The saltmarsh vegetation of the lower Berg River

The lower Berg River supports approximately 250 ha o f estuarine saltmarsh vegetation. Species distribution patterns, as sampled along six transects, are described. Elevation above mean sea level (MSL) is proposed as a strong determinant o f these patterns. However, there are no typical patterns. The patchy and irregular distribution patterns possibly result from an inconsistent relationship between species distribution and salinity, tidal inundation and/or competitive interactions.


INTRODUCTION METHODS
Much anthropogenic development has taken place around the lower Berg River.The effects that these developments have had on the marshes vary from total destruction (e.g. the development of housing and saltworks), through partial destruction (e.g. the dumping of dredge spoil) to subtle effects brought about by altered tidal and flow patterns with the construction of dams in the catchment and the manipulation of the river mouth to allow access to fishing boats.Nevertheless, Anderson (1991) estimates that the marshes have decreased by only 13% from 1938 to 1986.Unfortunately, he does not dis tinguish between saltmarshes and reed beds.Saltmarshes currently cover approximately 250 ha around this estuary which constitutes just less than 2% of the saltmarshes of southern Africa.
The Berg River is one of the largest rivers on the western Cape coast (Day 1981) and numerous aspects of its ecology have been investigated (Harrison 1958a(Harrison & b. 1974;;Harrison & Elsworth 1958;Scott 1958;Ratte 1976Ratte , 1977;;Coetzer 1976Coetzer , 1978;;Summers et al. 1976;Gaigher 1979).However, apart from superficial descriptions by Giliomee (1973).Day (1981) and Van Wyk (1983), very little information concerning the saltmarsh vegetation is available.These saltmarshes are increasingly being threatened by the expansion and further development of recreational, housing and water storage (McDowell 1992(McDowell . 1993)).This investigation aims to describe the floristic structure of the saltmarshes around the lower Berg River estuary.
After studying aerial photographs, orthophotographic maps and following field reconnaissance, six transects were demarcated across the marshes of the lower Berg River (Figure 1).The siting of these transects was deter mined subjectively according to variability in species composition and the relatively undisturbed nature of the vegetation.Details of these transects are presented in Table 1.Elevation profiles of the transects were surveyed using a theodolite, and at least one point on each transect was surveyed to sea level.
Sampling took place on four occasions during 1987 (March.June.September and November) in order to in clude all bulbous and annual plants.Contiguous 1 x 1 m plots were laid along each transect.The cover-abundance of each species within the plots was estimated according to normal phytosociological methods (Braun-Blanquet 1965).Excessive repetition was avoided by not sampling plots in which it was deemed that the floristic data were simply repetitions of data already recorded from adjacent plots.Taxon names follow Arnold & De Wet (1993) and  2. To plot the distribution of species, each transect was divided into elevation classes of 10 cm.The converted factors were averaged within each 10 cm class and further averaged over the four sampling periods.As some of the species have annual geophytic or hemicryptophytic life-cycles, the number next to the species name on Figures 2 to 7 indicates the number of times this species was located through the year.The order in which the species occur along the transect is primarily deter mined by its lowest starting point and secondarily by its termination point along the elevation gradient.

RESULTS A N D DISCUSSION
The distribution of species along elevation gradients on Transects B 1 to B6 is shown in Figures 2-7.

Transects Bl and B6 (Figures 2 & 7)
Aerial photographs reveal great variability in the dis tribution of saltmarshes in the Blind Lagoon area (Tran sects B1 & B6).In 1942 (Job No. 168;Photo No. 38133), the marshes in this area consisted of disjunct patches with some submerged macrophytes (probably Zostera capensis).By 1960 (Job No. 137;Photo No. 5110 & 5111).the marshes were greatly reduced and there is no evidence of  species on the top of the transect indicate a moist dune environment which abuts onto the marsh.
The rate of elevational increase at Transect Bl was greater than at B6 (5.15 cm/m for Bl; 3.62 cm/m for B6).The marshes at Bl are also younger, dating from 1966 at the earliest.The distinction between the marsh vegetation and adjacent terrestrial vegetation was not as clear as at B6.
At Bl, Zostera capensis was found to ± 53 cm below MSL, especially during summer.From -50 cm to 63 cm, no angiospermous vegetation was present.These mud flats were rich in microflora and filamentous green algae, espe cially during early summer.

Transects B2 and B3 (Figures 3 & 4)
A saltmarsh cliff limited the development of the lower marsh at both these transects.The area immediately above this cliff was slightly elevated compared with the rest of the marsh.This would impede the drainage and a number of creeks of varying size traversed the marsh.With the exception of these creeks, the marsh consisted of flat ex panses with little variation in elevation.Under these con ditions, the distribution of the species tended to form patches and seemed to be related to soil drainage charac teristics, rather than to changes in elevation alone.
At B3, the elevational distributions of most species overlapped somewhat.Zostera capensis was found in a large creek which crossed the marsh.Juncus kraussii and Spartina maritima were found near a smaller creek where the soils were almost constantly waterlogged.The saltmarsh at B3 can be roughly divided into four sections: 1, the deeper creeks containing Zostera capensis; 2, the zone from Sarcocomia perennis to Limonium depauperation constituted 90% of the remaining spatial distribution of the marsh.These areas were flooded rela tively frequently, usually from water entering and leaving via creeks; 3, the June us/Spa rtina zone was found near a smaller creek where the soils were almost constantly waterlogged; 4, the remainder of the marsh constituted less than 10% of the spatial area of this transect.It was characterized by Salicornia meyeriana and, except for Sarcocomia pillan sii (x S. perennis), contained only annual species.These upper areas are seldom flooded, although the soils were rather saline.The top of the transect had been disturbed by residential developments and was largely devoid of vegetation except after rains, during spring.
The physiographic structure of the marsh at B2 was similar to B3.Again, the lower part of the transect (from Schoenoplectus triqueter to Limonium depauperatum) constituted more than 90% of the spatial distribution of this transect and elevational distribution of most species overlapped.Numerous creeks of varying size traversed this transect, but none were developed to a depth able to support Zostera capensis.The soils of B2 seemed to be somewhat wetter than at B3, but the flooding waters were less saline.This decreased salinity was confirmed by the presence of S. triqueter, Trig loch in striata, Scirpus venustulus and Juncus kraussii.
The upper part of B2 had been disturbed by the development of the Salt Works evaporation pans.The soils were highly saline in summer and vegetation was sparse (Sarcocomia pillansii x S. perennis and Salicornia meyeriana) or absent.
Transect B4 (Figure 5) The effects of fresher water inputs were more evident at this transect.Schoenoplectus triqueter a n d Juncus kraussii were found in a large creek which crossed the marsh.Zostera capensis survived at the riverine end of this transect throughout the year, but disappeared from the large creek during spring when fresher waters after the rainy season would limit its growth.Other species which were supported by the lowered salinities include Crassula natans, Samolus porosus and Juncus scabriusculus.
Juncus kraussii covers more than 90% of the spatial area of B4 and Schoenoplectus triqueter covers more than 70%, often forming dense monospecific stands.The dis tribution of some species usually found in more saline marshes (as at B6) did not overlap with the distributions of Juncus kraussii and S. triqueter.However.Chenolea diffusa and Sarcocomia pillansii occurred in association with the sedges as well as with more salt-tolerant species.It seems that these taller sedges have a selective competi tive effect on the distribution of other species.
The distribution patterns of the more salt-tolerant species were somewhat irregular.Species co-occurring with the Jailer sedges were lanky and required the sedges for mechanical support.
Although elevation decreased towards the top of this transect, relatively high water would be required to over top the surrounding higher lying areas.This upper end would therefore be flooded only occasionally bv relatively fresh water during the rainy season.A number of grasses and herbs were found here (Polypogon monspeliensis, Sporobolus virginicus, Disphyma crassifolium, Plantago crassifolia), reminiscent of the upper extremes of a marsh with a more regular elevation gradient.
Transect 115 (Figure 6) This transect starts at a creek which traversed the marsh south of Velddrif.The banks of this creek were lined with Phragmites australis.The bottom of this creek, which is not included in Figure 6. was constantly filled with fresher water in which Potamogeton pectinatus was found.The P. pectinatus indicated in Figure 6 occurred in a depres sion which is seasonally flooded with rain water.
Most of this transect was dominated by Sarcocomia pillansii.As the area immediately adjacent to the creek was slightly elevated, riverine water did not often fWx>d this transect.Furthermore, water which might have entered the marsh through rain, surface drainage or storm surges, would not easily drain away.This resulted in rela tively saline soils, due to evaporation rather than fUxxling by saline water.In areas where the soils were slightly less saline (e.g.immediately adjacent to the creek).Juncus kraussii co-dominated with Sarcocomia pillansii.The remaining species (including most of J. kraussii) were concentrically arranged around a depression which was seasonally filled with rain water.The majority of these species germinated and grew during spring and died back by midsummer.
Phragmites australis was again found on the top of the transect where it was supported by fresh water runoff from adjacent hard surfaces.

CONCLUSIONS
There was no 'typical' saltmarsh at the Berg River.The marshes in the Blind Lagoon showed the clearest zonation patterns, whereas the zonation patterns higher up the es tuary were not as easily discernible.
The conditions prevailing in the Blind Lagoon were somewhat different from those in the other parts of the estuary.There was no unidirectional riverine flow in this part.Rather, the vegetation was subject only to tidal in undation originating from the mouth of the estuary.These conditions are similar to those found in a marine lagoon, and the zonation of the vegetation is similar to that of Langebaan Lagoon (O'Callaghan 1994b).
Unlike Langebaan Lagoon, however, the distribution patterns of species along the elevation gradients in the remainder of the Berg River Estuary were not constant, but somewhat patchy.These inconsistencies could be at tributed to a number of factors: 1, the rate of elevational increase was very low as one moved away from the river channel.The elevations of B 1 and B6 increased by an average of 4.35%.In contrast the elevations of B2 and B3 increased by an average of 0.11 %.With this relatively flat topography, the flooding and drainage at a particular point on the marsh was not neces sarily directly related to the elevation at that point, but rather to the elevation relative to the surrounding marsh.Water could remain dammed in a local depression, result ing in a longer period of inundation.A local elevation would have the opposite result; 2, numerous creeks of varying sizes traversed these mar shes, adding to the complexity of inundation pattern ob served at any particular point.The water which flooded the marsh often did not originate from the main river channel.Usually, the part of the transect adjacent to the river channel was slightly elevated and flooding occurred via the creeks or ground water seepage.The inundation patterns across the marsh were thus rather complex: 3, the fresh water input into the system increased as one moved up the river.The effects of salinity on the saltmarsh plants were complex as they varied tidally as well as seasonally.
These complexities would result in an inconsistent relationship between salinity and tidal inundation, affect ing the relative distribution and competitive abilities of the various species.These interactions require further in vestigation.
The species composition of the marshes did vary ac cording to the period of inundation, remembering that the period of inundation is not necessarily related to elevation alone.Precise data concerning salinity and inundation tolerances for each species are not available.In general, as one moved away from the mouth, sedges and reeds started to dominate in less saline conditions and Zostera capensis was replaced by Potamogeton pectincitus.

FIGURE 6 .
FIGURE 2. Distribution ol species along an elevation gradient on Transcct B l. 1-4.number o f times species was located through the year

FIGURE 7 -
FIGURE 7 -Distribution o f species along an elevation gradient on Transect B6. 1-4, number o f times species was loeated through the year.