Oecologia (2006) 148: 373–383 DOI 10.1007/s00442-006-0379-2
E C O PH Y SI OL O G Y
Catherine Picon-Cochard Æ Lluis Coll Philippe Balandier
The role of below-ground competition during early stages of secondary succession: the case of 3-year-old Scots pine (Pinus sylvestris L.) seedlings in an abandoned grassland Received: 17 March 2005 / Accepted: 24 January 2006 / Published online: 18 February 2006 Springer-Verlag 2006
Abstract In abandoned or extensively managed grasslands, the mechanisms involved in pioneer tree species success are not fully explained. Resource competition among plants and microclimate modifications have been emphasised as possible mechanisms to explain variation of survivorship and growth. In this study, we evaluated a number of mechanisms that may lead to successful survival and growth of seedlings of a pioneer tree species (Pinus sylvestris) in a grass-dominated grassland. Threeyear-old Scots pines were planted in an extensively managed grassland of the French Massif Central and for 2 years were either maintained in bare soil or subjected to aerial and below-ground interactions induced by grass vegetation. Soil temperatures were slightly higher in bare soil than under the grass vegetation, but not to an extent explaining pine growth differences. The tall grass canopy reduced light transmission by 77% at ground level and by 20% in the upper part of Scots pine seedlings. Grass vegetation presence also significantly decreased soil volumetric water content (Hv) and soil nitrate in spring and in summer. In these conditions, the average tree height was reduced by 5% compared to trees grown in bare soil, and plant biomass was reduced by 85%. Scots pine intrinsic water-use efficiency (A/g), measured by Communicated by Marilyn Ball C. Picon-Cochard (&) Grassland Ecosystem Research Team, INRA, Agronomy Research Unit, 234 Avenue du Bre´zet, 63039 Clermont-Ferrand Ce´dex, France E-mail:
[email protected] Tel.: +33-4-73624584 Fax: +33-4-73624457 L. Coll Æ P. Balandier Applied Ecology of Woodlands Research Team, CEMAGREF, 24 Avenue des Landais, BP 50085, 63172 Aubie`re Ce´dex, France Present address: L. Coll Groupe de Recherche en E´cologie Forestie`re Interuniversitaire (GREFi), Universite´ du Que´bec a` Montre´al, Succ. Centre-Ville, C.P. 8888, Montre´al, QC, H3C3P8, Canada
leaf gas-exchange, increased when Hv decreased owing to a rapid decline of stomatal conductance (g). This result was also confirmed by d13C analyses of needles. A summer 15N labelling of seedlings and grass vegetation confirmed the higher NO3 capture capacity of grass vegetation in comparison with Scots pine seedlings. Our results provide evidence that the seedlings’ success was linked to tolerance of below-ground resource depletion (particularly water) induced by grass vegetation based on morphological and physiological plasticity as well as to resource conservation. Keywords Light Æ Morphological and physiological plasticity Æ Soil water and N Æ d13C Æ 15N
Introduction Understanding the mechanisms driving woody plant seedlings establishment and growth is an important first step in the comprehension of shifts from grass to woody plant domination during secondary succession caused by cessation of agricultural activities in grassland ecosystem. Natural patterns of seedling establishment (germination and emergence) in grasslands are well documented and have been related to different biotic (seed bank and dormancy, predation) and abiotic variables (climate change, fire regimes, livestock grazing frequency) (Van Auken 2000). In many cases, seedling emergence can be severely restricted in an intact grass canopy and many species require gaps to establish (McConnaughay and Bazzaz 1991). Once established, survival and growth are further influenced by interactions with the grass vegetation (McConnaughay and Bazzaz 1991). The role of such interactions during succession is well documented (Connell and Slayter 1977; Tilman 1985; Scholes and Archer 1997). However, these interactions are complex: they can be positive, negative or neutral, and they can change through time and space
374
in relation to specific phenologies, resource availability and climatic conditions (Balandier et al. 2005a). Negative interactions are generally described by the word ‘‘competition’’, even if used to describe such interactions at vastly different study levels (i.e. ecosystem, community, species or individuals; Grace and Tilman 1990). At the individual scale, competition between plants occurs when the supply of shared resource is limited, leading to a reduction of growth, survival or reproduction (Begon et al. 1990). Here, we used this term in accord with a functional meaning: resource competition occurs when plant individuals utilize the same pool of growth-limiting resources (Grime 2001). In the case of the fertile grassland ecosystems, it has largely been shown that exploitative grasses (which are the dominant species) are highly competitive and are generally assumed to be more competitive for belowground resource (Grime et al. 1990; Harmer 1996), but their effect above ground is less evident. For example, the influence of grasses on light availability is unclear because most have thin erect leaves that may or may not significantly alter light penetration to ground level. In fertile grasslands characterised by a tall and dense grass canopy, competition for light is more conspicuous (Soussana and Lafarge 1998). Others have also suggested that extreme temperature fluctuations, aboveground and below-ground, induced by grass vegetation, may have or not dramatic consequences for seedlings survival and growth depending on the considered species and climate (Ball et al. 1997, 2002; Provendier and Balandier 2004). From these examples, we can conclude that in many cases exploitative grasses are probably unfavourable to tree establishment and growth, but the determinants need to be identified. Pioneer tree species are characterised as having broad physiological responses to environmental variations (Bazzaz 1979; Sands and Nambiar 1984; Caldwell and Richards 1986; Casper and Jackson 1997). Their success despite the presence of grasses could be realised if they survive and grow by avoiding or tolerating resource depletion through morphological and physiological plasticity (e.g. lower tissue turnover, reduction of nutrient losses, lower stomatal conductance, etc.) (Goldberg 1990; Ko¨chy and Wilson 2000; Connolly et al. 2001; Peltzer and Ko¨chy 2001). In the present experiment, we studied Scots pine (Pinus sylvestris L.), which is able to rapidly colonise open areas (Hansen et al. 2002; de Chantal et al. 2003), and during the second half of the past century has naturally established in both unfertile or fertile abandoned pastures of the French Massif Central (Bazin et al. 1983; Pre´vosto 1999; Pre´vosto et al. 2000). Our main objective was to understand why and how this species (considered a pioneer tree) succeeds in grasslands during secondary succession while other species like beech (Fagus sylvatica), a late successional species, does not (Coll et al. 2003; Balandier et al. 2005b). Determination of these mechanisms was achieved by measurements of soil temperature, aerial and below-
ground resource availability and acquisition and resource use efficiency of the different competitors. Evaluations were based on leaf functional traits (leaf gas exchange, needles bulk d13C, SLA, N%), plant growth (height, biomass) and short-term 15N labelling experiments for N mineral acquisition. Specific questions were addressed: (1) what was the effect of vegetation on aerial and below-ground resources availability and microclimate modifications, and (2) how did Scots pine seedlings respond to this resource depletion or microclimate modification?
Materials and methods Site and experimental plots In March 2000, a grassland dominated by herbaceous species (mainly graminaceae) and extensively grazed by sheep prior to the experiment was fenced (20·20 m2) to prevent predation by wild animals. The experiment was situated in the southern part of the ‘Chaıˆ ne des Puys’ in the French Massif Central (900 m a.s.l.; 4543¢N, 259¢E), where the montane climate experiences oceanic influences (820 mm annual rainfall, 7C annual mean air temperature). Soil is developed on a substrate of basaltic ash-fall deposits or lava blocks. It is characterised by a loamy silt texture (pHwater=6) with a rich-organic upper horizon (average thickness of 23 cm). Scorias appear between 40 and 60 cm deep, which constitute a welldrained layer, whereas the upper horizon presents risk of rapid summer dehydration. Thirty-six small plots (2·2 m2) separated by 0.5-m width pathways, in order to avoid trampling of the treatments, were established to create four experimental treatments: (1) bare soil (six replicates), (2) pine (P) by planting 3-year-old bare rooted Scots pine (Pinus sylvestris, St-Bonnet-le-Chaˆteau’ Provenance, France) in bare soil (ten replicates), (3) pine + vegetation (P+veg) by planting Scots pine within the grass vegetation (ten replicates), and (4) vegetation corresponding to the natural grass vegetation of the plot (ten replicates). The year before the pine seedlings were planted, the bare soil was obtained by a glyphosate application (RoundUp, Monsanto). One month later, roots of vegetation localised in the upper soil layers were manually removed to avoid mineralisation of decaying material, and thereafter bare soil was maintained by manual weeding. Three-year-old Scots pine seedlings were used to avoid the establishment stage already documented and to focus on the period where grasses and pine have a close stature. In the two treatments with pines, five individuals were planted per replicate. The planting occurred late in the season (May 2001) because the first planting (November 2000) failed probably due to the winter dryness (only 16.4 mm in December 2000). The replicates were randomly assigned to the four treatments within the site.
375
Vegetation composition and management The botanical composition of the vegetation was evaluated in June 2001 in order to quantify the proportion of herbaceous groups. All the species observed in a circle of 50 cm diameter in the central area of each replicate were recorded. The vegetation was composed of 61% grasses (mainly Arrhenatherum elatius, Agrostis capillaris, Festuca heterophylla, Holcus lanatus, Dactylis glomerata, Poa trivialis), 26% non N-fixing dicots (Achillea millefolium, Galium verum, Taraxacum officinale) and 13% legumes (Vicia sativa, Trifolium pratense). The herbaceous vegetation was cut with manual battery powered clippers at 6 cm height in November 2000 and again in March 2002 in order to simulate extensive management of the grassland.
Sigma, France) were placed in nylon mesh bags (5·10 cm2, mesh size 50 lm). Three (bare soil, P treatments) and four (vegetation, P+veg) bags were buried vertically at 15 cm depth at a distance of 20 cm from the stem of the central Scots pine seedling in five replicates of P, P+veg and vegetation treatments and in three replicates of the bare soil treatment. Two sets of bags were incubated for 28 and 22 days, May–June (DOY 136–164, spring) and June–July (DOY 175–197, summer), respectively. After the incubation period, all bags for a given replicate were pooled and the resin washed with 50 ml of de-ionised water. Inorganic N was extracted by shaking the resin for 5 min in 100 ml of 2 M NaCl in 0.1 M HCl. The extract was filtered through ashless paper and frozen before determination of NO 3 with a flow analyser (Aquatec, France). Nitrate exchanged in the resin bag was expressed in microgram per gram resin per day for the two periods.
Temperature and resources measurements Soil temperature
Light
Soil temperature was measured with 10-cm-long copperconstantan probes (107L type, Campbell Scientific, UK) vertically inserted in one plot of each treatment and logged at 30-s intervals with a data logger (Model 10X, Campbell Scientific) and averaged over 30-min periods. Maximum, minimum and average daily values are presented.
The photosynthetic active radiation (PAR) was measured with a sunfleck ceptometer (Decagon Devices, Pullman, Wash., USA) at three vertical heights, 10 and 20 cm from the apex of the pine and at ground level. The transmitted PAR was calculated as the ratio of PAR measured in the grass vegetation to the incident PAR above the grass vegetation. In the grass vegetation treatment, measurements were only performed at the ground level. Two vertical profiles were performed on either side of the Scots pine seedlings in an east–west orientation, and the mean of these two measurements was used. Measurements were performed at noon (solar time) on cloudless days four times during the 2002 growing season and on ten replicates for each treatment. No measurements were performed in P and bare soil treatments.
Soil water content In February 2001, thin-walled plastic TDR tubes (80 cm length, 4 cm diameter) were inserted vertically in the soil at 20 cm distance from the seedling located in the centre of the replicate. These were used to measure the profile of soil volumetric water content (Hv, water volume by soil volume, in percent) with a TDR probe (Trime T3, IMKO, Ettlingen, Germany). Because the scoria layer at around 40 cm depth disrupts too much the Hv measurement, values were only used from the 0–20 cm layer of the soil, layer containing most of plant roots (Coll et al. 2003). Measurements were done weekly or every 2 weeks from day of year (DOY) 137 until DOY 225 in six to seven replicates of P, P+veg and vegetation treatments and in three replicates for the bare soil treatment. For each replicate, the mean of three measurements, performed in three different directions, was used for the calculation of Hv. Only the results of 2002 are shown in this paper because those of 2001 showed the same pattern of variation among treatments. Soil nitrate availability Ion-exchange resin bags were incubated in treatment soils in 2002 and NO 3 captured on the resins was measured. According to Gloser et al. (2000), 6.19– 6.205 g of wet mixed bed resin (Amberlite IRN-150,
Plant growth measurements and functional traits Growth and biomass The heights of the pines and of the grass vegetation were measured weekly during the two growing seasons (April–August 2001 and 2002). For P+veg and grass vegetation treatments, height of the vegetation was measured with a sward stick in a square (25 contact points in 50·50 cm2) around the central pine or in the centre of the replicate. In July and December 2002, three and five Scots pine individuals per treatment, respectively, were harvested and oven-dried (60C for 48 h) for needle, stem and root biomass determination. In the laboratory, needles were sorted by age class (current-, 1 and 2 years old). Projected needle area was measured on sub sample of current- and 1-year-old needles with an electronic planimeter (LI 3100, Li-Cor, Lincoln, Neb., USA).
376
These measurements were used to calculate specific leaf area (SLA, m2 kg1) after needles were oven-dried (60C for 48 h) and weighed. Roots were washed and separated by diameter class (fine 2 mm). Root to shoot ratio (R/S) was calculated as ratio of root to shoot dry biomass. Leaf gas exchange In June (DOY 170 and 177) and July (DOY 200) 2002, leaf net CO2 assimilation rate (A, lmol m2 s1) and stomatal conductance for water vapour (gw, mmol m2 s1) were measured in situ in natural light conditions (full light conditions, mean PAR=1,527± 36 lmol m2 s1, n=28) with an open gas-exchange system (LiCor6400, Lincoln, Neb., USA). At noon (solar time), measurements were performed on 1-year-old needles (11–33 cm2 projected leaf area) from a lateral shoot. Values were recorded five times during the 20–30 min period following chamber closure. Thus, the mean of five measurements was used for each of the three individuals per treatment. The same individuals were used for each date. On DOY 177, leaf gas-exchange was also measured on the grass Dactylis glomerata located in the replicates of the P+veg treatment. This species was chosen as representative of exploitative grasses common in the studied grassland. Measurements were performed with the same chamber after the pine measurements. Two fully expanded leaves were chosen (2–4 cm2) in the vicinity of the pine used in each of three replicates per treatment. Scots pine isotopic composition of
13
Pine needles collected in July and December 2002 were sorted by age class (current, 1 and 2 years old), ovendried (48 h, 60C) and finely milled. Five to seven micrograms were weighed and used for carbon isotope composition determination by mass spectrometry (FISONS/ISOCHROM). Results are expressed in d notation (Eq. 1), i.e. relative to the Pee Dee Belemnite standard: Rs d¼ 1 1000 ð‰Þ; ð1Þ Rst 13
Data analysis With a complete random experimental design, variance analysis (ANOVA) was performed using a General Linear Model (Statgraphics Plus, v 4.1, Manugistics, USA) for all variables to test treatment effects with one(vegetation presence) or two-factor(s) (species, vegetation presence). Means were separated from least square deviation method (LSD). Root-to-shoot ratio was logtransformed before ANOVA.
Results
C (d13C)
where Rs and Rst are the molar fractions of the sample and the standard, respectively.
N corresponds to a supply of 0.1875 g 15N m2. Twenty-six days later (DOY 205), Scots pine seedlings were collected as well as the vegetation within the labelled area. Most of the pine roots were excavated in each treatment, especially for P+veg where the root system was confined in the first 20 cm. However, in P treatment, roots expanded horizontally and into deeper layers, thus it was impossible to get the whole root system. For the grass vegetation of each treatment, the roots of the first 40 cm in the soil were collected and washed. All organs of the Scots pines and the grass vegetation were oven-dried (48 h, 60C) and finely milled, and 5–7 mg of each sample were analysed for total nitrogen concentration (mass basis) and 15N by mass spectrometry (FISONS/ISOCHROM). Results are expressed as isotopic excess (%), which corresponded to the difference between sample 15N abundance and air 15N abundance (0.3663%). The quantity of absorbed 15N was calculated at the plant level (lg 15N g1 DW).
C–12C for
Plant N measurements At the end of June 2002 (DOY 179), one day after a 21 mm rainfall, 500 ml of 15NH15 4 NO3 (10% isotopic excess) were applied to a 40-cm square area centred on a target pine seedling in three replicates each of the P, P+veg and vegetation treatments. This amount of
Soil temperature Mean soil temperature differences between bare soil and P+veg treatments were 0.7, 1.2 and 0.4C for average, maximum and minimum daily values, respectively (Fig. 1). These differences increased in summer and reached 2.6C for the maximum daily values and less than 1C for the minimum values. Resources availability Soil water Seasonal fluctuations of soil volumetric water content (Hv) of the upper horizon (0–20 cm) were observed with maximum and minimum values in spring and summer periods, respectively (Fig. 2). The bare soil treatment, without or with pine (bare soil and P, respectively), exhibited the highest values of Hv from DOY 176 to 225 (P
