akq_summary: Summary Table from the Asquith-Knight Discharge Decay...

In wasquith-usgs/akqdecay: Many Distributional Statistics of USGS Daily Streamflow Hydrographs and Powerful Utilities

Description

Extract the summary table from Asquith–Knight discharge decay analyses returned by akqdecay for a sequence of USGS streamgages contained within the R environment that has been previously populated by fill_akqenv.

Usage

akq_summary(akqenv, silent=FALSE, ...)

Arguments

akqenv	The R environment populated by fill_akqenv. This argument can also be an R list directly from akqdecay; special accommodation is made;
silent	Suppress informative calls to message(); and
...	Additional arguments to pass (see source code to ascertain flexible usage).

Value

An R data.frame containing the summary for each streamgage.

site	The streamgage identification number;
beg_year	First year in period of record data retrieval;
end_year	Last year in period of record data retrieval;
yr_range_str	A neat string representation that might be useful in formal tables for publication;
total_count	The total number of daily values;
count	The count of daily values processed by the settings of akqdecay;
kendall_tau	Kendall's Tau between the days per log-cycle changes and the streamflow from the flow-duration curve;
spearman_rho	Spearman's Rho between the days per log-cycle changes and the streamflow from the flow-duration curve;
median	The median (Ψ_\mathrm{med}) of the processed daily values;
L1L2	The mean plus square-root pi L-scale: λ_1 + λ_2√{π} (see Note);
gfactor	The G_f-factor from a fitted distribution for the probability f used by akqdecay; and
gfactor_emp	The G_f-factor from the empirical distribution for probability f through a call to the standard (default) built-in R quantile() function.

Note

As yet, Asquith and Knight have definitive opinion on what the optimal “Gfactor” is in regards to depth into the distribution tail of days per log-cycle change on the recession limb of hydrographs. Please contact the authors for further information.

The L1L2 as a term or concept is expected to have little direct meaning to most users. The first L-moment (arithmetic mean) is λ_1. The second L-moment (L-scale) is λ_2 and is directly interpretable as, but not numerically equal to, the well-known standard deviation (σ). Thus, it has units of days per log-cycle and matches those of the mean. The σ is the second parameter of the normal distribution, and this parameter in terms of L-scale is σ = λ_2√{π}. Thus, this is an estimate of the standard deviation via L-moments. The percentile pnorm(sqrt(pi), sd=sqrt(pi)) is the 84 percentile of the normal for λ_1 = 0—hence, L1L2 can loosely be thought of as a parametric estimate of a G_{\mathrm{nor}}(F=0.84) Gfactor. This would be a Gfactor defined as about “one standard deviation into the right tail of the distribution.”

Author(s)

W.H. Asquith

See Also

akqdecay, akq_lmom, akq_counts, akq_summary

Examples

sites <- c("05403500", "05405000") # Two USGS streamgages in Wisconsin
WisExample <- new.env(); fill_dvenv( sites,              envir=WisExample,
                                     sdate="1945-01-01", edate="2016-12-31")
WisAKQ <- new.env(); fill_akqenv(dvenv=WisExample, envir=WisAKQ)
akq_summary(WisAKQ) # These line-wrapped values will change as record increases.
#     site beg_year end_year yr_range_str total_count count kendall_tau spearman_rho
# 05403500     1945     2016   1945--2016       18444  9242 -0.20468258   -0.3083903
# 05405000     1945     2016   1945--2016       26297 13724 -0.08380238   -0.1295017
#   median     L1L2  gfactor gfactor_emp
# 37.98102 125.1716 136.3617    139.3032
# 34.95030 155.0216 161.0996    176.1453 #

wasquith-usgs/akqdecay documentation built on Nov. 9, 2020, 1:13 p.m.