Handcart Gulch is an alpine watershed along the Continental Divide in the Colorado Rocky Mountain Front Range. It contains an unmined mineral deposit typical of many hydrothermal mineral deposits in the intermountain west, composed primarily of pyrite with trace metals including copper and molybdenum. Springs and the trunk stream have a natural pH value of 3 to 4. The U.S. Geological Survey began integrated research activities at the site in 2003 with the objective of better understanding geologic, geochemical, and hydrologic controls on naturally occurring acid-rock drainage in alpine watersheds. Characterizing the role of groundwater was of particular interest because mountain watersheds containing metallic mineral deposits are often underlain by complexly deformed crystalline rocks in which groundwater flow is poorly understood. Site infrastructure currently includes 4 deep monitoring wells high in the watershed (300– 1,200 ft deep), 4 bedrock (100–170 ft deep) and 5 shallow (10–30 ft deep) monitoring wells along the trunk stream, a stream gage, and a meteorological station. Work to date at the site includes: geologic mapping and structural analysis; surface sample and drill core mineralogic characterization; geophysical borehole logging; aquifer testing; monitoring of groundwater hydraulic heads and streamflows; a stream tracer dilution study; repeated sampling of surface and groundwater for geochemical analyses, including major and trace elements, several isotopes, and groundwater age dating; and construction of groundwater flow models. The unique dataset collected at Handcart Gulch has yielded several important findings about bedrock groundwater flow at the site. Most importantly, we find that bedrock bulk permeability is nontrivial and that bedrock groundwater apparently constitutes a substantial fraction of the hydrologic budget. This means that bedrock groundwater commonly may be an underappreciated component of the hydrologic system in studies of alpine watersheds. Additionally, despite the complexity of the fracture controlled aquifer system, it appears that it can be represented with a relatively simple conceptual model and can be treated as an equivalent porous medium at the watershed scale. Interpretation of existing data, collection of new monitoring data, and efforts to link geochemical and hydrologic processes through modeling are ongoing at the site.