Data for: A fiery past: a comparison of glacial and contemporary fire regimes on the Palaeo-Agulhas Plain, Cape Floristic Region
We calculated daily FDI from daily measurements recorded at 14h00 (local time) of temperature, relative humidity, wind speed, and rainfall (of the past 24 hours) (cf. Kraaij et al., 2013b). Conditions at this time of day approximate daily maximum temperature, minimum relative humidity, wind at its strongest, and thus maximum daily FDI. This index also incorporates a drought factor and the Keetch Byram Drought Index (Keetch and Byram, 1968; Noble et al., 1980). For each of the eight locations we calculated daily FDI spanning 30 years for three types of series: (i) ‘Contemporary actual’, i.e. weather data actually observed during the past four decades (the exact dates of the series depended on the availability and completeness of weather records for the different locations); (ii) ‘Contemporary model’, i.e. data modelled for contemporary conditions (1981-2010) by the AMIPw simulation (Engelbrecht et al., 2019) but bias-corrected as described by Cowling et al. (this issue); and (iii) ‘LGM models’, i.e. data modelled for conditions during the LGM, obtained from the eight dynamically downscaled climate simulations (CCSIM4a, CNRM, FGOALS, GISS, IPSL, MIROC, MPI, and MRI) of Engelbrecht et al. (2019) that were bias-corrected for rainfall and temperature as described by Cowling et al. (this issue). For each location and series type we then calculated mean monthly FDI and the mean number of days per month with at least moderate (FDI > 5) and at least high (FDI > 12) fire weather (respectively) to assess the seasonality and severity of fire danger weather and how this differed among locations and periods. We also computed mean monthly values for each of the input variables, i.e. maximum temperature, minimum relative humidity, wind speed at mid-day, and daily rainfall, to aid our interpretation of the FDI results.