Cristian SUTEANU, Saint Mary's University, Canada
Many natural processes are strongly variable both in space and in time. Observing them on different time scales can reveal distinct aspects of the studied phenomena (Lovejoy and Schertzer 2013). While increasingly powerful methods have been developed for the characterization of irregular patterns, their results may depend on the time scale ranges applied in the analysis. Moreover, the sampling design and its impact on time scale often vary from one study to another, limiting the relevance of comparisons regarding different experimental investigations. This paper focuses on the implications of time scales and time scale changes on our understanding of strongly variable processes.
We introduce the “guided cut”, a selection process that is operating in successive stages on data acquired from the environment. The guided cut, set representation, and mapping are shown to lie at the core of procedures involving scale in general, and temporal scale in particular. We show how three key types of time scale - scale as size, scale as ratio, and scale as rank - can be used in order to identify pattern-specific properties, contributing to a clearer and more comprehensive picture of time-scale-dependent aspects of natural processes.
This approach is designed to support a rigorous comparison and a better understanding of strongly variable processes occurring at different times, in different locations. Such an approach is particularly useful when the studied patterns are subject to significant change over time (Carvalho et al. 2017, Karanauskas et al. 2018). The resulting unifying framework is illustrated with practical applications regarding wind speed variability in relation to renewable energy and natural hazards.
Mots clés : time scale|variability|wind speed|renewable energy|natural hazards
A102883CS