Climatic and anthropogenic influences on vegetation changes during the last 5000 years in a seasonal dry tropical forest at the northern limits of the Neotropics

Abstract

Seasonal dry tropical forest (STDF) is a widespread vegetation type in western Mexico. Generally, this type of forest develops in semihumid habitats where lacustrine basins are scarce, preventing documentation of the vegetation history and forest dynamics using a paleoecological approach. Here, we present a palynological record from the Santa Maria del Oro crater lake, located within the distribution area of the STDF and the adjacent dry oak forest which gives insight into the changes in diversity and human impact at this tropical site. Pollen data, in combination with geochemical analysis and microcharcoal data from a lacustrine sequence, are used to discuss vegetation change and diversity in the plant assemblages related to drought or anthropogenic activity over the last 5000 years. Our results show three distinct periods of drought in the basin, from 4200 to 3850, 3100 to 2300, 1570 to 1100 and 300 cal year BP, with changes in vegetation composition mainly related to a decrease in taxa diversity during these periods. Based on the presence of maize pollen, two periods of human activity (3790–2160 and 280 cal year BP to present) were detected with an increase in herbaceous pollen used as a surrogate for deforestation. These two disturbance periods, pre-Columbian and postcolonial, occurred during wet conditions in the basin. Our results highlight the correlation between El Niño events, drought and fire with changes in the composition and diversity of STDF.

Keywords

diversity drought Hill numbers palynology tropical dry forest western Mexico

Introduction

Seasonally dry tropical forests (SDTF) are one of the most diverse and widespread vegetation types in the dry tropics (Pérez-García et al., 2012), inhabited by plant types that are adapted to an extended dry season (4–8 months). This vegetation is characteristic of the lowlands (<1400 m a.s.l.) in the northern Neotropics (Rzedowski, 1978), forming an almost continuous band from southern Sonora to central America (Pennington et al., 2000; Rzedowski and Calderon de Rzedowski, 2013). This vegetation generally grows on moderate to steep slopes in warm sub-humid climates, with annual mean precipitation less than 1600 mm/year, mean annual temperatures between 17°C and 27.4°C (Janzen, 1988; Pennington et al., 2000; Trejo and Dirzo, 2002). It poses high turnover rates, suggesting high local diversification (Rzedowski, 1962, 1991).

According to Rzedowski (1991), SDTF distribution in Mexico encompasses high diversity with 60% of endemic species and 11% of endemic genera. The high diversity can be related to the complex geologic and phytogeographical history of the region (Kohlman and Sánchez-Colon, 1984). Time-calibrated molecular phylogeny studies of Bursera, a dominant member of the STDF, suggest that the establishment of the STDF in western Mexico occurred 30–20 Ma, corresponding with the formation of the Trans-Mexican Volcanic Belt (TMVB) and the Sierra Madre Occidental (Becerra, 2005). It has also been suggested that climate changes during the Pleistocene drove fragmentation and expansion processes associated with plant speciation (Pennington et al., 2000; Trejo and Dirzo, 2002). Nonetheless, research on the vegetation history and forest dynamics of SDTF in western Mexico during the Holocene using a paleoecological approach is limited. One of the reasons for the lack of paleoecological studies related to STDF is the paucity of lacustrine basins with good pollen records and, the use of forest hollows to reconstruct vegetation using fossil pollen, is challenging due to the sub-humid climate habitats where this forest develops.

Lacustrine sediment sequences provide information regarding past landscape changes. These natural terrestrial archives offer information regarding changes in plant communities, hydrological and depositional conditions and climate change. Due to the complex geological history of the TMVB, many of lakes have formed since the Neogene and efforts have been made to reconstruct the Pleistocene and Holocene environmental history based on the study of sediment cores from lakes in this region. The Holocene palynological records from the central region of the TMVB mainly focused on temperate ecosystems, documenting changes in plant composition and distribution in response to several drivers such as climate variability, (e.g. ENSO, droughts) and other disturbances like early agriculture, deforestation and fires. During early Holocene (11,700–8200 cal year BP), a trend to low lake levels and saline environments in several lacustrine records of the TMVB have been recognised, this condition has been associated to high summer insolation (Bhattacharya et al., 2017; Del Castillo-Batista et al., 2018; Figueroa-Rangel et al., 2008; Lozano-García and Vázquez-Selem, 2005; Lozano-García et al., 2020; Park et al., 2010). By mid-Holocene (8200–4200 cal year BP), some records with low lake levels showed a recovery (Caballero et al., 2002; Lozano-García et al., 2013; Vázquez-Castro et al., 2017); at some sites freshwater conditions were documented although no paleolimnological records are available for the western end of the TMVB. For late-Holocene (4200 cal year BP to the present), a dry period that started at the end of the mid-Holocene, continue until 3000–2000 cal year BP; after this period, a slight recovery in lake levels, is detected. According to the records, another dry period occurs at around 1.5 cal year BP (AD 600) to 0.8 cal year BP (AD 1100) that correlates with the demise of Mesoamerican cultures Lozano-Garcia et al., 2020).

For the western tropical end of the TMVB, scarce palynological Holocene records are available regarding climate variability and long-term vegetation changes (Brown, 1985). A trend towards drier conditions together with the occurrence of agriculture activities are documented from Lago Guzman; pollen data suggests drier conditions and agriculture at 750 cal year BP (AD 1200) and for Laguna San Pedro pollen record, lying within the STDF zone, shows a period of dry conditions and or agriculture from 3000 year BP to ca 800 year BP (1050 BC to AD 1150) although interpretation is problematic due to low resolution in pollen counting (Brown, 1985).

The Santa Maria del Oro (SMO) is a crater lake located at the northern limits of the Neotropics, it lies in the western portion of the TMVB, at the intersection of the TMVB and the Sierra Madre Occidental (Figure 1a); it represents a suitable site to study SDTF evolution given its location at the climatic boundary between the northern Mexico arid climates and the temperate sub-humid to humid climates of central Mexico. Previous paleoecological studies of lacustrine sediments in SMO using diatoms and ostracods, along with magnetic mineralogy and inorganic geochemistry, have provided evidence on significant environmental changes during the last 2000 years, particularly those regarding climate change and human impact (Rodríguez-Ramírez et al., 2015; Vázquez-Castro et al., 2008); however, there are no paleoecological records of the tropical plant communities around the basin and their responses to climate variability and human disturbances. Therefore, the main aim of this study was the reconstruction of plant assemblages in SMO in order to answer the following questions: (i) based on taxa composition, was the SDTF an stable plant community over the last 5000 years? (ii) is the plant community responding to periods of climate change (droughts) and/or to anthropogenic activity? and (iii) what were the main drivers climate change?

Study site

SMO crater lake is located in west-central Mexico (21°22′N, 104°34′W at 730 m a.s.l.), 65 km from the Pacific coast in the state of Nayarit (Figure 1a). The crater walls are dacitic-rhyolitic in composition with some basaltic flows on the northern slopes. According to Vázquez-Castro et al. (2008), the crater is probably of Pleistocene age based on its alignment with other volcanic cones in the area. The lake inside the crater is almost circular, with an area of 3.7 km² and a maximum depth of 65 m (Caballero et al., 2013; Serrano et al., 2002). Most of the lake lacks a shallow littoral zone because the crater walls are very steep down to the bottom of the lake, which is nearly flat (Figure 1b).

Previous paleoenvironmental studies based on magnetic mineralogy, geochemistry, diatoms and ostracods were undertaken on a littoral core (SMO02V, 880 cm) (Figure 1b) retrieved from a small shallow (<20 m) bay in the southwestern part of the lake (Rodríguez-Ramírez et al., 2015; Vázquez-Castro et al. 2008). Another sediment sequence was retrieved from the central flat area (MOLE SMO03, 894 cm), for which geochemical data have been published for the top 160 cm (Sosa-Nájera et al., 2010) and shown in Figure 1b. The lake is warm monomictic with stable stratification for 9–10 months; the water is slightly alkaline with a pH of 8.6 and evidence of evaporative concentration, Ca²⁺ depletion and Cl⁻ enrichment of lake waters (Caballero et al., 2013).

The modern climate in SMO is tropical sub-humid with predominantly summer precipitation. Mean annual precipitation is 1214 mm, and the rainy season occurs from June to October responding to the northward migration of the Intertropical Convergence Zone (ITCZ) and the onset of the North America Monsoon (NAM) system. Precipitation in western Mexico is modulated by the ITCZ position and the intensity of NAM (Metcalfe et al., 2015). Summer precipitation is reduced during El Niño events in central and southern Mexico and the Pacific coast of Central America (Castro et al., 2012; Magaña et al., 2003). The dry season lasts between 5 and 8 months. The mean annual temperature is 21°C with monthly mean temperatures ranging from 16.6°C (January and February) to 25°C (June and July) (SMN, 2019, 18005-Cerro Blanco meteorological station).

At present, SMO is surrounded by SDTF, a deciduous community that loses its leaves for 5–8 months each year (Figure 1c and d). Tree species include Bursera simaruba (L) Sarg., Enterolobium cyclocarpum (Jacq.) Griseb., Piscidia piscipula Sarg., Pseudobombax ellipticum (Kunth) Dugand., Haematoxylum brasiletto H. Karst., Parkinsonia praecox (Ruiz&Pav.) Hawkins, Lysiloma acapulcense (Kunth) Benth., Alvaradoa amorphoides Liebm., Diphysa suberosa S. Watson and Guettarda macrosperna Donn.Sm. Sub-deciduous Tropical forest (SDTF) elements are present in ravines with tree species such as Acacia polyphylla DC., Astronium graveolens Jacq., Brosimum alicastrum Sw., Ceiba pentandra (L.) Gaertn., Cocoloba barbadensis Jack., Ficus glabrata Kunth, Lonchocarpus hermannii M Sousa and Randia armata (Sw.) DC.. Numerous vines plants and epiphytes are important elements as well (Tellez, 1995). Oak tropical forest is conspicuous on the steep northern slopes with Quercus castanea Née as the dominant species, and common shrubs including Acaciella angustissima (Mill.) Britton& Rose, Acacia tequilana S. Watson, Bolanosa coulteri A.Gray, Rhamnus palmeri S. Watson, Tephrosia sinapou (Buc′hoz) A. Chev and Mimosa somnians Humb. & Bonpl. Ex Willd. Herbaceous cover is highly diverse with species belonging to the Leguminosae and Asteraceae species such as Desmodium angustifolium (Kunth) DC., D.cinereum, . Poepp. Ex Grisb. Eriosema multiflorum Robinson, Crotalaria micans Link, Tephrosia nicaraguense Oerst., Tagetes lucida Cav., Verbesina tequilana J.R. Coleman, as well as numerous graminoids (Tellez, 1995).

Archaeological framework

Some of the earliest human remains in western Mexico belongs to Matanchén bay (Nayarit), dating back to 2200–1730 BC (Williams, 2020) and, according to Mountjoy (2015), there is an archaeological gap between the last Archaic record and the first dates of the Formative period (1500 BC–AD 300). For west Mexico, the first farmer groups of the Middle Formative period (1200 BC arrived into an uncolonised territory (Mountjoy, 2015), while sedentary groups appear to establish during a period of wet conditions (Beekman, 2010). Maize pollen found in lake sediments from Laguna San Pedro in the Nayarit state, indicates agriculture activities at ca. 3000 year BP to ca 800 year BP (1050 BC–AD 1150 (Brown, 1985)). During the Classic period (AD 300–900), considered as the cultural blooming of the Maya region and central Mexico civilisations, the western area of Mesoamerica also shows a cultural development; this region played an intermediary role between central and southern Mesoamerica (Williams, 2020). Several records indicate that the late Classic (600–1000 AD) was a period of significant droughts and extreme changes, with the decline of the Teotihuacan city-state (Park et al., 2019). For the western region, there is evidence of decreases in human occupation (Beekman, 2010; Vazquez-Castro et al., 2019). This trend continues during the early Postclassic (1000–1522 AD) with the abandonment of ceremonials sites but with the rise of the Pacific Coast communities, which developed intensive floodplain agriculture and significant technological advances such as metallurgy (Beekman, 2010). Soon after the Spaniards arrival (1532), western Mexico was recognised as a mineral-rich area and a series of roads and bridges were constructed in the XVII century for silver and gold exploitation. During the XVIII century, near the town of Santa Maria del Oro, gold mining was carried out and this activity continues until today.

Materials and methods

Sample collection and chronology

Sediment cores were taken from the central part of the lake (57 m deep; 21°21.901′N, 104°34.093′W) (Figure 1) using the Kullenberg coring system from the National Lacustrine Core Facility (LacCore) at the University of Minnesota. Three continuous parallel cores 1A, 1B and 2A, 6.2, 7.3 and 8.2 m deep, respectively, and one short core 1BM (71 cm) were retrieved and transported to LacCore for documentation including high-resolution photographs, magnetic susceptibility and density measurements. A master sequence (MOLE-SMO03, 894 cm long) was constructed by visual comparison of high-resolution photographs and magnetic susceptibility data of the four cores. Sediment sub-sampling for pollen and other proxy analyses was done only down to 868 cm depth at LacCore at approximately 10–12 cm intervals.

The chronological model was established based on seven radiocarbon dates (Table 1), assuming that the top of the sequence dates to the year of core collection (i.e. 2003). Samples for age determination were dried and sent to Beta Analytic for AMS radiocarbon dating. Dates were calibrated with IntCal13 (Reimer et al., 2013) and an age-depth model was elaborated with the Bacon-v2.3.3 (Blaauw and Christen, 2011) package available in R v3.4.3 (R Development Core Team, 2017) using Bayesian statistics.

Table 1.

AMS radiocarbon dates of the Santa María del Oro master sequence MOLE-SMO03.

Laboratory ID	Depth (cm)	Radiocarbon¹⁴date (year BP) and error	Calibrated yr BP (2 σ)	Material dated
Beta 198312	170	750 ± 40	652–739	Bulk sediment organic matter
Beta 227823	180	780 ± 40	666–768	Bulk sediment organic matter
Beta 227824	248	3530 ± 40	3910–3700	Bulk sediment organic matter
Beta 227826	321	2520 ± 40	2740–2470	Bulk sediment organic matter
Beta 198313	368	2010 ± 50	1869–2114	Bulk sediment organic matter
Beta 227825	510	2810 ± 40	2791–3005	Bulk sediment organic matter
Beta 198315	733	3870 ± 50	4151–4420	Bulk sediment organic matter

Geochemical analysis

The MOLE-SMO03 master sequence sediments down to 761 cm, were analysed using an ITRAX X-ray sediment core scanner at 1 mm resolution at Southampton University, UK. For this study we only selected Ti and Ca data, and the intensities data count per second (cps) were transformed to percentages following the protocol established previously by Sosa-Najera et al. (2010). This transformation was based on a linear regression between the cps data and elemental concentration of selected samples that were analysed using traditional X-ray fluorescence (XRF) equipment (Siemens SRS 3000) at the Institute of Geology, UNAM (Sosa-Nájera et al., 2010). Titanium is an insoluble element present in minerals of catchment rocks and it was used as a run-off indicator (Metcalfe et al., 2010; Sosa-Nájera et al., 2010). Sources of Ca are the catchment rocks and authigenic carbonates. Previous studies in SMO lake (Rodríguez-Ramírez et al., 2015; Sosa-Nájera et al., 2010) demonstrated that higher Ca concentrations are related to historical droughts; they used Ca normalisation against Ti to reduce the detritical input signal in the detection of drought periods.

Pollen and microcharcoal analysis

Samples for pollen extraction were analysed only down to 868 cm depth because of poor preservation in the calcareous laminations at depths > 868 cm in the MOLE-SMO03 master sequence. Sixty-three 0.5 cm³ sediment samples were processed for pollen extraction according to Batten (1999), including the addition of two Lycopodium clavatum spore tablets for pollen concentration calculations. Residues were mounted in glycerin jelly and examined with a Zeiss microscope at 40× and 100×. We counted 400 pollen types, excluding tracer spores, pollen of aquatic taxa, fern spores, microalgae and testaceae. The pollen assemblage included terrestrial pollen types. Identification was undertaken using comprehensive pollen reference collection from the Laboratorio de Paleoecología, Paleoambientes y Cambio Climático at UNAM which included regional material from a vegetation survey undertaken in the SMO basin in 2004 (Supplemental Material, available online). Charcoal particles >100 µm were counted on the pollen slides to calculate concentration (particles/cm3).

Statistical analysis

Tilia software (Grimm, 1991) was used to generate the percentage pollen diagrams and concentration calculations. Four zones were established based on a stratigraphically constrained cluster analysis (CONISS, Grimm, 1987) of the pollen taxa with percentages >1%, and including only trees, herbs, fern spores and pollen types (Figure 3).

Taxa diversity was estimated using four diversity metrics:

(a) Hill Numbers (Hill, 1973) were estimated as follow: (1) Hill N0: the number of species (S) found in a sample regardless of their abundance; (2) Hill N1: the exponential of the Shannon-Weaver diversity index represents the number (abundance) of taxa in each sample and weighting each taxon by its relative abundance; (3) Hill N2: the inverse of Simpson’s index and representing the number of very abundant (dominant) fossil pollen taxa (Gotelli and Ellison, 2013). Evenness was estimated as the ratio N2/N0, the dominant taxa relative to all taxa.

(b) Rate of Change: estimated using the Bray-Curtis dissimilarity index to calculate the dissimilarities between two pairs of samples to record changes over time.

Hill Numbers and rate of change were estimated using the vegan package (Oksanen et al., 2018) available in R-v3.4.3 (R Development Core Team, 2017).

To identify the temporal correlation of anthropogenic activity, drought occurrences and changes in the plant assemblages of the dry tropical forest, cross-correlation analysis was performed between Quercus and Zea mays with Ca/Ti as a proxy of high evaporation rate and low precipitation (Rodríguez-Ramírez et al., 2015). Cross-correlation analysis was also carried out between the metrics (N0, N1 and N2) with Ca/Ti, and microcharcoal particles with Zea mays and the three-diversity metrics; significant cross-correlation were tested at 95% confidence level. The analyses were developed with the cross-correlation function (CCF) available in R-v3.6.0 (R Core Team, 2019).

Results

Core description and age model

According to the age model, the base of the core (894 cm) was determined to be ca. 5000 cal year BP, with a mean sedimentation rate of ~2 mm/year and an average pollen sample resolution of ~80 years (Figure 2).

The sediments are characterised by banded to laminated calcareous muds, fine sands, dark organic muds (sapropels) and clayey diatomaceous mud, with characteristic bundles of ~5–30 calcareous laminae separated by 2–8 cm of massive brown silt and fine sands, allowing for correlation between the four cores.

X-ray fluorescence

Titanium varied along the core from 0.17% to 0.83%, with values generally lower than average (<0.40%) between 761 and 688, 560–410 and 330–110 cm; higher than average values corresponding to 690– 580, 375–320 and 57–45 cm. The Ca/Ti ratio, with an average of 20.8, ranged between 8 and 101; values between 48 and 101 were present from 722–658, 320–240 and 27 cm and the top 10 cm (Figure 4).

Pollen record

Pollen composition comprised 88 taxa identified at genera or family level; they included 21 arboreal pollen and 53 non-arboreal pollen (36 herbaceous, 17 fern spores and 14 aquatics-not presented in this work) (Table 2). Besides, 24 taxa were counted as pollen types and 9 taxa as fern spore types. The pollen stratigraphy is presented in percentages showing the downcore changes (Figure 3). Based on the CONISS cluster analysis (Grimm, 1987), the sequence from SMO can be divided into four pollen zones (SMO-1–SMO-4), one of which is subdivided (Figure 4) into two subzones (SMO-2a and 2b).

Table 2.

Identified pollen from lake Santa María del Oro core. Taxa are grouped according to ecological preferences based on the plant survey of SMO basin. Pollen types and fern spores are not listed.

Seasonal tropical dry forest

Acacia

Acalypha

Acanthaceae

Agave

Amaranthaceae

Ambrosia

Anacardiaceae

Asteraceae

Brassica

Bocconia

Bombacaceae

Boraginaceae

Borreira

Bursera

Celastraceae

Celtis

Convolvulaceae

Cuphea

Euphorbiaceae

Fabaceae

Ficus

Heliocarpus

Iresine

Lamiaceae

Malpighiaceae

Meliaceae-Sapotaceae

Mimosa

Myrtaceae

Piperaceae

Poaceae

Psychotria

Rubiaceae

Sapindaceae

Verbesina

Trema

Zanthoxylum

Sub-deciduous tropical forest

Brosimum

Carpinus

Juglans

Liliaceae

Melastomataceae

Moraceae

Poulsenia

Salix

Ulmus

Urticaceae

Tropical Oak forest

Quercus

Alnus

Artemisia

Apiaceae

Begonia

Caryophyllaceae

Cupressaceae

Eryngium

Fraxinus

Thalictrum

Pinus

Human impact indicator

Zea mays

SMO-1

868–656 cm (4940–3800 cal year BP). Pollen percentages are dominated by Pinus (40%), Quercus (45%), Moraceae (10%) and Bursera (5%), with Poaceae (70%) and Amaranthaceae (8%) as the most important herbaceous pollen. The first occurrence of maize pollen corresponds to the sample at 656 cm (3990 cal year BP). The highest concentrations of charcoal (1500 particles/cm³) were recorded at the base of the zone, with a trend towards lower values with fluctuations.

SMO-2

623–220 cm (3600–1063 cal year BP) is primarily characterised by a continuous increase in herbaceous pollen percentages. This zone is subdivided into two subzones:

Subzone SMO-2a

623–321 cm (3593–1797 cal year BP) is represented by a sudden increase in Moraceae (20%) at the base of the subzone, a reduction in Quercus percentages (mostly below 20%), and an increase in Poaceae (52%), Asteraceae (25%) and Ambrosia (9 %) with the continuous presence of maize. Higher values in charcoal concentration are reached in the middle and the upper part of this subzone, after a decline at the base.

Subzone SMO-2b

270–220 cm (1450–1063 cal year BP). Pollen percentage spectra are dominated by Quercus (28%), Pinus (13%) and Poaceae (44%) together with a decline of Asteraceae (7%), Amaranthaceae (10%) and Ambrosia (1.7%). The absence of maize differentiates this subzone, and charcoal concentration declines considerably compared to the previous subzone SMO-2a.

SMO-3

212–78 cm (998–317 cal year BP). Pollen spectra show a decline in Quercus (19%) and herbaceous taxa such as Poaceae, Asteraceae and Amaranthaceae, while other taxa increase: Pinus (16%), Moraceae (12%), Brosimum (4%), Bursera (3%), Poulsenia (4%) and Piperaceae (7%). Maize was only recorded at 176 cm; charcoal concentrations continued with values from 1200 to 7000 particles. Quercus and Poaceae increase their percentages in the last two samples.

SMO-4

69.2 cm (279–3 cal year BP). A sudden increase in Amaranthaceae (40%) and Asteraceae (20%) and a decline in Poaceae (18%) with the continuous presence of maize characterises this zone. Microcharcoal concentrations are similar to SMO-3.

Taxa diversity

Hill number N0, which expresses the effective number of pollen taxa, varies through the sequence (with a 95% confidence interval) from 28 to 31 taxa; lower numbers (23 and 21 pollen taxa) corresponded to SMO-1 and SMO-2b pollen zones, respectively, and the highest number (43 pollen taxa) was found at SMO-3 (Figure 4).

Both N1 and N2 showed the same pattern of peaks and drops along the sequence; low values (4–11 pollen taxa) were observed in all pollen zones, except in SMO-3 where they presented the highest values (10–18 pollen taxa). N1, which counts taxa equally in proportion to their abundance, varied slightly from only 8 to 9 taxa; N2, which gives greater weight to the most abundant taxa, oscillated from 5 to 6 taxa (Figure 4).

Rate of change

The rate of change based on the pollen data was low (<0.1) in SMO-1 and in most of SMO-2a with an increase at 1800 cal year BP; lower values (0.05) were also present in SMO-2b. An increase in vegetation turnover was recorded in SMO-3, with a maximum (0.3) at 600 year cal BP. The rate of change decreased in the first section of zone SMO-4 (between 0.1 and 0.2) while an abrupt increase (0.4) was evident in the last section of this zone.

Cross-correlation

Cross-correlation results (Figure 5) revealed a positive correlation between Quercus and Ca/Ti at lag 0, indicating that they change together. Correlation between Ca/Ti and the diversity metrics (N0, N1, N2) is negative, so an increase in Ca/Ti coincides with a decrement in N0, N1 and N2. The same pattern of negative correlation was obtained between charcoal concentration with Zea mays, N0, N1 and N2 indicating that they shift inversely at the same time. Also, a negative correlation was obtained between Zea mays and the Ca/Ti ratio.

Discussion

Vegetation change during the last 5000 years

The 5000 years pollen record of SMO provides historical information on the evolution of the SDTF that grows at the northern limit of the Neotropics and also provides an insights of the evolution of the tropical oak forests adjacent to the SDTF. The fossil pollen assemblage in the small SMO basin was a combination of the SDTF and SCTF represented by a mixture of tropical taxa with relatively low pollen percentages (<5%) (e.g. Bombacaceae, Heliocarpus and Poulsenia) and the dry tropical oak forest. In many of the tropical taxa pollination is entomophilous, causing these taxa to be poorly represented in the pollen spectra. On the other hand, anemophilous taxa such as Quercus and Pinus showed the highest percentages in the sedimentary record.

Pollen spectra from ca. 4940–3800 cal year BP have a distinct signal with high percentages (60–70%) of Quercus and Pinus, while herbaceous pollen such as Asteraceae and Amaranthaceae showed low values; Poaceae had values of 30% except for two peaks. At present, pine forests and mixed pine-oak forests are the dominant vegetation in the highlands of central Mexico. Studies from several lake sequences, reconstructing Pleistocene and Holocene history, report Pinus with percentages fluctuating from 60% to 90% (Lozano-García et al., 2005; Park et al., 2010; Ruiz-Córdova et al., 2019; Torres-Rodríguez et al., 2012). In comparison, Quercus pollen generally ranged between 10% and 30%. Currently, on the northern slopes of SMO basin, oak groves are dominated by Quercus castanea. Therefore, late-Holocene fossil pollen spectra in this study point to a larger presence of this plant community at the site. This agrees with evidence on the permanence of a large population of Q. castanea in the TMVB since the Pleistocene (Peñaloza-Ramírez et al., 2020).

On the other hand, Pinus is absent from the present-day vegetation of the SMO basin, although a few Pinus individuals were observed on the eastern slopes outside the SMO basin. Studies involving the floristic composition of oak forests in the state of Nayarit, where SMO lake is located, reported the occasional presence of Pinus individuals along with several oak species (Tellez, 1995).

Changes in the pollen spectra between 3800 and 1700 cal year BP show an increase in non-arboreal pollen, and high values of Poaceae, Asteraceae, Amaranthaceae and Ambrosia, along with the presence of maize pollen (the first record was observed at 3900 cal year BP). These pollen assemblages evidence that human activities started in the catchment of SMO around this time. We hypothesise that the reduction in the percentages of Pinus observed after ca. 4000 cal year BP in the pollen diagram could be associated with the initiation of human impact in the region that resulted in logging targeting the extraction of wood from the Pinus species.

The herbaceous assemblage (Ambrosia, Asteraceae and Poaceae) reduced their percentages between 1500 and 300 cal year BP. In the first part of this period, between 1500 and 1000 cal year BP, Poaceae continued with high values and only Quercus showed an increase in values. The rate of change was higher from 1000 to 300 cal year BP) (Figure 4), Poaceae and Quercus percentages showed a reduction, and the taxa belonging to the SDTF such as Bursera, and the SCTF taxa as Brosimum, Ficus, Poulsenia, Moraceae and Piperaceae increased. During this period, maize pollen was only recorded in one sample at 760 cal year BP (AD 1200).

During the last 300 cal year BP (since AD1650), an increase of Amaranthaceae and Asteraceae marked a significant change in vegetation coupled with an abrupt rise in the rate of change. Evidence of agriculture, as indicated by the presence of maize pollen, correlates with a significant increase in the non-arboreal assemblage but with low charcoal concentration values.

Correlation between the plant assemblages and periods of drought

The dry tropics have lower diversity in terms of species richness compared to the humid tropics, but in western Mexico, the species richness of the SDTF is very high (Durán et al., 2006). Studies regarding the relationship between diversity and climate parameters such as precipitation and seasonality have shown a positive correlation between the diversity of the Mexican STDF with evapotranspiration (Trejo and Dirzo, 2002).

In the record of SMO, the cross-correlation analysis of Quercus and Ca/Ti shows that these two variables shifted together; confirming that Quercus is a drought indicator. Based on the combination of the Quercus pollen spectra and the Ca/Ti ratio, we inferred four main periods of drought conditions in the SMO basin during the last 5000 years: from ca. 4200–3850 cal year BP, 3100–2300, 1570–1100 cal year BP and 300 cal year BP. (A) During the first drought period, Quercus attained values of 50% and the Ca/Ti ratio was high. (B) Throughout the second period, ca. 3100–2300 cal year BP, Quercus attained irregular percentages, and there was an increase in Poaceae, possibly related to agricultural practices inferred from the presence of maize pollen. Agriculture probably took place on the slopes near the inlet of the south-western part of the catchment, as it occurs there today. In the other sectors of the crater, the slopes are too steep to carry out agricultural practices. (C) For the third drought period, dated from 1570 to 1000 cal year BP (AD 380–950), low concentrations of Ti and high values of Ca/Ti suggest severe droughts. Maize pollen was absent in these sediments leaving no evidence of human activity during this period in the basin. This time corresponds to the period of the Classic drought, which has been documented in several paleoecological records in central Mexico (Bhattacharya et al., 2017; Caballero et al., 2002; Metcalfe and Davies, 2007). In SMO, the record is characterised by an increase in Quercus percentages, low run-off with low Ti concentrations and high evaporation, as suggested by the Ca/Ti ratios. Paleoecological evidence for the same period obtained from the littoral core shows low lake levels from 1350–1150 cal year BP (500–1000 AD) and intense evaporation and reduced surface run-off from 600 to 800 AD (Rodríguez-Ramírez et al., 2015). (D) The most recent and intense drought at ca. 300 cal year BP with an episode of a sudden increase in Quercus percentages along with a high Ca/Ti ratio.

Previous high-resolution XRF of Ti and Ca concentration data and magnetic susceptibility for the last 700 year BP (since AD 1250) in the SMO sequence (top 160 cm), together with historical information and tree-ring chronologies, allowed us to document significant drought events. The drought of 295 cal year BP (AD1655–1670) was one of the most intense, characterised by an increase in Ca and a reduction in Ti percentages (Sosa-Nájera et al., 2010). According to Stahle et al. (2016), this drought reconstructed on tree-ring chronologies at 1666–1669, was one of the two most extreme of the past 600 years in Mexico. Other analyses, including a combination of magnetic mineralogy, organic and inorganic carbon and geochemistry have been carried out in a sequence that covers the last 2600 years (SMO02-V) drilled in the littoral zone of SMO lake (Figure 1b, Vázquez-Castro et al., 2008). Data in this earlier publication shows warmer and drier conditions between 1350 and 810 cal year BP (AD 600–1140) and promoting high pH in the lake and the precipitation of carbonates. Rodríguez-Ramírez et al. (2015) using ostracod and diatom content in combination with magnetic susceptibility, total inorganic carbon and Ca/Ti concentration values in the same core showed periods of lower lake levels associated with higher Ca/Ti values which were associated with reduced precipitation. The SMO littoral record corroborated that high Ca/Ti values in this lake can be taken as an indicator of high carbonate formation during periods of high evaporation, and therefore lower effective moisture.

Two periods of wetter conditions at SMO can be inferred based on the Ca/Ti record and high percentages of Moraceae pollen. The first event centered at 3700 cal year BP and the second between 1050– 650 cal year BP (AD 900–1300). Evenness values /N2/N0 ratio) are high during the periods of more moisture availability.

Anthropogenic activity in the SMO basin

It has been suggested that early agriculture dispersed inland from coastal western Mexico through river basins (Zizumbo-Villareal and Colunga-GarcíaMarín, 2010) and established in various lake basins, such as SMO. In several lacustrine sequences of the TMVB fossil maize is reported after 5000 cal year BP (Goman and Byrne, 1998; Lozano-García et al., 2010, 2013; Park et al., 2010; Sluyter and Dominguez, 2006) indicating the expansion of agriculture in central Mexico. Most of the late-Holocene paleoecological records from Mexico show evidence of anthropogenic activities, including the presence of agricultural taxa such as maize pollen, reduction in arboreal pollen due to deforestation, increases in charcoal particles pointing to frequent fires, increases in magnetic susceptibility suggesting higher erosion rates, and/or hiatuses in sedimentation (Caballero et al., 2002; Del Castillo-Batista et al., 2018; Figueroa-Rangel et al., 2008; Lozano-García et al., 2013; Park et al., 2010). The pollen record of Tixtla and Huitziltepec karstic lakes, located within the distribution area of the SDTF in southern Mexico (Berrío et al., 2006), shows the evidence of agricultural activities starting at 2700 cal year BP (750 BC). At Laguna San Pedro near SMO, preliminary pollen analysis suggested that maize agriculture could occur at ca. 3000 year BP (1050 BC) (Brown, 1985). Maize pollen at SMO was recorded before at 3990 cal year BP (2040 BC), suggesting an earlier agricultural development for this site.

In the pollen record from SMO, two periods of human disturbance based on the presence of maize pollen were observed, one between 3790 and 2160 cal year BP (1820–190 BC), and the other from 280 cal year BP to present (i.e. 1700 AD to present). The first interval falls within the Formative period (2000 BC–AD 100) and the second corresponds with the postcolonial period starting after the AD 1650 drought. Similarly, the pollen spectra in both periods of disturbance show an increase in non-arboreal pollen suggesting deforestation, but during the first period, Poaceae dominated the herbaceous assemblage, while during the second period, Amaranthaceae was the most abundant pollen taxa.

In several studies, high concentrations of microcharcoal, related to slash and burn agricultural practices have been reported for Mesoamerica (Kennett et al., 2010). However, in the SMO record, the association between the microcharcoal concentration and agriculture is not evident; this is verified by the negative correlation between charcoal and maize in the cross-correlation analysis (Figure 5). The Ca/Ti ratio is used as an indicator of warmer climate and lower precipitation. The cross-correlation analysis between this ratio and Zea mays, established a negative correlation which suggests that agricultural activities in the basin may have occurred during wet periods rather than dry periods, an idea consistent with the lower values of Ca/Ti and the high percentages of Moraceae. Archeological data for western Mexico documents an increase in population growth during the late Formative and early Classic periods (300 BC–600 AD) in the shaft and chamber tomb tradition (Beekman, 2010); however, in the SMO record there is no evidence of agriculture during this period. In the littoral core, some evidence of human presence is inferred during this time ending as the climate became dryer after AD 400 (Rodríguez-Ramírez et al., 2015). Our data suggest the disruption of agriculture by droughts at this site.

Climate, fires and diversity

Paleoecological studies are useful in assessing the responses on the STDF to natural disturbances such as fires and droughts (Stan et al., 2019). The present study showed changes in diversity in the pollen assemblages over the last 5000 cal year BP in SDTF in west-central Mexico using different metrics. When looking at N1 and N2 values, the higher diversity in the sequence corresponded to SMO-3 (998–317 cal year BP); this zone was characterised by humid conditions as the Ti and Ca/Ti proxies revealed. The resemblance in values and pattern of these two numbers (N1 and N2) exhibit that rare and abundant taxa responded similarly to changes in the environment, but rare taxa dominated the assemblage of the STDF. Also, the higher N2/N0 ratio is presented during the humid periods in the record.

Two drivers for diversity change (climate and human disturbance) intermingle throughout the vegetation history of SMO, both have effects on the pollen taxa diversity and the Ti, Ca/Ti and charcoal records. Although the dry forests are well-adapted to long periods of dryness (6–8 months), according to our results, the diversity of STDF was affected by droughts throughout the last 5000 years. Actual floristic research in the tropical dry forests of South America and Central America show that species diversity is lower away from the equator and with indications of negative correlation between the number of dry months with diversity and turnover (Golicher et al., 2012). Diversity and turnover decline when dry conditions are established in SMO, this pattern is clearly observed in the record during the period of significant droughts such as during the late Classic and for the AD 1650 drought.

Another disturbance factor related to droughts is fire (Torres-Rodríguez et al., 2015). Fire is an important disturbance factor that can affect diversity in the dry tropics (Rodríguez-Trejo, 2008). The role of fires in the STDF dynamics is poorly studied. Although natural fires are not considered as part of the Neotropics dry forests ecology (Middelton et al., 1997) recent studies in a dry forest in Nicaragua suggest that for this site, the forest is dominated by species that can survive infrequent and low-intensity fires (Otterstrom et al., 2006). More frequent are the anthropogenic fires, when agriculture activity promotes it for field preparation. The cross-correlation results between charcoal concentration and the diversity metrics showed a negative and synchronic correlation, implying that diversity decreases when fire occurrences are high. Our data show two periods of increase in microcharcoal, the first at the end of mid-Holocene with no evidence of agriculture in the record and the second with the presence of maize from 3100 to 1600 cal year BP. Since the SMO basin is close to the Pacific coast, El Niño events could promote wet-season drought, which could increase the occurrence of fire events in this ecosystem, altering the plant community composition for long-term periods with important changes in diversity. According to a synthesis by Metcalfe et al. (2015) related to Holocene climate change and its consequences in the continental environments of Mexico, the late-Holocene climatic pattern is complex, but generally dry conditions are established in most of the lacustrine basins in the TMVB. The main driver for the reduction in summer precipitation seems to be the weakening of the monsoon strength associated with a more southerly mean position of the ITZC. Today, the ITCZ is displaced to the south during El Niño events, reducing summer precipitation in Mexico (Magaña et al., 2003). In the paleorecords, an increase in strength and frequency of El Niño events during the mid to late-Holocene is observed in the Pallcacocha record (Moy et al., 2002, Figure 4). Also, in the palynological record of Zirahuen, a temperate site located east of the SMO in the TMVB, a dry event at ca. 4000 cal year BP was related to an increase in El Niño events. (Lozano-García et al., 2013). Tree-ring analyses of Mimosa acantholoba, a pioneer species of the dry tropical forest in southern Mexico, exhibited how large-scale climatic drivers as El Niño, influences tree growth through effects of local climate such as the reduction in rainfall and warmer climates. This research showed that growth was negatively affected by El Niño episodes with a reduction of 37% in growth during El Niño years (Brienen et al., 2010). Although this study referred only to one species, making it difficult to generalise to the tropical dry forest communities, it is evident that climate change affects the survival of the species and modify diversity.

We hypothesised that because SMO is close to the Pacific coast and receives rain during the summer months, El Niño events could promote significant drought conditions, as shown in the Ca/Ti values (Figure 5), which in turn alters the composition of the plant community by decreasing the diversity of its taxa as evidenced in this study.

Conclusion

In this paper, we have explored the history of the SDTF, one of the most widespread tropical vegetation types in Neotropics, and the adjacent oak forest in the SMO basin. Using pollen data and microcharcoal concentration and analysing geochemical proxies from the MOLE-SMO03 lacustrine sediments, past environmental conditions were reconstructed for the last 5000 years. The palynological data show that the main changes in the plant assemblages are represented by a combination of low SDTF pollen percentages and a high representation of the anemophilous Quercus and Pinus sp. in the fossil pollen spectra. Our data suggest a wider presence of the tropical oak forest at the site. Herbaceous pollen percentages of Poaceae, Amaranthaceae and Asteraceae correlate with the presence of maize pollen since 3900 cal year BP.

We found a correlation between the Ca/Ti ratio and Quercus, indicating low precipitation, and we used it as a drought indicator for this site. Based on these data, we inferred four main drought events (4200–3850, 3100–2300, 1570–1100 and ca. 300 cal year BP). The most severe drought occurred during the late Classic (1570–1000 cal year BP), with no evidence of agriculture at the site. Our results indicate the disruption of agriculture by droughts in SMO. Periods of wet conditions were documented in the sedimentary record based on high values of Moraceae pollen and Ti percentages; during these periods, the presence of maize pollen indicates human activities in the basin. We found a negative correlation between charcoal data and diversity metrics (Hill numbers and palynological richness), suggesting that diversity declined when fire events were high. Although the tropical dry forests are adapted to long periods of dryness, our results indicate that droughts affected the diversity of the STDF. Reduction in summer precipitation seems to be promoted by El Niño events, causing changes in the plant assemblages mainly inferred through changes in the diversity of taxa.

Our data reveal that vegetation of the STDF has mainly responded to climate change rather than human impact in the SMO lake during the last 5000 years.

Supplemental Material

sj-pdf-1-hol-10.1177_0959683620988054 – Supplemental material for Climatic and anthropogenic influences on vegetation changes during the last 5000 years in a seasonal dry tropical forest at the northern limits of the Neotropics

Supplemental material, sj-pdf-1-hol-10.1177_0959683620988054 for Climatic and anthropogenic influences on vegetation changes during the last 5000 years in a seasonal dry tropical forest at the northern limits of the Neotropics by Socorro Lozano-García, Blanca Figueroa-Rangel, Susana Sosa-Nájera, Margarita Caballero, Anders J Noren, Sarah E Metcalfe, Oswaldo Tellez-Valdés and Beatriz Ortega-Guerrero in The Holocene

Footnotes

Acknowledgements

We would like to thank those who contributed to the fieldwork in 2003 during the MOLE expedition: Douglas Schnurrenberger, Mark Shapley, Alejandro Ramírez-Rodríguez, Isabel Israde and Sarah Davies. Rodrigo Martinez Abarca helped with the map editing.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was founded by UNAM DGAPA-PAPIIT (México grant no. IN203102 and IN110106).

ORCID iD

Socorro Lozano-García

Supplemental material

Supplemental material for this article is available online.

References

Batten

(1999) Small palynomorphs. In: Jones

Rowe

(eds) Fossil Plants and Spores Modern Techniques. London: The Geological Society, pp.15–19.

Becerra

(2005) Timing the origin and expansion of the Mexican dry forest. Proceedings of the National Academy of Sciences 102(31): 10919–10923.

Beekman

(2010) Recent research in western Mexican archaeology. Journal of Archaeology Research 18: 41–109.

Berrío

Hooghiemstra

van Geel

, et al. (2006) Environmental history of the dry forest biome of Guerrero, Mexico, and human impact during the last c. 2700 years. The Holocene 16(1): 63–80.

Bhattacharya

Chiang

Cheng

(2017) Ocean-atmosphere dynamics linked to 800–1050 CE drying in Mesoamerica. Quaternary Science Reviews 169: 263–277.

Blaauw

Christen

(2011). Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6: 457–474.

Brienen

RJW

Lebrija-Trejos

Zuidema

, et al. (2010) Climate-growth analysis for a Mexican dry forest tree shows strong impact of sea surface temperatures and predicts future growth declines. Global Change Biology 16: 2001–2012.

Brown

(1985) A summary of late Quaternary pollen records from Mexico west of the Isthmus of Tehuantepec. In: Bryant

Holloway

(eds) Pollen Records of Late Quaternary North America Sediments. Dallas TX: America Association of Stratigraphic Palynologists Foundation, pp.71–93.

Caballero

Ortega

Valadez

, et al. (2002) Sta. Cruz Atizapán: A 22-ka lake level record and climatic implications for the late-Holocene human occupation in the Upper Lerma Basin, Central Mexico. Palaeogeography, Palaeoclimatology, Palaeoecology 186: 217–235.

10.

Caballero

Rodríguez

Vilaclara

, et al. (2013) Hydrochemistry, ostracods and diatoms in a deep, tropical, crater lake in Western Mexico. Journal of Limnology 72: 512–523.

11.

Castro

Chang

Dominquez

, et al. (2012) Can a regional climate model improve the ability to forecast the North American Monsoon? Journal of Climate 25: 8212–8237.

12.

Del Castillo-Batista

Figueroa-Rangel

Lozano-García

, et al. (2018) 1580 years of human impact and climate change on the dynamics of a Pinus-Quercus-Abies forest in west-central Mexico. Revista Mexicana de Biodiversidad 89: 208–225.

13.

Durán

Meave

Lott

, et al. (2006) Structure and diversity patterns at landscape level in a Mexican tropical deciduous forest. Boletín de la Sociedad Botánica de México 79: 43–60.

14.

Figueroa-Rangel

Willis

Olvera-Vargas

(2008) 4200 years of pine-dominated upland forest dynamics in west-central Mexico: Human or natural legacy? Ecology 89: 1893–1907.

15.

Goman

Byrne

(1998) A 5000-year record of agricultural and tropical forest clearance in the Tuxtlas, Veracruz, Mexico. The Holocene 8: 83–89.

16.

Golicher

Cayuela

Newton

(2012) Effects of climate change on the potential on the species richness of Mesoamerican forest. Biotropica 44(3): 284–293.

17.

Gotelli

Ellison

(2013) A Primer of Ecological Statistics, 2nd edn. Sunderland: Sinauer.

18.

Grimm

(1991) Tilia and Tiliagraph. Illinois State Museum: Springfield.

19.

Grimm

(1987) Coniss: A fortrand 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computers and Geosciences 13(1): 13–35.

20.

Hill

(1973) Diversity and evenness: A unifying notion and its consequences. Ecology 54: 427–432.

21.

Janzen

(1988) Tropical dry forest. The most endangered major tropical ecosystems. In: Wilson

(ed.) Biodiversity. Washington DC: National Academy Press, pp.130–137.

22.

Kennett

Piperno

Jones

, et al. (2010) Pre-pottery farmers on the Pacific coast of southern Mexico. Journal of Archaeological Science 37: 3401–3411.

23.

Kohlman

Sánchez-Colon

(1984) Estudio aerográﬁco del género Bursera en México. In: Ezcurra

Equihua

Colman

, et al. (eds) Métodos cuantitativos en la Biogeografía, México: Instituto de Ecología, pp.41–120.

24.

Lozano-García

Vázquez-Selem

(2005) A high-elevation Holocene pollen record from Iztaccíhuatl volcano, central Mexico. The Holocene 15: 329–338.

25.

Lozano-Garcı́a

Caballero

Ortega

, et al. (2010) Late-Holocene palaeoecology of Lago Verde: Evidence of human impact and climate change in the northern limit of the neotropics during the late formative and classic periods. Vegetation History and Archaeobotany 19: 177–190.

26.

Lozano-García

Torres-Rodríguez

Ortega

, et al. (2013) Ecosystem responses to climate and disturbances in western central Mexico during the late Pleistocene and Holocene. Palaeogoegraphy Palaeoclimatology Palaeoecology 370: 184–195.

27.

Lozano-García

Caballero

Ortega-Guerrero

, et al. (2020) Insights into the Holocene Environmental History of the Highlands of Central Mexico. In: Torres-Cano

Islebe

Roy

(eds) The Holocene and Anthropocene Environmental History of Mexico. Switzerland: Spinger Nature, pp.97–114.

28.

Magaña

Vázquez

Pérez

, et al. (2003) Impact of El Niño on precipitation in Mexico. Geofísica Internacional 42: 313–330.

29.

Metcalfe

Davies

(2007) Deciphering recent climate change in central Mexican lake records. Climatic Change 83: 169–186.

30.

Metcalfe

Barron

Davies

(2015) The Holocene history of the North American Monsoon: ‘known knows’ and ‘known unknowns’ in understanding its spatial and temporal complexity. Quaternary Science Reviews 120: 1–27.

31.

Middelton

Sáncez-Rojas

Suedmeyer

, et al. (1997) Fire in a tropical dry forest of central America: A natural part of the disturbance regime? Biotropica 29: 515–517.

32.

Mountjoy

(2015) La colonización del lejano occidente de México por agricultores sedentarios durante el Formativo medio, 1200 a 400 a.C. Revista Occidente (junio 2015): 1–18.

33.

Morrone

(2019) Regionalización biogeográfica y evolución de la biótica de México: encrucijada de la biodiversidad del nuevo mundo. Revista Mexicana de Biodiversidad 90: e9002980.

34.

Moy

Seltzer

Rodbell

, et al (2002) Variability of El Niño/Southern oscillation activity at millennial timescales during the Holocene epoch. Nature 420: 162–165.

35.

Oksanen

Blanchet

Friendly

(2018) Available at: https://cran.r-project.org/web/packages/vegan/index.Html (accessed 3 May 2020).

36.

Otterstrom

Schwartz

Velázquez-Rocha

(2006) Responses to fire in selected tropical dry forest trees. Biotropica 38(5): 592–598.

37.

Park

Byrne

Bohnel

(2019). Late-Holocene change in Central Mexico and the decline of Teotihuacan. Annals of the American Association of Geographers 109(1): 104, 120.

38.

Park

Byrne

Bohnel

, et al. (2010) Holocene climate change and human impact, central Mexico: A record based on maar lake pollen and sediment chemistry. Quaternary Science Reviews 29: 618–632.

39.

Pennington

Prado

Pendry

(2000) Neotropical seasonally dry forest and Quaternary vegetation changes. Journal of Biogeography 27: 261–273.

40.

Peñaloza-Ramírez

Rodríguez-Correa

Gonzales-Rodríguez

, et al. (2020) High genetic diversity and stable Pleistocene distributional ranges in the widespread Mexican oak Quercus castanea Née (1801) Fagaceae. Ecology and Evolution 10(10): 4204–4219.

41.

Pérez-García

Meave

Cevallos-Ferriz

SRS

(2012) Flora and vegetation of the seasonally dry tropics in Mexico: Origin and biogeographical implications. Acta Botánica Mexicana 100: 149–193.

42.

Reimer

Bard

Bayliss

, et al. (2013) IntCal13 and marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55(4): 1869–1887.

43.

R Core Team (2019) R: A language and environment for statistical computing. Available at: http://www.R-project.org (accessed 2 May 2020).

44.

Rodríguez-Ramírez

Caballero

Roy

, et al. (2015) Climatic variability and human impact during the last 2000 years in western Mesoamerica: Evidence of late Classic (AD 600-900) and little ice age drought events. Climates of the Past 11: 1239–1248.

45.

Rodríguez-Trejo

(2008) Fire regimes, fire ecology and fire management in Mexico. Ambio 37(7/8): 548–566.

46.

Ruiz-Córdova

Lozano-García

Caballero

, et al. (2019) Historia de la vegetación, ambiente y evidencia de actividad humana de los últimos 6,000 años en el lago alpino La Luna, Nevado de Toluca. Revista Mexicana de Biodiversidad 90: e902996.

47.

Rzedowski

(1962) Contribuciones a la fitogeografía florística e histórica de México I. Algunas consideraciones acerca del elemento endémico en la flora mexicana. Boletín de la Sociedad Botánica de México 27: 52–65.

48.

Rzedowski

(1978) Vegetación de México. México: Ed. Limusa México.

49.

Rzedowski

(1991) Diversidad y orígenes de la flora fanerogámica de México. Acta Botánica 14: 3–21.

50.

Rzedowski

Calderón de Rzedowski

(2013) Datos para apreciación de la flora fanerogámica del bosque tropical caducifolio de México. Acta Botánica Mexicana 102: 1–28.

51.

Serrano

Filonov

Tereshchenki

(2002) Dynamic response to valley breeze circulation in Santa Maria del Oro, a volcanic lake in Mexico. Geophysical Research Letters 29(13): 1649.

52.

Sluyter

Dominguez

(2006) Early maize (Zea mays L.) cultivation in Mexico: Dating sedimentary pollen records and its implications. Proceedings of the National Academy of Sciences 103: 1147–1151.

53.

Sosa-Nájera

Lozano-García

Roy

, et al. 2010. Registro de sequías históricas en el occidente de México con base en el análisis elemental de sedimentos lacustres: El caso del lago de Santa María del Oro. Boletín de la Sociedad Geológica Mexicana 62: 437–451.

54.

Stahle

Cook

Burnette

, et al. (2016) The Mexican drought atlas: Tree-ring reconstructions of the soil moisture balance during the late pre-hispanic, colonial, and modern eras. Quaternary Science Reviews 149: 34–60.

55.

Stan

Sánchez-Azofeifa

(2019) Tropical dry forest diversity, climate response and resilience in a changing climate. Forests 10: 443.

56.

Tellez

(1995) Flora, Vegetación y Fitogeografía de Nayarit. PhD Thesis, Universidad Nacional Autónoma de México, México.

57.

Torres-Rodríguez

Lozano-García

Figueroa-Rangel

, et al. (2012) Cambio ambiental y respuestas de la vegetación de los últimos 17,000 años en el centro de México: el registro del lago de Zirahuén. Revista Mexicana de Ciencias Geológicas 29(3): 764–778.

58.

Torres-Rodríguez

Lozano-García

Roy

, et al. (2015) Last glacial droughts and fire regimes in the central Mexican highlands. Journal of Quaternary Science 30(1): 88–99.

59.

Trejo

Dirzo

(2002) Floristic diversity of Mexican seasonally dry tropical forest. Biodiversity and Conservation 11: 2063–2048.

60.

Vázquez-Castro

Ortega-Guerrero

Rodríguez

(2008) Mineralogía magnética como indicador de sequía en los sedimentos lacustres de los últimos ca. 2600 años de Santa María de Oro, occidente de México. Revista Mexicana de Ciencias Geológicas 25: 21–38.

61.

Vázquez-Castro

Roy

Solís

, et al. (2017) Holocene paleohydrology of the Etzatlán-Magdalena Basin in western-central Mexico and evaluation of main atmospheric forcings. Palaeogeography, Palaeoclimatology, Palaeoecology 487: 149–157.

62.

Williams

(2020) Ancient West Mexico in the Mesoamerican Ecumene, Archaeopress Precolumbian Arachaeology 12. Oxford: Archaeopress Publising Ltd., p. 442.

63.

Zizumbo-Villareal

Colunga-GarcíaMarín

(2010). Origin of agriculture and plant domestication in West Mesoamerica. Genetic Resources and Crop Evolution 57: 813–825.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.23 MB