Thèse-EliseBuisson-Discussion

Elise Buisson

Description

General Discussion The discussion is organized following the introduction outline. I answer the basic questions set in the introduction by summarizing A) the results of experiments in La Crau B) the results of experiments in California coastal prairies C) I then draw conclusions for both sites together and detail how these results contribute to ecological theory and/or to ecological restoration. 1. Impact What is the impact of strong exogenous disturbances on Mediterranean herbaceous plant communities that have evolved with regular endogenous disturbances for centuries? A) Exogenous disturbances have a great impact on both Mediterranean herbaceous plant communities studied. Cultivation has a great impact on the steppe of La Crau which varies depending on cultivation intensity. Cereal cultivation represents a moderate degree of disturbance compared to melon on, and this is still observable decades after cultivation abandonment. Both types of cultivation changed a) the physical and chemical soil parameters; b) the vegetation composition, richness and structure; and c) the species composition of the actual seed bank (Fig. 41). a) changes in physical and chemical soil parameters: Cereal cultivation involved surface plowing, fertilization and some stone removal, and has had a moderate impact on soil parameters: it has increased pH, potassium, and clay. Melon cultivation involved deep plowing, fertilization, irrigation, treatments and stone removal, ...

Informations

• Publié par: Niseng
• Nombre de lectures: 33
• Langue: English

Extrait

General Discussion The discussion is organized following the introduction outline. I answer the basic questions set in the introduction by summarizing A) the results of experiments in La Crau B) the results of experiments in California coastal prairies C) I then draw conclusions for both sites together and detail how these results contribute to ecological theory and/or to ecological restoration.   1. Impact What is the impact of strong exogenous disturbances on Mediterranean herbaceous plant communities that have evolved with regular endogenous disturbances for centuries? A) Exogenous disturbances have a great impact on both Mediterranean herbaceous plant communities studied. Cultivation has a great impact on the steppe of La Crau which varies depending on cultivation intensity. Cereal cultivation represents a moderate degree of disturbance compared to melon cultivation, and this is still observable decades after cultivation abandonment. Both types of cultivation changed a) the physical and chemical soil parameters; b) the vegetation composition, richness and structure; and c) the species composition of the actual seed bank (Fig. 41). a) changes in physical and chemical soil parameters: Cereal cultivation   involved surface plowing, fertilization and some stone removal, and has had a moderate impact on soil parameters: it has increased pH, potassium, and clay. Melon cultivation involved deep plowing, fertilization, irrigation, treatments and stone removal, and has had a great impact on soil parameters: it has increased pH, phosphorus, potassium, clay, fine silt and coarse sand. b) changes in vegetation composition, richness and structure: As observed by Wells et al. (1976), Graham & Hutchings (1988a), Willems et al. (1993), Dutoit (1996) and Janssens et al. (1998), cultivation has strongly influenced vegetation composition by changing soil structure and fertility. We find that the vegetation of both types of abandoned fields is dominated by species like hordeaceus BromusandCalamintha nepeta   (ruderal competitor) which are classified as competitive-ruderal species by Grime 2001 (annual with rapid plant growth and maximized seed production) and BASECO database (BASe de données ECOlogiques; Gachet et al.  134

2005). Cultivation has homogenized vegetation at the landscape level: contrary to steppe patches whose plant composition depends on soil characteristics and stone coverage which are different in the north and in the south of the plain, abandoned melon fields have a relatively similar composition whether they are located in the south or the center of the plain. Plant composition also varies depending on the type of cultivation. While some steppe species are found on abandoned cereal fields (Crepissp.,   Dactylis glomerata, Polycarpon tetraphyllum, Trifolium scabrum), abandoned melon fields do not have steppe species and have the lowest species richness. However, melon cultivation in large plastic greenhouses (tunnel, 1980s) has left some steppe patches intact in between the greenhouse rows; they have functioned as refuges and are important in determining community composition and species richness after abandonment (Römermann et al. 2005). My results confirm the findings of Wells et al. (1976) and of Austrheim & Olsson (1999): floristic differences can be explained by differences in historical trajectories, such as cultivation time and type of cultivation.

R veg = 23.5 R sb = 20 N = 1.1 g/kg P = 0.005 g/kg K = 0.09 g/kg

R veg = 40.9 R sb  R = 24 veg = 13 6 N = 1.4 g/kg  . R sb = 14 P K     = =     0 0 . . 0 1 1 1     g g / / k k g g             N = 1.3 g/kg       P = 0.04 g/kg K = 0.17 g/kg     Fig. 41. Main values of the three fields studied in Peau de Meau (Chapter 3). Cereal cultivation occurred from 1960 to 1966. Melon cultivation occurred twice between 1967 and 1984. Rveg = vegetation species richness in 4m² (mean of ten 4m² quadrats). Rsb = total number of seeds in the seed bank for 24L of soil. N, P, K in soil. c) changes in species composition of the actual seed bank: Like in many other dry grasslands (Dutoit & Alard 1995; Bakker et al. 1996b; Jackel 1999), most steppe species have a transient seed bank,  135

probably because they have co-evolved with grazing which creates favorable conditions for the establishment of seedlings (Willems 1995). Conversely, the seed bank of abandoned fields is composed of species with persistent seeds: arable weeds, such asChenopodium album(maximum longevity observed: >660 years - Thompson et al. 1997), Kickxia elatine(>30 years)   andPortulaca oleracea (40 years) and ruderals, such asCalamintha nepeta(not in database), Lobularia maritima(5 years), Polygonum aviculare(>460 years), Senecio vulgaris(>5 years), and nigrum Solanum(>39 years).This result is consistent with those of Graham & Hutchings (1988a,b). The steppe species are not found in the seed bank of abandoned fields anymore because transient seeds respond negatively to disturbances such as burial or deep plowing, fertilization and pesticide applications (Marshall 1989). Arable weed species are not found in the above-ground vegetation because the conditions necessary for their germination and growth no longer exist (e.g. irrigation). Such disparity between the composition of the vegetation and the seed bank has been shown by other authors (Lavorel et al. 1991; Dutoit & Alard 1995). B) In California, exogenous disturbances occurring after post-European settlement have a great impact on Mediterranean herbaceous plant communities such as central coast prairies. Changes in the vegetation composition are due to various combinations of the following effects: changes in fire regimes, intensive cultivation, introduction of competitive exotic plant species, heavy cattle grazing and recurrent years of drought around 1862-64 (Heady et al. 1988). Direct comparison of changes in soil parameters is impossible because there are very few intact remnants and most of them are on serpentine soils that have very different characteristics. More recent intensive cultivation has a greater impact on coastal prairies (e.g. degradation of the seed bank) than extensive cultivation, such as hay cultivation in the Mission period (Hayes 2004). Coastal prairie vegetation may include Danthonia californica, Nassella pulchra, Deschampsia danthonioides, Deschampsia caespitosa, Festuca rubra and various wildflowers, such as Lasthenia californica, Holocarpha macradenia, Clarkia purpurea, Chorizanthe robusta, Plantago erecta, Viola pedunculata, Cammisonia ovata, Sidalcea malvaeflora. Common invasive species of degraded prairies includeAvena barbata, Brachypodium distachyon, Bromus hordeacus, Erodium botrys, Erodium cicutarium, Plantago lanceolata, Rumex acetosella, Trifolium subterraneaum, Vulpia myuros(Stromberg et al. 2001; Hayes 2004). C) Whether in La Crau or in California, exogenous disturbances have had a great impact on the herbaceous plant communities, and have changed the vegetation and seed bank composition and richness. These results corroborates other studies in different climates (Wells et al. 1976; Graham &  136

Endogenous disturbance = grazing + drought

Cultivat

Cereal field: R veg = 40.9 R sb 24 = ? Steppe: veg = 23.5 R sb = 20

Hutchings 1988a,b; Marshall 1989; Willems et al. 1993; Dutoit 1996). The vegetation developing on fields after cultivation is greatly influenced by the type of cultivation (Wells et al. 1976; Austrheim & Olsson 1999).                             1960-1966 Melon  field:   R veg = 13.6   R sb = 14   1967-1984   2005 Time   Fig. 42. Model showing the effects of a disturbance (cultivation) on the plant species richness of the steppe community. Cereal cultivation occurred from 1960 to 1966. Melon cultivation occurred twice between 1967 and 1984. Rveg = vegetation species richness in 4m² (mean of ten 4m² quadrats). Rsb = total number of seeds in the seed bank for 24L of soil.                        

 

137

2. Resilience Are these communities degraded? Are these communities resilient? A) In La Crau, the observation of the abiotic conditions, the vegetation and the seed bank of abandoned fields 20 years after abandonment suggests that succession on abandoned fields towards the steppe community is unlikely.Our results in Mediterranean climate show that the seed bank plays a minimal role in the re-establishment of former steppe on abandoned fields. corroborate most results found in Northern Europe This (Graham & Hutchings 1988a,b; Hutchings & Booth 1996a) (Fig. 42). B) In California, although grasslands seem to have an increasing resilience potential on an interior to coastal gradient (e.g. Heady et al. 1988), the resilience potential of California central coast prairies is still extremely low (Hamilton et al. 1999) or even unlikely (Stromberg & Griffin 1996). The seed bank cannot play a role in the restoration of prairies rich in native species on areas that have been intensively cultivated or on actual prairies that were converted from other community types, such as coastal scrub (Hayes 2004). On areas which were coastal prairies before European settlement and which were relatively less disturbed or which have only been extensively cultivated (e.g. hay during the Mission period), the role of the seed bank remains unclear, although some species recovery from the seed bank are documented: e.g. on the grassland of the City of Santa Cruz's Arana Gulch Greenbelt Land, weeds were removed andHolocarpha macradenia, Danthonia   californica,Lupinus bicolor,Castilleja densiflora andPlagiobothrys chorisianusrecruited from the seed bank (CNPS 2005). C) Vegetation succession of abandoned fields may be very long, but is hard to assess. In north-western Europe, Critchley & Fowbert (2000) observe relatively fast succession towards the reference grasslands, while Gibson & Brown (1992) state that succession towards the reference grasslands can take from decades to a century. In the Mediterranean region, many studies on vegetation dynamics after cultivation abandonment have been carried out (Escarré et al. 1983; Tatoni & Roche 1994; Desbussche et al. 1996; Ne'eman & Izhaki 1996; Bonet 2004). However, whether in northern Europe or in the Mediterranean Basin, studies have been carried out in woody landscapes, e.g. matorrals in the Mediterranean Basin.

 

138

In open landscapes, such studies have been carried out, but not in Mediterranean climatic contexts: e.g. in the shortgrass prairies ranging from the Rocky Mountains east to the central Great Plains, where climate is arid, and precipitation occur mainly between May and July (Booth 1941; Reichhardt 1982; Coffin et al 1996). Coffin et al. (1996) show that grassland recovery is highly variable depending on management practices and/or fine-scale climate, but can be achieved in 50 years. Conversely, Hobbs’ (2001) observe that there may not be a secondary succession on abandoned fields in Mediterranean climate which corroborates our results that show that, in a Mediterranean open-landscape, vegetation dynamics after cultivation abandonment is extremely slow, and recovery unlikely. I assessed resilience more precisely by carrying out a study at the margins of abandoned fields. Aronson & Le Floc'h (1996a) insist on the fact that any irreversibility thresholds, which have been crossed in the history of the ecosystems to-be- restored, need to be known, because the quantity and the types of thresholds crossed determine resilience and thus restoration efforts. They compare two specific areas with similar Mediterranean climate and suggest that in southern France, thresholds may have been crossed in the distant past, and are no longer relevant to the time frame considered. Hobbs & Hopkins (1990in   Aronson & Le Floc'h 1996a) also state that ecosystems of the Mediterranean basin are more likely to be resilient than other Mediterranean ecosystems, because they have evolved with anthropogenic disturbances. In southern France, the environment is adapted to some disturbances that have been occurring for thousands of years, such as fire. After fire or traditional cultivation abandonment, matorral species recolonize, following by pine trees, sclerophyllous oaks, and eventually deciduous oaks. Unless fires are too frequent, the ecosystems are resilient and we can say that irreversibility thresholds have not been crossed. However, our results show that the steppe La Crau is not resilient to a disturbance such as cultivation. I suggest, like Grubb & Hopkins (1986), that some ecosystems are not prepared for disturbance such as recent cultivation practices and although La Crau is the result of a long interaction between humans and nature, it does not make it resilient to a disturbance different from that it has evolved with. I also show that the re-introduction of grazing, the endogenous disturbance to the plain of La Crau, does not help vegetation dynamics towards the steppe, which corroborates results in Californian prairies and even slows it down greatly. The chances for a seed of a grassland species to be produced, to disperse to degraded grasslands, to germinate and establish and to not be grazed are close to zero.

 

139

3. Boundaries Does spontaneous restoration at the boundaries between remnant patches of undisturbed Mediterranean herbaceous plant communities and degraded communities occur and if it does, at what rate? Can we assess resilience precisely? A) In the steppe of La Crau, the spontaneous restoration at the boundaries between remnant patches of undisturbed steppe and degraded abandoned fields seems to occur in some cases (Fig. 43). The composition of above-ground vegetation at the margins of the abandoned fields studied and adjacent to steppe patches   changes with distance from boundaries: I find more steppe species close to boundaries and more arable weed species and mesophilous species further away. I recognize the fact that this experiment was carried out only on the sheepfold land of Peau de Meau at the margins of three abandoned fields. However, during my search for abandoned fields adjacent to steppe patches for this experiment, I found only four (one of which on military land) that seemed to show patterns of steppe species colonization. I thus believe that steppe species colonization is even lower on the fields that I have not studied. Steppe species colonize the field margins mainly through seed rain. Few steppe species are found in the seed bank (e.g.T. vulgaris), and seed dispersal by ants on the margins is limited to small distances, as observed by Harrington and Driver (1995). Although seed rain contributes to the colonization of most steppe species on the field margin, it is very limited: i) steppe species colonized less than 3 meter in ~30 years; ii) typical steppe species, such asStipa capillata, Asphodelus ayardii,Plantago holosteum, andHyssopus canescens are not found on the margins. This confirms Verkaar et al. (1983)'s work showing that dispersal distances of forbs are short. These results are completely different from what was expected by Etienne et al.’s (1998) and do not confirm those of Usher & Jefferson (1990), Gibson & Brown (1992), Willems & Bik (1998) and Critchley & Fowbert (2000). Instead, my results corroborate the observation of both Graham & Hutchings (1988a,b) and Hutchings & Booth (1996a) that succession is relatively slow or that of Hobbs (2001) that there is no secondary succession on abandoned fields in Australia. The Mediterranean climate induces conditions that are very different from those of north-western Europe or similar wetter and cooler climate, dramatically slowing down succession processes and thus leading to very different results (Blondel & Aronson 1999; Hobbs 2001).

 

140

B) In California, I did not study the spontaneous restoration from remnant patches of species-rich coastal prairies, but I assess it as highly unlikely due to landscape fragmentation and the relatively rarity of this ecosystem. C) I find that there is increasing soil pH and soil concentrations of phosphorus on the transition zones going from steppe patches to abandoned fields. I do not think that this weak environmental gradient can explain the change of vegetation composition as stated by Van der Maarel (ecocline 1990). These transition zones are more likely edge effects (Frochot 1987) created by steppe species, species of a more mature ecosystem, colonizing field margins (younger ecosystem)   (Whittaker 1980). However, colonization is extremely slow and thus this edge effect is small. This is probably relatively common in Mediterranean open landscapes because the low capacity of herbaceous stress-tolerant species to disperse influences greatly their spatial dynamics while weak environmental gradients only have a small effect on species recruitment. A 0m 10m   Rveg 18.3 14.0    12.6 9R ssb R wsb 5.3 4.3 10   R sr 12.5 8.6 0    6.6 9.5R ant   0   10       B 0m 10m Rveg 24.6 17.1   R ssb 12.6 12.3    3.7 1.3R wsb R sr 12.5 13.2 0   R ant 8 7.4 10   C 0m 10m   Rveg 21.5 15.8   R ssb 12.6 15   R wsb 7.3 3 R sr 8.3 11.6   R ant 4.2 8.1     Fig. 43. Main characteristics of the margins of three abandoned fields. Rveg = above- ground vegetation species richness /1m², R ssb = summer seed bank species richness /500cm³ , R wsb = winter seed bank species richness /500cm³, R sr = seed rain species richness /462cm², R ant = ant-born seed species richness /59cm².

 

141

4. Spontaneous restoration Do remnant undisturbed patches of Mediterranean herbaceous plant communities play a role in the resilience process within the context of a fragmented landscape? Does the rate of spontaneous restoration justify ecological restoration for the area?   A) While in many landscapes, remnant patches of native vegetation can have a positive effect on the secondary succession of abandoned fields (Coffin et al 1996; Willems & Bik 1998; Critchley & Fowbert 2000), I show that, in a Mediterranean fragmented open-landscape, remnant patches of steppe adjacent to abandoned fields do not seem to play much of a role in the colonization processes of steppe species as seed dispersal (by wind and ants) is limited. sheep grazing, which was reintroduced Even immediately after cultivation abandonment and which is known to be the only management for the steppe vegetation (Borck 1998; Fabre & Pluvinage 1998), cannot efficiently disperse seeds further onto degraded fields. In northern Europe plant and insect dispersal in fragmented landscape is successfully achieved by livestock (cattle, goats and especially sheep; Poschlod et al. 1998) through epizoochory and endozoochory (Fisher et al. 1996; Poschlod et al. 1996; Gibson & Brown 1992). In Mediterranean grasslands, endozoochory has proven to be an effective mechanism in the dispersal of viable seeds (Malo & Suarez 1995; Traba et al. 2003), although additional sowing is recommended (Traba et al. 2003). In La Crau, although the daily ranging patterns of itinerant sheep could easily be planned so that they graze on steppe patches first, then on abandoned fields in order to disperse steppe species seeds onto fields, most steppe species grow low, which reduces their adhesive dispersal (Fischer et al. 1996) and most steppe species disseminate in summer (Bourrelly et al. 1983) when all flocks are in transhumance in the Alps (Rinschede 1979; Fabre 1998).     Ecological restoration is needed to repair the degraded fields and re-establish the pre-existing steppe species composition and community structure because of the slow rate at which species colonize field margins and because remnant patches of steppe do not play a significant role in the recovery processeven with traditional itinerant sheep grazing.    Some bird species seem to benefit from various habitats, such as cereal fields, hay fields and abandoned fields (Fabre 1997; Dureau 1998; Vivat 1998; Wolff 1998). Restoration aims are thus certainly not to restore all abandoned fields in the Special Conservation Area. The restoration of select abandoned fields may allow patches of steppe to be reconnected. This may be an opportunity for steppe species likePrionotropis hystrix rhodanica,a listed endemic apterous grasshopper, which has only two populations left on separate steppe patches (Foucart & Lecoq 1998). Knowing precisely the real difficulty and cost of restoring the steppe is important as it serves as a basis to  142

assess mitigation measures. While remaining steppe patches are protected and are included in the Nature Reserve of La Crau, in the Special Protection Area and in the Special Conservation Area, destruction of this habitat still occurs. B) California coastal prairies are recognized as a rare habitat, but destruction of this habitat is also still allowed through mitigation measures which usually include habitat restoration (CNPS 2005). Unfortunately, so far, prairie restoration remains unsatisfactory and more work and studies are necessary.  C) In both cases, the impact of exogenous disturbances on herbaceous plant communities is strong: i) over the long-term, as changes in soil parameters, vegetation and seed bank composition are still observable 30 years later; and ii) over a large spatial scale, as remnant patches of grasslands do not play much of a role in the recovery process.   Recovery is highly unlikely and ecological restoration is thus needed. The restoration of these communities implies the re-introduction of grassland species and the control of plant propagation in the field. To do so, I discriminate the factors that most influence plant establishment among plant species strategies and various components of the degraded habitat.

 

143

5. Restoration Can we hierarchize the factors which limit restoration among various factors of the habitat, the disturbance regime and plant dissemination? Can irreversibility thresholds be reversed? 5a. Reference grassland seeds hardly disperse onto degraded grasslands, but when they do get there, do they find appropriate environmental conditions to germinate? I show that while field emergence ofNassella pulchra seeds reaches 37%, emergence ofD. californica, T. vulgarisandB. retusumis extremely low. None of the experiments I carried out to try to understand low field emergence showed any clear patterns. 5b. Reference grassland species can germinate but cannot survive and/or establish. Experiment with plug-planting. Whether in La Crau or in California, results show that in the degraded grasslands in as-they-are conditions the perennial species studied can survive the spring of planting. However,I can thus say that irreversibility thresholds have been passedbecause perennial species need to establish in favorable conditions (e.g. low grazing or low competition) to grow as much above- and/or below-ground biomass before the first summer to withstand summer drought. It should be noted that these experiments were conducted in 2003 that included a very extreme summer in France and that was not a particularly favorable year in California. In Mediterranean climates, notorious for high inter-annual variability, experiments should be carried out over several years, but this could not be done in the time allocated to a Ph.D. research. A) In La Crau,Thymus vulgaris andBrachypodium retusum the make most of the nutrients available on abandoned fields, but the combination of heavy grazing, ruderal species competition exacerbated the absence of large stones makes the fields a harsh environment to establish (Fig. 44). The ideal combination of treatments would be to exclude sheep grazing during the first spring to allow planted seedlings to establish, to reduce arable weed competition and to restore the original 50% stone cover (Fig. 45; Fig. 46). B) In California, exotic annual grasses are the main cause of transplanted native perennial death. This corroborates results showing that annuals  144