R package for handling iMHEA hydrometeorological data
Install imhea from github:
library(devtools)
devtools::install_github("https://github.com/simonmoulds/imhea")Load the library:
library(imhea)
Import some data files included with the package:
iMHEA_Catchment_AREA = readr::read_csv(
system.file("extdata", "iMHEA_Data_Areas.csv", package = "imhea"),
show_col_types = FALSE
)
q1_raw <- readr::read_csv(
system.file("extdata", "LLO/iMHEA_LLO_01_HI_01_raw.csv", package = "imhea"),
show_col_types = FALSE
)
p1_raw = readr::read_csv(
system.file("extdata", "LLO/iMHEA_LLO_01_PO_01_raw.csv", package = "imhea"),
show_col_types = FALSE
)
p2_raw = readr::read_csv(
system.file("extdata", "LLO/iMHEA_LLO_01_PO_02_raw.csv", package = "imhea"),
show_col_types = FALSE
)
The package implements an object-oriented design using the S3 system. We define three classes: rain_gauge, stream_gauge and catchment. These extend the tsibble class by including additional metadata on catchment properties.
To begin with we convert the raw preciptation data to rain_gauge objects:
p1 <- p1_raw |> rain_gauge(id = "LLO_01_P0_01", event_units = "mm")
p2 <- p2_raw |> rain_gauge(id = "LLO_01_P0_02", event_units = "mm")
is_rain_gauge(p1)
is_rain_gauge(p2)
Convert raw streamflow data to stream_gauge object:
q1 <-
q1_raw |>
stream_gauge(
id = "LLO_01_HI_01",
discharge_units = "l/s",
level_column = "Level cm",
level_units = "cm"
)
Now we prepare the additional metadata needed to build a catchment object, which groups together the hydrometeorological data for a specific catchment. We define the catchment ID and retrieve the catchment area from the reference dataset we obtained above:
catchment_id <- "LLO_01"
catchment_area <-
iMHEA_Catchment_AREA |>
dplyr::filter(Catchment %in% catchment_id)
catchment_area <- catchment_area[1, 2, drop=TRUE]
Now we have everything we need to create a catchment object (this takes a few minutes):
x <- catchment(q1, p1, p2, id = catchment_id, area = units::set_units(catchment_area, km^2))
The catchment object contains various attributes which describe the catchment properties. We can look at some catchment indices:
indices(x)
We can use ggplot2 to make some plots:
library(ggplot2)
library(units)
library(tsibble)
p1 <- ggplot(data = daily(x), aes(x = Date, y = P)) +
geom_line()
p2 <- ggplot(data = daily(x), aes(x = Date, y = Q)) +
geom_line()
Use patchwork to join plots together:
library(patchwork)
p1 + p2 + plot_layout(ncol = 1, nrow = 2)
Plot monthly/annual summaries:
p1 <- ggplot(data = monthly(x), aes(x = Date, y = Q)) +
geom_bar(stat = "identity")
p2 <- ggplot(data = annual(x), aes(x = Date, y = Q)) +
geom_bar(stat = "identity")
p1 + p2 + plot_layout(ncol = 1, nrow = 2)
Plot the intensity duration and flow duration curves:
ggplot(data = idc(x), aes(x = as.numeric(D), y = Intensity)) +
geom_line() +
scale_x_continuous(trans = "log10") +
xlab("Duration [min]")
ggplot(data = fdc(x), aes(x = Exceedance_Pct, y = as.numeric(Q))) +
geom_line() +
scale_y_continuous(trans = "log10") +
ylab(expression(paste("Discharge [", m^3/s, "]")))
Plot baseflow:
d <- baseflow_data(x)[["UKIH"]]
d <- d %>% tidyr::pivot_longer(cols = starts_with("Q"), names_to = "name", values_to = "values")
ggplot(data = d, aes(x = Date, y = values, col = name)) +
geom_line()
Plot the monthly climatology:
d <- monthly(x) %>% dplyr::mutate(Month = lubridate::month(Date))
p1 <- ggplot(data = d, aes(x = factor(Month), y = Q)) +
geom_boxplot()
p2 <- ggplot(data = d, aes(x = factor(Month), y = P)) +
geom_boxplot()
p1 + p2 + plot_layout(ncol = 1, nrow = 2)
This package is a work in progress. There are some parts of the package that need further work:
- The package does not currently compute all the indices from the MATLAB scripts
- The package still needs a nice
printgeneric function to display the various indices - Some of the functions are rather slow, especially the internal function which interpolates precipitation data to regular intervals using cubic splines
I will be working to resolve these issues in the coming weeks. Other parts of the package may need changing depending on user needs. For example, I do not currently provide any specific plot functions because the ggplot2 methods are relatively easy to implement given the right data. I therefore focused on developing the catchment class so that it easy to retrieve various data summaries needed for plotting. However, I could develop functions to produce specific plots that are needed repeatedly.
It would also be helpful to know whether there are any aspects of the MATLAB scripts which are not used, so that the R package can be as lean as possible.
Simon Moulds ([email protected]).
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent