Mangrove litter fall: Extrapolation from traps to a large tropical macrotidal harbour

Mangrove litter is a major source of organic matter for detrital food chains in many tropical coastal ecosystems, but scant attention has been paid to the substantial challenges in sampling and extrapolation of rates of litter fall. The challenges arise due to within-stand heterogeneity including incomplete canopy cover, and canopy that is below the high tide mark. We sampled litter monthly for three years at 35 sites across eight mapped communities in the macrotidal Darwin Harbour, northern Australia. Totals were adjusted for mean community canopy cover and the occurrence of canopy below the high tide mark. The mangroves of Darwin Harbour generate an estimated average of 5.0 t ha−1 yr−1 of litter. This amount would have been overestimated by 32% had we not corrected for limited canopy cover and underestimated by 11% had we not corrected for foliage that is below the high tide mark. Had we made neither correction, we would have overestimated litter fall by 17%. Among communities, rates varied 2.6-fold per unit area of canopy, and 3.9-fold among unit area of community. Seaward fringe mangroves were the most productive per unit of canopy area but the canopy was relatively open; Tidal creek forest was the most productive per unit area of community. Litter fall varied 1.1-fold among years and 2.0-fold among months though communities exhibited a range of seasonalities. Our study may be the most extensively stratified and sampled evaluation of mangrove litter fall in a tropical estuary. We believe our study is also the first such assessment to explicitly deal with canopy discontinuities and demonstrates that failure to do so can result in considerable overestimation of mangrove productivity.

Figure optionsDownload high-quality image (123 K)Download as PowerPoint slideHighlights
► the mangroves of a macrotidal harbour yielded 5.0 t ha−1 yr−1 of leaf litter.
► this estimate accounts for floristic diversity, canopy gaps and canopy inundation.
► estimation requires mapping of community and extensive stratified sampling.
► failure to allow for gaps and inundation would generate an overestimate of 17%.