quantile_doc: quantile

In Climandes/ClimIndVis: Calculation, verification and visualisation of climate indices

Description

quantile

Arguments

baseperiod	Vector of start and end year for the calculation of quantiles. If not provided, whole range of years in dataset will be used for the calculation of quantiles.
q	Vector of quantiles, default to seq(0,1,by=0.01)
th_object	Climindvis_index object, default=NULL. if calculating a quantile index for forecast data, a climindvis_index object of hindcasts data for the same spatial dimensions and forecast months has to be provided. Threshold values will then be taken from the hindcast object (object$index_info$quantiles). In this case all other arguments (q, baseperiod, threhold,..) are ignored and taken from the climindvis_index object. In case of autoplot functions, these values are handed over automatically and th_object should be left at default value of NULL..
q_threshold	Percentile threshold (numerical). Value between 0 and 100.
percentile	Percentile threshold (numerical). Value between 0 and 100.
percentile1	Lower percentile threshold (numerical). Value between 0 and 100.
percentile2	Upper percentile threshold (numerical). Value between 0 and 100.
n	Window size (in days) for running window in temperature quantile calculation. Default =5. See details for further information.
inbase	For temperature quantiles, calculate quantiles inside baseperiod and outside baseperiod follwoing the boostrapping method of Zhang 2005. Only applicable, when baseperiod is selected. Default = TRUE. For hindcast data, inbase is set to FALSE because boostrapping of ensemble and years is computational intensive.
NAmaxQ	Minimim number of days needed for calculation of quantiles. Default = 365 [days]. For monthly aggregation of precipitation, consider lowering the threshold or expanding the baseperiod.
min_base_fraction	Minimum fraction of base data that must be present for quantile to be calculated for a particular day (only applies to temperature quantiles). Default = 0.1.
qens_all=TRUE	logical. Should quantiles of ensembles be calculated over all ensemble members (TRUE) or for each member individually (FALSE)? Default = TRUE

Details

Quantiles are calculated identically to the calculations performed in the R-package climdex.pcic. For precipitation, quantiles are calculated with the function quantile(..., type = 8) from R stats using all values inside the selceted aggregation.
For temperature, quantiles are calculated for every day of the year independently based on a moving window with default 5 days. Following the suggestiosn of Zhang (2005) it is possible to calculate a bootstrapping for quantile thresholds within the baseperiod by setting inbase=TRUE. If inbase is set to FALSE, the same thresholds are used inside and outside the reference period.
For ensemble data (hindcasts), it is not possible to calculate the boostrapping inside the reference period. Consider therefore using the full period as reference period. Please note, that the quantile calculation of ensembles stil may be subject to errors. In case you note inconsistencies, please report these to the package authors. For a detailed description of this method, see: Zhang, X, Hegerl, G, Zwiers, F.W and Kenyon, J. (2005). Avoiding inhomogeneity in percentile-base indices of temperature extremes. Journal of Climate, 18.

Climandes/ClimIndVis documentation built on Oct. 24, 2021, 10:52 a.m.