Developing a systematic simulation-based approach for selecting indicators in strategic cumulative effects assessments with multiple environmental valued components

Indicator selection is a critical step in conducting effective strategic cumulative effects assessments. Selecting an appropriate set of indicators to represent multiple and sometimes disparate values is particularly challenging because the interpretation of impacts depends on indicator roles and relationships among indicators. However, systematic approaches for selection of indicators for strategic cumulative effects assessments (CEA) are unclear. For a 909,000 ha case study area involving 214 watersheds in coastal British Columbia, we defined a suite of twenty indicators linked to six Valued Components (VCs) that could be forecasted for forest, riparian and species at risk as three key values consistent with present land-use planning policies in British Columbia, Canada. We used spatio-temporal process-based models to project and integrate the stressor-response relationships between forest harvesting and run-of-river power resource management activities and the suite of selected indicators. For a likely development scenario, we assessed the correlative structure among projected indicator responses and, using a PCA-based analysis of outcomes, identified both patterns of potential redundancies and ecological processes linking indicators and dominant processes influencing VCs. Our results suggest that strategic CEAs will benefit if indicator selections are not made independently for each VC. Identifying the type of indicator, i.e., pressure or condition, and scale of its representation was important in determining if assessed impacts for individual indicators could be appropriately integrated to quantify overall impacts in the landscape. Consideration of indicator-indicator relationships both within and among VCs clarifies the influences of spatial scale, potential redundancies among indicators, and the role of underlying ecological processes in interpreting and aggregating indicator responses. Our case study demonstrates that relative scales of ecological processes, disturbances and management actions can limit how cohesive the interpretations of impacts may be across VCs in strategic CEA. Analysis of correlative structures among the twenty indicators suggested criteria-based statistical redundancies occurred between only two indicator pairs, however PCA suggested three ecological processes (road disturbance, Spotted Owl habitat state, retention and recruitment of old forest) were operating to relate behaviors of multiple indicators. Careful consideration of the interacting roles of ecological processes as they relate to values is required when determining appropriate indicators and designing how best to aggregate indicator results into an effective strategic CEA. A three step systematic and generalizable approach to forecasting present and future states of indicators will improve efficiencies and effectiveness of strategic CEA.