The overall objective of this project was to develop and demonstrate genetic approaches for assessing the condition (i.e., integrity) of Pacific island watersheds. Traditional assemblage-based protocols provide little basis for discrimination because the biota of Pacific island streams is naturally depauperate and tightly linked to oceanic environments. Genetic assessment protocols can be more effective tools for quantifying watershed condition because genetic analyses can reveal how environmental stressors influence individuals and populations. However, developing and implementing genetic assessment approaches for Pacific island watersheds requires more thorough knowledge of the demographic and ecological processes—especially dispersal and sensitivity to environmental stressors—that give rise to patterns of genetic variation within species that serve as biotic indicators.

A suite of studies was undertaken that focused on native amphidromous fishes of the Hawaiian Islands to support the development and use of genetic approaches for assessing the integrity of Pacific island watersheds. In the first study, the historical biogeography and contemporary dispersal of native amphidromous fishes across the Hawaiian archipelago were examined. Amphidromy is a form of diadromy, where adults inhabit and spawn in freshwater, larvae drift downstream to near-shore or oceanic environments where they mature for periods of up to six months, after which they recruit back to freshwater streams as postlarvae. The specific objectives of this study were to (1) determine whether different islands harbor distinct evolutionary lineages, and through a multi-disciplinary approach (2) determine whether local recruitment draws from mixed immigrant pools due to larval exchange across the archipelago. This work was intended to help identify the most appropriate spatial scale for management of native fishes and stream ecosystems.

In the second study, the project team assessed among-watershed patterns of genetic variation in relation to local and watershed-scale environmental conditions. This study was intended to illustrate the degree to which key measures of genetic variation (e.g., genetic diversity) within local populations reflect stressor exposure resulting from in-stream habitat degradation or watershed land use patterns. This study was also intended to provide an unprecedented, integrative assessment of watershed integrity across the Hawaiian archipelago. In a related study, genetic indices and how they vary longitudinally (i.e., from headwaters to the stream mouth) within watersheds was examined. This work was intended to demonstrate whether genetic approaches can discriminate potential effects arising from activities on Department of Defense (DoD) installations from effects arising from natural processes or other anthropogenic activities elsewhere in the same watershed.

New methods for assessing the aquatic environmental condition of oceanic island streams were developed from the analysis of genetic variation and otolith microchemistry of native amphidromous fishes across Hawaii.

In the first study, the project team examined within- and among-island patterns of mitochondrial deoxyribonucleic acid (mtDNA) haplotype frequencies in Awaous stamineus and Sicyopterus stimpsoni to infer the presence of island-specific evolutionary lineages and geographic patterns of historical colonization of the archipelago. To assess contemporary dispersal, an existing archipelago-wide model of ocean circulation was coupled to a particle-tracking model to determine the probability that larvae entering the ocean would re-enter their natal stream, streams on the same island, or streams on different islands over two years assuming a short or long larval duration. Model predictions were tested by examining nuclear microsatellite genotype frequencies in both species. The project team also reconstructed movement potential and life history variation of A. stamineus from otolith microchemistry, which involved examining standard isotopic ratios, novel oxygen isotope ratios, and panels of trace elements.

In the second study, the project team assessed watershed-scale patterns of covariance among environmental variables and measures of genetic diversity, population densities, and species diversity. This involved comparison of mtDNA and microsatellite-based estimates of genetic variation to native and non-native species population densities, assemblage structure, water chemistry, and watershed land use. Similar comparisons were conducted to assess relationships across longitudinal transects within four watersheds.

To support both studies, estimates of population densities derived from snorkel surveys were validated through comparison to estimates derived from mark-recapture approaches.

The two studies generated a rich portfolio of findings, including the following:

• A. guamensis in Hawaii is a distinct species (A. stamineus) relative to Indo-Pacific congeners.
• Neither A. stamineus nor S. stimpsoni exhibit island-specific evolutionary lineages across the Hawaiian archipelago; however, neither species exhibits panmictic population structure.
• A. stamineus exhibits a facultative rather than an obligatory amphidromous life history.
• A large proportion of A. stamineus larvae remain in or near their natal stream and ocean-going larvae occupy both near-shore and off-shore marine habitats.
• Genetic diversity of A. stamineus and S. stimpsoni declines with increasing densities of invasive species and watershed conditions related to nutrient loading.
• Population densities and genetic diversity of native fishes in watersheds that host military activity are comparable to levels in forest and ag-urban dominated watersheds on Oahu.
• Population densities of native fishes are depressed in watersheds across Oahu, regardless of land use or stewardship.

These and additional findings yielded the following conclusions:

• Immigrant pools may draw from populations across the archipelago, and the influx of postlarvae may sustain local populations in degraded waterways, but among-island dispersal cannot sustain at-risk populations as it has little influence on local population dynamics.
• Site-specific and watershed-scale conditions may supersede the importance of physiographic conditions in structuring genetic, demographic, and assemblage variation of native amphidromous species in oceanic island streams.
• Aggregate effects can arise from local and watershed-scale degradation, where the cumulative influence of biotic or abiotic stressors can disrupt processes that promote the persistence of native fishes across entire islands.

This research has increased fundamental knowledge of insular stream ecology and responses of native amphidromous fishes to environmental stressors. By demonstrating the extent to which core biological measures of watershed integrity reflect local degradation and the loss of connectivity between streams and adjoining near-shore habitats, these studies also show that integrative protocols tailored to capture estimates of genetic variation alongside measures of population density and species diversity can provide a robust basis for watershed management on oceanic islands. Accordingly, information gained from this research will better enable DoD resource managers and collaborators to maintain habitat for at-risk aquatic species, thereby further ensuring the sustainability of military training activities and guiding the restoration of degraded streams across the Pacific.