Developing stream network topologies for the Yukon and Kuskokwim rivers

A central challenge to developing a comprehensive understanding of the river systems of the Arctic-Yukon-Kuskokwim (AYK) region, and the natural resources they support, is a robust analytical framework that accounts for and takes advantage of the spatial connectivity of rivers. The basis of such a framework is a topologically correct stream network, which explicitly defines the multi-dimensional spatial relationships of dendritic river networks, such as longitudinal connectivity, branching networks of streams, and abrupt changes at tributary confluences. In the last five years, a new class of powerful geostatistical models have been developed for river systems – spatial stream network models (SSNMs). Using the Alaska National Hydrography Dataset (NHD) and the Canadian National Hydrological Network (NHN) as our stream baseline data, we developed topological correct stream networks suitable for use in spatial stream network model (SSNMs) for the Yukon and Kuskokwim drainages. Two datasets were generated from for us in the SSN models: 1) the first dataset retains the original segmentation and attributes from the source data (NHD and NHN networks), 2) the second dataset consists of a simplified network that eliminates 1st and 2nd order streams and generates only one stream segment between confluences. The simplified stream network reduces network size and greatly increases processing time. The addition of these networks throughout the AYK region should pave the way for substantial advances in the research and management of this region’s natural resources, including Chinook salmon. For example, the data products generated from these networks can be used to develop ‘isoscapes’ of strontium isotope ratios against which the strontium isotopes in Chinook salmon otoliths can be compared to assign individual adult salmon to their natal origins across the river basins. The Yukon and Kuskokwim stream networks that maintain the original segmentation and attributes will be available for public dissemination at the National Streams Internet (NSI) project (http://www.fs.fed.us/rm/boise/AWAE/ projects/NationalStreamInternet.html).

The development of topological correct stream networks for the Yukon and Kuskokwim plays an integral role in the generating the strontium isoscapes, which we are using to determine the natal origins and movement patterns of Chinook salmon across the Yukon and Kuskokwim basins. Initial applications of the stream networks will include integration with our existing AYK SSI-funded project using strontium isotope ratios recorded in Chinook salmon otoliths to assign individual fish to their birth locations throughout the Yukon River (AYK#1515). From previous work, (Brennan and Schindler 2017) developed an analytical approach, which integrates i) isoscapes built via SSNMs, and ii) a continuous Bayesian assignment framework within the Nushagak River to map the production patterns of returning adult Chinook salmon. A similar approach is being used to model Chinook salmon production patterns across the Yukon and Kuskokwim Rivers. In addition, the stream networks will provide integral baseline data and overall framework for two other AYK funded projects (“Integrating genetics with otolith microchemistry to reconstruct production patterns of Chinook salmon throughout the Yukon River Basin”, #1705, and “Building a strontium isotope baseline of the Kuskokwim River able to reconstruct the production patterns and life histories of Chinook salmon at fine spatial scales”, #1704). Furthermore, in the Kuskokwim, we are using this newly developed stream network and associate covariates to model the controls on spatial variation in bioavailable mercury in slimy sculpin collected from across the Kuskokwim, and also to explore spatial controls on dissolved organic carbon and nutrient transport through this river. Thus, the clean stream networks developed herein, are already opening the doors to valuable spatially analyses and research that previously was not possible within these vast river networks. The stream networks have wide-ranging implications for numerous other objectives and goals throughout the AYK region, including tracing pollutant transport, other chemical constituents in rivers, temperature patterns, and flow regimes. To this end, the topologically correct stream networks will be made publically available at the National Stream Internet (NSI) project http://www.fs.fed.us/rm/boise/AWAE/ projects/NationalStreamInternet.html) in order to facilitate interdisciplinary use of the resulting data products across the AYK region.