- Top of page
Sustaining environmental flows for riparian ecosystems requires knowledge of linkages between hydrological and biological processes (Arthington et al. 2006). One of the many components of a flow regime is the timing of floods. Flood timing ranges widely among rivers, from predictable spring snowmelt run-off to more stochastic patterns in rivers whose flows derive from seasonally shifting storm tracks. Flows are also widely altered by human actions. Although elimination of floods on dam- and flow-regulated rivers is known to induce many changes in riparian plant community structure, the effects of changes in flood timing are not as well understood (Beauchamp & Stromberg 2007; Poff & Zimmerman 2010; Greet et al. 2011).
One way in which floods influence vegetation is through disproportionately dispersing propagules to safe sites, e.g. directed dispersal (Grubb 1977). As habitat patches are created and destroyed in the riparian landscape mosaic, various mechanisms disperse plants to new areas. Hydrochory is one such mechanism, particularly for aquatic, wetland and near-channel plants (Drezner et al. 2001; Boedeltje et al. 2003; Merritt & Wohl 2006). Although many of the riparian plants that grow on floodplains and terraces show adaptations for anemochory or zoochory, secondary dispersal by flood flows can be an important dispersal mechanism as well (Drezner et al. 2001; Pettit & Froend 2001; Moggridge & Gurnell 2010).
Numbers and richness of hydrochores often peak at highest discharge (Boedeltje et al. 2004; Moggridge & Gurnell 2010), but questions remain regarding the degree to which a plant's life cycle is coupled with the typical seasonal pattern of discharge. For a few well-studied species, the timing of seed dispersal from the parent plant is known to be coupled to winter/spring flow peaks; such conditions expose bare wet mineral soil during the short time when seeds are viable, and allow plants to grow to a size that enables over-winter survival (Stella et al. 2006). There may also be species that disperse and colonize during summer floods, although plants that disperse late in the growing season tend to produce dormant seeds (Boedeltje et al. 2004). Many such seeds over-winter in soil and drift seed banks and are then entrained by ensuing floods to colonize new sites (Goodson et al. 2003; Gurnell et al. 2006; Merritt & Wohl 2006).
The connection between phenological events in a plant's life and the seasonal pattern of peak flows is of importance from a management standpoint (Merritt & Wohl 2002). An understanding of such connections can aid managers in determining environmental flows (e.g. the timing, magnitude and duration of flows) to facilitate seed germination and establishment of particular plant taxa. In particular, identifying temporal flood specialists (those where phenology of dispersal and germination are linked with seasonal peak flows) and flood generalists (those capable of responding to flood pulses that clear or moisten soils at any time during the growing season) is important for managing river flows, given that temporal dispersal specialists likely are more at-risk from climate change or from direct human alterations to stream flow patterns.
Focusing on a semi-arid region river with a bimodal (but variable) flood pattern, we asked several questions relating to linkages between reproductive phenology, hydrochory and stream flows: (1) do flowering and primary seed release at the community level exhibit a bimodal pattern; (2) does the diversity and abundance of hydrochores increase during flood events; (3) how do seed release phenology and stream hydrology interact to influence hydrochore patterns; (4) are distinct hydrochore guilds present, and do these encompass vernal specialists and aestival specialists; and (5) are hydrochores a subset of the extant riparian vegetation, composed primarily of the same small-seeded wetland species that form persistent soil seed banks?
- Top of page
In this study, we asked specific questions relating to interrelationships between phenology of primary seed dispersal, hydrochory patterns and stream hydrology of a semi-arid region river with bimodal flow peaks. Our first question related to temporal correlations between primary seed release and seasonal floods. In some riverine ecosystems, many taxa release seed during the seasonal period of flood inundation, providing both for long-distance transport of seeds and for deposition of seeds on safe sites for seedling establishment (Schneider & Sharitz 1988; Kubitzki & Ziburski 1994). On the Verde River, a bimodal pattern of flowering and seed release was apparent only for the dryland subset of riparian species, many of which were annuals. Seasonal moisture pulses provide a critical resource for germination and flowering of this group, with seed release following afterwards, a sequence consistent with patterns in arid and semi-arid settings (Chesson et al. 2004; Bowers 2005). Wetland species, in contrast, had more constant flowering and seed release at the community level (with numbers high from July through October) indicating looser coupling between seasonal wetting events and seed production and dispersal.
The second question related to hydrochory and peak flows. The ability of hydrochory to influence riparian plant communities is dependent on fluctuating flows to allow exchange of seeds between the water and riverbank (Merritt & Wohl 2002; Jansson et al. 2005). As found for many rivers, hydrochory in the Verde River did peak during high flow periods both for number of individuals and species; during such times, increased velocities are able to entrain and transport large numbers and types of propagules (Boedeltje et al. 2004; Vogt et al. 2006; Gurnell et al. 2008). In contrast to some rivers, which show a strong single or bi-seasonal hydrochore peak (Moggridge et al. 2009), three richness/abundance peaks were evident on the Verde River, reflecting the hydrology of our atypical study year. Patterns of hydrochory on the Verde likely vary between years, depending on inter-annual variability in flood timing and magnitude, as found elsewhere (Andersson & Nilsson 2002).
For our third question, we found that dispersal phenology interacted with patterns of discharge to produce varying outcomes in hydrochore richness and composition, as evidenced by compositional contrasts of the three hydrochore richness peaks. The species richness peak in September reflected the coincidence of late summer seed ripening with a late summer flood. During the other two periods of high flow, many of the hydrochores that were captured were not undergoing primary dispersal. The May floodwaters contained many wetland species that were not actively dispersing, so the increased richness was probably due to re-suspension of past seasons' diaspores. This pattern for dispersal to peak during the spring run-off season owing to remobilization of previous seasons' propagules is common in rivers throughout the world (Gurnell et al. 2006; Merritt & Wohl 2006). The flood event in late July captured a different set of species. This flood originated from rainfall in the headwaters (as indicated by the lack of flow increases for the lower tributaries of the Verde River during this time period). Dryland species were abundant in this flood, which we attribute to headwater input upstream of the study area, or to heavy rains that cause lateral overland flow (Vittoz & Engler 2007; Moggridge & Gurnell 2010). Convective rain storms may serve to hydrochorically connect otherwise disconnected rivers and to facilitate downstream accumulation of seeds and species from tributaries, enhancing the potential for higher river margin plant species richness as a function of downstream distance (‘river accumulator hypothesis’ sensu Nilsson et al. 1994).
These results differ in two ways from the trimodal hydrochore pattern recognized along streams in the Rocky Mountains, in which there is a spring flush of propagules re-entrained from previous years occurring with the first flood pulse, early dispersers occurring in synchrony with the descending limb of the snowmelt hydrograph, and the third occurring during summer when most plants are setting seed (Merritt & Wohl 2006). First, we did not find a clear temporal separation between re-entrained snowmelt hydrochores and vernal specialists with transient seeds that disperse during the spring snowmelt, although our study design did not allow for capture of the two major peaks of the winter/spring snowmelt (which occurred in February and March). Second, hydrochores of the summer dispersers in the Verde River were most abundant during summer floods (likely owing to remobilization of seeds as the floodwaters spread across the channel margins) rather than during low flow periods (Merritt & Wohl 2006). Hydrochore studies spanning a large range of hydroclimatic river types are needed to tease apart the processes driving these divergent patterns.
We also asked whether distinct phenological hydrochore dispersal guilds are apparent on the Verde River. We recognized (1) pulse hydrochore taxa, three of which were spring flood specialists and one of which appears to be a late summer (aestival) flood specialist; (2) constant and semi-constant hydrochore species, which can be considered as opportunists or generalists in that their seeds were in the stream throughout the growing season, during and after their primary dispersal period; and (3) flood incidentalists. One of the vernal hydrochore species was Carex senta, a perennial wetland graminoid that is restricted to streams in Arizona and California and that is in need of study with respect to germination and seedling ecology. For two other vernal specialists, Salix gooddingii and P. fremontii, primary seed release, hydrochory and germination are all tightly coupled with the flow regime. These members of the Salicaceae have short-lived seeds that decrease to 50% viability in less than 2 mo (Stella et al. 2006). Seeds can travel to the stream via barochory (if overhanging the water) or by anemorchory. The short-lived seeds and restricted dispersal windows of Salix sp. and P. fremontii restricts their presence as viable hydrochores to a narrow time window in spring, which corresponds with the timing of winter flood drawdown in the study river and other regional rivers. Both of these species have a limited range, being restricted to the American Southwest.
Despite the (long-term) bimodal flood regime, we found only one taxon that had summer-restricted hydrochory. Summer flood specialists may be more prominent on rivers with more pronounced summer monsoon floods, such as those in the Chihuahuan Desert. Designation of a species as a seasonal hydrochore specialist is complicated, as the timing of seed maturation, longevity of seeds and dormancy all influence hydrochory patterns (Boedeltje et al. 2004) and can vary within the range of a species. Polygonum lapathifolium, P. punctatum and Polygonum sp., the summer-restricted pulse taxon in our study, began to disperse seeds in June but was recorded from water samples only in September. Polygonum spp. have multiple types of seed dormancy (including innate) and it is possible that seeds were present in the surface stream in June but detected by us only as they lost dormancy (i.e. after-ripened; Staniforth & Cavers 1979). Seeds of P. lapathifolium can persist for several years but often have high overwinter mortality from desiccation and submergence (Staniforth & Cavers 1976). Apparently, there was little overwinter seed survival of this species during our study, making the seeds functionally transient.
Over half of the common hydrochores – including Typha domingensis, Cyperus odoratus and Polypogon monspeliensis – were generalists or opportunists in the sense that their hydrochores were essentially continuously present throughout the year and less tightly coupled with flow events of particular timing. All of the generalists have seeds that are persistent in soil seed banks, and all were found in the near-stream soil and litter seed bank. We suspect that these species have long-lived seeds that are stored in the stream bed and banks and re-suspended and deposited throughout the year, allowing for colonization after disturbance that occurs at various times throughout the year. Such a strategy provides resilience to temporal fluctuations in the timing of flood events (Merritt & Wohl 2006; Stromberg et al. 2008). This phenological dispersal plasticity also may contribute to the wide geographic ranges of these species (Gordon 2002; Neff & Baldwin 2005; Ruaux et al. 2009; Osland et al. 2011) and their capacity to adapt or shift ranges in response to climate change (Perry et al. 2012).
The final question examined trait similarity between hydrochores, extant riparian vegetation and riparian seed banks. Although the community of hydrochores was biased toward small-seeded wetland species, it did encompass a wide range of species with respect to wetland class, seed mass and growth habit. Of note, there was one group of large-seeded mesic and dryland species for which the primary functional role of floods appears to be periodic broad dispersal. Such widespread dispersal can significantly increase local species richness in a wide range of communities (Moggridge et al. 2009; Myers & Harms 2009) and, by expanding their range, may enable species to respond to changing environmental conditions.