The Many Hazards of Trend Evaluation

Geoffrey M Henebry, Michigan State University, Department of Geography, Environment, and Spatial Sciences & Center for Global Change and Earth Observations, East Lansing, MI, United States, Kirsten de Beurs, University of Oklahoma, Geography and Environmental Sustainability, Norman, OK, United States, Xiaoyang Zhang, South Dakota State University, Geospatial Sciences Center of Excellence (GSCE), Department of Geography & Geospatial Sciences, Brookings, United States, John S Kimball, University of Montana, Numerical Terradynamic Simulation Group, Missoula, United States and Chris Small, Columbia University of New York, Lamont Doherty Earth Observatory, Palisades, NY, United States
Abstract:
Given the awareness in the scientific community of global scale drivers such as population growth, globalization, and climatic variation and change, many studies seek to identify temporal patterns in data that may be plausibly related to one or more aspect of global change. Here we explore two questions: “What constitutes a trend in a time series?” and “How can a trend be misinterpreted?” There are manifold hazards—both methodological and psychological—in detecting a trend, quantifying its magnitude, assessing its significance, identifying probable causes, and evaluating the implications of the trend. These hazards can combine to elevate the risk of misinterpreting the trend. In contrast, evaluation of multiple trends within a biogeophysical framework can attenuate the risk of misinterpretation. We review and illustrate these hazards and demonstrate the efficacy of an approach using multiple indicators detecting significant trends (MIDST) applied to time series of remote sensing data products.
Back to: Assessing the State of the Art in Change Detection and Analysis for Geophysical Time Series II