Evaluate and develop modeling (process-and empirical-based) approaches for management scenarios and alternative futures
This task will develop a framework to assess the interaction of watershed-based stressors with aquatic ecosystem responses. The goal is a developed approach, generalizable and transferable to a broad range of watershed settings and across a diversity of aquatic resources. A successful approach would enable the identification of significant watershed drivers that (individually and in concert) affect the integrity and sustainability of hydrologically-linked receiving water systems and allow us to reveal multi-metric stressor-response expectations for abiotic and biotic endpoints. We couple several ongoing, mature watershed-based case studies to leverage existing data/models and develop supplementary evaluations which can explore aspects of efficacy, scaling, multiple stressor issues, and the level of detail (e.g., processes, dynamics, spatial resolution) to include in model approaches for sustainable management practices.
Rationale and Research Approach:
The purpose of this task is to develop datasets, models, and decision support tools that can more accurately and efficiently inform water quality policy and management decisions at regional and national scales. Our research will focus on the effects of land use and climate on abiotic aquatic stressors (nutrients, sediments, temperature, etc.) and their impact on biotic endpoints (food webs, endangered and commercially important species, etc.) within streams, estuaries, and coastal margins. A major goal will be to identify sustainable land-water management practices. We propose to use several selected case studies, for which we presently have significant data and/or working models, to explore development and advancement of a more general framework. These case studies, in the Great lakes basin, coastal Caribbean, watersheds of the Pacific Northwest, and Chesapeake Bay offer a diversity of settings and thus are a place to start given limited resources, and we can take advantage of the investments already made in these efforts. Across these settings, we have a variety of watershed characteristics, different decisions asked of management, a range of stressors and aquatic response and a mosaic of aquatic resources being considered; in common is the purpose of deriving relationships between watersheds and the waters they influence. The array of case studies we will take advantage of are not a comprehensive set, nor a single facet embedded in a specific classification context, but given limited resources, we can use this set in evaluating the ability to capture and model the limits to predictability as a function of the spatial scale of resolution.
This problem is well known in ecological management; for example, at coarse levels, we know there are strong statistical relationships between nutrient loading and lake responses, but not each case is well predicted. At the finest scales, we may be able to develop data intensive, well-calibrated and predictive models, but we do not have the resources to do this in each individual case; for example, there are ~6000 small catchments draining to the US Great Lakes coast. The relatively simple, correlative representation in statistical models enables rapid, large-scale assessments with a minimal amount of forcing data and computational resources. However, insights obtained through correlative means cannot unambiguously link effects to stressors, and are bounded by the data available to develop the models. The latter constraint limits the usefulness of statistical models for assessing the effects of future land use and climate scenarios for which there are no historical analogs. While mechanistic watershed models can address these limitations by capturing important interactions among hydrological and biogeochemical processes, their predictive and explanatory power comes with greater data requirements and computational costs that limit their applicability to the regional scales that most interest policymakers. Therefore, there is a clear need for a hybrid approach that combines the strengths of finer-scale and coarser scale models (either mechanistic or statistical), while overcoming their respective limitations.We will develop a multi-model approach that considers a wide variety of model concepts-- mechanistic, statistical (correlative, multivariate, and conditional probability approaches), GIS- aided spatially-explicit formulations, and approaches that can determine response thresholds and tipping points--to more accurately and efficiently assess the effects of watershed features and watershed development on aquatic resources. The envisioned approach is stepwise, along two fronts. First, each case study is developing model concepts and these will individually be evaluated. Second, we will jointly consider how best to fashion demonstration applications of multiple model approaches; this was the focus of a mini-workshop in 2012 to develop combined tests of different models and scales of management interest. We currently envision a mechanistic ecohydrological model--VELMA (McKane et al. 2010, Abdelnour et al. 2011, Abdelnour et al. in press)--which will be applied to data-rich watersheds (for which we already have developed coarse level statistical relationships, such as the Great Lake tableau) and implemented at comparatively fine-spatial scales (headwater catchments to HUC 10 watersheds) to develop a process-level understanding of how watershed features control stressor sources within watersheds. The multi-model evaluation aims to provide scientifically-defensible approaches for formulating forward-looking watershed management decisions.
Larson, J.H., A.S. Trebitz, A.D. Steiman, M.J. Wiley, M. Carlson-Mazur, V. Pebbles, H. Braun, and P. Seelbach. 2013. Great Lakes rivermouth ecosystems: Scientific synthesis and management implications. Journal of Great Lakes Research 39:513-524.
Rowe, M.D., R.G. Kreis Jr, and D.M. Dolan. 2013. A reactive nitrogen budget for Lake Michigan. Journal of Great Lakes Research, online: http://www.sciencedirect.com/science/article/pii/S038013301300172X#.