We propose a novel approach to study heat waves that focuses on characterizing key components of their definition: threshold temperature and persistence. Ambiguity about these criteria often arises from the lack of a specific impact of interest and results in questionable assumptions.
To present our idea we focus on summertime heat waves over the La Plata Basin (LPB) in South America and employ a high quality dataset of daily maximum temperature recently become available for the period 1961 to 2000. We start adopting the 90th percentile of the maximum daily distribution as threshold temperature and a minimum persistence of two days to classify an event as a heat wave. The latter is chosen because 2 to 3 days is the typical persistence of these temperatures and longer persistence events are rare.
We show that two events occur on average every year; most years have 0 to 4 events, and only 15% of the years have at least 5. We identified only three hot summers, which are defined as those in which a large number of events occurred across much of the region; they are: 1967/68, 1971/72, and 1988/89. Their limited number prevents any possible search for statistically significant large scale conditions associated with their occurrence.
The sensitivity of outcomes to the choice of threshold temperature is minimal: when the 80th percentile is employed, heat waves are more frequent but persist as long as those characterizing the 90th percentile, and only one additional hot summer is found: 1985/86. Results indicate the lack of a climatological circulation that sets the stage for heat waves in LPB; alternatively, this is too rare to be appreciated over a 40 year period.