(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Irish surface water response to the 2018 drought [1]

['Devin F. Smith', 'School Of Earth Sciences', 'The Ohio State University', 'Columbus', 'Ohio', 'United States Of America', 'W. Berry Lyons', 'Tiernan Henry', 'Earth', 'Ocean Sciences']

Date: 2023-12

Intense weather events are projected to increase as a consequence of climate change. The summer 2018 drought in Europe impacted human health, ecosystems, and economic prosperity. Even locations with an abundance of fresh water, like Ireland, faced water restrictions due to depleted supplies. To characterize the effect of the 2018 drought on Irish rivers, we collected surface water samples from rivers across the island at the drought onset and termination. We analyzed samples for stable water isotopes δ 18 O and δ 2 H and calculated the fraction of evaporation from river groundwater and precipitation inflow (E/I) of rivers. We extended river δ 18 O and δ 2 H analysis to 2020 for rivers in two catchments, Corrib and Shannon, to investigate how Irish river systems respond to high precipitation events, and the role of loughs (lakes) in the system. River δ 18 O and δ 2 H values showed progressive depletion from west to east in response to precipitation depletion from airmasses arriving off the Atlantic Ocean. From onset to termination of the 2018 drought, river δ 18 O and δ 2 H values were enriched and the calculated E/I value increased for most rivers. D-excess were negatively correlated with E/I value, providing support for E/I calculations. Extended analysis of loughs along the Corrib and Shannon river systems showed that lough Corrib consistently induced isotopic enrichment, while loughs in the Shannon catchment inconsistently caused isotopic enrichment. Both systems exert control over river isotopic composition in hydrologic extremes. Findings promote additional research in hydrologic patterns in response to increasing frequency of floods and droughts.

Funding: Sample collection was conducted while WBL was supported Fulbright-GSI Fellowship at University of Galway. Funds through a Faculty Professional Leave from The Ohio State University College of Arts and Sciences also supported WBL.DFS was partially supported by internal fellowships at The Ohio State University. Funding for instrument purchase came from NSF GEO EAR/IF 0930016 and 1707989 (AEC, WBL). The authors received no specific funding for this work. The funders had no role in study design, data collection and analysis, decision to publish or preparation in the manuscript.

Data Availability: All river δ18O and δ2H values are reported in this study. Upon publication content from S1 and S2 Tables will be available for open access through the Consortium of Universities for the Advancement of Hydrologic Science, Inc. HydroShare ( https://www.hydroshare.org/ ). The data can be accessed upon manuscript publication at https://doi.org/10.4211/hs.aa3a731c51f44e4aba637e36e802117e . Data will also be uploaded to waterisotopes.org database. Data Further description of evaporation fraction calculation is provided in Diamond and Jack (2018).

The goal of this research is to evaluate the change in river water isotopic composition in relation to surface water deficit (supply < demand) and calculate potential evaporative water loss during the 2018 drought in Ireland. Sampled rivers were used to document shifts in Irish river isotopic composition at the onset and termination of the 2018 drought. We hypothesized that river water isotopic composition would not significantly change from the onset to termination of the 2018 drought from an evaporative effect. Objectives to test this hypothesis for this research were to (1) identify differences in Irish river δ 18 O and δ 2 H values and (2) calculate the evaporative fraction for Irish rivers in May and August 2018. We also present a δ 18 O and δ 2 H dataset from March 2018 –March 2020 for rivers in the Corrib and Shannon catchments to identify the control of loughs (lakes) on Irish surface water systems in drought and flood conditions. Rivers were sampled after extratropical storm Lorenzo passed over Ireland in October 2019 and after one of the wettest winter months on record in March 2020. We hypothesized that rivers draining large loughs in the Corrib and Shannon surface water systems would have enriched isotopic composition from evaporative enrichment.

Stable water isotopes, δ 18 O and δ 2 H, are a powerful natural hydrologic tracers that can be used to trace water source and investigate hydrologic response to seasonal meteorological changes or events. We utilize stable water isotopes to investigate surface water response to the 2018 drought. Precipitation amount and temperature have strong control on precipitation event δ 18 O values, and North Atlantic Oscillation (NAO) has strong control on monthly precipitation δ 18 O values [ 8 , 22 ]. Previous work conducted by Diefendorf & Patterson [ 23 ] presented a snapshot of Irish surface water isotopic composition in 2005, and showed that the rivers are a reflection of local precipitation. More recent work by Regan et al. [ 24 ] supported these results by showing that groundwater δ 18 O and δ 2 H values are also a reflection of local precipitation, but are influenced by orographic lifting, precipitation volume, biased winter recharge, and hydrogeologic setting. Global-scale modelling has shown that the isotopic composition of precipitation in Ireland is ~6‰ for δ 18 O and ~40‰ for δ 2 H [ 25 ]. These studies provide a foundation for the work presented herein.

In 2018, the continent of Europe experienced extreme drought conditions, with the most severe precipitation deficits and elevated temperatures occurring in northern and central Europe [ 16 ]. In Ireland, precipitation deficits and heat waves were as extreme, and the conditions were categorized as a meteorological drought [ 17 – 19 ]. These conditions were consistent in the UK, which led to excess water demand that over extended supply [ 20 ]. Surface waters provide 80% of drinking water in the Republic of Ireland [ 21 ], and elevated water demand coincided with extreme low river discharge, which raises the question: how did surface waters respond to compounding environmental stress and human demand in the 2018 drought?

The island of Ireland, which includes the Republic of Ireland and Northern Ireland and will hereon be referred to as Ireland, can be classified as a “water rich” island, as it receives ~700 and >1200 mm precipitation annually. It is an energy-limited environment with no distinct wet or dry season, however precipitation amount is weighted toward winter months, and there is a precipitation a gradient that decreases from west to east [ 8 , 9 ]. Despite annually consistent patterns and abundant precipitation, three centuries of droughts and multi-season precipitation deficits in Ireland have been reconstructed [ 10 – 13 ]. Research shows that Ireland will have a higher incidence of summer droughts and winter floods due to more uneven delivery of annual precipitation [ 14 , 15 ], threatening water security in Ireland.

As a result of future climate change, the incidence of extreme heat and drought around the world will increase [ 1 ]. Droughts have negative effects on terrestrial and aquatic communities, food and water resources, human health, socio-economic prosperity, infrastructure stability, and alter biogeochemical processes [ 2 , 3 ]. Up to 50% of the world’s regions, both arid and humid, could experience more frequent drought conditions by 2050 according to modelled water balance changes [ 2 ]. However, the increase extent and severity is uncertain because it is difficult to quantify the temporal and spatial extent of drought [ 4 , 5 ]. The onset of a drought can be slow, on the order of months to seasons, and the spatial distribution depends upon local climate and antecedent conditions [ 3 ]. Alternatively, flash droughts can occur, where the onset of drought is quick, on the order of weeks, caused by precipitation deficits and high evapotranspiration [ 6 ]. Because of the uncertainty in quantifying drought there is no universal definition for the term. Instead, droughts are classified by precipitation deficits from atmospheric evaporative demand indices (meteorological drought), soil moisture indices (agricultural drought), and hydrological indices (hydrological drought) [ 3 , 7 ]. Less attention has been paid to drought in regions that are typically thought of as “water-rich”. The occurrence of water scarcity is not only determined by the total amount of water delivery, but it is also dependent on natural and human population water demand. Since drought is amplified by human demand on water resources, “water rich” countries can experience water scarcity in drought conditions even though the quantity of water delivered is greater than arid regions.

Discharge data were available for 24 of the sampled rivers. Data were obtained from the Office of Public Works (OPW) Hydro-data Archive [ 31 ] and the Irish Environmental Protection Agency (EPA) Hydronet [ 49 ]. Discharge data for Athlone were calculated from a rating curve data generated from OPW discharge and gauge height data. Fitted rating curve flows may deviate from flow measurement by 10–20% and high values are determined by extrapolation of the curve (personal communication, OPW Hydrometric Section). Discharge percentiles were calculated and reported for the 24 rivers with discharge measurements provided by the Office of Public Works, with the exception of the river Shannon at Dowra [ 31 ].

An evaporative fraction , also referred to as the E/I values, was calculated for all rivers sampled in both May and August 2018. Calculations were modelled after Diamond and Jack [ 43 ], which are founded on the fundamentals of Craig and Gordon [ 44 ]. The fraction of water evaporated compared to groundwater and precipitation inflow water (E/I) was calculated with equations detailed in S3 Text [ 43 – 48 ]. The analytical propagated error for E/I calculations was 8.3% of reported E/I values. Additional uncertainties were included to calculate a total propagated uncertainty for May and August E/I values with Valentia and Armagh precipitation: relative humidity, temperature, groundwater δ 18 O values [ 48 ], and GNIP 2018 precipitation δ 18 O values. Uncertainty for May E/I values was 32% of the E/I reported values for Valentia and Armagh precipitation inputs. Uncertainty for August E/I values was 81% (Valentia) and 55% (Armagh) of the reported E/I values. Propagated error for calculated August E/I values was greater because the July 2018 precipitation δ 18 O values were more enriched or depleted compared to July averages at Valentia (-4.05‰) and Armagh (-6.79‰) [ 27 ].

Maps showing the distribution of river water δ 18 O and δ 2 H composition (isoscapes) were generated for Ireland in May and August 2018. Isoscapes were created for the Corrib catchment in March 2019, October 2019, and March 2020 and for the Shannon catchment in March 2018, June 2018, October 2019, and March 2020 were also constructed. Isoscapes were generated at 100 m 2 resolution with a minimum-curvature spline methods with barriers with a smoothing factor of 0.1 in ArcMap 10.6.1 [ 42 ].

Samples were analysed on a Picarro Wavelength Scanned-Cavity Ring Down Spectroscopy Analyzer Model L1102-I for δ 18 O and δ 2 H. A sample aliquot of 2 mL was used for analysis, where seven injections of 2.3 μL were made per sample. The first three injections were discarded to avoid memory effects and the last four injections were averaged to obtain the sample δ 18 O and δ 2 H raw values. The sample δ 18 O and δ 2 H were corrected by internal laboratory standards that had been calibrated to VSMOW (δ 18 O = 0‰, δ 2 H = 0‰) at the Institute of Arctic and Alpine Research (INSTAAR) at University of Colorado at Boulder by a dual inlet mass spectrometer. Internal laboratory standards were Colorado (δ 18 O = -16.53‰; δ 2 H = -126.3‰), Nevada (δ 18 O = -14.20‰; δ 2 H = 104.80‰), Ohio (δ 18 O = -8.99‰; δ 2 H = -61.80‰), and Florida (δ 18 O = -2.09‰; δ 2 H = -9.69‰). Two internal standards were run at the beginning and end of each sample run and after every fifth sample. Sample duplicates were run for every tenth sample. Instrument precision was determined by Picarro precision tests, where 150 deionized (DI) water samples were run to calculate instrument precision. The precision was 0.016‰ for δ 18 O and 0.15‰ δ 2 H. Accuracy for samples was determined with sample duplicates and values were ≤0.83‰ and ≤2.5‰ δ 18 O and δ 2 H, apart from one run where the δ 2 H precision of March 2019 samples is ≤3.9‰. Deuterium excess (d-excess) values (d-excess = δ 2 H − 8 × δ 18 O) were calculated for all samples. Local evaporation line was calculated for all river sample locations draining major loughs in the Corrib and Shannon catchments [ 37 ] ( Fig 2 ). Sample populations from each sampling campaign were not normally distributed and a Kruskal-Wallis test [ 38 ] was used in MATLAB 2019b to identify statistical differences among sample collection groups. All data are reported in S1 Table .

River sample locations from this study and groundwater δ 18 O sample locations from Regan et al. [ 24 ]. Samples collected in the Corrib and Shannon catchments are displayed as yellow circles. Ireland synoptic weather stations and Northern Ireland weather stations with historical data are shown as green diamonds [ 19 , 26 ]. GNIP δ 18 O and δ 2 H data were collected and measured at Valentia and Armagh, which are labelled on the map [ 27 ]. Lough (lake) names in the Corrib and Shannon catchments are noted next to the loughs in blue. All projected spatial data, including surface waters, catchments, and island boundary, were sourced from the Irish Environmental Protection Agency (EPA) and downloaded from EPA Geoportal (gis.epa.ie/GetData/Download) (CC-BY 4.0; https://creativecommons.org/licenses/by/4.0/legalcode ).

Surface water samples (n = 148) were collected from 40 rivers during 2018–2020 ( Fig 2 ). Rivers across Ireland were sampled in the summer (May–August) of 2018, while additional sampling was conducted in the Corrib and Shannon catchments in March 2018, March and October 2019, and March 2020 ( S1 Table ). Detail of Corrib and Shannon river systems and catchment characteristics can be found in S1 Text [ 32 – 36 ]. Not all locations were sampled in each sampling campaign. Samples were collected by hand in 20mL HDPE scintillation vials with urea caps or with a high-density polyethylene (HDPE) sampler. Samples collected in the sampler were transferred to 20 mL HDPE scintillation vials for transport and storage. Samples had no headspace to prevent evaporation and were shipped back to The Ohio State University, Columbus Ohio for analysis and analysed within a week of arrival. River samples were taken from public access points and no permits or approval were required for this research. Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the supporting information ( S1 Checklist ).

February 2020 was one of the wettest winter months on record [ 19 ]. Three extratropical cyclones pass over Ireland, and precipitation totals were 155 to 332% of 1981–2010 year long–term average [ 32 ]. In response to a high volume of precipitation in February, river flows were between 10 percentile (Q10) and 1–percentile (Q1) values and flooding was widespread across the island [ 31 ].

Met Éireann (the Irish Meteorological Service) has identified three classes of drought: (1) a dry spell is a period of +15 consecutive days with less than 1 mm of precipitation; (2) an absolute drought is a period of +15 consecutive days with less than 0.2 mm precipitation; and (3) a partial drought is a period of +29 consecutive days with <0.2 mm precipitation per day [ 30 ]. In summer 2018 heat wave conditions were recorded at 15 weather stations between June 24 and July 4. From May 22 to July 14, 21 stations recorded absolute droughts conditions. From May 28 to July 25, ten stations recorded partial drought conditions and five stations recorded dry spell conditions between June 18 and July 14 [ 17 ].

Precipitation amounts from May, June, and July 2018 are represented as pie charts for Irish synoptic weather stations and Northern Ireland weather stations [ 19 , 26 ]. Numbers in boxes next to each pie chart show the total precipitation amount (mm) from May, June, and July 2018. The color of the box corresponds to the month when most precipitation was measured, as shown on the pie chart. The table shows the percent of summer (May, June, July) precipitation measured at each station as the percent long–term average. Long–term average is classified as 30-year average (1981–2010) by Met Éireann Meteorological Services. LTA for Northern Irish weather stations was calculated over the same period with monthly total precipitation (mm). All projected spatial data, including surface waters, catchments, and island boundary, were sourced from the Irish Environmental Protection Agency (EPA) and downloaded from EPA Geoportal (gis.epa.ie/GetData/Download) (CC-BY 4.0; https://creativecommons.org/licenses/by/4.0/legalcode ).

In mid-May two compounding atmospheric circulation patterns (1) positive summer North Atlantic Oscillation and (2) northern hemisphere high amplitude Rossby Waves, called Wave–7 pattern, formed anticyclonic conditions over Ireland [ 28 , 29 ], causing prolonged dry weather. Few storms passed over the island but the timing of precipitation was not evenly distributed throughout the island. June consistently had the lowest monthly cumulative amount of precipitation measured at stations across the island ( Fig 1 ). Ireland received rainfall again towards the end of July that resulted in the termination of the summer drought [ 17 ].

Additional rivers were sampled in the Corrib (n = 2) and Shannon (n = 3) catchments in 2019 and 2020 ( S1 Table ), which provided more information for isoscape generation. Samples were collected below two large loughs in the Corrib catchment and 3 large loughs in the Shannon catchment ( Fig 2 ). River Corrib consistently had the most enriched δ 18 O and δ 2 H values in the Corrib catchment ( Fig 7 , S3 Fig ). October 2019 sampling in the Corrib and Shannon catchments occurred after extratropical Storm Lorenzo, and elevated river discharge ranged between Q50 and Q25 ( Fig 3 ). Rivers had the largest range in isotopic composition during this sampling campaign. River δ 18 O and δ 2 H values varied from -4.60‰, -28.38‰ (River Owenriff) to -7.61‰, -50.53‰ (Shannon Pot) ( Fig 7 , S1 Table ). The western Corrib catchment rivers had enriched δ 18 O and δ 2 H values compared to east catchment rivers. River Shannon also had enriched δ 18 O and δ 2 H values below the three loughs compared to tributary δ 18 O and δ 2 H values (Figs 2 and 7 , S3 Fig ). These isotopic patterns were consistent in samples collected in March 2020 during extremely high flows (Q10–Q1) ( Fig 3 ) and after one of the wettest winter months since 1850 [ 19 , 52 ].

Isoscapes were created for March 2018 (n = 10), May 2018 (n = 9), June 2018 (n = 9), August 2018 (n = 13), March 2019 (n = 6), October 2019 (n = 23), and March 2020 (n = 23). The color gradient is the same as Fig 5 and indicates a shift from enriched to depleted δ 18 O values. All corresponding δ 2 H values (‰) are shown in S3 Fig . Dashed lines and corresponding numbers show contour intervals of river water δ 18 O (1‰). Corrib and Shannon catchment rivers are displayed in black. All projected spatial data, including surface waters, catchments, and island boundary, were sourced from the Irish Environmental Protection Agency (EPA) and downloaded from EPA Geoportal (gis.epa.ie/GetData/Download) (CC-BY 4.0; https://creativecommons.org/licenses/by/4.0/legalcode ).

Isoscapes were also generated for the Corrib and Shannon catchments for additional sampling campaigns in 2018, 2019, and 2020 ( Fig 7 , S3 Fig ) to show the extent of change in 2018 and identify the role that loughs play in low and high flow events. In early March 2018, Shannon catchment rivers were sampled after unusual snow fall in February [ 51 ]. All sampled rivers, including locations below loughs, had δ 18 O and δ 2 H values within 0.18‰ and 5.46‰ of February precipitation mean δ 18 O and δ 2 H values (6.61‰, -36.4‰) ( Fig 7 , S3 Fig ), showing that surface water systems reflected the isotopic composition precipitation [ 27 ].

In May, the rivers Corrib and Annascaul (southwest) had the greatest evaporative fraction (0.05). May E/I values were negligible across the rest of southern Ireland. The E/I values in the Corrib catchment rivers (0.02–0.05), Shannon tributaries (0.02), and River Bann (0.03) resulted in a band of elevated E/I values from west to northeast ( Fig 6A ). The band of elevated E/I values from west to northeast was not observed in August, but E/I values along the west coast increased and extended toward the southeast of Ireland ( Fig 6B ). Rivers in western Ireland and in the Shannon catchment had greater E/I values in August than in May. In August, the River Corrib had the greatest calculated E/I value (0.10), and negligible E/I values were calculated for locations in the north and southeast (≤0.01) ( Fig 6B , S4 Table ). The There was a notable E/I value increase in the southwest from May to August ( Fig 6A and 6B ). The E/I value was greatest for the River Corrib and the River Shannon above Lough Ree in both May and August. A significant (p<0.05) negative relationship between d-excess and E/I values calculated from average groundwater δ 18 O values (r = -0.7) from May to August provided robust support for E/I value calculations despite uncertainty associated with the E/I calculations ( S3 Text ).

Maps show interpolated Irish river E/I values [ 43 ] of river water in (a) May and (b) August 2018. Average drainage basin groundwater δ 18 O values were used as inflow values for surface water prior to evaporation [ 24 ]. Values are reported in S4 Table . The color gradient of red to blue show a shift from a greater E/I values (more evaporation) to lower E/I values (less evaporation). Dashed lines are contour intervals for E/I intervals of 0.01. E/I values ≤0.01 were considered negligible. All projected spatial data, including surface waters, catchments, and island boundary, were sourced from the Irish Environmental Protection Agency (EPA) and downloaded from EPA Geoportal (gis.epa.ie/GetData/Download) (CC-BY 4.0; https://creativecommons.org/licenses/by/4.0/legalcode ).

The calculated evaporation fraction (evaporation/inflow, E/I) of rivers increased between the onset of the drought in late May and the termination of the drought in August ( Fig 6A and 6B ). The E/I values were greatest when the minimum (depleted) groundwater δ 18 O compositions were used as the river inflow value ( S3 Text , Eq. 2). Conversely, E/I values were lowest when the maximum (enriched) groundwater δ 18 O composition were used for inflow value ( S1 Fig , S4 Table ) [ 24 ]. Rivers with E/I ≤ 0.01 were considered to have a negligible evaporative fraction of water. The E/I values groups calculated from minimum and maximum groundwater δ 18 O values varied in statistical difference ( S5 Table ). The E/I values calculated for May and August with average groundwater δ 18 O inflow values were significantly different (p<0.05) ( S5 Table ). Average groundwater δ 18 O values were used to calculate E/I value shown in Fig 6 . Ranges of maximum, minimum, and average groundwater δ 18 O inflow values [ 24 ] used to calculate E/I values are shown in S2 Fig . Hereon, results presented in Fig 6 are discussed.

In August, River Corrib had the most enriched δ 18 O and δ 2 H values (-3.56‰, -25.26‰) and the lowest d-excess value (3.2‰). The upper Shannon also had enriched δ 18 O and δ 2 H values (-4.31‰, -29.42‰) and a low d-excess (5.0‰) ( Fig 5 ). The River Bann, located in northeast Ireland, had the most depleted δ 18 O and δ 2 H values (-7.34‰, -49.60‰) ( Fig 5 ). Overall, there was enrichment in river δ 18 O and δ 2 H values from May to August ( Fig 5 ). The d-excess values for August ranged from 3.2–11.8‰ ( S1 Table ), where lower d-excess values were correlated with the enriched δ 18 O and δ 2 H sample values.

There was less spatial variation in river δ 18 O and δ 2 H in May than August. In May, the most enriched δ 18 O and δ 2 H values were -5.08, 30.27 ‰ (southwest) and the most depleted values were -7.14‰, -47.35‰ (east) ( Fig 5 ). The range of southwest river isotopic composition (δ 18 O = -5 to -6‰; δ 2 H = -33 to -40‰) was similar to the May 2018 GNIP precipitation sample that was reported at Valentia (-4.65‰, -32.00‰) [ 27 ] ( Fig 5 ). In May, rivers in central Ireland had δ 18 O values of -6‰ to -7‰ and δ 2 H values of -40‰ to -45‰ and shifted further east in August ( Fig 5 ). River d-excess values ranged from 7.5‰ to 12.1‰ with the majority of river d-excess values between 9‰ to 11‰ ( S1 Table ).

The isoscapes generated from May and August sampling campaigns showed a pattern of δ 18 O and δ 2 H depletion from west to east ( Fig 5 ). The samples collected along each river system reflect the mixed signature of all water upstream of the sample point. Samples collected from rivers in the north and northeast reflected May 2018 GNIP precipitation δ 18 O and δ 2 H values that were reported at Armagh (-7.16‰, -50.9‰), while rivers in the south and southwest reflected δ 18 O and δ 2 H values recorded at Valentia (-4.65‰, -32.00‰) [ 27 ]. These data follow the general pattern of prevailing winds and decreasing precipitation amount across the island ( Fig 5 ). In August, the same pattern was observed but southwest rivers were depleted compared to July GNIP precipitation (-2.70‰, -13.60‰; Valentia) and northeast rivers were enriched (-8.21‰, -56.1‰; Armagh). River water isotopic composition showed inconsistent insignificant relationships with aquifer vulnerability and soil drainage ( S4 Text ).

River samples fall on or above the LMWLs. The distribution of river sample δ 18 O values and groundwater δ 18 O were significantly different (p<0.05) than GNIP precipitation data [ 27 ]. River sample δ 18 O values and groundwater δ 18 O were not significantly different. River data plotted in isospace show that river samples reflect historical precipitation isotopic composition despite statistical differences in the datasets ( Fig 4 ). The surface water line is δ 2 H = 6.06 × δ 18 O – 1.42 and the local evaporation line is δ 2 H = 5.58 × δ 18 O – 4.47 ( Fig 4 ). Statistical differences of δ 18 O and δ 2 H among sample groups were inconsistent ( S3 Table ). The range of δ 18 O and δ 2 H data for most sampling campaign datasets overlapped, particularly for campaigns with a greater number of samples collected ( S2 Fig ). The majority of data from summer months plot below the GMWL and fell into the same range of summer 2003 data (δ 18 O = -7.4 to -2.4‰ δ 2 H = -53 to -17‰) [ 23 ], but the most enriched sample values in summer 2018 were -5.08‰, -30.27‰ in May and -3.56‰, -25.26‰ in August ( S1 Table , Fig 5 ). The most depleted δ 18 O and δ 2 H values of rivers were measured in March and October ( Fig 4 , S2 Fig ).

(a) Samples collected 2018–2020 plotted with the global meteoric water line (GMWL) [δ 2 H = 8 × δ 18 O + 10] GNIP Armagh [δ 2 H = 7.49(±0.14) × δ 18 O + 5.38(±1.10)] and Valentia [δ 2 H = 7.01(±0.09) × δ 18 O + 2.99(±0.5)] local meteoric water lines (LMWL) in Northern Ireland and the Republic of Ireland, respectively [ 27 ]. A calculated local evaporation line [δ 2 H = 5.58 × δ 18 O ‒ 4.47] from sample locations draining major loughs in the Corrib and Shannon catchments is presented. (b) River Corrib samples collected 2018–2020 plotted with GMWL and LMWLs. Samples collected in 2018 are shown as circles, sample collected in 2019 are shown as squares, and samples collected in 2020 are shown as diamonds.

Local meteoric water lines (LMWLs), which define precipitation isotopic composition in relation to water sources, local geographic, and topographic variables have been published by the IAEA at Armagh [δ 2 H = 7.49(±0.14) × δ 18 O + 5.38(±1.10)] (R 2 = 0.97) and Valentia [δ 2 H = 7.01(±0.09) × δ 18 O + 2.99(±0.5)] (R 2 = 0.91) [ 27 ] (Figs 2 and 4 ). Both LMWLs were calculated with a least squares regression methods [ 27 ]. The Armagh LMWL was generated from 51 samples collected monthly since 2012 to 2019 and has a standard error of 2.6‰. The Valentia LMLW has a standard error of 3.4‰, and it is generated from 542 samples collected monthly from 1960 to 2018, with consistent monthly collection since 1977 [ 27 ].

Data were obtained from the Office of Public Works [ 31 ] and EPA Hydronet [ 49 ]. (a) Discharge data for 2018 presented for major Irish rivers outside of the Corrib and Shannon catchment with low percentile flow Q99, Q90 and Q75 values displayed in red, blue, and yellow lines. Vertical orange bars show dates of sampling campaigns. (b) Discharge data for 2018–2020 for Corrib and Shannon catchment rivers with Q10 and Q1 flow values displayed in purple and pink in addition to the Q99, Q90, Q75 discharge thresholds. River Athlone only shows Q1 flows. Vertical orange bars show dates of campaigns when rivers were sampled. Gaps in hydrograph indicate missing data. Discharge measurements displayed as a grey line indicate discharge data was estimated.

River flows began to decline in April 2018 and continued to fall their lowest point in late June and July ( Fig 3 ) [ 31 ]. All rivers had at least 26 days of average daily discharge rates below 75 percentile (Q75) flow values. Thirteen rivers had more than 90 days of average daily discharge below Q75. All the rivers had more than 10 days of average daily discharge lower than Q90, except for Rivers Suck and Slaney. The greatest number of days with average daily discharge values less than Q90 was the River Boyne. Rivers on the western side of the island, in the Corrib catchment and River Boyle, had the fewest days when discharge <Q90. River discharge of <Q75 is denoted as low flow rather than baseflow, which is defined as slow, consistent water input from multiple sources that sustain river flow between water–input events [ 50 ]. Flow values <Q75 are referred to as low flows instead of baseflow because several precipitation events occurred across the island between May and August. Rivers had notable low flows throughout the drought period, but events with small precipitation volumes may have contributed to river flow ( Fig 3 ).

Discussion

[END]
---
[1] Url: https://journals.plos.org/water/article?id=10.1371/journal.pwat.0000197

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.