Implementation of a watershed modelling framework to support adaptive management in the Canadian side of the Lake Erie basin

Abstract

The present study aims to provide guidelines with respect to the implementation of a watershed modelling framework that can support adaptive management in the Canadian side of the Lake Erie basin. Our primary objective is to highlight the establishment of a multimodel ensemble as an appealing strategy that can address a wide range of conceptual and operational uncertainties typically underlying any modelling exercise. We identify three ensemble strategies that have different implications for the granularity of the analysis, empirical knowledge and data input demands to constrain the individual models, and required timeframe to operationalize them. The first option partly resembles the watershed modelling work in the Maumee River watershed, where the characterization of the watershed attributes and functioning that modulate nutrient loading will be based on multiple independent SWAT applications with different process characterizations. Before implementing SWAT, we propose the use of SPARROW —a data-driven model— to delineate hot-spots, examine landscape predictors for nutrient mobilization and retention, and use these results to inform the SWAT models. Alongside the SPARROW and SWAT applications, the second option considers two additional models (HYPE, HSPF) that are conceptually sound, have been applied in the area, and can increase the structural and functional diversity of the ensemble of catchment models by offering alternative representations of fundamental processes/functions associated with the water cycle, soil erosion, biogeochemical cycles, and soil-plant interactions. The third option introduces local (farm, urban or natural vegetation) and edge-of-field monitoring and modelling (EPIC or APEX tools) to refine the spatial resolution and design in-field conservation practices that mitigate nonpoint source pollution from agriculture and other sources. Our study also pinpoints data monitoring needs to elucidate critical unknowns of the watershed functioning, such as the role of legacy phosphorus (P), the causes and consequences of the increasing long-term trends in dissolved reactive P loading, the challenges in reproducing spring-freshet or event-flow conditions, and the dynamic characterization of water/nutrient cycles under the non-stationarity of a changing climate. The strong linkages between environmental processes and management decisions that occur along the watershed-lake continuum require the presence of a holistic model ensemble to provide rigorous BMP efficiency assessment and simulation of the human-environment system responses to current and future watershed nutrient loads, land uses, climatic forcing, and socioeconomic conditions. The proposed model ensemble framework can be used as an integrative tool for ecosystem policy analysis and design of proper management action plans in Lake Erie and elsewhere.