Prioritizing management goals for stream biological integrity within the developed landscape context

• Published in: Freshwater science Vol. 38; no. 4; pp. 883 - 898
• Main Authors: Beck, Marcus W., Mazor, Raphael D., Johnson, Scott, Wisenbaker, Karin, Westfall, Joshua, Ode, Peter R., Hill, Ryan, Loflen, Chad, Sutula, Martha, Stein, Eric D.
• Format: Journal Article
• Language: English
• Published: Lawrence The University of Chicago Press 01.12.2019; University of Chicago Press
• Subjects: Agricultural management; Calibration; Constraints; Context; Decisions; Evaluation; Integrity; Land use; Landscape; Macroinvertebrates; Management decisions; Predictions; Protection; Restoration; Rivers; Stream classification; Streams; Urban agriculture; Watersheds; Zoobenthos; United States > US; California; stakeholder group; streams; biotic integrity; modified channels; random forests; Bioassessment; landscape stressors; data visualization; urbanization; prioritization
• ISSN: 2161-9549; 2161-9565
• DOI: 10.1086/705996

Stream management goals for biological integrity may be difficult to achieve in developed landscapes where channel modification and other factors constrain in-stream conditions. To evaluate potential constraints on biological integrity, we developed a statewide landscape model for California that estimates ranges of likely scores for a macroinvertebrate-based index that are typical at a site with the observed level of landscape alteration. This context can support prioritization decisions for stream management, like identifying reaches for restoration or enhanced protection based on how observed scores relate to model predictions. Median scores were accurately predicted by the model for all sites in California with bioassessment data (Pearson correlation r = 0.75 between observed and predicted for calibration data, r = 0.72 for validation). The model also predicted that 15% of streams statewide are constrained for biological integrity within their present developed landscape, particularly for urban and agricultural areas in the South Coast, Central Valley, and Bay Area regions. We worked with a local stakeholder group from the San Gabriel River watershed (Los Angeles County, California) to evaluate how the statewide model could support local management decisions. To achieve this purpose, we created an interactive application, the Stream Classification and Priority Explorer, that compares observed scores with predictions from the landscape model to assign priorities. We observed model predictions consistent with the land-use gradient from the upper to lower watershed, where potential limits to achieving biological integrity were more common in the heavily-urbanized lower watershed. However, most of the sites in the lower watershed scored within their predicted ranges, and were therefore given a low priority for restoration. In contrast, 2 low-scoring sites in the undeveloped upper watershed were prioritized for causal assessment and possible future restoration, whereas 3 high-scoring sites were prioritized for protection. The availability of geospatial and bioassessment data at the national level suggests that these tools can easily be applied to inform management decisions at other locations where altered landscapes may limit biological integrity.