi-natural vegetation of the Crau, representing the last xeric Mediterranean steppe in France. Today, the ex-cultivated melon and cereal fields are characterised by different vegetation compositions, species richness and evenness compared to the undisturbed steppe community. Also the abiotic conditions (N, P, K, pH, soil granule fractions) have been changed by former cultivation practices. The rather transient seed bank of the steppe was depleted during the cultivation periods; ancient weed species and ruderals now determine the seed bank of the ex-cultivated fields. It is concluded that the conservation of the last parts of undisturbed steppe must have absolute priority. A re-development of the original and unique

 

49

steppe community on formerly cultivated fields may take decades or centuries, if at all. Keywords :   Crau; dry grassland; restoration ecology; seed bank; secondary succession Introduction   Within the last few decades, agricultural intensification led to a considerable floristic change and decrease of semi-natural ecosystems in Europe (Hodgson & Grime 1990; Akinola et al. 1998; Poschlod et al. 1998; WallisdeVries et al. 2002). Many formerly continuous types of grassland have become extinct or fragmented, and in most cases only small and isolated patches of ancient remnants remain. In such fragmented landscapes, restoration of ex-arable fields on ancient grasslands plays a major role as restored sites could enlarge and connect these remnants. The restoration depends on the abiotic (e.g. soil fertility) and biotic conditions of the abandoned arable fields (e.g. modified competitive interactions; Saunders et al. 1991) and on temporal and spatial dispersal abilities of the characteristic species of the target community (Poschlod et al. 1998). In north-western Europe, the short-term (Austrheim & Olsson 1999) and long-term (Wells et al. 1976) effects of ploughing on present-day dry calcareous vegetation were clearly identified. Ex-cultivated plots showed a decrease in species-richness and changes in their botanical composition compared with adjacent dry grasslands. Cultivation, however, also influences the direction taken by post-cultural plant succession (Gibson & Brown 1991). This phenomenon is linked to increased soil fertility of ex-cultivated plots because of ancient fertiliser applications during the cultivation period (Gough & Marrs 1990) and the presence of ruderal species (sensu Grime 1979), which are known to have persistent seeds in the soil (Graham & Hutchings 1988a,b; Dutoit & Alard 1995; Dutoit et al. 2003). However, also seeds of target species might survive the cultivation period in deeper soil layers, providing the fact that ploughing were not too deep (Bekker et al. 1997). Hence, it is of great importance, to what extent such a memory of the ancient vegetation exists (Bakker et al. 1996b), and if it could act as a source of propagules in the restoration of fallow lands (Willems & Bik 1998). However, most of the characteristic grassland species do not form persistent soil seed banks (Thompson & Grime 1979; Graham & Hutchings 1988a,b; Bakker et al. 1996a), and would, therefore, not survive cultivation periods. The restoration value of dry grassland seed banks is, therefore, very small (Dutoit & Alard 1995; Bakker et al. 1996b). Hence, successful reestablishment of the target species depends on the function of dispersal vectors as for example sheep grazing (Marshall & Hopkinson 1990; Fischer et al. 1996) since the species own dispersal capacities in space are very low (Poschlod et al. 1998). The introduction of sheep herding would also positively influence vegetation dynamics by limiting competition between stress tolerant target species and  50

ruderal ancient weed species (Gibson & Brown 1991), especially in those ecosystems where plant communities have been determined by sheep grazing for many centuries. Though past agricultural land uses have affected the composition and structure of the dry herbaceous ecosystems throughout the landscape around the Mediterranean Sea (Grove & Rackham 2001), effects of former ploughing on present day vegetation have received little attention in this region compared to north-western Europe. Formerly ploughed sites are often difficult to locate because they are masked by the effects of a high number of wild fires, homogenising the vegetation during secondary succession following abandonment of cultivation (Trabaud & Galtie 1996). The Crau represents the last xeric steppe of south-eastern France (Devaux et al., 1983) which has been shaped by sheep grazing for centuries (Rinschede 1979; Fabre 1998). The steppe vegetation represents a plant species association which is unique in Europe (Devaux et al. 1983). It provides a habitat for many endangered animal species, e.g. for the endemic grasshopper Prionotropis hystrix rhodanica  (Foucart & Lecoq 1998) and the rare bird species Pterocles   alchata  (sandgrouse; Wolff 1998). However, destruction by agricultural practices started in the 16th century, when irrigation systems were brought up and the northern parts were transformed into intensively used grasslands. At the beginning of the twentieth century, large parts of the dry semi-natural grasslands (called coussous) have been transformed into arable, industrial or military land, leading to the fragmentation of the former 60 000 ha steppe vegetation (Etienne et al. 1998). After abandonment of most arable fields in the 1980s, traditional transhumance has been reintroduced, connecting ex-cultivated fields with undisturbed steppe vegetation (Dureau & Bonnefon 1998). Hence, the aims of the current study were: (i) to describe the vegetation of coussous, ex-cultivated melon and ex-cultivated cereal fields and (ii) to investigate the impact of historical cereal and melon cultivation on the biotic and abiotic conditions of the ex-cultivated fields (e.g. species richness, evenness, soil chemistry) and (iii) to investigate the impact of historical cultivation on the seed bank of the steppe plant community, and its possible contribution for the restoration of the ex-cultivated fields. The general aim from a conservational point of view is, therefore, to find out if a re- development of the original coussous vegetation is possible on arable fields after abandonment.

 

51

Materials and methods Study area The study was carried out in the Crau, located about 50 km north-west of Marseille, France (Fig. 12). The plain is an old, stony river delta which was shaped by several changes of the Durance river bed between 650 000 and 120 000 years ago (Devaux et al. 1983). The dryness of the ecosystem is induced by the Mediterranean climate (mean temperature 14.5 °C, yearly precipitations of 500600 mm; Cherel 1986); precipitation does rarely occur within the vegetation period. Soil properties are exceptional: a water impermeable conglomerate in 4060 cm depth (called locally taparas) separates the ground water from the pebbly soil (Devaux et al. 1983). While the steppe plant community of central Crau is determined by Brachypodium   retusum  Poaceae), Thymus vulgaris  (Lamiaceae) is the dominant plant species in the western regions (Loisel et al. 1990). Between 1950 and 1960 relatively moderate cereal cultivation (without deep ploughing) and fertilisation took place (Devaux et al. 1983). Melon cultivation (1965 until now) was more intensive as it destroyed the taparas by deep ploughing, simplifying irrigation afterwards (Borck 1998). Furthermore, huge amounts of fertilisers and pesticides were applied (Le Gloru 1956). Before 1975, melon cultivation took place in small plastic tunnels (1.7 m width, 0.7 m height; Borrey 1965). Afterwards, large plastic tunnels were used (3 m width, 2 m height), leaving always track ways of about 2 m between the tunnels. At most sites, cultivation stopped around 1990 because of melon overproduction, non-profitable cultivation and hard working conditions. Study sites As vegetation composition of central and western regions of the Crau is different, four study areas in central Crau (Magnan, Valigne, Peau de Meau, Pt. Abondoux) and three areas in western Crau (Gamadou, Negreiron, Négrès) were chosen (Fig. 12). The sampling areas provide always an ex-cultivated field with adjacent undisturbed steppe. At Pt. Abondoux, only one ex-cultivated cereal field was sampled. In total six plots of coussous, six ex-cultivated melon fields and three ex-cultivated cereal fields were sampled. At each site, the same flock of sheep was grazing both steppe and ex-cultivated plots. With the help of documentary evidence and aerial photographs (1947 1998, with gaps), the histories of the sampling sites were reconstructed (Table 2). Vegetation and soil sampling Vegetation and soil analyses were carried out in the centre of all study sites. For vegetation analyses, 400 m 2  (20m ×  20 m) of each site were  52

sampled with 10 randomly placed quadrats of 2 m × 2 m, divided into 25 sub- quadrats of 0.16 m 2  using the frequency method (Mueller-Dombois & Ellenberg 1974). Percentage cover of vegetation, stones and bare ground were surveyed.  

Peau de Meau Gamadou Magnan Negreiron Pt. Abondoux F r a n c e Négrès Valigne

Crau

N   Fig. 12 Map of the Crau (according to Devaux et al., 1983, modified). Western (Gamadou, Negreiron, Négrès) and central (Peau de Meau, Magnan, Pt. Abondoux, Valigne) study sites are marked. For soil analyses, five replicate samples (200 g each) were taken per site. After drying and sieving (200 mm), nutrient analyses (C, N, P 2 O 5 , CaO, MgO, K 2 O, O.M., C:N, pH) were carried out by the INRA (Institut National de la Recherche Agronomique) using the methods illustrated in Baize (2000). Percentage content of clay (<0.002 mm), fine silt (0.0020.02 mm), coarse silt (0.020.05 mm), fine sand (0.050.2 mm) and coarse sand (0.22 mm) were investigated at the University of Aix-Marseille III (Baize 2000). Seed bank sampling and analysis At the site of Peau de Meau (central Crau), the seed bank of the steppe plant community (PC), the ex-cultivated cereal field (PFc1) and the ex- cultivated melon field (PM) were sampled in April 2001 before the input of fresh seeds and after the natural stratification of dormant seeds in winter (persistent viable seed bank).  53

For each land-use type, 10 replicate soil samples were taken consisting each of 10 pooled soil cores according to the recommendations in Bakker et al. (1996a). After taking the soil cores (4 cm in diameter, 20 cm depth), they were divided into two depth layers (010 and 1020 cm) in order to account for the cultivation impact on seed distribution in the soil (ploughing regimes, up to 2040 cm depth!). The soil of the two depth layers was pooled, respectively, to one sample each of 1256 cm3 (1.26 litre) per replicate and depth. Table 2. Cultivation history of the sampled sites of Central and Western Crau.     Sites Sampled plotsAbbr. Cultivation type  ar T a i b m l e e     p c e ul r t i i o v d a   ti o o f  n  Coussous MC - - a nan Mg Ex m - e c lo ult n i   v fi a e t l e d d     MM  Me t l u o n n n s  e ( l l s a ) r   ge  1993-98  Coussous VC - - Valigne  Ex-c l ult n i   v fi ated  VM  M meoeld  e t l u o n n n s  e ( l l s a ) r   ge  1978, 1984  Central Coussous  PC - - Crau Peau de  Ex m - e c lo ul n ti  f v i a e t l e d  d  PM Melons 1967-1984  E Meau  c x e - r c e u a lt l i  f v i a e t l e d  d  PFc1 Cereals 1948-64  E c x e - r c e u a lt l i  f v i a e t l e d d  PFc2 Cereals 1974-75    Petit Ex-cultivat l e d d  AFc Mels, Cereals 1979-1984 Abondoux cereal fie  on Coussous GC - - Gamadou Ex-cult rge  198 melon i  f v i a e t l e d d     GM  Me t l u o n n n s  e ( l l s a )  6  Coussous NC - - W C e r s a te u r   n  Negreiron  Exultivated  1978/79  m - e c lon field  NM  Me t l u o n ns n   e ( l l s a ) r   ge  Négrès Coussous BRC - - ( R B o e m rg a e in ri e e )     E m x e - l c o u n lt   i fi v e a l t d e s d     BRM  Me t l u o n n n s  e ( l s s) m   all  1975, 1982/83  54

 

For seed bank cultivation, the seedling emergence method (ter Heerdt et al. 1997) was chosen to qualify and quantify germinable seeds in the soil seed bank. Soil samples were concentrated by washing them with water on two different sieves (2 mm and 200 lm) to reduce bulk and clay. The concentrated soil was spread in a thin layer (0.5 cm) in plastic trays filled with 2 cm vermiculite topped with medical compresses (mesh size about 100 lm). The trays were covered with fine gauze material to prevent contamination. All trays were watered frequently from below. Emerged seedlings were identified (Muller 1978; Marmarot 1997), counted and removed weekly. Unknown seedlings were grown for identification. The seed bank samples were cultivated for in total six months including a six week chilling period after the third month to break secondary dormancy of the seeds. Data analyses Vegetation and soil analyses A detrended correspondant analysis (DCA) was carried out on the raw vegetation data (PC-ORD 4.0, McCune & Mefford 1999). Plots and species frequencies were ordinated and, to present a bi-plot, species were correlated with it. Comparisons of vegetation and soil parameters were carried out on different levels:  for western and central sites (coussous and ex-cultivated melon fields with 30 quadrats each),  for the three land-use types (coussous, ex-cultivated cereal fields and ex-cultivated melon fields with 30 quadrats each),  for all different plots (for each site and each vegetation type with 10 quadrats each). The parameters species richness and evenness were calculated (formulas see Mühlenberg 1993). These two parameters and vegetation, stone and bare ground cover, soil nutrients and granule sizes were compared on the different levels using univariate tests. Normal distributed data with homogeneous variances were analysed by single-factor ANOVAs with subsequent post-hoc LSD tests (least significant difference). The remaining data were compared by KruskalWallis H tests with following MannWhitney U tests. For the comparison of only two data sets, pair-wise t tests were applied. For all statistical tests, SPSS Inc. (Release 10.0) for Windows was used. Combination of vegetation data with abiotic parameters To identify the factors which have influenced the species compositions of the various plots, a principal correspondent analysis was carried out. Vegetation data was the main matrix; the standardised second matrix contained the parameters soil chemistry, soil granule fractions, vegetation, stones and bare ground cover, time of abandonment, time period of land use, cultivation types (cereals, melons, none) and melon cultivation regime (small vs. large tunnels).  55

Seed bank analyses Total number of species and seeds were counted and recorded for each plot. Mean number of species per sample and mean number of seeds p s e a r m   p s l p e e   c o i f e   s  12 w 6 e   re c   m c 2 alc w u e l r a e t   ed c   al a c n u d l   at c e o d  m a p n ar d e   d s . t   at Mean number of seeds per   istically compared. For the different depths levels (020, 010 ample were extrapolated to 1 m 2  and com a p n a d r   e 1 d 0 .   F 2 o 0 r     c al m l  ) st s a e ti e s d ti   c n a u l  m c b o e m rs p   a p ri e s r  o s ns, the  same methods of analyses were applied as described above using SPSS 10.0. Evenness was calculated for each plot. Combination of seed bank data with vegetation data To compare seed bank composition with vegetation composition, a DCA was carried out on a power transformed matrix containing the combined data of seed bank and established vegetation. Results Vegetation attributes In the bi-plot (Fig. 13) the 30 quadrats (4 m²) of western and central coussous are clumped, respectively, confirming different species compositions of western and central undisturbed steppe vegetation. Differently, the quadrats (4 m2) of central and western ex-cultivated melon plots are mixed. The ex-cultivated cereal plots PFc2 and AFc have intermediate positions between coussous and ex-cultivated melon plots. Species being strongly correlated with the described plot (r 2  = 0.55) are presented as vectors pointing to the corresponding direction. Neatostema apulum, Trifolium scabrum, Sherardia arvensis and Thymus vulgaris  are characteristic for western coussous. Representatives for the central coussous are B. retusum and Linum strictum . None of the typical species of the ex-cultivated plots were strongly correlated with the axes. However, Bromus madritiensis, B. rubens, Rostraria cristata, Lobularia maritima and Calamintha nepeta were very abundant in the ex-cultivated plots; they were dominant species in ex-cultivated melon fields. Characteristic and dominant species in ex-cultivated cereal plots were Aegilops ovata , Trifolium subterraneum, Hedypnois cretica  and Geranium molle . Species richness and evenness Concerning species richness (S) and evenness (E) the western coussous had significantly higher values than central coussous (S west  = 45.8 vs. S centre  = 36.1; U S  = 199.5, p < 0.001; E west  0.91 vs. E centre  = 0.88; U E  = = 183, p < 0.001). Significant differences between western and central ex- cultivated melon plots were not detected (S west  = 29.97 vs. S centre  = 32.53; E west = 0.86 vs. E centre = 0.87). Significant differences were found when  56

MC, PC, VC

Axis 2

Legend : ▲ coussous Centre (MC, PC, VC) Δ coussous West (GC, NC, BRC) Brachypodium retusum   Linum strictum   ● ex-cultivated melon fields Centre BRC (MM, PM, VM) PFc2, GC, NC, ○ ex-cultivated melon fields West MM, VM,  AFc Thymus vulgaris (GM, NM, BRM) GM, NM  ex-cultivated cereal fields Centre Axis 1 (PFc1, PFc2, AFc) Neatostema apulum Sherardia arvensis Trifolium scabrum

PM BRM  PFc1 Fig. 13 DCA ordination of the established vegetation of Central and Western Crau. Matrix: 160 species, 150 quadrates of 15 sites (see Table 2 for abbreviations of the study sites; first letter: site; second/third letter: land use type C = Coussous, M =ex- cultivated melon field; Fc = ex-cultivated cereal field)); r²-cut-off-value= 0.55. Species correlated with axis 1: Trifolium scabrum (r²= 0.698), Sherardia arvensis (r²= 0.574), Neatostema apulum (r²= 0.573) and Thymus vulgaris (r²= 0.55). Species correlated with axis 2: Brachypodium retusum (r²= 0.748) and Linum strictum (r² 0.664). = comparing all studied plots ( χ 2 S = 123.09, p < 0.001; F E = 20.451, p < 0.001; see Table 3). Generally, undisturbed steppe vegetation had higher values in both species richness and evenness (Table 3). Abiotic factors influencing vegetation dynamics Vegetation, stone and bare ground cover Vegetation and stone cover were higher in central compared to western coussous (vegetation: 55% vs. 49%, T = 2.04, p < 0.05; stones: 50% vs. 15%, T = 19.2, p < 0.001). Bare ground cover was higher in western coussous (39% vs. 11%, T = -0.918, p < 0.001). Western and central ex- cultivated melon plots showed equal values for stone and bare ground cover (stones: 44.87% vs. 44.88%; bare ground: 19.65% vs. 22.96%). In vegetation cover, western ex-cultivated melon plots had even higher values than central ones (43.4% vs. 39.5%; T = -2.1, p < 0.05). As shown in Fig. 14., melon and cereal cultivation resulted in a modification of vegetation (F = 19.11, p < 0.001), stone (F = 47.0, p < 0.001) and bare ground cover (F = 25.24, p < 0.001).  57

Table 3. Mean species richness (S) and mean species evenness (E) of the different sites. Maximums (>45 species) are bold; minimums (<20 species) are bold and cursive.  Site S E GC 48.2 + 0.76 0.92 + 0.00 Western coussous  NC 46.7 + 1.02 0.90 + 0.01 BRC 42.5 + 0.95 0.91 + 0.00   MC 47.5 + 0.93 0.90 + 0.00 Central coussous  VC 37 0.89 0.90 + 0.01 .2 + PC 23.5 + 1.44 0.84 + 0.01 GM 30.9 + 1.53 0.85 + 0.01 Western 40.7  + 1.81 0.88 + 0.00 ex-cultivated melon plots  NM BRM 18.3 + 1.83 0.84 + 0.01 MM 45.2 + 2.17 0.88 + 0.00 melon plots Central ex-cultivated  VM 38.8 + 1.9 0.88 + 0.01 PM 13.6 + 0.43 0.83 + 0.01 PFc1 40.9 + 0.72 0.89 + 0.00 lots Central ex-cultivated cereal p  PFc2 32.9 + 1.02 0.87 + 0.01 AFc 36.4 + 1.35 0.89 + 0.00

b b a

70 60 a a a 50 c b 40 b 30 20 10 0 % vegetation cover % stone cover % bare ground cover coussous ex-cultivated cereal plots ex-cultivated melon plots Fig. 14 Percentage cover of vegetation, stones and bare ground each compared between coussous, ex-cultivated cereal and ex-cultivated melon fields. Different letters indicate significant differences within each block.  58

Soil chemistry and soil physics As shown in Table 4, western coussous had higher nitrogen, phosphorous and potassium contents compared to central coussous. The central coussous were richer in magnesium and C:N content. Regarding the physical soil conditions, the western coussous had higher contents of fine sand and the central coussous was superior in the percentage of clay and coarse sand content. These differences between central and western sites have been disappeared after melon cultivation; only nitrogen and carbon content and percentage coarse sand were different for central and western ex-cultivated melon fields (Table 4). Coussous, ex-cultivated cereal and melon fields have significantly different phosphorous and potassium contents, C:N ratios and pH-values (Table 5). Compared to the coussous, the soil of both ex-cultivated cereal and melon plots had increased values in pH and potassium content. The ex- cultivated melon fields had elevated phosphorous contents. C:N ratio was lowest in ex-cultivated fields; the ex-cultivated melon fields had higher values than the ex-cultivated cereal fields. Table 4. Results of chemical and physical soil analyses of central and western coussous and central and western ex-cultivated melon plots. Means and standard errors and the test values of t-tests (T) or Mann-Whitney-U-tests (U) and the p-values (* P < 0.05, ** P < 0.01, *** P < 0.001, n.s. not significant) are given.  Coussous  centre west statistics Nitrogen (g/kg) 1.29 + 0.05 1.54 + 0.06  T=  * 3 ** .21    Phosphorous (g/kg) 0.005+0.002 0.01 + 0.00  U=  * 4 * 0.5    Potassium (g/kg) 0.09 + 0.01 0.15+ 0.01  U= *   * 1 * 5.5    Magnesium (g/kg) 0.19 + 0.01 0.16 + 0.01  T=  * - * 2.98    = C : N 9.48 + 0.08 8.54 + 0.23  U *   * 2 * 1.0    Carbon (g/kg) 12.28 + 0.62 13.10 + 12.28 T= n  1 s . . 0   3  . Fine sand (%) 21.96 + 0.56 24.94 + 0.71  T=  * 3 * .28    Coarse sand (%) 20.4 + 0.89 17.47 + 0.81  T= -2.43 *   Clay (%)  22.57 + 0.56 17.11 + 1.89  U=  * 61.0           

 

59