| Title: | Daily Streamflow Trend and Change Point Screening |
|---|---|
| Description: | Screens daily streamflow time series for temporal trends and change-points. This package has been primarily developed for assessing the quality of daily streamflow time series. It also contains tools for plotting and calculating many different streamflow metrics. The package can be used to produce summary screening plots showing change-points and significant temporal trends for high flow, low flow, and/or baseflow statistics, or it can be used to perform more detailed hydrological time series analyses. The package was designed for screening daily streamflow time series from Water Survey Canada and the United States Geological Survey but will also work with streamflow time series from many other agencies. |
| Authors: | Jennifer Dierauer [aut, cre], Paul Whitfield [aut] |
| Maintainer: | Jennifer Dierauer <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 1.2.6 |
| Built: | 2025-02-24 03:32:01 UTC |
| Source: | https://github.com/cran/FlowScreen |
Create custom axis starting on hyrologic year start month
axis_doy.internal(hyrstart = 10)axis_doy.internal(hyrstart = 10)
hyrstart |
numeric indicating month for start of the hydrologic year (water year). |
Paul Whitfield
This function estimates baseflow
bf_boughton(discharge, k, C)bf_boughton(discharge, k, C)
discharge |
Nnumeric vector of daily flow data |
k |
Numeric value of the recession constant (dimensionless). |
C |
Numeric value of the partitioning factor (dimensionless). |
Returns a numeric vector of the estimated baseflow.
Paul H. Whitfield
Boughton, WC. 1993. A hydrograph-based model for estimating the water yield of ungauged catchments.In Hydrology and Water Resources Symposium, Institution of Engineers Australia, Newcastle, NSW; 317-324.
data(cania.sub.ts) res <- bf_boughton(cania.sub.ts$Flow, k=0.9, C=0.1) plot(cania.sub.ts$Date, cania.sub.ts$Flow, xlab="", ylab="Q (m3/s)", type="l") points(cania.sub.ts$Date, res, type="l", col="blue")data(cania.sub.ts) res <- bf_boughton(cania.sub.ts$Flow, k=0.9, C=0.1) plot(cania.sub.ts$Date, cania.sub.ts$Flow, xlab="", ylab="Q (m3/s)", type="l") points(cania.sub.ts$Date, res, type="l", col="blue")
This function takes vector of discharge data and estimates the baseflow
bf_eckhardt(discharge, a, BFI)bf_eckhardt(discharge, a, BFI)
discharge |
vector of daily discharge observations |
a |
Numeric value. |
BFI |
Numeric value. |
Returns
Paul Whitfield
Eckhardt, K. 2012. Technical note: Analytical sensitivity analysis of two parameter recursive digital baseflow separation filter. Hydrology and Earth System Sciences 16: 451-455.
data(cania.sub.ts) bf <- bf_eckhardt(cania.sub.ts$Flow, 0.97, 0.8) plot(cania.sub.ts$Date, cania.sub.ts$Flow, type="l") points(cania.sub.ts$Date, bf, type="l", col="blue")data(cania.sub.ts) bf <- bf_eckhardt(cania.sub.ts$Flow, 0.97, 0.8) plot(cania.sub.ts$Date, cania.sub.ts$Flow, type="l") points(cania.sub.ts$Date, bf, type="l", col="blue")
This function estimates baseflow.
bf_oneparam(discharge, k)bf_oneparam(discharge, k)
discharge |
Numeric vector of daily flow data |
k |
Numeric value for the recession constant (dimensionless). |
Returns a numeric vector of the estimated baseflow.
Paul H. Whitfield
Eckhardt, K. 2005. How to construct recursive digital filters for baseflow separation methods. Journal of Hydrology 352: 168-173.
data(cania.sub.ts) res <- bf_oneparam(cania.sub.ts$Flow, k=0.9) plot(cania.sub.ts$Date, cania.sub.ts$Flow, xlab="", ylab="Q (m3/s)", type="l") points(cania.sub.ts$Date, res, type="l", col="blue")data(cania.sub.ts) res <- bf_oneparam(cania.sub.ts$Flow, k=0.9) plot(cania.sub.ts$Date, cania.sub.ts$Flow, xlab="", ylab="Q (m3/s)", type="l") points(cania.sub.ts$Date, res, type="l", col="blue")
This function estimates the percentage of baseflow in a given period relative to the total annual baseflow.
bf.seas(TS, seas = c(6:8))bf.seas(TS, seas = c(6:8))
TS |
output from |
seas |
Integers representing months of the year. Default is c(6:8), i.e. June-August. |
This function calls bf_eckhardt to complete the
baseflow separation.
Returns a vector containing the calculated percentage for each year in the input time series. The "times" attribute provides the corresponding year for each calculated value.
Jennifer Dierauer
See bf.stats to calculate additional baseflow metrics.
data(cania.sub.ts) res <- bf.seas(cania.sub.ts) res2 <- screen.metric(res, "Percent Annual Baseflow in Jun-Aug")data(cania.sub.ts) res <- bf.seas(cania.sub.ts) res2 <- screen.metric(res, "Percent Annual Baseflow in Jun-Aug")
This function estimates the baseflow and calculates the mean, max, and min baseflow and baseflow index for a user defined time period.
bf.stats(TS, by = "hyear")bf.stats(TS, by = "hyear")
TS |
output from |
by |
summary period. Options are "year", "hyear", "month", or "doy". Default is "hyear". |
This function calls bf_eckhardt to complete the
baseflow separation.
Returns a data.frame with the following columns:
By - Unique values representing the summary periods, e.g. a list of unique years, months, or days of year
MeanQ - Mean daily streamflow for the summary period, in m3/s
MeanBF - Mean daily baseflow for the summary period, in m3/s
MaxBF - Maximum daily baseflow for the summary period, in m3/s
MinBF - Minimum daily baseflow for the summary period, in m3/s
BFVol - Baseflow volume for the summary period, in km3
MeanBFI - Mean daily baseflow index for the summary period, dimensionless
MaxBFI - Maximum daily baseflow index for the summary period, dimensionless
MinBFI - Minimum daily baseflow index for the summary period, dimensionless
Jennifer Dierauer
data(cania.sub.ts) res <- bf.stats(cania.sub.ts) res2 <- screen.metric(res[,2], "m3/s")data(cania.sub.ts) res <- bf.stats(cania.sub.ts) res2 <- screen.metric(res[,2], "m3/s")
This data set includes a subset of the mean daily streamflow for the Caniapiscau Rivers. It includes observations from 1970-1995 (hydrologic years). The code used to subset and modify the original data is shown below.
data(caniapiscau)data(caniapiscau)
Formatted as a data.frame with the following columns:
ID - Water Survey Canada Station ID
Date - Date of observation, formatted as YYYY-mm-dd
Flow - Mean daily streamflow, measured in m3/s
Code - Data Quality Code
Agency - Source Agency (Water Survey Canada)
Year - Calendar year
month - Calendar month
doy - Calendar day of year
hyear - Hydrologic year
hmonth - Hydrologic month
hdoy - Hydrologic day of year
Environment Canada. 2010. EC Data Explorer V1.2.30.
Water Survey of Canada V1.2.30 https://www.ec.gc.ca/rhc-wsc/
# Code used to subset and modify original Caniapiscau series: ## Not run: data(caniapiscau) cania.ts <- create.ts(caniapiscau, hyrstart=3) cania.sub.ts <- subset(cania.ts, cania.ts$hyear %in% c(1970:1995)) ## End(Not run) # example use of example subset flow series data(cania.sub.ts) head(cania.sub.ts) str(cania.sub.ts)# Code used to subset and modify original Caniapiscau series: ## Not run: data(caniapiscau) cania.ts <- create.ts(caniapiscau, hyrstart=3) cania.sub.ts <- subset(cania.ts, cania.ts$hyear %in% c(1970:1995)) ## End(Not run) # example use of example subset flow series data(cania.sub.ts) head(cania.sub.ts) str(cania.sub.ts)
This data set includes the mean daily streamflow for the Caniapiscau
River. The file has been read from the original .csv format using
read.flows. The Caniapiscau River is located in Nunavik, Quebec,
Canada, and flows northward. The headwaters (representing 45 percent of the
total flow) were dammed to create the Caniapiscau Reservoir, which started
filling in 1981. In 1985, the reservoir was diverted to the west into the
La Grande hydroelectric complex.
This flow time series is used as an example of a river with a known
change point to demonstrate the package's screening capabilities.
data(caniapiscau)data(caniapiscau)
Formatted as a data.frame with the following columns:
ID - Water Survey Canada Station ID
PARAM - Parameter ID (1 indicates flow)
Date - Date of observation, formatted as YYYY-mm-dd
Flow - Mean daily streamflow, measured in m3/s
Agency - Source Agency (Water Survey Canada)
Environment Canada. 2010. EC Data Explorer V1.2.30.
Water Survey of Canada V1.2.30 https://www.ec.gc.ca/rhc-wsc/
data(caniapiscau) head(caniapiscau) str(caniapiscau)data(caniapiscau) head(caniapiscau) str(caniapiscau)
Contains the results from metrics.all for the full
Caniapiscau River daily flow series. Data set created as indicated below. This
data set is used in the example documentation for the screen.frames,
screen.summary, and screen.cpts functions in order to
reduce example run times.
data(caniapiscau)data(caniapiscau)
Formatted as indicated in the documentation for metrics.all
Original flow series from Environment Canada. 2010. EC Data Explorer V1.2.30.
Water Survey of Canada V1.2.30 https://www.ec.gc.ca/rhc-wsc/
# Code used produce this data set: ## Not run: data(caniapiscau) caniapiscau.ts <- create.ts(caniapiscau, hyrstart=3) caniapiscau.ts <- subset(caniapiscau.ts, caniapiscau.ts$hyear > 1962) caniapiscau.res <- metrics.all(caniapiscau.ts) ## End(Not run) # example use of example subset flow series data(caniapiscau.res)# Code used produce this data set: ## Not run: data(caniapiscau) caniapiscau.ts <- create.ts(caniapiscau, hyrstart=3) caniapiscau.ts <- subset(caniapiscau.ts, caniapiscau.ts$hyear > 1962) caniapiscau.res <- metrics.all(caniapiscau.ts) ## End(Not run) # example use of example subset flow series data(caniapiscau.res)
This function creates a daily time series formatted for use with the functions in this package.
create.ts(Flows, hyrstart = 10)create.ts(Flows, hyrstart = 10)
Flows |
Data.frame containing daily streamflow time series loaded
with the |
hyrstart |
define start month of hydrologic year. Defaults to 10 (October). |
Returns a data.frame with year, month, doy, and hyear columns appended to the original input data.frame.
Jennifer Dierauer
data(caniapiscau) # subset flow series for shorter example run time caniapiscau.sub <- caniapiscau[300:1800,] caniapiscau.sub.ts <- create.ts(caniapiscau.sub)data(caniapiscau) # subset flow series for shorter example run time caniapiscau.sub <- caniapiscau[300:1800,] caniapiscau.sub.ts <- create.ts(caniapiscau.sub)
This function extracts the partial duration series for all streamflow droughts based on a moving window quantile threshold. Also returns summary information (start date, end date, duration, deficit volume) for each drought event.
dr.events(TS, Qdr = 0.2, WinSize = 30, IntEventDur = 10, EventDur = 15)dr.events(TS, Qdr = 0.2, WinSize = 30, IntEventDur = 10, EventDur = 15)
TS |
output from |
Qdr |
Numeric value of the drought threshold quantile. Default is 0.2. |
WinSize |
Numeric value specifying the size of the moving window in days. Default is 30. |
IntEventDur |
Numeric value for the minimum inter-event duration in days. Drought events with less than the specified number of days between will be pooled and considered as one event. |
EventDur |
Numeric value for the minimum drought duration in days. Default is 15. |
Returns a list with the following elements:
DroughtEvents: A data.frame with the following columns:
Event - Integer indicating the original event number assigned before minor drought events were removed.
Start - Date of the start of the drought event.
End - Date of the end of the drought event
maxDef - Numeric value of the maximum streamflow deficit.
Severity - Numeric value indicating the drought severity, calculated as the cumulative daily streamflow deficit in m3/s.
Duration - Numeric value of the drought duration in days.
Magnitude - Numeric value indicating the drought magnitude, which is calculated as the mean daily streamflow deficit in m3/s.
stdtotDef - Numeric value indicating the standardized cumulative streamflow deficit, calculated as the drought severity divided by the mean annual daily streamflow.
DroughtPDS: A data.frame of the original input TS that has been subset to include only the days on which the streamflow was below the drought threshold. The data.frame also has the following columns appended:
Thresh - Numeric value indicating the streamflow drought threshold,
as calculated by mqt
BelowThresh - Logical indicating whether the observed streamflow was below the streamflow drought threshold.
Def - Numeric value of the streamflow defict, calculated as the streamflow drought threshold (m3/s) minus the observed streamflow (m3/s).
Jennifer Dierauer
See dr.seas to calculate metrics for droughts
occurring in a user-defined season.
This function calls dr.pds which calls mqt.
data(cania.sub.ts) res1 <- dr.events(cania.sub.ts) events <- res1$DroughtEvents plot(events$Start, events$Duration, pch=19, ylab="Drought Duration (days)", xlab="")data(cania.sub.ts) res1 <- dr.events(cania.sub.ts) events <- res1$DroughtEvents plot(events$Start, events$Duration, pch=19, ylab="Drought Duration (days)", xlab="")
This function returns the partial duration series for streamflow droughts based on a moving window quantile threshold.
dr.pds(TS, Qdr = 0.2, WinSize = 30)dr.pds(TS, Qdr = 0.2, WinSize = 30)
TS |
output from |
Qdr |
Numeric value of the drought threshold quantile. Default is 0.2. |
WinSize |
Numeric value specifying the size of the moving window in days. Default is 30. |
This function defines a daily streamflow threshold and finds the partial duration series of streamflow droughts. Drought events are identified in the daily streamflow time series with the threshold level approach. In this function, the threshold is defined by a moving quantile, where daily threshold values are based on the 80th percentile of the flow duration curve from a 30-day moving window (Beyene et al. 2014). With this method, every day of the year has a different threshold based on the streamflow measured on the day, the 15 days before the day, and the 15 days after the day. The size of the moving window can be modified with the WinSize argument, and the percentile can be modified with the Qdr argument.
Returns the input TS data.frame with "Thresh" and "BelowThresh" columns appended. The Thresh column contains the daily flow threshold, and the BelowThresh column is a binary indicating whether the flow on each day was below the drought threshold.
Jennifer Dierauer
Beyene, B.S., Van Loon, A.F., Van Lanen, H.A.J., Torfs, P.J.J.F., 2014. Investigation of variable threshold level approaches for hydrological drought identification. Hydrol. Earth Syst. Sci. Discuss. 11, 12765-12797. http://dx.doi.org/10.5194/hessd-11-12765-2014.
See create.ts to format the input flow series.
See mqt to return only the daily moving quantile threshold.
See dr.events to pool drought events, remove minor events,
and calculate metrics.
See dr.seas to calculate metrics for streamflow droughts
that start in a specific month or months.
data(cania.sub.ts) pds <- dr.pds(cania.sub.ts) pds <- subset(pds, pds$BelowThresh==TRUE) # plot the flow time series with black and the drought events in red plot(cania.sub.ts$Date, cania.sub.ts$Flow, ylab="m3/s", xlab="", type="l") points(pds$Date, pds$Flow, pch=19, cex=0.7, col="red")data(cania.sub.ts) pds <- dr.pds(cania.sub.ts) pds <- subset(pds, pds$BelowThresh==TRUE) # plot the flow time series with black and the drought events in red plot(cania.sub.ts$Date, cania.sub.ts$Flow, ylab="m3/s", xlab="", type="l") points(pds$Date, pds$Flow, pch=19, cex=0.7, col="red")
This function returns the day of year for the start, middle, and end of seasonal droughts. It also returns the duration and severity of each drought event. The function allows for seasonal analysis by defining a season argument which lists months during which droughts of interest may start.
dr.seas(TS, Qdr = 0.2, WinSize = 30, IntEventDur = 10, EventDur = 15, Season = c(4:9))dr.seas(TS, Qdr = 0.2, WinSize = 30, IntEventDur = 10, EventDur = 15, Season = c(4:9))
TS |
output from |
Qdr |
Numeric value for drought quantile. Default is 0.2. |
WinSize |
Numeric value for moving window size in days. Default is 30. |
IntEventDur |
Numeric value for the minimum inter-event duration in days. Drought events with less than the specified number of days between will be pooled and considered as one event. Default is 10. |
EventDur |
Numeric value for the minimum drought duration in days. Default is 15. |
Season |
Numeric vector of months during which droughts start. Default is c(4:9) for non-frost season droughts. |
This function calls dr.events which calls
dr.pds and mqt
Returns a data.frame of drought event metrics; the columns are:
StartDay - day of year that the drought event started on
MidDay - day of year for the middle of the drought event, which is defined as the day when the cumulative drought deficit reached 50 total cumulative daily streamflow deficit. Total cumulative streamflow deficit is also referred to as drought severity in this package.
EndDay - day of year that the drought ended on
Duration - length of the drought event, in days
Severity - severity of the drought event, calculated as the total cumulative daily streamflow deficit
The "times" attribute provides the start date to preserve year information and aid in plotting the time series.
Jennifer Dierauer
See create.ts to format the input flow series.
See dr.events and mqt for details on how drought
events are defined.
data(cania.sub.ts) res <- dr.seas(cania.sub.ts) res2 <- screen.metric(res[,1], "Day of Year")data(cania.sub.ts) res <- dr.seas(cania.sub.ts) res2 <- screen.metric(res[,1], "Day of Year")
Produces a flow duration curve plot with optional Gustard type-curves that can be used to estimate catchment permeability.
FDC(flow, title = NULL, normal = FALSE, gust = TRUE)FDC(flow, title = NULL, normal = FALSE, gust = TRUE)
flow |
daily streamflow time series |
title |
character string for plot title |
normal |
boolean indicating whether to plot on normal probability axis (normal=TRUE) or linear probability axis (default, normal=FALSE) |
gust |
boolean indicating whether to plot Gustard type curves. |
Paul Whitfield
Gustard, A., Bullock, A., and Dixon, J.M. (1992). Report No. 108: Low flow estimation in the United Kingdom. Oxfordshire, United Kingdom: Institute of Hydrology.
data(caniapiscau) caniapiscau <- subset(caniapiscau, !is.na(caniapiscau$Flow)) FDC(caniapiscau$Flow, title="Caniapiscau River")data(caniapiscau) caniapiscau <- subset(caniapiscau, !is.na(caniapiscau$Flow)) FDC(caniapiscau$Flow, title="Caniapiscau River")
This package can be used to calculate more than 30 different streamflow metrics and identify temporal trends and changepoints. It is intended for use as a data quality screening tool aimed at identifying streamflow records that may have anthropogenic impacts or data inhomogeneity.
| Package: | FlowScreen |
| Type: | Package |
| Version: | 1.2.6 |
| Date: | 2019-04-05 |
| License: | GPL (>= 2) |
Daily streamflow time series downloaded with the Environment Canada Data Explorer can be loaded with read.flows. The read.flows function can also be used to load daily streamflow time series from the USGS. The streamflow regime can be visualized with regime. A list of 30 streamflow metrics that describe high flows, low flows, and baseflows can be calculated using metrics.all. The temporal occurrence of changepoints for all metrics or for only the high flow, baseflow, or low flow metrics can be analyzed using screen.cpts. If the streamflow time series has multiple metrics exhibiting changepoints within the same year (or few years), the time series can be further analyzed using screen.summary which creates a summary plot showing the significant temporal trends and changepoints for the high flow, low flow, or baseflow metrics. The screen.metric can be used to create a time series plot for one metric at a time. The screen.metric function works with individual metrics output from the following functions: pk.max, pk.max.doy, Qn, pk.bf.stats, dr.seas, MAMn, bf.stats, pk.cov, and bf.seas.The screen.frames function creates individual plots from the screen.summary function. The screen.frames function can also be used to create custom summary plots, see the example code in the function documentation.
Jennifer Dierauer, Paul H. Whitfield
Maintainer: Jennifer Dierauer <[email protected]>
Bard, A., Renard, B., Lang, M. 2011. The AdaptAlp Dataset: Description, guidance, and analyses. In AdaptAlp WP 4 Report, 15. Lyon, France: Cemagraf.
Bard, A., Renard, B., Lang, M., Giuntoli, I., Korck, J., Koboltschnig, G., Janza, M., d'Amico, M., Volken, D. 2015. Trends in the hydrologic regime of Alpine rivers. Journal of Hydrology online.
Svensson, C., Kundzewicz, Z.W., Maurer, T. 2005. Trend detection in river flow series: 2. Flood and low-flow index series. Hydrological Sciences Journal 50:811-824.
Whitfield, P.H. 2012. Why the provenance of data matters: Assessing "Fitness for Purpose" for environmental data. Canadian Water Resources Journal 37:23-36.
Whitfield, P.H. 2013. Is 'Center of Volume' a robust indicator of changes in snowmelt timing? Hydrological Processes 27:2691-2698.
pot, decluster, cpt.meanvar, zyp.trend.vector, Kendall
## Not run: # load daily streamflow time series for the Caniapiscau River data(caniapiscau) # summary plot of the annual flow regime caniapiscau.ts <- create.ts(caniapiscau) regime(caniapiscau.ts) # calculate high flow, low flow, and baseflow metrics res <- metrics.all(caniapiscau.ts) # plot histogram of changepoints for high flow, low flow, and baseflow metrics screen.cpts(res, type="h") screen.cpts(res, type="l") screen.cpts(res, type="b") # or plot all changepoints together cpts <- screen.cpts(res) # create screening plots for high, low, and baseflow metrics screen.summary(res, type="h") screen.summary(res, type="l") screen.summary(res, type="b") ## End(Not run)## Not run: # load daily streamflow time series for the Caniapiscau River data(caniapiscau) # summary plot of the annual flow regime caniapiscau.ts <- create.ts(caniapiscau) regime(caniapiscau.ts) # calculate high flow, low flow, and baseflow metrics res <- metrics.all(caniapiscau.ts) # plot histogram of changepoints for high flow, low flow, and baseflow metrics screen.cpts(res, type="h") screen.cpts(res, type="l") screen.cpts(res, type="b") # or plot all changepoints together cpts <- screen.cpts(res) # create screening plots for high, low, and baseflow metrics screen.summary(res, type="h") screen.summary(res, type="l") screen.summary(res, type="b") ## End(Not run)
Get station information for USGS or WSC hydrometric stations
get.station.internal(stnID)get.station.internal(stnID)
stnID |
Character string of station ID. |
Returns a list of station information
Jennifer Dierauer
Returns plot titles and labels based on plot type and language preference
get.titles.internal(type, language = "English", Qmax)get.titles.internal(type, language = "English", Qmax)
type |
character indicating the type of summary plot |
language |
"English" or "French" |
Qmax |
the flow quantile used to define peaks of threshold, e.g. 0.95 |
Jennifer Dierauer
Add hydrologic Year, month, and doy columns to a daily time series
hyear.internal(TS, hyrstart = 10)hyear.internal(TS, hyrstart = 10)
TS |
Output from |
hyrstart |
define start month of hydrologic year. Defaults to 10 (October). |
Returns a data.frame with hyear, hmonth, and hdoy columns appended to the original input data.frame.
Jennifer Dierauer
This function calculates the calculates the mean annual minimum n-day flow by calendar year or by hydrologic year. This function can also be used to find the annual minimum series by setting n=1.
MAMn(TS, n = 7, by = "hyear")MAMn(TS, n = 7, by = "hyear")
TS |
output from |
n |
Numeric value for the number of days in the n-day flow period. Default is 7. |
by |
Character string indicating whether to use hydrologic years or calendar years. Default is "hyear". Other option is "year". |
Returns a numeric vector containing the calculated MAM n-day flow for each year in the input time series. The "times" attribute provides the corresponding year for each calculated value.
Jennifer Dierauer
data(cania.sub.ts) # find the annual minimum series and plot res <- MAMn(cania.sub.ts, n=1) res2 <- screen.metric(res, "Q (m3/s)") # do the same with MAM 7-day flow instead of annual minimum res <- MAMn(cania.sub.ts, n=7) res2 <- screen.metric(res, "Q (m3/s)")data(cania.sub.ts) # find the annual minimum series and plot res <- MAMn(cania.sub.ts, n=1) res2 <- screen.metric(res, "Q (m3/s)") # do the same with MAM 7-day flow instead of annual minimum res <- MAMn(cania.sub.ts, n=7) res2 <- screen.metric(res, "Q (m3/s)")
Calculates 30 different flow metrics, 10 each for high flows, low flows, and baseflow.
metrics.all(TS, Qmax = 0.95, Dur = 5, Qdr = 0.2, WinSize = 30, Season = c(4:9), NAthresh = 0.5, language = "English")metrics.all(TS, Qmax = 0.95, Dur = 5, Qdr = 0.2, WinSize = 30, Season = c(4:9), NAthresh = 0.5, language = "English")
TS |
output from |
Qmax |
Numeric value for peaks over threshold quantile. Default is 0.95. |
Dur |
Numeric value for minimum number of days between flood peaks. Default is 5. |
Qdr |
Numeric value for drought quantile. Default is 0.2, i.e. the 80th percentile of the flow duration curve. |
WinSize |
Numeric value for moving window size (in days) for the moving
window quantile drought threshold. See |
Season |
Numeric vector of months during which droughts start. Default is c(4:9) for non-frost season droughts. |
NAthresh |
Numeric value indicating the threshold for missing data points in any one year. Default is 0.5, indicating that years with more than 50 percent missing data will be omitted from the metric calculations. This value should always be set to greater than 0.1, as years with fewer observations than approximately 1 month will cause errors. |
language |
Character string indicating the language to be used for naming
the different plot metrics. These names are used in |
This function calculates streamflow metrics and calculates the
prewhitened trend using zyp.trend.vector and looks
for changpoints in mean and variance using cpt.meanvar
This function is intended for use as a data quality screening tool aimed
at identifying streamflow records with anthropogenic impacts and should not be used
to complete a temporal trend analysis, as the calculated metrics may not be
appropriate for all catchments. See the functions linked in the following section
for details on how each metric is calculated.
Returns a list with the following elements:
metricTS: a list containing a vector of each metric calculated. Each vector has a times attribute providing either the year for metrics with one observation per year or a date for metrics that may have more than one observation per year (e.g., Peaks Over Threshold). This list has the following elements:
Annual Maximum Series - calculated with pk.max
Day of Annual Maximum - calculated with pk.max.doy
Peaks Over Threshold (Qmax) - calculated with pks
Inter-Event Duration - calculated with pks.dur
Q80 - calculated with Qn
Q90 - calculated with Qn
Day of Year 25 percent Annual Volume - calculated with pk.cov
Center of Volume - calculated with pk.cov
Day of Year 75 percent Annual Volume - calculated with pk.cov
Duration between 25 percent and 75 percent Annual Volume - calculated with cov
Q10 - calculated with Qn
Q25 - calculated with Qn
Drought Start - calculated with dr.seas
Drought Center - calculated with dr.seas
Drought End - calculated with dr.seas
Drought Duration - calculated with dr.seas
Drought Severity - calculated with dr.seas
Annual Minimum Flow - calculated with MAMn
Mean Annual Minimum 7-day Flow - calculated with MAMn
Mean Annual Minimum 10-day Flow - calculated with MAMn
Mean Daily Discharge - calculated with bf.stats
Annual Baseflow Volume - calculated with bf.stats
Annual Mean Baseflow - calculated with bf.stats
Annual Maximum Baseflow - calculated with bf.stats
Annual Minimum Baseflow - calculated with bf.stats
Mean Annual Baseflow Index - calculated with bf.stats
Day of Year 25 percent Baseflow Volume - calculated with pk.bf.stats
Center of Volume Baseflow - calculated with pk.bf.stats
Day of Year 75 percent Baseflow Volume - calculated with pk.bf.stats
Duration between 25 percent and 75 percent Baseflow Volume - calculated with pk.bf.stats
tcpRes: this list contains the results of the trend and changepoint analysis for each of the metrics in the metricTS list described above. Each list element is a list containing the following elements:
MetricID - integer used to identify the metric
MetricName - Name of the metric.
Slope - numeric vector containing the intercept and slope of the
prewhitened linear trend calculated using the Yue Pilon method. See
zyp.trend.vector
ci1 - upper bound of the trend's 95 percent confidence interval
ci2 - lower bound of the trend's 95 percent conficence interval
pval - Kendall's P-value computed for the detrended time series
cpts - Most probable location of a changepoint, if one is detected.
means - Mean before and after the changepoint
NumObs - The number of data points for the metric
inData: A data.frame of the original input daily streamflow time series.
OmitYrs: A data.frame containing the years and the number of observations for any years omitted from the analysis due to insufficient data. If no years were omitted, NA is returned.
Jennifer Dierauer
See the documentation for individual functions linked in the output description for a details on methods.
See screen.metric to create individual plots for each metric.
# load subset of daily streamflow time series for the Caniapiscau River data(cania.sub.ts) ## Not run: # calculate low flow, high flow, and baseflow metrics res <- metrics.all(cania.sub.ts) ## End(Not run)# load subset of daily streamflow time series for the Caniapiscau River data(cania.sub.ts) ## Not run: # calculate low flow, high flow, and baseflow metrics res <- metrics.all(cania.sub.ts) ## End(Not run)
This function calculates the daily moving window quantile threshold for use in identifying the partial duration series of streamflow droughts.
mqt(TS, Qdr = 0.2, WinSize = 30)mqt(TS, Qdr = 0.2, WinSize = 30)
TS |
output from |
Qdr |
Numeric value of the drought threshold quantile. Default is 0.2. |
WinSize |
Numeric value specifying the size of the moving window in days. Default is 30. |
The threshold is defined by a moving quantile, where daily threshold values are based on the 80th percentile of the flow duration curve (i.e. 0.2 quantile) from a 30-day moving window (Beyene et al. 2014). With this method, every day of the year has a different threshold based on the streamflow measured on the day, the 15 days before the day, and the 15 days after the day.The size of the moving window can be modified with the WinSize argument, and the percentile can be modified with the Qdr argument.
Returns a numeric vector containing the streamflow drought threshold in m3/s for each day of the year.
Jennifer Dierauer
Beyene, B.S., Van Loon, A.F., Van Lanen, H.A.J., Torfs, P.J.J.F., 2014. Investigation of variable threshold level approaches for hydrological drought identification. Hydrol. Earth Syst. Sci. Discuss. 11, 12765-12797. http://dx.doi.org/10.5194/hessd-11-12765-2014.
See create.ts to format the input flow series.
The following functions use this function: dr.pds,
dr.events, dr.seas
data(cania.sub.ts) res <- mqt(cania.sub.ts) # subset one year of the flow series flow.sub <- cania.sub.ts[cania.sub.ts$year == 1990,] # plot the 1990 observed flows in dark blue and the daily drought threshold in red plot(flow.sub$doy, flow.sub$Flow, ylab="Q (m3/s)", xlab="Day of Year", pch=19, col="darkblue", type="b") points(res, pch=19, cex=0.7, col="red")data(cania.sub.ts) res <- mqt(cania.sub.ts) # subset one year of the flow series flow.sub <- cania.sub.ts[cania.sub.ts$year == 1990,] # plot the 1990 observed flows in dark blue and the daily drought threshold in red plot(flow.sub$doy, flow.sub$Flow, ylab="Q (m3/s)", xlab="Day of Year", pch=19, col="darkblue", type="b") points(res, pch=19, cex=0.7, col="red")
This function takes a data.frame from create.ts and returns a data.frame of missing data runs.
NA.runs(TS)NA.runs(TS)
TS |
output from |
Returns a data.frame with the following columns:
Start - Date of the start of the missing data period
End - Date of the end of the missing data period
Duration - number of days in the missing data period
Jennifer Dierauer
create.ts to create input, NA.sum to sum the
the missing data occurrences by year or month.
data(caniapiscau) cania.sub <- caniapiscau[300:1200,] cania.ts <- create.ts(cania.sub) res <- NA.runs(cania.ts)data(caniapiscau) cania.sub <- caniapiscau[300:1200,] cania.ts <- create.ts(cania.sub) res <- NA.runs(cania.ts)
Counts the number of missing data points by calendar year, hydrologic year, or month
NA.sum(input, by = "hyear", hyrstart = 1)NA.sum(input, by = "hyear", hyrstart = 1)
input |
output from |
by |
character string identifying the time period to summarize by. Defaults is hydrologic year ("hyear"), other choices are "year" and "month". The "month" option will return the number of missing data points for each month in the time series. |
hyrstart |
optional argument, define start month of hydrologic year |
Returns a numeric vector of the number of missing observations per summary period. The "times" attribute of the returned vector provides the corresponding year, hyear, or month.
Jennifer Dierauer
data(caniapiscau) cania.sub <- caniapiscau[300:1200,] cania.ts <- create.ts(cania.sub) res <- NA.runs(cania.ts) res2 <- NA.sum(res)data(caniapiscau) cania.sub <- caniapiscau[300:1200,] cania.ts <- create.ts(cania.sub) res <- NA.runs(cania.ts) res2 <- NA.sum(res)
This function finds the start, middle, end, and duration of the baseflow peak based on percent of the total annual baseflow volume. A value of 0 is returned for years with no flow. Hydrologic years with fewer than normal observations (outliers) are excluded from the analysis, and for stations with seasonal flow records, additional seasonal subsetting is done to include only days with observations in all years.
pk.bf.stats(TS, bfpct = c(25, 50, 75))pk.bf.stats(TS, bfpct = c(25, 50, 75))
TS |
output from |
bfpct |
numeric vector of percentages used to define the start, middle, and end of the baseflow peak. Default is c(25, 50, 75) |
This function calculates metrics intended to focus on snowmelt-related
streamflow occuring in spring and summer. For catchments in cold climates,
the baseflow peak can be interpreted as snowmelt-induced. Baseflow is estimated with
bf_eckhardt. If total annual flow is equal to 0, returns NA for that
year.
Returns a data.frame with the following columns:
Start - day of year defining the start of the baseflow peak
Mid - day of year defining the middle of the baseflow peak
End - day of year defining the end of the baseflow peak
Dur - duration of the baseflow peak, in days
Jennifer Dierauer
data(cania.sub.ts) res1 <- pk.bf.stats(cania.sub.ts) # trend and changepoint plot for baseflow peak start doy res2 <- screen.metric(res1[,1], "Day of Year")data(cania.sub.ts) res1 <- pk.bf.stats(cania.sub.ts) # trend and changepoint plot for baseflow peak start doy res2 <- screen.metric(res1[,1], "Day of Year")
This function calculates center of volume metrics, including the day of the hydrologic year that 25 percent, 50 percent, and 75 percent of the total annual streamflow is reached. A value of 0 is returned for years with no flow. Hydrologic years with fewer than normal observations (outliers) are excluded from the analysis, and for stations with seasonal flow records, additional seasonal subsetting is done to include only days with observations in all years.
pk.cov(TS)pk.cov(TS)
TS |
output from |
Returns a data.frame with the following columns:
hYear - Hydrologic Years
Q25 - day of hydrologic year for 25 percent of the total annual streamflow
Q50 - day of hydrologic year for 50 percent of the total annual streamflow, i.e. Center of Volume
Q75 - day of hydrologic year for 75 percent of the total annual streamflow
Dur - duration of between the 25 percent and 75 percent day of year, in days
Jennifer Dierauer
data(cania.sub.ts) res1 <- pk.cov(cania.sub.ts) # trend and changepoint plot for baseflow peak start doy res2 <- screen.metric(res1[,2], "Day of Year")data(cania.sub.ts) res1 <- pk.cov(cania.sub.ts) # trend and changepoint plot for baseflow peak start doy res2 <- screen.metric(res1[,2], "Day of Year")
This function returns the annual maximum series from a daily streamflow time series.
pk.max(TS)pk.max(TS)
TS |
output from |
Returns a numeric vector containing the annual maximum flow (m3/s) series, by hydrologic year. The "times" attribute contains the hydrologic year for each element in the vector.
Jennifer Dierauer
See create.ts to format the input flow series.
See pk.max.doy to find the day of year for each annual
maximum flow event.
data(cania.sub.ts) res <- pk.max(cania.sub.ts) res2 <- screen.metric(res, "Q (m3/s)")data(cania.sub.ts) res <- pk.max(cania.sub.ts) res2 <- screen.metric(res, "Q (m3/s)")
This function returns the day of the hydrologic year for each annual maximum flow.
pk.max.doy(TS)pk.max.doy(TS)
TS |
output from |
Returns a numeric vector containing the day of the (hydrologic) year for each annual maximum flow. The "times" attribute contains the hydrologic year for each element in the vector.
Jennifer Dierauer
See create.ts to format the input flow series.
See pk.max for the annual maximum flow series.
data(cania.sub.ts) res <- pk.max.doy(cania.sub.ts) res2 <- screen.metric(res, "Day of Year")data(cania.sub.ts) res <- pk.max.doy(cania.sub.ts) res2 <- screen.metric(res, "Day of Year")
This function finds the flow peaks over a user defined threshold and declusters to remove dependent peaks.
pks(TS, Dur = 5, Qmax = 0.95)pks(TS, Dur = 5, Qmax = 0.95)
TS |
output from |
Dur |
numeric value of the minimum number of days between peaks |
Qmax |
numeric value for peaks over threshold quantile. Default is 0.95. |
Peaks Over Threshold values are calcuated as mean daily streamflow (m3/s)
minus the threshold streamflow value (m3/s) defined by the
input quantile (Qmax). Peaks are identified with pot and
the minimum inter-event duration (Dur) is applied by
decluster.
Returns a numeric vector of peaks of threshold values in m3/s. The "times" attribute contains the date by calendar year, and the "names" attribute contains the hydrologic year and hydrologic day of year, e.g., 2012 55.
Jennifer Dierauer
data(cania.sub.ts) res <- pks(cania.sub.ts) res2 <- screen.metric(res, "Peak Over Threshold (m3/s)")data(cania.sub.ts) res <- pks(cania.sub.ts) res2 <- screen.metric(res, "Peak Over Threshold (m3/s)")
This function calculates duration (in days) between flow peaks.
pks.dur(Peaks)pks.dur(Peaks)
Peaks |
Output from |
Returns a numeric vector containing the duration (in days) between peaks
over threshold from pks. The "times" attribute contains the calendar
year date of the earlier peak. The "names" attribute contains the hydrologic year and
the day (1-366) of the hydrologic year.
Jennifer Dierauer
data(cania.sub.ts) res1 <- pks(cania.sub.ts) res2 <- pks.dur(res1) res3 <- screen.metric(res2, "Inter-Event Duration (days)")data(cania.sub.ts) res1 <- pks(cania.sub.ts) res2 <- pks.dur(res1) res3 <- screen.metric(res2, "Inter-Event Duration (days)")
This function calculates flow quantiles by hydrologic year, calendar year, month, or doy.
Qn(TS, n = 0.1, by = "hyear")Qn(TS, n = 0.1, by = "hyear")
TS |
output from |
n |
Numeric value of the quantile. Default is 0.1. |
by |
Character string indicating time unit to summarize by. Default is
"hyear" for hydrologic year, see |
Returns a numeric vector of the calculated flow quantile for the time periods indicated with the "by" argument. The "times" attribute contains the hydrologic year, calendar year, month, or day of year for each data point.
Jennifer Dierauer
data(cania.sub.ts) # 50% quantile, i.e. mean, by calendar year res <- Qn(cania.sub.ts, n=0.5, by="year") res2 <- screen.metric(res, "Q (m3/s)") # Default 10% quantile, by hydrologic year res <- Qn(cania.sub.ts) res2 <- screen.metric(res, "Q (m3/s)")data(cania.sub.ts) # 50% quantile, i.e. mean, by calendar year res <- Qn(cania.sub.ts, n=0.5, by="year") res2 <- screen.metric(res, "Q (m3/s)") # Default 10% quantile, by hydrologic year res <- Qn(cania.sub.ts) res2 <- screen.metric(res, "Q (m3/s)")
Reads .csv or .Rdata files of daily streamflow time series. Recognizes several formats, including those used by Water Survey Canada and the United States Geological Survey. Uses read.csv() or load() functions from base package, returns data frame with ID, Date, and Flow, and, if available, associated quality codes and source agency. Replaces negative values that are sometimes used to denote missing data with NAs.
read.flows(filename, convert = FALSE)read.flows(filename, convert = FALSE)
filename |
name of .csv file to be read. |
convert |
Boolean indicating whether or not to convert USGS flow values from cubic feet per second to cubic meters per second. (Axes are automatically labeled in cubic meters per second) |
Jennifer Dierauer
This function plots the min, max, mean, and two user-defined quantiles of daily streamflow to provide visual summary of the flow regime. Area between the upper and lower quantile is shaded grey, the dark blue line represents the mean daily discharge, gray line represents the median daily discharge, and the period of record daily maximum and minimum are shown with the blue points.
regime(TS, q = c(0.9, 0.1), text = "d", by = "hdoy", y.lims = NA)regime(TS, q = c(0.9, 0.1), text = "d", by = "hdoy", y.lims = NA)
TS |
output from |
q |
Numeric vector of the upper and lower quantile values. Default is c(0.9, 0.1). |
text |
optional character string for margin text, e.g. for station name,
location, or other notes. Set to NULL if no margin text is wanted, or set to "d"
to use default text containing the station ID, station name, and province/state
returned from |
by |
Character string indicating whether to plot the regime by day of
the hydrologic year (defined using |
y.lims |
optional user-defined y-axis minimum and maximim. e.g. c(0, 500) |
Jennifer Dierauer
# plot the flow regime of the Caniapiscau River data(cania.sub.ts) regime(cania.sub.ts)# plot the flow regime of the Caniapiscau River data(cania.sub.ts) regime(cania.sub.ts)
Compiles change point information for all metrics and outputs a daily flow time series plot overlain with a bar plot of changepoint counts by year.
screen.cpts(metrics, type = "a", text = NULL)screen.cpts(metrics, type = "a", text = NULL)
metrics |
output from |
type |
character indicating which type of metric to compile change points for. Options are "h" for high flow metrics, "l" for low flow metrics, "b" for baseflow metrics, or "a" for all 30 metrics (10 high, 10 low, 10 baseflow). |
text |
optional character string for margin text, e.g. for station name,
location, or other notes. Set to NULL if no margin text is wanted, or set to "d"
to use default text containing the station ID, station name, and province/state
returned from |
When type="a", returns a data.frame of changepoint counts by metric type and year.
Jennifer Dierauer
# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) # plot changepoints for all metrics screen.cpts(caniapiscau.res, type="l")# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) # plot changepoints for all metrics screen.cpts(caniapiscau.res, type="l")
This function plots one or more frames (i.e. time series plot) from any of the three plot.screening summary plots at a time. It can be used to create custom summary plots - see the example code.
screen.frames(metrics, type = "h", element = NULL, language = "English", StnInfo = NULL, mmar = c(3, 4, 0.5, 0.5), text = "d", multi = F, xaxis = T)screen.frames(metrics, type = "h", element = NULL, language = "English", StnInfo = NULL, mmar = c(3, 4, 0.5, 0.5), text = "d", multi = F, xaxis = T)
metrics |
output from |
type |
Character string indicating the set of metrics to plot. Options are "h" for high flow metrics, "l" for low flow metrics, or "b" for baseflow metrics. |
element |
Numeric index(es) (1-10) of the frame(s) to plot, see details of this function for the list of metrics for each category (high, low, baseflow). Each category has ten different metrics that can be plotted individually. Default is NULL, which creates individual plots for all ten metrics. A list of elements c(1, 5, 10) can be specified or a range c(1:5). |
language |
Language for plot labels. Choice of either "English" or "French". Default is "English". |
StnInfo |
Optional data.frame containing user-supplied station info for plot.
data.frame must have 7 columns containing station info in the following order:
Station ID, Station Name, Prov/State, Country, Latitude, Longitude, Catchment Area
If any of the information is unavailabe, fill with NA. The Station ID column must
match the Station ID in column 1 of the data.frame input from |
mmar |
Numeric vector specifying plot margins. Default is c(3,4,0.5,0.5) |
text |
Character string containing text for margin. This can be set to NULL if no
margin text is wanted, or set to "d" to use default text containing the station ID, station
name, and the prov/state output from |
multi |
Boolean indicating whether the function is being used to create one plot in a multi-plot layout. Default is F. If T, suppresses the reset of plot parameter settings. This plot function will only work for a multi-plot layout if text=F |
xaxis |
Boolean indicating whether to plot an x-axis. Default = T. |
High flow metrics include:
Annual Maximum Series
Annual Maximum Day of Year
Peaks Over Threshold (Qmax)
Inter-Event Duration
Q80
Q90
Day of Year 25 percent Annual Flow
Center of Volume
Day of Year 75 percent Annual Flow
Duration between 25 percent and 75 percent Annual Flow
Low flow metrics include:
Q10
Q25
Drought Start
Drought Center
Drought End
Drought Duration
Drought Severity
Annual Minimum Flow
Mean Annual Minimum 7-day Flow
Mean Annual Minimum 10-day Flow
Baseflow metrics include:
Mean Daily Discharge
Annual Baseflow Volume
Annual Mean Baseflow
Annual Maximum Baseflow
Annual Minimum Baseflow
Mean Annual Baseflow Index
Day of Year 25 percent Baseflow Volume
Center of Volume Baseflow
Day of Year 75 percent Baseflow Volume
Duration between 25 percent and 75 percent Baseflow Volume
Jennifer Dierauer and Paul Whitfield
# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) caniapiscau.ts <- caniapiscau.res$indata # plot one frame from the baseflow screening plot screen.frames(caniapiscau.res, type="b", element=1) # plot three frames from the low flow screening plot screen.frames(caniapiscau.res, type="l", element=c(1:3)) # create a custom summary plot opar <- par(no.readonly = TRUE) layout(matrix(c(1,2,3,4), 2, 2, byrow=TRUE)) par(oma=c(0,0,3,0)) stninfo <- station.info(caniapiscau.ts, Plot=TRUE) screen.frames(caniapiscau.res, type="h", element=1, text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="l", element=1, text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="b", element=1, text=NULL, multi=TRUE) mtext(paste("Station ID: ", caniapiscau.ts[1,1], ", Agency: WSC, Country: CA", sep=""), side=3, line=1, outer=TRUE, cex=0.9) par <- opar layout(1,1,1) # or plot everything! opar <- par(no.readonly = TRUE) layout(matrix(c(1:30), 5, 6, byrow=TRUE)) screen.frames(caniapiscau.res, type="h", text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="l", text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="b", text=NULL, multi=TRUE) par <- opar layout(1,1,1)# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) caniapiscau.ts <- caniapiscau.res$indata # plot one frame from the baseflow screening plot screen.frames(caniapiscau.res, type="b", element=1) # plot three frames from the low flow screening plot screen.frames(caniapiscau.res, type="l", element=c(1:3)) # create a custom summary plot opar <- par(no.readonly = TRUE) layout(matrix(c(1,2,3,4), 2, 2, byrow=TRUE)) par(oma=c(0,0,3,0)) stninfo <- station.info(caniapiscau.ts, Plot=TRUE) screen.frames(caniapiscau.res, type="h", element=1, text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="l", element=1, text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="b", element=1, text=NULL, multi=TRUE) mtext(paste("Station ID: ", caniapiscau.ts[1,1], ", Agency: WSC, Country: CA", sep=""), side=3, line=1, outer=TRUE, cex=0.9) par <- opar layout(1,1,1) # or plot everything! opar <- par(no.readonly = TRUE) layout(matrix(c(1:30), 5, 6, byrow=TRUE)) screen.frames(caniapiscau.res, type="h", text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="l", text=NULL, multi=TRUE) screen.frames(caniapiscau.res, type="b", text=NULL, multi=TRUE) par <- opar layout(1,1,1)
Internal wrapper for creating trend and change-point plots
screen.frames.internal(input, mparam, mylab, DataType, maf, mmar, text, xaxis, Year1, YearEnd, hyrstart)screen.frames.internal(input, mparam, mylab, DataType, maf, mmar, text, xaxis, Year1, YearEnd, hyrstart)
input |
metric time series |
mparam |
trend and change point info |
mylab |
y axis label |
DataType |
numeric indicating data type |
maf |
mean annual flow series |
mmar |
plot margins |
text |
boolean indicating whether to add text |
xaxis |
boolean indicating whether to plot the x axis |
Year1 |
start year of original time series |
YearEnd |
end year of original time series |
hyrstart |
numeric indicating month for start of the hydrologic year |
Jennifer Dierauer
This function plots a time series of a streamflow metric with the prewhitened linear trend and any detected changepoints in mean and variance.
screen.metric(y, ylabel = "", text = NULL)screen.metric(y, ylabel = "", text = NULL)
y |
Numeric vector with "times" attribute |
ylabel |
Character string for the y-axis label |
text |
optional character string for margin text, e.g. for station name, location, or other notes. |
This function plots detected changepoints as a vertical dashed line. The means on either side of a changepoint are plotted as solid black lines. If the temporal trend is significant (p-value < 0.1), the trend is plotted as a blue or red line for an increasing or decreasing trend, respectively. The upper and lower 95 dotted red or blue lines. If a trend is not significant, it is not plotted.
Returns a list containing results from the trend and changepoint analysis. This list has the following elements:
slope - Numeric vector containing the intercept and slope of the
prewhitened linear trend computed with zyp.trend.vector
using Yue Pilon's method
ci1 - numeric vector containing the intercept and slope of the
upper confidence bound. See confint.zyp
ci2 - numeric vector of length 2 containing the intercept and slope
of the lower confidence bound. See confint.zyp
pval - numeric value indicatng the significance value of the detected
trend, Kendall test computed within zyp.trend.vector
cpts - numeric vector of changepoints if any are found, computed
with cpt.meanvar
means - numeric vector of means computed with
cpt.meanvar
Jennifer Dierauer
See screen.summary to create a summary screening plot of
high flow, low flow, or baseflow metrics.
See metrics.all to calculate 30 different streamflow metrics at once.
The screen.metric function could then be used to loop through the metrics and
create an individual plot for each.
data(cania.sub.ts) # calculate and plot the annual maximum series res <- pk.max(cania.sub.ts) res1 <- screen.metric(res, ylabel="Q (m3/s)", text="Caniapiscau River, Annual Maximum Series") # calculate and plot the annual minimum series res <- MAMn(cania.sub.ts, n=1) res1 <- screen.metric(res, ylabel="Discharge (m3/s)", text="Caniapiscau River, Annual Minimum Series")data(cania.sub.ts) # calculate and plot the annual maximum series res <- pk.max(cania.sub.ts) res1 <- screen.metric(res, ylabel="Q (m3/s)", text="Caniapiscau River, Annual Maximum Series") # calculate and plot the annual minimum series res <- MAMn(cania.sub.ts, n=1) res1 <- screen.metric(res, ylabel="Discharge (m3/s)", text="Caniapiscau River, Annual Minimum Series")
Plots the daily streamflow time series and color codes points by data quality codes if the data are from Water Survey Canada. Also highlights date ranges with missing observations.
screen.series(TS, StnInfo = NULL, text = "d")screen.series(TS, StnInfo = NULL, text = "d")
TS |
output from |
StnInfo |
Optional data.frame containing user-supplied station info for plot.
data.frame must have 7 columns containing station info in the following order:
Station ID, Station Name, Prov/State, Country, Latitude, Longitude, Catchment Area
If any of the information is unavailabe, fill with NA. The Station ID column must
match the Station ID in column 1 of the data.frame input from |
text |
optional character string for margin text, e.g. for station name,
location, or other notes. Set to NULL if not margin text is wanted, or set to "d"
to use default text containing the station ID, station name, and province/state
returned from |
Jennifer Dierauer and Paul Whitfield
# load flow time series for the Caniapiscau River data(cania.sub.ts) # plot daily time series with default margin text screen.series(cania.sub.ts)# load flow time series for the Caniapiscau River data(cania.sub.ts) # plot daily time series with default margin text screen.series(cania.sub.ts)
Produces summary screening plots of high flow, low flow, or baseflow metrics. Each plot shows significant temporal trends and step changes. Intended for use as a data quality screening tool aimed at identifying streamflow records with anthropogenic impacts or data inhomogeneities.
screen.summary(metrics, type = "h", language = "English", StnInfo = NULL)screen.summary(metrics, type = "h", language = "English", StnInfo = NULL)
metrics |
output from |
type |
Character indicating the set of metrics to plot. Options are "h" for high flow metrics, "l" for low flow metrics, or "b" for baseflow metrics. |
language |
Language for plot labels. Choice of either "English" or "French". Default is "English". |
StnInfo |
Optional data.frame containing user-supplied station info for plot.
data.frame must have 7 columns containing station info in the following order:
Station ID, Station Name, Prov/State, Country, Latitude, Longitude, Catchment Area
If any of the information is unavailabe, fill with NA. The Station ID column must
match the Station ID in column 1 of the data.frame input from |
For the center of volume (COV) plots on the high flow and baseflow screening plots, the correlation coefficients for COV and years and for mean annual flow (MAF) and years are added to the plot. The ratio of the correlation coefficients (r COV-years / r COV-MAF) is included as a rudimentary indication of whether or not the temporal trend in COV is meaningful. See Whitfield (2013) for a discussion of COV.
Drought metrics for the low flow plot may not be applicable to intermittent streams, and plots will be empty in this case.
Important note: If "French" is the language wanted for the plot labels, the language
option must also be specified in metrics.all, as this plotting function
pulls the metric names from the output metrics.all output.
Jennifer Dierauer
Whitfield, P.H. 2013. Is 'Center of Volume' a robust indicator of changes in snowmelt timing? Hydrological Processes 27:2691-8.
# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) # create a summary flow screening plot of the high flow metrics screen.summary(caniapiscau.res, type="l")# load results from metrics.all function for the Caniapiscau River data(caniapiscau.res) # create a summary flow screening plot of the high flow metrics screen.summary(caniapiscau.res, type="l")
Internal wrapper for creating trend and change-point summary plots
screen.summary.internal(input, mparam, type, ylabs, i, DataType, maf, Year1, YearEnd, hyrstart)screen.summary.internal(input, mparam, type, ylabs, i, DataType, maf, Year1, YearEnd, hyrstart)
input |
metric time series |
mparam |
trend and change point info |
type |
character indicating type of summary plot |
ylabs |
y axis labels |
i |
plot position |
DataType |
numeric indicating data type |
maf |
mean annual flow series |
Year1 |
start year of original time series |
YearEnd |
end year of original time series |
hyrstart |
numeric indicating month for start of the hydrologic year |
Jennifer Dierauer
Returns station information for Water Survey Canada or United States Geological Survey stream gauges.
station.info(TS, StnInfo = NULL, Plot = F, language = "English")station.info(TS, StnInfo = NULL, Plot = F, language = "English")
TS |
output from |
StnInfo |
Optional data.frame containing user-supplied station info for plot.
data.frame must have 7 columns containing station info in the following order:
Station ID, Station Name, Prov/State, Country, Latitude, Longitude, Catchment Area
If any of the information is unavailabe, fill with NA. The Station ID column must
match the Station ID in column 1 of the data.frame input from |
Plot |
Boolean indicating whether a plot of station information should be created.
Default is F. Plot is intended for use as the upper-left panel of the plot produced
by |
language |
Language for plotting when Plot = T. Choice of either "English" or "French". Default is "English". |
Returns a list of the following station information:
$StationID
$StnName
$Prov/State - Abbreviation for the province or state in which the station is located
$Country - Abbreviation for the country in which the station is located
$Lat - Latitude of the station
$Long - Longitude of the station
$Area - Catchment area, in square kilometers
$RHN - Boolean indicating whether the station is part of a reference hydrologic network
Jennifer Dierauer
data(cania.sub.ts) StnInfo <- station.info(cania.sub.ts)data(cania.sub.ts) StnInfo <- station.info(cania.sub.ts)
Add calendar year, month, and day of year columns
YMD.internal(TS)YMD.internal(TS)
TS |
Output from |
Returns a data.frame with year, month, and doy columns appended.
Jennifer Dierauer