Jeju Island is located approximately 100 km south of the Korean Peninsula. It covers 1845 km2 and is dominated by Mt. Halla (1950 ma.s.l.), with terrain rising from coastal plains to the central summit (Fig. 1). The island consists predominantly of fractured basalt, with localized trachyte and pyroclastic deposits. The high permeability of the basalt promotes rapid infiltration and limited surface runoff, contributing to groundwater-dominated hydrology (Hahn et al., 1997; Koh et al., 2012; Kim et al., 2026a). Jeju Island has a monsoon-influenced climate situated between temperate continental and subtropical maritime regimes. Regional studies describe marked seasonal contrasts in atmospheric circulation and moisture transport (Lee et al., 2013; Kim et al., 2019). These climatic conditions provide context for the observation period, although individual precipitation samples are not assigned to specific moisture sources or atmospheric mechanisms in the present study.

Climatically, Jeju Island lies in a transitional zone between temperate continental and subtropical maritime regimes. Its climate is strongly shaped by monsoon circulation, seasonal shifts in moisture sources, and interactions between the Siberian High and the North Pacific High. Winters (November to April) are influenced by cold, dry continental air masses, while summers (June–September) are dominated by humid maritime tropical air advected from the Pacific Ocean. During the sampling period from September 2000 to December 2003, the site-specific mean monthly air temperatures associated with the isotope-monitoring network ranged from -3.4 to 27.9 °C, with an overall mean of 12.7 °C. Annual precipitation exceeds 1600 mm in most regions, with over half occurring during the summer monsoon period. These synoptic-scale seasonal contrasts contribute to seasonal changes in precipitation δ18O and δ2H values and d-excess, contributing to well-defined seasonal patterns in δ18O, δ2H, and d-excess values (Lee et al., 2013; KMA, 2021). The combined influence of steep altitudinal gradients, varied moisture source trajectories, and distinct monsoon circulation patterns produces spatial and temporal variations in precipitation δ18O, δ2H, and d-excess that cannot be captured using single-station monitoring approaches. Coastal regions are predominantly influenced by marine vapor and convective rainfall, while precipitation at higher elevations is generally depleted in 18O and 2H relative to precipitation at lower elevations. These conditions make Jeju Island an ideal setting for investigating altitude effects, moisture-source variability, and monsoon-driven isotopic seasonality through multi-site, long-term monitoring. Consequently, the 15 monitoring stations were distributed across elevations ranging from 5 to 1230 m a.s.l. to capture these spatial gradients.

Previous studies on Jeju Island examined stable isotopes in precipitation, groundwater, and cave water, but the early precipitation records were generally limited in duration or spatial coverage (Lee et al., 1999, 2003; Woo et al., 2013). Recent work analyzed stable isotopes in monthly precipitation and groundwater together with isotope mixing to examine the seasonal and elevational origins of groundwater recharge, identifying high-elevation summer precipitation as the dominant recharge source (Kim et al., 2026a). The present data paper complements that application-focused study by documenting and openly publishing a station-resolved precipitation-isotope record and associated metadata from September 2000 to December 2003, facilitating transparent reuse and evaluation of seasonal and elevation-related patterns.

Subsequent research expanded to include the isotopic characterization of groundwater, cave dripwater, and speleothems, demonstrating that modern precipitation isotope data are essential for calibrating paleoclimate proxies in Jeju's volcanic landscapes (Woo et al., 2008; Jo et al., 2014). These studies identified that groundwater recharge is predominantly derived from high-elevation winter precipitation, while speleothem isotopic signals preserve a mixture of seasonal infiltration and moisture source variability (Lee et al., 2007; Kim et al., 2003). However, these investigations largely relied on extrapolated isotope lapse rates or interpolated GNIP data due to the absence of a spatially resolved precipitation isotope dataset collected simultaneously across multiple elevations and climatic regimes. This limitation has hindered the ability to accurately quantify spatial heterogeneity in isotopic elevation relationships, moisture source transitions, and monsoon driven fractionation processes, thereby constraining the development of reliable isoscapes and robust paleoclimate interpretations for Jeju Island (Peng et al., 2010; Ha et al., 2013).

More recent studies employed isotope-enabled models and atmospheric satellite observations to analyze moisture source trajectories and synoptic-scale vapor dynamics affecting Jeju Island (Yoshimura et al., 2011; Lee et al., 2013). These works highlighted the importance of a seasonal pattern characterized by lower precipitation δ18O and δ2H values in summer and higher values in winter, but validation of these model outputs was limited due to sparse in situ observations and the lack of a high-density monitoring network. The previous studies could therefore characterize either temporal changes or spatial differences, but not both concurrently. No prior study has produced a multi-year, multi-site dataset capturing both spatial and temporal variability of precipitation isotopes across volcanic terrain, which is essential for quantifying altitude effects, seasonal moisture source changes, monsoon circulation patterns, and Rayleigh distillation across elevation gradients.

Recent work on Jeju Island combined precipitation and groundwater isotope observations with mixing analysis to identify seasonal and high-elevation contributions to groundwater recharge (Kim et al., 2026a). The present data paper complements that application-focused study by openly providing the complete station-resolved precipitation isotope record and associated metadata, enabling independent reuse, consistent evaluation of isotope–elevation relationships, and comparison with subsequent monitoring. The present dataset fills this critical gap by providing simultaneous spatial and temporal precipitation isotope observations from 15 stations over a continuous 39-month period. It enables, for the first time on Jeju Island, quantitative assessment of both seasonal isotope cycles and decreases in precipitation δ18O and δ2H values with increasing elevation, facilitating more accurate characterization of recharge dynamics, topographic gradients, and paleoenvironmental interpretations. The dataset also offers a robust observational basis for calibrating isotope-enabled climate models and for developing regional isoscape frameworks.

Precipitation samples were collected continuously from September 2000 to December 2003 at 15 strategically selected stations distributed across Jeju Island (5–1230 m a.s.l.), covering distinct topographic, climatic, and geographic features (Fig. 1). The sampling network was designed to represent windward and leeward slopes, coastal-to-inland transitions, and altitudinal gradients, allowing simultaneous characterization of spatial and temporal variations in precipitation δ18O, δ2H, and d-excess. Stations were classified into four elevation zones (coastal < 100 m, lowland 100–300 m, mid-elevation 300–800 m, and high-elevation > 800 m) to quantify altitude effects and orographic precipitation processes.

In total, δ18O and δ2H were measured in 462 precipitation samples, and d-excess was calculated as d = δ2H - 8 × δ18O. Analyses were conducted using an Isoprime isotope ratio mass spectrometer (IRMS) at the Korea Basic Science Institute (KBSI), following IAEA protocols. δ18O measurements were performed using CO2–H2O equilibration and cryogenic condensation, whereas δ2H was analyzed by chromium reduction (Cr-Pyrolysis) to generate H2 gas. Analytical precision was ± 1.0 ‰ for δ2H and ± 0.1 ‰ for δ18O, based on repeated internal standards calibrated against VSMOW2, SLAP2, and GISP reference waters. All reported values follow the conventional δ-notation relative to VSMOW: 1δ=RxRVSMOW-1×1000 where Rx is the ratio of heavy to light isotopes (18O/16O or 2H/H) in the sample.

During compilation of the released dataset, we did not apply a statistical outlier threshold to δ18O, δ2H, or d-excess. We retained the single exclusion reported by Lee et al. (2002). A May 2001 sample identified as ST-3 in that study had δ18O = -1.9 ‰ and δ2H = -33 ‰, with a d-excess value of -17.8 ‰. Lee et al. (2002) reported that the sample departed strongly from the meteoric-water-line relationship, interpreted this as evidence of substantial evaporation before analysis, and excluded it from their calculations. We therefore also excluded this sample from the released observations and from all analyses reported here. Its identifier, values, cross-reference, and exclusion rationale are retained in the Excluded_Samples worksheet for traceability. No additional archived measurement was removed during the present compilation solely because its δ18O, δ2H, or d-excess value was relatively high or low. The released dataset contains the remaining 462 paired δ18O and δ2H measurements. Site–month combinations without a valid archived isotope measurement remain missing and were neither reconstructed nor imputed. Precipitation and air-temperature records were not used as numerical isotope-screening criteria.

Precipitation δ18O and δ2H values and d-excess (d = δ2H - 8 × δ18O) varied seasonally during the 39-month observation period from September 2000 to December 2003 (Fig. 2). Monthly mean δ18O and δ2H values were generally lower during summer, particularly from June to September, and relatively higher during winter from November to February although the timing and magnitude varied among years. The seasonal minima in δ18O and δ2H broadly coincided with the main summer-monsoon precipitation period, whereas the seasonal maxima occurred during the relatively dry winter period. Across the 462 individual samples, δ18O ranged from -13.2 ‰ to -2.0 ‰, δ2H from -93.3 ‰ to -5.8 ‰, and d-excess from 3.3 ‰ to 29.1 ‰. The individual station-level δ18O observations are shown in Fig. S1 in the Supplement, and island-wide monthly means are shown with the corresponding meteorological series in Fig. 2. This pattern indicates that precipitation during the peak summer monsoon carries the most isotopically depleted signatures, reflecting intensified convective activity, long-distance moisture transport, and enhanced Rayleigh distillation associated with summer precipitation. In contrast, isotopic enrichment during winter precipitation is likely related to lower atmospheric humidity, stronger sub-cloud evaporation, and increased contributions from drier continental air masses during the cold season (Tao et al., 2025).

Unlike δ18O and δ2H, the secondary isotope parameter d-excess shows a more complex seasonal evolution, characterized by a quadratic rather than sinusoidal pattern, indicating a stronger sensitivity to kinetic fractionation processes (Dansgaard, 1964). Monthly d-excess values are lowest during summer months, consistent with highly humid maritime moisture sources, whereas significantly higher winter d-excess values are consistent with seasonal changes in source-region humidity and evaporation conditions; however, no unique moisture-source attribution is made from d-excess alone. These seasonal behaviors were further supported by correlation analysis, which showed that d-excess exhibited the strongest inverse relationship with air temperature (R2 = 0.77, p < 0.0001) compared to δ18O (R2 = 0.07) and δ2H (R2 = 0.29), confirming its higher sensitivity to moisture source humidity and evaporation conditions (Table 1). The broad seasonal cycle is consistent with changes in East Asian monsoon circulation, precipitation amount, transport history, and condensation along air-mass pathways (Lee et al., 2003, 2013). The present monthly dataset does not provide event-scale trajectories or process diagnostics, however, so individual samples are not assigned to specific moisture sources and the observed variations are not attributed to a single mechanism (Guo et al., 2021).

Seasonal analysis of the local meteoric water lines (LMWLs) further reveals systematic variations in δ18O vs δ2H relationships associated with changes in atmospheric circulation and moisture source conditions over Jeju Island (Fig. 3; Table 2). The annual LMWL derived from the full dataset (δ2H = 8.17 × δ18O + 15.72, R2 = 0.89) reflects the integrated influence of both summer monsoon and winter precipitation regimes, but seasonal regressions demonstrate clear contrasts between warm and cold seasons (Fritz et al., 2022). During the summer monsoon period (JJAS), the LMWL exhibits a relatively higher slope and a markedly lower intercept, consistent with precipitation dominated by humid maritime moisture sources and extensive rainout under convective conditions, which suppresses kinetic fractionation during evaporation. In contrast, the winter LMWL (NDJF) is characterized by a reduced slope and an elevated intercept, indicating enhanced kinetic effects associated with lower humidity, stronger evaporation at the moisture source, and increased contributions from continental air masses transported over the Yellow Sea. Transitional seasons display intermediate LMWL characteristics, reflecting the gradual shift between these two dominant circulation regimes. These seasonal differences in LMWL geometry are consistent with the observed seasonal behavior of d-excess and corroborate the interpretation that moisture source humidity and atmospheric transport pathways exert a first-order control on isotope relationships in precipitation. The well-resolved seasonal LMWLs presented here provide a robust reference framework for interpreting groundwater recharge processes, cave dripwater isotopes, and paleoclimate archives in monsoon-influenced volcanic island environments, where mixing of seasonal signals can otherwise obscure the climatic interpretation of isotope records.

Recurring annual variability in monthly δ18O and δ2H was summarized empirically as δ(t)=C+Asin⁡[2π(t-φ)/12], where t is calendar month, C is the fitted baseline, A is amplitude, and φ is phase (Fig. 4). Unlike conventional raw isotopic time-series, the seasonal cycle parameters (baseline, amplitude, and phase) extracted from the periodic fitting approach allow direct input into hydrological and land-surface models that require temporally smoothed isotope forcing data (Nelson et al., 2021). This is particularly valuable in monsoon climates, where conventional amount-weighted averages do not capture seasonal variability in moisture source or atmospheric circulation. The documented seasonal cycle also defines a reference framework for distinguishing between temperature-based fractionation (winter isotopic enrichment) and moisture-source depletion (summer monsoon rainout), facilitating attribution of isotope variation to underlying atmospheric processes rather than simple temperature effects. The secondary isotopic parameter, d-excess, however, exhibited a more complex seasonal pattern, deviating from pure sinusoidal behavior. Its strong inverse relationship with temperature and relative humidity confirms that d-excess can serve as an effective indicator of moisture source humidity, ocean–atmosphere evaporation conditions, and sub-cloud kinetic fractionation during transport (Schoenemann and Steig, 2016). This sensitivity makes d-excess particularly valuable for identifying shifts between continental (high d-excess) and maritime (low d-excess) moisture sources, especially during winter and summer seasons, respectively. When integrated with satellite water vapor isotope datasets (e.g., TES, TROPOMI) or isotope-enabled reanalysis, the dataset can help improve representations of monsoon moisture recycling, boundary layer humidity, and airmass trajectory modeling (Worden et al., 2007; Lee et al., 2011; Lee et al., 2015). The fitted δ18O and δ2H baselines are −7.09 ‰ and −43.86 ‰, with amplitudes of 1.60 ‰ and 17.42 ‰, respectively. The corresponding R2 values are 0.40 and 0.63, respectively (Fig. 4). They are therefore descriptive summaries, not mechanistic representations of isotope processes or monthly climatological normals. d-excess was summarized with a quadratic calendar-month curve because its observed seasonal distribution was not symmetric. d-excess may be influenced by source-region humidity and evaporation, cloud processes, and sub-cloud exchange, among other factors. Low or high d-excess values are consequently not assigned here to a binary maritime-versus-continental source interpretation.

The spatial distribution of precipitation isotopes across Jeju Island demonstrates strong topographic control, primarily driven by orographic uplift, Rayleigh distillation, and varying exposure to maritime air masses. As shown in Fig. 5, δ18O and δ2H values exhibit progressive depletion with increasing elevation, forming a distinct altitudinal gradient characteristic of mountainous volcanic terrains. Based on 462 precipitation samples across 15 stations, δ18O decreased at an average rate of -1.5 ‰ per 1000 m, with a strong linear fit (R2 = 0.88, p < 0.005), while δ2H showed a corresponding trend of -11 ‰ per 1000 m (R2 = 0.77). These lapse rates are broadly consistent with those reported for other high-relief island environments in East Asia and the western Pacific, consistent with progressive rainout during orographic uplift.

While δ18O and δ2H showed clear elevation-driven depletion trends, d-excess displayed weaker spatial dependence, with only modest changes across elevation groups (R2 = 0.39). Higher d-excess values at mid- and high-altitude locations indicate potential influence of drier continental air masses and moisture recycling during winter months, while lower values in coastal stations reflect more humid maritime moisture and stronger kinetic fractionation at the ocean–atmosphere interface. The combined analysis of δ18O, δ2H, and d-excess suggests that spatial variability in precipitation isotopes on Jeju Island is shaped not only by orographic lifting but also by seasonal shifts in moisture sources and varying humidity conditions during condensation and transport. The spatially resolved dataset presented here is the first to quantify the simultaneous effects of altitude, topographic orientation, and monsoon circulation on precipitation isotopes across Jeju Island. These results provide a robust observational basis for estimating groundwater recharge elevations, refining local isoscape models, and improving isotope-enabled hydrological simulations in volcanic island settings. Furthermore, the observed spatial gradients serve as critical calibration inputs for interpreting isotopic signals preserved in speleothems, cave dripwater, and other paleoclimate archives, contributing to improved proxy-based climate reconstructions in East Asia.

To examine spatial heterogeneity, the monitoring stations were grouped into northern, southern, eastern, and western sectors (Fig. 6). To avoid confounding the comparisons with unequal station record lengths, station means were calculated over the common 2002–2003 network period. Across all 15 stations spanning 5–1230 ma.s.l., precipitation δ18O and δ2H values decreased significantly with elevation at rates of -0.146 ± 0.013 ‰ and -1.043 ± 0.128 ‰ per 100 m, respectively (δ18O: R2 = 0.913, p < 0.001; δ2H: R2 = 0.837, p < 0.001; Table 3). The sector-specific δ18O gradients ranged from -0.105 ‰ to -0.155 ‰ per 100 m, while the corresponding δ2H gradients ranged from -0.682 ‰ to -1.158 ‰ per 100 m. However, each sector comprised only three or four stations and covered a different elevation range. These sector-specific regressions should therefore be regarded as exploratory, and the separate fits do not demonstrate that the slopes differ significantly among sectors. Moreover, because sample-resolved wind and atmospheric-trajectory information is unavailable, the observed sector patterns cannot be attributed uniquely to marine influence, prevailing wind direction, topographic shielding, or windward–leeward position. The network-wide regressions provide the most robust first-order description of elevation-related variation in precipitation δ18O and δ2H during the common observation period (Wang et al., 2024b).

Examining altitudinal gradients along multiple cardinal directions provides further clarity on the spatial complexity of isotope fractionation across Jeju. This multidirectional perspective disentangles purely altitude-driven depletions from other spatial influences, such as the proximity to marine air masses or the prevailing wind trajectories. By constructing isotopic transects from north, south, east, and west, it becomes possible to pinpoint anisotropic depletion patterns and better discern localized controls on isotopic variability. These data, in turn, enhance the spatial resolution of isotope-based models and strengthen our capacity to predict recharge elevations and groundwater flow paths. The ability to distinguish between low-elevation recharge derived from coastal precipitation and high-elevation recharge sourced from more depleted rainfall is especially valuable for managing volcanic island aquifers, which often exhibit complex stratigraphy and vulnerability to contamination. Ultimately, integrating cardinal-direction transects with altitude effect analyses offers a comprehensive approach to understanding both the spatial and altitudinal dimensions of precipitation isotope variability, supporting improved water resource management and more nuanced paleoclimate reconstructions.

Previous satellite, IsoGSM, and 120 h HYSPLIT backward-trajectory analyses demonstrated seasonally contrasting atmospheric transport pathways over Jeju Island (Lee et al., 2013). These results provide regional atmospheric context for the possible influence of circulation on the directional lapse rates. However, because event-specific wind and trajectory information is not available for each monthly precipitation sample, the differences among the directional lapse rates cannot be attributed uniquely to particular airflow regimes. They are therefore interpreted as being consistent with combined differences in slope exposure, orographic ascent, and seasonal circulation rather than as direct evidence of a specific circulation mechanism.

Although precipitation isotope composition is affected by multiple atmospheric and meteorological processes, the observed isotope–elevation relationships provide a first-order empirical basis for developing a regional precipitation isoscape. The network-wide lapse rates can be used to estimate expected modern precipitation δ18O and δ2H values at elevations where direct observations are unavailable. The directional analyses additionally indicate whether separate relationships may be required for different slope sectors, although such subdivision should be applied only where supported by the regression results and station coverage. A regional precipitation isoscape would be particularly useful for cave and speleothem studies on Jeju Island because precipitation isotope composition is rarely measured directly at individual cave sites. Applying the observed isotope–elevation relationships to the elevation and geographical setting of a cave can provide an initial estimate of the modern precipitation isotope input to its recharge system. This estimate can help define the plausible precipitation end-member from which soil water, epikarst water, and cave dripwater are derived.

The estimated precipitation input does not directly calibrate a speleothem record because the transfer of an isotope signal from precipitation to speleothem calcite is additionally influenced by seasonal recharge weighting, evaporation, soil and epikarst storage, cave hydrology, water–rock interaction, and calcite precipitation processes. Nevertheless, the isoscape provides an important initial constraint for evaluating this transfer when direct precipitation isotope measurements at the cave elevation are unavailable.

The 39-month dataset across Jeju Island provides one of the few datasets in East Asia that captures both the amplitude and phase of seasonal isotopic cycles, allowing for robust quantification of seasonal offsets in δ18O, δ2H, and d-excess. Unlike short-duration event-based sampling campaigns or single-site monitoring records, this dataset enables assessment of multi-year seasonal stability, which is essential for calibrating isotope-enabled climate models, such as IsoGSM and iCESM, and for validating isotopic proxies preserved in groundwater and speleothem archives (Yoshimura et al., 2011; Brady et al., 2019). The consistent occurrence of isotopic minima during summer (June to September) and maxima during winter (November to February) supports its application in seasonal weighting of recharge, particularly in volcanic aquifers where infiltration timing determines groundwater residence and isotope signatures.

To evaluate how well an isotope-enabled general circulation model reproduces the observed seasonal isotope variability over Jeju Island, we compared the observations with output from IsoGSM4 (IsoGSM-HR), the latest version of the isotope-incorporated Global Spectral Model, which simulates the fractionation and transport of stable water isotopes within a global atmospheric general circulation model (Yoshimura et al., 2008; Bong and Yoshimura, 2025a, b). The simulation was run at T248 spectral resolution, corresponding to a horizontal grid spacing of approximately 50 km, and was nudged to the ERA5 reanalysis so that the simulated synoptic-scale circulation follows the observed atmospheric evolution; this allows a direct, month-by-month comparison between the simulated and observed precipitation isotope compositions over the observation period (September 2000 to December 2003). Monthly mean precipitation δ18O and δ2H values were obtained from the published model output (Bong and Yoshimura, 2025a, b; 10.5281/zenodo.14681370 and 10.5281/zenodo.15073235). At this resolution, Jeju Island is covered by three model grid cells, centered at 33.31° N and 126.09° E, 126.56° E, and 127.03° E, corresponding to the western, central, and eastern parts of the island, respectively. Monthly values with simulated precipitation below 0.2 mmd-1 were excluded, because the isotopic composition of near-zero precipitation amounts is numerically unstable and not physically meaningful. For each month, the mean and standard deviation of δ18O, δ2H, and d-excess across the three grid cells were computed, and the ± 1 standard deviation range was used to construct the shaded band shown in Fig. 7, which is compared with the monthly mean observations averaged over all fifteen stations.

The mean model-minus-observation differences during May–October were +0.3 ‰ for δ18O, +1.2 ‰ for δ2H, and -1.2 ‰ for d-excess, whereas the corresponding differences during November–April were +1.4 ‰, +5.9 ‰, and -5.1 ‰, respectively (Fig. 7). The differences were therefore smaller during May–October than during November–April. However, the approximately 50 km grid cannot resolve Jeju Island's detailed topography, station-scale isotope–elevation relationships, or local-to-mesoscale precipitation processes. The comparison does not determine whether the seasonal differences arise from unresolved topography, land–sea contrasts, spatial representativeness, or parameterized cloud and precipitation processes. Resolving these contributions would require additional process diagnostics and an isotope-enabled convection-permitting regional model or an appropriately configured non-hydrostatic framework such as NICAM-WISO (Tanoue et al., 2023). The IsoGSM4 comparison is therefore used only to assess broad seasonal consistency and is not considered a validation of the observations or a process-attribution analysis.

During summer, particularly during the early-summer Changma season, precipitation over Jeju Island is strongly influenced by the Changma front, the Korean sector of the Meiyu–Baiu frontal system, and its associated large-scale moisture transport. Southwesterly flow along the western to northwestern flank of the western North Pacific subtropical high transports warm, moist air toward the Korean Peninsula, promoting moisture convergence and precipitation over Jeju Island and adjacent regions (Ninomiya and Akiyama, 1992; Lee et al., 2003; Kim et al., 2010).

When combined with independent cave-hydrological information, the dataset may also help assess how precipitation-isotope variability is transferred into cave dripwater. For example, comparisons between the precipitation record and contemporaneous or subsequent dripwater measurements may assist in evaluating whether a cave system integrates precipitation approximately across the year or preferentially reflects recharge during particular seasons. The observed range of precipitation δ18O values can provide an initial reference against which the magnitude and variability of dripwater δ18O are evaluated. Similarly, the elevation dependence of precipitation δ18O may help constrain the range of possible recharge elevations. These comparisons could support the development and testing of conceptual or numerical models describing the transfer of precipitation through soil, epikarst, and cave systems. They cannot, however, determine recharge seasonality or elevation without independent hydrological evidence.

The temporal coverage of the dataset imposes an additional limitation. The record extends for 39 months from September 2000 to December 2003 and can characterize spatial patterns and recurring seasonal variability during that interval. It is too short to quantify interannual-to-decadal variability, establish a long-term precipitation-isotope climatology, or determine whether the observed seasonal and elevation-related relationships remained stationary outside the monitoring period. The monthly samples may also integrate multiple precipitation events and atmospheric pathways, preventing event-specific moisture sources or transport histories from being identified. Moreover, the dataset does not include contemporaneous observations of cave dripwater, cave temperature, infiltration, or speleothem calcite. It therefore cannot provide a stand-alone calibration or validation of paleoclimate reconstructions.