Comparisons of riparian vs. over stream trap location in the estimation ofvertical litterfall inputs

Vertical inputs of allochthonous organic matter to streams can be estimated directly with traps; traps may be placed over the stream, in the riparian forest or at other locations in the catchment. In this study, we compared l itter inputs estimated from traps located over the stream with those obtain ed from traps placed on the stream bank in two stretches (orders 2 and 3) o f the Mera stream, with similar tree species composition and riparian fores t structure. The amount of annual litter input measured with the traps on t he stream was similar in the two stretches, 748 and 730 g m(-2) yr(-1), for orders 2 and 3, respectively. However, there appeared to be differences in the composition of litter; a lower percentage of leaves and a higher perce ntage of wood was found in the stretch of order 3 than in the stretch of or der 2. In both sites, the mean input measured under the riparian forest was lower than the input recorded with the traps placed over the stream by app roximately 10%. In relation to the traps located over the stream, the traps located on the bank overestimated the input from adjacent terrestrial tree species and, therefore, underestimated the input from riparian species.