Sam Albers

Ilaria Prosdocimi

Sam Zipper

### Importance of Hydrology

Given that liquid water is [essential to life on
Earth](https://www.amnh.org/explore/science-topics/water-and-life-on-earth),
water research cuts across numerous disciplines including hydrology,
meteorology, geography, climate science, engineering, ecology, and more.
Numerous R packages have emerged from this diversity of approaches, and
we recently gathered many of them into a new rOpenSci [task
view](https://github.com/ropensci/Hydrology) which we broadly titled
'Hydrology' and published to
[CRAN](https://CRAN.R-project.org/view=Hydrology). Our intent is to be
exhaustive and compile R packages to access, model, and summarise
information related to the movement of water across the Earth's
landscape. We hope to contribute to a [nascent community of hydrological
R users](https://www.hydrol-earth-syst-sci-discuss.net/hess-2019-50/)
and develop an infrastructure of packages that provide a comprehensive
toolkit for water practitioners who use R as their preferred
computational analysis tool. Making sense of water data is critical to
understanding the response of landscapes to a changing climate.
Consolidating water-related packages will promote their usage and
discovery and ultimately facilitate reproducible workflows for water
research. Since this is a new task view, it serves our purpose to
evaluate the current *State of Hydrology* in R and look at the
interdependency of hydrology packages relative to some better known
collections of packages in the general R ecosystem.

### State of Hydrology in R

We will conduct a high-level assessment of the state of hydrology in R
by visualizing network connectivity among package dependencies -- which
packages use code from another package to accomplish a task. For
context, we will evaluate our Hydrology task view against one of the
most well-known collection of packages in R, the
[tidyverse](https://www.tidyverse.org/), as well as the venerable
[Environmetrics task
view](https://CRAN.R-project.org/view=Environmetrics) which includes
many packages for environmental analysis. To do that we need to write
some R code and load our required packages:

![](https://media.giphy.com/media/xT8qBmSnYDXS21ZvHO/giphy.gif)

::: {.highlight}
``` {.chroma tabindex="0"}
library(tidyverse)library(tidygraph)library(ggraph)library(ctv)library(tools)library(cranlogs)
```
:::

We will use the handy `CRAN_package_db` function from the `tools`
package (part of base R) which conveniently grabs information from the
DESCRIPTION file of every package on CRAN and turns it into a dataframe.
Here we are only after a few columns so we will extract those right
away:

::: {.highlight}
``` {.chroma tabindex="0"}
all_cran_packages <- CRAN_package_db()[,c("Package", "Imports")] 
```
:::

We are also going to parse some of the info we receive from CRAN to make
it a little easier to work with:

::: {.highlight}
``` {.chroma tabindex="0"}
tidied_cran_imports <- as_tibble(all_cran_packages) %>%   separate_rows(Imports, sep = ",") %>%   mutate(Imports = str_replace(Imports,"\\s*\$[^\$]+\\)","")) %>%  mutate(Imports = str_trim(Imports, side = "both"))
```
:::

#### Package Connectivity

First, let's take a look at the tidyverse. We can take the rare step of
actually using a function from the tidyverse package (appropriately
named `tidyverse_packages`), which unsurprisingly identifies those
packages officially belonging to the tidyverse. We filter for those
packages and their imports, convert to `tbl_graph` and plot using
`ggraph` to ultimately visualize package connectivity:

::: {.highlight}
``` {.chroma tabindex="0"}
tidyverse_tbl <- tidied_cran_imports %>%   filter(Package %in% tidyverse_packages()) %>%  filter(Imports %in% tidyverse_packages()) %>%  as_tbl_graph() ggraph(tidyverse_tbl, layout = 'nicely') +   geom_edge_link(colour = "grey50") +   geom_node_text(aes(label = name), colour = "#F0F921FF", size = 6) 
```
:::

![](/img/blog-images/2019-04-02-hydrology-task-view/tidyverse.png)

As you might expect, we see many intersecting lines traversing in all
directions; many packages in the tidyverse import other packages in the
tidyverse. This is expected as tidyverse development is coordinated.

Next, let's have a look at the Environmetrics task view where we might
not expect to see the same degree of connectivity. The `ctv` package,
the infrastructure behind CRAN task views, (sort of) provides us with a
function to access a vector of package names for a given task view,
which we can use to similarly create a network plot:

::: {.highlight}
``` {.chroma tabindex="0"}
env_packages <- ctv:::.get_pkgs_from_ctv_or_repos(views = "Environmetrics") %>%   unlist(use.names = FALSE)env_tbl <- tidied_cran_imports %>%  filter(Package %in% env_packages) %>%  filter(Imports %in% env_packages) %>%  as_tbl_graph()ggraph(env_tbl, layout = 'nicely') +   geom_edge_link(colour = "grey50") +   geom_node_text(aes(label = name), colour = "#FDB32FFF", size = 6) 
```
:::

![](/img/blog-images/2019-04-02-hydrology-task-view/environmetrics.png)

While there are clearly fewer connections than in the tidyverse package
ecosystem, there is still quite a large amount of connectivity in
Environmetrics task view In particular the `mgcv`, `MASS`, `vegan` and
`zoo` packages all provide critical infrastructure for those developing
packages in this field. This presumably can at least partially be
attributed to the age (and therefore the maturity) of the Environmetrics
task view.

Finally, we can evaluate the new Hydrology task view by a similar means:

::: {.highlight}
``` {.chroma tabindex="0"}
hyd_packages <- ctv:::.get_pkgs_from_ctv_or_repos(views = "Hydrology") %>%   unlist(use.names = FALSE)hyd_tbl <- tidied_cran_imports %>%   filter(Package %in% hyd_packages) %>%   filter(Imports %in% hyd_packages) %>%   as_tbl_graph() ggraph(hyd_tbl, layout = 'nicely') +   geom_edge_link(colour = "grey50") +   geom_node_text(aes(label = name), colour = "#ED7953FF", size = 6)
```
:::

![](/img/blog-images/2019-04-02-hydrology-task-view/hydrology.png)

Clearly some connectivity among water-related packages is already
happening, which serves as a reminder that CRAN task views aren't the
only place people discover packages to use. Some tools form a hub of
connected packages because they share a common package maintainer (see
`hydroGOF`) while others like `CityWaterBalance` draw upon a diverse
suite of `Hydrology` packages to study urban water systems. Many
rOpenSci packages have been included in the task view
([`bomrang`](https://github.com/ropensci/bomrang),
[`clifro`](https://github.com/ropensci/clifro),
[`dbhydroR`](https://github.com/ropensci/dbhydroR),
[`GSODR`](https://github.com/ropensci/GSODR),
[`hddtools`](https://github.com/ropensci/hddtools),
[`hydroscoper`](https://github.com/ropensci/hydroscoper),
[`MODIStsp`](https://github.com/ropensci/MODIStsp),
[`rnoaa`](https://github.com/ropensci/rnoaa),
[`smapr`](https://github.com/ropensci/smapr),
[`tidyhydat`](https://github.com/ropensci/tidyhydat),
[`weathercan`](https://github.com/ropensci/weathercan)). `rnoaa` in
particular is a key piece of infrastructure that underpins many analyses
in hydrological research. The plot below summarises the number of
downloads from the RStudio CRAN mirror over the past year via the
`cranlogs` package.

::: {.highlight}
``` {.chroma tabindex="0"}
pkgcount <- cran_downloads(packages = hyd_packages,                            from = Sys.Date()-1*365,                            to = Sys.Date()) %>%  group_by(package) %>%  summarise(downloads = sum(count)) %>%  filter(downloads >= 5000)pkgcount %>%  ggplot(aes(x = reorder(package, downloads), y = downloads)) +  geom_col(aes(fill = downloads)) +  geom_text(aes(y = 1000, label = package), size = 5.5, colour = "white",             hjust = 0, nudge_y = -500, angle = 90) +  scale_fill_viridis_c(name = "Downloads", option = "D", end = 0.75) 
```
:::

It is illustrative that the top four packages feature two modelling
packages and two data packages. These are the basic tenets of
hydrological research and indicate how those in the hydrological space
are using R.

![](/img/blog-images/2019-04-02-hydrology-task-view/countDownloads-1.png)

### Future of R for Hydrology

Hydrology, like many other sciences, is becoming more and more data
driven and programmatic ways to access, manipulate and model data in a
reproducible workflow are becoming a necessity. Water related data
though has its own specific challenges, which are similar but different
from those of other data sources. It can be expected that with a growing
uptake of R in the hydrological community more packages will be
available, with more interconnectivity between them. At the moment the
landscape of R for hydrologists is quite fragmented, but in time it
might be that some *core* packages to manipulate data will emerge and
several computational approaches to hydrological models will be
centralised in a unique package. Regardless of how the landscape of
hydrology packages will evolve, R and open source tools are becoming
increasingly relevant in the hydrological community, and we are happy to
have made a small contribution to make it easier for everybody to know
what can already be done with water data in R.

If the Hydrology CRAN task view can connect developers and hydrologists,
we hope it can leverage the efforts of others in the hydrology space. A
topic in R can live in many places without detracting from others.
Twitter, Stack Overflow and GitHub are all obviously great places to
strengthen the community of water practitioners using R. Our task view
hopefully will join those spaces as a resource to enable shared efforts
towards water management. You can contribute to the task view by opening
an issue or pull request at our [project
repository](https://github.com/ropensci/Hydrology).


Importance of Hydrology   Given that liquid water is essential to life on Earth, water research cuts across numerous disciplines including hydrology, meteorology, geography, climate science, engineering, ecology, and more. Numerous R packages have emerged from this diversity of approaches, and we recently gathered many of them into a new rOpenSci task view which we broadly titled ‘Hydrology’ and published to CRAN.

Getting your toes wet in R: Hydrology, meteorology, and more

Claudia Vitolo

I've worked for over 12 years in hydrology and natural hazard modelling
and one of the things that still fascinates me is the variety of factors
that come into play in trying to predict phenomena such as river floods.
From local observations of meteorological and hydrological variables and
their spatio-temporal patterns to the type and condition of soils and
vegetation/land use as well as the geometry and state of river channels
and engineering structures affecting the flow.

This large number of well known predictors directly translates in the
need to collect a wide variety of information such as in-situ sensor
observations, river survey data, satellite imagery and, more recently,
social media information (reporting in real-time the evolution of
on-going events). In this field, data collection is challenging on many
levels: data are available from different providers under a variety of
licences, in various data types and formats that need to be handled and
homogenised.

On the one side, not much can be done to overcome licensing issues and
the learning curve to become an expert hydrologist is rather steep. On
the other side, the R community is working hard to provide tools to make
data standards more user-friendly and the convenience of data APIs
available to everyone, not only web developers. Here is where `hddtools`
comes into play! This R package is a proof of concept that hydrological
data can be made more accessible and consists of a collection of
functions to retrieve and homogenise hydrological information. Let me
walk you through the main functionalities!

## The R package `hddtools`

The name `hddtools` stands for **H**ydrological **D**ata **D**iscovery
**Tools**. This is an open source project designed to facilitate access
to on-line data sources. This typically implies the download of a
metadata catalogue, selection of information needed, formal request for
dataset(s), de-compression, conversion, manual filtering and parsing.
All those operation are made more efficient by re-usable functions.

Depending on the data license, functions can provide offline and/or
on-line modes. When redistribution is allowed, for instance, a copy of
the dataset is cached within the package and updated twice a year. This
is the fastest option and also allows offline use of functions. When
re-distribution is not allowed, only on-line mode is provided.

## Installation

The `hddtools` package and the examples in this blog depend on other
CRAN packages. Before attempting to install `hddtools`, solve any
missing dependencies using the commands below:

::: {.highlight}
``` {.chroma tabindex="0"}
packs <- c("zoo", "sp", "RCurl", "XML", "rnrfa", "Hmisc", "raster",           "stringr", "devtools", "leaflet")new_packages <- packs[!(packs %in% installed.packages()[,"Package"])]if(length(new_packages)) install.packages(new_packages)
```
:::

The package is available from the Comprehensive R Archive Network
([CRAN](https://cran.r-project.org/index.html)):

::: {.highlight}
``` {.chroma tabindex="0"}
install.packages("hddtools")
```
:::

However, the development of this package is very dynamic. Data sources
become often unavailable or migrate to new systems. URLs change and, as
a consequence, functions may stop working. For this reason, I always
suggest to use the development version of this package, which is
available from github using
[devtools](https://github.com/hadley/devtools):

::: {.highlight}
``` {.chroma tabindex="0"}
devtools::install_github("ropensci/hddtools")
```
:::

Load the `hddtools` package:

::: {.highlight}
``` {.chroma tabindex="0"}
library("hddtools")
```
:::

## Data sources and Functions

In my quest for hydrological data I found that there are tons of open
datasets available but the problem is that **you need to know where to
look!**

Here is the list of data sources available within the `hddtools`
package:

- The Koppen Climate Classification map
- The Global Runoff Data Centre
- NASA's Tropical Rainfall Measuring Mission
- The Top-Down modelling Working Group:
  - Data60UK
  - MOPEX-SEPA river level data

For each source, functions are available to obtain and/or filter
relevant data. Their usage is described below.

### The Koppen Climate Classification map

The Koppen Climate Classification is the most widely used system for
classifying the world's climates. Its categories are based on the annual
and monthly averages of temperature and precipitation. This climate
classification is used, in hydrological studies, to explain the
continental-scale variability inannual runoff. The `hddtools` package
contains a function to identify the updated Koppen-Greiger climate zone,
given a bounding box. In the example below I'm getting the climate class
for my beautiful home town in Italy: Pompeii.Country borders are
retrieved using the `getData()` function from the raster package, which
retrieves global administrative areas from the following website:
<http://www.gadm.org/>.

::: {.highlight}
``` {.chroma tabindex="0"}
# Plot Pompeii on a maplibrary(raster)Italy <- getData("GADM", country = "IT", level = 0)Pompeii <- SpatialPoints(coords = data.frame(x = 14.4870, y = 40.7510))plot(Italy, col = NA, border = "darkgrey")plot(Pompeii, add = TRUE, col = "red")# Define and plot a bounding box centred in Pompeii (Italy)areaBox <- raster::extent(Pompeii@coords[[1]] - 0.5, Pompeii@coords[[1]] + 0.5,                          Pompeii@coords[[2]] - 0.5, Pompeii@coords[[2]] + 0.5)plot(areaBox, add = TRUE, col = "red")
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-5-1.png)

::: {.highlight}
``` {.chroma tabindex="0"}
# Extract climate zones from Kottek's map:KGClimateClass(areaBox = areaBox, updatedBy = "Kottek")
```
:::

``` {tabindex="0"}
##   ID Class Frequency## 1 34   Csa       131
```

From the column `Class` in the above table, it can be derived that the
area falls in a warm temperate climate (C) with dry (s) and hot summers
(a). A description of the retrieved class and related criterion can be
printed setting the argument `verbose = TRUE` in the function
`KGClimateClass()`.

### The Global Runoff Data Centre

The Global Runoff Data Centre (GRDC) is an international archive hosted
by the Federal Institute of Hydrology in Koblenz, Germany. The Centre
operates under the auspices of the World Meteorological Organisation and
retains services and datasets for all the major rivers in the world. The
data catalogue, kml files and the Long-Term Mean Monthly Discharges are
open data and accessible via `hddtools`.

Information on all the GRDC stations can be retrieved using the function
`catalogueGRDC` with no input arguments, as in the example below:

::: {.highlight}
``` {.chroma tabindex="0"}
# GRDC full catalogueGRDC_catalogue_all <- catalogueGRDC()
```
:::

However, there are a number of options to filter the catalogue and
return only a subset of stations. For instance, the catalogue can be
filtered based on a geographical bounding box as in the function
`KGClimateClass()`. It might also be interesting to subset only stations
in a given country, e.g. Italy. As the 8th column in the catalogue lists
the country codes, the catalogue can be filtered passing two input
arguments to the `catalogueGRDC()` function:
`columnName = "country_code"` and `columnValue = "IT"`, as in the
example below.

::: {.highlight}
``` {.chroma tabindex="0"}
# Filter GRDC catalogue based on a country codeGRDC_IT <- catalogueGRDC(columnName = "country_code", columnValue = "IT")# Convert the table to a SpatialPointsDataFramehydro <- SpatialPointsDataFrame(coords = GRDC_IT[, c("long", "lat")],                                data = GRDC_IT)# Plot the stations on the mapplot(Italy, col = NA, border = "darkgrey")plot(hydro, add = TRUE, col = "blue", pch = 1)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-7-1.png)

The arguments `columnName` and `columnValue` can be used to filter over
other columns. For instance, the example below shows how to subset
stations along the Tanaro River which source in the Ligurian Alps.

::: {.highlight}
``` {.chroma tabindex="0"}
# Filter GRDC catalogue based on rivername, search is not case sensitive!Tanaro <- catalogueGRDC(columnName = "river", columnValue = "Tanaro")# Convert the table to a SpatialPointsDataFrameTanaroSP <- SpatialPointsDataFrame(coords = Tanaro[, c("long", "lat")],                                   data = Tanaro)# Highight in red the stations on the Tanaro Riverplot(Italy, col = NA, border = "darkgrey")plot(hydro, add = TRUE, col = "blue", pch = 1)plot(TanaroSP, add = TRUE, col = "red", pch = 1)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-8-1.png)

If `columnName` refers to a numeric field, in the `columnValue` other
terms of comparison can be specified. For example, it is straightforward
to find out that the country with the most longstanding monitoring
stations is Germany (with 11 out of 15 stations with more than 150 years
of recordings).

::: {.highlight}
``` {.chroma tabindex="0"}
# Filter GRDC stations with more than 150 years of recordingsGRDC150 <- catalogueGRDC(columnName = "d_yrs", columnValue = ">150")# Which country has the most longstanding monitoring stations?table(GRDC150$country_code)
```
:::

``` {tabindex="0"}
#### DE FI LT RO SE## 11  1  1  1  1
```

::: {.highlight}
``` {.chroma tabindex="0"}
# Where is the oldest station?oldest <- GRDC150[which(GRDC150$d_yrs == max(GRDC150$d_yrs)),                  c("grdc_no", "river", "station", "d_yrs")]oldest
```
:::

``` {tabindex="0"}
## # A tibble: 1 × 4##   grdc_no      river station d_yrs##     <chr>      <chr>   <chr> <dbl>## 1 6340120 ELBE RIVER DRESDEN   208
```

The oldest stastion is in Dresden on the Elbe River. Let's now find out
whether monthly data is available. For this we need the GRDC
identification number (this is stored in the column grdc_no of the
catalogue) and the function `tsGRDC()`.

::: {.highlight}
``` {.chroma tabindex="0"}
# Monthly data extractionDresdenStation <- tsGRDC(stationID = oldest$grdc_no, plotOption = TRUE)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-10-1.png)

The plot above, shows the time on the x-axis and the flow (in m^3^/s) on
the y-axis. The river flow has had huge oscillations over time, is this
an effect of anthropogenic changes and/or climate change? This is well
out of the scope of this blog post, I leave the reader to look at trends
from the mean monthly values `DresdenStation$mddPerYear`.

### NASA's Tropical Rainfall Measuring Mission

The Tropical Rainfall Measuring Mission (TRMM) is a joint mission
between NASA and the Japan Aerospace Exploration Agency (JAXA) that uses
a research satellite to measure precipitation within the tropics in
order to improve our understanding of climate and its variability.

The TRMM satellite records global historical rainfall estimation in a
gridded format since 1998 with a daily temporal resolution and a spatial
resolution of 0.25 degrees (spatial extent goes from -50 to +50 degrees
latitude). This information is openly available for educational purposes
and downloadable from an FTP server. The `hddtools` provides a function,
called TRMM(), to download and convert a selected portion of the TRMM
dataset into a raster-brick that can be opened in any GIS software.

As an example, I'm going to download precipitation maps for three months
in 2016 using the `areabox` defined previously to locate Pompeii and
surrounding areas. But remember, values become less reliable moving away
from the tropics!

::: {.highlight}
``` {.chroma tabindex="0"}
# Define a temporal extenttwindow <- seq(as.Date("2016-01-01"), as.Date("2016-03-31"), by = "months")# Retrieve mean monthly precipitations based on a bounding box and time extentTRMMfile <- TRMM(twindow = twindow, areaBox = areaBox)plot(TRMMfile)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-11-1.png)

Each of the above plots shows the average precipitation over Pompeii
during January (top left), February (top right) and March (bottom left)
2016. Precipitation seems generally more abundant in the south (coastal
areas) where the plots show prevalence of green, especially during
January and February.

### Top-Down modelling Working Group

The Top-Down modelling Working Group (TDWG) for the Prediction in
Ungauged Basins (PUB) Decade (2003-2012) is an initiative of the
International Association of Hydrological Sciences (IAHS) which
collected datasets for hydrological modelling free-of-charge, available
[here](http://tdwg.catchment.org/datasets.html). This package provides a
common interface to retrieve, browse and filter two datasets: Data60UK
and MOPEX.

#### The Data60UK dataset

The Data60UK initiative collated datasets of areal precipitation and
streamflow discharge across 61 gauging sites in England and Wales (UK).
The database was prepared from source databases for research purposes,
with the intention to make it re-usable. This is now available in the
public domain free of charge.

The `hddtools` contain two functions to interact with this database: one
to retrieve the catalogue and another to retrieve time series of areal
precipitation and streamflow discharge.

::: {.highlight}
``` {.chroma tabindex="0"}
UK <- getData("GADM", country = "GBR", level = 0)plot(UK, col = NA, border = "darkgrey")# Data60UK full catalogueallData60UK <- catalogueData60UK()hydroALL <- SpatialPointsDataFrame(coords = allData60UK[, c("Longitude",                                                         "Latitude")],                                data = allData60UK)plot(hydroALL, add = TRUE, col = "blue", pch = 1)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-12-1.png)

::: {.highlight}
``` {.chroma tabindex="0"}
head(allData60UK)
```
:::

``` {tabindex="0"}
##   stationID        River          Location gridReference Latitude## 1     22001       Coquet           Morwick      NU234044 55.33310## 2     22006        Blyth   Hartford Bridge      NZ242800 55.11381## 3     23006   South Tyne      Featherstone      NY672610 54.94257## 4     24004 Bedburn Beck           Bedburn      NZ117321 54.68382## 5     25005        Leven      Leven Bridge      NZ444120 54.50139## 6     25006        Greta Rutherford Bridge      NZ033122 54.50511##   Longitude## 1 -1.632691## 2 -1.622163## 3 -2.513539## 4 -1.820046## 5 -1.315911## 6 -1.950550
```

Given the station identification number (in the column stationID), time
series of areal precipitation and streamflow discharge can be downloaded
using the function `tsData60UK()`. In the example below I show how to
get the time series for the first station in the table, for a given
temporal window.

::: {.highlight}
``` {.chroma tabindex="0"}
# Extract time series for the first stationstationID <- catalogueData60UK()$stationID[1]# Extract time series for a specified temporal windowtwindow <- seq(as.Date("1988-01-01"), as.Date("1989-12-31"), by = "days")MorwickTSplot <- tsData60UK(stationID = stationID,                            plotOption = TRUE,                            twindow = twindow)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-13-1.png)

The above figure is divided into two plots: precipitation over time
(top) and river flow over time (bottom). High flows and related
precipitation events seem to be clearly identifiable. The baseflow is
around 10 m^3^/s, but during the most important events the flow reached
120 m^3^/s. The shape of the hydrograph, however, suggests that the flow
was contained by the embankments and did not cause floods.

#### MOPEX

This source contains historical hydrometeorological data and river basin
characteristics for hundreds of river basins and from a range of
climates in the US. As with the previous dataset, `hddtools` contains
functions to download the catalogue and time series. The example below
shows how to download the MOPEX catalogue and the time series for the
first station in the catalogue.

::: {.highlight}
``` {.chroma tabindex="0"}
# MOPEX full catalogueallMOPEX <- catalogueMOPEX()# Extract time seriesBroadRiver <- tsMOPEX(stationID = allMOPEX$stationID[1],                      plotOption = TRUE)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-14-1.png)

In the plots above P stands for precipitation, E for potential
evapotranspiration, Q for river flow, T~min~ for minimum temperature and
T~max~ for maximum temperature.

### SEPA river level data

The Scottish Environment Protection Agency (SEPA) manages river level
data for hundreds of gauging stations in the UK. The catalogue of
stations was derived from a list that's no longer available, [but
there's an equivalent dataset of river levels from the
SEPA](http://apps.sepa.org.uk/database/riverlevels/SEPA_River_Levels_Web.csv).
The time series of the last few days is available from the SEPA website
and can be downloaded using the following function `tsSEPA()`, as in the
example below. Plese note that this data is updated every 15 minutes and
the code will always generate different plots.

::: {.highlight}
``` {.chroma tabindex="0"}
# SEPA unofficial catalogueallSEPA <- catalogueSEPA()# Single time series extractionKilphedir <- tsSEPA(stationID = catalogueSEPA()$stationId[1],                    plotOption = TRUE)
```
:::

![](/assets/blog-images/2013-03-07-hddtools/unnamed-chunk-15-1.png)

## Example applications

There are a number of possible research applications for the
`hddtools`package. Retrieved precipitation and streamflow data could,
forinstance, be used to draw spatial trends, as done by (Vitolo et
al.,2016b) using the `rnrfa` package but over larger areas. The
packagecould also be used to compare hydrological behaviours in
different areasof the world, to run and calibrate hydrological models
such as fuse(Clark et al., 2008, Vitolo et al. 2016a) as well as to
undertakeregionalisation studies.

## Future developments

If you have suggestions, please add them to the
[issuetracker](https://github.com/ropensci/hddtools/issues) on github.
Also,feel free to contribute to the package sending a
[pullrequest](https://github.com/ropensci/hddtools/pulls), that would
begreatly appreciated!

### Acknowledgments

I'm very grateful to [Erin Le Dell](https://github.com/ledell)
and[Michael Sumner](https://github.com/mdsumner) who reviewed, on behalf
ofrOpenSci, the `hddtools` package and the related paper published in
theJournal of Open Source Software (Vitolo, 2017). Both reviewers
providedvery constructive suggestions that grately improved this
package. I'dlike to also thank [StefanieButland](/about/#staff) and
[ScottChamberlain](https://github.com/sckott) for providing invaluable
advicewhen reviewing this post.

### References

Clark, M. P., Slater, A. G., Rupp, D. E., Woods, R. A., Vrugt, J.
A.,Gupta, H. V., Wagener, T. and Hay, L. E.: Framework for
understandingstructural errors (fuse): A modular framework to diagnose
differencesbetween hydrological models, Water Resources Research,
44(12),
n/a--n/a,doi:[10.1029/2007WR006735](https://doi.org/10.1029/2007WR006735),
2008.

Vitolo, C.: Hddtools: Hydrological data discovery tools, The Journal
ofOpen Source Software,
2(9),doi:[10.21105/joss.00056](https://doi.org/10.21105/joss.00056),
2017.

Vitolo, C., Wells, P., Dobias, M. and Buytaert, W.: fuse: An R
packagefor ensemble Hydrological Modelling, The Journal of Open
SourceSoftware,
1(8),doi:[10.21105/joss.00052](https://doi.org/10.21105/joss.00052),
2016a.

Vitolo, C., Fry, M. and Buytaert, W.: rnrfa: An R package to
Retrieve,Filter and Visualize Data from the UK National River Flow
Archive, The RJournal, 8(2) \[online\] Available
from:<https://journal.r-project.org/archive/2016-2/vitolo-fry-buytaert.pdf>,2016b.


Framework for Understanding Structural Errors (FUSE): A modular framework to diagnose differences between hydrological models

hddtools: Hydrological Data Discovery Tools

fuse: An R package for ensemble Hydrological Modelling

None

I’ve worked for over 12 years in hydrology and natural hazard modelling and one of the things that still fascinates me is the variety of factors that come into play in trying to predict phenomena such as river floods.

Messaggi di Rogue Scholar

Getting your toes wet in R: Hydrology, meteorology, and more

Discover hydrological data using the hddtools R package