Drip water δ¹⁸O variability in the northeastern Yucatán Peninsula, Mexico: Implications for tropical cyclone detection and rainfall reconstruction from speleothems

Abstract

This study examines the oxygen isotopic composition (δ¹⁸O values) of drip water, rainfall, and groundwater in the Río Secreto cave system, located in the Yucatán Peninsula, Mexico. The main motivation of this study was to determine the implications of drip water hydrology for the reconstruction of rainfall, droughts and tropical cyclone activity from stalagmite δ¹⁸O records. Monitoring of environmental and isotopic conditions was conducted for two years, from June 2017 to April 2019. This study provides the first instrumental evidence of an “amount effect” on interannual timescales in the Yucatán Peninsula. Observed bi-weekly to interannual variability in drip water δ¹⁸O values can be explained for individual drips by different integrations of rainfall amount in the time domain. Drip sites in two chambers (Stations A and B) integrate 4–15 months of rainfall accumulation. In a third chamber (Station LF) one drip site reflects the annual rainfall isotopic cycle with a positive offset and another, the largest rainfall events. During epikarst infiltration, the integration of rainfall amount by drip water source reservoirs determines the degree to which they “dilute” a tropical cyclone (TC) isotopic signature. TCs can be detected particularly when: (1) the water volume of the reservoir is low, such as during a persistent meteorological drought, and; (2) TCs have a sufficiently distinct isotopic signal relative to that of the reservoir prior to the event. TC isotopic signals can be masked or attenuated when drip water samples integrate more than a week and if significant rainfall events proceed the TC. In Río Secreto cave, reconstructing precipitation amount and detecting the TC isotopic signatures from stalagmite δ¹⁸O records is possible. Our analysis shows that stalagmite δ¹⁸O records are more likely to underestimate the magnitude of annual-scale droughts following normal hydroclimate conditions and more likely to record TCs during multiyear droughts than during normal or wet periods. Drip water monitoring results suggest that available stalagmite δ¹⁸O records from the Maya lowlands might be underestimating the intensity of paleo-drought events, such as the Terminal Classic droughts associated with the disintegration of the Maya civilization. This study complements the results from Lases-Hernandez et al. (2019) comparing two different sampling protocols of drip water collection. This study shows that a discrete sampling protocol is expected to approximate the amount-weighted isotopic composition of a drip, as long as it is conducted at a temporal resolution higher than the rainfall integration time by the drip reservoir. We highlight the importance of conducting multiyear monitoring of drip water and rainfall in order to interpret stalagmite δ¹⁸O as a paleoclimate proxy.

Introduction

The frequency and intensity of precipitation extremes are expected to increase globally as human greenhouse gas emissions (GHG) continue to rise over the 21st century (Huntington, 2006, Williams et al., 2007; Déry et al., 2009, O'Gorman and Schneider, 2009, Lu and Fu, 2010, Seager et al., 2010, Wu et al., 2010, Kao and Ganguly, 2011, Müller et al., 2011, Durack et al., 2012). Intergovernmental Panel of Climate Change (IPCC-AR5) projections of precipitation variability in response to GHG forcing over the 21st century show large uncertainties, particularly in regions between tropical and subtropical climates. Paleoclimate studies have the potential to help reduce these uncertainties by providing empirical estimates of precipitation responses to shifts in internal modes of climate variability and the atmospheric composition of GHGs.

Stalagmite calcite or aragonite oxygen isotope (δ¹⁸O value) records represent one of the most promising terrestrial paleoclimate archives with the potential to yield semi-quantitative records of precipitation variability and records of tropical cyclone (TC) activity in tropical and subtropical regions (e.g. Wang et al., 2001, Frappier et al., 2007a; Fleitmann et al., 2009; Medina-Elizalde et al., 2010, Kennett et al., 2012, Partin et al., 2012, Akers et al., 2016, Baldini et al., 2016, Medina-Elizalde et al., 2017).

In these regions, stalagmite δ¹⁸O records have been interpreted either explicitly or implicitly to reflect the “amount effect”; that is, the inverse relationship between rainfall amount and rainfall δ¹⁸O described by Dansgaard (1964), that occurs on seasonal and interannual time scales in low and mid-latitude regions (Vuille et al., 2003, Lachniet and Patterson, 2009, Medina-Elizalde et al., 2016a, Lases-Hernandez et al., 2019). The working hypothesis of stalagmite hydrological records from these regions is often that calcite and aragonite deposited under isotopic equilibrium conditions preserve the rainfall δ¹⁸O composition thus recording rainfall amount (Burns et al., 2003, Medina-Elizalde et al., 2010). This approach assumes that other potential effects on rainfall δ¹⁸O within these regions may be negligible such as temperature, evaporation, source moisture δ¹⁸O and plant transpiration (Fairchild and Treble, 2009, Wang et al., 2017, Wang et al., 2001).

Although seldom acknowledged in studies of paleohydrological records based on stalagmites, the isotopic relationship between rainfall and drip water is crucial when comparing relative isotopic variations within a single stalagmite and among different stalagmite δ¹⁸O records, and particularly if the goal is to reconstruct precipitation amount quantitatively (e.g. Medina-Elizalde et al., 2010, Lachniet et al., 2012, Lachniet et al., 2017, Medina-Elizalde and Rohling, 2012, Aharon and Dhungana, 2017, Medina-Elizalde et al., 2017). The δ¹⁸O signature of rainfall may be altered between the ground surface and the cave interior due to processes including isotopic fractionation in the soil, epikarst and/or vadose zone, driven by evaporation (Ayalon et al., 1998, Bradley et al., 2010, Cuthbert et al., 2014, Beddows et al., 2016, Hartmann and Baker, 2017), and by mixing with other meteoric water reservoirs in the epikarst, which essentially attenuates the isotopic signal of a rainfall event (Yonge et al., 1985; Ayalon et al., 1998, Williams and Fowler, 2002, McDermott, 2004, Fairchild et al., 2006; Lachniet and Patterson, 2009, Genty et al., 2014, Hartmann and Baker, 2017). The resolution of a stalagmite δ¹⁸O-derived rainfall record, importantly, is not determined solely by the stalagmite sampling resolution, but also by the time integration of the rainfall signal during drip water infiltration (Lases-Hernandez et al., 2019). Drip-specific infiltration pathways can potentially integrate the amount-weighted isotopic signal of rainfall accumulated over days (Luo et al., 2014, Duan et al., 2016), a season (Cruz, 2005, Cobb et al., 2007, Fuller et al., 2008, Genty, 2008, Beddows et al., 2016 Duan et al. 2016), a year or even multiple years (Yonge et al., 1985; Williams and Fowler, 2002, Onac et al., 2008, Genty et al., 2014, Riechelmann et al., 2011, Riechelmann et al., 2017, Czuppon et al., 2018, Jean-Baptiste et al., 2019). Moreover, the drip sites within a single cave can exhibit different responses to the same isotopic signal of the rainfall infiltrating water (Treble et al., 2013, Moerman et al., 2014, Pérez-Mejías et al., 2018, Lases-Hernandez et al., 2019).

Drip water isotopic information, therefore, is critical to validate studies with high-resolution stalagmite δ¹⁸O records that seek to characterize seasonal to interannual precipitation variability (Medina-Elizalde et al., 2010, Medina-Elizalde and Rohling, 2012, Kennett et al., 2012, Lachniet et al., 2012, Lachniet et al., 2017) or to discern the negative isotopic anomalies of TC rainfall (Frappier et al., 2007a, Baldini et al., 2016).

The present study presents new drip water isotopic data from drip sites monitored from June 2017 to April 2019 in the Río Secreto cave system, located in the northeastern coast of the Yucatán Peninsula (YP), Mexico (Fig. 1). These drip sites are 7 of the 16 previously examined by Lases-Hernandez et al. (2019) (hereafter LH19), and were selected because they represent distinctive isotopic patterns with variable implications for paleoclimate reconstruction. In this study we applied a new sampling protocol different from that of LH19 which: (i) enabled the characterization of all the water discharged at these 7 drip sites over two years; (ii) help test the influence of LH19’s sampling protocol on a positive isotopic bias observed at a drip site with a small reservoir size (labelled LF1) (details in Section 2.2); (iii) enabled estimates of rainfall integration times and degree of homogenization, stratified by reservoir, of six of these seven drip sites, from bi-weekly, monthly and annual drip water samples, and; (iv) help test the notion that the isotopic composition of drip water at different drip sites converge into a single value when drip water is integrated over a sufficiently long period and provided they reflect the same water source and no other process, such as evaporation, has altered the isotopic composition of the water source. In addition, this study explores the implications of observed drip water variability and residence times for the detection of significant precipitation reductions (i.e. droughts) and TCs from stalagmite δ¹⁸O records. Lastly, we provide the first instrumental evidence of the existence of a rainfall amount effect on interannual timescales for the YP.