Estimate Variance for Annual Maximum Series Data — data_mw

Generates a time series of standard deviations using a moving window algorithm, which can be used to explore potential evidence of nonstationarity in the variability of a dataset. It returns a list that pairs each window’s mean year with its window standard deviation. The hyperparameters size and step control the behaviour of the moving window. Following the simulation findings from Vidrio-Sahagún and He (2022), the default window size and step are set to 10 and 5 years respectively. However, these can be changed by the user.

Usage

data_mw_variability(data, years, size = 10L, step = 5L)

Arguments

data: Numeric vector of observed annual maximum series values. Must be strictly positive, finite, and not missing.
years: Numeric vector of observation years corresponding to data. Must be the same length as data and strictly increasing.
size: Integer scalar. The number of years in each moving window. Must be a positive number less than or equal to length(data) (default is 10).
step: Integer scalar. The offset (in years) between successive moving windows. Must be a positive number (default is 5).

Value

A list with two entries:

years: Numeric vector containing the mean year within each window.
std: Numeric vector of standard deviations within each window.

References

Vidrio-Sahagún, C. T., and He, J. (2022). The decomposition-based nonstationary flood frequency analysis. Journal of Hydrology, 612 (September 2022), 128186. doi:10.1016/j.jhydrol.2022.128186

Examples

data <- rnorm(n = 100, mean = 100, sd = 10)
years <- seq(from = 1901, to = 2000)
data_mw_variability(data, years)
#> $std
#>  [1]  7.979322 11.091337 10.834792  7.833379  6.918857  8.207443  9.093049
#>  [8] 11.037079 10.865376  8.024952  8.950729 10.082149  9.835107  8.926596
#> [15] 13.342132 15.260867 11.322819  8.959743  7.579580
#> 
#> $year
#>  [1] 1905.5 1910.5 1915.5 1920.5 1925.5 1930.5 1935.5 1940.5 1945.5 1950.5
#> [11] 1955.5 1960.5 1965.5 1970.5 1975.5 1980.5 1985.5 1990.5 1995.5
#>