Reassessing Ireland’s Hottest Temperature Record

The highest currently recognised air temperature (33.3°C) ever recorded in the Republic of Ireland was logged at Kilkenny Castle in 1887. The original observational record however no longer exists. Given that Ireland is now the only country in 10 Europe to have a national heat record set in the 19th century, a reassessment of the verity of this record is both timely and valuable. The present analysis undertakes a fundamental reassessment of the plausibility of the 1887 temperature record using methods similar to those used to assess various weather extremes under WMO auspices over recent years. Specifically, we undertake an inter-station reassessment using sparse available records and make recourse to the new and improved 20CRv3 sparse-input reanalysis product. Neither surrounding available stations nor the reanalysis offer substantive support for the 15 Kilkenny record of 33.3°C being correct. Moreover, recent data rescue efforts have uncovered several earlier extreme values, one of which exceeds the Kilkenny value (33.5ᵒC on 16 th July 1876 recorded at the Phoenix Park). However, the sparsity of early observational networks, a distinct lack of synoptic support from 20CRv3 for many of the extreme heat values, and the fact that these measurements were obtained using non-standard exposures leads us to conclude that there is grossly insufficient evidence to support any of these 19 th Century extremes as robust national heat record candidates. Data from the early 20 th 20 Century onwards benefits from a denser network of stations undertaking measurements in a more standardised manner, many under the direct auspices of Met Éireann and its predecessors, adhering to WMO guidance and protocols. This enables more robust cross-checking of records. We argue that the Met Éireann recognised 20 th Century heat record from Boora in 1976 verifies as the most plausible robust national temperature record based upon the synoptic situation and comparisons with nearby neighbouring stations. This measurement of 32.5°C thus likely constitutes the highest reliably recorded temperature 25 measurement in the Republic of Ireland. Ultimately, the formal decision on any reassessment and reassignment of the national record rests with the national meteorological service, Met Éireann. lies between -2.5°C and 2.5°C with Kilkenny generally measuring slightly higher temperatures than Phoenix Park. The difference on 26 th June 1887 between the two sites at 9°C, lies entirely outside the distribution of modern inter-site behaviour. The highest recorded temperature difference in the modern era was 7.5°C, indicating the heat record to be an extreme anomaly. Most (95%) of differences between Kilkenny and Roches Point (2010 -2019) lie within a range between 0°C - 5°C, showing that here again Kilkenny generally measures higher temperatures than Roches Point. The temperature record of the 26 th of June 1887 was 6.6°C warmer and lies within the upper 200 2% of the differences. Similarly, the difference between Armagh and Kilkenny (2010 - 2018) shows that 90% of the data lies within -2.5°C and 3°C. Here, the temperature difference of 6.4°C also lies within the upper 2% showing that these temperature differences are unlikely to occur but not impossible.


Introduction and Context
The highest recognised shaded air temperature recorded in Ireland was observed at Kilkenny Castle on 26th June 1887 (Rohan, 30 1986). The reported temperature was observed from a "standard site," reading 33.3°C (Rohan, 1986:133). Unfortunately, the original records from Kilkenny Castle for the period are missing and unavailable for scrutiny (Mary Curley, pers comm).
Curiosity surrounds this temperature record as Ireland remains the only country in Europe to have an existing heat record from the 19th century, with almost all other European countries having had their heat records set much more recently (Table 1).
Given this, and the relative paucity of metadata supporting the Irish record, it is appropriate to re-examine this record using 35 modern investigative techniques. Several recent investigations have been undertaken under the auspices of the WMO which have either recognised (Queterlard et al., 2009, Courtney et al., 2012, Purevjav et al., 2015, Lang et al., 2016, Cerveny et al., 2017, Laska et al., 2018, Merlone et al., 2020, Peterson et al., 2020, Weidner et al., 2020 or occasionally removed and replaced (El Fadli et al., 2013) various global or regional records for meteorological phenomena including heat and cold records. The present analysis uses similar techniques to reassess the Kilkenny heat record. 40

Country
Temperature (

45
The national temperature record was logged at Kilkenny Castle using a Negretti and Zambia thermometer, a standard thermometer in use at the time. This thermometer was tested by the Meteorological Council in 1890 which showed the instrument required no corrections, thus reporting the true value (Report of the Meteorological Council, 1890). During the 1880s the Stevenson Screen (Stevenson, 1864) became the preferred standard screen to shield thermometers from sunlight and radiation. Dolores Gaffney, the collections officer at Kilkenny Castle, believes the temperature in question was recorded within 50 a Stevenson Screen (pers. comm.). The original Kilkenny Castle site was surveyed in 1890 but not considered a 'second order' (maintained by trained staff (UCAR, 2021)) station (RMS, 1890) and eventually closed in 1933 (Niall Dollard, pers. comm.).
The exact location of the weather station in Kilkenny Castle at the time of the record is unknown. Station siting is known to have substantial potential impacts upon the representativity of resulting measurements and modern guidance from WMO specifies criteria for siting (WMO, 2018). Kilkenny Castle is in the centre of Kilkenny Town beside the River Nore ( Figure  55 1). Given ample grounds it is possible that the instrument was well sited but, equally, there are many potential sitings which would today be considered to yield the possibility of biased records. Met Éireann explained that a station inspection report from approximately 1911 stated that Kilkenny Castle only kept the records for four years before destroying them (Mary Curley, pers. comm.). This both means that the original record from which the national temperature record arises is lost for perpetuity and that necessary metadata to understand the measurement context is also unavailable. Unfortunately, the timeseries of daily 60 recordings around the record heat event was not saved and so the single 'observation' of record heat must be analysed in isolation.
3 https://doi.org/10.5194/cp-2021-139 Preprint. Offaly. This reading came in the midst of the long heatwave and summer drought which affected the British and Irish Isles along with much of NW Europe (Stubbs, 1977, Noone et al., 2017. This station recorded data from 1950 to 2016 and was maintained by Met Éireann. Based upon available digital data, other notable heat events in the instrumental record are summarised in Table 2. These include a suite of new potential heat records arising from recent highly valuable data rescue 70 efforts as part of a collaboration between Met Éireann and National University of Ireland, Galway as part of a PhD project (Mateus et al., 2020). Notable heat extremes have occurred exclusively in meteorological high summerlate June through early August. Given the uncertainty associated with the Irish national high temperature record, together with its novelty in being the earliest national temperature record in Europe, the 2019-20 class of the MSc Climate Change at Maynooth University were set a group assignment to re-evalaute the record. Having introduced the national context and aspects of the recognised all-time and 20 th 80

Location
Century national heat records, as well as selected additional heat extremes in available records, the rest of the paper is structured as follows. Section 2 introduces key methodological approaches and results from past WMO extremes verification efforts which this exercise was designed to mimic. Section 3 goes on to apply these approaches to propose that the long-standing Kilkenny Castle national heat record should be rejected. Section 4 then considers additional candidates arising from Table 2, until alighting upon a recommendation for the most plausible highest reliable heat record in the available observations. Section 85 5 concludes.

Assessments and Certifications of Climatic Extremes
The WMO has instigated a formal process to verify and certify a broad range of extremes. A rapporteur, when advised of a candidate record extreme that is desired to be assessed convenes a team of domain area and regional experts to investigate the event and recommend either acceptance or rejection as a bona fide observation. If the event is rejected the team is asked to 90 recommend an alternative record if they can do so. Teams have been instigated for temperature (El Fadli et al., 2013, Laska et al., 2018, Merlone et al., 2020, Weidner et al., 2020, precipitation (Quetelard et al., 2009), winds (Courtney et al., 2012) and pressure (Purevjav et al., 2015) amongst other records. The WMO teams consider global or WMO region records and have not, to date, assessed national records.

95
Teams assess all aspects of the potential record including available metadata, data, the synoptic situation and, if possible, the instrumentation, to come to a conclusion. The extent of the investigation depends upon the nature and timing of the event being analysed. More recent events can benefit from more thorough analyses which may include instrument characterisation and site visits (Merlone et al, 2020). Whereas revisiting old records, as is the case herein, has to rely more on event characterisation and reference to nearby instrumental records (e.g. El Fadli et al., 2013). Increasingly, recourse is made to meteorological 100 reanalysis products, including sparse-input centennial scale reanalyses (Slivinski et al., 2019) that can provide information on the evolving synoptic meteorological situation surrounding the record event.
Several analyses of various meteorological records worldwide have been undertaken over recent years leading to their removal (WMO, 2021). Perhaps most famously, what was cited to be the world's highest recorded temperature for over 90 years was 105 invalidated following a careful WMO-team reappraisal. A temperature of 58°C was recorded at El Azizia (modern day Libya) but was disproved due to various issues brought to light after an in-depth investigation of the record. The issues surrounding the reading included problematic instrumentation, siting, observer interpretation and, notably in the context of the present study, that the record did not correspond to other nearby locations (El Fadli et al., 2013). The South African Weather Service (SAWS, 2019) reported a temperature of 50.1°C from the Vioolsdrif weather station on the 28th of November 2019 (Austral 110 summer). This reading exceeded all previous heat records for South Africa. However, the temperature sensor in the Vioolsdrif weather station was replaced two days prior to the record temperature and upon tracking the behaviour of the station over the following days, it was decided that the behaviour of the temperature sensor was questionable. This resulted in the 50.1 °C record being deemed invalid. A temperature of 42.9 °C was recorded at the Deelen weather station, in the Netherlands, on the 25 th of July 2019 (Schildkamp, 2019). The sensors at the Deelen station were inspected and any technical faults were ruled 115 out. However, another instrument on the same site, a few hundred metres away, didn't record the same high values, which  Table 2). The stations on the island of Ireland all consist of long-running sites that have been well maintained either by Met Éireann, its precursors or the Armagh Observatory. To augment the records on the island of Ireland recourse is also made to records from Sheffield in the UK. There are other British stations for the same time but they are even further from the Kilkenny site (e.g. Oxford, Kew Gardens). With the notable exception of a station situated south-east of Kilkenny, the stations provide a 135 reasonable geographical spread sufficient to infer spatial temperature gradient behaviour on the day of the record. Barrington also notes widespread crop failure across Ireland in 1887, while newspaper articles from the time indicate reduced harvests and crop failure throughout the country, industrial activity, particularly the Linen industry in Northern Ireland was also adversely affected (Noone et al., 2017). Although the main focus of Barrington (1888) is on precipitation deficits, he does also note the temperature "On Sunday, June 5th, the temperature rose, and for the remainder of the month we had a 155 combination of heat and drought, which lasted until July 10th. No record exists of such a hot and dry June in the south and southeast of Ireland". Barrington (1888) goes on to mention that "… in June and July 1887, the S. of Ireland suffered more from excess of heat than any part of England or Scotland." further cementing the plausibility that the Island of Ireland was experiencing extreme hot and dry conditions at the time of the Kilkenny record high. near continent. On 26 th June the high pressure collapses and is subsequently replaced by a building high pressure from the near Atlantic initially advecting cooler air from the north which becomes centred over Ireland by the end of the month. The evolving pressure situation is highly consistent with the timing and relative phasing of changes reported from each site in Figure 3. 170

175
On the basis of the synoptic situation ( Figure 4) and the contiguous station series (Figure 3) it is certain that the period was indeed a notable heatwave event that could, plausibly, be consistent with a record high temperature. However, such a synoptic situation is not uncommon in the summer season over Ireland. In isolation this evidence is insufficient to determine the validity of the record. When looking at the modern-day differences between Markree and Kilkenny 90% of the distribution lies between a 0°C -5°C temperature difference (Kilkenny warmer than Markee on average, Figure 5). The difference recorded between Markree and 190 Kilkenny on 26 th June 1887 was 9.9°C which is a substantial outlier. The bulk of of the differences (95%) between Kilkenny and Sheffield (2010-2017) lie between -5°C and 5°C (broader because of the far greater distance between the sites). Despite this increased dispersion of modern era differences, the difference on 26 th June 1887 of 9.4°C again indicates a substantial outlier relative to modern inter-site characteristics. Only a handful of higher differences have been reached within the recent era. For Kilkenny and Phoenix Park (2013-2019), 95% of the data lies between -2.5°C and 2.5°C with Kilkenny generally 195 measuring slightly higher temperatures than Phoenix Park. The difference on 26 th June 1887 between the two sites at 9°C, lies entirely outside the distribution of modern inter-site behaviour. The highest recorded temperature difference in the modern era was 7.5°C, indicating the heat record to be an extreme anomaly. Most (95%) of differences between Kilkenny and Roches Point (2010 -2019) lie within a range between 0°C -5°C, showing that here again Kilkenny generally measures higher temperatures than Roches Point. The temperature record of the 26 th of June 1887 was 6.6°C warmer and lies within the upper 200 2% of the differences. Similarly, the difference between Armagh and Kilkenny (2010 -2018) shows that 90% of the data lies within -2.5°C and 3°C. Here, the temperature difference of 6.4°C also lies within the upper 2% showing that these temperature differences are unlikely to occur but not impossible. The June 26 th 1887 difference is highly anomalous being in the extreme upper tail at Markee, Sheffield, Roches Point and Armagh and entirely outside the distribution for Phoenix Park. Differences overall are highly anomalous in a broad arc from 210 SSW through N to due E. While there is a change in circulation via advection of a northerly airmass this occurred on June 27 th (Figures 3 and 4). There is no active frontal system that may support a multi-degree thermal gradient between Kilkenny and its neighbours. The slack pressure gradient also does not support such a strong temperature gradient between Kilkenny and neighbouring sites. While many of the Irish stations are coastal, Birr is not and would have been less impacted by any sea breezes that may have set up than Kilkenny. Overall, therefore, while June 1887 undoubtedly was an unusually hot month and 215 there is strong support for a build-up of heat breaking on or around June 26 th /27 th , the very anomalous differences compared to modern behaviour argue strongly against the validity of the reported Kilkenny Castle record value. The question therefore is what value may constitute the 'true' Irish national heat record and when did it occur? Table 2 noted a number of other candidate dates and locations in which the reported temperature exceeded 32C. Unlike the Kilkenny Castle value, many of these observations have been made as part of long-term series that are digitally available.
These permit, in addition to the analyses performed for Kilkenny Castle, a consideration of the evolution of the station series around the event, including for some cases the hour-by-hour evolution of temperatures on the day. Also, at least for more recent events there is a denser neighbour series network available to perform the neighbour-based comparison and the 225 difference series can be calculated directly rather than via a replacement site as is the case in Kilkenny. Their recency also means there is better metadata, and the records are generally made in a more standardised and uniform manner consistent with WMO guidance on methods of observation (WMO, 2018). However, first it is necessary to consider the viability of the other records unearthed by Mateus et al. (2020) many of which pre-date the Kilkenny Castle record.

Early period of record high temperatures 230
Mateus et al. (2020) have digitised a range of early Irish meteorological stations and made these data available. In constructing Table 2 the majority of the remaining observations above 32C arise from this source, including an observation of 33.5C from Phoenix Park, Dublin in July of 1876 which is even warmer than the Kilenny Castle observations and a set of observations in the 1850s.

1850s candidates 235
The suite of observations made in the 1850s preceded the formation of the UK Met Office (at the time the Republic of Ireland had not gained independence from the UK). Metadata on who exactly took these observations is incomplete but suggests they were undertaken by military engineers to whom it is assumed some degree of training had been given. This period precedes efforts to shield instruments using a Stevenson Screen (Stevenson, 1864) and instruments may typically have been housed on north facing walls or placed in ventilated rooms. There is substantial literature pointing to potential biases in these early records 240 (Parker, 1994, Camuffo, 2002, Bohm et al., 2010, Trewin, 2010, Brunet et al., 2011. Modern-day comparisons either of original instruments and exposures (Bohm et al., 2010) or reconstructed instruments and exposures (Brunet et al., 2011) with standard Stevenson Screens highlight important potential biases in summertime daily maximum temperatures. These are particularly marked for mid-latitude north wall exposures that would, especially if not oriented to true north, catch significant solar radiation in mid-summer (Parker, 1994, Bohm et al., 2010. 245 A further concern over the validity of several of these 1850s high temperature records relates to their geographical situation.
A reading of 32.2C on Scattery Island, Clare, an Island off the west coast of less than 1km by 1km seems implausible when climatological sea surface temperatures are in the high teens. Markee in Sligo also appears multiple times. As a coastal location in the NW of Ireland it is difficult to envisage how temperatures as warm as 32C or higher could be attained on such a regular 250 13 https://doi.org/10.5194/cp-2021-139 Preprint. basis. They also are very much bunched around late June which may indicate the presence of radiative effects around the solstice upon this measurement series or something as simple as annual leave measurements being taken by another observer.
Kilrish, in Clare is similarly coastal and similar questions would pertain around how plausible such extreme heat was so close to the Atlantic Ocean. The same concerns pertain to Dunmore East which is on the Irish Sea coast of SE Ireland. The Royal College of Surgeons in Dublin is relatively close to the quay so, again, it is questionable whether such a warm temperature 255 could be attained even allowing for urban heat island effects.
The 1850s candidates also are sometimes associated with implausible synoptic situations according to 20CRv3 reanalysis reconstructions ( Figure 6). Even in the 1850s the available pressure constraint is sufficiently robust over NW Europe to provide a robust synoptic scale reconstruction. The three synoptic charts in 1851 (Top row, Figure 6) are associated with high pressure 260 and a slack flow of air from the continent which would, generally, be associated with climatologically hot conditions. Of note is that over 28 th /29 th of June, Kilrish, Markee and Dunmore East all reported temperatures in excess of 32C. Combined with the synoptic chart reconstructed from 20CRv3 there is little doubt that 28 th /29 th June 1851 was, indeed, a heatwave event across the island of Ireland. Newspapers at the time reflected these conditions reporting on the 'hot' and 'oppressive' weather that was occurring in Ireland (Freemans Journal, 1851;Leinster Express, 1851). 265 The remaining dates in the 1850s are associated synoptically with conditions that would tend to be climatically normal or even below normal temperatures. The 27/06/1852 chart has a low pressure to the north of Ireland with a slack north-westerly flow which would be an onshore wind at Markee, Sligo, and completely inconsistent with a temperature in excess of 32C. The 26/06/1853 event, again for a temperature at Markee, Sligo, is represented by a southwesterly flow, which is grossly 270 inconsistent with temperatures as high as the low 30s in this location. The situation on 29/06/1854 is of a low pressure centred over England with a northerly flow, again inconsistent with temperatures in excess of 32C at Markee. The slack pressure gradient on 03/08/1856 leads to potentially a slight easterly onshore wind inconsistent with a high temperature at the Royal College of Surgeons in Dublin.

275
Taken together, the uncertainty relating to early instrumental bias, climatological locations, and in many cases lack of substantive synoptic situation support mean that the 1850s values contained in Table 2 almost certainly should be precluded as robust candidates for record temperatures. That is not to say that, at least in some cases, the values may not be correct or at the very least indicative of, climatologically speaking, extreme heat. But Markee in particular, which occurs the most frequently, has values that are often not supported at all by the synoptic situation. 280 were housed in a Stevenson screen, the data may be considered to have a high level of accuracy and reability [sic]" (Irish Meteorological Service, 1983: 4). It is hence probable, but not certain, that the measurement was made by a thermometer housed in something other than a Stevenson Screen although exactly what is hard to ascertain precisely. The site is about 5km inland from Dublin Bay such that, in the absence of a sea-breeze, climatologically speaking, a temperature in the 30s could potentially be attained. The synoptic situation as reconstructed by 20CRv3 is supportive of very warm conditions with a high 295 pressure of 1025hPa centred over the British and Irish isles and very high 500hPa geopotential heights extending across Ireland from Northern Africa (Figure 8). Preceding days had had a gentle flow of air from the near continent. There is thus clear synoptic support for very high temperatures.
Having discounted the long-standing Kilkenny Castle national record in Section 3 and, in Section 4.1, ruled out as sufficiently 300 robust any values from the 1850s, this leaves the Phoenix Park temperature a full 1C warmer than any other remaining candidate value in Table 2. Most of these remaining candidate values are from far further inland where such high temperatures a priori would be more easily attainable. When analysed against other Dublin city temperature observations for the month of June 1876 (Figure 7) it can be seen that, while a similar temperature evolution is mostly followed throughout the month, the Phoenix Park station is consistently hotter than the other Dublin stations (except for the 3 rd and 26 th day). Dixon (1953) reported 305 a sudden and rapid temperature high just a few years earlier, on the 21 st July 1868; "the temperature in Phoenix Park one day during July shot up to 88.4 degrees F. (31.3C), a record which has yet to beaten". At this time, the Botanic Gardens similarly recorded a very hot July, however the temperature recorded on the 21 st of July is 81 degrees F (27.2C), several degrees below the Phoenix Park recording despite their close proximity.

310
It is clear from the 20CRv3 reconstruction and the surface temperature recording in Figure 7 that the synoptic situation in July 1876 supports very warm conditions, however the near constant systematic elevation of the temperature readings in Phoenix Park relative to other stations casts doubt on wether it was reporting the actual surface temperature at the time. Phoenix Park is in an elevated position so on lapse rate basis alone would be expected to record slightly lower and not higher temperatures than nearby sites. While the value at the Phoenix Park cannot be discounted there is equally no obvious way to sufficiently 315 robustly confirm the observation for it to be adequately verified as a national heat record.

Candidates since the beginning of the 20th Century
The remaining candidate events in which the reported temperature attained or exceeded 32C are all associated with 325 synoptically broadly similar conditions. High pressure is situated over or to the north / north east of the island of Ireland, often with weak advection of air from the near continent (Figure 9). The synoptic situations thus do not call into immediate question any of the remaining candidate values realism.  Boora is the hottest value of the remaining candidates at 32.5C, attained on 29/06/1976. Boora is situated in the centre of the island of Ireland, which means it is highly unlikely to suffer from marine influences which may cap diurnal maxima closer to the coast. The summer of 1976 was famous across the British and Irish Isles for its drought and hot summer following on from 335 18 dry months previously (Noone et al. 2017;O'Laoghog, 1979). Therefore, the partitioning of heat between sensible (temperature) and latent (evaporation) terms would have been climatologically skewed toward sensible heating. The value at Boora sits within a string of very warm days with many of the observations exceeding 30C.
The hot and dry conditions were widely reported in newspapers at the time; The Irish Press reported on 24 th June 1976 'The 340 Irish have been greeted by a heatwave which saw the temperature hover around the 88 degree (31.1°C) mark for a long time' (Redmond, 1976 (Cahill, 1976).
These elevated temperatures, experienced at Boora, were also experienced at the nearby stations which sit within tens of kms of the Boora site (Table 3). 350  A comparison to neighbouring sites benefits from the standardisation of meteorological observations during the 20 th Century and the much denser network of Met Éireann sites since the mid-20 th Century. Comparisons can now be made to sites at tens 355 of kilometres distance rather than some hundreds of kilometres. Commensurately the distribution of statistical differences in summertime maximum temperatures are considerably smaller. Repeating the methodology of the Kilkenny neighbour-based analysis but using Boora and nearby stations for a period of up to 20 years either side of the record, the observation in question sits well within the distributions of expected differences in daily maxima over summer ( Figure 10). In the modern-day differences between Boora and its four neighbouring stations -Mullingar, Clonsast, Derrygreenagh and Ballybrittas, most of 360 the inter-site distribution lies between ± 1.5°C temperature difference. On 29 th June 1976 the difference recorded between Boora and Mullingar was 3.7°C which falls within the upper 2% of its distribution. Differences between Boora and Clonsast are generally between -2°C and 2.5°C and the heat record difference on 29 th June 1976 was 1.7°C, falling within the upper 10%. For Boora and Derrygreenagh, most of the data lies between -1.5°C -2°C with Boora typically measuring slightly higher temperatures than Derrygreenagh. The difference on 29 th June 1976 between the two sites at 3.8°C, lies within the upper 2% 365 of the expected differences but has still been exceeded on several occassions. Lastly, Boora and Ballybrittas have very strong similarities in the recorded temperature with most of the differences between -2°C and 2°C. The June 29 th 1976 difference between these stations was 0.5°C completely within the distribution of modern inter-site behaviour. A combination of geographical location, the synoptic situation, antecedent conditions, neighbour comparisons and use of 375 standard modern-day meteorological observations leads us to conclude the Boora observation is likely to be a credible reading.

ID
We would thus conclude logically that this represents the highest reliably recorded temperature in the Republic of Ireland since observations began. If the observation were to be called into question any of several candidates between 32C and 32.3C since 1900 based upon the synoptic situation on each date would be viable alternatives, but we do not consider these in greater depth herein. 380

Conclusions
Ireland is now highly anomalous in having its recognised national heat record set in the 19 th Century. The record is from a long-closed site in Kilkenny Castle and the associated series is not available, presumed lost, as is the bulk of relevant metadata.
The Kilkenny record of 33.3C in June 1887 exceeds that reached since the beginning of the 20 th Century by 0.8C. Results from our inter-station reassessment and reanalysis comparisons leads us to question if this truly represents the highest reliably 385 measured national temperature for Ireland. Based upon the few contemporaneous station records available on the island of A search of candidate records since 1887 yields that the highest likely defensible observation is 32.5C recorded at Boora on 29 th June 1976. The synoptic situation is consistent with a significant heatwave, the series evolution is realistic and comparison with nearby (within 10s of kms) neighbouring stations does not suggest any substantial issues in the reported reading. We 395 would thus recommend that the Boora observation be recognised as the national heat record for the Republic of Ireland.
Ultimately, however, the official recognition of the national temperature record rests with Met Éireann as the national meteorological service.
Finally, this analysis only considered the national all-time heat record. Met Éireann maintains all time annual and monthly 400 records, several of which date back to similarly early, pre-standardisation, measurements (Met Éireann, 2021). Most notably the all time cold record is -19.1°C, recorded in Markee, Sligo in January 1881 which is close to the Atlantic coast in NW Ireland. Reassessment of these records using a range of techniques such as those used here would likely be valuable. It is probable that Ireland is far from alone in having records that may benefit from a reassessment using these new techniques piorneered by WMO record assessment teams. 405