Lost in translation: Reconciling different streamflow permanence data products

The objective of this study is to provide a framework to reconcile streamflow permanence products that provide information on whether a stream is perennial or nonperennial. Accurate classification of streams as perennial or nonperennial is important for a variety of land and water resource management decisions. However, resource managers are challenged with how to reconcile different streamflow permanence information to determine if it is sufficient to avoid field verification, which can be prohibitively costly and time consuming. The study evaluates two datasets for the Pacific Northwest Region, U.S., the National Hydrography Dataset Plus High Resolution (NHDPlus HR) hydrographic classification and the PRObability of Streamflow PERmanence (PROSPER) model output, which differ in structure, output type, and temporal and spatial resolution. The framework includes a two-level evaluation that involves assessing agreement between the two datasets and evaluating the reliability of the two datasets. Summarized at a regional scale, the two datasets agree for 68% of flowlines with higher agreement for nonperennial streams. PROSPER nonperennial stream classifications may be considered generally reliable in arid regions relative to other ecoregions within the Pacific Northwest study area but less reliable for high elevation mountain regions and larger streams and rivers. More than 75% of NHDPlusHR classifications may be considered reliable when evaluating reliability based on approximate climate conditions associated with the year that the NHDPlus HR classification was assigned. The decision procedure is reproducible, flexible to varying criteria of end user applications, and is intended to help provide cost saving opportunities for land managers by providing information for more strategic field verification of streamflow class determination based on available yet imperfect data sources.