Novel vegetation and establishment of Chihuahuan Desert communities in response to late Pleistocene moisture availability in the Cuatrociénegas Basin,NE Mexico

Abstract

With over 200 pools, lakes and rivers supporting over 70 species of endemic flora and fauna, the Cuatrociénegas Basin (CCB), Coahuila, NE Mexico, is an extremely important area for conservation studies. However, the palaeoenvironment of this unique area has been relatively neglected. Here, pollen data are presented alongside U-series dating and ¹⁴C AMS dating techniques from a 15-m sediment core taken from Poza Tierra Blanca in the CCB. These data suggest the CCB contains palaeoenvironmental information spanning at least the late Pleistocene (84.5 ka BP) to the present and has undergone extensive environmental change, possibly controlled by stadial–interstadial cycles. The CCB is currently functioning as a hydrologically closed system, established around 4 ka BP synchronously with regional drying of the Chihuahuan Desert. Pollen data suggest similar closed hydrology conditions from ~33 to 23.13 ka BP – before the onset of full glacial conditions at the LGM. Hydrologically open system characteristics with a dominance of wetter, winter monsoon climate punctuate the long-term record. The wetter conditions observed in these units appear to have facilitated the downslope movement of montane taxa and the expansion of wetland taxa. These data illustrate that novel vegetation assemblages are not just products of deglaciation but represent the interaction of the individualistic response of taxa with the unique climate spaces formed by millennial-scale variability during both glacial and interglacial times.

Keywords

Chihuahuan Desert late-glacial Mexico novel communities palaeoecology pollen

Introduction

Evidence for late-glacial woodlands in the North American Deserts has been well documented in the palaeoecological record (Betancourt et al., 1990; Thompson et al., 1993). Packrat (Neotoma sp.) midden data spanning over 40,000 years provide some of the best evidence of a broader distribution for many coniferous taxa away from mountain ranges where they are extant today (Betancourt et al., 1990; Wells and Jorgensen, 1964). During the late-glacial period, Chihuahuan Desert piñon pines (e.g. Pinus remota, Pinus edulis) had a more southerly and lower elevation distribution than at present (Lanner and Van Devender, 1981). Co-mingling with these taxa were other woodland and montane species (e.g. Pinus strobiformis, Pseudotsuga menziesii, Juniperus sp., Quercus sp.) as well as now common desert taxa (Opuntia sp., Celtis reticulata, Ephedra sp., Agave sp., Euphorbia antisyphilitica) (Lanner and Van Devender, 1981; Minckley and Jackson, 2008; Van Devender, 1990; Van Devender and Burgess, 1985). Indeed, glacial woody plant encroachment into the Chihuahuan Desert was likely sourced from multiple directions, from temperate to subtropical regions (Van Devender and Burgess, 1985).

The process driving late-glacial floristic patterns has been attributed to temperature and hydroclimatic differences from present (Metcalfe, 2006; Metcalfe et al., 2002; Van Devender and Burgess, 1985). Estimations of the impact of lowered temperatures on effective moisture were enough to allow for woodland expansion into now desert regions (Brakenridge, 1978). Differences in hydroclimate are an equally probable mechanism for promoting woodland expansion during the late-glacial period. In northern Mexico, many basins record deep water conditions during the last glacial cycle. Metcalfe et al. (2002) interpreted diatom assemblages as indicative of deep lake cycles in the Alta Babícora Basin, Mexico, from ~65¹ to 57 ka and 38 to 29 ka. Between 57 and 38 ka and after 29 ka, lake depths in the Alta Babícora Basin were highly variable, but still suggested greater moisture flow into northern Mexico. Musgrove et al. (2001) noted increased stalagmite growth rates in the Trans-Pecos region, Texas, US, from 71 to 60 ka, 39 to 33 ka and 24 to 12 ka, also suggesting greater moisture flux episodes into the north-central Chihuahuan Desert during those times. Both studies indicate increased regional moisture flux consistent with a southward displaced polar jet-stream, caused by the Laurentide Ice Sheet (Bartlein et al., 1998; Metcalfe et al., 2000), tropical moisture transport into northern Mexico (Holmgren et al., 2007; Lyle et al., 2012) and generally cooler summers lowering net evapotranspiration (Van Devender and Burgess, 1985).

Located within the Chihuahuan Desert, the Cuatrociénegas Basin (CCB) of Coahuila Mexico was long thought to be a centre of environmental stability based on the high endemism of the basin (Abell et al., 2000) and early paleoenvironmental work that suggested floristic stability from ~31.4 ka to the present (Meyer, 1973). The environmental stability of this valley has since been challenged by both packrat midden data and reanalysis of the original pollen record (Minckley and Jackson, 2008; Van Devender and Burgess, 1985). Here, we present a >50,000-year pollen record from the CCB, specifically to address how the composition of glacial woodlands related to the hydroclimate of that time, the role of desert taxa in these woodlands and environmental change prior to the formation of the modern Chihuahuan Desert.

Site description

The CCB (102°W, 26°N; 742 m a.s.l.) is a closed evaporitic basin located in the southeastern Chihuahuan Desert of Coahuila, Mexico (Figure 1). The average rainfall is 260 mm yr⁻¹, falling mostly (73%) in the summer, while evaporative loss exceeds 2000 mm yr⁻¹ (Badino et al., 2004; Lawrimore et al., 2011). The average daily temperatures range from 6.3°C to 24.6°C with peaks up to 40°C in June and July and troughs of 0°C during December and January. Despite the dry climate of the valley, this limestone basin has an extensive system of springs, streams and pools fed largely by regional groundwater systems (Felstead et al., 2015; Johannesson et al., 2004; Wolaver et al., 2008).

Figure 1.

(a) Location of the Cuatrociénegas Basin in NE Mexico. (b) Location of the Cuatrociénegas Basin in Coahuila State. (c) Location of the Poza Tierra Blanca (PTB) sediment core in the Cuatrociénegas Basin indicated by a star. Location of surrounding limestone mountains and major hydrologic features are also shown.

Poza Tierra Blanca (PTB) has a wetted surface area of approximately 50 m² and a maximum depth of 1 m. The open pool is located in the central marsh (ciénega) area of the main CCB through-flow system, with an emergent underground inflow stream entering the pool to the south and an outflow stream exiting to the north, connecting to surface flow throughout the marsh area. Today, lake water δ¹⁸O is −5.68‰ (Felstead et al., 2015), greater than the value for regional precipitation (–8.3‰) (Wassenaar et al., 2009) and discharging regional groundwater (–7‰) (Johannesson et al., 2004). The inflow spring of PTB emerges from underneath the CCB floor suggesting the pool is sourced from a mix of cold karst reservoir water and warmer regional groundwater; however, the exact water source of PTB remains unclear due to complex subterranean flow of the CCB water (Felstead et al., 2015).

Field and laboratory methods

In March 2008, a 15-m-long core was collected next to PTB with an Acker percussion device, which uses a 60-cm-long split-spoon barrel. The sediments were described and subsampled contiguously for dating, pollen, stable isotope and loss on ignition analyses at Centro de Investigación Científica de Cuatrociénegas. Subsamples were then taken to John Moores University, Liverpool, UK, and the University of Wyoming, US, for analysis and archiving.

Chronology was determined using Uranium series (U-series) and AMS radiocarbon dating (Table 1). A total of five samples for U-series dating were prepared from tufas (see core description below). Samples were cut into 1-cm slices for ease of transport, storage and sampling. Material for analysis was removed from several tufa pieces using a stainless steel end cutter. Carbonates were cleaned under a binocular microscope. Samples were broken down into ⩽3mm chunks using a medical scalpel and tweezers and rinsed in distilled water while using a fine-tipped artists brush to remove contamination from any surfaces and accessible pore spaces before being treated in an ultrasound bath for 1 h to remove any inaccessible contamination. AMS radiocarbon dating was conducted on bulk organic matter. A full age model was not calculated because of the complex lithology (peat-tufa-clay-tufa-clay). However, we did apply a linear age model for the late-glacial section described in this paper.

Table 1.

U-series and ¹⁴C AMS dates obtained for core PTB (Felstead, 2012). PTB 317-318, 321-322, 339-341 and 355-356 give an average age of 22,130 ± 880 cal. yr BP for the 317-356 tufa within core PTB which is used as a minimum age for this tufa.

Lab	Lab code	Sample depth (cm)	Material dated	Technique employed	Conventional age (yr BP)	±2σ calibration (yr BP)	Mid Calibrated age (cal. yr BP)
BETA	PTB36	36	Bulk Peat	¹⁴C AMS	70 ± 40	215–267	241 ± 26
BETA	PTB160	160	Bulk Peat	¹⁴C AMS	8440 ± 50	9399–9535	9467 ± 68
NIGL	PTB317-318	317–318	CaCO₃	U-series	22,710 ± 910	21,800–23,620	22,710 ± 910
NIGL	PTB321-322	321–322	CaCO₃	U-series	23,250 ± 930	22,320–24,180	23,250 ± 930
NIGL	PTB339-341	339–341	CaCO₃	U-series	22,150 ±890	21,260–23,040	22,150 ± 890
NIGL	PTB355-356	355–356	CaCO₃	U-series	20,400 ± 820	19,580–21,220	20,400 ± 820
BETA	PTB368	368	Bulk Peat	¹⁴C AMS	23,130 ± 130	27,634–28,469	28,051 ± 417
NIGL	PTB1013-1020	1013–1020	CaCO₃	U-series	56,180 ± 2250	53,930–58,430	56,180 ± 2250

Beta: Beta Analytic, Miami, US. NIGL: NERC Isotope Geosciences Laboratory, Nottingham, UK.

For pollen analysis, sediments went through standard acid–base–acid digestion (Faegri et al., 1989) with exotic Lycopodium tracers added to each sample. Samples were examined at least every metre of core, with more intensive sampling occurring where pollen preservation was best (between 367 and 700 cm). Pollen counts of 300 terrestrial grains (excluding Poaceae and Amaranthaceae per Minckley and Jackson, 2008) or 1000 tracers were conducted for each sample. Asteraceae counts were included in both upland (terrestrial) and wetland (local) sums, as this pollen type may be a significant contributor to the pollen spectra based on either an upland or wetland assumption. Terrestrial and local pollen percentages were only calculated for samples whose pollen sum exceeded 100 grains, respectively; otherwise data were presented as presence/absence (Minckley et al., 2011).

Results

The use of a percussion coring device resulted in sediment loss and sediment resampling, making complete recovery of the section impossible. Sediment resampling was caused by sediment drawdown between drives and the lack of a piston in the end of the barrel, resulting in some liquefaction on the top of each section of core. Unfit sediment determination was made in the field and that material was disposed of prior to wrapping the remaining core material for latter subsampling. The resulting recovery was ~60% of a 1500-cm-long section.

The full Core PTB stratigraphy is shown in Figure 2. Below 1020 to the end of pollen sampling at 1420 cm, sediments were marls with dark organic banding that showed evidence of soft-sediment deformation. Tufa was present from 1013 to 1020 cm that contained abundant mollusc shells. Sediments from 360 and 1013 cm were marl with some organic banding and precipitate inclusions. A 43-cm-thick tufa occurred between 317 and 360 cm. The uppermost sediments (0–317 cm) consisted of inter-bedded carbonate marls and peats (Figure 2).

Figure 2.

Core PTB stratigraphy with sediment descriptions and structures. Photos of Nymphophilus minckleyi are taken from www.desertfishes.org.

Sediment organic matter consisted of well-humidified peat indicating dry periods and low lake levels, when expansion of terrestrial wetland was enhanced, similar to today in the CCB (Badino et al., 2004; Felstead et al., 2015). In contrast, deposition of carbonate marl or tufa indicates wet periods and high lake levels, when catchment run-off and groundwater discharge, through increased hydrostatic pressure, would be enhanced. Mollusc shells present in both tufa and carbonate marl sediments were of the species Nymphophilus minckleyi, which is endemic to the CCB (Hershler, 1984).

Age control

Table 1 shows the results of AMS and U-series dates on organic and tufaceous materials (Felstead, 2012). Thin section analysis of the core PTB tufas indicated very little organic contamination or secondary calcite precipitation before cleaning and preparation. Organic detritus was removed during the ultrasonic disintegration technique and possible secondary calcite precipitation was not indicated by detrital thorium ratios. All U-series samples obtained 1.5–1.9 ppm U with ²³⁰Th/²³²Th activity between 77 and 594 (reliable ²³⁰Th/²³²Th is accepted to be >20) and are displayed in Table 1 alongside ¹⁴C AMS dates (from peats). All samples for ¹⁴C AMS analysis contained sufficient uncontaminated material to produce reliable dates.

In total, eight dates were obtained for the sediment core (Table 1). All dates are in good chronological order. An average age of 22.1 ka BP is obtained for the four U-series dates, giving a reasonable paired date with the underlying ¹⁴C AMS date of 28 ka BP obtained from bulk peat. The average age of 22.1 ka BP for the 317-357 tufa in the core, combined with the four additional dates, provides an overall chronology to the sedimentary sequence. Linear interpolation of these dates suggests an average accumulation rate of 64 yr/cm. The section between 368 and 1020 cm spans ~53.8–23.1 ka with a sedimentation rate of 50 yr/cm. Based on sedimentological similarity of the materials below the second tufa to the materials in the above section, we extrapolate a bottom age for the core between 72 and 83 ka. For graphical and discussion purposes, we use the older age for material below 1020 cm (>56.2 ka), but acknowledge that these data are unconstrained.

The core PTB chronology is shown in Figure 3 and is based on cm-scale age/depth points generated from the three ¹⁴C AMS and two U-series dates (the coeval date of 22,130 ± 880 cal. yr BP was used for the upper U-series date) from core PTB. All radiocarbon dates are presented as calibrated ages using IntCal13 and CALIB 7.1 (Reimer et al., 2013; Stuiver et al., 2017) and radiocarbon and U-Series dates were imported into CLAM 2.2 (Blaauw, 2010) for age modelling. The age model was based on a smooth spline regression that includes the underlying non-Gaussian uncertainty in the calibrated age probability distribution. Our age model and interpretation of the oldest sediments are conservative because of the lack of temporal constraint for sediments below the second tufa. However, CLAM produces multiple point age estimates through the model, so the spline curve applied through unknown points will provide a reliable basal age approximation (Blaauw, 2010). Interpolation of the spline regression gives core PTB a maximum basal age of 84,900 ± 8500 cal. yr BP; however, the authors acknowledge the uncertainty with this and all other estimated dates through core PTB as the age model created using CLAM is in a large part driven by the uneven interval of the ¹⁴C AMS and U-series dating points. The error bound ages provide a reliable age estimate based on the weighted means of point age estimates through the model.

Figure 3.

Calibrated ¹⁴C AMS and U-series dates versus depth using CLAM (Blaauw, 2010). Black line represents best mid-calibrated age and the grey shaded areas represent a 95% confidence interval range.

Pollen

Sedimentary units were determined through core description (Felstead, 2012) and pollen preservation. Below 694 cm depth (units 1–3, >37.4 ka) total pollen preservation was poor (Figure 4). Cupressaceae, Pinus and Quercus were the most common woodland taxa in this section of the record. However, Picea, Betula, Carya and Fraxinus were occasional, particularly in sediments above 1000 cm depth. Pollen of common woody desert taxa was not observed in this section. Artemisia and Asteraceae pollen types were the most common shrub and upland forbs. Most wetland taxa were present throughout the lowermost section.

Figure 4.

(a) Terrestrial and (b) wetland pollen percentage data from Core PTB. White secondary plots indicate 5× exaggerations for taxa with low abundances. Plus (+) symbols indicate presence/absence data of pollen types from counts that did not exceed 100 pollen grains. Unit determinations are based on sedimentary and isotopic compositional changes suggesting hydrological open or closed system conditions.

Between 694 and 570 cm (unit 4, 37.4–31.1 ka), woodlands were characterized by Cupressaceae (5–23%), Picea (1–4.5%) and Pinus (30–70%) pollen (Figure 4). Traces of Abies pollen were observed. Other significant arboreal taxa include Acer (0–3%), Carya (<1%) and Fraxinus (0–2%). Common desert taxa Prosopis (0–3%) and Quercus (3–9%) were present in this section. Upland shrubs and forbs were principally Artemisia (2–7%), Fabaceae (1–2%) and Asteraceae (8–22%). Wetlands had their highest Poaceae abundances (25–42%), with the co-dominants Amaranthaceae (22–45%) and Asteraceae (16–22%) contributing much of the remaining lowland pollen. Ambrosia (1–3%) and Cyperaceae (5–15%) were also present.

From 570 to 372 cm (unit 5, 31.1–28 ka), woodland taxa were reduced (Figure 4). Cupressaceae (4–21%) and Pinus (21–65%) pollen percentages were lower than previous. Picea (1–2%) was not observed in sediments above 512 cm depth, and Abies was absent from this interval. Other arboreal taxa were largely absent, with the exception of Quercus (6–12%). Of the shrubs and forbs, Artemisia (2–13%) and Asteraceae (11–40%) were the principal significant pollen types. Wetland pollen data extended further upcore to 367 cm depth. Amaranthaceae (22–91%) replaced Poaceae (5–35%) as the dominant pollen type. Asteraceae (2–40%) percentages were also generally lower. Ambrosia (1–3%) remained constant, while Cyperaceae (2–40%) percentages increased in this interval.

Between 372 and 200 cm (unit 6, 28–12.1 ka), pollen was rare or not observed. From 200 to 0 cm (unit 7, 12.1 ka to present), woodlands were represented by Cupressaceae (2–16%) and Pinus (13–41%) pollen (Figure 4). Deciduous trees were absent, while desert taxa Prosopis (<1%), Acacia (<1%) and Quercus (1–10%) remained in the record.

Discussion

The pollen data reveal the complex palaeoenvironmental history of the CCB. Preservation of pollen during inferred dry periods was poor, but those taxa that were identified indicate a persistence of woodland taxa prior to 59 ka. During wet intervals, pine–juniper–oak woodlands dominated, but these woodlands also contained relatively abundant maple, birch, ash, hickory, mesquite and acacia species. Determining causal relationships of vegetation change to hydroclimate is not possible because the pollen record is largely silent during periods of inferred aridity.

A long-standing question about late-glacial vegetation difference-from-present in the North American Deserts has been what role lower temperatures had on changing effective moisture within the region (Brakenridge, 1978). Indeed, lower temperatures, southerly displacement of the polar jet-stream, increased cloudiness, reduced evaporation and recirculation of terrestrial moisture all likely promoted a greater spatial extent of woodlands throughout these now desert regions (Bartlein et al., 1998; Braconnot et al., 2007; Brakenridge, 1978; Holmgren et al., 2007; Hostetler and Bartlein, 1990; Hostetler et al., 1994; Minckley et al., 2004; Thompson et al., 1993). The case for a temperature control on woodland development in northern Mexico is less clear, as subtropical circulation patterns might have allowed for greater advection of moisture from the Gulf of Mexico (Hostetler and Bartlein, 1999) and greater westerly subtropical flow bringing moisture in from the Pacific Ocean (Metcalfe et al., 2000, 2015). Model simulations for the ‘Last Glacial Maximum’ indicate northern Mexico was cooler- and drier-than-present (Braconnot et al., 2007). However, that interpretation is largely based on inference given the paucity of data for the region (Farrera et al., 1999).

Northern Mexico and the western US regions experienced cooler, wetter stadial periods during the last glacial period (Metcalfe et al., 2015). Records of the Sierra Nevada (Bischoff and Cummins, 2001), Edwards Plateau (Bryant and Holloway, 1985; Musgrove et al., 2001) and Alti Babícora Basin (Metcalfe et al., 1997, 2002) all indicate greater westerly subtropical flow bringing greater moisture flux from the Pacific Ocean. These cooler, wetter conditions in northern Mexico, including the CCB, would have allowed for the expansion of mesic and montane vegetation species observed in the pollen record as well as increasing groundwater recharge from regional aquifers, bringing greater moisture flux through the CCB. During stadial periods, increased rainout from Pacific-sourced moisture and North American Monsoon (NAM) may have occurred due to the presence of the mountains, recharging regional aquifers and subsequently increasing groundwater flow into the CCB. Coupled with increased groundwater flow, an increased moisture body from the west may have also provided rainout in the CCB itself, increasing run-off and local precipitation on the main recharge zone of the Sierra San Marcos y Pinos. Data from the CCB suggest greater changes in millennial-scale effective moisture in northern Mexico during the last glacial period than model estimations (Braconnot et al., 2007) might suggest. The mountainous topography of the Sierra’s Madre Occidental and Madre Oriental surrounding the CCB and large grid scale of models used for millennial-scale moisture reconstructions often result in low level flows from the Pacific and Gulf of Mexico being underestimated, therefore underrepresenting the influence, and strengthening, of NAM during the early-mid Holocene over the north of Mexico (Metcalfe et al., 2015).

>56.3 ka

Sediment lithology throughout this period is dominated by organic banded marl suggesting PTB experienced rapid wetting and drying events indicative of a wetland system, highly sensitive to moisture flux and hydrologic change (Minckley et al., 2009; Pigati et al., 2009). Continued presence of desert woodland (Pinus, Quercus) and possibly eastern deciduous forest (Fraxinus) indicates conditions in the CCB became increasingly wet after this time. Indeed, records for the Trans-Pecos and Chihuahuan desert regions indicate substantially wetter conditions with open oak and pine woodland (Metcalfe et al., 2002; Musgrove et al., 2001), around the time of the maximum extent of Mono Lake stadial (Ehlers and Gibbard, 2004).

~56.3–24 ka

Vegetation data indicate the CCB to be a hydrologically open basin as the time intervals 56.18–53.8 ka and 23.1–11.1 ka may reflect simply greater water flow through the spring complex, rather representing the outflow from the valley based on the lack of independent evidence (i.e. pollen data). However, pollen data for the period 39.2–33 ka (unit 4) suggest a flora indicative of greater moisture availability in the valley. The median ratio of Poaceae (mesic) to Amaranthaceae (xeric) pollen abundance was 0.95, indicating less halophytic taxa on the basin floor during this time (Minckley and Jackson, 2008).

Periodic desiccation and deflation, and decreased stalagmite growth are reported in Chihuahua between 54.6 and 38.5 ka (Metcalfe et al., 2002; Musgrove et al., 2001). Conditions during this period are inferred to be similar to those prior to ~56 ka, although perhaps slightly drier with punctuated periods of wetter conditions. The presence of key mesic vegetation taxa Picea, Fraxinus and Betula at ~49, ~44 and ~43 ka, respectively, suggests these three points to be associated with greater groundwater recharge within the CCB and coinciding with periods of deep water in Chihuahua (Metcalfe et al., 2002). Increases of wetland Aster, Amaranthaceae and Poaceae also suggest these three points to be much wetter, despite the period as a whole displaying generally much drier conditions. Complex interplay between the circulation patterns of the Pacific and Gulf of Mexico due to climate forcing (Bernal et al., 2011; Musgrove et al., 2001) could be important, as shifts between summer and winter atmospheric patterns change source waters as well as change lake levels, which could lead to periods of eutrophication and desiccation.

Upland taxa observed between 39.2 and 33 ka (unit 4) suggest an admixture of western montane forest (Abies, Picea), eastern deciduous forest (Acer, Betula, Carya, Fraxinus) and warm desert woodland taxa (Pinus, Prosopis, Quercus, Cupressaceae) (Figure 4). The overlap in modern climate space for these genera (Thompson et al. (1999) indicates moderate temperature extremes (4.1–24.4°C), annual precipitation between 410 and 1755 mm, and a moisture index (actual evapotranspiration/potential evapotranspiration) ranging between 0.77 and 0.94. The genera constraining these climatic boundaries are Carya and Prosopis.

The median ratio of Poaceae to Amaranthaceae pollen abundance was 0.95, indicating less halophytic taxa on the basin floor during this time (Minckley and Jackson, 2008). Marginal aquatic vegetation (Typha) and upland mesic vegetation appear for the first time in the pollen record (Figure 4). Of the aquatic vegetation in unit 4, Typha is present for the first time in core PTB along with higher numbers of Poaceae and Amaranthaceae, suggesting that this period was much wetter than the previous hydrologically open period ~54.3–53.3 ka.

The mixed woodland of unit 4 persisted into the first half of unit 5 as CCB returned to a hydrologically closed system. The median ratio of Poaceae to Amaranthaceae pollen abundance was 0.53, indicating increases in halophytic taxa on the basin floor, suggestive of drying (Minckley and Jackson, 2008). Western montane taxa, present early, were absent by 23.1 ka (Figure 4). With the exception of Betula, eastern deciduous taxa had disappeared by the top of this unit, while desert woodland taxa persisted. Climate estimations based on the modern climate space for the remaining genera indicate increases in potential temperature ranges (4.1–30.8°C), a greater potential annual precipitation range (410 and 2555 mm) and a greater moisture index range (0.42 and 0.94). The genera constraining the climatic boundaries were Betula, Prosopis and Quercus.

~24 ka to present

Mesic woodlands may have reformed during the full glacial period (unit 6, 23.1–11.1 ka), based on the original pollen counts by Meyer (1973) and a single pollen sample from a packrat midden dated to ~14 ka (Minckley and Jackson, 2008). These data provide evidence of Abies and Picea presence near the basin floor, but no indication of the eastern deciduous forest constituents observed in unit 4, indicating regional drying and greater influence of local moisture flux at these times.

Upland vegetation in unit 7 (11.1 ka to present) is strikingly similar to unit 5 despite the local vegetation indicating greater mesic conditions. The median ratio of Poaceae to Amaranthaceae pollen abundance was 0.71, indicating decreases in halophytic taxa on the basin floor and suggestive of wetter-than-previous conditions. Upland changes suggest greater diversity of desert taxa indicative of greater aridity than that indicated by the pollen composition of unit 5. Additions to the pollen flora include Acacia, Celtis, Fouquieria and Agavaceae, all supporting the interpretation of modern floristic composition establishment over the past 11 ka (Anderson and Vandevender, 1995; Bryant, 1977; Bryant and Holloway, 1985; Elias and Vandevender, 1990; Huang et al., 2001; Metcalfe et al., 2002; Minckley and Jackson, 2008; Musgrove et al., 2001; Van Devender and Burgess, 1985). The establishment of modern floristic composition is roughly synchronous with the earliest evidence of human occupation in the CCB at 10.5 ka (Felstead et al., 2014), suggesting this floristic composition was favourable for these early settlers, known as the Coahuiltecans. Indeed, the appearance of Agavaceae, used to construct textiles, sandals and fishing nets (Taylor, 1956, 1966, 2003) at 11 ka in this pollen record, may have ushered in the beginning of the common cultural tradition that spanned the Holocene period in the CCB (Browman, 2003; Felstead et al., 2014). Further to this, the appearance of Opuntia, Salix and Carya in the CCB pollen record at ~8 ka (Felstead et al., 2014), though in small numbers, suggests that the floristic composition became slightly more temperate in the early-mid Holocene, allowing early humans to become established in the basin before the onset of more arid conditions at ~7 ka.

While temperature changes may have been enough to change effective moisture for terrestrial vegetation during the late-glacial period (Brakenridge, 1978), the temporal coincidence of deep water conditions in the Alta Babícora Basin and increased stalagmite growth rates in the Trans-Pecos region, Texas provides strong evidence for greater moisture flux into the Chihuahuan Desert (Metcalfe et al., 2002; Musgrove et al., 2001). Given the distance from the Laurentide Ice Sheet, circulation causing moisture to be advected from the Gulf of Mexico may have been a controlling factor on the hydroclimate of the region. Greater moisture availability allowed for novel, or non-analogue, vegetation assemblages to form prior to the full glacial period – ca. 21 ka (Jackson and Williams, 2004; Williams et al., 2007; Williams and Jackson, 2007). The novel assemblage (Figure 4) of western montane forest, eastern deciduous forest and warm desert woodland taxa has not been observed in paleoenvironmental records. The admixture of taxa in CCB suggests biogeographic range shifts for eastern deciduous forest taxa that have not been previously observed in other records. While we only see this unique assemblage occurring between 39.2 and 33 ka, weak evidence allows us to infer that these vegetation associations may have formed when CCB was hydrologically open, with the last occurrence possibly during the last glacial maximum − 21 ka (Meyer, 1973; Minckley and Jackson, 2008).

The time period covered by this study, based on the maximum relative age of 84.5 ka, encompasses the last glacial period, ca. 85–11 ka BP, almost in its entirety. Stable isotope (Felstead, 2012) and pollen data in units 2, 4 and 6 suggest that these units were hydrologically open and climatically wetter with a dominance of winter precipitation. Wet periods, particularly units 2 and 4, coincide with the Tahoe and Tenaya stadial periods of the last glacial period, before the maximum extension of the Laurentide Ice Sheet occurred (Tioga stadial) during unit 6 (Ehlers and Gibbard, 2004).

Summary of the CCB palaeoenvironment

The palaeoenvironmental record from CCB suggests episodic extra-local connectivity of the basin to other watersheds, possibly 30% of the time. Rather than being a model of persistent isolation, the isotopic (Felstead, 2012) and pollen data suggest CCB had external hydrologic connections, which would allow for species to migrate into and out of the basin over centuries to millennia before climatological conditions would force isolation similar to present-day conditions. Over the past 60,000 years, CCB may have had external aquatic and riparian connections over a cumulative 18,000 years where species and genetic exchanges would have occurred. Periods of connectivity lasted millennia. However, more common were periods of hydrologic isolation that would have allowed evolutionary pressures to act on biogeographically separated populations within the CCB. This ebb and flow of connectivity may suggest a mechanism for promoting the high biodiversity that characterizes the region today.

The downslope movement of western montane taxa has been well established throughout the American Deserts through packrat midden analysis (Betancourt et al., 1990). The distribution of desert taxa has been less constrained with these data. However, the climate space of the woody genera identified in our study suggests that similar to today, freezing temperatures were not common in CCB during the past 50–83 ka, potentially removing that environmental constraint on the distribution of desert taxa. Increased moisture, on the other hand, may have been a greater determinant for the presence of some taxa, particularly succulents in the region. Unfortunately, these taxa are largely silent in our pollen data. The novel pollen assemblages seen in the CCB record illustrate that no-analogue vegetation assemblages are not just products of deglaciation. Rather, the strength of individualistic responses of taxa to millennial-scale climate variability allows for unique plant associations during both glacial and interglacial times. In CCB, pollen assemblages indicate vegetation admixtures of taxa presently distributed in the American deserts, Mexican subtropics, western montane and eastern deciduous forest regions.

While temperature changes may have been enough to change effective moisture for terrestrial vegetation during the late-glacial period (Brakenridge, 1978), the temporal coincidence of deep water conditions in the Alta Babícora Basin and increased stalagmite growth rates in the Trans-Pecos region, Texas, provides strong evidence for greater moisture flux into the Chihuahuan Desert (Metcalfe et al., 2002; Musgrove et al., 2001). Given the distance of the CCB from the Laurentide Ice Sheet, the relative contributions of displaced polar jet-streams, westerly storm tracks, ITCZ and Pacific subtropical air masses to seasonal atmospheric circulation (Bernal et al., 2011; Lyle et al., 2012; Metcalfe et al., 2000; Wagner et al., 2010) are unclear. What is more clear is that PTB (and more widely CCB) spring discharge appears closely linked with moisture variations to the west (Bischoff and Cummins, 2001; Ehlers and Gibbard, 2004; Metcalfe et al., 2002; Ortega-Guerrero et al., 1999), which suggests, given that the regional aquifer system recharges in the Sierra Madre Occidental and Bolson de Mapimi (Wolaver et al., 2008), stadial/interstadial controls on the CCB hydroclimate. Greater moisture availability allowed for novel, or non-analogue, vegetation assemblages to form prior to the full glacial period − ca. 21 ka (Jackson and Williams, 2004; Williams et al., 2007; Williams and Jackson, 2007). This novel assemblage of Mexican subtropical forest, western montane forest, eastern deciduous forest and warm desert woodland taxa has not been observed in palaeoenvironmental records and suggests a large biogeographic range shift for many taxa (Figure 4). While we only see this unique assemblage occurring between 39.2 and 33 ka, weak evidence allows us to infer that these vegetation associations may have formed when CCB was hydrologically open, with the last occurrence possibly during the last glacial maximum − 21 ka (Meyer, 1973; Minckley and Jackson, 2008).

The establishment of modern floristic composition in the early-Holocene period provided favourable conditions for nomadic humans to settle in the CCB at ~10.5 ka (Felstead et al., 2014). The importance of cooler, wetter conditions persisting up to ~7 ka, allowing the development of the desert wetland ecosystem, cannot be underestimated in terms of human settlement in the north of Mexico. Indeed, desert wetland habitats are fragile, but important, locations of early human settlement (Nicholas, 1998; Springer et al., 2015). The novel vegetation assemblages observed in this record from 11 ka onwards may have been the ‘spark’ necessary for the establishment of the common cultural tradition in the CCB, and wider Trans-Pecos region, providing the desert wetland habitat conducive to tethered nomadism (Taylor, 1964).

Conclusion

Desert wetlands provide a refugia for both preserving evolutionary lineages and promoting diversification of species (Murphy et al., 2015). Desert wetlands have been identified as biodiversity hotspots, often with associated endemism that argues for their continued conservation and protection (Hendrickson and Minckley, 1985; Minckley, 1969, 1992; Minckley et al., 2013a, 2013b). Desert wetlands also represent critical environments used to settle the Americas (Felstead et al., 2014), providing water, food and fibres (Taylor, 1956, 1966, 2003). As rare resources within arid landscapes, studies into these environments can serve as focal points for transdisciplinary collaborations.

Increased moisture flux into the Chihuahuan Desert region of North America allowed for a novel mixture of taxa during the last glacial period. Changes in effective moisture allowed for significant biogeographic shifts in the terrestrial vegetation as well, allowing early humans to settle in the CCB. Biogeographic changes suggest not only the downslope migration of conifers from the western montane forests but also a southwestward migration of eastern deciduous taxa. Characteristic desert taxa persisted and mixed with these more mesic elements during the last glacial period. These data illustrate that novel vegetation assemblages are not just products of deglaciation but represent the interaction of the individualistic response of taxa to the unique climate spaces formed by millennial-scale variability during both glacial and interglacial times.

Footnotes

Acknowledgements

The authors thank the land managers and property owners within the Área de Protección de Flora y Fauna de Cuatrociénegas (APFFCC) and also its director Ivo Garcia Gutierrez for allowing research access to the APFFCC. Assistance in the field was provided by David Huddart, Jason Kirby, Sarah Metcalfe and Charles Minckley. Three anonymous reviewers provided valuable feedback, and their efforts are appreciated.

Funding

This work was funded by the Natural Environment Research Council (NERC) in the UK awarded to SG, with a PhD studentship to NJF (Project NE/F006772/1). Funding for TAM was provided through NSF (Award No. 1125532/1125549).

Notes

References

Abell

Olson

Dinerstein

et al . (2000) Freshwater Ecoregions of North America: A Conservation Assessment. Washington, DC: Island Press.

Anderson

Vandevender

(1995) Vegetation history and paleoclimates of the coastal lowlands of Sonora, Mexico – Pollen records from pack-rat middens. Journal of Arid Environments 30: 295–306.

Badino

Bernabei

De Vivo

et al . (2004) Under the Desert: The Mysterious Waters of Cuatro Ciénegas. Treviso: Associazione Geografica La Venta.

Bartlein

Anderson

et al . (1998) Paleoclimate simulations for North America over the past 21,000 years: Features of the simulated climate and comparisons with paleoenvironmental data. Quaternary Science Reviews 17: 549–585.

Bernal

Lachniet

McCulloch

et al . (2011) A speleothem record of Holocene climate variability from southwestern Mexico. Quaternary Research 75: 104–113.

Betancourt

VanDevender

Martin

(1990) Packrat Middens: The Last 40,000 Years of Biotic Change. Tucson, AZ: The University of Arizona Press.

Bischoff

Cummins

(2001) Wisconsin glaciation of the Sierra Nevada (79,000–15,000 yr BP) as recorded by rock flour in sediments of Owens Lake, California. Quaternary Research 55: 14–24.

Blaauw

(2010) Methods and code for ‘classical’ age-modelling of radiocarbon sequences. Quaternary Geochronology 5: 512–518.

Braconnot

Otto-Bliesner

Harrison

et al . (2007) Results of PMIP2 coupled simulations of the Mid-Holocene and last glacial maximum, part 1: Experiments and large-scale features. Climate of the Past 3: 261–277.

10.

Brakenridge

(1978) Evidence for a cold, dry full-glacial climate in the American Southwest. Quaternary Research 9: 22–40.

11.

Browman

(2003) The archaeology of drylands: Living at the margin. American Anthropologist 105: 179–180.

12.

Bryant

Jr (1977) A 16,000 year pollen record of vegetational change in central Texas. Journal of Palynology 1: 143–156.

13.

Bryant

Jr Holloway

(1985) A late-Quaternary paleoenvironmental record of Texas: An overview of the pollen evidence. In: Bryant

Holloway

(eds) Pollen Record of Late Quaternary North American Sediments. Calgary: American Association of Stratigraphic Palynologists, pp. 39–70.

14.

Ehlers

Gibbard

(2004) Extent and Chronology of Glaciation. Amsterdam: Elsevier Science.

15.

Elias

Vandevender

(1990) Fossil insect evidence for late Quaternary climatic-change in the Big Bend region, Chihuahuan Desert, Texas. Quaternary Research 34: 249–261.

16.

Faegri

Kaland

Kzywinski

(1989) Textbook of Pollen Analysis. New York: John Wiley.

17.

Farrera

Harrison

Prentice

et al . (1999) Tropical climates at the Last Glacial Maximum: A new synthesis of terrestrial palaeoclimate data. I. Vegetation, lake levels and geochemistry. Climate Dynamics 15: 823–856.

18.

Felstead

(2012) Palaeoenvironmental reconstruction and geoarchaeology of the Cuatro Ciénegas Basin, NE Mexico, from the late Pleistocene to the present. Liverpool John Moores University. Available at: http://researchonline.ljmu.ac.uk/6171/1/572046.pdf.

19.

Felstead

Gonzalez

Huddart

et al . (2014) Holocene-aged human footprints from the Cuatrociénegas Basin, NE Mexico. Journal of Archaeological Science 42: 250–259.

20.

Felstead

Leng

Metcalfe

et al . (2015) Understanding the hydrogeology and surface flow in the Cuatrociénegas Basin (NE Mexico) using stable isotopes. Journal of Arid Environments 121: 15–23.

21.

Hendrickson

Minckley

(1985) Ciénegas-vanishing climax communities of the American Southwest. Desert Plants 6: 130–176.

22.

Hershler

(1984) The hydrobiid snails (Gastropoda: Rissoacea) of the Cuatro Cienegas Basin: systematic relationships and ecology of a unique fauna. Journal of the Arizona-Nevada Academy of Science 19: 61–76.

23.

Holmgren

Norris

Betancourt

(2007) Inferences about winter temperatures and summer rains from the late Quaternary record of C₄ perennial grasses and C₃ desert shrubs in the northern Chihuahuan Desert. Journal of Quaternary Science 22: 141–161.

24.

Hostetler

Bartlein

(1990) Simulation of lake evaporation with application to modeling lake level variations of Harney-Malheur Lake, Oregon. Water Resources Research 26: 2603–2612.

25.

Hostetler

Bartlein

(1999) Simulation of the potential responses of regional climate and surface processes in western North America to a canonical Heinrich Event. In: Clark

Webb

Keigwin

(eds) Mechanisms of Global Climate Change at Millennial Time Scales. Washington, DC: American Geophysical Union, pp. 313–327.

26.

Hostetler

Giorgi

Bates

et al . (1994) Lake-atmosphere feedbacks associated with paleolakes bonneville and lahontan. Science 263: 665–668.

27.

Huang

Street-Perrott

Metcalfe

et al . (2001) Climate change as the dominant control on glacial-interglacial variations in C-3 and C-4 plant abundance. Science 293: 1647–1651.

28.

Jackson

Williams

(2004) Modern analogs in Quaternary paleoecology: Here today, gone yesterday, gone tomorrow? Annual Review of Earth and Planetary Sciences 32: 495–537.

29.

Johannesson

Cortes

Kilroy

(2004) Reconnaissance isotopic and hydrochemical study of Cuatro Cienegas groundwater, Coahuila, Mexico. Journal of South American Earth Sciences 17: 171–180.

30.

Lanner

Van Devender

(1981) Late Pleistocene pinon pines in the Chihuahuan Desert. Quaternary Research 15: 278–290.

31.

Lawrimore

Menne

Gleason

et al . (2011) An overview of the Global Historical Climatology Network monthly mean temperature data set, version 3. Journal of Geophysical Research: Atmospheres 116: 1984–2012.

32.

Lyle

Heusser

Ravelo

et al . (2012) Out of the tropics: The Pacific, Great Basin lakes, and late Pleistocene water cycle in the western United States. Science 337: 1629–1633.

33.

Metcalfe

(2006) Late Quaternary environments of the northern deserts and central transvolcanic belt of Mexico. Annals of the Missouri Botanical Garden 93: 258–273.

34.

Metcalfe

Barron

Davies

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

35.

Metcalfe

Bimpson

Courtice

et al . (1997) Climate change at the monsoon/Westerly boundary in Northern Mexico. Journal of Paleolimnology 17: 155–171.

36.

Metcalfe

O’Hara

Caballero

et al . (2000) Records of Late Pleistocene-Holocene climatic change in Mexico – A review. Quaternary Science Reviews 19: 699–721.

37.

Metcalfe

Say

Black

et al . (2002) Wet conditions during the last glaciation in the Chihuahuan Desert, Alta Babicora basin, Mexico. Quaternary Research 57: 91–101.

38.

Meyer

(1973) Late-quaternary paleoecology of Cuatro Ciénegas Basin, Coahuila, Mexico. Ecology 54: 983–995.

39.

Minckley

Jackson

(2008) Ecological stability in a changing world? Reassessment of the palaeoenvironmental history of Cuatrocienegas, Mexico. Journal of Biogeography 35: 188–190.

40.

Minckley

Bartlein

Shinker

(2004) Paleoecological response to climate change in the Great Basin since the Last Glacial Maximum. In: Jenkins

Connolly

Aikens

(eds) Early and Middle Holocene Archeology of the Northern Great Basin. Eugene, OR: Department of Anthropology, Museum of Natural History, pp. 21–30.

41.

Minckley

Brunelle

Blissett

(2011) Holocene sedimentary and environmental history of an in-channel wetland along the ecotone of the Sonoran and Chihuahuan Desert grasslands. Quaternary International 235: 40–47.

42.

Minckley

Brunelle

Turner

(2013b) A paleoenvironmental framework for understanding the development, stability and state-changes of ciénegas in the American Deserts. In: Gottfried

Ffolliott

Gebow

et al . (eds) Merging Science and Management in a Rapidly Changing World: Biodiversity and Management of the Madrean archipelago III, Tucson, AZ, 1–5 May 2012. Fort Collins, CO: Rocky Mountain Research Station, U.S. Forest Service, U.S. Department of Agriculture, pp. 77–83.

43.

Minckley

Turner

Weinstein

(2013a) The relevance of wetland conservation in arid regions: A re-examination of vanishing communities in the American Southwest. Journal of Arid Environments 88: 213–221.

44.

Minckley

Clementz

Brunelle

et al . (2009) Isotopic analysis of wetland development in the American Southwest. The Holocene 19: 737–744.

45.

Minckley

(1969) Environments of the Bolsón of Cuatro Ciénegas, Coahuila, México. El Paso, TX: Texas Western Press.

46.

Minckley

(1992) Three decades near Cuatro Ciénegas, México: Photographic documentation and a Plea for area conservation. Journal of the Arizona–Nevada Academy of Science 26: 89–118.

47.

Murphy

Guzik

Cooper

SJB

et al . (2015) Desert spring refugia: Museums of diversity or evolutionary cradles? Zoologica Scripta 44: 693–701.

48.

Musgrove

Banner

Mack

et al . (2001) Geochronology of late Pleistocene to Holocene speleothems from central Texas: Implications for regional paleoclimate. Geological Society of America Bulletin 113: 1532–1543.

49.

Nicholas

(1998) Wetlands and hunter-gatherers: A global perspective. Current Anthropology 39: 720–731.

50.

Ortega-Guerrero

Caballero Miranda

Lozano Garcia

et al . (1999) Palaeoenvironmental record of the last 70,000 yr in San Felipe Basin, Sonora desert, Mexico: Preliminary results. Geofísica Internacional 38: 153–163.

51.

Pigati

Bright

Shanahan

et al . (2009) Late Pleistocene paleohydrology near the boundary of the Sonoran and Chihuahuan Deserts, southeastern Arizona, USA. Quaternary Science Reviews 28: 286–300.

52.

Reimer

Bard

Bayliss

et al . (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55: 1869–1887.

53.

Springer

Manker

Pigati

(2015) Dynamic response of desert wetlands to abrupt climate change. Proceedings of the National Academy of Science 112: 14522–14526.

54.

Stuiver

Reimer

(2017) CALIB 7.1 [WWW program]. Available at: http://calib.org/calib/.

55.

Taylor

(1956) Some implications of the Carbon-14 dates from a cave in Coahuila, Mexico. Bulletin of the Texas Archaeological Society 27: 215–234.

56.

Taylor

(1964) Tethered nomadism and water territoriality: An hypothesis. Actas y Memorias del XXXV International Congress of Americanists. Mexico City: Instituto Nacional de Antropologia e Historia, pp. 197–203.

57.

Taylor

(1966) Archaic cultures adjacent to the northeastern frontiers of Mesoamerica. In: Wauchope

Ekholm

Willey

(eds) Handbook of Middle American Indians, vol. 4. Austin, TX: University of Texas Press, pp. 59–94.

58.

Taylor

(2003) Sandals from Coahuila Caves (Studies in Pre-Columbian Art and Archaeology Book 35). Washington, DC: Dumbarton Oaks Research Library and Collection.

59.

Thompson

Anderson

Bartlein

(1999) Atlas of Relations between Climatic Parameters and Distributions of Important Trees and Shrubs in North America. Denver, CO: US Geological Survey.

60.

Thompson

Whitlock

Bartlein

et al . (1993) Climatic changes in the western United States since 18,000 yr BP. In: Wright

Jr. Kutzbach

Webb

et al . (eds) Global Climates since the Last Glacial Maximum. Minneapolis, MN: University of Minnesota Press, pp. 468–513.

61.

Van Devender

(1990) Late Quaternary vegetation and climate of the Chihuahuan Desert, United States and Mexico. In: Betancourt

Van Devender

Martin

(eds) Packrat Middens: The Last 40000, Years of Biotic Change. Tucson, AZ: The University of Arizona Press, pp. 104–133.

62.

Van Devender

Burgess

(1985) Late Pleistocene woodlands in the Bloson de Mapimi – A refugium for the Chihuahuan Desert biota. Quaternary Research 24: 346–353.

63.

Wagner

Cole

Beck

et al . (2010) Moisture variability in the southwestern United States linked to abrupt glacial climate change. Nature Geoscience 3: 110.

64.

Wassenaar

Van Wilgenburg

Larson

et al . (2009) A groundwater isoscape (δD, δ18O) for Mexico. Journal of Geochemical Exploration 102: 123–136.

65.

Wells

Jorgensen

(1964) Pleistocene wood rat middens and climatic change in Mohave Desert: A record of juniper woodlands. Science 143: 1171–1174.

66.

Williams

Jackson

(2007) Novel climates, no-analog communities, and ecological surprises. Frontiers in Ecology and the Environment 5: 475–482.

67.

Williams

Jackson

Kutzbach

(2007) Projected distributions of novel and disappearing climates by 2100 AD. Proceedings of the National Academy of Sciences of the United States of America 104: 5738–5742.

68.

Wolaver

Sharp

Rodriguez

et al . (2008) Delineation of regional arid karstic aquifers: An integrative data approach. Ground Water 46: 396–413.