Cyclic changes in total solar irradiance (TSI) during the Holocene are known to affect global climatic conditions and cause cyclic climatic oscillations, e.g., Bond events and related changes of environmental conditions. However, the processes how changes in TSI affect climate and environment of the Southern Hemisphere, especially in southernmost South America, a key area for the global climate, are still poorly resolved. Here we show that highly sensitive proxies for aquatic productivity derived from sediments of a lake near the Chilean South Atlantic coast (53 °S) strongly match the cyclic changes in TSI throughout the Holocene. Intra-lake productivity variations show a periodicity of ~200–240 years coherent with the time series of TSI-controlled cosmogenic nuclide 10Be production. In addition TSI dependent periodicity of Bond events (~1500 years) appear to control wetness at the LH site indicated by mineral matter erosion from the catchment to the lake assumingly through shifts of the position of the southern westerly wind belt. Thus, both intra-lake productivity and wetness at the southernmost South America are directly or indirectly controlled by TSI.
The recent climate between 41° and 53°S is mainly controlled by the southern westerly wind belt (SWW), which affects the southern ocean circulation and consequently the global climate1,2 and changes in seasonal insolation cause variations in the position and intensity of the SWW3,4,5,6. Moreover, precipitation at the hyper-humid western side of the Andes has been linked to solar activity through SWW shifting during the Late Holocene, although no cyclicity could be found2,7.
During the Holocene, one of the strongest effects on the Northern Hemisphere (NH) climate is attributed to the 1500 yrs cycle in solar output, known as Bond Cycles8,9. Bond Cycles lead to periodic cooling and variations in the production of North Atlantic deep water and have also affected biogenic productivity in terrestrial aquatic systems10. In the Southern Hemisphere several studies have associated abrupt centennial to millennial scale variations in the mid-latitude records with Bond events such as the South America monsoon system (SAMS), although the physical mechanisms linking both are yet unclear11,12,13,14.
Recently, the relevance of a 200–250 years periodicity in the Southern Hemisphere climate has been demonstrated. A stratigraphic record from Lago Cipreses in Southwestern Patagonia (51°S) reveals recurrent ~200 years positive Southern Annual Mode-like (SAM) phases (dry/warm) during the last three millennia15. The authors noted the common structure of centennial changes in SAM with paleoclimate records from the NH, and an in-phase interhemispheric coupling thought atmospheric teleconnections15. Similar, a recent study of enhanced westerly intensity during the last 2,600 years at the Falkland Islands (Patagonia Atlantic coast) show the same periodicity, coherent with radiocarbon production rates, suggesting that solar forcing plays a dominant role in modulating the strength of the SWW16. Moreover, a short (1,500 years) but high-resolution speleothem record from south-central Brazil showed the sensitivity of the SAMS to solar forcing with a cyclicity of 208 years.
All these studies are based on paleoclimate records reaching back in time for not more than 3,000 years, so no information of previous periods is yet available and the processes behind Holocene climatic changes associated with short-term cyclic variations in TSI are still unclear. One reason might be that high-resolution proxies, sensitive to TSI are not easily available for the southernmost SH because of the small landmass and the extreme climatic zonation of this key region for the ocean–atmospheric circulation system of southern mid to high latitudes.
Here, we report results from a multi-proxy (Fourier transform infrared spectroscopy (FTIR) and geochemical data) investigation of a 12,500 cal BP lake sediment record from Southern Patagonia (Strait of Magellan). We used multiple principal component analysis (PCA) to extract productivity indicators as well as climate-dependent chemical signals from the entire FTIR data and from the upper core section (~ 5,500 cal BP). Moreover, evaluation of Holocene TSI cyclicity has been based on 10Be concentrations derived from ice cores17 to evaluate the effect of TSI variations on aquatic productivity.
Sedimentary and climatic history of Lago Hambre
Lago Hambre (LH) (Fig. S1) is a small 17 m deep lake, located 50 km south of Punta Arenas (53°36′ 13.19″S, 70°57′ 8.77″ W). The lake lacks surface inflows, but gullies reflect intermittent surface inflow during periods of intense precipitation. The sedimentary and climatic history of LH is characterized by steadily increasing rates of organic carbon accumulation from the early Holocene until ~5,500 cal BP, when the record is interrupted by a series of intense erosion events (Fig. 1). These have been interpreted as storm events during a period of climate transition to warmer and wetter conditions6,18. A global set of environmental archives support a major global climate change between 5,600 and 5,000 cal BP e.g.19,20,21. Moreover, tephra layers at 4,250, 8,200, 8,800, and 14,300 cal BP further are intercalated within the sedimentary record, which are no direct indicators of climatic change.
In the past 4,000 cal BP, organic carbon accumulation in LH was generally higher and more variable than in previous periods (Fig. 1). Carbon/nitrogen ratios (C/N) increased from an average value of 8 to about 18, indicating organic matter (OM) sources, shifted from a dominance of algae-derived OM between 12,500 cal BP and 5,500 cal BP to increasing fluxes of terrestrial OM to the lake due to the development of Nothofagus forests in the lake’s catchment after 11,000 cal BP22 (Fig. 1). Beside the general trend of increasing organic carbon accumulation, a strong short-term variability in OM fluxes from the two different carbon sources becomes most pronounced in the past 4,500 cal BP (Fig. 1). Accumulation of eroded allochthonous minerogenic material (MM) is inversely related to that of carbon, and show a continuous decrease since about 11,000 cal BP until present. This general decrease in MM fluxes to the lake is mainly attributed to the development of the organic-rich forests soils in the past 4,500 cal BP which diminish erosion of MM.
Productivity proxies from FTIR data
Most of the variance in the FTIR data extracted by PCA from the entire data set was explained by changes in OM- (58%, PCA-Cp1) and MM accumulation (40%, PCA-Cp2) (see supplementary discussion). Using the FTIR data of the past 5,500 cal BP, PCA1 extracted a signal (PCA1-Cp9) based on spectroscopic bands that are characteristic for aliphatic compounds associated with green algae (Figure S3). This specific spectrum is related to algaenan23, an insoluble, non-hydrolysable, and highly aliphatic macromolecule that serves as a structural component in the cell wall of freshwater green algae23. This component shows a significant (80%) cyclicity of 203 years since 5,500 cal BP (Fig. S4). Accordingly, positive loadings of PCA1-Cp9 indicate increased aquatic productivity. This interpretation is supported by a significant positive correlation (r = 0.74, n = 226) of the algaenan signal with the hydrogen index (HI) (Fig. 1), a proxy of the source of OM24, where high values of HI are typical for autochthonous OM25. PCA components extracted from FTIR data of sediments accumulated between 12,500 cal BP and 5,500 cal BP differed from those extracted from the past 5,500 cal BP. Most importantly, the signal of algaenan (PCA1-Cp9) was absent in the PCA components extracted from the data prior to 5,500 cal BP. Instead, a component, based on bands typical for biogenic silica (PCA2-Cp1) (Fig. S3) was detected indicating predominance of diatoms as primary producers.
At the end of the Pleistocene and early Holocene (~12,500 cal BP to ~11,000 cal BP), average C/N ratios below 10 indicate autochthonous productivity as the major source of OM in the sediments. Simultaneously, carbon accumulation and HI values both were very low indicating relatively little productivity and high OM decomposition, respectively (Fig. 1). Carbon accumulation and increasing BSi and HI values suggest that in-lake productivity increased from 11,000 to 6,800 cal BP followed by increasing fluxes of OM from the catchment until 5,500 cal BP (Fig. 1). Despite the general increase in terrestrial organic matter fluxes, in the upper part of the record, the C/N ratios and average HI values indicate a mix between contributions from aquatic productivity and terrestrial sources and oscillation in the prevalence of the two sources. The correlation between the two sources is negative (r = −0.64) supporting the interpretation of change of the major source: low C/N ratio and high HI values indicate greater proportion of aquatic (higher productivity) than terrestrial OM and vice versa. Changes in the prevalence of the OM source point to changes in the major controlling factors TSI, temperature, and wetness (precipitation). No positive correlation between productivity proxies (PCA1-Cp9, HI and PCA2-Cp1) and inputs of inorganic matter from the catchment could be found, which excludes nutrient fluxes as the major driver behind changes in intra-lake productivity. Moreover, no trends or cyclic fluctuations in temperature in the area during the Holocene have been reported26 which could explain the observed changes in intra-lake productivity. Thus, the variability in OM accumulation and especially in proxies indicating changes in the intensity of aquatic production, algaenan (PCA1-Cp9) and BSi, (PCA2-Cp1) imply that the cyclic changes in solar insolation must be the major factor driving these changes similar to what has been observed in lakes in the Arctic10.
Solar insolation changes and long-term aquatic productivity
When the long–term development of summer solar insolation at 50°S27 is compared to C/N ratios, PCA2-Cp1 (BSi) and carbon accumulation between 12,000 and 5,500 cal BP it becomes evident that summer (Dec-Feb) insolation is the main driver of long-term productivity in Lago Hambre (Fig. 1). After 5,500 cal BP, the generally high accumulation of OM in the sediments is attributed to a further increase in summer insolation as well as wetter and warmer conditions favouring both aquatic productivity and fluxes of catchment-derived OM (Fig. 1).
The long-term development of summer insolation from low values during the early- to higher values during the middle- and late Holocene (Fig. 1) also explains the long-term evolution of the two different signals for aquatic productivity (algaenan (PCA1-Cp9) and BSi (PCA2-Cp1)) extracted from the FTIR data. Because green algae require higher irradiance to grow than most cyanoprokaryotes and diatoms, they generally possess a higher photosynthesis saturation parameter Ik. Accordingly, lower insolation during early- and mid-Holocene at 50°S favours diatoms (BSi) over green algae abundance and vice versa during the past 5,500 cal BP, when insolation was generally higher (Fig. 1).
Changes in averaged total solar irradiance (ΔTSI)17 show the same trend than changes in aquatic productivity as indicated by the intensity of the algaenan (PCA1-Cp9) and BSi (PCA2-Cp1) signals (Fig. 2). This relationship is pronounced during the past 4,500 cal BP when summer insolation was generally high and green algae were dominant. However, during some periods (3,000–2,000 cal BP and 5,500 to 4,500 cal BP), this relationship appears to be partly overwritten by regional or local erosion events and becomes blurred. Between 9,300 cal BP and 5,500 cal BP, the relation between aquatic productivity (BSi, PCA2-Cp1) and ΔTSI remains, but is less pronounced due to the lower resolution in the lower core section. Additionally, the BSi (PCA2-Cp1) and ΔTSI records seem to show an offset of ~200 years, probably due to the low resolution of the BSi-data at this core section. However this ~200-year offset is within the 2σ uncertainty of the calibrated 14C dates, why we shifted the data here by 200 years
Spectral analyses of productivity proxies reveal significant periodicities of 203 years for PCA1-Cp9 and 201 and 240 years for HI (see S4). Both periodicities have been previously observed in records of the TSI controlled records of atmospheric 14C and 10Be28,29, which points to a relation between intra-lake productivity in LH and the Vries solar cycle30. Previous studies reported this periodicity for productivity records in the Southern Ocean31,32.
Location and Sampling
Lago Hambre (Fig. S1) is located 50 km south of Punta Arenas, near the Strait of Magellan, in southernmost Patagonia, Chile (53°360 13.1900 S, 70°570 8.7700 W). It is situated about 80 m a.s.l. and is a small lake with a surface area of 13,700 m2 and a maximum depth of 17 m. The catchment-to-lake ratio is about eight. The lake is situated in the zone of sub-Antarctic deciduous forest with Nothofagus pumilio and Nothofagus antarctica.
A long sediment core was recovered from the deepest part of the lake in 2008 using a 5-m-long piston corer. The core was recovered in three single drives, with an offset of about 90 cm to ensure overlaps. Sediment cores were stored under dark and cool (4 °C) conditions in the laboratory before they were split lengthwise using a stainless steel saw. Sub-samples were taken at 1 cm intervals in the upper 470 cm of the piston core, and at 2-cm intervals below 470 cm.
Concentrations of carbon (C) and nitrogen (N) in the carbonate-free samples were determined by GC-TCD after thermal combustion in an elemental analyzer (Euro EA3000, Eurovector, Germany). Samples were analyzed for zirconium (Zr), using an energy-dispersive XRF mini-probe multi-element analyzer (EMMA), see refs 24 and 46.
The hydrogen index (HI) i.e. the quantity of pyrolyzable hydrocarbons (S2) per gram TOC (mg HC gC−1), was determined by means of a Rock–Eval-II-plus-S3-unit at the Alfred Wegener Institut für Polar- und Meeresforschung in Bremerhaven, Germany, using sample aliquots of 30–50 mg.
The age-depth model is based on 18 Accelerator Mass Spectrometry (AMS) 14C dates on terrestrial macro-fossils, i.e. leaves and wood, as well as a tephra layer of known age (Mt Burney, 4,250 cal yr BP, ref. 39). For a more detailed description of the age model see refs 24 and 46 and supporting information.
Infrared Spectroscopy (IR)
A total of 650 samples were analysed by Infrared Spectroscopy (IR). FTIR spectra of freeze-dried and ground samples of lake sediment were obtained using a Vector 22 FTIR spectrometer (BrukerOptik, Ettlingen, Germany) in absorption mode, with subsequent baseline subtraction on KBr pellets (200 mg dried KBr and 2 mg sample). Measurements were recorded from 4,500 to 300 cm−1 using a resolution of 2 cm−1. Thirty-two scans were taken per sample and averaged to obtain the final spectra. In order to improve the comparison between samples, each FTIR spectrum was re-scaled to relative absorbance (i.e. dividing the absorbance value of each point of the spectrum by the summed values of the analysed spectral region).
Principal component analysis (PCA) was performed on the IR data to identify the main chemical signals of the lake core and relate them to underlying environmental processes in the lake and its catchment. To maximize the loadings of the original variables, a varimax rotation was applied. In this study, we performed the PCA on a transposed matrix (samples as columns and absorbance bands as rows). In contrast to PCA on direct matrices, that on the transposed matrix provides scores for the spectral bands and loadings for the samples. Each extracted component is represented by a spectrum of scores, which permits to identify the most relevant signals characterizing the composition of the samples, whereas sample’s loadings (or their squared values) indicate the weight of the identified compounds on the samples’ spectroscopic signal (related to its abundance in the samples).
We performed a PCA using all samples. In addition, due to the large compositional differences in the sediment along the core found in previous studies22,30, we also analyzed the main two sections separately: PCA1 for the samples between 94–6,273 cal BP and PCA2 for the samples 4,591–15,445 cal BP.
How to cite this article: Pérez-Rodríguez, M. et al. Solar Output Controls Periodicity in Lake Productivity and Wetness at Southernmost South America. Sci. Rep. 6, 37521; doi: 10.1038/srep37521 (2016).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported by a DFG-grant to H.B. (DFG-BI 734/15-1) and support to M.P.-R. by DAAD (grant ref no. 91588482) and TU-Braunschweig. We thank R. Kilian and H. Arz for LH core samples, A. Calean and P. Schmidt for technical assistance.
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/