- Email: [email protected]
Germination of an obligate seeder (Ulex parviflorus) and consequences for wildfire management
Forest Ecology and Management 256 (2008) 685–693
Contents lists available at ScienceDirect
Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco
Germination of an obligate seeder (Ulex parviflorus) and consequences for wildfire management M. Jaime Baeza a,b,*, Jacques Roy c a
Centro de Estudios Ambientales del Mediterra´neo, CEAM, Parque Tecnolo´gico Paterna, Charles Darwin, 14, 46980 Valencia, Spain Departamento de Ecologı´a, Universidad de Alicante, Ap. 99, 03080 Alicante, Spain c Centre d’Ecologie Fonctionnelle et Evolutive, CNRS, UMR5175, 34293 Montpellier cedex 5, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 March 2008 Received in revised form 12 May 2008 Accepted 13 May 2008
In fire-prone ecosystems, reducing the risk of wildfire is generally attempted through vegetation clearing using controlled fires or, less often, mechanical techniques. Management practices, however, can be poorly efficient when the disturbances they introduce install environmental conditions that are similar to the ones under which the undesirable species evolved. Ulex parviflorus is a Mediterranean obligate seeder with physical dormancy forming large amounts of highly combustible standing necromass. In the present study, combining field and laboratory experiments, we determined the seedling recruitment of this species under different management practices (burning, mechanical clearing, slash/no slash on soil surface), we measured the environmental conditions (temperature, light, and moisture) enforced by these practices and we tested their individual and combined impact on germination in order to determine the most appropriate control method for this species. Germination is low under intact canopies, but it is strongly stimulated by both brush-chipping and fire. This is partly related to the inhibiting effect of the low red:far-red ratio of the light filtered by the canopy which is removed by brush-chipping and fire. The other factor involved is moderate heat, either fire-generated or resulting from solar radiation on bare soil, which breaks seed coat impermeability. Indeed exposing seeds on bare soil in summer resulted in a significant increase in their germination capacity and germination was reduced when the brush chips remained on the soil. Moisture fluctuations did not enhance germination. The summer heat impact affects management practices. When the brush-chipping treatment occurred before summer, the germination flush appeared the following autumn, but when the treatment occurred after summer, the germination flush did not appear until the autumn of the subsequent year, when interspecific competition with regenerating vegetation is likely to be more intense. We demonstrated that brushchipping, especially when done after summer, is a better technique than fire for controlling U. parviflorus because it creates environmental conditions that are less favourable for its germination. This technique also has the potential to favour late-successional species less vulnerable to fire. By combining fuel reduction and land restoration, this technique is useful to the sustainable management policies that need to be developed. ß 2008 Elsevier B.V. All rights reserved.
Keywords: Clearing Control burning Fuel reduction Light quality Moisture fluctuations Slash Temperature fluctuations
1. Introduction Wildfires constitute a serious risk in many regions of the world, and their frequency is expected to increase as a result of climate change and land abandonment (fuel accumulation), in particular in ˜ ol et al., 1998; Cramer, 2001; Pausas, Mediterranean regions (Pin 2004). Vegetation control for reducing fuel loads in forest and
* Corresponding author at: Centro de Estudios Ambientales del Mediterra´neo, CEAM, Parque Tecnolo´gico Paterna, Charles Darwin, 14, 46980 Valencia, Spain. Fax: +34 96 590 98 25. E-mail address: [email protected] (M.J. Baeza). 0378-1127/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2008.05.014
shrubland ecosystems is often applied to decrease the fire risk and improve the probability of successful fire control by reducing fire intensity. Several management techniques have been developed to control undesirable species, including mechanical clearing, herbicides, fire, grazing, biological control agents or a combination of these (Paynter and Flanagan, 2004; Buckley et al., 2004). However, despite the stressfulness of these management techniques, the fact that natural selection in many species occurred under drastic disturbance regimes (Pickett and White, 1985) can reduce the effectiveness of these efforts. This is particularly the case in biomes where fire has been part of the selective forces shaping biodiversity. Such species have developed very effective regenera-
686
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
tion mechanisms, which are triggered by the environmental conditions associated with fire. Nevertheless, Zedler and Zammit (1989) pointed out that few of these conditions are specific to fire (very high temperatures, ash or charcoal covering) while most are shared with other disturbances, including management practices (changes in light intensity and quality, temperature, moisture, competition). Thus, fuel reduction management could have the counterproductive effect of facilitating the germination of undesirable species. For a successful vegetation management programme, understanding the ecological processes involved in recruitment are of key importance to generate unfavourable opportunities for undesirable species regeneration (Freckleton, 2004). Ulex parviflorus Pourr., Mediterranean gorse, is a shrub of the Fabaceae family found all along the Mediterranean coasts (and up to 100 km inland) of Southwestern Europe and North Africa (Tutin et al., 1964–1980). It is an obligate seeder with a physical exogenous dormancy on the seed (Baeza and Vallejo, 2006), which develops massively after land abandonment and fire and can constitute the dominant species of dense thickets, as is the case in Southern Spain. When senescence occurs (after 20–25 years), these stands constitute large amounts of highly combustible standing necromass (Baeza et al., 2006). Management strategies to control the extension of U. parviflorus are now being studied and applied; among these, controlled fires and clearings are the most common (Lloret and Vila`, 1996; Baeza et al., 2002b). However, experimental burning and brush-chipping have shown the difficulty of controlling U. parviflorus due to its massive regeneration through seed germination (Baeza, 2001). The elimination of existing vegetation and the high temperature generated in the fire can scarify the seed coat and break the dormancy of seeds stored in the soil (Baskin and Baskin, 1998). Similar environmental problems are caused on a larger scale by other leguminous species like Ulex europaeus L. and Cytisus scoparius (L.) Link, both of which are native to Europe and are now regarded as noxious weeds in New Zealand, Chile, Hawaii, North America and Australia (Rees and Hill, 2001; Sheppard et al., 2002). Leguminous species have an impermeable seed coat that imposes physical dormancy on the seed; to germinate the waterimpermeable layer must become permeable (Baskin and Baskin, 1998). In Mediterranean-type ecosystems, heat from fire has been directly correlated with seedling recruitment because high temperatures can break hard seed coat dormancy, making seeds permeable and ready to germinate under favourable environmental conditions. This temporal sequence indicates that the environmental factors involved in seed dormancy breakage may differ from those involved in germination (Thompson et al., 2003). In nature, fire is the most extreme example of temperature overcoming dormancy in hard-seeded species (Auld, 1986); other factors associated with temperature could also be related to dormancy loss (Ooi et al., 2006). Under field conditions, combined temporal sequences of high temperatures in summer followed by moist conditions in autumn occur seasonally. By uncoupling these seasonal effects management programmes may be able to delay the emergence flush in species with physical dormancy and improve control efficiency. In this paper we analyse the germination conditions of a largely undesirable shrub in relation to its current management practices and we discuss the suitability of these practices. We study (i) the in situ germination of U. parviflorus after different management treatments, (ii) the environmental soil surface conditions created by these treatments and (iii) the germination requirements of the seeds in controlled conditions. We then compare the relative efficiency of fire-specific and general-disturbance germination clues and discuss the techniques used for managing this species in the light of its germination ecology.
2. Material and methods 2.1. Study area and sites The study was conducted in the Mariola and Aitana interior mountains in the northwest of the Alicante province, Spain. The typically Mediterranean climate is characterised by a mean annual precipitation of 495 mm (Onil, 388 390 N, 08 400 W, 800 m above sea level, 40 km from the sea) with a summer drought from mid-May to mid-September. The mean temperature is 14.5 8C with mean monthly temperatures ranging from 7.5 8C in January to 23.5 8C in ˜ eres, Castell and Onil), were chosen July. Four sites (Alcoy, Ban within a 40 km distance. Their climatic conditions, bedrock (marl) and soil (deep loamy Calcaric Cambisol, FAO 1998) are similar. All 4 sites were cultivated terraces abandoned after the Spanish Civil war (1939). They were covered with Aleppo pine (Pinus halepensis Mill.) forests until a wildfire in 1985 (Alcoy and Onil) or 1991 ˜ eres and Castell). At the beginning of the field experiments (Ban (1994), the shrubland vegetation was dominated by U. parviflorus and to a lesser extend by other obligate seeders such as Cistus albidus L. and Rosmarinus officinalis L. A few isolated individuals of woody resprouters were present (Quercus coccifera L., Juniperus oxycedrus L. and Rhamnus alaternus L.) as well as a few small isolated individuals of P. halepensis. The herbaceous layer, covering approximately 70% of the ground, was dominated by Brachypodium retusum (Pers.) Beau. 2.2. Field experiments The four sites were part of a larger experiment designed to study the impact of successional stage and fuel load on fire behaviour and its implications for using control burning or clearing to reduce wildfire risk (Baeza et al., 2002b). In each site, three treatments were applied in May 1994, each on a single 1000 m2 plot: control, brush-chipping (slash remaining on the soil) and burning. Once a month until January 1996, U. parviflorus seedlings occurrence was recorded in 20 randomly distributed 0.5 m 0.5 m sub-plots within each plot. In Onil, three 1000 m2 plots were brush-chipped for a fuelbreak experiment in September 1996. The effect of keeping the chips on the ground (slash with 100% cover) or removing them (no slash, bare soil) on U. parviflorus germination was tested. In each plot, the chips were removed in five randomly distributed 1 m 1 m sub-plots and kept in five other randomly distributed 1 m 1 m sub-plots. U. parviflorus seedling occurrence was recorded in all these sub-plots from October 1996 until May 1998. In Onil, the impact of seed exposure to soil surface summer conditions was tested using seeds collected in June 2001 from ˜ eres and in Castell. For each seed origin, 20 sets mature fruits in Ban of 50 seeds were enclosed in bags made with nylon mesh (0.5 mm). Five sets were kept as control while fifteen were put on the bare ground of a fuel-break on July 1, 2001. Five sets were removed after 30 days, five sets after 60 days and the remaining 5 sets after 90 days. Soil surface temperature was measured and recorded every hour with a temperature probe (Hobo1 Event, Onset Computer Corporation, Bourne MA, US). Each set of seeds was then put in a Petri dish on filter paper and enclosed in a germination chamber (12 h light at 22 8C, 80% moisture and 12 h dark at 20 8C and 85% moisture). Filter paper moisture and seedling germination were checked every other day for 60 days. These field experiments suggested that fluctuations in soil temperature and moisture as well as heat generated by fire or change in light quality due to green leaves removal could stimulate U. parviflorus germination. These physical characteristics were then documented in the field before testing their isolated or
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
interacting effects in controlled conditions. Fluctuations of the soil surface temperature in a fire-break (with no slash) and under an adjacent mixed forest was compared in Onil using one temperature probe in each environment (Hobo1 Event, Onset Computer Corporation, Bourne MA, US). The probe thermocouple was inserted in the first mm of the soil and the measurements were taken during August and September 1998. The number of rainy days between seed set in May and seed germination in November was measured at the weather station ˜ eres for the period 1950–1990. The relation between light of Ban quality and vegetation cover was studied in November 2002 in the vicinity of the Castell site in two Mediterranean gorse shrublands 4 and 17 years old respectively. Every 25 cm along one 30 m transect in each shrubland, vegetation cover was estimated by the number of contacts along a vertical point quadrat (Greig-Smith, 1983). Light quality (ratio between 660 and 730 nm wavelength) was measured at the base of each quadrat using the red/far-red ratio sensor described in Me´thy et al. (1987). 2.3. Germination in controlled conditions ˜ eres Seeds were collected in June 2002 from mature fruits in Ban and stored in paper bags at room conditioned temperature. Seed pre-treatments and germination were conducted at the Centre of Functional and Evolutionary Ecology in Montpellier from October 2002 to February 2003. 2.3.1. Seed pre-treatments Petri dishes (9 cm diameter) were filled with the following layers: 1 cm of soil, one filter paper, 50 seeds, one filter paper and 0.5 cm of soil. Four Petri dishes (replicates) were used for each treatment level. - Temperature fluctuations (simulating summer bare soil surface temperature): Two germination chambers were set up with a 15 8C night temperature for 11 h and 45 and 55 8C day temperature respectively for 13 h. Temperature within each chamber was measured with a thermocouple and recorded every hour with a data-logger (CR21X, Campbell Scientific, Utah, USA). Room temperature (control) varied between 18 and 22 8C. - Moisture fluctuations (simulating wetting of the seeds by rainfall events): Half the Petri dishes enclosed in the germination chambers, as well as half the ones kept as control for the temperature fluctuations were watered every other day with 30 ml of distilled water. The other Petri dishes were left dry. - Length of these pre-treatments: The above two fluctuation treatments were applied for 5, 15 or 30 days. The corresponding number of dry/wet periods was 2, 7 and 15 respectively. The 30 days treatment started first, the 15 days one started 15 days later and the 5 days one started 25 days after the first one so that the germination could start simultaneously for all treatments with
687
an identical lap of time since the end of the pre-treatments. The treatments temperature, moisture and length of fluctuations were crossed. - Heat (simulating fire heat): The Petri dishes (with dry seeds) were introduced for 10 min in a preheated oven at 80 8C, the optimum short term heating for triggering germination. Due to a lack of ˜ eres, this optimum had been determined on seeds from Ban another population (Confrides) whose seeds were introduced the same way in a chamber at 60, 80, 100, 120 or 140 8C for 10 min. - Mechanical scarification (to estimate the potential maximum germination): Seeds were individually scarified with the blade of a scalpel. 2.3.2. Seed germination Seed were recovered from in between the filter papers of the pre-treatment Petri dishes. Each set of 50 seeds was then enclosed in a small mesh bag and soaked in a 1% sodium hypochlorite solution for 10 min to eliminate fungal contamination. The seeds were then rinced twice with distilled water. Each set of 50 seeds was spread on a filter paper in Petri dishes (9 cm diameter). The filter paper had been deposited on one layer of 0.6 mm glass beads allowing for a water reserve and anti-fungal Tachigaren LS1 (Hymexazol 0.5%). The Petri dishes were enclosed in three germination chambers set up with a 17 8C night temperature for 11 h and a 23 8C day temperature for 13 h. Day light was provided by two 8 W fluorescent bulbs in each chamber. In order to simulate light quality typical of a deep under canopy shade, far-red wavelengths were added in one of the germination chamber by a 40 W red-coloured incandescent lamp (Claude) surrounded by a blue filter (Altuglas S300, Altulor). Red/far-red ratio was 0.4 in this chamber compared to 1.1 in the other two chambers. The impact of this increased radiation was tested on seeds which had received the pre-treatment temperature and moisture fluctuations for 15 days. The 24 combinations of treatments (with 4 Petri dishes per combination) are given in Table 1. Germination (i.e. visible radicle) was recorded for 97 days, twice a week during 60 days and then once a week. 2.4. Statistical analysis The effect of vegetation treatments (control, brush-chipping and burning) and the presence of remains on the ground (slash-noslash) on U. parviflorus germination dynamics were analysed using repeated measures ANOVA. Mauchly’s W test statistic was used for testing sphericity assumptions (independence of data among repeated measures). For data found to be correlated among years, the Geisser-Greenhouse correction was used to adjust the degrees of freedom of the within-subjects. The time course of germination was investigated by fitting the curve of the cumulative germination (y) across time (t) with a sigmoid function (Scott et al., 1984)
Table 1 ˜ eres seeds (n = 4) Pre-treatment combinations for the Ban Length of fluctuations
5 days 15 days 15 days 30 days Heat 80 8C Mechanical
Light quality
z = 1.1 z = 1.1 z = 0.4 z = 1.1 z = 1.1 z = 1.1
Temperature and moisture fluctuations Room temperature
15–45 8C
Dry
Dry/wet
Dry
Dry/wet
Dry
Dry/wet
15–55 8C
688
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Fig. 1. Time course of field germination of U. parviflorus under intact canopies or after brush-chipping (slash remaining on the soil) or fire. The canopies were chipped or ˜ eres and (b) Castell, and a 9-year-old regrowth after fire at (c) Onil and (d) burned in May 1994 (arrows). At that time, vegetation was a 3-year-old regrowth after fire at (a) Ban Alcoy.
using Sigma Plot 7.0 programme (SPSS Inc., Chicago, IL, USA). The following sigmoid function for which y = 0 when t = 0 was chosen: y ¼ A½1 expðBtÞc The final number of germinated seeds (A, the asymptote of the curve) and the mean germination time (the time to 50% germination, t50 = [ln(1 exp(0.51/c))]/B, were derived from this function rather than from the raw data because in some treatments the germination was not completely finished when the experiment was stopped. The impact of seed exposition to soil surface summer conditions with four levels of duration (0, 30, 60 and 90 days) and two levels per seed origin was analysed by twoway ANOVA. Linear regression analysis was used to determine the relationship between light quality and presence of living vegetation components. Germination percentage and half-time for germination were analysed by three-way (temperature, moisture and length of the fluctuations pre-treatment) ANOVAs with all factors as fixed effects (temperature fluctuations with three levels (20–20, 15–45 and 15–55 8C), moisture fluctuations with two levels (dry and dry/wet) and length fluctuations with three levels (5, 15 and 30 days). We used two-way ANOVAs to determine the significance of the two fixed factors temperature fluctuation during 15 days with tree levels (20–20, 15–45 and 15– 55 8C) and light quality with two levels (red/far-red ratio 0.4 and 1.1), on germination percentage and half-time for germination. Finally we compared with ANOVA the impact of the different factors (scarification, fire simulation, light quality, intensity fluctuations and temperature + moisture fluctuations) on germination percentage and half-time for germination. Data were ln transformed (number of seedling per quadrat and the time to 50% germination) or arc-sine transformed (germination percentage) to correct deviations from normality. Post hoc comparisons of means were done by Tukey’s multiple range test. Standard error
bars are shown on the graphs. All the statistical analyses were performed using the SPSS 10.0 package (SPSS Inc., Chicago, IL, USA). 3. Results 3.1. Germination after in situ vegetation management treatments Germination of U. parviflorus occurs in autumn during the month following the first rainfall events. It was low under intact canopies but was strongly and significantly stimulated by both brush-chipping and fire (d.f. 2; F = 30.15; p < 0.001, Fig. 1). Germination was extremely low the second autumn after the treatments although September to December rainfall was higher in 1995 than in 1994 (187 and 129 mm respectively). After brush-chipping, germination was significantly reduced when the brush chips (slash) remain on the soil as compared to bare soil (d.f. 1; F = 17.19; p < 0.001, Fig. 2). When the brushchipping treatment occurred before summer, germination flush appeared the autumn of the same year (Fig. 1), but when the treatment occurred when the summer was over, germination flush did not appear until the autumn of the subsequent year (Fig. 2). Since rainfall was not particularly limiting germination in 1996 (September to December rainfall was 179 and 155 mm in 1996 and 1997 respectively), results of Figs. 1 and 2 suggest that the summer environmental conditions are needed to trigger the germination of U. parviflorus when the canopy is eliminated. Exposure the seeds on the soil surface of a fire-break during the summer indeed resulted in a significant increase in germination percentage measured in a germination chamber (d.f. 3; F = 12.1; p < 0.05, no difference between 30, 60 and 90 days of exposure) (Fig. 3).
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Fig. 2. Time course of field germination of U. parviflorus after brush-chipping when slash is removed or when it remains on the soil. The canopy (at Onil, 11 years old) was chipped in September 1996 (arrow).
689
Fig. 5. Relationship between light quality (red/far-red ratio) and the density of the vegetation measured as the number of contacts of vertical point quadrats with living plants. 240 measurements were taken along 30 m transects in 4 and 17 years old gorse shrublands.
(Fig. 4). Under the canopy of an adjacent mixed forest, these fluctuations were buffered, typically between 16 and 30 8C with peaks at 33 8C. The top soil layer and consequently the seed bank also experienced alternating periods of wetness and dryness. In average during the dry period, 2 days with rainfall >5 mm were ˜ eres between recorded per month at the weather station of Ban 1950 and 1990 (3.2; 1.8; 0.6; 0.8; 2; 2.9 mm from May to October respectively).
Fig. 3. Effect on subsequent germination of exposing the seeds to the summer soil surface conditions in a fire-break without slash. Seed were not exposed (0 days) or ˜eres and Confrides were used. exposed 30, 60 and 90 days. Seeds of two origins Ban Different letters indicate significant differences (p < 0.05) between exposure length.
3.2. Microsite conditions during or after vegetation removal 3.2.1. Summer soil surface conditions Temperature in August and early September typically fluctuated between 14 8C at night and 45 8C during the day at the soil surface of a fire-break (with no slash). Peaks at 50 8C were reached
Fig. 4. Daily maximum and minimum temperatures in August and September on the soil surface of a fire-break (with no slash) and of an adjacent mixed forest.
3.2.2. Light quality and vegetation cover The ratio between the 660 and 730 nm wavelengths was measured in two Mediterranean gorse shrublands 4 and 17 years old respectively. This ratio decreased as expected with the density of the vegetation, from a value of 1.1 in the incident radiation to a value of 0.6 where the layers of living vegetation were maximum (Fig. 5). 3.3. Environmental requirement for U. parviflorus germination Without any pre-treatments germination rate of U. parviflorus was 10% (half-time for germination = 43 days), while mechanically scarified seeds germinated at a rate of 97% (half-time for germination = 4 days) confirming the germination inhibition by the impermeability of the seed coat.
Fig. 6. Cumulative percentage germination of U. parviflorus seeds with heat pretreatments ranging from 20 8C (control) to 140 8C for 10 min. Different letters indicate significant differences (p < 0.001) between treatments.
690
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Table 2 Results from ANOVA comparing changes in germination percentage and germination rate (half-time for germination) among temperature and moisture fluctuation treatments applied for different length of time (d.f. error = 54) Treatment
Degree of freedom
Germination percentage F
p
F
p
Temperature Moisture Length Temperature Moisture Temperature Length Moisture Length Temperature Moisture Length
2 1 2 2 4 2 4
744.87 1.71 15.39 2.40 7.27 1.68 3.32
<0.001 0.19 <0.001 0.10 <0.001 0.19 0.02
7.95 14.84 14.89 3.34 3.03 4.55 2.85
0.001 <0.001 <0.001 0.04 0.03 0.02 0.03
3.3.1. Heat pulse response (simulating fire temperature) The final germination percentage as well as the half-time for germination were significantly affected by heat (Fig. 6) (d.f. 5; F = 217.22; p < 0.001 and d.f. 5; F = 132.10; p < 0.001 respectively). Ten minutes of heating at 80 or 60 8C resulted in the highest germination rates. 3.3.2. Temperature and moisture fluctuations (simulating summer soil surface conditions) Germination percentage was significantly affected by temperature fluctuations and its length but not by moisture fluctuations (Table 2 and Fig. 7). Germination percentages with 15–45 and 15– 55 8C fluctuations did not differ between each other but were significantly higher than in the control treatment. Moisture fluctuations increased germination percentage only in combination with the highest and longest temperature fluctuations 15– 55 8C. The length of the fluctuations affected germination but did so in strong interaction with temperature. At the highest temperature fluctuations, increasing the length significantly increased germination percentage, while at the lowest temperature fluctuations, the opposite was observed. The half-time for germination was significantly affected by all the treatments and their binary interactions, although only slightly so in the latter case (Table 2). At the highest temperature fluctuations, moisture decreased the half-time for germination while at the lowest temperature fluctuations, the opposite effect was observed. This same behaviour was seen with respect to the interactions between temperature and moisture fluctuations applied for different lengths of time. The interaction between moisture (dry and dry/wet) and moisture fluctuations indicates that the moisture fluctuations decrease the half-time for germination. The seeds that were not submitted to moisture fluctuation treatments showed the opposite effect.
Half-time for germination
3.3.3. Light quality (simulating vegetation removal) Temperature fluctuations and light quality significantly affected both the germination percentage and its half-time. Shade light quality (z = 0.4) resulted in the lowest germination percentages and increased germination half-times (Table 3). The interaction between temperature fluctuations and light quality is at the limit of significance for the germination percentage and is slightly significant for the germination half-time (Table 4). At the highest temperature fluctuations, half-time for germination increases slightly with z = 0.4, while at the lowest temperature fluctuations, half-time for germination is two times higher at z = 0.4 than at z = 1.1. 3.3.4. Comparison of the factors simulating different management options Comparing the most effective treatments for each of the different factors associated with pre- and post-management environmental conditions gives the following ranking: (1) premanagement condition, canopy z = 0.4, no temperature fluctuation (germination 8%); (2) no canopy z = 1.1, no temperature fluctuation (germination 12.4%); (3) simulation of vegetation removal, 15–45 8C for 30 dry days (germination 75.7%); (4) fire simulation, 10 min at 80 8C (germination 80.3%); (5) simulation of vegetation removal, 15–55 8C for 30 dry/wet days (germination 85%). These treatments had significant effects on the germination percentage (d.f. 5; F = 171.45; p < 0.001) and the half-time for germination (d.f. 5; F = 115.92; p < 0.001). The simulations of removing vegetation (with no slash) were not significatively
Table 3 Germination percentage and germination rate (half-time for germination) (mean S.D.) among temperature fluctuation treatments and light quality for 15 days moisture fluctuation Light quality/ temperature fluctuation
Germination percentage
Half-time for germination
z = 1.1
z = 0.4
z = 1.1
z = 0.4
20 8C (control) 15–45 8C 15–55 8C
12.4 (2.3) 63.7 (8.0) 60.1 (5.2)
8 (2.3) 48.2 (2.6) 39.2 (5.0)
27.9 (6.7) 22.8 (2.6) 22.2 (2.6)
53.4 (13.7) 27.8 (4.6) 25.1 (3.7)
Table 4 Results from ANOVA comparing changes in germination percentage and germination rate (half-time for germination) among temperature fluctuation treatments and light quality applied for 15 days moisture fluctuation (d.f. error = 18) Treatment
Fig. 7. Effect on germination percentage of daily temperature and every other day moisture fluctuations applied for 5, 15 or 30 days.
Temperature Light quality Temperature Light
Degree of freedom
2 1 2
Germination percentage
Half-time for germination
F
p
F
p
239.98 48.29 3.48
<0.001 <0.001 0.05
14.95 17.29 4.54
<0.001 0.001 0.02
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
different from the one simulating fire. The premanagement treatment and the treatment with ambient light quality and no temperature fluctuation were not significantly different, but both were significantly lower than all the other treatments. 4. Discussion 4.1. Germination ecology of U. parviflorus In Mediterranean fire-prone ecosystems, the enhanced germination and seedling establishment observed in the wet season following wildfires has been referred to as a fire-adaptive response (Zammit and Westoby, 1988; Moreno and Oechel, 1991; Keeley, 1994; Keeley and Bond, 1997). Temperature is the most important environmental factor involved in the softening of hard seeds (Baskin and Baskin, 1989) and heat treatments simulating wildfires have resulted in a strong increase of the germination of many Mediterranean species (Auld and O’Connell, 1991; Thanos et al., 1992; see also review in Keeley, 1994). Our field experiments (Fig. 1) and laboratory tests (Fig. 6) show that this is also the case for U. parviflorus, confirming the results of Baeza et al. (2002a) and Baeza and Vallejo (2006). Several mechanisms have evolved by plant species to regenerate after disturbances. Responding to microclimate fluctuations, which often increases after disturbances, allows plants to take advantage of gaps created in mature vegetation. Baskin and Baskin (1989) argued that high fluctuating temperatures associated with daily heating and cooling of the soil during summer render hardseeded species permeable. Our results show that daily cycles of temperature alternation (ranges of 15–45 and 15–55 8C) also enhance U. parviflorus germination as it was shown for U. europaeus (Zabkiewicz and Gaskin, 1978; Ivens, 1983) or Mimosa pigra (Lonsdale, 1993). Soil moisture fluctuations, which are also larger in gaps than under developed canopies, did not enhance germination of U. parviflorus. Gaps microclimate is also characterised by a higher red to farred ratio in the light spectrum reaching the soil surface. The low red to far-red ratio found under canopies has been shown to inhibit germination of many species (Gorski et al., 1978; Rees, 1997) through phytochrome conversion. Our results (Table 3) show that a low red to far-red ratio reduces germination by 25–35% whatever the temperature fluctuation regime. Light quality is then an additional factor limiting U. parviflorus regeneration under developed canopies. This was shown for a few other fire-proned species in Mediterranean climate (Roy and Arianoutsou-Faraggitaki, 1985; Roy and Sonie´, 1992). Most probably a composite germination syndrome developed in some Mediterranean species. We suggest that predictable seasonal climatic patterns (dry summer period with high temperatures followed by an autumn wet and mild period) together with recurrent fires could have participated to the selection of hard seed dormancy (broken by either fire heat or temperature fluctuations). Both these temperature responses as well as light quality response constitute a composite adaptive syndrome allowing species to germinate preferentially in open spaces were competition is reduced and when the dry summer period is over. 4.2. Implications for the control of Mediterranean gorse Our results indicate that the current method for reducing U. parviflorus populations, based on controlled fires, modifies the abiotic factors regulating seed germination in a way which facilitates the regeneration of this species. Baeza (2001) not only reported higher seedling establishment in burned areas than in
691
brush-chipped, but 2 years after treatment application, more seedlings reached reproductive maturity and more seeds were produced on the burned plots. This strong population growth may be promoted by ash fertilization and increased light availability. Since the processes involved in invasions by exotic species and colonizations by native species can be assumed to be essentially the same (Huston, 1994), this population development may be an illustration of the theory of fluctuating resource availability which predicts that a plant community becomes more susceptible to invasion whenever there is an increase in the amount of unused resources (Davis et al., 2000). Disturbances and clearing in particular have been shown to increase seedling recruitment in other species (Ivens, 1978; Lee et al., 1986; Sheppard et al., 2002). A reduction of the soil seed bank by prescribed burns of the standing vegetation as shown by Holmes (1989) for Acacia cyclops is rare. Our results also indicate that not only heat generated by fire can break seed dormancy in Mediterranean gorse. Other factors such as soil temperature fluctuations and light quality, which are modified by diverse vegetation clearing methods, are also efficient germination triggers. In mature and senescent stages after fire or land abandonment, U. parviflorus shrublands reach a continuous vertical and horizontal structure with standing necromass resulting in high fire risk (Baeza et al., 2006). Conducting clearing treatments at this development stage results in an accumulation of high amounts of litter in the soil surface that eventually may hamper U. parviflorus germination, similarly to brush-chipping treatment in our study. Indeed, germination of U. parviflorus was lower when slash resulting from brush-chipping was not removed. Similarly, Izhaki et al. (2000) in a study of Mediterranean Aleppo pine-forest soil-seedbank observed a reduction of Cistus spp. seed germination when soil was covered. Slash or mulch reduces germination through a buffering of soil surface temperature fluctuations and a slight reduction of the red/far-red ratio, but improved soil humidity can have an opposite effect, slash also reduces seedling development through physical impedance and reduction of light intensity (Teasdale and Mohler, 1993, 2000; Teasdale and Daughtry, 1993). However, slash or mulch effect on recruitment is differential. It reduces less the establishment of species with larger seeds (Teasdale and Mohler, 2000), a characteristic of late-successional species (Hewitt, 1998; Morin and Chuine, 2006). Recent investigations have shown a higher abundance of Quercus ilex seedlings in open shrublands as a result of acorn dispersal by the European jay (Garrulus glandarius) (Pons and Pausas, 2007). Slash also favours the establishment of resprouting species which have stored reserved allowing them to grow through the mulch layer. This is a crucial point in the restoring ecosystems after invasion by U. parviflorus since desirable species of later successional stages such as Q. ilex, Q. coccifera, and R. alaternus are favoured by this management treatment (Pausas, 2001; Baeza et al., 2005). For an effective management of U. parviflorus, our results suggest that neither standing vegetation nor slash should be burned. Instead clearing should be done by selective mechanical treatments in middle-aged or mature shrubland with slash remaining on the soil surface (Baeza and Vallejo, 2008). Our results also suggest that such vegetation clearing should preferably be done at the end of summer or beginning of autumn. When clearing was applied at the end of spring, germination took place that same autumn coinciding with the return of rains. When applied at the end of summer, seedling recruitment was minor the following autumn and spring despite abundant rain (144 mm for the months of October, November and December and 206 mm for the months of March, April and May). The largest seedling recruitment was delayed until the following autumn, which experienced only a marginally higher rainfall rate than in the
692
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
previous spring (265 mm for the months of October, November and December). This result originates from the strong impact of the large temperature fluctuations in summer which break seed dormancy and allow abundant germination in the following autumn rainy period. This remains true even when slash remains on the soil (Fig. 2) probably due to an incomplete buffering effect of slash or the heterogeneity of its distribution. Such an impact of the season at which disturbance occurs has been shown to affect seedling recruitment in the case of other disturbances (Bond and van Wilgen, 1996). In the case of fire, for example, season determines its intensity (greater during the dry period) (Hodgkinson, 1991; Moreno and Oechel, 1991) and the post-fire rainfall pattern (Whelan and Tait, 1995). It has been suggested that the present situation of high wildfire incidence is a consequence of extreme weather conditions (Moritz, 2003), and climate predictions point to increases in annual and summer temperatures in the Eastern Iberian Peninsula (Pausas, 2004), conditions that will enhance wildfire. Mediterranean gorse shrublands accumulate large amounts of dead biomass and they are dominated by seeder species with very few individuals of the more resilient late-successional sprouter species. Both of these factors make these communities particularly vulnerable to degradation processes, especially in areas of Southern Europe with a strong water deficit that limits ecosystems resilience (Puigdefa´bregas and Mendizabal, 1998; Vallejo et al., 2000, 2005). It has been argued that all fuel-control techniques should take into account the ecological characteristics and requirements that ensure the stability and sustainability of the ecosystem (Whelan, 1995; Dimitrakopoulos, 1997). Ideally, they should also alter the disturbance regime and/or the succession trajectory to create favourable establishment opportunities for late-successional species and unfavourable conditions for noxious species (Buckley et al., 2004). We suggest that a mechanical clearing of U. parviflorus shrublands which leaves mulch on the soil should be associated with the re-introduction by planting of lower flammability sprouter species to increase system resilience (Valdecantos et al., in press). This would be a sustainable management strategy that integrates both fuel reduction and land restoration. Acknowledgements We acknowledge L. Sonie´ and J. Fabreguettes for their technical collaboration and J. Scheiding for her linguistic revision. We thank Peter Fule´ and V.R. Vallejo for their comments and improvements to the manuscript. This work has been supported by Generalitat Valenciana post-doc contract to M.J. Baeza (Ref. CTESPP/2002/53), the program Consolider-Ingenio 2010 (GRACCIE CSD2007-00067) and Spanish Ministry of Education (FIREMAP CGL2004-06049C04-04). CEAM was supported by Generalitat Valenciana and Bancaixa. References Auld, T.D., 1986. Populations dynamics of the shrub Acacia suaveolens (Sm) Willd.: fire and the transition to seedlings. Aust. J. Ecol. 11, 373–385. Auld, T.D., O’Connell, M.A., 1991. Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Aust. J. Ecol. 16, 53–70. Baeza, M.J., 2001. Aspectos ecolo´gicos y te´cnicas de control del combustible (roza y quema controlada) en matorrales con alto riesgo de incendio dominados por Ulex parviflorus (Pourr.). Ph.D. Thesis, University of Alicante. http://www.cervantesvirtual.com/FichaObra.html?ReF=5920. Baeza, M.J., Vallejo, V.R., 2006. Ecological mechanisms involved in dormancy breakage in Ulex parviflorus seeds. Plant Ecol. 183, 191–205. Baeza, M.J., Vallejo, V.R., 2008. Vegetation recovery alter fuel management in Mediterranean shrublands. Appl. Veg. Sci. 11, 151–158. Baeza, M.J., Ravento´s, J., Escarre´, A., 2002a. Ulex parviflorus germination after experimental burning: effects of temperature and soil depth. In: Trabaud, L., Prodon, R. (Eds.), Fire and Biological Processes. Backhuys Publishers, Leiden, The Netherlands, pp. 83–91.
Baeza, M.J., De Luı´s, M., Ravento´s, J., Escarre´, A., 2002b. Factors influencing fire behaviour in shrublands of different stand ages and the implications for using prescribed burning to reduce wildfire risk. J. Environ. Manage. 65, 199–208. Baeza, M.J., Valdecantos, A., Vallejo, V.R., 2005. Management of Mediterranean Shrubland for Forest Fire Prevention. In: Burk, A.R. (Ed.), New Research on Forest Ecosystems. Nova Science Publishers Inc., New York, pp. 37–60. Baeza, M.J., Ravento´s, J., Escarre´, A., Vallejo, V.R., 2006. Fire risk and vegetation structural dynamics in Mediterranean shrublands. Plant Ecol. 187, 189–201. Baskin, M., Baskin, C., 1989. Physiology of dormancy and germination in relation to seed bank ecology. In: Leck, M.A., Parker, V.T., Simpson, R.L. (Eds.), Ecology of Soil Seed Banks. Academic Press Inc., San Diego, California, pp. 53–65. Baskin, C., Baskin, M., 1998. Seeds, Ecology, Biogeography and Evolution of Dormancy and Germination. Academic Press, London. Bond, W.J., van Wilgen, B.W., 1996. Fire and Plants. Chapman and Hall, London. Buckley, Y.M., Rees, M., Paynter, Q., Lonsdale, M., 2004. Modelling integrated weed management of an invasive shrub in tropical Australia. J. Appl. Ecol. 41, 547–560. Cramer, W., 2001. Fire ecology, Mediterranean forest and global change. Forest Ecol. Manage. 147, 1–2. Davis, M.A., Grime, J.P., Thompson, K., 2000. Fluctuating resources in plant communities: a general theory of invasibility. J. Ecol. 88, 528–534. Dimitrakopoulos, A.P., 1997. Wildland fire hazard reduction through natural resources management. In: Balabanis, P., Eftichidis, G., Fantechi, R. (Eds.), Forest Fire Risk and Management. European Commission, Brussels, pp. 207–214. Freckleton, R.P., 2004. The problems of prediction and scale in applied ecology: the example of fire as a management tool. J. Appl. Ecol. 41, 599–603. Gorski, T., Gorska, K., Rybicki, J., 1978. Studies on the germination of seeds under leaf canopy. Flora 167, 289–299. Greig-Smith, P., 1983. Quantitative Plant Ecology, 3rd ed. Blackwell, Oxford. Hewitt, N., 1998. Seed size and shade-tolerance: a comparative analysis of North American temperate trees. Oecologia 114, 432–440. Hodgkinson, K.C., 1991. Shrub recruitment response to intensity and season of fire in a semi-arid woodland. J. Appl. Ecol. 28, 60–70. Holmes, P.M., 1989. Effects of different clearing treatments on the seed bank dynamics of an invasive Australian shrub, Acacia cyclops, in the southwestern cape, South Africa. Forest Ecol. Manage. 28, 33–46. Huston, M.A., 1994. Biological Diversity. Cambridge University Press, Cambridge, UK. Ivens, G.W., 1978. Some aspects of seed ecology of gorse. In: Proceedings 31st New Zealand Weed and Pest Control Conference. pp. 53–57. Ivens, G.W., 1983. The influence of temperature on germination of gorse (Ulex europaeus L.). Weed Res. 23, 207–216. Izhaki, I., Henig-Sever, N., Ne’eman, G., 2000. Soil seed banks in Mediterranean Aleppo pine forest: the effect of heat, cover and ash on seedling emergence. J. Ecol. 88, 667–675. Keeley, J.E., 1994. Seed-germinations patterns in fire-prone Mediterranean-climate regions. In: Arroyo, M.T.K., Zedler, P.H., Fox, M.D. (Eds.), Ecology and Biogeography of Mediterranean Ecosystem in Chile, California and Australia. SpringerVerlag, New York, pp. 239–273. Keeley, J.E., Bond, W.J., 1997. Convergent seed germination in South African fynbos and Californian chaparral. Plant Ecol. 133, 153–167. Lee, W.G., Allen, R.B., Johnson, P.N., 1986. Succession and dynamics of gorse (Ulex europaeus L.) communities in the Dunedin Ecological District South Island, New Zealand. N Z J. Bot. 24, 279–292. Lonsdale, W.M., 1993. Losses from the seed bank of Mimosa pigra: soil microorganisms vs. temperature fluctuations. J. Appl. Ecol. 30, 654–660. Lloret, F., Vila`, M., 1996. Clearing of vegetation in Mediterranean garrigue: response after a wildfire. Forest Ecol. Manage. 93, 227–234. Me´thy, M., Fabreguettes, J., Jardon, F., Roy, J., 1987. Design of a simple instrument for the measurement of red/far-red ratios. Acta Oecol. 8, 281–290. Moreno, J.M., Oechel, W.C., 1991. Fire intensity effects on the germination of shrub and herbaceous species in southern California chaparral. Ecology 72, 1993– 2004. Morin, X., Chuine, I., 2006. Niche breadth, competitive strength and range size of tree species: a trade-off based framework to understand species distribution. Ecol. Lett. 9, 185–195. Moritz, M.A., 2003. Spatio temporal analysis of control on shrubland fire regimes: age dependency and fire hazard. Ecology 84, 351–361. Ooi, M.K., Auld, T.D., Whelan, R.J., 2006. Dormancy and the fire-centric focus: germination of three Leucopogon species (Ericaceae) from South-eastern Australia. Ann. Bot. 98, 421–430. Pausas, J., 2001. Resprouting vs seeding—a Mediterranean perspective. Oikos 94, 193–194. Pausas, J., 2004. Changes in fire and climate in the Eastern Iberian Peninsula (Mediterranean Basin). Clim. Change 63, 337–350. Paynter, Q., Flanagan, G.J., 2004. Integrating herbicide and mechanical control treatments with fire and biological control to manage an invasive wetland shrub Mimosa picra. J. Appl. Ecol. 41, 615–629. Pickett, S.T., White, P.S., 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York. ˜ ol, J., Terradas, J., Lloret, F., 1998. Climate warming, wildfire hazard and wildfire Pin occurrence in coastal eastern Spain. Clim. Change 38, 345–357. Pons, J., Pausas, J.G., 2007. Acorn dispersal estimated by radio-tracking. Oecologia 153, 903–911. Puigdefa´bregas, J., Mendizabal, T., 1998. Perpectives on desertification: western Mediterranean. J. Arid Environ. 39, 209–224.
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693 Rees, M., 1997. Seed dormancy. In: Crawley, M. (Ed.), Plant Ecology. Blackwell Science Ltd., London, pp. 214–238. Rees, M., Hill, R.L., 2001. Large-scale disturbances, biological control and the dynamics of gorse populations. J. Appl. Ecol. 38, 364–377. Roy, J., Arianoutsou-Faraggitaki, M., 1985. Light quality as the environmental trigger for the germination of the fire-promoted species Sarcopoterium spinosum L. Flora 177, 345–349. Roy, J., Sonie´, L., 1992. Germination and populations dynamics of Cistus species in relation to fire. J. Appl. Ecol. 29, 647–655. Scott, S.J., Jones, R.A., Williams, W.A., 1984. Review of data analysis methods for seed germination. Crop Sci. 24, 1192–1199. Sheppard, A.W., Hodge, P., Paynter, Q., Rees, M., 2002. Factors affecting invasion and persistence of broom Cytisus scoparius in Australia. J. Appl. Ecol. 39, 721–734. Teasdale, J.R., Daughtry, C.S.T., 1993. Weed suppression by live and dessicated hairy vetch (Vicia villosa). Weed Sci. 41, 207–212. Teasdale, J.R., Mohler, C.L., 1993. Light transmittance, soil temperature and soil moisture under residue of hairy vetch and rye. Agron. J. 85, 673–680. Teasdale, J.R., Mohler, C.L., 2000. The quantitative relationship between weed emergence and the physical properties of mulches. Weed Sci. 48, 385–392. Thanos, C.A., Georghiou, K., Kadis, C., Pantazi, C., 1992. Cistaceae: a plant family with hard seeds. Isr. J. Bot. 41, 251–263. Thompson, K., Ceriani, R.M., Bakker, J.P., Bekker, R.M., 2003. Are seed dormancy and persistence in the soil related? Seed Sci. Res. 13, 97–100.
693
Tutin, T.G., Heywood, V.H., Burges, N.A., Moore, D.M., Valentine, D.H., Walters, S.M., Webbs, D.A. (Eds.), 1964–1980. Flora Europaea, vol. 5. Cambridge Univ. Press, Cambridge. Valdecantos, A., Baeza, M.J., Vallejo, V.R. Vegetation management for promoting ecosystem resilience in fire-prone Mediterranean shrublands. Rest. Ecol., in press. Vallejo, V.R., Bautista, S., Cortina, J., 2000. Restoration for soil protection after disturbances. In: Trabaud, L. (Ed.), Life and Environment in the Mediterranean. Wit Press, Southampton, Boston, pp. 301–343. Vallejo, V.R., Aronson, J., Pausas, J.G., Cortina, J., 2005. Restoration of Mediterranean woodlands. In: Van Andel, J., Aronson, J. (Eds.), Restoration Ecology from European Perspective. Blackwell Science, pp. 193–207. Whelan, R.J., 1995. The Ecology of Fire. Cambridge University Press, New York. Whelan, R.J., Tait, I., 1995. Responses of plant populations to fire: fire season as an under-studied element of fire regime. CALM Sci. Suppl. 4, 147–150. Zabkiewicz, J.A., Gaskin, R.E., 1978. Effect of fire on gorse seeds. In: Proceedings of 31st New Zealand Weed and Pest Control Conference. pp. 47–52. Zammit, C.A., Westoby, M., 1988. Pre-dispersal seed losses, and survival of seeds and seedlings of two serotinous Banksia shrubs in burnt and unburnt heath. J. Ecol. 76, 200–214. Zedler, P.H., Zammit, C.A., 1989. A populations-based critique on concepts of change in the chaparral. Sciences series N8 34. In: Keeley, S.C. (Ed.), Chaparral Ecology: Paradigms Re-examined. Natural History Museum of Los Angeles Country, Allen Press, Los Angeles, California.
Contents lists available at ScienceDirect
Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco
Germination of an obligate seeder (Ulex parviflorus) and consequences for wildfire management M. Jaime Baeza a,b,*, Jacques Roy c a
Centro de Estudios Ambientales del Mediterra´neo, CEAM, Parque Tecnolo´gico Paterna, Charles Darwin, 14, 46980 Valencia, Spain Departamento de Ecologı´a, Universidad de Alicante, Ap. 99, 03080 Alicante, Spain c Centre d’Ecologie Fonctionnelle et Evolutive, CNRS, UMR5175, 34293 Montpellier cedex 5, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 March 2008 Received in revised form 12 May 2008 Accepted 13 May 2008
In fire-prone ecosystems, reducing the risk of wildfire is generally attempted through vegetation clearing using controlled fires or, less often, mechanical techniques. Management practices, however, can be poorly efficient when the disturbances they introduce install environmental conditions that are similar to the ones under which the undesirable species evolved. Ulex parviflorus is a Mediterranean obligate seeder with physical dormancy forming large amounts of highly combustible standing necromass. In the present study, combining field and laboratory experiments, we determined the seedling recruitment of this species under different management practices (burning, mechanical clearing, slash/no slash on soil surface), we measured the environmental conditions (temperature, light, and moisture) enforced by these practices and we tested their individual and combined impact on germination in order to determine the most appropriate control method for this species. Germination is low under intact canopies, but it is strongly stimulated by both brush-chipping and fire. This is partly related to the inhibiting effect of the low red:far-red ratio of the light filtered by the canopy which is removed by brush-chipping and fire. The other factor involved is moderate heat, either fire-generated or resulting from solar radiation on bare soil, which breaks seed coat impermeability. Indeed exposing seeds on bare soil in summer resulted in a significant increase in their germination capacity and germination was reduced when the brush chips remained on the soil. Moisture fluctuations did not enhance germination. The summer heat impact affects management practices. When the brush-chipping treatment occurred before summer, the germination flush appeared the following autumn, but when the treatment occurred after summer, the germination flush did not appear until the autumn of the subsequent year, when interspecific competition with regenerating vegetation is likely to be more intense. We demonstrated that brushchipping, especially when done after summer, is a better technique than fire for controlling U. parviflorus because it creates environmental conditions that are less favourable for its germination. This technique also has the potential to favour late-successional species less vulnerable to fire. By combining fuel reduction and land restoration, this technique is useful to the sustainable management policies that need to be developed. ß 2008 Elsevier B.V. All rights reserved.
Keywords: Clearing Control burning Fuel reduction Light quality Moisture fluctuations Slash Temperature fluctuations
1. Introduction Wildfires constitute a serious risk in many regions of the world, and their frequency is expected to increase as a result of climate change and land abandonment (fuel accumulation), in particular in ˜ ol et al., 1998; Cramer, 2001; Pausas, Mediterranean regions (Pin 2004). Vegetation control for reducing fuel loads in forest and
* Corresponding author at: Centro de Estudios Ambientales del Mediterra´neo, CEAM, Parque Tecnolo´gico Paterna, Charles Darwin, 14, 46980 Valencia, Spain. Fax: +34 96 590 98 25. E-mail address: [email protected] (M.J. Baeza). 0378-1127/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2008.05.014
shrubland ecosystems is often applied to decrease the fire risk and improve the probability of successful fire control by reducing fire intensity. Several management techniques have been developed to control undesirable species, including mechanical clearing, herbicides, fire, grazing, biological control agents or a combination of these (Paynter and Flanagan, 2004; Buckley et al., 2004). However, despite the stressfulness of these management techniques, the fact that natural selection in many species occurred under drastic disturbance regimes (Pickett and White, 1985) can reduce the effectiveness of these efforts. This is particularly the case in biomes where fire has been part of the selective forces shaping biodiversity. Such species have developed very effective regenera-
686
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
tion mechanisms, which are triggered by the environmental conditions associated with fire. Nevertheless, Zedler and Zammit (1989) pointed out that few of these conditions are specific to fire (very high temperatures, ash or charcoal covering) while most are shared with other disturbances, including management practices (changes in light intensity and quality, temperature, moisture, competition). Thus, fuel reduction management could have the counterproductive effect of facilitating the germination of undesirable species. For a successful vegetation management programme, understanding the ecological processes involved in recruitment are of key importance to generate unfavourable opportunities for undesirable species regeneration (Freckleton, 2004). Ulex parviflorus Pourr., Mediterranean gorse, is a shrub of the Fabaceae family found all along the Mediterranean coasts (and up to 100 km inland) of Southwestern Europe and North Africa (Tutin et al., 1964–1980). It is an obligate seeder with a physical exogenous dormancy on the seed (Baeza and Vallejo, 2006), which develops massively after land abandonment and fire and can constitute the dominant species of dense thickets, as is the case in Southern Spain. When senescence occurs (after 20–25 years), these stands constitute large amounts of highly combustible standing necromass (Baeza et al., 2006). Management strategies to control the extension of U. parviflorus are now being studied and applied; among these, controlled fires and clearings are the most common (Lloret and Vila`, 1996; Baeza et al., 2002b). However, experimental burning and brush-chipping have shown the difficulty of controlling U. parviflorus due to its massive regeneration through seed germination (Baeza, 2001). The elimination of existing vegetation and the high temperature generated in the fire can scarify the seed coat and break the dormancy of seeds stored in the soil (Baskin and Baskin, 1998). Similar environmental problems are caused on a larger scale by other leguminous species like Ulex europaeus L. and Cytisus scoparius (L.) Link, both of which are native to Europe and are now regarded as noxious weeds in New Zealand, Chile, Hawaii, North America and Australia (Rees and Hill, 2001; Sheppard et al., 2002). Leguminous species have an impermeable seed coat that imposes physical dormancy on the seed; to germinate the waterimpermeable layer must become permeable (Baskin and Baskin, 1998). In Mediterranean-type ecosystems, heat from fire has been directly correlated with seedling recruitment because high temperatures can break hard seed coat dormancy, making seeds permeable and ready to germinate under favourable environmental conditions. This temporal sequence indicates that the environmental factors involved in seed dormancy breakage may differ from those involved in germination (Thompson et al., 2003). In nature, fire is the most extreme example of temperature overcoming dormancy in hard-seeded species (Auld, 1986); other factors associated with temperature could also be related to dormancy loss (Ooi et al., 2006). Under field conditions, combined temporal sequences of high temperatures in summer followed by moist conditions in autumn occur seasonally. By uncoupling these seasonal effects management programmes may be able to delay the emergence flush in species with physical dormancy and improve control efficiency. In this paper we analyse the germination conditions of a largely undesirable shrub in relation to its current management practices and we discuss the suitability of these practices. We study (i) the in situ germination of U. parviflorus after different management treatments, (ii) the environmental soil surface conditions created by these treatments and (iii) the germination requirements of the seeds in controlled conditions. We then compare the relative efficiency of fire-specific and general-disturbance germination clues and discuss the techniques used for managing this species in the light of its germination ecology.
2. Material and methods 2.1. Study area and sites The study was conducted in the Mariola and Aitana interior mountains in the northwest of the Alicante province, Spain. The typically Mediterranean climate is characterised by a mean annual precipitation of 495 mm (Onil, 388 390 N, 08 400 W, 800 m above sea level, 40 km from the sea) with a summer drought from mid-May to mid-September. The mean temperature is 14.5 8C with mean monthly temperatures ranging from 7.5 8C in January to 23.5 8C in ˜ eres, Castell and Onil), were chosen July. Four sites (Alcoy, Ban within a 40 km distance. Their climatic conditions, bedrock (marl) and soil (deep loamy Calcaric Cambisol, FAO 1998) are similar. All 4 sites were cultivated terraces abandoned after the Spanish Civil war (1939). They were covered with Aleppo pine (Pinus halepensis Mill.) forests until a wildfire in 1985 (Alcoy and Onil) or 1991 ˜ eres and Castell). At the beginning of the field experiments (Ban (1994), the shrubland vegetation was dominated by U. parviflorus and to a lesser extend by other obligate seeders such as Cistus albidus L. and Rosmarinus officinalis L. A few isolated individuals of woody resprouters were present (Quercus coccifera L., Juniperus oxycedrus L. and Rhamnus alaternus L.) as well as a few small isolated individuals of P. halepensis. The herbaceous layer, covering approximately 70% of the ground, was dominated by Brachypodium retusum (Pers.) Beau. 2.2. Field experiments The four sites were part of a larger experiment designed to study the impact of successional stage and fuel load on fire behaviour and its implications for using control burning or clearing to reduce wildfire risk (Baeza et al., 2002b). In each site, three treatments were applied in May 1994, each on a single 1000 m2 plot: control, brush-chipping (slash remaining on the soil) and burning. Once a month until January 1996, U. parviflorus seedlings occurrence was recorded in 20 randomly distributed 0.5 m 0.5 m sub-plots within each plot. In Onil, three 1000 m2 plots were brush-chipped for a fuelbreak experiment in September 1996. The effect of keeping the chips on the ground (slash with 100% cover) or removing them (no slash, bare soil) on U. parviflorus germination was tested. In each plot, the chips were removed in five randomly distributed 1 m 1 m sub-plots and kept in five other randomly distributed 1 m 1 m sub-plots. U. parviflorus seedling occurrence was recorded in all these sub-plots from October 1996 until May 1998. In Onil, the impact of seed exposure to soil surface summer conditions was tested using seeds collected in June 2001 from ˜ eres and in Castell. For each seed origin, 20 sets mature fruits in Ban of 50 seeds were enclosed in bags made with nylon mesh (0.5 mm). Five sets were kept as control while fifteen were put on the bare ground of a fuel-break on July 1, 2001. Five sets were removed after 30 days, five sets after 60 days and the remaining 5 sets after 90 days. Soil surface temperature was measured and recorded every hour with a temperature probe (Hobo1 Event, Onset Computer Corporation, Bourne MA, US). Each set of seeds was then put in a Petri dish on filter paper and enclosed in a germination chamber (12 h light at 22 8C, 80% moisture and 12 h dark at 20 8C and 85% moisture). Filter paper moisture and seedling germination were checked every other day for 60 days. These field experiments suggested that fluctuations in soil temperature and moisture as well as heat generated by fire or change in light quality due to green leaves removal could stimulate U. parviflorus germination. These physical characteristics were then documented in the field before testing their isolated or
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
interacting effects in controlled conditions. Fluctuations of the soil surface temperature in a fire-break (with no slash) and under an adjacent mixed forest was compared in Onil using one temperature probe in each environment (Hobo1 Event, Onset Computer Corporation, Bourne MA, US). The probe thermocouple was inserted in the first mm of the soil and the measurements were taken during August and September 1998. The number of rainy days between seed set in May and seed germination in November was measured at the weather station ˜ eres for the period 1950–1990. The relation between light of Ban quality and vegetation cover was studied in November 2002 in the vicinity of the Castell site in two Mediterranean gorse shrublands 4 and 17 years old respectively. Every 25 cm along one 30 m transect in each shrubland, vegetation cover was estimated by the number of contacts along a vertical point quadrat (Greig-Smith, 1983). Light quality (ratio between 660 and 730 nm wavelength) was measured at the base of each quadrat using the red/far-red ratio sensor described in Me´thy et al. (1987). 2.3. Germination in controlled conditions ˜ eres Seeds were collected in June 2002 from mature fruits in Ban and stored in paper bags at room conditioned temperature. Seed pre-treatments and germination were conducted at the Centre of Functional and Evolutionary Ecology in Montpellier from October 2002 to February 2003. 2.3.1. Seed pre-treatments Petri dishes (9 cm diameter) were filled with the following layers: 1 cm of soil, one filter paper, 50 seeds, one filter paper and 0.5 cm of soil. Four Petri dishes (replicates) were used for each treatment level. - Temperature fluctuations (simulating summer bare soil surface temperature): Two germination chambers were set up with a 15 8C night temperature for 11 h and 45 and 55 8C day temperature respectively for 13 h. Temperature within each chamber was measured with a thermocouple and recorded every hour with a data-logger (CR21X, Campbell Scientific, Utah, USA). Room temperature (control) varied between 18 and 22 8C. - Moisture fluctuations (simulating wetting of the seeds by rainfall events): Half the Petri dishes enclosed in the germination chambers, as well as half the ones kept as control for the temperature fluctuations were watered every other day with 30 ml of distilled water. The other Petri dishes were left dry. - Length of these pre-treatments: The above two fluctuation treatments were applied for 5, 15 or 30 days. The corresponding number of dry/wet periods was 2, 7 and 15 respectively. The 30 days treatment started first, the 15 days one started 15 days later and the 5 days one started 25 days after the first one so that the germination could start simultaneously for all treatments with
687
an identical lap of time since the end of the pre-treatments. The treatments temperature, moisture and length of fluctuations were crossed. - Heat (simulating fire heat): The Petri dishes (with dry seeds) were introduced for 10 min in a preheated oven at 80 8C, the optimum short term heating for triggering germination. Due to a lack of ˜ eres, this optimum had been determined on seeds from Ban another population (Confrides) whose seeds were introduced the same way in a chamber at 60, 80, 100, 120 or 140 8C for 10 min. - Mechanical scarification (to estimate the potential maximum germination): Seeds were individually scarified with the blade of a scalpel. 2.3.2. Seed germination Seed were recovered from in between the filter papers of the pre-treatment Petri dishes. Each set of 50 seeds was then enclosed in a small mesh bag and soaked in a 1% sodium hypochlorite solution for 10 min to eliminate fungal contamination. The seeds were then rinced twice with distilled water. Each set of 50 seeds was spread on a filter paper in Petri dishes (9 cm diameter). The filter paper had been deposited on one layer of 0.6 mm glass beads allowing for a water reserve and anti-fungal Tachigaren LS1 (Hymexazol 0.5%). The Petri dishes were enclosed in three germination chambers set up with a 17 8C night temperature for 11 h and a 23 8C day temperature for 13 h. Day light was provided by two 8 W fluorescent bulbs in each chamber. In order to simulate light quality typical of a deep under canopy shade, far-red wavelengths were added in one of the germination chamber by a 40 W red-coloured incandescent lamp (Claude) surrounded by a blue filter (Altuglas S300, Altulor). Red/far-red ratio was 0.4 in this chamber compared to 1.1 in the other two chambers. The impact of this increased radiation was tested on seeds which had received the pre-treatment temperature and moisture fluctuations for 15 days. The 24 combinations of treatments (with 4 Petri dishes per combination) are given in Table 1. Germination (i.e. visible radicle) was recorded for 97 days, twice a week during 60 days and then once a week. 2.4. Statistical analysis The effect of vegetation treatments (control, brush-chipping and burning) and the presence of remains on the ground (slash-noslash) on U. parviflorus germination dynamics were analysed using repeated measures ANOVA. Mauchly’s W test statistic was used for testing sphericity assumptions (independence of data among repeated measures). For data found to be correlated among years, the Geisser-Greenhouse correction was used to adjust the degrees of freedom of the within-subjects. The time course of germination was investigated by fitting the curve of the cumulative germination (y) across time (t) with a sigmoid function (Scott et al., 1984)
Table 1 ˜ eres seeds (n = 4) Pre-treatment combinations for the Ban Length of fluctuations
5 days 15 days 15 days 30 days Heat 80 8C Mechanical
Light quality
z = 1.1 z = 1.1 z = 0.4 z = 1.1 z = 1.1 z = 1.1
Temperature and moisture fluctuations Room temperature
15–45 8C
Dry
Dry/wet
Dry
Dry/wet
Dry
Dry/wet
15–55 8C
688
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Fig. 1. Time course of field germination of U. parviflorus under intact canopies or after brush-chipping (slash remaining on the soil) or fire. The canopies were chipped or ˜ eres and (b) Castell, and a 9-year-old regrowth after fire at (c) Onil and (d) burned in May 1994 (arrows). At that time, vegetation was a 3-year-old regrowth after fire at (a) Ban Alcoy.
using Sigma Plot 7.0 programme (SPSS Inc., Chicago, IL, USA). The following sigmoid function for which y = 0 when t = 0 was chosen: y ¼ A½1 expðBtÞc The final number of germinated seeds (A, the asymptote of the curve) and the mean germination time (the time to 50% germination, t50 = [ln(1 exp(0.51/c))]/B, were derived from this function rather than from the raw data because in some treatments the germination was not completely finished when the experiment was stopped. The impact of seed exposition to soil surface summer conditions with four levels of duration (0, 30, 60 and 90 days) and two levels per seed origin was analysed by twoway ANOVA. Linear regression analysis was used to determine the relationship between light quality and presence of living vegetation components. Germination percentage and half-time for germination were analysed by three-way (temperature, moisture and length of the fluctuations pre-treatment) ANOVAs with all factors as fixed effects (temperature fluctuations with three levels (20–20, 15–45 and 15–55 8C), moisture fluctuations with two levels (dry and dry/wet) and length fluctuations with three levels (5, 15 and 30 days). We used two-way ANOVAs to determine the significance of the two fixed factors temperature fluctuation during 15 days with tree levels (20–20, 15–45 and 15– 55 8C) and light quality with two levels (red/far-red ratio 0.4 and 1.1), on germination percentage and half-time for germination. Finally we compared with ANOVA the impact of the different factors (scarification, fire simulation, light quality, intensity fluctuations and temperature + moisture fluctuations) on germination percentage and half-time for germination. Data were ln transformed (number of seedling per quadrat and the time to 50% germination) or arc-sine transformed (germination percentage) to correct deviations from normality. Post hoc comparisons of means were done by Tukey’s multiple range test. Standard error
bars are shown on the graphs. All the statistical analyses were performed using the SPSS 10.0 package (SPSS Inc., Chicago, IL, USA). 3. Results 3.1. Germination after in situ vegetation management treatments Germination of U. parviflorus occurs in autumn during the month following the first rainfall events. It was low under intact canopies but was strongly and significantly stimulated by both brush-chipping and fire (d.f. 2; F = 30.15; p < 0.001, Fig. 1). Germination was extremely low the second autumn after the treatments although September to December rainfall was higher in 1995 than in 1994 (187 and 129 mm respectively). After brush-chipping, germination was significantly reduced when the brush chips (slash) remain on the soil as compared to bare soil (d.f. 1; F = 17.19; p < 0.001, Fig. 2). When the brushchipping treatment occurred before summer, germination flush appeared the autumn of the same year (Fig. 1), but when the treatment occurred when the summer was over, germination flush did not appear until the autumn of the subsequent year (Fig. 2). Since rainfall was not particularly limiting germination in 1996 (September to December rainfall was 179 and 155 mm in 1996 and 1997 respectively), results of Figs. 1 and 2 suggest that the summer environmental conditions are needed to trigger the germination of U. parviflorus when the canopy is eliminated. Exposure the seeds on the soil surface of a fire-break during the summer indeed resulted in a significant increase in germination percentage measured in a germination chamber (d.f. 3; F = 12.1; p < 0.05, no difference between 30, 60 and 90 days of exposure) (Fig. 3).
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Fig. 2. Time course of field germination of U. parviflorus after brush-chipping when slash is removed or when it remains on the soil. The canopy (at Onil, 11 years old) was chipped in September 1996 (arrow).
689
Fig. 5. Relationship between light quality (red/far-red ratio) and the density of the vegetation measured as the number of contacts of vertical point quadrats with living plants. 240 measurements were taken along 30 m transects in 4 and 17 years old gorse shrublands.
(Fig. 4). Under the canopy of an adjacent mixed forest, these fluctuations were buffered, typically between 16 and 30 8C with peaks at 33 8C. The top soil layer and consequently the seed bank also experienced alternating periods of wetness and dryness. In average during the dry period, 2 days with rainfall >5 mm were ˜ eres between recorded per month at the weather station of Ban 1950 and 1990 (3.2; 1.8; 0.6; 0.8; 2; 2.9 mm from May to October respectively).
Fig. 3. Effect on subsequent germination of exposing the seeds to the summer soil surface conditions in a fire-break without slash. Seed were not exposed (0 days) or ˜eres and Confrides were used. exposed 30, 60 and 90 days. Seeds of two origins Ban Different letters indicate significant differences (p < 0.05) between exposure length.
3.2. Microsite conditions during or after vegetation removal 3.2.1. Summer soil surface conditions Temperature in August and early September typically fluctuated between 14 8C at night and 45 8C during the day at the soil surface of a fire-break (with no slash). Peaks at 50 8C were reached
Fig. 4. Daily maximum and minimum temperatures in August and September on the soil surface of a fire-break (with no slash) and of an adjacent mixed forest.
3.2.2. Light quality and vegetation cover The ratio between the 660 and 730 nm wavelengths was measured in two Mediterranean gorse shrublands 4 and 17 years old respectively. This ratio decreased as expected with the density of the vegetation, from a value of 1.1 in the incident radiation to a value of 0.6 where the layers of living vegetation were maximum (Fig. 5). 3.3. Environmental requirement for U. parviflorus germination Without any pre-treatments germination rate of U. parviflorus was 10% (half-time for germination = 43 days), while mechanically scarified seeds germinated at a rate of 97% (half-time for germination = 4 days) confirming the germination inhibition by the impermeability of the seed coat.
Fig. 6. Cumulative percentage germination of U. parviflorus seeds with heat pretreatments ranging from 20 8C (control) to 140 8C for 10 min. Different letters indicate significant differences (p < 0.001) between treatments.
690
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
Table 2 Results from ANOVA comparing changes in germination percentage and germination rate (half-time for germination) among temperature and moisture fluctuation treatments applied for different length of time (d.f. error = 54) Treatment
Degree of freedom
Germination percentage F
p
F
p
Temperature Moisture Length Temperature Moisture Temperature Length Moisture Length Temperature Moisture Length
2 1 2 2 4 2 4
744.87 1.71 15.39 2.40 7.27 1.68 3.32
<0.001 0.19 <0.001 0.10 <0.001 0.19 0.02
7.95 14.84 14.89 3.34 3.03 4.55 2.85
0.001 <0.001 <0.001 0.04 0.03 0.02 0.03
3.3.1. Heat pulse response (simulating fire temperature) The final germination percentage as well as the half-time for germination were significantly affected by heat (Fig. 6) (d.f. 5; F = 217.22; p < 0.001 and d.f. 5; F = 132.10; p < 0.001 respectively). Ten minutes of heating at 80 or 60 8C resulted in the highest germination rates. 3.3.2. Temperature and moisture fluctuations (simulating summer soil surface conditions) Germination percentage was significantly affected by temperature fluctuations and its length but not by moisture fluctuations (Table 2 and Fig. 7). Germination percentages with 15–45 and 15– 55 8C fluctuations did not differ between each other but were significantly higher than in the control treatment. Moisture fluctuations increased germination percentage only in combination with the highest and longest temperature fluctuations 15– 55 8C. The length of the fluctuations affected germination but did so in strong interaction with temperature. At the highest temperature fluctuations, increasing the length significantly increased germination percentage, while at the lowest temperature fluctuations, the opposite was observed. The half-time for germination was significantly affected by all the treatments and their binary interactions, although only slightly so in the latter case (Table 2). At the highest temperature fluctuations, moisture decreased the half-time for germination while at the lowest temperature fluctuations, the opposite effect was observed. This same behaviour was seen with respect to the interactions between temperature and moisture fluctuations applied for different lengths of time. The interaction between moisture (dry and dry/wet) and moisture fluctuations indicates that the moisture fluctuations decrease the half-time for germination. The seeds that were not submitted to moisture fluctuation treatments showed the opposite effect.
Half-time for germination
3.3.3. Light quality (simulating vegetation removal) Temperature fluctuations and light quality significantly affected both the germination percentage and its half-time. Shade light quality (z = 0.4) resulted in the lowest germination percentages and increased germination half-times (Table 3). The interaction between temperature fluctuations and light quality is at the limit of significance for the germination percentage and is slightly significant for the germination half-time (Table 4). At the highest temperature fluctuations, half-time for germination increases slightly with z = 0.4, while at the lowest temperature fluctuations, half-time for germination is two times higher at z = 0.4 than at z = 1.1. 3.3.4. Comparison of the factors simulating different management options Comparing the most effective treatments for each of the different factors associated with pre- and post-management environmental conditions gives the following ranking: (1) premanagement condition, canopy z = 0.4, no temperature fluctuation (germination 8%); (2) no canopy z = 1.1, no temperature fluctuation (germination 12.4%); (3) simulation of vegetation removal, 15–45 8C for 30 dry days (germination 75.7%); (4) fire simulation, 10 min at 80 8C (germination 80.3%); (5) simulation of vegetation removal, 15–55 8C for 30 dry/wet days (germination 85%). These treatments had significant effects on the germination percentage (d.f. 5; F = 171.45; p < 0.001) and the half-time for germination (d.f. 5; F = 115.92; p < 0.001). The simulations of removing vegetation (with no slash) were not significatively
Table 3 Germination percentage and germination rate (half-time for germination) (mean S.D.) among temperature fluctuation treatments and light quality for 15 days moisture fluctuation Light quality/ temperature fluctuation
Germination percentage
Half-time for germination
z = 1.1
z = 0.4
z = 1.1
z = 0.4
20 8C (control) 15–45 8C 15–55 8C
12.4 (2.3) 63.7 (8.0) 60.1 (5.2)
8 (2.3) 48.2 (2.6) 39.2 (5.0)
27.9 (6.7) 22.8 (2.6) 22.2 (2.6)
53.4 (13.7) 27.8 (4.6) 25.1 (3.7)
Table 4 Results from ANOVA comparing changes in germination percentage and germination rate (half-time for germination) among temperature fluctuation treatments and light quality applied for 15 days moisture fluctuation (d.f. error = 18) Treatment
Fig. 7. Effect on germination percentage of daily temperature and every other day moisture fluctuations applied for 5, 15 or 30 days.
Temperature Light quality Temperature Light
Degree of freedom
2 1 2
Germination percentage
Half-time for germination
F
p
F
p
239.98 48.29 3.48
<0.001 <0.001 0.05
14.95 17.29 4.54
<0.001 0.001 0.02
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
different from the one simulating fire. The premanagement treatment and the treatment with ambient light quality and no temperature fluctuation were not significantly different, but both were significantly lower than all the other treatments. 4. Discussion 4.1. Germination ecology of U. parviflorus In Mediterranean fire-prone ecosystems, the enhanced germination and seedling establishment observed in the wet season following wildfires has been referred to as a fire-adaptive response (Zammit and Westoby, 1988; Moreno and Oechel, 1991; Keeley, 1994; Keeley and Bond, 1997). Temperature is the most important environmental factor involved in the softening of hard seeds (Baskin and Baskin, 1989) and heat treatments simulating wildfires have resulted in a strong increase of the germination of many Mediterranean species (Auld and O’Connell, 1991; Thanos et al., 1992; see also review in Keeley, 1994). Our field experiments (Fig. 1) and laboratory tests (Fig. 6) show that this is also the case for U. parviflorus, confirming the results of Baeza et al. (2002a) and Baeza and Vallejo (2006). Several mechanisms have evolved by plant species to regenerate after disturbances. Responding to microclimate fluctuations, which often increases after disturbances, allows plants to take advantage of gaps created in mature vegetation. Baskin and Baskin (1989) argued that high fluctuating temperatures associated with daily heating and cooling of the soil during summer render hardseeded species permeable. Our results show that daily cycles of temperature alternation (ranges of 15–45 and 15–55 8C) also enhance U. parviflorus germination as it was shown for U. europaeus (Zabkiewicz and Gaskin, 1978; Ivens, 1983) or Mimosa pigra (Lonsdale, 1993). Soil moisture fluctuations, which are also larger in gaps than under developed canopies, did not enhance germination of U. parviflorus. Gaps microclimate is also characterised by a higher red to farred ratio in the light spectrum reaching the soil surface. The low red to far-red ratio found under canopies has been shown to inhibit germination of many species (Gorski et al., 1978; Rees, 1997) through phytochrome conversion. Our results (Table 3) show that a low red to far-red ratio reduces germination by 25–35% whatever the temperature fluctuation regime. Light quality is then an additional factor limiting U. parviflorus regeneration under developed canopies. This was shown for a few other fire-proned species in Mediterranean climate (Roy and Arianoutsou-Faraggitaki, 1985; Roy and Sonie´, 1992). Most probably a composite germination syndrome developed in some Mediterranean species. We suggest that predictable seasonal climatic patterns (dry summer period with high temperatures followed by an autumn wet and mild period) together with recurrent fires could have participated to the selection of hard seed dormancy (broken by either fire heat or temperature fluctuations). Both these temperature responses as well as light quality response constitute a composite adaptive syndrome allowing species to germinate preferentially in open spaces were competition is reduced and when the dry summer period is over. 4.2. Implications for the control of Mediterranean gorse Our results indicate that the current method for reducing U. parviflorus populations, based on controlled fires, modifies the abiotic factors regulating seed germination in a way which facilitates the regeneration of this species. Baeza (2001) not only reported higher seedling establishment in burned areas than in
691
brush-chipped, but 2 years after treatment application, more seedlings reached reproductive maturity and more seeds were produced on the burned plots. This strong population growth may be promoted by ash fertilization and increased light availability. Since the processes involved in invasions by exotic species and colonizations by native species can be assumed to be essentially the same (Huston, 1994), this population development may be an illustration of the theory of fluctuating resource availability which predicts that a plant community becomes more susceptible to invasion whenever there is an increase in the amount of unused resources (Davis et al., 2000). Disturbances and clearing in particular have been shown to increase seedling recruitment in other species (Ivens, 1978; Lee et al., 1986; Sheppard et al., 2002). A reduction of the soil seed bank by prescribed burns of the standing vegetation as shown by Holmes (1989) for Acacia cyclops is rare. Our results also indicate that not only heat generated by fire can break seed dormancy in Mediterranean gorse. Other factors such as soil temperature fluctuations and light quality, which are modified by diverse vegetation clearing methods, are also efficient germination triggers. In mature and senescent stages after fire or land abandonment, U. parviflorus shrublands reach a continuous vertical and horizontal structure with standing necromass resulting in high fire risk (Baeza et al., 2006). Conducting clearing treatments at this development stage results in an accumulation of high amounts of litter in the soil surface that eventually may hamper U. parviflorus germination, similarly to brush-chipping treatment in our study. Indeed, germination of U. parviflorus was lower when slash resulting from brush-chipping was not removed. Similarly, Izhaki et al. (2000) in a study of Mediterranean Aleppo pine-forest soil-seedbank observed a reduction of Cistus spp. seed germination when soil was covered. Slash or mulch reduces germination through a buffering of soil surface temperature fluctuations and a slight reduction of the red/far-red ratio, but improved soil humidity can have an opposite effect, slash also reduces seedling development through physical impedance and reduction of light intensity (Teasdale and Mohler, 1993, 2000; Teasdale and Daughtry, 1993). However, slash or mulch effect on recruitment is differential. It reduces less the establishment of species with larger seeds (Teasdale and Mohler, 2000), a characteristic of late-successional species (Hewitt, 1998; Morin and Chuine, 2006). Recent investigations have shown a higher abundance of Quercus ilex seedlings in open shrublands as a result of acorn dispersal by the European jay (Garrulus glandarius) (Pons and Pausas, 2007). Slash also favours the establishment of resprouting species which have stored reserved allowing them to grow through the mulch layer. This is a crucial point in the restoring ecosystems after invasion by U. parviflorus since desirable species of later successional stages such as Q. ilex, Q. coccifera, and R. alaternus are favoured by this management treatment (Pausas, 2001; Baeza et al., 2005). For an effective management of U. parviflorus, our results suggest that neither standing vegetation nor slash should be burned. Instead clearing should be done by selective mechanical treatments in middle-aged or mature shrubland with slash remaining on the soil surface (Baeza and Vallejo, 2008). Our results also suggest that such vegetation clearing should preferably be done at the end of summer or beginning of autumn. When clearing was applied at the end of spring, germination took place that same autumn coinciding with the return of rains. When applied at the end of summer, seedling recruitment was minor the following autumn and spring despite abundant rain (144 mm for the months of October, November and December and 206 mm for the months of March, April and May). The largest seedling recruitment was delayed until the following autumn, which experienced only a marginally higher rainfall rate than in the
692
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693
previous spring (265 mm for the months of October, November and December). This result originates from the strong impact of the large temperature fluctuations in summer which break seed dormancy and allow abundant germination in the following autumn rainy period. This remains true even when slash remains on the soil (Fig. 2) probably due to an incomplete buffering effect of slash or the heterogeneity of its distribution. Such an impact of the season at which disturbance occurs has been shown to affect seedling recruitment in the case of other disturbances (Bond and van Wilgen, 1996). In the case of fire, for example, season determines its intensity (greater during the dry period) (Hodgkinson, 1991; Moreno and Oechel, 1991) and the post-fire rainfall pattern (Whelan and Tait, 1995). It has been suggested that the present situation of high wildfire incidence is a consequence of extreme weather conditions (Moritz, 2003), and climate predictions point to increases in annual and summer temperatures in the Eastern Iberian Peninsula (Pausas, 2004), conditions that will enhance wildfire. Mediterranean gorse shrublands accumulate large amounts of dead biomass and they are dominated by seeder species with very few individuals of the more resilient late-successional sprouter species. Both of these factors make these communities particularly vulnerable to degradation processes, especially in areas of Southern Europe with a strong water deficit that limits ecosystems resilience (Puigdefa´bregas and Mendizabal, 1998; Vallejo et al., 2000, 2005). It has been argued that all fuel-control techniques should take into account the ecological characteristics and requirements that ensure the stability and sustainability of the ecosystem (Whelan, 1995; Dimitrakopoulos, 1997). Ideally, they should also alter the disturbance regime and/or the succession trajectory to create favourable establishment opportunities for late-successional species and unfavourable conditions for noxious species (Buckley et al., 2004). We suggest that a mechanical clearing of U. parviflorus shrublands which leaves mulch on the soil should be associated with the re-introduction by planting of lower flammability sprouter species to increase system resilience (Valdecantos et al., in press). This would be a sustainable management strategy that integrates both fuel reduction and land restoration. Acknowledgements We acknowledge L. Sonie´ and J. Fabreguettes for their technical collaboration and J. Scheiding for her linguistic revision. We thank Peter Fule´ and V.R. Vallejo for their comments and improvements to the manuscript. This work has been supported by Generalitat Valenciana post-doc contract to M.J. Baeza (Ref. CTESPP/2002/53), the program Consolider-Ingenio 2010 (GRACCIE CSD2007-00067) and Spanish Ministry of Education (FIREMAP CGL2004-06049C04-04). CEAM was supported by Generalitat Valenciana and Bancaixa. References Auld, T.D., 1986. Populations dynamics of the shrub Acacia suaveolens (Sm) Willd.: fire and the transition to seedlings. Aust. J. Ecol. 11, 373–385. Auld, T.D., O’Connell, M.A., 1991. Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Aust. J. Ecol. 16, 53–70. Baeza, M.J., 2001. Aspectos ecolo´gicos y te´cnicas de control del combustible (roza y quema controlada) en matorrales con alto riesgo de incendio dominados por Ulex parviflorus (Pourr.). Ph.D. Thesis, University of Alicante. http://www.cervantesvirtual.com/FichaObra.html?ReF=5920. Baeza, M.J., Vallejo, V.R., 2006. Ecological mechanisms involved in dormancy breakage in Ulex parviflorus seeds. Plant Ecol. 183, 191–205. Baeza, M.J., Vallejo, V.R., 2008. Vegetation recovery alter fuel management in Mediterranean shrublands. Appl. Veg. Sci. 11, 151–158. Baeza, M.J., Ravento´s, J., Escarre´, A., 2002a. Ulex parviflorus germination after experimental burning: effects of temperature and soil depth. In: Trabaud, L., Prodon, R. (Eds.), Fire and Biological Processes. Backhuys Publishers, Leiden, The Netherlands, pp. 83–91.
Baeza, M.J., De Luı´s, M., Ravento´s, J., Escarre´, A., 2002b. Factors influencing fire behaviour in shrublands of different stand ages and the implications for using prescribed burning to reduce wildfire risk. J. Environ. Manage. 65, 199–208. Baeza, M.J., Valdecantos, A., Vallejo, V.R., 2005. Management of Mediterranean Shrubland for Forest Fire Prevention. In: Burk, A.R. (Ed.), New Research on Forest Ecosystems. Nova Science Publishers Inc., New York, pp. 37–60. Baeza, M.J., Ravento´s, J., Escarre´, A., Vallejo, V.R., 2006. Fire risk and vegetation structural dynamics in Mediterranean shrublands. Plant Ecol. 187, 189–201. Baskin, M., Baskin, C., 1989. Physiology of dormancy and germination in relation to seed bank ecology. In: Leck, M.A., Parker, V.T., Simpson, R.L. (Eds.), Ecology of Soil Seed Banks. Academic Press Inc., San Diego, California, pp. 53–65. Baskin, C., Baskin, M., 1998. Seeds, Ecology, Biogeography and Evolution of Dormancy and Germination. Academic Press, London. Bond, W.J., van Wilgen, B.W., 1996. Fire and Plants. Chapman and Hall, London. Buckley, Y.M., Rees, M., Paynter, Q., Lonsdale, M., 2004. Modelling integrated weed management of an invasive shrub in tropical Australia. J. Appl. Ecol. 41, 547–560. Cramer, W., 2001. Fire ecology, Mediterranean forest and global change. Forest Ecol. Manage. 147, 1–2. Davis, M.A., Grime, J.P., Thompson, K., 2000. Fluctuating resources in plant communities: a general theory of invasibility. J. Ecol. 88, 528–534. Dimitrakopoulos, A.P., 1997. Wildland fire hazard reduction through natural resources management. In: Balabanis, P., Eftichidis, G., Fantechi, R. (Eds.), Forest Fire Risk and Management. European Commission, Brussels, pp. 207–214. Freckleton, R.P., 2004. The problems of prediction and scale in applied ecology: the example of fire as a management tool. J. Appl. Ecol. 41, 599–603. Gorski, T., Gorska, K., Rybicki, J., 1978. Studies on the germination of seeds under leaf canopy. Flora 167, 289–299. Greig-Smith, P., 1983. Quantitative Plant Ecology, 3rd ed. Blackwell, Oxford. Hewitt, N., 1998. Seed size and shade-tolerance: a comparative analysis of North American temperate trees. Oecologia 114, 432–440. Hodgkinson, K.C., 1991. Shrub recruitment response to intensity and season of fire in a semi-arid woodland. J. Appl. Ecol. 28, 60–70. Holmes, P.M., 1989. Effects of different clearing treatments on the seed bank dynamics of an invasive Australian shrub, Acacia cyclops, in the southwestern cape, South Africa. Forest Ecol. Manage. 28, 33–46. Huston, M.A., 1994. Biological Diversity. Cambridge University Press, Cambridge, UK. Ivens, G.W., 1978. Some aspects of seed ecology of gorse. In: Proceedings 31st New Zealand Weed and Pest Control Conference. pp. 53–57. Ivens, G.W., 1983. The influence of temperature on germination of gorse (Ulex europaeus L.). Weed Res. 23, 207–216. Izhaki, I., Henig-Sever, N., Ne’eman, G., 2000. Soil seed banks in Mediterranean Aleppo pine forest: the effect of heat, cover and ash on seedling emergence. J. Ecol. 88, 667–675. Keeley, J.E., 1994. Seed-germinations patterns in fire-prone Mediterranean-climate regions. In: Arroyo, M.T.K., Zedler, P.H., Fox, M.D. (Eds.), Ecology and Biogeography of Mediterranean Ecosystem in Chile, California and Australia. SpringerVerlag, New York, pp. 239–273. Keeley, J.E., Bond, W.J., 1997. Convergent seed germination in South African fynbos and Californian chaparral. Plant Ecol. 133, 153–167. Lee, W.G., Allen, R.B., Johnson, P.N., 1986. Succession and dynamics of gorse (Ulex europaeus L.) communities in the Dunedin Ecological District South Island, New Zealand. N Z J. Bot. 24, 279–292. Lonsdale, W.M., 1993. Losses from the seed bank of Mimosa pigra: soil microorganisms vs. temperature fluctuations. J. Appl. Ecol. 30, 654–660. Lloret, F., Vila`, M., 1996. Clearing of vegetation in Mediterranean garrigue: response after a wildfire. Forest Ecol. Manage. 93, 227–234. Me´thy, M., Fabreguettes, J., Jardon, F., Roy, J., 1987. Design of a simple instrument for the measurement of red/far-red ratios. Acta Oecol. 8, 281–290. Moreno, J.M., Oechel, W.C., 1991. Fire intensity effects on the germination of shrub and herbaceous species in southern California chaparral. Ecology 72, 1993– 2004. Morin, X., Chuine, I., 2006. Niche breadth, competitive strength and range size of tree species: a trade-off based framework to understand species distribution. Ecol. Lett. 9, 185–195. Moritz, M.A., 2003. Spatio temporal analysis of control on shrubland fire regimes: age dependency and fire hazard. Ecology 84, 351–361. Ooi, M.K., Auld, T.D., Whelan, R.J., 2006. Dormancy and the fire-centric focus: germination of three Leucopogon species (Ericaceae) from South-eastern Australia. Ann. Bot. 98, 421–430. Pausas, J., 2001. Resprouting vs seeding—a Mediterranean perspective. Oikos 94, 193–194. Pausas, J., 2004. Changes in fire and climate in the Eastern Iberian Peninsula (Mediterranean Basin). Clim. Change 63, 337–350. Paynter, Q., Flanagan, G.J., 2004. Integrating herbicide and mechanical control treatments with fire and biological control to manage an invasive wetland shrub Mimosa picra. J. Appl. Ecol. 41, 615–629. Pickett, S.T., White, P.S., 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York. ˜ ol, J., Terradas, J., Lloret, F., 1998. Climate warming, wildfire hazard and wildfire Pin occurrence in coastal eastern Spain. Clim. Change 38, 345–357. Pons, J., Pausas, J.G., 2007. Acorn dispersal estimated by radio-tracking. Oecologia 153, 903–911. Puigdefa´bregas, J., Mendizabal, T., 1998. Perpectives on desertification: western Mediterranean. J. Arid Environ. 39, 209–224.
M.J. Baeza, J. Roy / Forest Ecology and Management 256 (2008) 685–693 Rees, M., 1997. Seed dormancy. In: Crawley, M. (Ed.), Plant Ecology. Blackwell Science Ltd., London, pp. 214–238. Rees, M., Hill, R.L., 2001. Large-scale disturbances, biological control and the dynamics of gorse populations. J. Appl. Ecol. 38, 364–377. Roy, J., Arianoutsou-Faraggitaki, M., 1985. Light quality as the environmental trigger for the germination of the fire-promoted species Sarcopoterium spinosum L. Flora 177, 345–349. Roy, J., Sonie´, L., 1992. Germination and populations dynamics of Cistus species in relation to fire. J. Appl. Ecol. 29, 647–655. Scott, S.J., Jones, R.A., Williams, W.A., 1984. Review of data analysis methods for seed germination. Crop Sci. 24, 1192–1199. Sheppard, A.W., Hodge, P., Paynter, Q., Rees, M., 2002. Factors affecting invasion and persistence of broom Cytisus scoparius in Australia. J. Appl. Ecol. 39, 721–734. Teasdale, J.R., Daughtry, C.S.T., 1993. Weed suppression by live and dessicated hairy vetch (Vicia villosa). Weed Sci. 41, 207–212. Teasdale, J.R., Mohler, C.L., 1993. Light transmittance, soil temperature and soil moisture under residue of hairy vetch and rye. Agron. J. 85, 673–680. Teasdale, J.R., Mohler, C.L., 2000. The quantitative relationship between weed emergence and the physical properties of mulches. Weed Sci. 48, 385–392. Thanos, C.A., Georghiou, K., Kadis, C., Pantazi, C., 1992. Cistaceae: a plant family with hard seeds. Isr. J. Bot. 41, 251–263. Thompson, K., Ceriani, R.M., Bakker, J.P., Bekker, R.M., 2003. Are seed dormancy and persistence in the soil related? Seed Sci. Res. 13, 97–100.
693
Tutin, T.G., Heywood, V.H., Burges, N.A., Moore, D.M., Valentine, D.H., Walters, S.M., Webbs, D.A. (Eds.), 1964–1980. Flora Europaea, vol. 5. Cambridge Univ. Press, Cambridge. Valdecantos, A., Baeza, M.J., Vallejo, V.R. Vegetation management for promoting ecosystem resilience in fire-prone Mediterranean shrublands. Rest. Ecol., in press. Vallejo, V.R., Bautista, S., Cortina, J., 2000. Restoration for soil protection after disturbances. In: Trabaud, L. (Ed.), Life and Environment in the Mediterranean. Wit Press, Southampton, Boston, pp. 301–343. Vallejo, V.R., Aronson, J., Pausas, J.G., Cortina, J., 2005. Restoration of Mediterranean woodlands. In: Van Andel, J., Aronson, J. (Eds.), Restoration Ecology from European Perspective. Blackwell Science, pp. 193–207. Whelan, R.J., 1995. The Ecology of Fire. Cambridge University Press, New York. Whelan, R.J., Tait, I., 1995. Responses of plant populations to fire: fire season as an under-studied element of fire regime. CALM Sci. Suppl. 4, 147–150. Zabkiewicz, J.A., Gaskin, R.E., 1978. Effect of fire on gorse seeds. In: Proceedings of 31st New Zealand Weed and Pest Control Conference. pp. 47–52. Zammit, C.A., Westoby, M., 1988. Pre-dispersal seed losses, and survival of seeds and seedlings of two serotinous Banksia shrubs in burnt and unburnt heath. J. Ecol. 76, 200–214. Zedler, P.H., Zammit, C.A., 1989. A populations-based critique on concepts of change in the chaparral. Sciences series N8 34. In: Keeley, S.C. (Ed.), Chaparral Ecology: Paradigms Re-examined. Natural History Museum of Los Angeles Country, Allen Press, Los Angeles, California.