1. Introduction
Analyses of the stable isotopes of oxygen (O) and hydrogen (H) in the hydrosphere can illuminate hydrologic connections between precipitation, groundwater, and surface water (Kendall and McDonnell 1998, and references therein; Gat 1996, and references therein). Meteoric waters consist of naturally occurring species of the water molecule (isotopologues), and recent advances in laser spectroscopy have opened new opportunities to characterize their occurrence in space and time. The distribution of these isotopologues in the hydrosphere is governed by kinetic and equilibrium fractionation processes (Dansgaard 1964; Craig 1961; Gat 1996), which result in spatial and seasonal patterns across the globe, including the continental effect, the rainout effect, the amount effect, and the altitude effect (Kendall and McDonnell 1998). These patterns can be used to constrain predictions of precipitation stable isotope composition for larger regions where rainfall sampling may not be feasible. Regional understanding of stable isotope composition in water can allow for the tracing of surface and groundwater back to source regions and thus illuminate water flow paths and residence times. In other words, the analysis of stable isotope values in precipitation can establish a chemical fingerprint that can be used to link water drawn from a well or stream to its atmospheric origin.
Naturally occurring stable isotopes of water have been used extensively to investigate hydrologic questions from the global or continent scale (van der Veer et al. 2008; Terzer et al. 2013) down to small watersheds (Scholl et al. 2002; Kendall and McDonnell 1998, and references therein). These studies generally focus on the ratios of heavy isotopes 18O and 2H (deuterium) to more abundant lighter isotopes of 16O and 1H. The abundances are expressed in delta-notation δ, which relates 18O:16O and 2H:1H ratios to that of standard mean ocean water (SMOW) following Craig (1961) and others. The δ18O and δ2H of meteoric waters are strongly correlated in precipitation globally and expressed as the global meteoric water line (GMWL), or δ2H = 8 δ18O + 10 (Dansgaard 1964). The δ18O and δ2H relationship also varies locally, and the establishment of a local meteoric water line (LMWL) usually precedes the application of stable isotope tracer methods (Scholl et al. 1995; Davis et al. 1970).
Another important dimension of isotope fractionation is captured by deuterium excess (d-excess), a second-order isotopic parameter quantifying a water sample’s excess 2H beyond the expected global 8:1 ratio of δ2H to δ18O, as d-excess = δ18H − 8 δ18O (Dansgaard 1964). The difference in mass-dependent diffusivities of 1H2H16O (molar mass = 19 g mol−1) and 1H218O (molar mass = 20 g mol−1) impacts kinetic fractionation processes and results in “excess” deuterium in the vapor above an evaporating surface (Kendall and McDonnell 1998). Precipitation d-excess is therefore strongly impacted by moisture source, sea surface relative humidity and temperature, and subcloud evaporation processes (Guan et al. 2013; Xia et al. 2022) and can vary strongly with precipitation altitude (i.e., pseudoaltitude effect; Natali et al. 2022; Bershaw 2018; Gonfiantini et al. 2001). Many studies have also documented seasonal patterns in the d-excess of precipitation, making it a useful tool in examining the seasonal dynamics of the water cycle (e.g., Natali et al. 2022; Dores et al. 2020; Booth et al. 2021; Guan et al. 2013).
The field of isotope hydrology has been dominated by research in continental settings (Bowen 2010) and global scales (Terzer et al. 2013; van der Veer et al. 2008). There is a growing body of research examining processes and spatial scales relevant to islands (Prada et al. 2015; Lanças et al. 2022; Sosa et al. 2011). Enhancing isotope hydrology’s relevance for oceanic islands, however, is crucial, as islands are particularly sensitive to hydrologic fluctuations associated with climate change (Veron et al. 2019) and are overwhelmingly reliant on groundwater for freshwater needs. The island of O‘ahu, Hawai‘i, is emblematic of these issues in several respects. Groundwater is a critical resource on O‘ahu, where it supplies more than 90% of freshwater needs and is replenished primarily from precipitation (Gingerich and Oki 2000; State of Hawai‘i 2019; Nichols et al. 1997). Urbanization and population growth are projected to increase the demand for freshwater on O‘ahu (State of Hawai‘i 2019) while simultaneously encroaching on critical groundwater recharge areas. O‘ahu is also experiencing a long-term drying trend (Frazier and Giambelluca 2017). Coupled with the island’s accessibility, diverse physiography, and dynamic climate, these factors make O‘ahu an ideal location for advancing isotope hydrology research for island-scale processes.
The Hawaiian Islands have a rich history of stable isotope hydrology research beginning in the 1950s (Friedman and Woodcock 1957). From 1963 to 1970, the International Atomic Energy Agency (IAEA) conducted a monthly precipitation sampling campaign in Hilo, Hawai‘i Island (IAEA/WMO 2023). In the 1990s, pioneering research led by Martha Scholl combined precipitation and groundwater stable isotope data to characterize groundwater flow paths, recharge areas, and residence times near the Kīlauea Volcano on the island of Hawai‘i (Scholl et al. 1995) and to differentiate fog and rainfall contributions to recharge in East Maui (Scholl et al. 2002, 2007). Findings from these works spurred a series of isotope hydrology studies on Hawai‘i Island (Tillman et al. 2014; Kelly and Glenn 2015; Fackrell et al. 2020; Tachera et al. 2021). All of these studies established reliable elevation stable isotope lapse rates to connect groundwater and surface water to source recharge areas on Maui and Hawai‘i Island.
Recent studies on O‘ahu (Dores et al. 2020; Booth et al. 2021; Torri et al. 2023) have advanced Hawai‘i isotope hydrology by emphasizing temporal variation of stable isotopes in precipitation. Dores et al. (2020) found that the drier leeward areas of O‘ahu received most precipitation from synoptic storms, which potentially accounted for disparities between precipitation and groundwater isotopic compositions. Studies sampling O‘ahu rainfall at event scale (Booth et al. 2021) and at weekly intervals (Torri et al. 2023) both identified systematic isotopic differences between orographic trade wind–derived rainfall and rainfall from synoptic, disturbance-based storms. While these studies established the influence of diverse rainfall-generating mechanisms on the isotopic composition of precipitation, and Torri et al. (2023) identified significant spatial variability, they fell short of explicitly characterizing this spatial variability, including the consideration of altitude effects.
These recent works indicate the need for more expansive and longer-term isotopic monitoring of precipitation to characterize the complex interaction of atmospheric and physiographic dynamics shaping O‘ahu’s isotopic landscape (i.e., isoscape; Bowen 2010). An accurate model of O‘ahu’s isoscape could be used to estimate the isotopic composition of precipitation in unmonitored or inaccessible sites, elucidate the connections between precipitation and groundwater, and advance research into the dominant atmospheric processes governing stable isotope compositions in precipitation.
Thus, the goals of this paper are to establish a more comprehensive model of O‘ahu’s water isoscape by 1) characterizing the spatial and temporal isotopic variability of precipitation on O‘ahu through a new high-density, multiyear precipitation sampling campaign. We use this same dataset to 2) evaluate the strength of the isotopic “amount effect” and 3) evaluate the isotopic altitude and pseudoaltitude effects by establishing reliable isotope lapse rates for δ18O and d-excess. Here, we present new stable isotope data from precipitation and high-elevation spring samples collected on O‘ahu, Hawai‘i, and synthesis of this new dataset with the body of published precipitation isotope data from the same region.
2. Methods and materials
a. Study region and climate
The study area is centered on the Pearl Harbor Sole Source Aquifer (PHSSA), which occupies the southern third of O‘ahu and encompasses the majority of the island’s urban sprawl (Fig. 1). Two steep volcanic ranges dominate the physiography: the older Wai‘anae Range in the west reaching approximately 1200 m MSL and the Ko‘olau Range in the east reaching 900 m MSL (Sherrod et al. 2021). These volcanic ranges are connected by a central saddle called the Schofield Plateau. Persistent northeasterly trade winds interact with O‘ahu’s mountainous terrain to create a rain shadow effect that divides the island into distinct windward and leeward regions. Northeast-facing windward slopes of the Ko‘olau Range receive an average of 3000 mm of rainfall a year and are characterized by deep amphitheater valleys with steep intervening ridgelines, abundant streams and waterfalls, and thick vegetation. The drier leeward regions in south-central and southwest O‘ahu generally receive less than 500 mm of rainfall a year and are characterized by semiarid scrub lands.
Precipitation collector locations on the island of O‘ahu sampled from March 2017 to August 2021. The boundaries of the PHSSA are outlined in blue. The 1000-mm rainfall isohyets are shown in gray. Fog zones are identified by the yellow highlighted areas above 600 m MSL. Collectors D3 and MT1 were collocated but sampled at different periods.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
Elevation gradients on O‘ahu are severe, and the topography exerts a strong influence on weather patterns and thus precipitation isotopic composition. Persistent northeasterly trade winds bring warm, humid air to the island, which is lifted and cooled as it passes over the Ko‘olau and Wai‘anae Ranges, lowering atmospheric vapor pressure and causing condensation (Wahl and Urey 1935). The trade winds persist approximately 90% of the time during the dry season (May–October) and 50% of the time during the wet season (November–April) (Giambelluca 1983) and generally bring precipitation with a δ18O signature in the 0‰ to −5‰ range (Torri et al. 2023; Booth et al. 2021; Dores et al. 2020). Trade wind–driven orographic precipitation is thought to constitute approximately 70% of O‘ahu’s total annual precipitation (Longman et al. 2021).
The remaining 30% of O‘ahu’s precipitation is generated by synoptic, disturbance-based weather patterns and tends to be highly depleted in heavy isotopes (Booth et al. 2021; Torri et al. 2023). Synoptic systems include cold fronts, upper-tropospheric troughs, Kona lows, and tropical cyclones and, excepting tropical cyclones, occur more frequently in the wet season (Torri et al. 2023) and are the major contributors to precipitation in leeward areas (Dores et al. 2020).
b. Precipitation monitoring sites and sample collection
Twenty precipitation collectors were deployed throughout the study region to assess stable isotope composition (Fig. 1; Table 1). Most collectors were sampled once every 2–3 months between March 2018 and August 2021. Precipitation collectors were of the same design used in previous Hawai‘i-based studies (Fackrell et al. 2020; Scholl et al. 1995, 2002; Dores et al. 2020; Booth et al. 2021; Tachera et al. 2021), consisting of a 19-L (5 gallon) high-density polyethylene (HPDE) bucket with a 76- or 110-mm Buchner funnel affixed to the lid (appendix A). A 2-cm-thick layer of high-purity mineral oil in each collector prevented the evaporation of the collected precipitation. Collectors were covered with black polyethylene bags for additional protection against ultraviolet light damage.
Precipitation collector location information (NAD83) and sampling period for the present study (3-month sampling frequency March 2018–August 2021) and previous studies, Dores et al. (2020) collectors D1–D20 (quarterly sampling, March 2017–July 2018), and Booth et al. (2021) collectors HH, NH, DH (monthly sampling, November 2017–February 2019). Mean annual precipitation (MAP) is taken from the Rainfall Atlas of Hawai’i (Frazier and Giambelluca 2017). Manana transect collectors are MT1, MT2, MT3, and MTS. Ka‘ala transect collectors are K1, K2, K3, and KS. Schofield Plateau collectors are SB, WAAF, ER, and KK.
We sampled three transects spanning elevational gradients with four collectors each to evaluate the variation in the stable isotope composition with altitude (Fig. 1). The first transect was deployed along the Manana Ridge trail within the ‘Ewa Forest and had collectors distributed between 291 m (MT1) and 806 m MSL (MTS). The second transect was deployed along Mount Ka‘ala Road and had collectors distributed between 290 m MSL (K1) and the summit of Mount Ka‘ala, which is the highest point on O‘ahu at 1212 m MSL (KS). A high-elevation perennial spring just below the summit of Mount Ka‘ala was also sampled from October 2019 to April 2021 (appendix B). A third transect was deployed along the Schofield Plateau, with four collectors distributed generally east to west, connecting the Wai‘anae Range to the foot of the Ko‘olau Range. Samples were collected in a triple-deionized-rinsed 500- or 60-mL HPDE bottle. The sample bottles were filled completely, forming a convex meniscus at the container mouth to minimize air bubbles and were refrigerated within 6 h of collection. To increase the sample point density, the precipitation stable isotope data we collected were combined with the data from Dores et al. (2020), and Booth et al. (2021) into one dataset, which we refer to hereafter as the combined dataset (Table 1). Dores et al. (2020) conducted quarterly sampling while Booth et al. (2021) sampled at monthly intervals. Booth et al. (2021) also collected 38 event-based precipitation samples (not included in the combined dataset for our analysis). Collectors beginning with letter “D” are from Dores et al. (2020). The three Honolulu collectors (HH, DH, and NH) are from Booth et al. (2021).
c. Isotopic analysis
Water samples were analyzed at one of two laboratories: the Isotope Biogeochemistry Laboratory at the University of Hawai‘i at Mānoa or the Stable Isotope Ratio Facility for Environmental Research (SIRFER) at the University of Utah. Both facilities determined the hydrogen and oxygen isotopic composition of water samples using a Picarro L2130-I cavity ring-down spectrometer equipped with a high-precision vaporizer (V1102-I, Picarro, Inc., Santa Clara, California) and autosampler (HTC PAL, Leap Technologies, Carrboro, North Carolina) with Chem-Correct acquisition software that monitors for the interference of isotopologues of water by organic compounds (Gupta et al. 2009). All measurements were performed in the nitrogen carrier mode, using ultrahigh-purity nitrogen (<10 ppm H2O, >99.99% N2; Matheson, Irving, Texas) and eight 1.2-μL injections. Analytical precision determined through duplicate sampling indicates the uncertainty of ±0.35‰ δ18O and ±1.32‰ δ2H. Propagated d-excess uncertainty is ±3.1‰ for all collectors.
d. Local meteoric water line and deuterium excess
An LMWL was derived from the combined dataset (present study; Dores et al. 2020; Booth et al. 2021) using two approaches: reduced major axis (RMA) and volume-weighted average reduced major axis (VWA-RMA), which utilized the precipitation volume-weighted average isotopic composition for each collector location, rather than individual sample values. These LMWLs were compared using three difference measures: root-mean-square error (RMSE), mean absolute error (MAE), and index of agreement (IA) following Willmott and Wicks (1980) and Willmott (1981, 1982a,b). The regression with the lowest RMSE and MAE and the highest IA was assumed to be the closest representation of O‘ahu’s LMWL. Seasonal LMWLs were also calculated from data subsets for the dry season (May–October) and wet season (November–April). Samples during collection intervals that included a month or more in both the dry and wet seasons were excluded from the seasonal LMWL analysis.
e. Statistical analysis
To investigate the potential effects of precipitation amount and precipitation altitude on isotopic variation, we analyzed both the combined dataset and data subsets based on the season (wet/dry) as described earlier and eight regional subsets based on the geographic location and potentially distinct climate regimes (see Table 1).
To test the amount effect, individual sample precipitation δ18O values were plotted against sample rainfall rates normalized as percentile scores between 0 and 1 following Torri et al. (2023). Correlation analyses were evaluated on the basis of the Spearman correlation coefficient and p value where subsets with correlation coefficients r ≥ 0.5 and p ≤ 0.05 were interpreted as non-negligible correlations. The intent of examining these subsets was to explore the amount effect, or the strength of the relationship between precipitation amount and stable isotope composition for certain regions and certain seasons.
We similarly assessed the altitude and pseudoaltitude effects by plotting precipitation VWA isotopic composition against collector elevation to develop isotope lapse rates. Precipitation volume-weighted averages were used because of our interest in the hydrologic connectivity between precipitation and groundwater, since precipitation amount and recharge volume tend to be strongly correlated on O‘ahu (Engott et al. 2017).
Last, the analysis of temperature controls on precipitation stable isotope composition was excluded in this study due to the study area’s narrow seasonal temperature variation and because of the stronger influence of precipitation source on stable isotope composition as compared to the season (Torri et al. 2017, 2023; Dores et al. 2020; Booth et al. 2021). We also found that seasonal temperature fluctuations were less pronounced than orographic ones and inferred from this that the altitude effect was a stronger influencer of stable isotope composition.
f. Spatial interpolation and isoscape characterization
Spatial interpolations were completed using the VWA stable isotope composition for each collector in the combined dataset to characterize O‘ahu’s isoscape. Spatial interpolations were completed using the gstat package in R (Pebesma 2004; Gräler et al. 2016). The interpolation method was ordinary kriging. We created an omnidirectional variogram and fit an exponential variogram model. The kriging parameters from δ18O were τ (nugget) = 0, σ (partial sill) = 0.38, and φ (range) = 3.87. The kriging parameters from d-excess were τ = 0, σ = 4.08, and φ = 3.40. This procedure was used to create maps showing the VWA isotopic composition of δ18O and d-excess in precipitation across the island of O‘ahu (Figs. 2 and 5). The visual classification scheme used to determine the stable isotope contour intervals was equal interval.
Precipitation VWA δ18O (‰) map for O‘ahu interpolated from precipitation samples collected between March 2018 and August 2021, plus data from Dores et al. (2020) and Booth et al. (2021). Contour intervals are shown in blue. The interpolation method was ordinary kriging, with coverage interpolated from data at the collector locations displayed. MT1 and D3 were collocated.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
g. Weather pattern analysis
Weather pattern analysis was used to assess major seasonal weather-related impacts on precipitation stable isotope composition. National Oceanic and Atmospheric Administration (NOAA) weather summaries were used to identify the instances of rainfall associated with cold fronts, Kona lows, upper-level lows, and tropical cyclones impacting O‘ahu (NOAA/NWS 2023). Weather data from the International Global Radiosonde Archive (IGRA) was used to identify trade wind periods (Durre et al. 2016). The IGRA data were filtered by isolating wind readings below 2500 m MSL, a height typically associated with the upper limit of the trade wind regime over Hawai‘i (Cao et al. 2007). Days in which the average wind direction below 2500 m MSL ranged from 22° to 112° from north were categorized as trade wind days. Weather trends were overlaid on the precipitation δ18O and δ2H composition data utilizing “rug” plots, which are a plot enhancement that display variable frequency as tick marks along an axis (Wickam et al. 2016). These were used to identify periods when synoptic events influenced rainfall stable isotope composition.
3. Results
a. Rainfall isotopic composition (δ18O, δ2H, and d-excess)
The isotopic composition of rainfall in the study area showed strong spatial and seasonal trends and was generally consistent with the narrow isotopic range of marine precipitation (Gat 1996). Precipitation VWA δ18O varied from −4.5‰ (Ka‘ala summit, KS) to −1.9‰ (Lyon Arboretum, D13) (see Table 2). Precipitation VWA δ2H varied from −22.9‰ (‘Ewa Beach, EB) to −4.1‰ (Lyon Arboretum, D13). Individual sample composition of δ18O ranged from −5.9‰ (‘Ewa Beach, EB, September 2019–January 2020) to −1.1‰ (Manana Trail, MT1, March 2021–April 2021), and δ2H ranged from −35.7‰ (‘Ewa Beach, EB, September 2019–January 2020) to 3.7‰ (Tripler Ridge, TR, March 2021–April 2021). Figure 2 displays an interpolated coverage of the VWA δ18O composition of rainfall. Rainfall tended to be more enriched in heavy isotopes in the Ko‘olau Range and in valleys and depleted in heavy isotopes for leeward O‘ahu and the Wai‘anae Range (Fig. 2). Figure 3 displays a time series of precipitation δ18O with weather patterns derived from the NOAA and IGRA data overlain. Figure 3 shows that wet seasons tended to bring isotopically lighter rainfall and higher rainfall rates, particularly during periods when there were disruptions to the trade wind regime. Dry seasons tended to produce isotopically heavy rainfall (i.e., closer to SMOW) and lower rainfall amounts, which is consistent with weather patterns reported by Longman et al. (2021) and other studies.
Rainfall VWA δ2H, δ18O, and d-excess for samples collected between March 2018 and August 2021 and published data from Dores et al. (2020) collectors D1–D20 and Booth et al. (2021) collectors HH, NH, and DH.
Time series of precipitation δ18O for 18 precipitation collectors sampled across the island of O‘ahu from March 2018 to August 2021. Precipitation collectors are identified by point shape, and the sample point size is scaled by normalized rainfall rate over the sample period. Weather trend data derived from IGRA weather balloon launches at Līhue, Kauai (Durre et al. 2016), and from monthly weather summaries provided by the NOAA (NOAA/NWS 2023) are shown in the background.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
b. Local meteoric water line
(a) LMWLs for O‘ahu derived from four different datasets compared to the GMWL. The combined LMWL presented in this study is generally consistent with those of other studies on O‘ahu and the GMWL. (b) Seasonal LMWLs for O‘ahu for the November–April wet season (blue line) and May–October dry season (orange line) and GMWL (dashed line). These were derived from the combined dataset and excluded split samples with more than 30 days in both wet and dry seasons.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
c. Deuterium excess
There was a strong correlation between VWA d-excess and elevation for collectors in several regions of the study area (Fig. 5). The strongest correlations were found among collectors in the leeward Ko‘olau and windward Wai‘anae ranges. For the leeward Ko‘olau collectors, the correlation analysis produced a R2 of 0.86 (p < 0.01), and a d-excess enrichment rate of 0.76‰ per 100 m of elevation gain (Fig. 6b). For the Wai‘anae collectors, the correlation analysis produced an R2 of 0.81, p < 0.002, and a d-excess enrichment rate of 0.45‰ per 100 m of elevation gain (Fig. 6b). There is a moderately strong correlation between islandwide VWA d-excess and collector elevation, with regression analysis producing an R2 of 0.53, p < 0.01, and a d-excess enrichment rate of 0.52‰ per 100 m of elevation gain (Fig. 6a). Deuterium excess also varied semipredictably with season, with wet seasons bringing higher d-excess in rainfall.
Precipitation VWA d-excess (‰) map for O‘ahu interpolated from the combined precipitation sample dataset, collected between March 2017 and August 2021, plus data from Dores et al. (2020) and Booth et al. (2021). Contour intervals are shown in blue. The interpolation method was ordinary kriging, with coverage interpolated from data at the collector locations displayed. MT1 and D3 were collocated.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
Precipitation VWA d-excess plotted against elevation for O‘ahu from the combined precipitation sample dataset collected between March 2017 and August 2021 for (a) all collectors and (b) select collectors located in key regions of the island (** indicates p < 0.05; * indicates 0.05 < p < 0.10).
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
4. Discussion
a. Amount effect: Correlating δ18O and rainfall rate
The correlation analysis of δ18O versus rainfall rates shows that rainfall isotopic composition did not vary appreciably with rainfall rate, which indicates that the amount effect was not a major factor controlling the stable isotope composition of rainfall on O‘ahu. Correlation analysis on the combined dataset produced poor results (R2 = 0.11, p < 0.001). Appendix C summarizes the results of this analysis on the subsets of the combined dataset broken down by the collector region and by the sample season. The results from this analysis show that for all subsets of the collector region and sample season, only two met the Spearman and p-test criteria: the dry season windward sample group and the split season North Shore sample group (which showed a positive correlation between rainfall rate and δ18O, opposite of what would be expected in samples impacted by the amount effect), although in both cases, sample sizes were small (n = 4). Overall, the correlation analysis results show that the amount effect was weak and localized, if not negligible, within the dataset.
b. Altitude effects meet source effects
Source-related differences in the isotopic composition of precipitation seem to strongly impact regional δ18O–elevation lapse rates on O‘ahu. Figure 7 shows the key trends observed in elevation–δ18O lapse rates within the study area. The δ18O–elevation relationships were similar to those found in other parts of the Hawaiian Islands (Tachera et al. 2021; Dores et al. 2020; Fackrell et al. 2020; Scholl et al. 1995; Booth et al. 2021; Torri et al. 2023) and to the average global lapse rate of −0.28‰ (100 m)−1 (Poage and Chamberlain 2011). Of the three elevation transects sampled in this study, the Manana Ridge transect showed the strongest δ18O lapse rate at −0.23‰ (100 m)−1 (R2 = 0.99). Several other collectors located in the leeward Ko‘olau Range plotted close to the Manana elevation lapse rate line (Fig. 7). The Ka‘ala transect produced a δ18O lapse rate half that of Manana at −0.11‰ (100 m)−1 (R2 = 0.88), and VWA precipitation δ18O values were generally 0.5‰–1.0‰ less than that of collections at the same elevation in the Manana transect. The Schofield Plateau transect did not produce a discernable lapse rate but seemed to straddle those of the Manana and Ka‘ala transects, suggesting a transition zone between two distinct isoscape regimes. Within the Schofield Plateau transect, the Kolekole collector (KK) produced an unexpectedly enriched VWA δ18O value, which created a dramatic kink in the elevation lapse rate, giving it a “U” shape (Fig. 7). Overall, these results confirm a strong altitude effect and near negligible amount effect shape precipitation stable isotope composition on O‘ahu over the study period. This points to the altitude and rainout effects as the primary relations impacting O‘ahu’s isoscape.
Precipitation VWA δ18O–elevation lapse rates from precipitation samples on the island of O‘ahu taken from March 2017 to August 2021. Three elevation transects are shown. The Manana transect is shown in purple and represents the leeward Ko‘olau Range, where trade wind–driven precipitation dominates stable isotope composition. The Ka‘ala transect is shown in red and represents the windward Wai‘anae Range, where synoptic storms have a stronger influence on stable isotope composition. The Schofield Plateau transect is shown in white and shows traits intermediate between the Ko‘olau and Wai‘anae isotopic regimes. The dashed lines show the ordinary least squares regressions for each transect (** indicates p < 0.05; * indicates 0.05 < p < 0.10).
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
One of the key regions where a predictable δ18O–elevation lapse rate would be useful is the leeward portion of the Ko‘olau Range, which supplies more than half of O‘ahu’s groundwater recharge (Engott et al. 2017). Ten precipitation collectors resided within this region: the Department of Health (DOH), Aiea (A), Waimano Upland (WU), the Manana transect (MT1/D3, MT2, MT3, MTS), Tripler Ridge (TR), East Range (ER), and Wahiawa Botanical Gardens (D17). Correlation analysis produced poor results (R2 = 0.26, p = 0.11) (Fig. 8a). We speculate that the variable performance of this regional δ18O–elevation regression is due to the seasonally dependent interactions between topography and the trade wind weather regime. At lower elevations on the leeward flank of the Ko‘olau Range, rainfall contributions from synoptic, disturbance-based storms are proportionally greater than the contributions from orographic precipitation. This is supported by isotope data from three locations within the leeward Ko‘olau region: Aiea (A, 13 m MSL), Department of Health (DOH, 154 m MSL), and Waimano Trail (WU, 330 m MSL) span an important transition zone where orographic precipitation begins to dominate local rainfall contributions. From sea level to approximately 100 m MSL, long-term average annual precipitation increases gradually from 600 to about 1000 mm yr−1 (Frazier and Giambelluca 2017). Above 100 m MSL, long-term average precipitation increases steeply, climbing to more than 5000 mm yr−1 over less than 400 m of elevation gain (Frazier and Giambelluca 2017). These three collectors represent a transition zone where synoptic versus orographic rainfall processes dominate the stable isotope composition of precipitation.
(a) Comparison of the VWA δ18O–elevation lapse rate from precipitation collectors located in the leeward Ko‘olau Range before and after a correction for synoptic storm contributions. The black regression line shows the raw, unadjusted δ18O–elevation lapse rate. The red regression line shows the model performance after five samples determined to be impacted by synoptic storms were removed from the calculations. Data for collectors D17 and D3 were taken from Dores et al. (2020). (b) Time series of precipitation δ18O compositions for three precipitation collectors sampled on the island of O‘ahu from March 2018 to August 2021. Samples identified as having been influenced by synoptic storms are outlined in red.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
Figure 8 depicts the interplay between these two precipitation zones. Figure 8a shows the leeward Ko‘olau δ18O–elevation lapse rate before and after a correction for synoptic storm rainfall contributions. Figure 8a shows that the Aiea (A) and DOH collectors had unexpectedly low VWA δ18O compositions, which distorted the regional lapse rate. These collectors were set apart as being more heavily influenced by synoptic rainfall events. Figure 8b displays a time series of δ18O composition with weather patterns overlain. This plot shows that the distortion of the regional lapse rate was due to five individual samples (outlined in red), all of which included a period of disturbance-based precipitation. These samples also stood out in that they generally had higher rainfall rate and volume, and mirrored each other in terms of stable isotope composition, which all point to a strong influence from synoptic rainfall events. Removing these samples from the VWA calculations produced a strong linear regression fit and a lapse rate nearly identical to that of the Manana transect (Fig. 8a). These findings indicate that precipitation sources must be considered when estimating elevation lapse rates from raw δ18O and δ2H composition.
c. Pseudoaltitude effect
We suggest that d-excess may be a reliable isotopic proxy for recharge elevation on O‘ahu when examined alongside δ18O lapse rates. Figure 6 shows that d-excess increased with elevation islandwide and that correlations were stronger when divided by region. We speculate that this correlation is related to temperature and humidity elevation gradients. The process of isotope fractionation is sensitive to temperature and humidity. This can be demonstrated by using the Craig–Gordon model for isotopic fractionation to predict the d-excess of vapor evaporated from the ocean surface at a range of temperatures and humidities, and then using the Rayleigh law to simulate rainout (Gat et al. 1996; Merlivat 1978; Kendall and McDonnell 1998, 66–68). The δ18O and δ2H composition of vapor, and therefore the d-excess of vapor, is strongly controlled by relative humidity and the kinetic isotope fractionation factor α (Kendall and McDonnell 1998, p. 68; Friedman and O’Neil 1977; Gat 1996). The parameter α is strongly dependent on temperature (Friedman and O’Neil 1977). Figure 9 shows modeled precipitation d-excess with changes in temperature and rainout. Figure 9a shows the d-excess change versus fraction of rainout with temperature changes denoted by color. This figure shows that d-excess change is most pronounced at higher temperatures and in the initial stages of rainout (rainout fraction < 0.15). Storms originating in cooler regions could theoretically experience minimal d-excess change as they progress toward Hawai‘i. Further, Fig. 9b shows that d-excess production is more sensitive to temperature and humidity changes than rainout. Intuitively, lower temperature and higher humidity result in lower precipitation d-excess.
Modeled d-excess production due to kinetic and equilibrium isotope fractionation. (a) d-excess production vs fraction of rainout with temperature changes denoted by color. (b) d-excess production vs relative humidity with temperature changes denoted by color. The figures were derived from equations given in Kendall and McDonnell (1998, 66–68) using equilibrium isotope fractionation factors from Friedman and O’Neil (1977).
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
Our data point to a pseudo-orographic control on precipitation d-excess within the study area. On O‘ahu, the mean annual temperature at sea level fluctuates between 25° and 29°C. Orographic temperature gradients can be steeper (∼1°C per 100 m of elevation gain) and can thus exert a stronger influence on isotope fractionation processes (Gat 1996; Natali et al. 2022). If a storm’s history has not caused a dramatic departure from the GMWL, a depleted air parcel from a synoptic storm could produce precipitation with a nearly identical d-excess signature to that of a more enriched parcel delivered by trade wind orographic processes. We speculate that after an air parcel has reached the island, subcloud evaporation and moisture recycling become the primary mechanisms controlling the d-excess of precipitation. This theory is supported by the findings of Bershaw and Lechler (2019), Bershaw (2018), Gonfiantini et al. (2001), Froehlich et al. (2008), and others. As an air parcel moves up the windward side of the Ko‘olau Range, rain droplets condense and undergo subcloud evaporation which preferentially removes 1H216O and 1H2H16O. Temperature and the saturation vapor pressure decrease, while relative humidity increases, which limits subcloud evaporation at the highest, coolest elevations. As the air parcel crests Ko‘olau and begins moving downslope toward leeward O‘ahu, temperature increases and rain droplets have a greater vertical distance to travel over which they experience subcloud evaporation. This removes even more 1H216O and 1H2H16O relative to precipitation that fell at higher elevations. In other words, precipitation falling at lower elevations has lower d-excess because it has undergone more subcloud evaporation than the precipitation falling at higher elevations. We speculate that fog may also contribute to the pseudoaltitude effect by incorporating 1H2H16O-rich fog droplets into precipitation at high elevations. These dynamics could be explored further by event-based high-elevation cloud water sampling, coupled with ground-level precipitation sampling and meteorological data collection, following Torri et al. (2023) and others.
d. Beyond rainfall: Assessing the potential isotopic influence of fog
Ka‘ala transect precipitation and high-elevation spring sample data with weather trends overlain. Data were collected between December 2018 and April 2021. The high-elevation spring stable isotope composition is depicted by the inverted triangle labeled SK1. Weather trend data derived from IGRA weather balloon launches at Līhue, Kauai (Durre et al. 2016), and from monthly weather summaries provided by the NOAA (NOAA/NWS 2023) are shown in the background.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
Figure 11 displays the results of the mixing model analysis. The mixing line passes through the VWA summit spring composition and extends to the average fog composition. Confidence intervals were added to the mixing line to account for the observational and instrumental uncertainty in both the spring and VWA rainfall composition measurements. The summit VWA rainfall stable isotope composition was −4.5‰ ± 0.7‰ δ18O and −19.7‰ ± 1.4‰ δ2H. The average isotopic composition of the high-elevation spring was −3.50‰ ± 0.02‰ δ18O and −10.4‰ ± 0.6‰ δ2H. The average isotopic composition of fog at the Ka‘ala summit was −2.00‰ ± 0.01‰ δ18O and 1.00‰ ± 0.06‰ δ2H. The average isotopic composition of fog on the mixing line is −2.19‰ δ18O and 1.39‰ δ2H. These values agree reasonably well with fog isotopic compositions presented by Scholl et al. (2002) from the eastern flank of Haleakalā on the island of Maui. Scholl et al. (2002) presented the fog isotopic data as raw fog composition and as the relative enrichment of fog compared to VWA rainfall isotopic composition. For comparison, Fig. 11 also depicts the Ka‘ala VWA rainfall isotopic composition enriched by 2.9‰ δ18O and 21‰ δ2H, which was the relative enrichment of Maui fog from Maui VWA rainfall stable isotope (Scholl et al. 2002).
KS rainfall–spring–fog mixing model analysis. The KS VWA rainfall stable isotope composition is −4.5 ± 0.7 δ18O and −19.7 ± 1.4 δ2H. The average Ka‘ala high-elevation spring composition is −3.5 ± 0.3 δ18O and −10.4 ± 2.4 δ2H. The average composition of fog at the KS was −2.0 δ18O and 1.0 δ2H (Tseng et al. 2018). The fog average composition was derived from the fog samples shown in light gray. Maui fog samples are taken from Scholl et al. (2002). Raw Maui fog composition was −2.6 δ18O and −4 δ2H. Relative enrichment depicts the departure from VWA precipitation at that location. The mixing model produced fog fractions of 0.41–0.45.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
The mixing model produced fog fractions of 0.41–0.45, which exceed regional fog drip estimates extrapolated from Ekern (1983). Other researchers have also noted that fog is a major component of the summit water budget. Tseng (2021) measured fog input due to interception by vegetation at the Ka‘ala summit and found that fog accounted for approximately 33.5% of precipitation. The findings presented here are preliminary but do suggest that more research is needed to fully understand the fog contributions to the water budget on Ka‘ala. Particularly of interest are temporal changes to fog contributions and stable isotope composition, and the isotopic behavior of moisture in the canopy and soil during the infiltration process. Mount Ka‘ala is an ideal place to carry out such research due to its pristine ecology, high elevation, and accessibility.
e. Influence of microclimates
The findings cumulatively point to strong, locale-specific connections between microclimates and the isotopic composition of rainfall; we identify six potentially distinct isotopic microclimate regimes. In high-elevation, high-rainfall areas in the leeward Ko‘olau Range, topographic influences on condensation during trade wind periods represent one regime, as evidenced by the strong δ18O–elevation lapse rate along the Manana transect (Fig. 7). In lower elevation, leeward areas, synoptic storms influence rainfall stable isotope composition in a fundamentally different way (i.e., the rainout effect), producing an isoscape characterized by higher depletion and moderate d-excess. Collectors D11 and D19 in the leeward Wai‘anae Range produced distinctly higher VWA δ18O and δ2H composition and moderate d-excess, possibly indicating the incorporation of vapor carried inland by land–sea-breeze cycles (Figs. 2 and 5). The strong influence of fog on Mount Ka‘ala points to yet another isotopic regime in high-elevation, fog-impacted areas. The Schofield Plateau and the Honolulu transect (Booth et al. 2021) present two more potentially distinct isoscape regions (Fig. 6). Last, we speculate that there may be a Ko‘olau valley effect. Collectors D13 and D20 produced precipitation that was isotopically enriched in δ18O with moderate d-excess (Figs. 2 and 5). We speculate that this could be due to increased moisture recycling in topographically isolated valleys. As moisture recirculates through valleys, 1H218O molecules would precipitate at a slightly greater rate than 1H2H16O, while 1H2H16O molecules would be preferentially evaporated from recirculated droplets. This could create a pattern with dampened d-excess and isotopically heavier rainfall compared to nonvalley locations. Understanding the strength and consistency of these microclimate stable isotope regimes will be an important step in examining connectivity between meteoric, ground, and surface waters. We suggest that focused sampling in the Ko‘olau Range, particularly in valleys and high-elevation fog-impacted areas, would provide valuable insights into these dynamics.
5. Conclusions
The data and findings presented here illuminate some key processes controlling the stable isotope composition of precipitation on O‘ahu. Analysis of precipitation samples indicates that the altitude effect exerts a distinct influence on precipitation stable isotope composition. However, systematic differences between orographic and synoptic precipitation values coupled with strong microclimate influences on rainfall stable isotope composition necessitate accounting for synoptic storm influences to establish reliable regional precipitation δ18O–elevation lapse rates. The data also reveal that rainfall d-excess is strongly connected to elevation, potentially more reliably than either δ18O or raw δ2H composition. We speculate that after an air parcel has reached the island, subcloud evaporation and moisture recycling become the primary mechanisms controlling the d-excess of precipitation. This connection is likely due to orographic controls on subcloud evaporation and moisture recycling. The exploration of a montane spring enabled the assessment of fog contribution to the Ka‘ala summit’s water budget. These results suggest that fog drip could constitute as much as 45% of the total groundwater recharge at the summit. Cumulatively, these findings indicate that the stable isotope composition of rainfall strongly reflects the regional microclimates on O‘ahu, and they highlight the strong influence of physiography on fractionation processes.
The findings of this study point to a need for more event-based precipitation sampling in connection with basic meteorological data collection to examine local controls on stable water isotope composition and d-excess. Fog sampling in both the Ko‘olau and Wai‘anae Ranges is also warranted. This study, along with previous work examining the major ion composition of precipitation in the same region and over the same timespan (Brennis et al. 2023), lays the groundwork for future research aimed at understanding the chemical connections between precipitation and groundwater recharge. Our ultimate goal is to constrain groundwater flow paths by means of natural, conservative geochemical tracers.
Acknowledgments.
This project has been funded by the NSF Hawai‘i EPSCoR Program through the National Science Foundation’s Research Infrastructure Improvement award (RII) Track-1: ‘Ike Wai: Securing Hawai‘i’s Water Future Award OIA-1557349. G. T. is supported by NSF Grant AGS-1945972. The project was also partially supported by the Department of Defense (DoD) Science, Mathematics and Research for Transformation (SMART) Service-for-Scholarship program. This project was also partially funded by the Office of Naval Research, Grant N00014-22-1-2403. The views expressed are those of the authors and do not necessarily reflect the views of any of the agencies listed. We thank Michael and April Wolfe, the State of Hawai‘i Department of Health, the State of Hawai‘i Department of Land and Natural Resources, the Schofield Barracks Directorate of Public Works, Mililani High School, and the Ko‘olau Mountains Watershed Partnership for land access and sampling assistance. This research was carried out as part of a PhD dissertation at the University of Hawai‘i at Mānoa by the lead author, with four of the six co-authors acting as committee members.
Data availability statement.
The data that support the findings of this study are openly available at https://hdl.handle.net/10125/106352.
APPENDIX A
Bulk Precipitation Collector Design
Figure A1 shows the design of the bulk precipitation collectors used in this study.
(left) Collector deployed along the TR hiking trail within the ‘Ewa Forest Reserve, O‘ahu. (center) Design of HPDE bucket style bulk precipitation collectors deployed throughout the study region. (right) Collector deployed at the summit of Mount Ka‘ala on O‘ahu.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
APPENDIX B
Mount Ka‘ala High-Elevation Spring
Figure B1 provides a visual reference for the perennial spring that was sampled near the summit of Mount Ka‘ala.
Perennial high-elevation spring along Mount Ka‘ala Road at 945 m MSL. Sunglasses are shown at the bottom for scale.
Citation: Journal of Hydrometeorology 25, 9; 10.1175/JHM-D-23-0193.1
APPENDIX C
Amount Effect Regression Analysis
Correlation analysis was used to evaluate the strength of the amount effect on precipitation stable isotope composition (Table C1). Individual sample precipitation δ18O values were compared to sample rainfall rates and evaluated on the basis of Spearman correlation coefficient and p value. The dataset was subdivided by season (dry, wet, or split) and collector location. Collectors were grouped into one of eight regions based on geographic location and general climate trends. Regression fits were scored by Spearman coefficient and p value. Subsets with p values less than or equal to 0.05 and Spearman scores greater than |0.5| were assumed to represent non-negligible correlations. Data from two other studies (Dores et al. 2020; Booth et al. 2021) were included in this analysis.
Results from amount effect correlation analysis. Bolded and underlined text show regressions that met the Spearman–p-test criteria. Abbreviations: n = sample group size; s = Spearman coefficient; p = p-test score.
APPENDIX D
Stable Isotope Composition of Fog Samples from Mount Ka‘ala
Fog, precipitation, and groundwater-fed spring samples were taken from the summit of Mount Ka‘ala and used to explore fog contributions to the water budget (Table D1). Fog samples were collected on Mount Ka‘ala on three days between April and July 2018 (Tseng et al., 2018). Fog samples were analyzed in the Isotope Biogeochemistry Laboratory at the University of Hawai‘i at Mānoa.
Stable isotope composition of six fog samples collected on Mount Ka‘ala.
REFERENCES
Bershaw, J., 2018: Controls on deuterium excess across Asia. Geosciences, 8, 257, https://doi.org/10.3390/geosciences8070257.
Bershaw, J., and A. R. Lechler, 2019: The isotopic composition of meteoric water along altitudinal transects in the Tian Shan of Central Asia. Chem. Geol., 516, 68–78, https://doi.org/10.1016/j.chemgeo.2019.03.032.
Booth, H., N. Lautze, D. Tachera, and D. Dores, 2021: Event-based stable isotope analysis of precipitation along a high resolution transect on the South Face of O’ahu, Hawai’i. Pac. Sci., 75, 421–441, https://doi.org/10.2984/75.3.9.
Bowen, G. J., 2010: Isoscapes: Spatial pattern in isotopic biogeochemistry. Annu. Rev. Earth Planet. Sci., 38, 161–187, https://doi.org/10.1146/annurev-earth-040809-152429.
Brennis, T., N. Lautze, R. Whitter, T. Giuseppe, and D. Thomas, 2023: Understanding the origins of and influences on precipitation major ion chemistry on the Island of O‘ahu, Hawai‘i. Environ. Monit. Assess., 195, 1265, https://doi.org/10.1007/s10661-023-11887-2.
Cao, G., T. W. Giambelluca, D. E. Stevens, and T. A. Schroeder, 2007: Inversion variability in the Hawaiian trade wind regime. J. Climate, 20, 1145–1160, https://doi.org/10.1175/JCLI4033.1.
Craig, H., 1961: Isotopic variations in meteoric waters. Science, 133, 1702–1703, https://doi.org/10.1126/science.133.3465.1702.
Dansgaard, W., 1964: Stable isotopes in precipitation. Tellus, 16A, 436–468, https://doi.org/10.1111/j.2153-3490.1964.tb00181.x.
Davis, G. H., C. K. Lee, E. Bradley, and B. R. Payne, 1970: Geohydrologic interpretations of a volcanic island from environmental isotopes. Water Resour. Res., 6, 99–109, https://doi.org/10.1029/WR006i001p00099.
Dores, D., C. R. Glenn, G. Torri, R. B. Whittier, and B. N. Popp, 2020: Implications for groundwater recharge from stable isotopic composition of precipitation in Hawai’i during the 2017–2018 La Niña. Hydrol. Processes, 34, 4675–4696, https://doi.org/10.1002/hyp.13907.
Durre, I., Y. Xungang, R. Vose, S. Applequist, and J. Arnfield, 2016: Integrated Global Radiosonde Archive (IGRA), version 2. NOAA National Centers for Environmental Information, accessed 4 January 2023, https://doi.org/10.7289/V5X63K0Q.
Ekern, P., 1983: Measured evaporation in high rainfall areas, Leeward Ko’olau Ranges, O’ahu, Hawai’i. WRRC Tech. Rep. 156, 70 pp., https://scholarspace.manoa.hawaii.edu/items/2ed4571f-bf6e-4655-851b-9f63287d8374.
Engott, J. A., A. G. Johnson, M. Bassiouni, S. K. Izuka, and K. Rotzoll, 2017: Spatially distributed groundwater recharge for 2010 land cover estimated using a water-budget model for the Island of O‘ahu, Hawai‘i (version 2.0, December 2017). U.S. Geological Survey Scientific Investigations Rep. 2015–5010, 49 pp., https://doi.org/10.3133/sir20155010.
Fackrell, J. K., C. R. Glenn, D. Thomas, R. Whittier, and B. N. Popp, 2020: Stable isotopes of precipitation and groundwater provide new insight into groundwater recharge and flow in a structurally complex hydrogeologic system: West Hawai’i, USA. Hydrogeol. J., 28, 1191–1207, https://doi.org/10.1007/s10040-020-02143-9.
Frazier, A. G., and T. W. Giambelluca, 2017: Spatial trend analysis of Hawaiian rainfall from 1920 to 2012. Int. J. Climatol., 37, 2522–2531, https://doi.org/10.1002/joc.4862.
Friedman, I., and A. H. Woodcock, 1957: Determination of deuterium-hydrogen ratios in Hawaiian waters. Tellus, 9A, 553–556, https://doi.org/10.3402/tellusa.v9i4.9119.
Friedman, I., and J. R. O’Neil, 1977: Compilation of stable isotope fractionation factors of geochemical interest. U.S. Geological Survey Professional Paper 440-KK, 117 pp., https://pubs.usgs.gov/publication/pp440KK.
Froehlich, K., M. Kralik, W. Papesch, D. Rank, H. Scheifinger, and W. Stichler, 2008: Deuterium excess in precipitation of Alpine regions – moisture recycling. Isot. Environ. Health Stud., 44, 61–70, https://doi.org/10.1080/10256010801887208.
Gat, J. R., 1996: Oxygen and hydrogen isotopes in the hydrologic cycle. Annu. Rev. Earth Planet. Sci., 24, 225–262, https://doi.org/10.1146/annurev.earth.24.1.225.
Gat, J. R., A. Shemesh, E. Tziperman, A. Hecht, D. Georgopoulos, and O. Basturk, 1996: The stable isotope composition of waters of the eastern Mediterranean Sea. J. Geophys. Res., 101, 6441–6451, https://doi.org/10.1029/95JC02829.
Giambelluca, T. W., 1983: Water balance of the Pearl Harbor-Honolulu Basin, Hawaii, 1946-1975. WRRC Tech Rep. 151, 163 pp., https://scholarspace.manoa.hawaii.edu/items/60d53166-b843-4c67-9df4-9f9bcf9aae0c.
Gingerich, S. B., and D. S. Oki, 2000: Ground water in Hawaii. U.S. Geological Survey Fact Sheet FS 126-00, 6 pp., https://pubs.usgs.gov/fs/2000/126/.
Gonfiantini, R., M.-A. Roche, J.-C. Olivry, J.-C. Fontes, and G. M. Zuppi, 2001: The altitude effect on the isotopic composition of tropical rains. Chem. Geol., 181, 147–167, https://doi.org/10.1016/S0009-2541(01)00279-0.
Gräler, B., E. Pebesma, and G. Heuvelink, 2016: Spatio-temporal interpolation using gstat. R J., 8, 204–218, https://doi.org/10.32614/RJ-2016-014.
Guan, H., X. Zhang, G. Skrzypek, Z. Sun, and X. Xu, 2013: Deuterium excess variations of rainfall events in a coastal area of South Australia and its relationship with synoptic weather systems and atmospheric moisture sources. J. Geophys. Res. Atmos., 118, 1123–1138, https://doi.org/10.1002/jgrd.50137.
Gupta, P., D. Noone, J. Galewski, C. Sweeney, and B. H. Vaughn, 2009: Demonstration of high-precision continuous measurements of water vapor isotopologues in laboratory and remote field deployments using Wavelength-Scanned Cavity Ring-Down Spectroscopy (WS-CRDS) technology. Rapid Commun. Mass Spectrom., 23, 2534–2542, https://doi.org/10.1002/rcm.4100.
IAEA/WMO, 2023: Global network of isotopes in precipitation. The GNIP Database, https://nucleus.iaea.org/wiser.
Ingraham, N. L., and R. A. Matthews, 1988: Fog drip as a source of groundwater recharge in Northern Kenya. Water Resour. Res., 24, 1406–1410, https://doi.org/10.1029/WR024i008p01406.
Ingraham, N. L., and R. A. Matthews, 1990: A stable isotopic study of fog: The Point Reyes Peninsula, California, U.S.A. Chem. Geol., 80, 281–290, https://doi.org/10.1016/0168-9622(90)90010-A.
Kelly, J. L., and C. R. Glenn, 2015: Chlorofluorocarbon apparent ages of groundwaters from West Hawaii, USA. J. Hydrol., 527, 355–366, https://doi.org/10.1016/j.jhydrol.2015.04.069.
Kendall, C., and J. J. McDonnell, 1998: Isotope Tracers in Catchment Hydrology. Elsevier, 839 pp.
Lanças, V. G., L. V. Santarosa, L. N. Garpelli, L. de Simone Borma, C. S. Quaggio, V. T. de Souza Martins, and D. Gastmans, 2022: Assessment of the changes in contributions from water sources to streamflow induced by urbanization in a small-sized catchment in Southeastern Brazil using the dual stable isotopes of water (18O and 2H). Environ. Monit. Assess., 194, 357, https://doi.org/10.1007/s10661-022-10040-9.
Longman, R. J., O. E. Timm, T. W. Giambelluca, and L. Kaiser, 2021: A 20-year analysis of disturbance-driven rainfall on O’ahu, Hawai‘i. Mon. Wea. Rev., 1767–1783, https://doi.org/10.1175/MWR-D-20-0287.1.
Merlivat, L., 1978: Molecular diffusivities of H2 16O, HD16O, and H2 18O in gases. J. Chem. Phys., 69, 2864–2871, https://doi.org/10.1063/1.436884.
Natali, S., M. Doveri, R. Giannecchini, I. Baneschi, and G. Zanchetta, 2022: Is the deuterium excess in precipitation a reliable tracer of moisture sources and water resources fate in the western Mediterranean? New insights from Apuan Alps (Italy). J. Hydrol., 614, 128497, https://doi.org/10.1016/j.jhydrol.2022.128497.
Nichols, W. D., P. J. Shade, and C. D. Hunt Jr., 1997: Summary of the Oahu, Hawaii, regional aquifer-system analysis. U.S. Geological Survey Professional Paper 1412-A, 74 pp., https://pubs.usgs.gov/pp/1412a/report.pdf.
NOAA/NWS, 2023: Hawaii monthly weather summaries. Hawai‘i Climate Data Portal, accessed 4 January 2023, https://www.hawaii.edu/climate-data-portal/publications-list/?collection_id=PZN8C7R9&collection_name=-%2520NOAA%252FNWS%2520Monthly%2520Precipitation%2520Summaries.
Pebesma, E. J., 2004: Multivariable geostatistics in S: The gstat package. Comput. Geosci., 30, 683–691, https://doi.org/10.1016/j.cageo.2004.03.012.
Poage, M. A., and C. P. Chamberlain, 2011: Empirical relationships between elevation and the stable isotope composition of precipitation and surface waters: Considerations for studies of paleoelevation change. Amer. J. Sci., 301, 1–15, https://doi.org/10.2475/ajs.301.1.1.
Prada, S., C. Figueira, N. Aguiar, and J. V. Cruz, 2015: Stable isotopes in rain and cloud water in Madeira: Contribution for the hydrogeologic framework of a volcanic island. Environ. Earth Sci., 73, 2733–2747, https://doi.org/10.1007/s12665-014-3270-1.
Scholl, M. A., S. E. Ingebritsen, C. J. Janik, and J. P. Kauahikaua, 1995: An isotope hydrology study of the Kilauea Volcano Area, Hawaii. U.S. Geological Survey Water-Resources Investigations Rep. 95-4213, 49 pp., https://doi.org/10.3133/wri954213.
Scholl, M. A., S. B. Gingerich, and J. W. Tribble, 2002: The influence of microclimates and fog on stable isotope signatures used in interpretation of regional hydrology: East Maui, Hawaii. J. Hydrol., 264, 170–184, https://doi.org/10.1016/S0022-1694(02)00073-2.
Scholl, M. A., T. W. Giambelluca, S. B. Gingerich, M. A. Nullet, and L. L. Loope, 2007: Cloud water in windward and leeward mountain forests: The stable isotope signature of orographic cloud water. Water Resour. Res., 43, W12411, https://doi.org/10.1029/2007WR006011.
Sherrod, D. R., J. M. Sinton, S. E. Watkins, and K. M. Brunt, 2021: Geologic map of the State of Hawai‘i. U.S. Geological Survey Scientific Investigations Map 3143, pamphlet, 5 sheets, scales 1:100,000 and 1:250,000, 78 pp., https://doi.org/10.3133/sim3143.
Sosa, E., J. C. Guerra, and M. T. Arencibia, 2011: Isotopic composition of rainwater in the subtropical island of Tenerife, Canary Islands. J. Environ. Hydrol., 19, 1–13.
State of Hawai‘i, 2019: Water Resource Protection Plan 2019 Update. Townscape Inc. and State of Hawai‘i, Commission on Water Resource Management Rep., 72 pp., https://files.hawaii.gov/dlnr/cwrm/planning/wrpp2019update/WRPP_201907.pdf.
Tachera, D. K., N. C. Lautze, G. Torri, and D. M. Thomas, 2021: Characterization of the isotopic composition and bulk ion deposition of precipitation from Central to West Hawai‘i Island between 2017 and 2019. J. Hydrol., 34, 100786, https://doi.org/10.1016/j.ejrh.2021.100786.
Terzer, S., L. I. Wassenaar, L. J. Araguás-Araguás, and P. K. Aggarwal, 2013: Global isoscapes for δ18O and δ2H in precipitation: Improved prediction using regionalized climatic regression models. Hydrol. Earth Syst. Sci., 17, 4713–4728, https://doi.org/10.5194/hess-17-4713-2013.
Tillman, F. D., D. S. Oki, A. G. Johnson, L. B. Barber, and K. R. Beisner, 2014: Investigation of geochemical indicators to evaluate the connection between inland and coastal groundwater systems near Kaloko-Honokōhua National Historic Park, Hawai‘i. Appl. Geochem., 51, 278–292, https://doi.org/10.1016/j.apgeochem.2014.10.003.
Torri, G., D. Ma, and Z. Kuang, 2017: Stable water isotopes and large-scale vertical motions in the tropics. J. Geophys. Res. Atmos., 122, 3703–3717, https://doi.org/10.1002/2016JD026154.
Torri, G., A. D. Nugent, and B. N. Popp, 2023: The isotopic composition of rainfall on a subtropical mountainous island. J. Hydrometeor., 24, 761–781, https://doi.org/10.1175/JHM-D-21-0204.1.
Tseng, H., 2021: Cloud water interception in Hawai’i. Ph.D. dissertation, University of Hawai‘i at Mānoa, 218 pp., https://hdl.handle.net/10125/81641.
van der Veer, G., S. Voerkelius, G. Lorentz, G. Heiss, and J. A. Hoogewerff, 2008: Spatial interpolation of the deuterium and oxygen-18 composition of global precipitation using temperature as ancillary variable. J. Geochem. Explor., 101, 175–184, https://doi.org/10.1016/j.gexplo.2008.06.008.
Veron, S., M. Mouchet, R. Govaerts, T. Haevermans, and R. Pellens, 2019: Vulnerability to climate change of islands worldwide and its impact on the tree of life. Sci. Rep., 9, 14471, https://doi.org/10.1038/s41598-019-51107-x.
Wahl, M. H., and H. C. Urey, 1935: The vapor pressures of the isotopic forms of water. J. Chem. Phys., 3, 411–414, https://doi.org/10.1063/1.1749690.
Wickam, H., and Coauthors, 2016: Rug plots in the margins. ggplot2: Elegant Graphics for Data Analysis, https://ggplot2.tidyverse.org/reference/geom_rug.html.
Willmott, C. J., 1981: On the validation of models. Phys. Geogr., 2, 184–194, https://doi.org/10.1080/02723646.1981.10642213.
Willmott, C. J., 1982a: On the climatic optimization of the tilt and azimuth of flat-plate solar collectors. Sol. Energy, 28, 205–216, https://doi.org/10.1016/0038-092X(82)90159-1.
Willmott, C. J., 1982b: Some comments on the evaluation of model performance. Bull. Amer. Meteor. Soc., 63, 1309–1313, https://doi.org/10.1175/1520-0477(1982)063<1309:SCOTEO>2.0.CO;2.
Willmott, C. J., and D. E. Wicks, 1980: An empirical method for the spatial interpolation of monthly precipitation within California. Phys. Geogr., 1, 59–73, https://doi.org/10.1080/02723646.1980.10642189.
Xia, Z., J. M. Welker, and M. J. Winnick, 2022: The seasonality of deuterium excess in non-polar precipitation. Global Biogeochem. Cycles, 36, e2021GB007245, https://doi.org/10.1029/2021GB007245.
