CC or super-CC? A critical appraisal of temperature-precipitation scaling rates at sub-hourly timescales.

Precipitation intensities have been shown to increase with temperature, according to scaling rates varying between approximately 7 and 14% (CC and super-CC scaling). While several studies have investigated the geographical differences in scaling rates across the world and their link to physical processes and precipitation climatology, there seems to be little agreement on how exactly scaling rates are to be calculated and how they ought to be interpreted. Studies have used different temperatures (dewpoint versus surface air,) and looked at different precipitation time scales (daily mean versus instantaneous etc). Methodological approaches also vary in how scaling rates are statistically inferred from observations, with different choices for binning and calculating wet or dry day percentiles.

Here, we investigate how methodological choices influence scaling rates retrieved from temperature and precipitation observations based on a large dataset from 353 weather stations in Germany, covering a period from 10 to 25 years. We show that scaling rates are strongly dependent on methodological choices and can vary by a factor of two depending on how observations are selected and statistically treated. This has important implications for interpretation of scaling rates in the context of global warming, and expected changes in precipitation. Single scaling rates have limited value when interpreted in isolation, given the strong sensitivity to data selection and variability in scaling rates with precipitation time scales. Instead, scaling rates should be viewed in the context of precipitation events and how the combination of their various properties change with temperature.