The saltmarsh vegetation of Langebaan Lagoon

The saltmarshes of Langebaan Lag<x>n arc the most extensive in southern Africa. These marshes, as sampled along six transects, are described. A general marsh, consisting of three species assemblages, was recognized and elevation above mean sea level (MSL) is discussed as a probable determinant of species distributions. However, minor variations in species distributions have been induced by changes in soil characteristics, the effects of wind on inundation depth and differences in water salinity. Langebaan se soutmoerasse is die mees uitgebreide in suidelike Afrika. Hierdie moerasse, soos bestudeer langs ses lynopnames. word beskryf. n Algemene soutmoeras, bestaande uit drie spesiegroepe. kan herken word. Hoogte bo seevlak (MSL) is waarskynlik bepalend vir die verspreidingspatrone van spesies. Nietemin kan klein variasies van hierdie patrone veroorsaak word deur veranderinge in grondeienskappe, die effek van w ind op waterdiepte en verskillende watersoutgehaltes.


INTRODUCTION
The saltmarshes of Langebaan Lagoon are the most extensive in the temperate zone of southern Africa. At 5 700 ha, this Lagoon contains over 30% of South Africa's saltmarsh areas (O'Callaghan 1990). Most of the Langebaan marshes have been protected since 1988 as part of a nature reserve.
These saltmarshes have developed under unique con ditions as there is no riverine flow into the Lagoon. All other saltmarsh development along the South African coast occurs in estuaries where salinity and tidal charac teristics result from interactions between marine and riverine water bodies. There is, however, an extensive fresh water seepage into Langebaan Lagoon at the southeastern comer (Shannon & Stander 1977) and the wetland vegetation in this area is different from the true saltmarshes. These less saline wetlands were excluded from this study. They have been described by Boucher & Jarman (1977) and Boucher (1987).
The saltmarshes are integral to the functioning of many biotic components of the Lag (x>n |Puttick 1980;Whitfield et al. 1989; see Christie (1981) tor estimates of macrophyte production]. The biology of this Lagoon was dis cussed in detail at a symposium on the area (Siegfried 1977).
The most extensive development of marshes occurs at the southern and southeastern parts of the Lagoon. In these areas, complex channel systems allow water to flow into large low-lying backshore areas where marshes develop. The inundation and salinity features of the waters flooding these parts are highly variable and the backshore areas were excluded from this study. Boucher & Jarman (1977) maintained that the marsh communities ot Langebaan Lagoon can be distinguished by the presence of Sarcocomia pillansii. They divided the marshes into Juncus kraussii Dense Sedgelands and Chenolea-Salicomia Dwarf Succulent Shrublands. The former describes the fresher water marsh areas and the latter describes the saltmarshes. Boucher (1987) included the above data to establish a class of halophytic com munities called Sarcocomietea pillansiae. although he mentioned that the establishment of clear-cut units is hindered by the low' species densities of these commu nities. The associations included in this order are slight modilications of the concepts of Boucher & Jarman (1977). Day (1959) briefly described the zonation of saltmar shes at Langebaan Lagoon. The top of the marsh is indi cated by the presence of Sporobolus virginicus and terrestrial species which are found at the extreme high water spring tide level. From this point to MHWS (mean high water spring), a mixed zone of Salicomia mexeriana, Limonium scabrum and Chenolea diffusa occurs. Between MHWS and MHWN (mean high water neap), SarccKornia perennis and Triglochin bulbosa dominate. Spartina maritima is found from the bottom of this zone to the top of the zone which occurs below MSL.
All the descriptions above correlate somew hat. Zostera capensis grow s below M SL. follow ed by Spartina maritima. A mixed zone is found up to the MHWS mark, above which terrestrial species make an appearance. How ever. there is some confusion when these patterns are compared to Boucher's (1987) communities, particularly the required presence of Sarcocomia pillansii. The pur pose of this paper is describe the marshes of Langebaan Lagoon in some detail.

M ETHODS
After studying aerial photographs, orthophotographic maps and following field reconnaissance, six transects were demarcated across the marshes of Langebaan Lagoon ( 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 each plot 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 voucher specimens are housed at the herbarium of the National Botanical Institute at Stellenbosch (STE), the Na tional Herbarium (PRE) and at the Stress Ecology Re search Unit at Kirstenbosch. These voucher specimens are listed by O'Callaghan (1994a).
As classical Braun-Blanquet values cannot be manipu lated mathematically, these values were converted accord ing to Table 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 indicates the number of times this species was located through the year. The order in which the species occur along the transect is primarily determined by its lowest starting point and secondarily by its termination point along the elevation gradient.

Species distributions along elevation gradients at Tran sects LI to L6 are shown in Figures 2-7.
The zonation patterns at Transects L4 and L5 are con sidered to be typical of the saltmarshes at Langebaan Lagoon, for the following reasons: 1, these profiles were topographically regular with very few gullies or rapid changes in elevation. The shore ex tended gently into the Lagoon without promontories, rocks or outcrops which could affect currents and tidal fluctuation; 2, as a result, the species were distributed continuously along the elevation gradient. At L4, a single specimen of Puccinellia angusta was found in the mid-marsh during September, resulting in a split distribution. By November, this plant had disappeared. At L5, a path existed from 1 (K) cm to 120 cm above MSL. This resulted in a split distribu tion for Chenolea diffusa. Sheep were removed from the area towards the end of this year and. by November, Sarcocomia pillansii and Suaeda inflata had expanded to cover most of the path; 3, the majority of species shown were present during the entire year, indicating that environmental conditions were relatively stable. However, all the species above an eleva tion of approximately 80 cm showed signs of grazing at  L5. This might have had a differential effect on the presence of some of the species: a) Puccinellia angusta is an annual grass which germinates in winter and flowers by spring. At L4, the dead plants remained in place throughout the whole year. It was only recorded during November at L5, once the seeds had germinated in spring; b) at L5, Limonium depauperation was not recorded in March. The growth form of L. depauperatum (a single stem growing higher than the surrounding vegetation) is such that it would have been most affected by grazing. This species was again evident from June onwards.
The distribution of species along these two transects will be discussed first and the other transects will be re lated to them.

3, the upper assemblage (Suaeda inflata. Sarcocomia pillansii. Disphyma crassifolium and others).
Although this terminology is similar to that of Chap man (1976. 1977), Beeftink (1977), Adam (1978Adam ( . 1981, and others, it does not correspond with their concepts of   for these changed relationships is uncertain, but they were observed where the species tended towards inversion.

Transects LI and L3 (Figures 2 & 4)
Relative to L4 and L5, the species distributions were displaced up the shore by 10 cm at LI and by 60 cm at L3. This can be accounted for by an interaction between major currents (Flemming 1977a) and the predominant southeasterly winds which continuously push the water high up onto the transect. The effect is more pronounced at L3 where a sand-spit to the north of the transect traps the water more readily.
Higher current speeds and coarser substrata along these western shores (Flemming 1977b) limited the growth of Zostera capensis. It was only found at L3 from July (i.e. with the seasonal abatement of the strong southeasterly winds) and had increased by November. It seems to have a seasonal or cyclical occurrence in this area. Furthermore, it was found relatively high on the marsh where some protection and possibly fine organic substrate was provided by Spartina maritima.
The presence of Sarcocomia littorea at LI is unusual (Tolken 1967). This species usually grows in cracks and crevices on rocky marine shores, just above extreme high water. On such a rocky shore, the species would seldom be inundated, but would often be exposed to salt spray. The soils would be coarse w ith little organic content. The upper layers of the soil would be well aerated, but deeper root development would be limited. At LI, this species was found on a mound above MHWS, but low down on the transect. It would have been exposed to salt spray from the southeasterly winds. It occurred on coarse sands, but soil depth would be limited by a high water table. It could be speculated that the conditions under which this species was found on this transect were similar to those under which it normally grows on rocky shores.
More than half the species above the level of Suaeda inflata at LI had an annual life cycle. Cotula eckloniana first appeared in June, Crassula decumbens and Senecio littoreus appeared in September. These three species are often found after spring and summer draw-downs in fresher water pans (pers. obs.). Heavy winter rains could briefly freshen the heavy soils of the upper parts of LI. As these species appeared above HAT. the environment could have briefly simulated fresher water pans.
The species at the top of LI are often found on ter restrial arid and/or salini/ed soils.

Transects L2 and I/) (Figures 3 & 7)
Much of the saltmarsh development along the eastern shores of the Lagoon was similar to that already described. However, seepage of fresh water into the Lagoon was an added factor influencing the vegetation, particularly at L2 and L6.
Juncus kraussii was found in depressions near the top of L2 where occasional incoming saline water was diluted by a high terrestrial water table. Fresh water seeps into the Lagixm immediately north of L6 ( Boucher & Jarman 1977;Boucher 1987). This fresher water floods L6 during parts of the year, supporting the growth of Schoenoplectus triqueter and restricting Spartina maritima to a narrow fringe higher up the marsh.
Juncus kraussii can survive in two distinct habitats: (1) around fresh water pans and wetlands (e.g. Kleinmond Lagoon, pers. obs.); and (2) on upper tidal marshes where salinities rarely exceed 209£c (e.g. Bree River. O 'Callaghan 1983). Although it prefers medium to low salinities, it can withstand occasional inundation by saline water. This species was found at 80 cm and again at 120 cm at L6. Conditions at the lower distribution might be interpreted as habitat 1 whereas conditions at the upper distribution might be similar to those in habitat 2. This split in the distribution of Juncus kraussii was also noted at the Berg River (pers. obs.).

CONCLUSIONS
A number of species recorded by Boucher (1987) were not found during the present study (Limoniurn equisetinum; Diplachne fusca, Boucher 2820. Drosanthemwn floribunda). Some of these inconsistencies are due to differences in species concept (e.g. Drvsanihemum floribunda and D. delicatula). However, some of the differences might be due to different sampling scales used in the different studies. Diplachne fusca. for example, is usually found in less saline areas (Kleinmond Lagoon. O'Callaghan 1994b). It is unlike ly to be found in the saltmarshes per se at Langebaan Lagoon. But it might have been found in the fresher marshes north of L6. The less intensive but laiger scale sampling techniques used by Boucher & Jarman (1977) and Boucher (1987) might have resulted in an erroneous inclusion of this species in the saline marshes.
With minor specific variations, the present description of the saltmarshes around Langebaan Lagoon corresponds well with those presented by Day (1959) and Macnae (1957). but not w ith Grindley's description of saltmarshes at Knvsna Lagoon (1985).
However, variations in the relative distributions of the species can be brought about by an instability in the relationship between environmental controlling factors. Further investigation is required, especially as these areas are often exposed to other disturbances such as the deposi tion of wracks and trampling. Although there was some evidence of direct human-induced disturbance (e.g. graz ing and trampling), this disturbance is relatively minor and should decrease as much of the area is managed as part of the West Coast Nature Reserv e.