Litter Dynamics in Karri (Eucalyptus Diversicolor) Forests of South- Western Australia

Publisher Website
Google Scholar
Abstract
(1) Decomposition and accumulation of litter were studied in karri (Eucalyptus diversicolor) forest, a tall open forest growing in south-western Australia. (2) Rates of weight loss from five leaf and five non-leaf litters were determined using the litter-bag technique over a period of eighty-two weeks. Leaves of karri and the four understorey species Trymalium spathulatum, Bossiaea laidlawiana, Acacia urophylla and Casuarina decussata, karri twigs, bark, and fruit, and seed pods of B. laidlawiana and A. urophylla were used in the study. (3) Early rates of weight loss were significantly related to the proportion of soluble materials initially present in each fraction. Amounts remaining after decomposition for eight-two weeks were significantly related to initial lignin plus cellulose content of the fraction. Weight loss from leaf litter was greater than from non-leaf litter. (4) After eight-two weeks on the forest floor, weight loss from leaf-litter fractions decreased in the order T. spathulatum > B. laidlawiana=karri=A. urophylla=C. decusata, and weight loss from non-leaf fractions decreased in the order A. urophylla seed pods=B. laidlawiana seed pods>karri bark=karri fruit>karri twigs. (6) Single and double exponential decay functions were used to describe weight loss in relation to time of exposure. For each litter type the double exponential model explained a greater proportion of the data variance than the single exponential model. Half-lives of the more resistant component of the litters were lowest for understorey leaf litters (53-127 weeks) intermediate for karri leaves (129 weeks) and greatest for karri non-leaf litters (394-492 weeks). (7) Decomposition constants were combined with published litterfall data to develop a mechanistic model of litter accumulation. The importance of non-leaf litter fractions to the development of the forest floor is demonstrated. The model predicts that the composition of the litter layer changes with time of accumulation, suggesting that average decay constants for the total litter layer may decrease with age. (8) The accumulation model was used to simulate forest floor development in regularly burnt forest. The predictions of the model are compared with measured values of litter accumulation and forest floor composition.