Tidal fluxes of total (TOC), dissolved (DOC) and suspended particulate (POC) organic carbon were measured between a high, Sarcocornia-Chenolea salt marsh and Kariega estuary, South Africa. Fluxes were measured over 42 tides, through four (7 to 14 d) sampling periods. The marsh showed an annual export of TOC of + l 6 g C m-2 yr.', with 80% of this occurring in dissolved form. The export was equivalent to 6% of the aerial macrophyte production of the marsh. In both absolute terms, and relative to macrophyte production, the fluxes were simllar to those reported for hlgh elevation, Spartina patens and Distichlis spicata marshes on the east coast of the USA. They were, however, an order of magnitude smaller than for most low, Spartina dlterniflora marshes. A carbon budget developed for the Kariega marsh showed that respiration (mainly by the sediments and to a lesser extent by resident crabs) accounted for 7 0 % of marsh production, leaving 30% for burial or export to the estuary. We speculate that the small export was a funct~on of the high elevation of the marsh, and that the Outwelling Hypothesis may be less applicable to high than low marshes.

Laboratory mesocosm experiments were conducted to quantify the effects of 2 crab species, Sesarma catenata and Cleistostoma edwardsii, on carbon exchanges across the surface of a south temperate salt marsh. The grapsoid crab S. catenata was most abundant in the vegetated marsh flats, and the ocypodoid C. edwardsii In the unvegetated tidal creek. Replicate mesocosms of the marsh flats and tidal creek were incubated with and without crabs of the species dominant in that reglon. Both species enhanced the losses of carbon relative to uptake by the marsh, but they did so vla different mechanisms. In the tidal creek. C. edwardsii increased the net fluxes of total organic carbon (TOC) by 60 % (or 448 mg C m-' d-l), but had no significant impact on community net production (NP). (Bioturbation accounted for 95 %of the enhanced TOC fluxes, and excretion only 5 X). In the marsh flats, S. catenata had no measurable impact on the fluxes of TOC, but decreased community NP by 1132 mg C m-' d-l (Ninety-two percent of this reduction was caused through reduction in epibenthic NP through crab grazlng of mlcroalgae. Crab respiration accounted for only 8 % of the reduction). The crab effects were sufficient to determine whether, for a tide of a particular tlde elevation, the marsh functioned as a carbon source or sink. When crabs were absent, the marsh functioned as a carbon sink at all except the very hlghest tides that, in turn, accounted for only 3 %of all tides inundating the marsh. When crabs were present, the marsh functioned as a carbon source at 87 %, and all except the very lowest tides.

No abstract