Use of structured expert judgment to forecast invasions by bighead and silver carp in Lake Erie

Abstract

Identifying which nonindigenous species will become invasive and forecasting the damage they will cause is difficult and presents a significant problem for natural resource management. Often, the data or resources necessary for ecological risk assessment are incomplete or absent, leaving environmental decision makers ill equipped to effectively manage valuable natural resources. Structured expert judgment (SEJ) is Resumen: Identificar cuáles especies no-nativas se volverán invasoras y predecir el daño que causarán es complicado y presenta un problema significativo para el manejo de recursos naturales. Con frecuencia los datos o recursos necesarios para la evaluación de riesgo ecológico están incompletos o son inexistentes, lo que deja mal equipados a quienes toman las decisiones ambientales para manejar efectivamente recursos naturales valiosos. El juicio experto estructurado (JEE) es un método con bases matemáticas y de desempeño para obtener, sopesar y agregar juicios expertos. En contraste con otros métodos de obtención y agregación de juicios expertos (donde, por ejemplo, se le pueden asignar pesos iguales a los expertos), JEE sopesa a cada experto con base en su asertividad estadística y capacidad de informar por medio de la medida de desempeñoPaper submitted November 12, 2013; revised manuscript accepted May 20, 2014. This is an open access article under the terms of the Creative Commons Attribution

Suburban watershed nitrogen retention: Estimating the effectiveness of stormwater management structures

“…Here we use SEJ to estimate the uncertainty in stormwater BMP performance under a range of environmental conditions. To date, SEJ has not been applied widely in ecological studies (but see Bamber and Aspinall, 2013;Rothlisberger et al, 2010Rothlisberger et al, , 2012Teck et al, 2010;Wittmann et al, 2014Wittmann et al, , 2015. Nonetheless, over the last 20 years, the approach has been used to successfully estimate uncertainty in many disciplines including nuclear safety, volcanic forecasting, and human health (Cooke and Goossens, 2008); and it is well established within the discipline of risk assessment.…”

“…This type of SEJ aggregates the estimates for all experts according to their performance on a subset of calibration variables, for which true nitrogen loads are known to the elicitors, but unknown to the experts. For each question posed in the interview (Appendix S3), the resulting performance-weighted uncertainty estimate reflects the optimal balance of accuracy and information for the entire group of experts (Cooke, 1991;Wittmann et al, 2015).…”

“…Calibration measures the statistical likelihood that the set of known values for the calibration variables correspond with an expert's assessment of those variables. Thus, the calibration score is calculated as the probability (p-value) of falsely rejecting the hypothesis that the expert's uncertainty estimates were accurate (Cooke, 1991;Wittmann et al, 2015). Calibration scores take values between 0 and 1 and signify the degree to which an expert's uncertainty estimates (5 th , 50 th , 95 th percentiles) are accurate relative to the calibration variables.…”

“…SEJ has been used widely in the field of decision science to inform management decisions where empirical data are lacking or are impossible to collect (e.g., Bamber and Aspinall, 2013;Cooke and Goossens, 2004;Rothlisberger et al, 2012;Wittmann et al, 2014Wittmann et al, , 2015.…”

Conceptual Models for Marine and Freshwater Systems in Alaska: Flexible Tools for Research Planning, Prioritization and Communication + Supplementary Appendices (See Article Tools)

“…Depending on the objectives of the exercise, participants may have similar expertise (e.g., salmon ecology), or the process may be designed to have a balance of expertise when working with ecosystemlevel models (e.g., models of intertidal ecology or physical oceanography). More formally, experts could be asked to rate their ability to contribute to parts of a conceptual model, and consensus models could summarize this information throughout different components of the final model (e.g., McDaniels et al, 2010), for example, by presenting linkage rating values weighted by experts' self-rating scores (or ratings based upon a set of calibration questions; cf., Wittmann et al, 2015).…”