Introduction

There are two functions available in clifro to create requisite cfStation objects to send queries to retrieve data via clifro. The first one is to search for stations using the cf_find_station function as detailed in the choose stations vignette. The other function that creates cfStation objects is the cf_station function that requires comma separated agent numbers as the only input. This vignette covers the construction of a cfStation object via the cf_station function, and then shows examples of plotting and visualising the station’s locations using KML files or within R using the ggmap package.

Creating a cfStation object from agent numbers

This is the simplest method to create a cfStation object, simply supply the cf_station function the comma separated agent numbers. The following stations are (or were) located around Lake Tekapo in Canterbury, in the South Island of New Zealand:

  1. Coal (Ski Field)
  2. Macaulay (Mt Gerald)
  3. South Opua
  4. Mount John
  5. Lake Tekapo Ews
  6. Godley Peaks
  7. Lilybank
lake.tekapo.st = cf_station(12709, 35567, 39557, 4630, 24945, 4616, 4602)
lake.tekapo.st[, c("name", "agent", "start", "end", "open")]
##                      name agent      start        end  open
## 1         Coal @ Skifield 12709 1989-02-01 2016-01-15  TRUE
## 2      Macaulay@Mt Gerald 35567 1990-07-04 2016-01-15  TRUE
## 3         Lake Tekapo Ews 24945 2003-06-18 2016-01-15  TRUE
## 4 South Opua @ South Opua 39557 2011-09-28 2016-01-15  TRUE
## 5        Lilybank Station  4602 1950-01-01 1992-09-30 FALSE
## 6                 Mt John  4630 1962-10-01 1988-01-01 FALSE
## 7    Godley Peaks, Tekapo  4616 1914-01-01 1976-06-01 FALSE

We can see that subsetting lake.tekapo.st acts just like a data.frame object, although it is technically a cfStation object. All the usual data.frame methods work on cfStation objects to maximise usability.

Adding more stations

To add more stations to this list the addition sign is used. Any repeated stations are removed and the resulting list is ordered by the end dates first and then by the stations’ start dates.

added.stations.st = lake.tekapo.st + 
  cf_station() + 
  cf_find_station("lighthouse", status = "all")
added.stations.st[, c("name", "agent", "start", "end", "open")]
##                       name agent      start        end  open
## 1              Reefton Ews  3925 1960-08-01 2016-01-15  TRUE
## 2          Coal @ Skifield 12709 1989-02-01 2016-01-15  TRUE
## 3       Macaulay@Mt Gerald 35567 1990-07-04 2016-01-15  TRUE
## 4          Lake Tekapo Ews 24945 2003-06-18 2016-01-15  TRUE
## 5  South Opua @ South Opua 39557 2011-09-28 2016-01-15  TRUE
## 6     Tiri Tiri Lighthouse  1401 1946-02-01 2016-01-15  TRUE
## 7         Lilybank Station  4602 1950-01-01 1992-09-30 FALSE
## 8                  Mt John  4630 1962-10-01 1988-01-01 FALSE
## 9    Cape Brett Lighthouse  1197 1934-11-01 1978-10-01 FALSE
## 10     Nugget Lighthouse B  5894 1975-03-01 1977-08-31 FALSE
## 11     Nugget Lighthouse A  5895 1975-03-01 1977-08-31 FALSE
## 12    Godley Peaks, Tekapo  4616 1914-01-01 1976-06-01 FALSE
## 13      Moeraki Lighthouse  5325 1935-10-01 1975-06-01 FALSE

The above code chunk adds the 7 stations around Lake Tekapo, the subscription-free reefton EWS station (cf_station()), and all stations located (currently or historically) on a lighthouse. Allowing multiple searches is not currently available using the web portal, CliFlo, but the above code demonstrates how easy it can be in clifro.

Visualising the station locations

CliFlo does not currently have any visualisation tools to aid in the selection of stations which can make the task of choosing geographically suitable stations a hard one.

Using KML files

The cf_save_kml functionality was introduced in the choose stations vignette and this function can be used on any cfStation object. To return a KML file showing all the stations within our added.stations.st object we just run cf_save_kml(added.stations.st) in R and the KML file is returned.

Showing station locations within R

Many useRs may prefer to plot the locations directly in R, or export the station information for use in another software. The as(object, "data.frame") function call is used for returning the station information as an R dataframe that can then be used for plotting or exporting.

This example is only one of the many ways to plot GIS data in R, and the ggmap package is chosen for it’s usability and quality plots. In this example we will plot all the past and present Auckland climate stations to assess the density of open and closed stations across the region.

# Conduct the search
auckland.st = cf_find_station("auckland", search = "region", status = "all")
library(ggmap)

# Add a column to colour the open and closed stations
auckland.st$colour = factor(auckland.st$open, labels = c("Closed", "Open"))

# Coerce to a data.frame and reverse the rows so the open stations get plotted 
# on top of the closed stations
auckland.df = as(auckland.st, "data.frame")[nrow(auckland.st):1, ]

# Obtain the map of the greater Auckland suitably scaled to fit the stations
auckland.map = ggmap(get_map("Auckland", maptype = "hybrid", zoom = 8))

# Plot the resulting map with the stations and station density
auckland.map %+% auckland.df + 
  stat_density2d(aes(colour = colour), alpha = .8) +
  geom_point(aes(colour = colour), alpha = .5) +
  scale_colour_discrete("Status", c("Closed", "Open")) +
  theme(legend.title = element_text(face = "bold"))


Remembering that closed stations in clifro are the ones that have end dates over four weeks ago. It appears that the open stations are reasonably confined to the city centre, whereas the more historic stations were placed further out in the rural areas.