hydroutils

An R package for handling hydrology data and plotting hydrologic frequency curves. Forked from DSSRip to provide these functions in a cross-platform environment. This package should not require the links to DSSVue and should be useful for more general hydrologic analysis.

wateryear, similar to the lubridate package's year function, returns the water year for a timestamp (starting 01 Oct of the previous calendar year). wymonth returns the month of the water year, starting with October, and the wy.month.abb constant contains the month abreviations in this order. wyday returns the day of the wateryear, with 1 being Oct 1st and 356 or 366 being Sept 30th depending on if it is a leap year or not.

For assessing the fit of hydrologic models, the functions nash.sutcliffe, rmse, and excelR2 are provided. The nash.sutcliffe function supports an argument x.alt to replace the mean of the x.obs dataset with a different alternative model to provide more resolution in the resulting score. This can either be the best 'dumb' model output, where no calibration is required to get a reasonable answer.

weibullProbs returns the corresponding Weibull plotting position for a given vector to plot on a flow-frequency graphic.

hydro_flow_trans and hydro_prob_trans can be used with ggplot's scale_continuous for axes to automatically apply flowBreaks and probBreaks. These call the methods flowBreaks and probBreaks for generating breaks on a plot with logarithmic and normal-deviate axes, such as those used for flow frequency graphics.

Examples:

The example below has been superceeded by one in the examples directory which retrieves data from the USGS via their dataRetrieval package. If you do not want to install (on Windows) or cannot install (on Mac and Linux) the dssrip library, the DSS references the files below can be replaced with either the nile dataset or data obtained through the USGS's dataRetrieval package)*

Frequency curve

annual_peaks_data.dss contains a record with a USGS flow frequency record.

require(hydroutils)
require(dssrip) # or use your own data or the R example 'nile' dataset
require(ggplot2)
require(scales)

testfile = opendss("C:/path/to/annual_peaks_data.dss")
peaks = getFullTSC(testfile, getPaths(testfile, "C=FLOW-*")[1])
colnames(peaks) = "FLOW"
peaks = fortify(peaks) ## xts to data.frame for ggplot2

ggplot(peaks, aes(y=FLOW, x=weibullProbs(FLOW))) + geom_point() + 
  theme_bw(base_size=11) + theme(legend.position = "bottom", panel.grid.minor=element_blank()) +
  scale_y_continuous(trans=hydro_flow_trans()) + 
  scale_x_continuous(trans=hydro_prob_trans(lines=c(1,2,5), labels=c(1,2,5), byPeriod=TRUE)) + 
  stat_smooth(method="glm", family=gaussian(link="log"))

Quick monthly summary hydrograph

test.dss contains a daily flow record downloaded with the USGS Import tool in DSSVue for the Delaware River at Trenton, NJ (USGS #01463500). This isn't the best example of a summary hydrograph, and definitely not the best way to do it, but gives an idea of seasonality and creating a WY based plot with wymonth and wymonth.abb.

require(hydroutils)
require(dssrip)
require(ggplot2) # for plotting and fortify command
mydssfile = opendss("F:/test.dss")

delawareflows = getFullTSC(mydssfile, getPaths(mydssfile, "A=DELAWARE RIVER B=TRENTON NJ C=FLOW"))
plot(delawareflows)
## uses fortify.xts from ggplot, creates column for index.
deflow = fortify(delawareflows)
## Let's label months and assign them a WY order (October thru Sept)
deflow$MONTH = factor(x=wymonth(deflow$Index), labels=wymonth.abb)
## plot
ggplot(deflow, aes(x=MONTH, y=FLOW, group=MONTH)) + geom_boxplot()