Gap size measurement: the proposal of a new field method

In the tropical Atlantic Forest, 42 canopy gaps had their areas estimated using four different field methods of measurement: Runkle, Brokaw and Green [Runkle, J.R., 1981. Gap formation in some old-growth forests of the eastern United States. Ecology 62, 1041-1051; Brokaw, N.V.L., 1982. The definition of treefall gap and its effect on measures of forest dynamics. Biotropica 14, 158-160; Green, P.T., 1996. Canopy Gaps in rain forest on Christmas Island, Indian Ocean: size distribution and methods of measurement. J. Trop. Ecol. 12, 427-434] and a new method proposed in this work. It was found that within the same gap delimitation, average gap size varied from 56.0 up to 88.3 m(3) while total sum of gap area varied from 2351.3 to 3707.9 m(3) Differences among all methods and between pairs of method proved to be statistically significant. As a consequence, gap size-class distribution was also different between methods. When one method is held as a standard, deviation on average values of gap size ranged between 11.8 and 59.7% as deviations on single gap size can reach 172.8%. Implications on forest dynamics were expressed by the forest turnover rate that was 24% faster or 15% slower depending on the method adopted for gap measurement. Based on my results and on methods' evaluation, the use of a new method is proposed here for future research involving the measure of gap size in forest ecosystems. Finally, it is concluded that forest comparisons disregarding the influence of different methods of gap measurement should be reconsidered. (c) 2005 Elsevier B.V. All rights reserved.