This README file provides original source context to the
MGBT package and other package design notes. These are
partial notes and the man/MGBT-package.Rd is contains much
more information. The ./inst/ subdirectory is standard for
R packages. The CITATION file is standard for the
R language. The documentation below provides a manifest of
sorts for the contents of this directory.
./MGBT/inst/FILE: AsquithLOT(1995).pdf — A hard to
find reference concerning low-outlier thresholds in Texas by the
author.
FILE: tim-low-outliers-2010-notes.pdf —
John F. England (JFE) visited Reston, Virginia (USGS Headquarters) in
Spring 2017 after the passing of Timothy A. Cohn (TAC). He scanned
Cohn’s personal note books. One is titled “Tim Cohn Low Outliers” and
contains 33 pages of mathematical notes. TAC, within
COHN_MGBT_LowOutliers(R).txt, describes “Orthogonal
evaluation of p-value” using “mathematical notes dated December 16,
2009” as spanning 39 pages and that these are “not for the faint of
heart ;-)” This note is not found in the file
./sources/P3_089(R).txt. It does not appear that the 33
pages scanned are the 39 pages being referenced.
FILE: USGSapproval20190924.pdf — This
is a copy of the approval email from USGS internal publication tracking
authorizing dissemination (release) of the product.
./MGBT/inst/helpers/Convenient research scripts related to experimental efforts by William H. Asquith (WHA). These will not necessarily do anything for end users.
./MGBT/inst/legend/This directory contains reference information and portable document
format (PDF) files related to explanation of the
MGBT::plotPeaks() function.
./MGBT/inst/sources/This directory holds original source code acquired from TAC at
various times by colleagues. It is advised to see also the documentation
./MGBT/man/MGBT-package.Rd.
./sources/FigureMacros_jfe(R).txtSource code exchanged between TAC and JFE focused on low-outlier thresholds (potentially influential low floods).
./sources/LowOutliers_jfe(R).txtSource code exchanged between TAC and JFE focused on low-outlier thresholds (potentially influential low floods).
./sources/LowOutliers_wha(R).txtOn or about August 21, 2013, TAC sent a file in an email attachment
titled COHN_MGBT_LowOutliers.R to WHA. At the time, WHA was
needing an MGBT implementation for a study of “modern periods of
statitic flood flow regulation” through flood frequency computations on
annual peak streamflow data for USACE Fort Worth. Asquith was using the
method of L-moments, but wanted the MGBT as protection against some
mixed population effects. This communication to WHA appears to be 100
percent independent of from TAC’s personal R code base related
to the Expected Moments Algorithm (EMA) and Bulletin 17C (log-Pearson
type III distribution) communicated with JFE in the years prior. This
file has been retitled to LowOutliers_wha(R).txt.
The source files have a *(R).txt in their name. This is
deliberate. They are R scripts, but R will complain on
an R CMD check MGBT when any .R file exits in
the ./inst/ subdirectory of a package. Cohn has many
functions related to low-outliers in his code bases, but many are not
directly related to MGBT for purposes of implementation. All of these
are ported into the MGBT package and documented. It is clear that Cohn
shared functions with Asquith in ./LowOutliers_wha(R).txt
that are not available in the file ./sources/P3_089(R).txt.
The typical design of R packages by WHA to break as much code
into separate .R files as possibly (logically) and
generally have about a one-to-one pairing of .Rd files.
./sources/P3_075_jfe(R).txtSource code exchanged between TAC and JFE at some point in time.
./sources/P3_085_wha(R).txtSource code exchanged between TAC and WHA in November 2015.
./sources/P3_089(R).txtOn May 16, 2017, Julie E. Kiang sent was she believes to be TAC’s
master .R file of the EMA-17C code. Within which separate
“MGBT” functions from the WHA communication appear present. The short
comments by TAC indicate changes as late as June 28, 2016. These are
left intact. The MGBT package will first pass into
maturity as CRAN-compliant through communication between TAC and WHA and
then port evidently TAC’s later developments. Asquith suggests appending
“2013” to the original functions (e.g. MGBT.2013) given to Asquith and
eventually use MGBT() as the canonical version as embedded
into 17C.
./MGBT/www/This directory is a holding ground for files to enrich the Internet presentation of the package.
Nancy A. Barth has encountered a situation where USGS PeakFQ software and the MGBT package at first glance appear divergent from each other in regards to the threshold defined by the MGBT. For now (summer 2019), this paragraph just lists USGS streamflow-gaging station number(s) of such sites; further study is needed.