Oecologia (2003) 136:193–204 DOI 10.1007/s00442-003-1259-7
ECOPHYSIOLOGY
Omar R. Lopez · Thomas A. Kursar
Does flood tolerance explain tree species distribution in tropical seasonally flooded habitats? Received: 19 November 2001 / Accepted: 17 March 2003 / Published online: 13 May 2003 Springer-Verlag 2003
Abstract In the tropics, seasonally flooded forests (SFF) harbor fewer tree species than terra firme (i.e. nonflooded) forests. The low species diversity of tropical flooded forests has been ascribed to the paucity of species with adaptations to tolerate flooding. To test the hypothesis that flooding is the only factor restricting most species from SFF, we compared plant morphological and physiological responses to flooding in 2-month old seedlings of 6 species common to SFF and 12 species common to terra firme forests. Although flooding impaired growth, total biomass, maximum root length and stomatal conductance in most species, responses varied greatly and were species-specific. For example, after 90 days, flooding reduced leaf area growth by 10– 50% in all species, except in Tabebuia, a common species from non-flooded habitats. Similarly, flooding had a 5– 45% negative effect on total biomass for all species, except in 1 SFF and 1 terra firme species both of which had more biomass under flooding. A principal component analysis, using the above responses to flooding, provided no evidence that SFF and terra firme species differed in their responses to flooding. Flooding also caused reductions in root growth for most species. Rooting depth and root: shoot ratios were significantly less affected by flooding in SFF than in terra firme species. Although flood tolerance is critical for survival in flooded habitats, we hypothesize that responses to post-flooding events such as drought might be equally important in seasonal O. R. Lopez ()) · T. A. Kursar Department of Biology, University of Utah, 257 S., 1400 E., Salt Lake City, UT 84112-0840, USA e-mail:
[email protected] Tel.: +1-828-4880178 T. A. Kursar Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Ancn, Panama Present address: O. R. Lopez, Department of Botany, University of Wisconsin-Madison, 430 Lincoln Dr., Madison, WI 53706-1381, USA
habitats. Therefore, we suggest that the ability to grow roots under anoxia might be critical in predicting success in inundated habitats that also experience a strong dry season. Keywords Panama · Root growth · Seasonality · Stomatal conductance · Swamps
Introduction In the tropics, low diversity forests are commonly associated with seasonally waterlogged conditions (Davis and Richards 1933; Janzen 1974; Peters et al. 1989; Hart 1990; Richards 1996). Shortly after flooding, oxygen depletion at the rhizosphere could result in reduced stomatal conductance, lower photosynthesis, plant hormonal imbalance, poor water and nutrient uptake, ultimately compromising plant growth and survival (Pezeshki 1994). Thus the reduced species richness of tropical swamps has been ascribed to the inability of terra firme species (species from non-flooded forests) to cope with the stressful conditions resulting from flooding (Black et al. 1950; Campbell et al. 1992; Duivenvoorden 1996). Contrary to this assertion, recent studies suggest that flood tolerance of tropical trees might be more common than previously thought, although the extent to which tropical trees are flood tolerant has not been thoroughly investigated. For example, Genipa americana L. (Rubiaceae), a widely distributed Neotropical species found in flood-prone habitats in Brazil, but most commonly occurring in terra firme elsewhere, is a flood-tolerant species (Andrade et al. 1999). Likewise, Hymenea courbaril L. var. stilocarpa (Caesalpinioideae), a dry forest species, and Corisia speciosa St. Hill (Bombacaceae), a terra firme species, were found to be highly flood tolerant (Joly and Crawford 1982). Thus, flood tolerance might be common among species that experience little or no flooding in their natural habitat.
194
To test the hypothesis that flood tolerance regulates species diversity in seasonally flooded forests (SFF), we evaluated flood tolerance in tropical tree species. For this purpose we investigated growth, physiological and morphological responses to flooding in 2-month-old seedlings of 18 Neotropical tree species. We studied 6 species from SFF, and 12 terra firme species that are common in lowland tropical forests, but that very rarely experience flooding (see Table 1). If the hypothesis of flood tolerance is true, we would expect terra firme species to be significantly less tolerant to flooding than SFF species. Furthermore, light demanding and pioneer species appear uncommon in some SFF (Parolin 2000; Lopez, personal observation). Thus, to see whether light-demanding species are more susceptible to flooding than shadetolerant species, we included 6 light-demanding species, 2 from SFF and 4 from terra firme forests.
Materials and methods Study site and species Flooding experiments were carried out in a screened house from December 1998 to September 2000 on Barro Colorado Island (BCI), a research facility administered by the Smithsonian Tropical Research Institute in central Panama (990 N, 79510 W). A complete description of the site and its vegetation can be found in Leigh et al. (1996) and Croat (1978), respectively. Mean annual rainfall on BCI is 2,600 mm with a pronounced dry season from mid-December to the beginning of May (Windsor 1990). A total of 18 species were studied (Table 1). Among the SFF species Prioria is unique as it forms extensive mono-dominant stands in SFF of Central America
Table 1 Species name, family, growth form, light environment and habitat. Light environment follows Welden et al. 1999, R. Condit personal communication and personal observation (CT canopy tree,
from Nicaragua to Colombia (Mayo-Melndez 1965; Linares and Martnez 1991). With the exception of Pachira, which is mainly found in riverine habitats, the other five SFF species also occur in non-flooded forests. Although terra firme species are most common in lowland non-flooded tropical forests, Anacardium, Tabebuia, Dipteryx and Virola also occur at very low densities in riparian or SFF (Linares and Martnez 1991; Grauel and Kursar 1999). The terra firme and SFF species include both shade-tolerant and lightdemanding species (Table 1). Experimental design Seeds of terra firme species were collected on BCI where trees never experience flooding, while seeds of SFF species were collected at Rio Balsas, Darin (8070 N, 77520 W) and brought back to BCI within 48 h. Seeds were germinated in 0.5-l plastic pots containing sandy loam soil (Utah State University Soils Laboratories, Logan, Utah). Seedlings were grown under shaded conditions (0.8 mol photons m2 day1) and transplanted 3 weeks prior to the experiment into plastic pots (0.5 m tall 0.2 m diameter), one plant per pot. Seedlings were about 10 weeks old and of similar size within each species (e.g. leaf area and seedling height) when assigned to the treatments. Flooded and non-flooded treatments comprised a total of nine or ten plants each, for all species, and were maintained under shaded conditions. Flooding was maintained in water-filled containers for 90 days providing a water level at least 3 cm above the soil surface. Non-flooded plants (=control) were well watered but never flooded. Measurements Soil redox potentials were subsampled for all species at a depth of 0.3 m after 1, 2, 3, 5, 6, 13, 21, 30, 45, 80, and 90 days in flooded and non-flooded pots. Soil redox potentials were taken using a Platinum (Pt) redox electrode, coupled with a silver (Ag/AgCl) electrode (Jensen Instruments, Tacoma, Wash.). Redox potentials
Sub-CT sub-canopy tree,SFFseasonally flooded forests,TF terra firme forests)
Family
Growth form
Light environment
Habitat
Reference
Meliaceae Bombacaceae Fabaceae Fabaceae
CT River edges CT CT
Shade-tolerant Light-demanding Shade-tolerant Shade-tolerant
SFF and TF SFF SFF and TF SFF and TF
Pterocarpus officinalis Pterocarpus sp. Terra firme species Anacardium excelsum
Fabaceae Fabaceae
CT
Light-demanding Unknown
SFF SFF
Jimnez 1994 Jimnez 1994 Hartshorn 1972 Condit et al. 1993, Grauel and Kursar 1999 Grauel and Kursar 1999 Personal observation
Anacardiaceae
CT
Light-demanding
TF and SFF
Aspidosperma cruenta Brosimum alicastrum Calophyllum longifolium Croton billbergianus Dipteryx panamensis
Apocynaceae Moraceae Clusiaceae Euphorbiaceae Fabaceae
CT CT CT Sub-CT CT
Shade-tolerant Shade-tolerant Shade-tolerant Light-demanding Light-demanding
TF TF TF TF TF and SFF
Guazuma ulmifolia Gustavia superba
Tiliaceae Lecythidaceae
CT Sub-CT
Light-demanding Shade-tolerant
TF TF
Tabebuia rosea
Bignoniaceae
CT
Shade-tolerant
TF and SFF
Tachigali versicolor Tetragastris panamensis Virola surinamensis
Fabaceae Burseraceae Myristicaceae
CT CT CT
Shade-tolerant Shade-tolerant Shade-tolerant
TF TF TF and SFF
SFF species Carapa guianensis Pachira aquatica Pentaclethra macroloba Prioria copaifera
Grauel and Kursar 1999, Condit et al. 1996 Condit et al. 1996 Condit et al. 1996 Condit et al. 1996 Personal observation Grauel and Kursar 1999, Condit et al. 1996 Personal observation Condit et al. 1996, Sork 1985 Grauel and Kursar 1999, Condit et al. 1996 Condit et al. 1996 Condit et al. 1996 Berry et al. 1995
195 were corrected (converted to Eh) by adding +242 mV to the potential given by the platinum electrode. Redox potentials were not adjusted for pH changes due to negligible pH differences between the two treatments (average pH =6.7). Leaf number and leaf area were measured every 15 days from the beginning of the experiment using a portable leaf area meter (LI-3000, Li-Cor, Lincoln, Neb.). At the end of the experiment (90 days), plants were harvested and seedling height, maximum root depth and stem diameter at 0.05 m above the soil were measured. Additionally, morphological responses to flooding were noted, specifically the formation of hypertrophied lenticels and adventitious roots. After 90 days, plants were harvested, divided into above- and belowground parts and dried at 70C for 72 h for biomass. Stomatal conductance (gst) was measured after 1, 2, 5, 12, 15 and 30 days of treatment with a steady-state porometer (LI-1600, Li-Cor, Lincoln, Neb.) on leaves that had matured prior to the treatments. Measurements were conducted in a growth chamber (Model M-Sun, Environmental Growth Chambers, Chagrin Falls, Ohio) between 0900 and 1300 hours at saturating light (500– 600 mmol m—2 s1, photosynthetic photon flux density, 25–30C and 70–75% relative humidity). gst was measured on a minimum of six plants per treatment and at least two leaves per plant. Plants were allowed to acclimate for 8–10 min prior to each measurement and measured in pairs, one control and one flooded plant per species. Data analysis The effect of flooding on soil redox potential was tested using an analysis of variance (ANOVA). However, the comparison at 90 days was made by Student’s t-test. Flooding effects on leaf area growth and gst in each species were analyzed by repeated measures ANOVA. After excluding several highly correlated parameters and normalizing the data when needed (log x), the effects of flooding on each species were analyzed by multivariate analysis-of-variance (MANOVA, F-test) using five parameters: height relative growth rate {RGRheight=[ln (final height cm)ln (initial height)]/3 months}, leaf mass per area, total biomass, root: shoot ratios and maximum root depth. Following this, each parameter was tested independently and the alpha level was adjusted by taking into consideration the number of tests performed (k=5). Therefore the effect of flooding on each parameter was tested at the P=0.01 level (Bonferroni correction; Sokal and Rohlf 1995; Cobin and Mitchel 2000). We used ANOVA to test for treatment (flooding, non-flooding) and habitat (SFF, terra firme) main effects and their interaction on seedling performance. Species effects were nested within habitat and treated as random effects. Principal component analysis (PCA) was used to determine species grouping patterns with respect to growth, morphological and physiological responses to flooding (Fernandez et al. 1993; Golluscio and Sala 1993). Data analysis was conducted with the statistical software JUMP v 3.2.1. (SAS Institute, Cary, N.C.).
Fig. 1 Soil redox potentials (Eh) for control (black circle) and flooded (clear circle) pots versus time. Each point represents the average Eh at a depth of 0.3 m across all species tested for each date (€SE). The dashed line represents the Eh at which soils become anaerobic (~300 mV)
Stomatal response gst over the first 20 days of flooding was highly variable among species. After 30 days, the stomatal responses of flooded seedlings can be broadly summarized by two distinctive patterns. Twelve species were “sensitive” to flooding showing a 21–62% reduction in gst due to the flooding treatment (Fig. 2A). In contrast, flooded seedlings of 6 other species appeared “insensitive”, having gst values that were undistinguishable from those of control seedlings after 30 days (Fig. 2B). Flooding significantly reduced gst in 5 out of the 12 terra firme species and 3 out of the 6 SFF species (Wilks’ Lambda repeated measures ANOVA, see Fig. 2A, B). In addition, the flooding by time interaction was only significant for 3 terra firme (Dipteryx, Anacardium and Virola) and 1 SFF (Pterocarpus officinalis) species (data not shown).
Results Aboveground growth responses Soil redox potential Flooding significantly decreased soil redox potentials (Eh) of potted plants (P
