Version 4.0.2

v4.0.2

Patch for percentiles-based indices when studied period is out of bas…

Version 4.0.1: Corrected computing percentiles for R75p, R95p, R99p, R75pTOT, R95pTOT

Corrected computing percentiles for R75p, R95p, R99p, R75pTOT, R95pTOT:
computing of precipitation percentile-based indices requires only one percentile value and not percentile values for each calendar day as for temperature percentile-based indices.

Version 4.0.0

• Implemented bootstrapping procedure (See "Removing the “jump”" section in this article: http://journals.ametsoc.org/doi/pdf/10.1175/JCLI3366.1).
• Percentile threshold computing is now directly inside icclim.indice() function:
  • added new parameters:
    • ignore_Feb29th (allow to ignore February 29th)
    • base_period_time_range for base period time range
    • window_width
    • only_leap_years
    • interpolation (2 interpolation methods: linear and interpolation proposed by `Hyndman and Fan (1996) (http://amherst.edu/media/view/129116/original/Sample+Quantiles.pdf)
    • save_percentiles_to_file
    • out_unit ('days' or '%' for indices: TG10p, TG90p, TX10p, TX90p, TN10p, TN90p)
• Possibility of setting user defined seasons.
• ETCCDI climate indices (http://etccdi.pacificclimate.org/list_27_indices.shtml) is used now as indices definition:
  • Indices R95pTOT and R99pTOT changed their definition (removed the division in the formula).
  • Indice RR changed its definition (before: sum of precipitation; now: sum of precipitation only of wet days) and renamed to PRCPTOT.