Quaternary Science Reviews 303 (2023) 107968 Contents lists available at ScienceDirect Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev Recent global warming induces the coupling of dissimilar long-term sedimentary signatures in two adjacent volcanic lakes (Azores Archipelago, Portugal) zquez-Loureiro a, Alberto Sa David Va ez b, Vítor Gonçalves c, d, Teresa Buchaca e, Armand Herna a  g, h, i, ndez , Pedro M. Raposeiro c, Erik J. de Boer b, f, Pere Masque j a, * Santiago Giralt , Roberto Bao a Universidade da Corun ~ a, GRICA Group, Centro Interdisciplinar de Química e Bioloxía (CICA), Rúa As Carballeiras, 15071, A Corun ~ a, Spain b Department of Earth and Ocean Dynamics, UB-Geomodels Research Institute, Universitat de Barcelona, Martí I Franqu es s/n, 08028, Barcelona, Spain c CIBIO, InBIO Laboratorio Associado, Centro de Investigaça~o Em Biodiversidade e Recursos Gen lo Dos Açores, Rua da Ma eticos, Po ~e de Deus, 9500-321, Ponta Delgada, Portugal d Faculdade de Ci^ encias e Tecnologia, Universidade Dos Açores, Rua da Ma ~e de Deus, 9500-321, Ponta Delgada, Portugal e Centre D'Estudis Avançats de Blanes (CEAB-CSIC), C/D'acc es a La Cala St. Francesc, 14, Blanes, 17300, Girona, Spain f Departamento de Estratigrafía y Paleontología, University of Granada, Avenida de La Fuente Nueva S/N, 18071, Granada, Spain g International Atomic Energy Agency, 98000, Principality of Monaco, Monaco h Institute of Environmental Science and Technology and Physics Department, Universitat Autonoma de Barcelona, Bellaterra, 08193, Spain i School of Natural Sciences, Centre for Marine Ecosystems Research, Edith Cowan University, Joondalup, WA 6027, Australia j Geosciences Barcelona (Geo3BCN-CSIC), Lluís Sol e I Sabarís S/n, 08028, Barcelona, Spain a r t i c l e i n f o a b s t r a c t Article history: Paleoclimatological information derived from the study of lacustrine sedimentary records is not only Received 24 September 2022 biased by taphonomical processes but also by potential differences in the expression of climate vari- Received in revised form ability in the sediments due to site-specific factors. Using a multiproxy approach (the elemental and 16 January 2023 isotopic compositions of organic matter, diatom assemblages, and marker pigments of algae and cya- Accepted 17 January 2023 Available online xxx nobacteria), we study the different environmental signatures recorded since the Little Ice Age (LIA) in the sediments of two volcanic lakes located within the same caldera on Sa ~o Miguel Island (Azores Archi- Handling Editor: P Rioual pelago). Lake Santiago is a crater lake whose eutrophic status in the last stage of the LIA was linked to external nutrient inputs associated with this humid period. Its post-LIA evolution was forced by changes Keywords: in the thermal structure of the water, which determined its degree of mixing and therefore nutrient Paleolimnology availability through recycling from the hypolimnion. In contrast, the decadal to centennial limnological Diatoms evolution of Lake Azul, a caldera lake 2.5 km from Lake Santiago, shows geochemical and micropale- Fossil pigments ontological signatures disconnected from climate variability until 1980/1990 CE due to its greater Global warming exposure to the fallout of tephra after a catastrophic eruption in c. 1290 CE. Only after 1980/1990 CE did a Holocene climate global warming scenario induce a common ecological restructuring of both lakes, involving the Remote islands replacement of turbulence-loving algal taxa by species adapted to strengthening water column strati- fication. Nevertheless, this shift was relatively gradual in Lake Azul but more sudden in Lake Santiago, indicating that the local site-specific components still had an effect on the expression of climate change in the sediments. Despite the short history of anthropogenic pressure (compared to their continental counterparts) and the large atmospheric patterns operating over the Azores Archipelago, the sedimen- tary records of these two adjacent oceanic volcanic lakes reacted quite differently to climate changes. © 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). * Corresponding author. Universidade da Corun ~ a, GRICA Group, Centro Interdisciplinar de Química e Bioloxía (CICA), Rúa As Carballeiras, 15071, A Corun ~ a, Spain. E-mail addresses: david.vazquez@udc.es (D. V azquez-Loureiro), a.saez@ub.edu (A. Sa ez), vitor.mc.goncalves@uac.pt (V. Gonçalves), buch@ceab.csic.es (T. Buchaca), armand.hernandez@udc.es (A. Hern andez), pedro.mv.raposeiro@uac.pt (P.M. Raposeiro), deboer.erikjan@gmail.com (E.J. de Boer), p.masque@ecu.edu.au (P. Masque ), sgiralt@geo3bcn.csic.es (S. Giralt), roberto.bao@udc.es (R. Bao). https://doi.org/10.1016/j.quascirev.2023.107968 0277-3791/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 1. Introduction inertia forced by this natural catastrophic episode (V azquez- Loureiro et al., 2019). An intersite comparison of lake paleo- Natural and human-induced impacts on ecosystems are char- records should therefore show either a good correspondence of acterized by large spatial and temporal variability in their nature, signals, thereby reflecting regional-scale climatic and environ- duration, and intensity (Jackson and Hobbs, 2009). When studying mental change, or a poor fit resulting from site-specific local processes operating at temporal scales that are much larger than components or ‘noise’ (Roberts et al., 2016). those typically addressed by neoecologists, paleoecological studies Thus, we can hypothesize that large differences in the sensitivity are highly relevant (Louys, 2012). With its decadal to millennial of volcanic lakes to climate variability should be attenuated when perspective, paleoecology provides a unique understanding of how the lakes are separated by short distances and fall within the same ecosystem services can be preserved over time, as well as infor- volcanic center. Here, we present the paleoecological evolution of mation on baseline environmental contexts and ecosystem alter- Lake Santiago, a deep volcanic lake located in the Caldeira das Sete native stable states (Jeffers et al., 2015). Cidades on S~ao Miguel Island (Fig. 1), and compare it to that of Lake However, the reliability of paleoecological archives depends on Azul, which is located in the same volcanic center, just 2.5 km from several factors. Taphonomical processes, such as selective preser- Lake Santiago. We use instrumental climate data, as well as the vation, postmortem transport, time averaging or diagenetic effects, independent paleoclimate record of Lake Empadadas (Hern andez are known to bias the information provided by the fossil record (e. et al., 2017), located 5.0 km outside of the Caldeira das Sete g., Raposeiro et al., 2018; Saraswati and Srinivasan, 2015). It is less Cidades, to determine the major climate phases of the archipelago known how the signatures of the environmental drivers recorded since 1350 CE and how the two deep lakes, Lake Santiago and Lake in the sedimentary archive are influenced by site-specific differ- Azul, have responded to climate changes. ences, despite being essential for addressing the effects of recent climate change (Hargan et al., 2016; Seddon et al., 2014). 2. Study site Lakes act as excellent sentinels of environmental change (Williamson et al., 2008, 2009), but they integrate local and global The Azores Archipelago comprises nine main islands of volcanic signals into sedimentary records previously modulated by the at- origin that have developed at the triple junction of the American, mosphere and lake catchments (Leavitt et al., 2009; Ritter et al., African and Eurasian tectonic plates (Andrade et al., 2021) (Fig. 1A). 2022). Site-specific variability affecting soil type, vegetation, The caldera of Sete Cidades represents an active volcanic center topography, or basin morphometry, among others (Jovanelly and located on the western part of the island of Sa~o Miguel (Fig. 1B) and Fritz, 2011), is already known to account for many of the differ- contains a set of lakes with contrasting morphologies and degrees ences in the expression of environmental drivers of change in of volcanic activity. Lake Santiago (37 500 N, 25 460 W) is in the lacustrine sedimentary records (Hargan et al., 2016; Rühland et al., southeasternmost part of the caldera, while Lake Azul (37 520 N, 2015). This expression can also differ in a single lake throughout its 25 46’ W) is in the northernmost part. development, resulting in a lack of linear responses of the sediment The basin of Lake Santiago (Fig. 1D) occupies the bottom of a records to climate forcing (Bao et al., 2015; Herna ndez et al., 2008). crater originating from hydromagmatic eruptions (360 m asl) and is Therefore, fundamental questions arise about the possibility of surrounded by steep walls composed of pumice rocks. The walls are disentangling climate vs. nonclimate forcings when analyzing lake covered by a woodland dominated by the introduced coniferous sediments (Mills et al., 2017). tree Cryptomeria japonica (Thunb. Ex L. f.) D. Don. The lake has an The sedimentary records of temperate low-elevation lakes on elliptical shape, with its maximum axis oriented NeS, an area of oceanic islands can potentially better preserve climatic information 0.25 km2, and maximum and average depths of 32 and 13 m than their continental counterparts, since the human occupation of (Fig. 1D), respectively (Pereira et al., 2014). The surface area to such areas often occurred in recent historical times and usually catchment ratio in Lake Santiago is 0.23 (Supplementary Table 1). began with low population densities (Keegan and Diamond, 1987; Lake Azul is the largest lake of the Azores, with a surface of Nogue  et al., 2021), resulting in much shorter time spans of human 3.59 km2 and a maximum depth of 28 m (Fig. 1C), and is located disturbances. In addition, the climate of oceanic islands is usually inside the caldera (259 m asl) yielding a surface area to catchment controlled by large atmospheric and oceanic patterns and is less ratio of 0.31. The lake basin has two main morphological areas: a impacted by the continental effects of local convection (Sa ez et al., shallow platform (0e12 m) and a deep plateau (25e28 m). Both 2009). Both conditions, namely, only recent anthropogenic pres- Lake Santiago and Lake Azul exhibit monomictic characteristics, sures and large atmospheric and oceanic patterns governing with stratification in the summer period. climate, are present in the Azores Archipelago, located in the The archipelago features an oceanic temperate climate, with middle of the Atlantic Ocean (Fig. 1). On the one hand, the coloni- mild temperatures, frequent winds, high relative air humidity, and zation of the Azores extends back to just ca. 700e850 CE, although a precipitation regime with a strong seasonal cycle and high it was not until ca. 1070e1280 CE that human presence became interannual variability (Hernandez et al., 2016). The position of the more prominent on the landscape (Raposeiro et al., 2021; Richter semipermanent high-pressure Azores Anticyclone and the Azores et al., 2022). On the other hand, the climate conditions of the ar- Current (a branch of the Gulf Stream) are the main factors driving chipelago are mostly controlled by the North Atlantic Oscillation climate (Volkov and Fu, 2010). The climate variability of the Azores (NAO), the main large-scale climate mode of variability in this re- Archipelago is further influenced by large-scale climate modes, gion (Herna ndez et al., 2016, 2020). A previous comparison of such as the NAO, the East Atlantic pattern (EA) and the Atlantic instrumental and proxy-based reconstructions for the pre- Multidecadal Oscillation (AMO) (Andrade et al., 2008; Hern andez instrumental period showed that climate was the main control in et al., 2016, 2017; Yamamoto and Palter, 2016). the long-term sedimentary dynamics of Lake Empadadas (Herna ndez et al., 2017), a shallow volcanic lake located on the west 3. Material and methods side of Sa~o Miguel Island (Fig. 1). In contrast, Lake Azul, a deep lake within a volcanic caldera (Caldeira das Sete Cidades), was exposed 3.1. Field methods and core lithologies to a massive influx of tephra after an eruptive event in ca. 1290 CE (Raposeiro et al., 2017) (Fig. 1) and has experienced a long-term A total of six sedimentary cores were retrieved from three ecological trajectory mainly determined by the oligotrophic different locations in Lake Santiago in September 2011 using a 2 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 ~o Miguel Island. B. Location of the Sete Cidades caldera on Sa Fig. 1. A. Location of the Azores Archipelago and Sa ~o Miguel Island with Lake Santiago (LS), Lake Azul (LA) and Lake Empadadas (LE). C. Bathymetric map of Lake Azul showing the location of core AZ11-02 described in Vazquez-Loureiro et al. (2019). D. Bathymetric map of Lake Santiago showing the location of the recovered cores. UWITEC© gravity corer (Fig. 1D): (1) offshore cores SA11-01 A and from bulk organic matter (Supplementary Table 2). SA11-01 B (Site 1: 37 500 5300 N e 25 460 2600 , 27 m), (2) offshore The age model was constructed by linear interpolation between cores SA11-02 A and SA11-02 B (Site 2: 37 500 5600 N e 25 460 2700, the tie-points resulting from the different dating methods. Addi- 31.5 m) and (3) cores SA11-03 A and SA11-03 B (Site 3: 37 500 4900 N tionally, to validate the constructed chronological model, samples - 25 46’ 21” W, 25.3 m). The description of the sediments includes were checked for the first appearances of pollen of the exotic Jap- lithology, texture, color, and structures that allowed the definition anese Cedar (Cryptomeria japonica) and Pinus spp. As their years of of facies and lithostratigraphic units and a correlation among the introduction on the island are known. six cores. A composite sequence (SA11-02) was constructed using cores SA11-02 A and SA11-02 B (Fig. 2), representing sedimentation 3.3. Pollen analyses in a pelagic environment. A multiproxy approach based on diatoms, organic pigments, and the elemental and isotopic compositions of Twelve pollen samples of 0.5 cm3 were strategically taken along the records provides the basis for paleoenvironmental the core. Sample preparation included treatments with KOH, HCl, reconstruction. HF, and acetolysis following standard procedures in unison with other paleoecological studies in the Azores (Raposeiro et al., 2021; 3.2. Dating methods Ritter et al., 2022). Slides were investigated for indicative pollen and spore taxa, as well as nonpollen palynomorphs, related to The relative concentrations of 210Pb were determined by high- ecological shifts, human impact, erosion, fire and other landscape resolution gamma spectrometry (Appleby, 2001), and a constant processes, at  400 magnification. flux:constant sedimentation (CF:CS) model was constructed for the top of the SA11-02 column (Supplementary Fig. 1). Accelerator 3.4. Diatom analyses mass-spectrometry (AMS) 14C dating from different cores, cali- brated to calendar years (cal yr CE) using the CALIB 7.1 software Samples for diatom analysis were treated following standard- (Stuiver, 1993), and the INTCAL13 curve (Reimer et al., 2013) were ized procedures (Renberg, 1990), and slides were prepared with also used for the remaining part of the composite column. Because Naphrax© mountant. At least 400 diatom valves were counted of variations in the reservoir effect in dissolved inorganic carbon at  1000 magnification using a Nikon Eclipse 600 microscope with (DIC) due to the contribution of CO2 of igneous origin (Andrade Nomarski differential interference contrast optics. Taxa were et al., 2019), only the 14C dates from terrestrial plant macrorest identified using standard references (e.g., Krammer, 2000, 2002, samples were used to construct the age model, disregarding those 2003; Krammer and Lange-Bertalot, 1986e1992; Lange-Bertalot, 3 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Fig. 2. Stratigraphic correlation panel of the six cores retrieved in Lake Santiago (location of cores indicated in Fig. 1). Sedimentary units, facies, onlap surfaces and age tie-points are represented. The arrow indicates the presumed slump transport of the deformed materials of Unit 4. Red rectangles in cores 2 A and 2 B indicate the stretches used for the construction of the composite column. 2001; Lange-Bertalot et al., 2011; Levkov, 2009; Levkov et al., 2013), ultrahigh-performance liquid chromatography (UHPLC) following and other sources for specific taxa (e.g., Houk and Klee, 2007; the method described by Buchaca et al. (2019). Pigments were Lange-Bertalot and Ulrich, 2014; Potapova et al., 2008). All identi- identified by comparison with a library of pigment spectra obtained fications were revised according to the available information from from extracts of pure algae cultures from the Culture Collection of the Azores Archipelago (Goncalves et al., 2010). Algae and Protozoa (CCAP, Oban, Scotland, UK) and pigment stan- dards (DHI Water and Environment, Hørsholm, Denmark). Pig- ments with a high taxonomic affinity were selected, including b- 3.5. Pigment analyses carotene (all groups), alloxanthin (cryptophytes), aphanizophyll (N2-fixing cyanobacteria), canthaxanthin and echinenone (mainly A total of 34 samples were used for pigment analyses, using cyanobacteria), diadinoxanthin (mainly dinoflagellates), diatox- freeze-dried sediments stored in dark conditions. The pigments anthin (diatoms), zeaxanthin (chlorophytes and cyanobacteria) and were extracted from sediments using 90% acetone with a probe lutein (chlorophytes). We also analyzed Chl-a and its derivative sonicator (Sonopuls GM70 Delft, The Netherlands) (50 W, 2 min). products, including chlorophyllide a, allomers and epimers of Chl-a, The extract was centrifuged (4 min at 3000 rpm, 4  C), filtered phaeophorbide-a and phaeophythins-a and -b. We defined the sum through Whatman ANODISC 25 (0,1 mm) and analyzed with 4 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 of Chl-a and Chl-a derivatives as a-phorbin. We based interpreta- (Queiroz, 1997). This major eruptive episode has been identified in tion on carotenoid profiles instead of phorbins because there is a the sedimentary record of Lake Azul and Lake Empadadas and has higher loss of phorbins than of carotenoids to undetectable col- been dated to ca. 1290 CE (Herna ndez et al., 2017; Raposeiro et al., ourless compounds when pigments are deposited and buried 2017). (Buchaca and Catalan, 2008). Furthermore, we calculated a Chl-a Unit 2. This unit is recovered in all cores and has a minimum preservation index as the Chl-a:a-phorbin ratio. The index is scaled thickness of 50 cm. It is made up of a banded, varicolored lumpy, from 0 (low Chl-a preservation) to 1 (high Chl-a preservation) organic mud and is rich in planktonic diatoms and the remains of (Buchaca and Catalan, 2007). Values for pigment abundances were other lacustrine organisms (Facies F3, Fig. 2). The colors of this unit expressed as mmol g1 TOC. The molecular weights of the different are brown (F3a), yellowish (F3b), olive green (F3c), and orange for a pigments were obtained from Jeffrey et al. (1997). single thin bed (F3d). It also shows several dark brown silty layers (F3e) and thin gray sandy horizons (Facies F4, Fig. 2). Considering 3.6. Geochemistry methods their characteristics and location in the modern lake, the deposits of Unit 2 accumulated in deep, offshore lacustrine conditions. The A Finningan delta Plus EAeCFeIRMS spectrometer at the Cen- top of this unit records the first appearance of pollen from the gics of the Universitat de Barcelona (CCiTUB) ters Científics i Tecnolo introduced coniferous trees Cryptomeria japonica and Pinus spp., was used to determine the total carbon (TC), total nitrogen (TN), indicating that the lower part of the unit was deposited before 1864 and isotopes of carbon (d13Corg) and nitrogen (d15Norg) of bulk CE (De Albergaria, 2000; Supico, 1995). organic matter. Since X-ray diffraction analyses did not detect any Unit 3. This unit is recorded in eastern core 3 A and consists of a carbonates, TC equals the total organic carbon (TOC) (Raposeiro nonvolcanic, highly erosive, brown, sandy gravel layer up to 60 cm et al., 2017). For the discrimination of the inorganically bound ni- thick. This facies is barren of fossil remains and displays debris flow trogen content from TN, we used the correction by Talbot (2001). features. We interpret this unit as the local arrival of a high-energy The TOC/TN ratio is therefore referred to as TOC/TNcorr. alluvial flood (Facies F5, Fig. 2). The age of this unit is expected to be close to the age of Unit 4, as both units correspond to the deposit of 3.7. Statistical methods a single high-energy event. Unit 4. This unit, recovered in all cores, corresponds to a gray Statistical analyses for the diatom data were performed on the interval up to 75 cm thick overlapping Units 2 and 3. There is a square-root transformed relative abundance matrix expressed as lateral change in facies in this unit. In cores 3 A and 3 B, it is percentages. Only taxa that reached >5% in at least one sample composed of pumice-rich tephra and disorganized sandy gravel were included in the analyses. The software Psimpoll 4.10 (Bennett, material that preserve traces of fallout structures (Facies F6, Fig. 2). 2002) was used to define the main diatom assemblage zones This facies shows plastic deformation features, attributed to a (DAZs) by means of stratigraphically constrained cluster analysis slump process that occurred in a pre-existing coarse pyroclastic based on squared Euclidean distance (CONISS (Grimm, 1987). Zo- deposit. These coarse deposits are barren of lacustrine fossils, nations with variances exceeding values generated by a broken- except for some terrestrial plant remains that must have been stick model of variance distribution were considered statistically incorporated into the mass of coarse material during their transport significant (Bennett, 1996). CANOCO 4.5 software (ter Braak and from slumping. Toward the western cores, these coarse deposits  Smilauer, 1998) was used to perform two ordination techniques, form 10-20 cm-thick distal laminated to massive gray muddy de- detrended correspondence analysis (DCA) and principal compo- posits, rich in lacustrine organism remains, deposited by decanta- nent analysis (PCA). The former was used to measure the length of tion of a cloud of fine suspended particles segregated from the  the primary environmental gradient (Smilauer and Leps, 2014), and coarse-grain mass slump episode (Facies F7, Fig. 2). A whitish the latter was used to determine the environmental controlling 1 cm-thick horizon (layer W) covers all deposits in this unit (Fig. 2). factors in the composition of the diatom assemblages. PCA was also The age of Unit 4 is constrained between the radiocarbon dates applied to organic pigments on a compositional data matrix, from plant macrorests found at its top and bottom in core SA11- including a selection of eight marker pigments following taxa af- 03 B corresponding to 1868 CE and 1893 CE, respectively. finity and preservation criteria. Both PCA and chronological clus- Unit 5. This unit is recovered in all cores, has a maximum tering were used to identify potential threshold responses in the thickness of 85 cm and is composed of light brown massive mud data series (Andersen et al., 2009). (Facies F8, Fig. 2), alternating with intervals of light brown mud with interbedded yellow bands (Facies F9, Fig. 2). These deposits, 4. Results rich in lacustrine organism remains, accumulated in deep offshore conditions similar to the present ones. 4.1. Sedimentary facies and units 4.2. Ageedepth model of the composite core From the stratigraphic correlation of six cores retrieved from Lake Santiago, we recognized five sedimentary units numbered To construct the ageedepth model for the studied composite chronologically from base to top (Fig. 2): core, we used a) the lithostratigraphic correlation of the six gravity Unit 1. This unit corresponds to volcanic deposits that were cores using specific tie-points, b) the 210Pb profile in core SA11-2 A retrieved from the lower 6 cm of core 1 B. It is composed of light (Supplementary Fig. 1), c) two AMS 14C dates on terrestrial plant gray massive tephra (fine ash) fallout deposits, has a muddy grain macrorests in core SA11-3 B, and d) the known age of the P17 size, and is barren of lacustrine fauna and flora (Facies F1, Fig. 2). volcanic eruption in core SA11-1 B (Fig. 2), which is the latest large The ash is intercalated with a gravel, reverse graded, pumiceous eruption in the Sete Cidades caldera (Cole et al., 2008). The first layer that corresponds to fine lapilli deposits (Facies F2, Fig. 2). appearances of pollen grains from the introduced Cryptomeria Correlation of the cores shows an onlap geometry of Unit 2 deposits japonica and Pinus spp. in core SA11-2 B were used to validate the covering this volcanic unit, indicating an undulating top geometry ages derived from the resulting age model (Supplementary (Fig. 2). The stratigraphic position and characteristics of these vol- Table 3). The lack of erosive surfaces and onlap geometries in canic deposits (Facies F1) suggest a correspondence with the upper cores SA11-2 A and SA11-2 B used to construct the composite core part of the eruptive episode P17 in the Sete Cidades caldera ensures its sedimentary and temporal continuity. 5 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Based on the CF:CS model of the 210Pb profile, the sedimentation zquez-Loureiro Lake Azul, also dated to 1820 CE (Rull et al., 2017; Va rate for the first 60 yr is 0.42 ± 0.02 cm yr1 (Supplementary Fig. 1). et al., 2019). The chronology of this interval covers the uppermost 25 cm. Ages for the top of the composite core (Unit 5) were then estimated by 4.3. Diatom assemblages linear interpolation from the topmost sample (dated to 2011 CE, the year of the field sampling campaign) and the base of the 210Pb The diatom record comprises 81 taxa and is dominated by profile (dated to 1950 CE), as well as from this level to the age at the species of the genus Aulacoseira, which constitute quasi- top of Unit 4 (Fig. 3). The nature of the coarse-grained Units 3 and 4 monospecific assemblages during several intervals (maximum matches those of the 1870 CE flood event deposits recorded in the percent abundance up to 89%) (Fig. 4). These dominant taxa adjacent Lake Azul (Va zquez-Loureiro et al., 2019), suggesting that become secondary in the uppermost 8 cm of the composite core, the same high-energy event generated both deposits and, thus, where the assemblages are dominated by Asterionella formosa and, they are of equal age. The sediments of Unit 2 were also dated by to a lesser extent, Fragilaria crotonensis and Fragilaria cf. tenera linear interpolation between the given age of deposition of Unit 4 (Fig. 4). The results of the broken-stick model of the distribution of (1870 CE) and the known age of the P17 tephra deposit in core variance define six DAZs (SAN-01 to SAN-06) based on CONISS SA11-01 B (Unit 1). This deposit corresponds to the most significant (Supplementary Table 4). The longest gradient length of taxa scores and recent eruption (P17, dated to ca. 1290 CE) that deposited resulting from the DCA is 1.94 SD units, indicating that ordination tephra in the lakes of the Sete Cidades complex (Queiroz et al., methods based on linear response models, such as PCA, are more 2008; Shotton and Williams, 1971). The orange organic layer cor- appropriate for further interpreting the underlying environmental responding to Facies F3d was used as a tie-point to correlate cores variables explaining the structure of the diatom assemblages SA11-01 A, SA11-01 B, SA11-02 B and SA11-03 A (Fig. 2). An addi-  (Smilauer and Leps, 2014). The first two axes of the PCA explain tional tie-point, reflected by the sharp change between the brown 65.1% of the total variance (Fig. 5A). Aulacoseira ambigua shows the lumpy and yellowish organic muds (Facies F3a and F3b) and the highest score for the first main axis of variation (PC1d, 42.9% of the olive green and dark brown organic muds intercalating the orange variance). Other Aulacoseira spp. and Fragilaria gracilis show high organic layer (Facies F3c, e and d, respectively), also allowed cor- scores, whereas the periphytic Achnanthidium minutissimum and relation among cores SA11-01 A, SA11-01 B, and SA11-03 A. As the Pseudostaurosira brevistriata show minimum scores. The second lowermost 20 cm of core SA11-02 B (Unit 2) could not be correlated axis of variation (PC2d, 22.2% of the variance) shows the highest with the rest of the cores, ages were extrapolated by extending the scores for F. crotonensis, A. formosa and F. cf. tenera. Downcore sedimentation rate to this short stretch of the core (Fig. 3). variations in the first two principal components are plotted in the The first recorded appearances of pollen from Cryptomeria diatom percent abundance diagram (Fig. 4). japonica and Pinus spp. (Fig. 2) almost exactly fit with a) the known introduction of Cryptomeria japonica to S~ ao Miguel Island dated to 4.4. Marker pigments of algae and cyanobacteria 1850-55 CE (De Albergaria, 2000; Supico, 1995) and b) the first noticeable plantations of Pinus on the island dated to 1820e1830 A total of 20 pigments were identified, including chlorophylls, CE, after the unsuccessful introduction of the taxon in 1750 CE chlorophyll degradation products and marker carotenoids. We (Supico, 1995). This finding is also consistent with the first selected pigments to be included in the paleoenvironmental appearance of this pollen in the sedimentary record of the adjacent reconstruction following two main criteria: pigment specificity to a taxonomic group or a process (e.g., grazing, diagenesis) and low lability, allowing the signal to persist in the sedimentary record (Supplementary Table 5). We explored the importance of in-lake planktonic and benthic primary production compared with other carbon sources derived from autochthonous and allochthonous matter and detritus, with the concentration of b-carotene (mmol b-carotene g1 TOC) (Fig. 6). The value along the sediment record can be used as a surrogate for the biomass of in-lake primary producers. The value is slightly higher in some levels of Unit 2, indicating a higher pigmentation in the organic matter deposited. There is a decrease at the top of Unit 2 with a minimum in Unit 4. The organic matter pigmentation in- creases again at the bottom of Unit 5, with minor changes recorded along this unit. All these results suggest a slightly higher biomass of in-lake primary producers during the deposition of Unit 2. The dominant pigment throughout the entire record is diatox- anthin, representing diatoms (Bacillariophyta) (Fig. 6). Low values of up to a maximum of 6.4 mmol g1 TOC are recorded in Unit 2, whereas the sediments of Unit 5 show a net increasing trend to- ward a maximum value of 17.4 mmol g1 TOC, followed by a decrease at the top of the sequence. The marker pigments of cyanobacteria, especially zeaxanthin, aphanizophyll (indicative of N2-fixing cyanobacteria) and echine- none and canthaxanthin, are mainly present in Unit 2. Cyanobac- teria are also represented in Unit 5, mostly by echinenone at the Fig. 3. Ageedepth model based on the 210Pb activityedepth profile and the AMS 14C bottom of DAZ SAN-05. dates. The expected age was calculated using linear interpolation and compared with documentary accounts of the introduction of the nonnative Cryptomeria japonica and Lutein, representing chlorophytes, is abundant in both Unit 2 expansion of Pinus spp. (De Albergaria, 2000; Supico, 1995). Corresponding sedimen- and the bottom half of Unit 5, sharply declining thereafter. The tary units are indicated. chlorophyte pattern revealed by lutein is confirmed by the 6 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Fig. 4. Diatom percentage abundance diagram for selected taxa (5% abundance in at least one sample) in Lake Santiago. The composite column on the left is based on the correlation of offshore cores SA11-02 A and SA11-02 B (see text). Diatoms are grouped according to their habitat preferences. The diatom assemblage zones (DAZs) generated by a broken-stick model of the distribution of variance (Bennett, 1996) are represented by discontinuous lines. The main lithological units and sedimentary facies are also shown. All data are plotted against age (cal yr CE). pheophytin-b curve, a degradation product of chlorophyll-b, a oscillations in the range of 2.51e10.16% for TOC and 0.29e0.87% for characteristic pigment of chlorophytes (Supplementary Fig. 2). The TN. In Unit 5, both TOC and TN show a net increase and stabiliza- topmost sample of the record shows a remarkable peak in allox- tion, with values ranging from 4.05 to 10.54% for TOC and anthin, representative of cryptophytes, and a peak in diadinox- 0.53e1.03% for TN (Fig. 7). anthin, representative of dinoflagellates. The TOC/TNcorr profile shows values between 10.58 and 17.25 for A PCA of pigment composition data reveals two major axes of Unit 2, with an average value of 13.78. The values in Unit 5 oscillate variation, which explain 76.6% of the variance (Fig. 5B). The first between 10.98 and 14.57, averaging 12.28. axis (PC1p; 62.0% of the variance) shows a strong positive relation The d13Corg curve exhibits values ranging from 25.92 with a diatom marker pigment (diatoxanthin) and a negative cor- to 20.18‰ in Unit 2 (Fig. 7). Unit 5 shows a peak value of 19.24‰, relation with the pigments of the other groups. The second axis exhibiting a progressive net decrease thereafter until a minimum (PC2p; 14.5% of the variance) is negatively related to the marker of 32.04‰ is reached at the top of the sequence. The d15Norg curve pigments of dinoflagellates (diadinoxanthin), mainly characteristic shows minimum and maximum values throughout the sequence of cyanobacteria (zeaxanthin, aphanizophyll, canthaxanthin, and of 2.63 and 2.55‰, respectively. The deposits overlying the echinenone), and that of cryptophytes (alloxanthin), whereas it is tephra, corresponding to Unit 5, show a net increasing trend, with a positively correlated with the marker pigment of chlorophytes marked positive excursion occurring in the first 8 cm (Fig. 7). (lutein). 4.6. Summary of proxy data from Lake Azul 4.5. Organic matter geochemistry The main proxy data trends from Lake Azul are shown in Fig. 8, The record shows TOC and TN percent values ranging from 2.51 zquez-Loureiro and the full results are explained in detail by Va to 10.54% and 0.29e1.03%, respectively, except for the coarse, vol- et al. (2019). The record, extending back to the P17 eruptive canic and reworked materials of Unit 4, which show values very episode of 1290 CE, shows a net increasing trend in TOC towards close to 0% for both parameters (Fig. 7). Unit 2 exhibits high the top of the record, with values ranging from 0.18 to 3.53, 7 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Fig. 5. Principal component analysis (PCA) ordination biplots of samples (numbers) and diatom taxa and pigments in Lake Santiago. Symbols indicate diatom assemblage zones (DAZs) defined by the stratigraphically constrained cluster analysis (see text). A) PCA of the diatom data matrix: Achmin ¼ Achnanthidium minutissimum, Astfor ¼ Asterionella formosa, Aulamb ¼ Aulacoseira ambigua, Aulgra ¼ Aulacoseira granulata, Aulspp ¼ Aulacoseira spp., Cycmen ¼ Cyclotella meneghiniana, Dispse ¼ Discostella pseudostelligera, Diswol ¼ Discostella woltereckii, Fracro ¼ Fragilaria crotonensis, Fragra ¼ Fragilaria gracilis, Fraten ¼ Fragilaria tenera, Nitgra ¼ Nitzschia gracilis, Psebre ¼ Pseudostaurosira brevistriata, Ulnacu ¼ Ulnaria acus, Ulndel ¼ Ulnaria delicatissima. B) PCA of the marker pigment composition data matrix: Alloxa ¼ alloxanthin (cryptophytes), Aphani ¼ aphanizophyll (N2- fixing cyanobacteria), Cantha ¼ canthaxanthin and Echine ¼ echinenone (mainly cyanobacteria), Diadin ¼ diadinoxanthin (mainly dinoflagellates), Diatox ¼ diatoxanthin (diatoms), Zeaxan ¼ zeaxanthin (mainly cyanobacteria) and Lutein (chlorophytes). Fig. 6. Pigment diagram (marker pigments) for Lake Santiago. The composite column on the left is based on the correlation of offshore cores SA11-02 A and SA11-02 B (see text). Discontinuous lines represent diatom assemblage zones (DAZs). The main lithological units and sedimentary facies are also shown. All data are plotted against age (cal yr CE). All marker pigments are expressed as mmol g1 TOC. A Chl-a preservation index was calculated and plotted as the Chl-a:a-phorbin ratio, and the index is scaled from 0 (low Chl-a preservation) to 1 (high Chl-a preservation). whereas a progressive reduction in the TOC/TNcorr values occurs almost continuous record of A. granulata from c. 1870 to 1980 CE, since c.1600 CE to the present. The diatom assemblages show an with a maximum value of 28.2% at c. 1940 CE. The most striking 8 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Fig. 7. Geochemical proxy data for Lake Santiago. Proxies include percent total organic carbon (%TOC), percent total nitrogen (%TN), TOC/TNcorr ratio (following Talbot (2001)), and carbon and nitrogen isotopes of organic matter (d13Corg, d15Norg). The composite column on the left is based on the correlation of offshore cores SA11-02 A and SA11-02 B (see text). Diatom assemblage zones (DAZs) are represented by discontinuous lines. The main lithological units and sedimentary facies are also shown. All data are plotted against age (cal yr CE). feature at the top of the record is the dominance or codominance of highest positive correlation for PC1d, whereas two periphytic taxa, an assemblage made up by A. formosa, F. cf. tenera, and F. Achnanthidium minutissimum and Pseudostaurosira brevistriata, crotonensis. show the highest negative scores for this axis. Additionally, the more terrigenous features of Unit 2 vs. Unit 5, as well as the recorded highest values of the TOC/TNcorr ratio in the former, 5. Discussion indicate enhanced inwash from the catchment (Meyers and Teranes, 2001) and, therefore, lateral advection of periphytic di- 5.1. Drivers of algal assemblage composition in Lake Santiago atoms from the littoral areas to this deep/inner position of the lake (e.g., Bao et al., 1999). We therefore interpret this component as a The ordination analyses of the main marker pigment composi- reflection of the relative importance of lacustrine vs. terrigenous tion (PCAp) and diatom species percent abundance data (PCAd), sedimentation. along with the geochemical data, explain the main environmental Asterionella formosa, F. crotonensis, and F. cf. tenera are positively factors controlling the long-term composition of primary producers correlated with PC2d, whereas A. granulata is negatively correlated. in Lake Santiago. The two opposing groups of diatoms show specific adaptations to The most abundant diatom species throughout most of the core thrive under different turbulence regimes. Asterionella formosa, sequence, the eutrophic Aulacoseira granulata, shows high absolute F. crotonensis, and F. tenera show buoyance capabilities, making scores for the first two components of the PCAp (PC1d and PC2d), them more competitive under more stable water column condi- indicating that both environmental gradients decisively affect the tions, whereas the heavily silicified Aulacoseira requires well-mixed abundance of this species in the record (Fig. 5A). The mesotrophic waters (Rühland et al., 2015). We therefore interpret this axis to be euplanktonic Aulacoseira ambigua and Aulacoseira spp. show the 9 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Fig. 8. Comparison of the main environmental patterns of change in Lake Santiago and Lake Azul under the climate stages defined by Hernandez et al. (2017) for the archipelago and the instrumental spring and summer (MAMJJA) maximum air temperature (MAT) record of Ponta Delgada (Cropper and Hanna, 2014). Data from Lake Azul are derived from 10 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 related to the frequency and/or duration of the lake mixing period regime during springs vs. intensified stratification during the late and thermal stratification strength. summers). The PCAp results of the pigment compositional data reveal that The peak abundances of the eutrophic Cyclotella meneghiniana the major axis of variation (PC1p) is positively related to diatoms in DAZ SAN-02, at the expense of A. granulata, can be related to and negatively related to chlorophytes, cyanobacteria, and crypto- significant changes in the element stoichiometry of the water. phytes (Fig. 5B). Following resource competition theory, chlor- C. meneghiniana is related to catchment disturbances (Roberts et al., ophytes and cyanobacteria are better competitors for N and light 2016), and any excess delivery of P from the basin might signifi- than diatoms and therefore dominate at low N:P and light:P ratios cantly reduce Si availability due to elevated diatom production. (Tilman et al., 1986). The allocation at different ends along this axis Under these eutrophic circumstances, C. meneghiniana would of algal groups showing distinct element uptake ratios suggests benefit from lower Si:P ratios over the highly silicified Aulacoseira that this axis is related to changes in the N:P stoichiometry of the (Rühland et al., 2015; Tilman et al., 1982). water column. The second major axis of pigment compositional The geochemical, pigment, and diatom data describe a scenario variation (PC2p) is positively related to chlorophytes and negatively of high nutrient availability during this phase. The potential related to dinoflagellates, cyanobacteria, and cryptophytes. Di- intensification of erosion since Portuguese colonization deforested atoms, characteristic of the spring early stage of phytoplankton the landscape (Raposeiro et al., 2021), besides the dominant wet succession in lakes, barely contribute to this axis (Fig. 5B), sug- and warm conditions on the island at the time (Hern andez et al., gesting that this component accounts for only the late-spring to 2017, Fig. 8), explain the export of characteristically large summer lake conditions, when silica depletion and thermal strat- amounts of P from watersheds of volcanic nature (Wetzel, 2001). ification favor the algal groups located along this axis (Margalef, 1983). The algal group positively correlated with PC2p is the Phase II (mesotrophic lake, DAZ SAN-03; ca. 1870e1900 CE) chlorophytes, a group with a high contribution of nonmotile taxa, which could outcompete cyanobacteria under a relatively more This phase shows a sharp change in diatom assemblages, turbulent regime due to their comparatively faster growth rates (Yu dominated by the euplanktonic A. ambigua, suggesting a change to et al., 2018). These data suggest a strong link between PC2p and mesotrophic conditions with higher light:P and N:P ratios (e.g., summer stratification strength, and PC2p follows the same overall Hassan, 2013; Kilham et al., 1986). The increase in the N:P ratio is downcore pattern exhibited by the turbulence-related PC2d (Figs. 4 also indicated by PC1p. This change is accompanied by a strong and 6). decrease in both the abundance of periphytic diatoms and the TOC/ TNcorr ratio (range 7.97e14.57, average 12.68), indicating a reduc- tion in allochthonous inputs to the lake. The lack of aphanizophyll, a 5.2. Ecological history of Lake Santiago and environmental forcing marker pigment of N2-fixing cyanobacteria, also supports the interpretation of a higher N:P ratio. On the other hand, high values Our multiproxy approach allows us to reconstruct the paleo- of PC2p indicate enhanced summer stratification, and this is ecological evolution of the lake as a sequence of five coherent compatible with the higher instrumental spring and summer phases (Fig. 8). (MAMJJA) maximum air temperature (MAT) record of Ponta Del- gada at the time (Cropper and Hanna, 2014). Phase I (eutrophic lake, DAZs SAN-01 and SAN-02; ca. 1580e1870 CE) Phase III (eutrophic lake, DAZ SAN-04; ca. 1900e1930 CE) This phase is characterized by the highest values of TOC in the This phase indicates a return to eutrophic conditions similar to entire record, a higher biomass of in-lake primary producers (b- Phase I, with TOC and in-lake primary production (b-carotene) carotene; Fig. 6), and dominance of the euplanktonic and eutrophic showing quite similar recovery trends. The abundance of lutein also A. granulata. The TOC/TNcorr profile shows values between 10.58 shows that chlorophytes replaced cyanobacteria during this phase. and 17.25 for Unit 2 (average 13.78), indicating that the origin of the A. granulata reaches very high values, which, together with deposited organic matter, although lacustrine, was influenced by pigment data, suggests enhanced P availability under lower irra- C3 land plants from the catchment (Meyers and Teranes, 2001). diance conditions (Kilham et al., 1986). However, the highly This is also supported by the high abundances of transported reduced numbers of periphytic diatoms and a forest-covered periphytic diatoms (Fig. 4). These data point to a lake largely catchment at this time (Raposeiro et al., 2021; Rull et al., 2017) influenced by the allochthonous delivery of nutrients, which suggest that other causative factors of eutrophication, apart from probably drove a large part of its high primary production at the external nutrient inputs, should be considered. Weakened stratifi- time. cation during the summer, as indicated by maximum values of The low values of PC1p suggest dominantly low N:P ratios fa- PC2p, is a major candidate, since it enhances internal P turnover, voring cyanobacteria and chlorophytes (Tilman et al., 1986). How- explaining the net increase in in-lake primary production (Wetzel, ever, despite the minimum concentrations of diatoxanthin 2001). Moreover, meteorological data show a prolonged reduction recorded throughout the core, it constitutes the most abundant in the spring and summer MATs during this phase (Fig. 8), which pigment in the sediments (Fig. 6). The diatom assemblages are would, not only extend the mixing period, but also facilitate deeper dominated by A. granulata, which, compared to other diatom taxa, mixing. Additionally, this phase also occurs after the stocking of have high P requirements (Kilham et al., 1986). The cooccurrence of cyprinid fish on the island, which could have altered top-down and the heavily silicified A. granulata, which requires a turbulent water bottom-up controls on primary production and lake food web dy- regime to avoid sinking from the euphotic zone (Margalef, 1978a,b) namics (Buchaca et al., 2011; Raposeiro et al., 2017; Skov et al., and, by contrast, algal groups favored by thermal stratification, such 2010; Va zquez-Loureiro et al., 2019). as cyanobacteria, is also indicative of the existence of a eutrophic lake with marked seasonality (a very contrasting strong turbulent zquez-Loureiro et al. (2019). The percent abundances of Aulacoseira granulata and Cyclotella meneghiniana for Lake Santiago were recalculated once the allochthonous benthic Va diatom component was removed from the data matrix. For explanations on the meaning of the different proxies, see the text. 11 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Phase IV (mesotrophic lake, DAZ SAN-05; ca. 1930e1990 CE) impermeable barrier which can inhibit nutrient return, mainly P, from the sediments to the water column (Barker et al., 2000). In the The TOC/TNcorr ratio values are maintained during this phase, case of thick tephra deposits, P-limitation due to massive tephra indicating the prevalence of lacustrine-dominated organic matter influx can have an impact on the long-term lake ontogeny, lasting sedimentation. Pigment data show that primary production is up to millennia (Telford et al., 2004). Besides, the 16th century dominated by diatoms, with a lower contribution of cyanobacteria, chronicles of Gaspar Frutuoso also show that more than two cen- cryptophytes, and dinoflagellates. Lutein abruptly disappears at the turies after the P17 eruption, at least the area of the catchment beginning of this phase and occasionally reappears at very low adjacent to the lake was still made up of “sterile sands” (Frutuoso, concentrations, indicating that chlorophytes barely contributed to 1977). Contrary to expectations, when this condition coincides with lake primary production. In addition, a new replacement of periods of enhanced runoff, the post-eruptive recovery of a volcanic A. granulata by A. ambigua is recorded, suggesting a return to lake can dramatically dilute any chemically enriched waters (Larson mesotrophic conditions similar to those of Phase II. Enhanced et al., 2006). Additionally, tephra deposition has the immediate thermally-induced stratification is indicated by the high values of effect of depauperating biological communities, restarting ecolog- both the PC2p and the maximum spring and summer MATs at Ponta ical succession. The periphytic algae characteristic of pioneering Delgada (Fig. 8). communities in oligotrophic lakes outcompete phytoplankton, because they can incorporate nutrients from both the water column Phase V (probable lake regime shift, DAZ SAN-06; ca. 1990e2011 and the sediments (Hansson, 1992). This mechanism could there- CE) fore also hinder the release of sufficient nutrients to support phytoplankton production (Va zquez-Loureiro et al., 2019). All these This phase is characterized by a sharp shift in diatom species synergetic effects forced a long-term state of oligotrophy for ca. 650 assemblages, with Aulacoseira being replaced mainly by A. formosa yr (1290e1940 CE), only moderately modified by fish stocking and and, secondarily, by F. crotonensis and F. cf. tenera, resulting in forestation practices and eventually interrupted by the extensive maximum values for PC2d. The change, therefore, points to use of fertilizers. In contrast, the Lake Santiago sedimentary record enhanced stability of the water column, compatible with the in- shows that eutrophic conditions were present only 290 yr after the crease in spring and summer MATs (Fig. 8). Diadinoxanthin (mainly P17 catastrophic event. characteristic of dinoflagellates) and alloxanthin (characteristic of This intersite comparison further confirms, at least for the cryptophytes) record their maximum values throughout the record, studied proxies, the major site-specific component of the paleo- with diatoxanthin exhibiting a strong reduction in concentration. environmental signal in these lakes. It can be hypothesized that the The greater contribution of motile groups, such as dinoflagellates different effects of the regional (climatic) signal vs. site-specific and cryptophytes, at the expense of diatoms also indicates ‘noise’ (Roberts et al., 2016) in lakes Azul vs. Santiago is a factor strengthened thermal stratification (Margalef, 1983). Oxygen of both a) the closer distance of Lake Azul to the volcanic focus of depletion at the hypolimnion, promoting microbial anaerobic the P17 eruption, which resulted in greater tephra fallout in Lake biomass, would explain the observed negative excursion in d13Corg Azul than in Lake Santiago (Cole et al., 2008), and b) the large dif- values (Fig. 7) (Teranes and Bernasconi, 2005). ferences in the bathymetric and topographic gradients of both lakes and their catchments (Fig. 1), a factor that can play a critical role in 5.3. The different signatures of climate in the Sete Cidades deep altering the phototrophic response to climate variations (Gushulak lakes et al., 2021). The latter can be partitioned into three main compo- nents, so that Lake Santiago, despite having, compared to Lake Azul, Despite their proximity and, therefore, the same climatic a similar depth (28.75 vs. 25.35 m) and lake area to catchment ratio context, the ecological histories of the two deep lakes in the Cal- (0.31 vs. 0.23) (Supplemental Table 1), shows 1) much steeper deira das Sete Cidades, Lake Santiago and Lake Azul (Fig. 1), show catchment slopes, 2) a water volume less than one order of large differences in their responses to climate variability. magnitude, and 3) a simple bathymetry without planar surfaces There is correspondence between the previously defined pale- which facilitates more linear responses (Stone and Fritz, 2004; Bao oclimate stages for S~ ao Miguel Island (Hernandez et al., 2017), the et al., 2015). These factors together would not only promote a larger meteorological data extending back to 1872 CE, and the predictable nutrient inwash to Lake Santiago, but also amplify the effects of ecological responses of Lake Santiago, as explained in section 5.2 variations in catchment-mediated climate forcings in this much and shown in Fig. 8. The long-term climate changes in the island smaller deep lake compared to Lake Azul. As a result, the sediments had two different effects on the ecological states of Lake Santiago: of Lake Azul fail as a reliable archive of changes in climate condi- a) the wet and warm last stage of the LIA (Phase I) created the tions in the archipelago for the selected proxy-based records at the conditions for in-excess P delivery from the catchment, and b) the considered temporal scale. post-LIA (phases II to V) differences in the thermal structure of the water and, thus, in its mixing regime, induced by fluctuations in the 5.4. The post-1980 CE climate forcing scenario in the Sete Cidades spring and summer MATs. In this regard, thermal contact with the deep lakes atmosphere imposes a temperature signal on the lake surface (Boehrer and Schultze, 2008), and spring and summer changes in The almost synchronous and identical major post-1980 CE shift the thermal structure of the water column are responsible for the in Lake Azul and Lake Santiago's diatom assemblages suggests a different mixing regimes that alter the algal composition in lakes common underlying control at the regional scale. (Margalef, 1983). The complete implications of global warming for lake systems Contrary to Lake Santiago, Lake Azul shows no direct corre- are still largely unresolved, including the restructuring of diatom spondence with the S~ ao Miguel Island climate stages and/or communities. According to several sediment studies, global changes in the spring and summer MATs mentioned above, except warming is forcing widespread diatom assemblage shifts by for the post-1980/90 CE period (Fig. 8). The catastrophic nature of altering thermal lake properties (e.g., Hadley et al., 2019; Rühland the P17 eruption overrode the climatic signature in Lake Azul et al., 2015; Sivarajah et al., 2016; Sorvari et al., 2002). Other sediments due to the disruption of the biogeochemical cycling of studies stress that an accelerated nutrient increase in recent de- nutrients (Vazquez-Loureiro et al., 2019). Tephra deposits create an cades, such as by atmospheric N2 deposition, is another critical 12 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 driver shifting diatom assemblages (e.g., Dong et al., 2012; Hu et al., deeper mixing depths (Hanna, 1990; Gillis et al., 2021), longer 2014; Wolfe et al., 2013). cooling periods (Nürnberg, 1988), and a more poorly developed The simultaneous replacement of Aulacoseira spp. As the thermocline (Oswald and Rouse, 2004) than smaller lakes. These dominant taxon in Lake Santiago and Lake Azul involves the characteristics make larger lakes require extra warming related to expansion of A. formosa, F. crotonensis and F. cf. tenera (Fig. 4 and their larger water volumes (Gillis et al., 2021), so a threshold Supplementary Fig. 3). Enhanced atmospheric N2 deposition could response of biotic communities is more likely in Lake Santiago be responsible for the widespread increases in the relative abun- where, with approximately. One order of magnitude less volume dance of A. formosa and F. crotonensis in recent decades (e.g., Saros (Supplemental Table 1), heating up the epilimnion requires less et al., 2005). This nutrient enrichment is manifested in the sedi- energy. mentary records by a sustained 20th century decline in d15Norg and a %N increase (Wolfe et al., 2006, 2013). However, there is no evi- 6. Conclusions dence of a clear net negative trend in d15Norg, as there is a lack of %N enrichment, in Lake Santiago (Fig. 7 and Supplementary Fig. 3). The Despite the proximity of the studied lakes, our results demon- post-1985 CE record of Lake Santiago shows negligible increases in strate pronounced differences in the sensitivity of the two deep TOC and TN in the uppermost sediments. The same is true for b- temperate volcanic lakes under equal climate forcing. From ca. 1580 carotene, an indicator of total algal biomass (Buchaca et al., 2019; CE and until ca. 1980/90 CE, the long-term ecological trajectory of Leavitt and Hodgson, 2001, Fig. 6). Nutrient enrichment associated the small Lake Santiago was mainly determined by climate vari- with land clearance as a likely cause for the A. formosa increase (e.g., ability. In contrast, other natural (volcanism) and nonnatural Boeff et al., 2016) can therefore also be disregarded. Additionally, (reforestation, invasive species introduction and artificial fertiliza- the high light requirements of F. crotonensis (Interlandi et al., 1999) tion) environmental drivers played a much larger role in the evo- are not compatible with a scenario involving turbid waters pro- lution of the large Lake Azul. The reconstructed ecological history of duced by eutrophication. Lake Santiago, based on a multiproxy approach, reveals that since As a result, changes in nutrient concentrations are unlikely ca. 1580 CE, the lake has experienced ecological shifts in phase with factors driving diatom community changes in Lake Santiago in the major climate stages known for Sa ~o Miguel Island. recent decades. Alternatively, a shift in thermal properties of the During the last warm and wet stage of the LIA (1580e1870 CE), water column could be the causative factor for the replacement of high P availability linked to runoff resulted in eutrophic conditions, the Aulacoseira-vs. Asterionella- and Fragilaria-dominated fossil characterized by the dominance of the diatom A. granulata, negli- assemblages. Thick-walled cells, such as those of Aulacoseira spp., gible presence of mesotrophic A. ambigua, and peak abundances of require well-mixed water column conditions to avoid sinking out cyanobacteria. The shifting dominance of A. granulata vs. meso- from the euphotic zone (Margalef, 1978a,b). The worldwide trophic A. ambigua characterizes the post-LIA ecological history of reduction of their abundances in recent sediments has been the lake, and it was related to the maximum spring and summer considered indicative of increased water column stability associ- atmospheric temperature variability, which provoked changes in ated with warmer climate conditions (Rühland et al., 2008, 2015) In the mixing regime. Higher temperatures induced a more stable contrast, the elongated shape and/or colonial capabilities of species water column during the 1870e1900 CE and 1930e1960 CE pe- such as A. formosa, F. crotonensis and F. tenera are key adaptations riods, with the dominance of A. ambigua and a higher relative that decelerate their sinking velocities, allowing these species to contribution of cyanobacteria over chlorophytes in the plankton. thrive in prevailing stratification regimes (Reynolds, 2006). Such Conversely, reduced temperatures in the 1900e1930 CE period changes in the thermal lake properties would also explain the peak favored deeper mixing, which benefited A. granulata over abundances of motile phytoplankton, such as dinoflagellates (dia- A. ambigua, as well as chlorophytes. dinoxanthin) and cryptophytes (alloxanthin). Moreover, the recent An intersite comparison with Lake Azul, located just 2.5 km net negative trend in d13Corg values could also be likely caused by away inside the caldera, showed major differences in the ecological the proliferation of anaerobic bacteria occurring under the anoxia response preserved in the sedimentary record under the same induced by prolonged and/or intensified thermal water stratifica- climate scenario until ca. 1980/90 CE. A disruption of biogeo- tion conditions. chemical cycles after the massive deposition of tephra induced by a The shift in diatom community structure in Lake Santiago in catastrophic volcanic eruption in ca. 1290 CE masked the climate recent decades resembles that of mesotrophic lakes where warmer signature in the sedimentary record. climates favor those species making the best trade-off between Despite this multistressor scenario, the post-1980/90 CE period decreasing nutrients and increased water column stability (Berthon showed a probable ecological regime shift in both lakes e also et al., 2014). Both A. formosa and F. crotonensis are favored by shown by the current limnological monitoring programs of other relatively high N:P and Si:P ratios (Interlandi et al., 1999; Saros lakes on Sa ~o Miguel Island e triggered by higher water tempera- et al., 2005), and F. tenera by low P (Interlandi et al., 1999) and tures under global warming conditions. high temperatures (Kienel et al., 2005), so the three taxa would This demonstration of different long-term responses to climate fulfill those requirements. The timing of this turnover in diatom variability in these close-proximity lakes not only implies that they species composition is consistent with the records of Lake Azul register dissimilar information for paleoclimate research but also (Vazquez-Loureiro et al., 2019) and Lake Empadadas (Marques, has management implications for the preservation of ecosystem 2021). It is also compatible with the available environmental services. The magnitude and pattern of response in the post-1980/ monitoring data of other lakes in the archipelago, such as Lake 90 CE period involve a disruptive and large replacement of Furnas and Lake Fogo, where a shift to Asterionella- and Fragilaria- turbulence-loving diatoms by taxa favoring warm stratified waters dominated assemblages has been observed (Gonçalves, 2008), in Lake Santiago, and a similar change occurred in Lake Azul, despite the large local-scale differences among these systems. although more gradual and less pronounced. This suggests that, However, this local component still plays its own significant role. despite their similarities and being forced by the same climate There are notable differences in the type of response, gradual in conditions, the two lakes require the adoption of different man- Lake Azul vs. abrupt in Lake Santiago (Supplementary Fig. 3), agement strategies under the current climate change scenario, with related to differences in the size and morphometry of each lake. Lake Santiago being more vulnerable to the effects of global Large lakes, such as Lake Azul, compared to Lake Santiago, have warming and Lake Azul being more resilient. 13 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Author contributions Benavente, M., Raposeiro, P.M., Giralt, S., 2021. Unraveling the Holocene erup- tive history of Flores Island (Azores) through the analysis of lacustrine sedi- mentary records. Front. Earth Sci. 9, 889. https://doi.org/10.3389/ All authors contributed to the manuscript and approved sub- feart.2021.738178. mission. DV-L and RB conducted the research, analyzed the diatom Appleby, P.G., 2001. Chronostratigraphic techniques in recent sediments. In: samples, processed and interpreted the data, made the tables and Last, W.M., Smol, J.P. (Eds.), Tracking Environmental Change Using Lake Sedi- ments: Volume 1: Basin Analysis, Coring, and Chronological Techniques. Kluwer figures and wrote the manuscript; AS described the lithofacies, Academic Publishers, Dordrecht, Netherlands, pp. 171e203. performed the stratigraphic correlations and interpreted data; VG Bao, R., Herna ndez, A., Saez, A., Giralt, S., Prego, R., Pueyo, J.J., Moreno, A., Valero- Garce s, B.L., 2015. Climatic and lacustrine morphometric controls of diatom reviewed the diatom taxonomy and processed and interpreted the paleoproductivity in a tropical Andean lake. Quat. Sci. Rev. 129, 96e110. https:// data; TB analyzed the pigment data and processed and interpreted doi.org/10.1016/j.quascirev.2015.09.019. the data; AH processed and interpreted the (paleo)climate data; PR Bao, R., Sa ez, A., Servant-Vildary, S., Cabrera, L., 1999. Lake-level and salinity processed and interpreted the data; EdB contributed pollen data for reconstruction from diatom analyses in Quillagua formation (late Neogene, Central Andean forearc, Northern Chile). Palaeogeogr. Palaeoclimatol. Palae- the ageedepth model; and PM contributed 210Pb and 137Cs data for oecol. 153, 309e335. https://doi.org/10.1016/S0031-0182(99)00066-8. the ageedepth model. VG, AS, AH, SG, PR and RB cored the lake. All Barker, P., Telford, R., Merdaci, O., Williamson, D., Taieb, M., Vincens, A., Gibert, E., authors revised the manuscript. AS, SG and RB led the research. 2000. The sensitivity of a Tanzanian crater lake to catastrophic tephra input and four millennia of climate change. The Holocene 10, 303e310. https://doi.org/10. 1191/095968300672848582. Funding Bennett, K.D., 1996. Determination of the number of zones in a biostratigraphical sequence. New Phytol. 132, 155e170. https://doi.org/10.1111/j.1469- 8137.1996.tb04521.x. This research was funded by the Spanish Ministry of Economy Bennett, K.D., 2002. Documentation for Psimpoll 4.10 and Pscomb 1.03. C Programs and Competitiveness projects PaleoNAO, RapidNAO and PaleoM- for Plotting Pollen Diagrams and Analysing Pollen Data. Uppsala University, odes (CGL 2010-15767, CGL 2013-40608-R and CGL 2016-75281-C2, Uppsala, Sweden. Berthon, V., Alric, B., Rimet, F., Perga, M.-E., 2014. Sensitivity and responses of di- respectively) and by the Fundaça ~o para a Cie ^ncia e Tecnologia atoms to climate warming in lakes heavily influenced by humans. Freshw. Biol. (PTDC/CTA-AMB/28,511/2017). David V azquez-Loureiro, Armand 59, 1755e1767. https://doi.org/10.1111/fwb.12380. Herna ndez, and Pedro Raposeiro benefited with grants from the Boeff, K.A., Strock, K.E., Saros, J.E., 2016. Evaluating planktonic diatom response to climate change across three lakes with differing morphometry. J. Paleolimnol. Xunta de Galicia (I2C Programme co-funded with the European 56, 33e47. https://doi.org/10.1007/s10933-016-9889-z. Social Fund), the Spanish Ministry of Science and Innovation Boehrer, B., Schultze, M., 2008. Stratification of lakes. Rev. Geophys. 46, RG2005. through the Ramo n y Cajal Scheme [RYC 2020-029253-I], and https://doi.org/10.1029/2006RG000210. ~o para a Cie ^ncia e Tecnologia (DL57/2016/ICETA/EEC 2018/ Buchaca, T., Catalan, J., 2007. Factors influencing the variability of pigments in the Fundaça surface sediments of mountain lakes. Freshw. Biol. 52, 1365e1379. https:// 25), respectively. Funding for open access charge: Universidade da doi.org/10.1111/j.1365-2427.2007.01774.x. Corun~ a/CISUG. Buchaca, T., Catalan, J., 2008. On the contribution of phytoplankton and benthic biofilms to the sediment record of marker pigments in high mountain lakes. J. Paleolimnol. 40, 369e383. https://doi.org/10.1007/s10933-007-9167-1. Declaration of competing interest Buchaca, T., Kosten, S., Lacerot, G., Mazzeo, N., Kruk, C., Huszar, V.L.M., Lotter, A.F., Jeppesen, E., 2019. Pigments in surface sediments of South American shallow lakes as an integrative proxy for primary producers and their drivers. Freshw. The authors declare that they have no known competing Biol. 64, 1437e1452. https://doi.org/10.1111/fwb.13317. financial interests or personal relationships that could have Buchaca, T., Skov, T., Amsinck, S.L., Gonçalves, V., Azevedo, J.M.N., Andersen, T.J., appeared to influence the work reported in this paper. Jeppesen, E., 2011. Rapid ecological shift following piscivorous fish introduction to increasingly eutrophic and warmer lake furnas (Azores Archipelago, Portugal): a Paleoecological Approach. Ecosystems 14, 458e477. https://doi.org/ Data availability 10.1007/s10021-011-9423-0. Cole, P.D., Pacheco, J.M., Gunasekera, R., Queiroz, G., Gonçalves, P., Gaspar, J.L., 2008. Data will be made available on request. Contrasting styles of explosive eruption at Sete Cidades, S~ ao Miguel, Azores, in the last 5000 years: hazard implications from modelling. J. Volcanol. Geoth. Res. 178, 574e591. https://doi.org/10.1016/j.jvolgeores.2008.01.008. Acknowledgements Cropper, T.E., Hanna, E., 2014. An analysis of the climate of Macaronesia, 18652012. Int. J. Climatol. 34, 604e622. https://doi.org/10.1002/joc.3710. De Albergaria, I.S., 2000. Quintas, jardins e parques da ilha de Sao Miguel, (1785- We thank María Jesús Rubio, Olga Margalef and Guiomar 1885). Quetzal Editores, Lisbon, Portugal. nchez-Lo Sa pez for field assistance. The IAEA is grateful for the Dong, X., Bennion, H., Maberly, S.C., Sayer, C.D., Simpson, G.L., Battarbee, R.W., 2012. Nutrients exert a stronger control than climate on recent diatom communities support provided to its Marine Environment Laboratories by the in Esthwaite water: evidence from monitoring and palaeolimnological records. Government of the Principality of Monaco. Our thanks are also Freshw. Biol. 57, 2044e2056. https://doi.org/10.1111/j.1365-2427.2011.02670.x. extended to two anonymous reviewers for their very useful and Frutuoso, G., 1977. Livro Quarto das Saudades da Terra. Instituto Cultural de Ponta constructive comments on an earlier version of the manuscript. Delgada, Ponta Delgada, Azores, Portugal. Gillis, D.P., Minns, C.K., Shuter, B.J., 2021. Predicting open-water thermal regimes of temperate North American lakes. Can. J. Fish. Aquat. Sci. 78, 820e840. https:// Appendix A. Supplementary data doi.org/10.1139/cjfas-2020-0140. Gonçalves, V., 2008. Contribuçao para o Estudo da Qualidade Ecolo  gica das Lagoas dos Açores. Fitopl^ ancton e Diatoma ceas Bento nicas. PhD. Dissertation. Uni- Supplementary data to this article can be found online at versidade dos Açores (Ponta Delgada, Portugal). https://doi.org/10.1016/j.quascirev.2023.107968. Goncalves, V., Marques, H., Fonseca, A., 2010. Lista das diatom aceas (Bacillar- iophyta). In: Borges, P.A.V., Bried, J., Costa, A., Cunha, R., Gabriel, R., Goncalves, V., Martins, A.F., Melo, I., Parente, M., Raposeiro, P., Rodrigues, P., References Santos, R.S., Silva, L., Vieira, P., Vieira, V. (Eds.), Listagem Dos Organismos Ter- restres e Marinhos Dos Acores. Princípia, Cascais, Portugal, pp. 81e97. Andersen, T., Carstensen, J., Herna ndez-García, E., Duarte, C.M., 2009. Ecological Grimm, E.C., 1987. CONISS: a FORTRAN 77 program for stratigraphically constrained thresholds and regime shifts: approaches to identification. Trends Ecol. Evol. 24, cluster analysis by the method of incremental sum of squares. Comput. Geosci. 49e57. https://doi.org/10.1016/j.tree.2008.07.014. 13, 13e35. https://doi.org/10.1016/0098-3004(87)90022-7. Andrade, C., Cruz, J.V., Viveiros, F., Coutinho, R., 2019. CO2 flux from volcanic lakes in Gushulak, C.A.C., Leavitt, P.R., Cumming, B.F., 2021. Basin-specific records of lake the western group of the Azores Archipelago (Portugal). Water 11, 599. https:// oligotrophication during the middle-to-late Holocene in boreal northeast doi.org/10.3390/w11030599. Ontario, Canada. Holocene 31, 1539e1554. https://doi.org/10.1177/ Andrade, C., Trigo, R.M., Freitas, M.C., Gallego, M.C., Borges, P., Ramos, A.M., 2008. 09596836211025972. Comparing historic records of storm frequency and the North Atlantic Oscilla- Hadley, K.R., Paterson, A.M., Rühland, K.M., White, H., Wolfe, B.B., Keller, W., tion (NAO) chronology for the Azores region. Holocene 18, 745e754. https:// Smol, J.P., 2019. Biological and geochemical changes in shallow lakes of the doi.org/10.1177/0959683608091794. Hudson Bay Lowlands: a response to recent warming. J. Paleolimnol. 61, Andrade, M., Ramalho, R.S., Pimentel, A., Herna ndez, A., Kutterolf, S., S aez, A., 313e328. https://doi.org/10.1007/s10933-018-0061-9. 14 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 Hanna, M., 1990. Evaluation of models predicting mixing depth. Can. J. Fish. Aquat. In: Lange-Bertalot, H. (Ed.), Diatoms of the European Inland Waters and Com- Sci. 47, 940e947. https://doi.org/10.1139/f90-108. parable Habitats, ume 6. A. R. G. Gantner Verlag, Ruggell, Liechtenstein, p. 747. Hansson, L.-A., 1992. Factors regulating periphytic algal biomass. Limnol. Oceanogr. Lange-Bertalot, H., Ulrich, S., 2014. Contributions to the taxonomy of needle-shaped 37, 322e328. https://doi.org/10.4319/lo.1992.37.2.0322. Fragilaria and Ulnaria species. Lauterbornia 78, 1e73. Hargan, K.E., Nelligan, C., Jeziorski, A., Rühland, K.M., Paterson, A.M., Keller, W., Larson, D.W., Sweet, J., Petersen, R.R., Crisafulli, C.M., 2006. Posteruption response of Smol, J.P., 2016. Tracking the long-term responses of diatoms and cladocerans to phytoplankton and zooplankton communities in spirit lake, mount st. Helens, climate warming and human influences across lakes of the Ring of Fire in the Washington. Lake reserv. OR Manag. 22, 273e292. https://doi.org/10.1080/ Far North of Ontario, Canada. J. Paleolimnol. 56, 153e172. https://doi.org/ 07438140609354362. 10.1007/s10933-016-9901-7. Leavitt, P.R., Fritz, S.C., Anderson, N.J., Baker, P.A., Blenckner, T., Bunting, L., Hassan, G.S., 2013. Diatom-based reconstruction of middle to late holocene paleo- Catalan, J., Conley, D.J., Hobbs, W.O., Jeppesen, E., Korhola, A., McGowan, S., environments in Lake Lonkoy, southern Pampas, Argentina. Diatom Res. 28, Rühland, K., Rusak, J.A., Simpson, G.L., Solovieva, N., Werne, J., 2009. Paleo- 473e486. https://doi.org/10.1080/0269249X.2013.851118. limnological evidence of the effects on lakes of energy and mass transfer from Herna ndez, A., Bao, R., Giralt, S., Leng, M.J., Barker, P.A., S aez, A., Pueyo, J.J., climate and humans. Limnol. Oceanogr. 54, 2330e2348. https://doi.org/ Moreno, A., Valero-Garce s, B.L., Sloane, H.J., 2008. The palaeohydrological 10.4319/lo.2009.54.6_part_2.2330. evolution of Lago Chungara  (Andean Altiplano, Northern Chile) during the Leavitt, P.R., Hodgson, D., 2001. Sedimentary pigments. In: Smol, J., Birks, H., Last, M. Lateglacial and early Holocene using oxygen isotopes in diatom silica. J. Quat. (Eds.), Tracking Environmental Change Using Lake Sediments Volume 3: Sci. 23, 351e363. https://doi.org/10.1002/jqs.1173. Terrestrial, Algal, and Siliceous Indicators. Kluwer Academic Press, Dordrecht, Herna ndez, A., Kutiel, H., Trigo, R.M., Valente, M.A., Sigro, J., Cropper, T., Santo, F.E., Netherlands, pp. 295e326. 2016. New Azores Archipelago daily precipitation dataset and its links with Levkov, Z., 2009. Diatoms of the European inland waters and comparable habitats. large-scale modes of climate variability. Int. J. Climatol. 36, 4439e4454. https:// Volume 5. Amphora sensu lato. In: Lange-Bertalot, H. (Ed.), Diatoms of the Eu- doi.org/10.1002/joc.4642. ropean Inland Waters and Comparable Habitats, Volume 4. A. R. A. Gantner Herna ndez, A., Martin-Puertas, C., Moffa-S anchez, P., Moreno-Chamarro, E., Verlag, Ruggell, Liechtenstein, p. 916. Ortega, P., Blockley, S., Cobb, K.M., Comas-Bru, L., Giralt, S., Goosse, H., Levkov, Z., Metzeltin, D., Pavlov, A., 2013. Diatoms of the European inland waters Luterbacher, J., Martrat, B., Muscheler, R., Parnell, A., Pla-Rabes, S., Sjolte, J., and comparable habitats. Volume 7. Luticola and luticulopsis. In: Lange- Scaife, A.A., Swingedouw, D., Wise, E., Xu, G., 2020. Modes of climate variability: Bertalot, H. (Ed.), Diatoms of the European Inland Waters and Comparable synthesis and review of proxy-based reconstructions through the Holocene. Habitats, ume 7. A. R. A. Gantner Verlag, Ruggell, Liechtenstein, p. 698. Earth Sci. Rev. 209, 103286. https://doi.org/10.1016/j.earscirev.2020.103286. Louys, J., 2012. Paleontology in Ecology and Conservation: Springer Earth System Herna ndez, A., Sa ez, A., Bao, R., Raposeiro, P.M., Trigo, R.M., Doolittle, S., Masque , P., Sciences. Springer Science & Business Media, Berlin, Germany. Rull, V., Gonçalves, V., V azquez-Loureiro, D., Rubio-Ingle s, M.J., Sanchez- Margalef, R., 1978a. Life-forms of phytoplankton as survival alternatives in an un- Lo pez, G., Giralt, S., 2017. The influences of the AMO and NAO on the sedi- stable environment. Oceanol. Acta 1, 493e509. mentary infill in an Azores Archipelago lake since ca. 1350CE. Global Planet. Margalef, R., 1978b. Phytoplankton communities in upwelling areas. The example of Change 154, 61e74. https://doi.org/10.1016/j.gloplacha.2017.05.007. NW Africa. Oceanol. Aquat. 3, 97e132. Houk, V., Klee, R., 2007. Atlas of freshwater centric diatoms with a brief key and Margalef, R., 1983. Limnología. Ediciones Omega, Barcelona, Spain. descriptions. Part II. Melosiraceae and Aulacoseiraceae (Supplement to Part I). Marques, H.S., 2021. Reconstruction of Past Environmental Changes in the Azores Fottea 7, 85e255. Based on Lacustrine Diatom Sedimentary Records. Universidade dos Açores, Hu, Z., Anderson, N.J., Yang, X., McGowan, S., 2014. Catchment-mediated atmo- Azores, Portugal. spheric nitrogen deposition drives ecological change in two alpine lakes in SE Meyers, P.A., Teranes, J.L., 2001. Sediment organic matter. In: Smol, J.P., Birks, H.J.B., Tibet. Global Change Biol. 20, 1614e1628. https://doi.org/10.1111/gcb.12435. Last, W.M. (Eds.), Tracking Environmental Change Using Lake Sediments. Vol- Interlandi, S.J., Kilham, S.S., Theriot, E.C., 1999. Responses of phytoplankton to varied ume 2: Physical and Geochemical Methods. Kluwer Academic Publishers, resource availability in large lakes of the Greater Yellowstone Ecosystem. Dordrecht, Netherlands, pp. 239e269. Limnol. Oceanogr. 44, 668e682. https://doi.org/10.4319/lo.1999.44.3.0668. Mills, K., Schillereff, D., Saulnier-Talbot, E., Gell, P., Anderson, N.J., Arnaud, F., Jackson, S.T., Hobbs, R.J., 2009. Ecological restoration in the light of ecological his- Dong, X., Jones, M., McGowan, S., Massaferro, J., Moorhouse, H., Perez, L., tory. Science 325, 567e569. https://doi.org/10.1126/science.1172977. Ryves, D.B., 2017. Deciphering long-term records of natural variability and hu- Jeffers, E.S., Nogue , S., Willis, K.J., 2015. The role of palaeoecological records in man impact as recorded in lake sediments: a palaeolimnological puzzle. WIREs assessing ecosystem services. Quat. Sci. Rev. 112, 17e32. https://doi.org/10.1016/ Water 4, e1195. https://doi.org/10.1002/wat2.1195. j.quascirev.2014.12.018. Nogue , S., Santos, A.M.C., Birks, H.J.B., Bjo €rck, S., Castilla-Beltran, A., Connor, S., de Jeffrey, S.W., Mantoura, R.F.C., Wright, S.W., 1997. Phytoplankton Pigments in Boer, E.J., de Nascimento, L., Felde, V.A., Fern andez-Palacios, J.M., Froyd, C.A., Oceanography: Guidelines to Modern Methods. UNESCO Publishing, Paris, Haberle, S.G., Hooghiemstra, H., Ljung, K., Norder, S.J., Pen ~ uelas, J., Prebble, M., France. Stevenson, J., Whittaker, R.J., Willis, K.J., Wilmshurst, J.M., Steinbauer, M.J., 2021. Jovanelly, T.J., Fritz, S.C., 2011. Morphometric and chemical response of two con- The human dimension of biodiversity changes on islands. Science 372, trasting lake systems to modern climate change. J. Paleolimnol. 46, 89e98. 488e491. https://doi.org/10.1126/science.abd6706. https://doi.org/10.1007/s10933-011-9521-1. Nürnberg, G.K., 1988. Prediction of phosphorus release rates from total and Keegan, W.F., Diamond, J.M., 1987. Colonization of islands by humans: a biogeo- reductant-soluble phosporus in anoxic lake-sediments. Can. J. Fish. Aquat. Sci. graphical perspective. In: Schiffer, M.B. (Ed.), Advances in Archaeological 45, 453e462. https://doi.org/10.1139/f88-054. Method and Theory. Academic Press, San Diego, CA, pp. 49e92. Oswald, C.J., Rouse, W., 2004. Thermal characteristics and energy balance of Kienel, U., Schwab, M.J., Schettler, G., 2005. Distinguishing climatic from direct various-size Canadian shield lakes in the Mackenzie river basin. anthropogenic influences during the past 400 years in varved sediments from J. Hydrometeorol. 5, 129e144. Lake Holzmaar (Eifel, Germany). J. Paleolimnol. 33, 327e347. https://doi.org/ Potapova, M.G., Bixby, R.J., Charles, D.F., Edlund, M.B., Enache, M.E., Furey, P., 10.1007/s10933-004-6311-z. Hamilton, P.B., Lowe, R.L., Manoylov, K.M., Ognjanova-Rumenova, N., Kilham, P., Kilham, S.S., Hecky, R.E., 1986. Hypothesized resource relationships Ponader, K.C., Ren, L., Siver, P.A., Spaulding, S.A., Zalack, J., 2008. Representatives among African planktonic diatoms. Limnol. Oceanogr. 31, 1169e1181. https:// of the Genus Aulacoseira Thwaites in NAWQA, Eighteenth NAWQA Workshop on doi.org/10.4319/lo.1986.31.6.1169. Harmonization of Algal Taxonomy. The Academy of Natural Sciences of Phila- Krammer, K., Lange-Bertalot, H., 1986-1991. Bacillariophyceae. In: Ettl, H., Gerloff, J., delphia, Philadelphia. April 27-29, 2007. Heynig, H., Mollenhauer, D. (Eds.), Sußwasserflora von Mitteleuropa. Fischer- Pereira, C.L., Raposeiro, P.M., Costa, A.C., Bao, R., Giralt, S., Gonçalves, V., 2014. Verlag, Stuttgart. Biogeography and lake morphometry drive diatom and chironomid assem- Krammer, K., 2000. Diatoms of the European Inland Waters and Comparable blages' composition in lacustrine surface sediments of oceanic islands. Hydro- Habitats. Volume 1. In: The Genus Pinnularia, Diatoms of the European Inland biologia 730, 93e112. https://doi.org/10.1007/s10750-014-1824-6. Waters and Comparable Habitats. Volume 1. A. R. G. Gantner Verlag, Ruggell, Queiroz, G., 1997. Vulca ~o das Sete Cidades (S. Miguel, Açores): Histo  ria Eruptiva e Liechtenstein, p. 703. Avaliaça ~o do Hazard. PhD Thesis. Universidade dos Açores. Departamento de Krammer, K., 2002. Diatoms of the European inland waters and comparable habi- Geocie ^ncias. tats. Volume 3. Cymbella. In: Lange-Bertalot, H. (Ed.), Diatoms of the European Queiroz, G., Pacheco, J.M., Gaspar, J.L., Aspinall, W.P., Guest, J.E., Ferreira, T., 2008. Inland Waters and Comparable Habitats, Volume 3. A. R. A. Gantner Verlag, The last 5000 years of activity at Sete Cidades volcano (Sa ~o Miguel island, Ruggell, Liechtenstein, p. 584. Azores): implications for hazard assessment. J. Volcanol. Geoth. Res. 178, Krammer, K., 2003. Diatoms of the European inland waters and comparable habi- 562e573. https://doi.org/10.1016/j.jvolgeores.2008.03.001. tats. Volume 4. Cymbopleura, Delicata, Navicymbula, Gomphocymbellopsis, Afro- Raposeiro, P.M., Hern andez, A., Pla-Rabes, S., Gonçalves, V., Bao, R., S aez, A., cymbella. In: Lange-Bertalot, H. (Ed.), Diatoms of the European Inland Waters Shanahan, T., Benavente, M., de Boer, E.J., Richter, N., Gordon, V., Marques, H., and Comparable Habitats, Volume 4. A. R. A. Gantner Verlag, Ruggell, Sousa, P.M., Souto, M., Matias, M.G., Aguiar, N., Pereira, C., Ritter, C., Rubio, M.J., Liechtenstein, p. 530. Salcedo, M., Va zquez-Loureiro, D., Margalef, O., Amaral-Zettler, L.A., Costa, A.C., Lange-Bertalot, H., 2001. Diatoms of the European inland waters and comparable Huang, Y., van Leeuwen, J.F.N., Masque , P., Prego, R., Ruiz-Fern andez, A.C., habitats. Volume 2. Navicula sensu stricto, 10 genera separated from Navicula Sanchez-Cabeza, J.-A., Trigo, R., Giralt, S., 2021. Climate change facilitated the sensu lato, Frustulia. In: Lange-Bertalot, H. (Ed.), Diatoms of the European Inland early colonization of the Azores Archipelago during medieval times. Proc. Natl. Waters and Comparable Habitats, Volume 2. A. R. G. Gantner Verlag, Ruggell, Acad. Sci. U.S.A. 118, e2108236118. https://doi.org/10.1073/pnas.2108236118. Liechtenstein, p. 526. Raposeiro, P.M., Rubio, M.J., Gonza lez, A., Hern andez, A., Sa nchez-Lo pez, G., Lange-Bertalot, H., Bak, M., Witkowski, A., 2011. Diatoms of the European inland Vazquez-Loureiro, D., Rull, V., Bao, R., Costa, A.C., Gonçalves, V., S aez, A., waters and comparable habitats. Volume 6. Eunotia and some related genera. Giralt, S., 2017. Impact of the historical introduction of exotic fishes on the 15 zquez-Loureiro, A. Sa D. Va ez, V. Gonçalves et al. Quaternary Science Reviews 303 (2023) 107968 chironomid community of Lake Azul (Azores Islands). Palaeogeogr. Palae- Shotton, F.W., Williams, R.E.G., 1971. Birmingham university radiocarbon dates V. oclimatol. Palaeoecol. 466, 77e88. https://doi.org/10.1016/j.palaeo.2016.11.015. Radiocarbon 13, 141e156. https://doi.org/10.1017/S0033822200008419. Raposeiro, P.M., Sa ez, A., Giralt, S., Costa, A.C., Gonçalves, V., 2018. Causes of spatial Sivarajah, B., Rühland, K.M., Labaj, A.L., Paterson, A.M., Smol, J.P., 2016. Why is the distribution of subfossil diatom and chironomid assemblages in surface sedi- relative abundance of Asterionella formosa increasing in a Boreal Shield lake as ments of a remote deep island lake. Hydrobiologia 815, 141e163. https:// nutrient levels decline? J. Paleolimnol. 55, 357e367. https://doi.org/10.1007/ doi.org/10.1007/s10750-018-3557-4. s10933-016-9886-2. Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Skov, T., Buchaca, T., Amsinck, S.L., Landkildehus, F., Odgaard, B.V., Azevedo, J., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Gonçalves, V., Raposeiro, P.M., Andersen, T.J., Jeppesen, E., 2010. Using inver- Haflidason, H., Hajdas, I., Hatte , C., Heaton, T.J., Hoffmann, D.L., Hogg, A.G., tebrate remains and pigments in the sediment to infer changes in trophic Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., structure after fish introduction in Lake Fogo: a crater lake in the Azores. Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A., Turney, C.S.M., van der Hydrobiologia 654, 13e25. https://doi.org/10.1007/s10750-010-0325-5. Plicht, J., 2013. IntCal13 and marine13 radiocarbon age calibration curves  Smilauer, P., Leps, J., 2014. Multivariate Analysis of Ecological Data Using CANOCO 5. 0e50,000 years cal BP. Radiocarbon 55, 1869e1887. https://doi.org/10.2458/ Cambridge University Press, Cambridge, UK. azu_js_rc.55.16947. Sorvari, S., Korhola, A., Thompson, R., 2002. Lake diatom response to recent Arctic Renberg, I., 1990. A procedure for preparing large sets of diatom slides from sedi- warming in Finnish Lapland. Global Change Biol. 8, 171e181. https://doi.org/ ment cores. J. Paleolimnol. 4, 87e90. https://doi.org/10.1007/BF00208301. 10.1046/j.1365-2486.2002.00463.x. Reynolds, C.S., 2006. The Ecology of Phytoplankton. Cambridge University Press, Stone, J.R., Fritz, S.C., 2004. Three-dimensional modeling of lacustrine diatom Cambridge, NY. habitat areas: improving paleolimnological interpretation of planktic : benthic Richter, N., Russell, J.M., Amaral-Zettler, L., DeGroff, W., Raposeiro, P.M., ratios. Limnol. Oceanogr. 49, 1540e1548. Gonçalves, V., de Boer, E.J., Pla-Rabes, S., Herna ndez, A., Benavente, M., Ritter, C., Stuiver, M., 1993. Editorial comment. Radiocarbon 35, iii. https://doi.org/10.1017/ Saez, A., Bao, R., Trigo, R.M., Prego, R., Giralt, S., 2022. Long-term hydroclimate S0033822200013771. variability in the sub-tropical North Atlantic and anthropogenic impacts on lake Supico, F., 1995. As Escavaço ~es. Instituto Cultural de Ponta Delgada, Ponta Delgada, ecosystems: a case study from Flores Island, the Azores. Quat. Sci. Rev. 285, Portugal. 107525. https://doi.org/10.1016/j.quascirev.2022.107525. Talbot, M.R., 2001. Nitrogen isotopes in paleolimnology. In: Smol, J.P., Birks, H.J.B., Ritter, C., Gonçalves, V., Pla-Rabes, S., de Boer, E.J., Bao, R., Sa ez, A., Herna ndez, A., Last, W.M. (Eds.), Tracking Environmental Change Using Lake Sediments. Vol- Sixto, M., Richter, N., Benavente, M., Prego, R., Giralt, S., Raposeiro, P.M., 2022. ume 2: Physical and Geochemical Methods. Kluwer Academic Publishers, The vanishing and the establishment of a new ecosystem on an oceanic island e Dordrecht, Netherlands, pp. 401e439. anthropogenic impacts with no return ticket. Sci. Total Environ. 830, 154828. Telford, R.J., Barker, P., Metcalfe, S.E., Newton, A., 2004. Lacustrine responses to https://doi.org/10.1016/j.scitotenv.2022.154828. tephra deposition: examples from Mexico. Quat. Sci. Rev. 23, 2337e2353. Roberts, N., Allcock, S.L., Arnaud, F., Dean, J.R., Eastwood, W.J., Jones, M.D., Leng, M.J., Ter Braak, C.J.F., Smilauer, P., 1998. CANOCO Reference Manual and User's Guide to Metcalfe, S.E., Malet, E., Woodbridge, J., Yig itbaşıog lu, H., 2016. A tale of two CANOCO for Windows: Software for Canonical Community Ordination (Version lakes: a multi-proxy comparison of Lateglacial and Holocene environmental 4). Microcomputer Power, Ithaca, NY. change in Cappadocia, Turkey. J. Quat. Sci. 31, 348e362. https://doi.org/10.1002/ Teranes, J.L., Bernasconi, S.M., 2005. Factors controlling d13C values of sedimentary jqs.2852. carbon in hypertrophic Baldeggersee, Switzerland, and implications for inter- Rühland, K., Paterson, A.M., Smol, J.P., 2008. Hemispheric-scale patterns of climate- preting isotope excursions in lake sedimentary records. Limnol. Oceanogr. 50, related shifts in planktonic diatoms from North American and European lakes. 914e922. https://doi.org/10.4319/lo.2005.50.3.0914. Global Change Biol. 14, 2740e2754. https://doi.org/10.1111/j.1365- Tilman, D., Kiesling, R., Sterner, R., Kilham, S., Johnson, F., 1986. Green, bluegreen 2486.2008.01670.x. and diatom algae: taxonomic differences in competitive ability for phosphorus, Rühland, K.M., Paterson, A.M., Smol, J.P., 2015. Lake diatom responses to warming: silicon and nitrogen. Arch. Hydrobiol. 106, 473e485. reviewing the evidence. J. Paleolimnol. 54, 1e35. https://doi.org/10.1007/ Tilman, D., Kilham, S.S., Kilham, P., 1982. Phytoplankton community ecology: the s10933-015-9837-3. role of limiting nutrients. Annu. Rev. Ecol. Systemat. 13, 349e372. https:// Rull, V., Lara, A., Rubio-Ingle s, M.J., Giralt, S., Gonçalves, V., Raposeiro, P., doi.org/10.1146/annurev.es.13.110182.002025. Hern andez, A., Sanchez-Lo  pez, G., Va zquez-Loureiro, D., Bao, R., Masque , P., Vazquez-Loureiro, D., Gonçalves, V., Sa ez, A., Hernandez, A., Raposeiro, P.M., Saez, A., 2017. Vegetation and landscape dynamics under natural and anthro- Giralt, S., Rubio-Ingle s, M.J., Rull, V., Bao, R., 2019. Diatom-inferred ecological pogenic forcing on the Azores Islands: a 700-year pollen record from the Sa ~o responses of an oceanic lake system to volcanism and anthropogenic pertur- Miguel Island. Quat. Sci. Rev. 159, 155e168. https://doi.org/10.1016/ bations since 1290 CE. Palaeogeogr. Palaeoclimatol. Palaeoecol. 534, 109285. j.quascirev.2017.01.021. https://doi.org/10.1016/j.palaeo.2019.109285. ez, A., Valero-Garce Sa s, B.L., Giralt, S., Moreno, A., Bao, R., Pueyo, J.J., Herna ndez, A., Volkov, D.L., Fu, L.-L., 2010. On the reasons for the formation and variability of the Casas, D., 2009. Glacial to Holocene climate changes in the SE pacific. The Azores current. J. Phys. Oceanogr. 40, 2197e2220. https://doi.org/10.1175/  Raraku lake sedimentary record (Easter Island, 27 S). Quat. Sci. Rev. 28, 2010JPO4326.1. 2743e2759. https://doi.org/10.1016/j.quascirev.2009.06.018. Wetzel, R.G., 2001. Limnology. Academic Press, San Diego, CA. Saraswati, P.K., Srinivasan, M.S., 2015. Micropaleontology: Principles and Applica- Williamson, C.E., Dodds, W., Kratz, T.K., Palmer, M.A., 2008. Lakes and streams as tions. Springer, Cham, Switzerland. sentinels of environmental change in terrestrial and atmospheric processes. Saros, J.E., Michel, T.J., Interlandi, S.J., Wolfe, A.P., 2005. Resource requirements of Front. Ecol. Environ. 6, 247e254. https://doi.org/10.1890/070140. Asterionella formosa and Fragilaria crotonensis in oligotrophic alpine lakes: im- Williamson, C.E., Saros, J.E., Schindler, D.W., 2009. Sentinels of change. Science 323, plications for recent phytoplankton community reorganizations. Can. J. Fish. 887e888. https://doi.org/10.1126/science.1169443. Aquat. Sci. 62, 1681e1689. https://doi.org/10.1139/f05-077. Wolfe, A.P., Cooke, C.A., Hobbs, W.O., 2006. Are current rates of atmospheric ni- Seddon, A.W.R., Mackay, A.W., Baker, A.G., Birks, H.J.B., Breman, E., Buck, C.E., trogen deposition influencing lakes in the Eastern Canadian Arctic? Arctic Ellis, E.C., Froyd, C.A., Gill, J.L., Gillson, L., Johnson, E.A., Jones, V.J., Juggins, S., Antarct. Alpine Res. 38, 465e476. https://doi.org/10.1657/1523-0430(2006)38 Macias-Fauria, M., Mills, K., Morris, J.L., Nogue s-Bravo, D., Punyasena, S.W., [465:ACROAN]2.0.CO;2. Roland, T.P., Tanentzap, A.J., Willis, K.J., Aberhan, M., van Asperen, E.N., Wolfe, A.P., Hobbs, W.O., Birks, H.H., Briner, J.P., Holmgren, S.U., Ingo  lfsson, O., Austin, W.E.N., Battarbee, R.W., Bhagwat, S., Belanger, C.L., Bennett, K.D., Kaushal, S.S., Miller, G.H., Pagani, M., Saros, J.E., Vinebrooke, R.D., 2013. Strati- Birks, H.H., Ramsey, C.B., Brooks, S.J., de Bruyn, M., Butler, P.G., Chambers, F.M., graphic expressions of the HoloceneeAnthropocene transition revealed in Clarke, S.J., Davies, A.L., Dearing, J.A., Ezard, T.H.G., Feurdean, A., Flower, R.J., sediments from remote lakes. Earth Sci. Rev. 116, 17e34. https://doi.org/10.1016/ Gell, P., Hausmann, S., Hogan, E.J., Hopkins, M.J., Jeffers, E.S., Korhola, A.A., j.earscirev.2012.11.001. Marchant, R., Kiefer, T., Lamentowicz, M., Larocque-Tobler, I., Lo pez-Merino, L., Yamamoto, A., Palter, J.B., 2016. The absence of an Atlantic imprint on the multi- Liow, L.H., McGowan, S., Miller, J.H., Montoya, E., Morton, O., Nogue , S., decadal variability of wintertime European temperature. Nat. Commun. 7, Onoufriou, C., Boush, L.P., Rodríguez-Sa nchez, F., Rose, N.L., Sayer, C.D., 10930. https://doi.org/10.1038/ncomms10930. Shaw, H.E., Payne, R., Simpson, G., Sohar, K., Whitehouse, N.J., Williams, J.W., Yu, Q., Liu, Z., Chen, Y., Zhu, D., Li, N., 2018. Modelling the impact of hydrodynamic Witkowski, A., 2014. Looking forward through the past: identification of 50 turbulence on the competition between Microcystis and Chlorella for light. Ecol. priority research questions in palaeoecology. J. Ecol. 102, 256e267. https:// Model. 370, 50e58. https://doi.org/10.1016/j.ecolmodel.2018.01.004. doi.org/10.1111/1365-2745.12195. 16