Cosmogenic nuclides have been used successfully to quantify long-term rates of sediment generation in arid lands. However, cosmogenic nuclides have been applied outside of arid environments in only a few locations. We have collected 37 samples from the tropical, humid Rio Chagres watershed in Panama as a methodological test. Since the headwaters of the Rio Chagres watershed are remote, and relatively undisturbed by humans, it is an ideal tropical location to test the cosmogenic method. We request modest funding for 24 analyses to investigate the utility of the cosmogenic nuclides as a sediment tracer and sediment generation monitor in tropical environments. Such a ‘proof of method’ study in tropical environments will broaden the geographic scope over which cosmogenic nuclides can be used to understand societal and environmentally relevant geologic processes including erosion and sediment transport. INTRODUCTION Quantifying rates of landscape change is fundamental to understanding the natural environment and the persistence of human impacts. Natural rates of landscape change are often slow and imperceptible. Field monitoring techniques, often used to measure landscape change, are labor intensive and time consuming (e.g. Abrahams et al., 1984; Lekach and Schick, 1999). Such methods require personnel and intensive monitoring over several years to decades to understand rates of landscape change. Even after several years of data collection, it is difficult to extrapolate such short-term rates to longer, geologic rates of change because it is difficult to determine if the geomorphic events captured during the study period are representative of the long-term events that shape the landscape (Baker and Twidale, 1991; Kirchner et al., 2001). Furthermore, in order to understand the human impact on landscape processes, it is necessary to compare the present day rates to longer-term average rates, which are often unavailable (Clapp et al., 2000; Trimble, 1999). Over the past several years many studies have demonstrated the utility of analyzing cosmogenic nuclides in sediment in order to determine long-term average sediment generation rates (Clapp et al., 2002, 2001, 2000; Nichols et al., 2002; Bierman et al., 2001a, 2001b; Kirchner et al., 2001; Matmon et al., 2001a, 2001b; Schaller et al., 2001; Bierman and Steig, 1996; Granger et al., 1996; Brown et al., 1995). Such long-term cosmogenic-nuclide-based sediment generation rates provide baseline data for comparison to present-day sediment generation rates. Without measuring such long-term average rates of sediment generation, it would be impossible to determine if the present day sediment export rates are greater (e.g. Clapp et al., 2000) or are less (e.g. Kirchner et al., 2001) than the long-term average. Furthermore, wide scale coverage allows one to determine areas of a basin that have high rates of sediment generation. By knowing the sediment generation rates at several locations in a basin, it is possible to develop a mixing model that tracks the sediment from the source areas through the drainage basin to the outlet (e.g. Clapp et al., 2002). Our proposed research, a ‘proof of method’ study, has the potential to expand the use of cosmogenic nuclides in the quantification of long-term, basin wide sediment generation rates (e.g., Brown et al., 1995, Clapp et al., 2000). Much of the initial application and method development for reliably using cosmogenic nuclides to measure such long-term, basin-wide sediment generation rates was supported by previous and current ARO funding. However, all the previous projects were conducted in arid regions, Yuma Proving Ground (Clapp et al., 2002), Nahal Yael in Israel (Clapp et al., 2000), the former Patton Camps of the Desert Training Center (Nichols et al., 2002), and Fort Irwin (Nichols et al., 2001). These projects have successfully quantified basin-wide erosion rates and sediment generation rates using cosmogenic nuclides. In this proposal, we request funds to analyze 24 samples from the humid, tropical Rio Chagres watershed in Panama, collected as part of the U.S. Army-supported Rio Chagres Expedition in the winter of 2002. The geomorphic processes that shape the tropical Rio Chagres watershed are different than the processes that modify the hyper-arid basins at the Yuma Proving Ground, Nahal Yael, the Patton Camps, and Fort Irwin. Proof of this method in tropical environments will greatly improve the geographic and climatic range over which cosmogenic nuclides can be used to determine basin-wide erosion and sediment generation rates. The analysis of the Rio Chagres samples has an additional benefit of supporting ongoing hydrologic modeling of the Rio Chagres watershed. Long-term average sediment generation rates will provide data that will complement hydrologic and paleohydrologic research. Integration of long-term basin sediment generation rates with hydrologic data allows modeling of basin behavior over longer time and spatial scales than either data set alone would allow. PROJECT DESCRIPTION Site location The Rio Chagres discharges into Lake Alhajuela, the reservoir for the Panama Canal (Figure 1). The Rio Chagres watershed is 400 km and the relief is 1500 m (Figure 1). The watershed supplies water for the Panama Canal and is the water supply for Panama City; thus, development is limited (Loewenberg, 1999). Such pristine and undisturbed conditions suggest that the Rio Chagres watershed is an ideal location to test the use of cosmogenic nuclides on tropical basins, because there is little human disturbance. Work Plan The work plan is divided into two distinct phases: sample collection and sample analysis. The first phase has been completed. Sample collection In the winter of 2002, one of Bierman’s graduate students, Nichols, was part of a team that collected 37 sediment samples from the Rio Chagres watershed for cosmogenic nuclide analysis (Figure 2). We collected samples from two different types of geomorphic settings: 1) small headwater tributaries and 2) major tributaries that flow into the Rio Chagres. The samples collected from 17 small tributaries will help us to understand sediment generation rates in the headwater basins (Figure 2). Samples collected downstream of the major confluences will allow us to develop mixing models to estimate the percentage of sediment supplied by each river branch or tributary (Figure 2). As indicated by the only other cosmogenic study of fluvial sediment in a tropical climate, Luquilo Rainforest in Puerto Rico, nuclide activity appears to depend on grain size (Brown et al., 1995). A recent study in the humid Great Smoky Mountains also demonstrates that different grain sizes have different nuclide activities (Matmon et al., 2001). Both studies show that larger grain sizes can have lower nuclide activities than smaller grain sizes. There are two explanations for the low nuclide activities in large grain sizes: 1) the larger grain sizes are delivered by deep-seated landslides in which cosmic ray dosing is minimal (Brown et al., 1995) and 2) the larger grain sizes have a local source and thus were dosed at a lower elevation (Matmon et al., 2001b). These finding demand that we analyze multiple grain sizes for several of the samples. Sample analysis Of the 37 samples collected, we request support to analyze Be in 16 samples and analyze an additional grain size in 8 of these samples. Analyzing at least 6 headwater tributary basin samples will allow us to understand the sediment generation rates in the steep headwater basins. Analyzing at least three major confluences (9 samples) spaced along the Rio Chagres will allow for estimates of the areas of the watershed that are major sediment inputs. A sample at the mouth of the Rio Chagres will estimate the sediment input into Lake Alhajuela. In addition, we will analyze multiple grain sizes in samples from several headwater tributary basin samples and from several confluence samples (a total of 8 analyses) to determine the grain size dependence of cosmic ray dosing. The samples will be processed at the University of Vermont Cosmogenic Isotope Extraction Laboratory. Successful accelerator mass spectrometry (AMS) analyses of two preliminary Rio Chagres samples, collected by Harmon of the ARO in 2001, proves that we can analyze lightly dosed samples of the Rio Chagres Basin. Due to the high erosion rates and short residence times of sediment it is not necessary, or cost effective, to analyze Al in addition to Be. We will follow methods that we have been developed in our laboratory over the last 8 years to measure low-level samples (Davis et al., 1999). We will run four batches of six samples, each batch having two procedural blanks. Such protocol allows for precise measurement of the blank and thus high precision analysis of the samples. Accelerator analyses will be made at Lawrence Livermore National Laboratory where Bierman has been collaborating for 10 years. We will present the data at a national meeting of the Geological Society of America meeting and in an article for publication. EXPECTED OUTCOMES We expect that our results will be an important data set leading to a ‘proof of method’ allowing the application and interpretation of cosmogenic nuclides in sediments collected from tropical rivers. Our widespread sampling of the Rio Chagres watershed may allow us to identify portions of the basin that are susceptible to large increases in sediment yield. In addition, we will be able to compare our long-term sediment yields to more than 20 years of total suspended sediment data to determine the anthropogenic impact on the watershed.