Search

Environmental scientists frequently rely on time series of explanatory variables to explain their impact on an important response variable. However, sometimes, researchers are less interested in raw observations of an explanatory variable than in derived indices induced by episodes embedded in its time series. Often these episodes are intermittent, occur within a specific limited memory, persist for varying durations, at varying levels of intensity, and overlap important periods with respect to the response variable. We develop a generic, parametrised, family of weighted indices extracted from an environmental signal called IMPIT indices. To facilitate their construction and calibration, we develop a user-friendly app in Shiny R referred to as IMPIT − a. We construct examples of IMPIT indices extracted from the Southern Oscillation Index and sea surface temperature signals. We illustrate their applications to two fished species in Queensland waters (i.e., snapper and saucer scallop) and wheat yield in New South Wales. • We developed indices from an environmental signal containing intermittent episodes. • The indices carry episode intensity, memory, persistence, intermittence and timing. • We illustrated indices' effectiveness and applications in fisheries and agriculture. • We developed IMPIT − a , an app to expedite index construction and calibration. • We expect IMPIT indices to be a useful addition to exploratory data analysis toolbox. [ABSTRACT FROM AUTHOR]