A Comparison of Snorkel Surveys and Mark-Recapture Density Estimations with Application in Hawaiian Streams

The extreme isolation of the Hawaiian archipelago has resulted in low diversity, high endemism and strong representation by amphidromous taxa among the freshwater fauna.  Due to the obligatory migration of the Islands’ freshwater fish species, modeling impacts of land use, in-stream habitat and exotic species presents additional and unique challenges not encountered in continental systems.  Further, electrofishing is generally not permitted by state agencies, leaving snorkel surveys as the most common population estimator used by stream biologists.  Population density represents a key measure of condition that can be quickly assessed by snorkel surveys and more intensively by mark-recapture.  Our research investigates the validity of snorkel surveys for population estimation in comparison to mark-recapture estimates and explores the most effective watershed land use model for predicting densities of native and exotic species across the Hawaiian archipelago. Snorkel surveys were conducted at 75 sites across 30 watersheds on five of the Hawaiian Islands between May and November, 2009.  Monthly snorkel surveys and mark-recapture experiments were then conducted from May 2010 to March 2011 at three sites in each of three watersheds on the Big Island and one watershed on Oahu.  The four watersheds were chosen to represent a gradient of land use and stewardship, as well as presence or absence of introduced poeciliid fishes.  Mark-recapture experiments on two native species, Sicyopterus stimpsoni and Awaous guamensis provided long-term estimates of population density.  Over 1800 individuals were marked.  Snorkel surveys and mark-recapture provided remarkably consistent estimates of adult population density for A. guamensis.  Long-term mark-recapture proved to be an inefficient means to estimate S. stimpsoni density at this scale, due to excessive dispersal by and/or densities of adults.  Analysis of snorkel survey data indicates that population densities of native fishes, as a group, are negatively correlated with percent of agriculture and urban land use in watersheds, whereas introduced poeciliid species are positively correlated.  Responses vary among species and between islands, with Oahu being particularly depauperate of native species even in relatively pristine watersheds.  Upstream and downstream watershed land use varies in magnitude of influence on species densities.  Densities of native and introduced poeciliid species are negatively correlated.  However, it is unclear, from this data, what additional impact introduced fishes are having beyond land use.   Future work will evaluate the use of simple Lincoln-Petersen estimations compared to snorkel surveys and long-term mark-recapture estimates.