EP31C-3579:
FORM FOLLOWS FUNCTION: CAN TROPICAL MOUNTAIN FOREST COMPETITION DRIVE THE GROWTH OF TOPOGRAPHY?

Wednesday, 17 December 2014
Gilles Y Brocard1, Jane K. Willenbring2 and Frederick N Scatena2, (1)Univ of Pennsylvania, Philadelphia, PA, United States, (2)University of Pennsylvania, Department of Earth & Environmental Sciences, Philadelphia, PA, United States
Abstract:
Forest succession theory maintains that trees drape existing landscapes as passive niche optimizers. Here, we explore the opposite view – that tree type can control the landscape morphology through canopy and soil structure differences between tree types. Using field observations combined with the analysis of a 1 m-resolution LiDAR DEM and cosmogenic nuclide geochemical techniques, we report links between topographic position, erosion rates, and tree distribution above 600 m in elevation in the pristine, tropical rainforest of the Luquillo Critical Zone Observatory, Puerto Rico. Above 600 m, flat ridges connect a subdued topographic surface and cap a relict landscape composed of 10s-of-meters-thick saprolite derived from a quartz dioritic bedrock. Erosion in the form of concave coves dissects this landscape, exposing the underlying corestones and bedrock and progressively creating a new landscape of greater local relief and higher erosion. To understand the origin of the increase in local relief we measured the respective contributions of ridges and coves to the stream signal by analyzing 10Be in soils and stream sediments. We find that, in the coves, erosion is systematically higher than on the ridges, confirming the increase in landscape dissection. Because the overall forest structure has not been altered by anthropogenic disturbances, the tree type represents a long-term feature of the landscape. We analyzed the distribution of tree associations over the LiDAR topography by classifying high-resolution multispectral images of the forest. Vegetation in the study area is dominated by Palm and Palo Colorado forests. We found an almost systematic association of the Palm forest with the coves and of the Palo Colorado with the ridges. These forest types vary in their potential to prevent or enhance soil erosion. The Palo Colorado generates greater soil coverage and rain interception rates. Therefore, we propose that soils are more vulnerable to erosion under the Palm forest, providing a positive feedback whereby the presence of the Palm forest favors the incision of the coves while the Palo Colorado forest favors the protection of the ridges, thereby increasing local relief.