The roles of grass leaf litter in streams draining tussock grassland in New Zealand: retention, food supply and substrate stabilization

1. We investigated the roles of grass litter in streams that drain upland New Zealand tussock grassland, paying particular attention to the ways in which grass leaves differed in their characteristics from much more intensively studied tree leaves. 2. The instantaneous retention rates of tussock grass leaves (Chionochloa rigida) on the bed of a second-order stream (0.157–0.515 m^?1) were significantly higher than those of the elliptical leaves of lemonwood trees (Pittosporum eugenioides; 0.068–0.180 m^?1). 3. Instantaneous retention rates of grass leaves in two third-order streams were very low. At high discharge, leaf retention rate was greater in Timber Creek (0.0040 m^?1), a braided, gravel bed stream, than in the adjacent Kye Burn (0.0010m^?1), with its well-defined channel and large, stable substrate particles. At baseflow, retention rates were similar in the two streams (0.0053–0.0064 m^?1 for Timber Creek; 0.0047–0.0058m^?1 for Kye Burn). Nevertheless, total coarse particulate organic matter (mainly derived from tussock litter) was present at lower densities in Timber Creek than Kye Bum, reflecting the instability of the bed of the former and its tendency to spread over a wide area at high discharge. 4. The results of a colonization tray experiment, in which substrate was mixed with tussock leaves, nylon ribbon or nothing, indicate that tussock leaves do not play an important role as microhabitat or food in the two third-order streams. This may reflect the poor food quality of grass litter, and/or the relatively low availability and predictability of its supply. 5. An in situ experiment revealed that tussock leaves play a role, analogous to that reported for certain seagrasses, in stabilizing substrate and reducing sediment transport, apparently by reducing bed roughness and therefore the force of friction on the bed.