Ireland’s Median Rainfall in January 1850-2014

Introduction

The goal for this blog is to create a map of the median rain fall levels in the the month of January between the period 1850 and 2014. Three R packages were used in order to complete the map seen below; Thematic Maps (TMAP) for map creation and interactivity; Simple Features (SF) in order to convert table data into spatial data; and Tidyverse for readability and additional features like the pipe (%>%) operator. The blog provides a brief description and analysis of the data being used, the methods employed in order to create the map (Fig. 1) and a short discussion of the results and any spatial patterns identified.

Data

The data used for this blog was sourced from The Irish Climate Analysis and Research UnitS (ICARUS) research center. The rainfall.RData file contains two objects: a rain table and a stations table. The rain table is a collection of monthly rainfall levels for twenty-five different weather stations between the period of January 1850 and December 2014 (49,500 total entries). There are 5 unique variables which can be seen in the plot below.

Year	Month	Rainfall	Station
1850	Jan	169.0	Ardara
1851	Jan	236.4	Ardara
1852	Jan	249.7	Ardara
1853	Jan	209.1	Ardara
1854	Jan	188.5	Ardara
1855	Jan	32.3	Ardara

The stations table contains the geographic information for each of the 25 weather stations present in the rain table. There are a total 9 variables within the table. The only shared variable between rain and stations is the station name (Station), this variable will later act as a primary key between the two tables. Note: Given that the Latitude and Longitude as doubles they must later be converted to coordinates using the SF package.

Station	Elevation	Easting	Northing	Lat	Long	County	Abbreviation	Source
Athboy	87	270400	261700	53.60	-6.93	Meath	AB	Met Eireann
Foulksmills	71	284100	118400	52.30	-6.77	Wexford	F	Met Eireann
Mullingar	112	241780	247765	53.47	-7.37	Westmeath	M	Met Eireann
Portlaw	8	246600	115200	52.28	-7.31	Waterford	P	Met Eireann
Rathdrum	131	319700	186000	52.91	-6.22	Wicklow	RD	Met Eireann
Strokestown	49	194500	279100	53.75	-8.10	Roscommon	S	Met Eireann

Methodology

This section outlines the 5 steps taken in order to create Fig.1 later in blog.
Step 1: In order to calculate the median rainfall for January, the irrelevant months must be removed. To do this the rain table was filtered into a new R object entitled jan_rain.

#Filter only for the month of January.
rain %>% filter(Month == "Jan")-> jan_rain

Year	Month	Rainfall	Station
1850	Jan	169.0	Ardara
1851	Jan	236.4	Ardara
1852	Jan	249.7	Ardara
1853	Jan	209.1	Ardara
1854	Jan	188.5	Ardara
1855	Jan	32.3	Ardara

Step 2: Is to calculate the median rainfall in January for each of the 25 weather stations across the study period. To achieve this the filtered jan_rain table was piped into a function which groups all 4125 entries by the Station variable. Then, this is piped into a summary function paired with a median calculation which complies our grouped entries to a single row for each station and stores it in a new variable entitled Median. Finally the output of this function is writeen into a new R frame titled median_jan_rain. The code executed and it’s resulting table can be seen below:

jan_rain %>%
  group_by(Station)%>%
  summarise(Median = median(Rainfall, na.rm = TRUE)) -> median_jan_rain
  #Note: na.rm ignores any missing values.

Station	Median
Ardara	171.6
Armagh	75.0
Athboy	87.1
Belfast	102.1
Birr	77.5
Cappoquinn	147.4

Step 3: Is to join the median January rainfall levels to the [stations]{style=“color:teal} table as this is the table that contains the geographic information needed to plot them using TMAP. As mentioned previously, the Station variable can be used as a primary key between rain and stations tables. This also works for median_jan_rain as it has retained the same variables as rain aswell as the new calculated Median. A left join was chosen as the join type as the goal is to only join the new Median variable from med_jan_rain while also keeping the geographical information present in stations. The code for this operation is below:

stations %>% left_join(median_jan_rain, by = "Station") -> stations

Station	Elevation	Easting	Northing	Lat	Long	County	Abbreviation	Source	Median
Athboy	87	270400	261700	53.60	-6.93	Meath	AB	Met Eireann	87.1
Foulksmills	71	284100	118400	52.30	-6.77	Wexford	F	Met Eireann	105.7
Mullingar	112	241780	247765	53.47	-7.37	Westmeath	M	Met Eireann	80.6
Portlaw	8	246600	115200	52.28	-7.31	Waterford	P	Met Eireann	123.3
Rathdrum	131	319700	186000	52.91	-6.22	Wicklow	RD	Met Eireann	126.6
Strokestown	49	194500	279100	53.75	-8.10	Roscommon	S	Met Eireann	102.5

Step 4: Is to convert stations into a SF object in order to be able to use the coordinate values (Lat and Long) present within the stations table. The achieve this the SF conversion function was used to convert coordinates to a point geometry by setting the coords parameter to Long and Lat. The Coordinate Reference System (CRS) was set to WGS1984 (EPSG:4326) as standard. The output of which was stored in a new spatial object called sf_stations.

stations %>% st_as_sf(coords = c("Long", "Lat"), crs = 4326) -> sf_stations

Station	Elevation	Easting	Northing	County	Abbreviation	Source	Median	geometry
Athboy	87	270400	261700	Meath	AB	Met Eireann	87.1	POINT (-6.93 53.6)
Foulksmills	71	284100	118400	Wexford	F	Met Eireann	105.7	POINT (-6.77 52.3)
Mullingar	112	241780	247765	Westmeath	M	Met Eireann	80.6	POINT (-7.37 53.47)
Portlaw	8	246600	115200	Waterford	P	Met Eireann	123.3	POINT (-7.31 52.28)
Rathdrum	131	319700	186000	Wicklow	RD	Met Eireann	126.6	POINT (-6.22 52.91)
Strokestown	49	194500	279100	Roscommon	S	Met Eireann	102.5	POINT (-8.1 53.75)

Step 5: Is to construct the map using the TMAP package. The TMAP mode was set to view to make the map interactive rather than static. The process is as follows: The tm_shape function extracts the geometry from the sf_stations table. The tm_bubbles aesthetic was chosen as the weather stations are represented at single point. The fill parameter was set to correspond with the Median of each station. Equal Interval was chosen as the classification method so that each class bin was similar to one another in size. The number of breaks chosen was 6 in order to evenly distribute all values. A scale bar was also added in order to align with cartographic best practice. The resultant map can be seen in Fig.1 in the section below:

tmap_mode("view")
tm_shape(sf_stations) +
tm_bubbles(fill = "Median", 
  fill.scale = tm_scale_intervals(style = "equal", n = 6),
  fill.legend = tm_legend(title = "Median Rainfall (mm)")) +
tm_scalebar(position = c("left", "bottom"))

Results and Discussion

Fig 1: Map of Ireland’s Median Rainfall in January between 1850 and 2014. Source: ICARUS Research Center

A number of spatial patterns can be observed when analysing Fig. 1 above. For example coastal weather stations tend to recieve a higher median January rainfall particularly in the South. Furthermore, There there is a East and West divide between rainfall levels. There are two reasons for this, one is their proximity to the sea means a higher air moisture content is present meaning more rain. The other is that Ireland’s prevailing wind brings in coastal showers that may be further out from the South West to the mainland (Met Éireann). Additionally, further inland weather stations have a lower median January rainfall likely becuase they are further away from the coast. There are a number of outliers to this observation, in the East, Dublin Airport (63mm) and Phoenix Park (67.6) have some of the lowest median rainfall despite being on the coast of the Irish Sea. Furthermore, in the West, Shannon Airport (92.9mm) weather station is also lower than expected when comparing to the surrounding West coast stations. This is due to the rain shadow or the orographic effect. This is a climatic process in which mountain slopes facing the wind receive more rainfall than on downward slopes due to changes in air temperature due to a quick rise in elevation (Britannica 2023). This process can be seen Fig. 1 as the weather stations just South of Dublin Airport (Rathdrum (126.6mm)) and Shannon Airport (Killarney (177.7mm)) receive nearly double the rainfall as they are located on the South facing slopes of the Wicklow and MacGillyCuddy’s Reeks respectively.

Conclusion

In conclusion this short blog described the processes required to create an interactive showing the median rainfall in January in Ireland between the period of 1850 and 2014. The inital 4 steps were executed using the SF and Tidyverse package in order to prepare the data. The TMAP package was also used to plot Fig. 1 and embed it in the section above. The spatial patterns that were identified were the disparity between coastal and inland weather stations including any outliers. The factors that influence these patterns are orographic effect, proximity to the coast and prevailing winds. One limitation is that the use of the Leaflet R package could allow for further interactivity in Fig. 1, for example a plot could be coded to pop-up when the user clicks on a particular weather station.

Bibliography

Britannica. (2023). Orograhpic Precipitation. [online]. Available at: https://www.britannica.com/science/orographic-precipitation. [Accessed 13/01/2026]

Met Éireann. (n.a.). Wind. [online]. Available at: https://www.met.ie/climate/what-we-measure/wind. [Accessed 13/01/2026]

Pebesma, E. (2018). Simple Features for R: Standardized Support for Spatial Vector Data. The R Journal. Vol 10(1), pgs. 439-446. Available at: https://doi.org/10.32614/RJ-2018-009.

Tennekes, M. (2018). tmap: Thematic Maps in R. Journal of Statistical Software. Vol 84(6), pgs. 1–39. Available at: doi:10.18637/jss.v084.i06.

Wickham, H. Averick, M. Bryan, J. Chang, W. McGowan, LD. François, R. Grolemund, G. Hayes, A. Henry, L. Hester, J. Kuhn, M. Pedersen, TL. Miller, E. Bache, SM. Müller, K. Ooms, J. Robinson, D. Seidel, DP. Spinu, V. Takahashi, K. Vaughan, D. Wilke, C. Woo, K. Yutani, H. (2019). Welcome to the tidyverse. Journal of Open Source Software. Vol 4(43), 1686. Available at: doi:10.21105/joss.01686.