Reconstructing precipitation in the tropical South Pacific from dinosterol 2H/1H ratios in lake sediment
Introduction
The South Pacific Convergence Zone (SPCZ) supplies freshwater to 11 million people on 3975 islands (Power et al., 2011, WHO, 2016). The majority of coupled climate models suggest a larger and wetter SPCZ in the future with a more extreme zonal structure during ENSO events (Brown et al., 2011, Cai et al., 2012, Brown et al., 2012, Borlace et al., 2014). It is difficult to assess these model projections because little is known about the natural variability of precipitation in the SPCZ region owing to a short instrumental record of just 40 years from satellites and no more than 130 years from gauges (Vincent, 1994). Efforts to extend the instrumental record of SPCZ precipitation with paleoclimate proxy data are needed to validate model hindcasts.
Speleothems from Vanuatu (Partin et al., 2013) and the Solomon Islands (Maupin et al., 2014) in conjunction with coral archives from Fiji and Rarotonga (Linsley et al., 2004, Linsley et al., 2006), Vanuatu (Quinn et al., 1993), New Caledonia (Quinn et al., 1998, DeLong et al., 2012), and the nearby Great Barrier Reef (Druffel and Griffin, 1993, Hendy et al., 2002) and Coral Sea (Calvo et al., 2007) provide insights into pre-instrumental-era climate changes, but do not extend prior to 600 years ago in this region. Sediment archives have successfully recorded longer-term changes in tropical vegetation and human activity (Southern, 1986, Parkes, 1994, Hope and Pask, 1998, Stevenson and Hope, 2005, Wirrmann et al., 2006, Prebble and Wilmshurst, 2009, Prebble et al., 2013, Combettes et al., 2015, Rull et al., 2015), and large scale precipitation during the past 10,000 years (Hassall, 2017). However, additional proxy development and calibration work is required to develop quantitative rainfall records that can be compared to climate model output.
In the maritime tropics, the hydrogen isotope ratio (2H/1H, expressed as δ2H (= [Rsample/RVSMOW] − 1, where R is 2H/1H and VSMOW is Vienna Standard Mean Ocean Water) of precipitation is principally controlled by the precipitation rate when averaged over monthly and longer time scales (Dansgaard, 1964, Rozanski et al., 1993, Risi et al., 2008, Kurita et al., 2009). Tropical island lake water δ2H values reflect this signal, plus the added effects of evaporative enrichment (Kebede et al., 2009, Garcin et al., 2012, Issa et al., 2015). Lipids from microalgae have δ2H values that track the isotopic composition of their environmental water with R2 values typically >0.9 (Paul, 2002, Englebrecht and Sachs, 2005, Schouten et al., 2006, Zhang and Sachs, 2007, Sachse et al., 2012), and their preservation in sediments offers a window into past lake hydrology, and by extension past precipitation. Despite this, limited region-specific calibration work has been conducted to understand and quantify the nature of the climate signal contained in the δ2H values of sedimentary lipids.
Phytoplankton produce a variety of taxon-specific lipids (Volkman et al., 1998, Volkman, 2005) that can be exploited to infer environmental conditions. Dinosterol (4α,23,24-trimethyl-5α-cholest-22E-en-3β-ol) is produced almost exclusively by dinoflagellates, although it has also been detected in a single marine diatom species (Volkman et al., 1993). The relative source-specificity of dinosterol offers an advantage over more generic lipids such as fatty acids and has been used to reconstruct pre-instrumental hydrology across the tropical Pacific (Sachs et al., 2009, Smittenberg et al., 2011, Atwood and Sachs, 2014, Nelson and Sachs, 2016, Richey and Sachs, 2016).
In addition to source water isotopic composition, algal lipid δ2H values are sensitive to environmental parameters including salinity (Schouten et al., 2006, Sachse and Sachs, 2008, Sachs and Schwab, 2011, Nelson and Sachs, 2014b, Maloney et al., 2016, Sachs et al., 2016, Weiss et al., 2017), growth rate and phase (Zhang et al., 2009b, Wolhowe et al., 2009, Chivall et al., 2014, Sachs and Kawka, 2015, Wolhowe et al., 2015), temperature (Zhang et al., 2009b, Wolhowe et al., 2009, Ladd et al., 2017), light (van der Meer et al., 2015, Sachs et al., 2017), and redox environment or metabolism (Schwab et al., 2015a). This highlights the need for careful lipid and site selection in attempts to use sedimentary biomarkers to reconstruct past environmental conditions.
Previous surficial lake sediment studies have demonstrated that δ2H values of sedimentary molecular fossils from a variety of terrestrial and aquatic photoautotrophs largely reflect modern spatial climate gradients. Sampling regions include global (Nelson and Sachs, 2014b) and continental transects across Europe (Sachse et al., 2004), the Americas (Polissar and Freeman, 2010), North America (Sauer et al., 2001, Huang et al., 2004, Nelson and Sachs, 2014a), southwest United States (Hou et al., 2008), Cameroon (Garcin et al., 2012, Schwab et al., 2015a, Schwab et al., 2015b), and the Tibetan Plateau (Mügler et al., 2008, Xia et al., 2008, Aichner et al., 2010). However, the fidelity with which lacustrine photoautotrophic lipid δ2H values track climate in lakes across the vast expanse of the tropical South Pacific Ocean has not yet been tested, nor has a calibration for rainfall been established that can be used to quantitatively reconstruct past precipitation rates.
This study develops and validates a quantitative rainfall proxy based on the δ2H value of dinosterol from recent lake sediments across the tropical South Pacific Ocean. Although variable secondary influences on 2H/1H fractionation have been used to enhance the level of paleoclimate information recovered from sedimentary biomarkers (Nelson and Sachs, 2016), in the present study we sought to limit potential impacts of salinity or temperature by targeting only freshwater lakes spanning a small range of seasonally stable temperatures (<4°C, Table A.1, Appendix A). To compensate for unconstrained parameters such as species composition, growth rate, metabolism, and light, which may influence fractionation, as well as changes in moisture source, temperature, and secondary evaporation, which may influence lake water isotopes, we sampled multiple lakes in each region. The results demonstrate that δ2Hdinosterol values in lake sediments can be used to reconstruct mean annual precipitation rates in the tropical South Pacific.
Section snippets
Sample collection
Sediment core top samples were collected from 21 lakes (Fig. 1; Table B.1, Appendix B). Most core top samples came from 6.6 cm diameter sediment cores collected in 2011–2012 with a universal corer device (Aquatic Research, Hope, ID), but the Lake Wanum sample was collected in 1999. The uppermost portion (typically 20–50 cm) was sectioned in the field at 0.5 or 1 cm intervals to a depth at which the sediment was sufficiently consolidated to transport and store without disturbance to
Study sites
The lakes in this study (Fig. 1; Table 1; Table A.1, Appendix A) were chosen to create a modern spatial calibration between δ2Hdinosterol and SPCZ precipitation. The lakes span 123.7° of longitude, 21.4° of latitude, and a narrow 23–27 °C range of NCEP/NCAR reanalysis mean annual air temperature where the largest (smallest) seasonal/inter-annual variability is 8.3 (3.7) °C in New Caledonia (Wallis) over the last 69 years (Fig. A.4, Appendix A). The lakes are all less than 250 m above sea level
Precipitation δ2H and δ18O values
There are 15 IAEA/GNIP stations and 1 JAMSTEC station (Palau) (Fig. 1) on tropical Pacific islands between 30°N and 30°S with at least 23 δ2H measurements from at least three years (Table A.2, Appendix A). The 9 mm d−1 range in station precipitation rates vs. the 44‰ range in mean annual amount-weighted δ2Hrain from the 16 stations yields a tropical Pacific amount effect with a slope of −4.5 ± 0.6‰ (mm d−1)−1 [r = −0.65, R2 = 0.42, p ≪ 0.001], note the Hilo station data fell well above the 95%
δ2Hdinosterol values
The slope of the δ2Hdinosterol–GPCP relationship (−12.1 ± 2.6‰ (mm d−1)−1) was similar to the δ2Hlakewater–GPCP relationship (−10.1 ± 2.4‰ (mm d−1)−1), and both were steeper than the δ2Hrain vs GPCP rainfall (amount effect) relationship from GNIP/JAMSTEC stations (−5.6 ± 0.8‰ (mm d−1)−1) (Fig. 2). This is due to the enriching effects of evaporation on δ2Hlakewater values that also manifests in δ2Hdinosterol values from dinoflagellates that use this lake water as a primary hydrogen source. Fig. 2
Conclusion
We investigated the environmental controls on the 2H/1H ratios of SPCZ lake waters and of the dinoflagellate sterol, dinosterol, purified from recent freshwater lake sediments. The 30 lakes in this study represent diverse freshwater ecosystems with varying physical and chemical parameters. The main controls on δ2Hlakewater were precipitation rate and evaporation. Additional lake-specific environmental properties such as lake surface area-to-volume, water residence time, trajectories of source
Acknowledgements
This material is based upon work supported by the US National Science Foundation under Grant Nos. 0823503 and 1502417 to J.P.S; a University of Washington Program on Climate Change Fellowship and IGERT Ocean Change Fellowship [grant #NSF1068839] provided partial support for A.E.M; radiometric dating was supported by AINSE (The Australian Institute of Nuclear Science and Engineering) [Award No ALNGRA14513 and ALNGRA16012] to M.P.; IGERT Ocean Change mini-research awards, and a University of
References (126)
- et al.
δD values of n-alkanes in Tibetan lake sediments and aquatic macrophytes – a surface sediment study and application to a 16ka record from Lake Koucha
Org. Geochem.
(2010) - et al.
Separating ITCZ- and ENSO-related rainfall changes in the Galápagos over the last 3 kyr using D/H ratios of multiple lipid biomarkers
Earth Planet. Sci. Lett.
(2014) - et al.
Characterization of unusual sterols and long chain diols, triols, keto-ols and n-alkenols in El Junco Lake, Galápagos
Org. Geochem.
(2014) - et al.
Interdecadal climate variability in the Coral Sea since 1708 A.D
Palaeogeogr. Palaeoclimatol. Palaeoecol.
(2007) - et al.
The effects of growth phase and salinity on the hydrogen isotopic composition of alkenones produced by coastal haptophyte algae
Geochim. Cosmochim. Acta
(2014) - et al.
High-resolution pollen record from Efate Island, central Vanuatu: Highlighting climatic and human influences on Late Holocene vegetation dynamics
Comptes Rendus Palevol
(2015) - et al.
Hydrogen isotope fractionation during lipid biosynthesis by Haloarcula marismortui
Geochim. Cosmochim. Acta
(2013) - et al.
Hydrogen isotope fractionation during lipid biosynthesis by Tetrahymena thermophila
Org. Geochem.
(2013) - et al.
Determination of sediment provenance at drift sites using hydrogen isotopes and unsaturation ratios in alkenones
Geochim. Cosmochim. Acta
(2005) - et al.
Fluvial carbon fluxes under extreme rainfall conditions: inferences from the Fly River, Papua New Guinea
Chem. Geol.
(2011)