Holocene climatic variability indicated by a multi-proxy record from southern Africa’s highest wetland

Abstract

The eastern Lesotho Highlands experience climate patterns distinct from those of surrounding lower altitude regions, representing a niche environment with a unique biodiversity, leading to well-adapted but restricted vegetation. This study explores changes in the Holocene composition of diatoms and pollen at southern Africa’s highest altitude wetland (Mafadi: 3390 m a.s.l.). The palaeoenvironmental record for Mafadi Wetland indicates fluctuations between cold, wet conditions, prevalent between ~8140 and 7580 cal. yr BP and between ~5500 and 1100 cal. yr BP, and warmer, drier periods between ~7520 and 6680 cal. yr BP and between ~6160 and 5700 cal. yr BP. Marked climatic variability is noted from ~1100 cal. yr BP with colder conditions at ~150 kyr BP. Notably, the first of these cold periods occurs soon after the Northern Hemisphere 8.2 kyr event, while a second period of notably cold conditions occurs around 1100 cal. yr BP. Variability exists between the moisture reconstructions presented in this study and those from adjacent lower altitude sites, which is hypothesised to reflect variations in the strength and extent of the Westerlies throughout the Holocene.

Keywords

diatoms eastern Lesotho palaeoclimate pollen rapid cold events sediments

Introduction

Southern Africa provides a valuable geographic setting for palaeoenvironmental and palaeoclimatic research owing to the large number of ocean-atmospheric driving forces on regional and local climate (Chase and Meadows, 2007). Situated at the confluence of the Indian and Atlantic Oceans, and spanning the sub-tropics to mid-latitudes, a wide range of climates and biomes comprise the contemporary environment (Mucina and Rutherford, 2006). Throughout the Holocene, spatial variation in climate changes has been noted for a range of sites, driven by varying strengths of the mid-latitude Westerlies and tropical Easterlies, and the position of the Inter Tropical Convergence Zone (ITCZ) (cf. Chase and Meadows, 2007; Chase et al., 2011, 2013, 2015; Norström et al., 2014). Questions have also been raised concerning the extent to which Holocene climatic changes in the Southern Hemisphere are consistent in timing, extent and magnitude with those of the Northern Hemisphere (Gasse, 2000). Although many studies have addressed the issue of past environments in southern Africa (Norström et al., 2009), the late Quaternary record remains uncertain and at best undefined for much of the Lesotho Highlands (Grab et al., 2005). The Lesotho Highlands offer a unique study region in southern Africa given their high altitude (>2200 m a.s.l.) and variable topography (Figure 1), influencing local and regional climate. Furthermore, palaeoenvironmental records from eastern Lesotho’s wetlands permit comparison with neighbouring lower altitude records.

Figure 1.

Topographic map indicating the location of the Mafadi Wetland site.

The majority of Quaternary environmental reconstructions for the Lesotho region stem from a rich archaeological heritage (cf. Cain, 2009; Carter, 1976; Mitchell, 1996; Mitchell et al., 1994, 1998, 2011; Plug, 1993; Stewart et al., 2012). This archaeological work has included palaeoclimatic inferences (Mitchell et al., 1998), but the temporal resolution and quantification of past climates lack detail. A variety of Quaternary periglacial and glacial studies from Lesotho’s eastern high mountain region have indicated colder and possibly relatively wet conditions during the late Pleistocene (Grab, 2002b; Harper, 1969; Mills and Grab, 2005; Mills et al., 2009) and Holocene neoglacial episodes (Grab, 2000). However, the absence of age determinations for many of these periglacial landforms, the lack of consistent temporal chronologies and the disputed interpretation of some geomorphological phenomena have limited their palaeoenvironmental value (Boelhouwers and Meiklejohn, 2002).

The only published record analysing fossil pollen from a sedimentary sequence in the eastern Lesotho Highlands (Van Zinderen Bakker, 1955) is limited in its palaeoenvironmental value owing to the absence of a chronology, a low sampling resolution and the examination of only five plant taxa. Although age determinations are available for a few Holocene sedimentary sequences in eastern Lesotho (Hanvey and Marker, 1994; Marker, 1994, 1995, 1998), these lack analyses on pollen, diatoms or other environmental proxy markers and thus are of limited value for detailed environmental and climatic reconstructions. The only detailed Holocene palaeoenvironmental records for Lesotho stem from the western lowlands, based on charcoal assemblages and grazer tooth enamel in the Phuthiatsana-ea-Thaba Bosiu Basin (Esterhuysen and Mitchell, 1996; Roberts et al., 2013; Smith et al., 2002), sedimentary and phytolith records from an exposed sedimentary sequence in a gully of the Tsoaing Basin (Grab et al., 2005) and from central Lesotho, based on phytoliths and stable isotopes (Parker et al., 2011). Many of these studies explore palaeoenvironmental proxies extracted from archaeological excavations and therefore are limited because of potential biases of material selection (Esterhuysen and Mitchell, 1996; Fitchett et al., 2016).

A more focussed effort is required to improve the palaeoenvironmental record for the region. The eastern Lesotho Highlands are particularly well positioned for exploring late Quaternary climate fluctuations and associated ecological changes in the alpine zone. The high altitude, relatively high latitude and frequent occurrence of snowfalls produce a particularly marginal environment for plant growth. Not only are a very specific group of species able to survive the harsh environmental conditions, but any climatic changes can potentially lead to the extirpation of plant groups (Carbutt and Edwards, 2006; Inouye, 2008). While plants at lower altitudes may respond to atmospheric warming by upslope succession, those at the mountain summit habitat edge are unable to relocate altitudinally during such warming and so may suffer heat stress (Parmesan and Yohe, 2003). Similarly, the contemporary sparse vegetation cover on the highest peaks of eastern Lesotho would have occurred further downslope during past colder or drier periods (Grab, 1996; Inouye, 2008). Consequently, vegetation at the habitat edge is more sensitive to climate changes, with more noticeable responses, than at lower altitude sites (Carbutt and Edwards, 2004).

This study presents a Holocene climate and environmental record based on diatom and pollen communities and sediment characteristics from southern Africa’s highest wetland (here termed Mafadi Wetland; see Figure 1). In so doing, we aim to provide the highest resolution palaeoenvironmental record thus far for the alpine zone of southern Africa.

Study region

Mafadi summit is located on the border between Lesotho and South Africa and is the highest peak in South Africa and the third highest in Lesotho (3450 m a.s.l.). Immediately northwest of the summit (29°11′58″S, 29°21′04.1″E) is a bowl-shaped depression, with a spring-fed wetland. According to Schwabe (1995), the wetland is classified as a ‘bog’ – the highest in southern Africa at 3390 m a.s.l. (Figures 1 and 2). Mean annual temperatures at this site are estimated at ~4°C, varying from 9°C in January to −2°C in July, with extreme minimum temperatures reaching −18°C in winter (Grab, 1996, 2002a). Annual precipitation is highly variable across the Lesotho Highlands, ranging between 740 and 1600 (Sene et al., 1998), with the Mafadi summit region likely receiving a mean median value of this (Grab, 1996). Most precipitation above ~3000 m a.s.l. between May and September falls as snow but accounts for less than 10% of total annual precipitation (Nel and Sumner, 2008). The highlands receive approximately 12 light-to-moderate snowfalls per annum (Grab and Linde, under review). There are considerable diurnal air temperature variations, with surficial soil freezing typically occurring from late March to mid-October (Grab, 2000). In the Mafadi summit region, contemporary frozen ground may persist from mid-May to early October and to depths exceeding 50 cm (Grab, 1996, 2004).

Figure 2.

Photograph of the Mafadi Wetland site, indicating the coring location and diatomite exposures. The sparse vegetation upslope from the wetland is demonstrated in the foreground.

The geology of the eastern Lesotho Highlands comprises Drakensberg Group flood basalts with occasional kimberlite pipe intrusions, some of which are diamondiferous, to the northwest of Mafadi summit. Although the basalts are underlain by Stormberg group sandstones of late Permian to Jurassic age (260–140 Ma), these are only exposed at considerably lower altitudes (below 2500 m) and thus have no influence on higher altitude sedimentary sequences. The contemporary Mafadi Wetland extends across a broad gently sloping upper valley head, where the wetland is ca. 180 m in maximum width, and continues in a northerly (downstream) direction for ca. 480 m, before terminating at a boulder-dominated slope deposit. The wetland surface topography comprises a predominantly periglacial hummocky (thufur) relief, which in places is interrupted by flarks, seasonal pans and shallow standing water (<30 cm deep). The wetland site is characterised by white patches where diatomite-rich sediments are exposed (Figure 2). The moist wetland conditions and the presence of diatomite, together with being located in a landscape depression, render the site ideal for palaeoenvironmental work, as both pollen and diatoms are well preserved.

Above ~3400 m a.s.l., vegetation becomes increasingly sparse and almost absent near the summit. The vegetation is broadly classified as alpine grassland (Carbutt and Edwards, 2004), comprising Erica–Helichrysum heath, with shrubs and grasses notably dwarfed compared to those at lower altitude. Contemporary vegetation within the wetland consists of meadow grasses and sedges but comprises predominantly Helichrysum shrubs on the slopes, with increasingly sparse vegetation cover towards the highest summits. This suggests the study site is in close proximity to the altitudinal limit of vegetation growth in this part of southern Africa (Grab, 1996). The sparsely vegetated higher summit interfluves host active micro-scale (several centimetre in diameter) and relic larger (>1 m diameter) periglacial sorted stone circles (Grab, 1996, 2002b).

Methods

A 1.03-m sediment core was extracted from the wetland in April 2014 using 75 mm PVC tubing (Figure 2) and subsampled in a pressure-sealed laboratory at a 3-cm sampling interval. These 38 samples were each divided into four subsamples of equal volume. Three of these were each prepared for pollen, diatom and sediment analysis, respectively, to ensure direct comparability between the results from each of the proxies. The remaining quarter of each subsample was stored for age determination. As identifiable macrofossils were not preserved, bulk organic material obtained from six subsamples at relatively equally spaced depths throughout the core was sent for accelerator mass spectrometer (AMS) radiocarbon dating at Beta Analytic (Table 1). These dates were calibrated using the Southern Hemisphere SHCal13 model (Hogg et al., 2013). The BACON model (Blaauw and Christen, 2011) was used to interpolate dates for the remainder of the profile, which employs millions of self-adjusting Markov Chain Monte Carlo iterations. This model was selected given the improved performance of Bayesian over linear regression models and the inclusion of information on sample thickness. No outliers were identified by the BACON model. Interpolated dates from the BACON model are used throughout this paper to ensure consistent interpretation.

Table 1.

AMS radiocarbon dates obtained from Beta Analytic for the Mafadi Wetland sequence.

Lab code	Laboratory ID	Conventional age (yr BP)	1σ uncertainty (year)	2σ calibrated age range (BP)	Mean depth (cm)	Sample thickness (cm)	δ¹³C (‰)
Beta-405423	M3	870	±30	790–680	7	3	−26.6
Beta-405424	M13	1280	±30	1270–1205 and 1190–1065	41	3	−26.8
Beta-405425	M19	4960	±30	5710–5675 and 5665–5595	59	3	−28.5
Beta-405426	M32	6650	±30	7565–7435	84	3	−25.8
Beta-393713	M38	7140	±30	7970–7925 and 7900–7865	103	3	−26.3

Palaeoenvironments were investigated using a range of analyses. At the broadest scale, sediment properties were used to determine moisture availability, demonstrated predominantly by relative changes in the percentage organic content, and the proportions of sand-sized particles to silt- and clay-sized particles. Distinct variations in the skewness:kurtosis ratio indicate changes in depositional environment (Masselink et al., 2014; Saarinen and Petterson, 2006). Skewness indicates periods during which a disproportionately large component of the particle size distribution comprises coarser particles, indicative of drier conditions with less active weathering (Masselink et al., 2014). Kurtosis measures the broad spread of particle sizes and thus the variability in the inferred moisture regime. Organic and carbonate contents of each sample were determined using loss-on-ignition (LOI) at 550°C and 925°C, respectively (Heiri et al., 2001). Sediment particle size distributions for each sample, together with the mean, skewness and kurtosis, were measured using a Malvern Mastersizer 3000.

Pollen was used to reconstruct past vegetation composition and to study the presence and absence of indicator species for alternating wetland and grassland conditions in the Highlands region. Pollen preparation followed standard procedures outlined by Faegri et al. (1989). Once the pollen had been isolated and slides prepared, a minimum of 250 pollen grains were counted per sample at a magnification of 400× using an Olympus BX51 light microscope. Identification was made with reference to the African Pollen Database and the reference collection at the University of the Free State, South Africa.

Diatoms were used to reconstruct the aquatic conditions and biodiversity within the wetland. Diatom preparation was undertaken using the procedures outlined by Battarbee et al. (2001). A minimum of 300 diatom valves were counted per sample at a magnification of 1000× using an Olympus BX51 light microscope. Diatoms were identified through consultation with both local (Harding and Taylor, 2011; Matlala et al., 2011; Schoeman, 1973; Schoeman and Archibald, 1976) and international (Camburn and Charles, 2000; Krammer, 2002; Krammer and Lange-Bertalot, 1986; Patrick and Reimer, 1975; Snoeijs and Balashova, 1998) literature.

Important gradients in the pollen and diatom data sets were explored using indirect ordination techniques. Initially, each data set was analysed using detrended correspondence analysis (DCA) to determine whether species responses along the main ordination axes were unimodal or linear. Axis 1 gradients for DCA for pollen and diatoms were under 3 (2.1 and 1.7, respectively), so ordination analyses were re-run using principal component analysis (PCA) on square-root-transformed raw percentage composition data. Major changes in vegetation composition were delimited into zones, determined using Constrained Incremental Sum of Squares (CONISS), with significance tested using the Broken Stick Model (Grimm, 1987). We then superimpose these zones onto the sediment and diatom record to facilitate discussion of environmental changes during discrete, yet statistically relevant, time periods. The pollen record was selected as the target variable for zonation because of its more regional representation. All statistical analyses were undertaken using the code-based statistical platform R (Venables and Smith, 2015), and stratigraphic plots were produced using C2 (Juggins, 2007).

Results

Sediment core lithology reflects a relatively shallow wetland environment throughout much of the period (Figures 3 and 6). The six AMS radiocarbon ages constrain the profile to the mid-Holocene with a basal age of 7140 ± 30 yr BP (Table 1). Sediments sampled at a depth of 7 cm are dated to 870 ± 30 yr BP (Table 1).

Figure 3.

BACON output of the age–depth profile for Mafadi Wetland.

The sedimentation rate, calculated by the BACON model, is averaged for the sequence at 100 yr cm⁻¹ (0.01 cm/yr). The upper sediments were likely deposited within the last century, while the bottommost sampled sediments were likely deposited since 8140 cal. yr BP. There is a marked period of particularly slow sedimentation, which extends from 1280 ± 30 to 4960 ± 30 cal. yr BP (Figure 3). During this period, only 18 cm of sediment accumulated, yielding a much slower sedimentation rate (mean: 205 yr cm⁻¹), resulting in a lower temporal frequency of samples (as they were sampled at consistent depth) and generally less well-resolved profiles. Before and after this period of slow sedimentation, periods of consistent and comparatively more rapid sedimentation are found (Figure 3).

CONISS revealed four statistically significant zones in the pollen stratigraphy (M1–M31; surface: 84 cm) (Figure 4). Samples spanning a depth of 103–84 cm contained insufficient pollen for counting (<20 pollen grains counted across duplicate slides from each sample) and therefore were excluded from zonation analyses. We assigned these pollen-depleted samples at the base of the profile to a non-numerical zone MP5. Zone MP5 extends from a depth of 103 to 84 cm (~8140–7580 cal. yr BP). Zone MP4 covers the sequence from 81 to 72 cm (~7380–6680 cal. yr BP), marking the terminal depth of samples containing sufficient pollen for counting. Zone MP3 includes seven samples and extends from 71 to 62 cm depth (~6610–6010 cal. yr BP). Zones MP1 and MP2 comprise larger sets of samples than zones MP3 and MP4 and are separated at a mean depth of 26.5 cm (~1080 cal. yr BP).

Figure 4.

CONISS output separating the Mafadi Wetland profile into zones based on the pollen results.

The sediment profile is dominated by silt-sized sediment particles (>40%), with relatively low percentages of organic matter through zones MP5–MP2 (>15%; Figure 5). The percentage of organic matter rapidly increases in zone MP1 (~1080 cal. yr BP to present), with a peak in the contemporary material (Figure 5). The percentage of sand-sized particles peaks in zone MP4 and at the termination of zone MP2 (Figure 5). The relationship between skewness and kurtosis reflects no statistically significant changes in sedimentation environment.

Figure 5.

Stratigraphic diagram representing the variations in sediment characteristics.

The pollen record is dominated by Poaceae (43.6%), Cyperaceae (23.5%) and Asteraceae (22.2%). A total of 28 taxa were identified, of which 20 taxa appeared with a frequency of more than 1% at any point throughout the profile. Because of morphological and environmental similarities, pollen counts from the plant families Chenopodiaceae and Amaranthaceae are summed to form a single group ‘Cheno-Am’ (Scott et al., 2012). The pollen composition of the sequence is largely representative of the contemporary wetland environment, comprising semi-aquatic species, shrubs and herbs, and succulents (Figure 6). Occasional Podocarpus pollen grains were counted (<2% maximum occurrence; Figure 6). As Mafadi Wetland is situated above the treeline, these must have been transported to the site by wind from adjacent forests at lower altitudes. The ratio of Asteraceae:Poaceae is presented (Figure 6) as a proxy for the strength of seasonality of precipitation, which is suggested to represent changes in the latitudinal extent, or strength of influence, of the Westerlies (Coetzee and Vogel, 1967; Norström et al., 2009). However, Norström et al. (2009) argue that local and regional components cannot be separated; thus, the ratio may provide locally biased indications of regional moisture availability. As the bowl-shaped depression characterising the site is likely to result in a relatively regional representation of the pollen sum, a ratio value >0.5 is interpreted as a seasonal shift to summer precipitation, and in the case of Lesotho, a decrease in snowfall; a value <0.5 suggests a proportional increase in winter precipitation (Coetzee and Vogel, 1967; Norström et al., 2009; Scott and Nyakale, 2002). Such shifts in seasonality are indicative only of the timing and not the quantity of precipitation (Coetzee and Vogel, 1967). Principal component 1 accounts for a statistically significant 29.1% (p > 0.0001) of the variance of the pollen distribution in the samples, separating at extremes Poaceae, Indigofera and Apiaceae, which have the strongest negative scores (−1.65, −0.58 and −0.29, respectively), from Cyperaceae and Asteraceae, with the strongest positive scores (1.6 and 0.5, respectively).

Figure 6.

Stratigraphic diagram of the pollen results for Mafadi Wetland.

A total of 37 diatom species were identified in the Mafadi sequence, of which 16 have a proportional representation of >1%. The diatom record is dominated by Staurosirella (Fragilaria) pinnata (22.4%), Aulacoseira ambigua (19.3%) and Fragilaria construens (19.0%). Because of similarities in both the morphology and habitat preferences (Ohlendorf et al., 2000; Wang et al., 2013), Staurosirella (Fragilaria) pinnata and Fragilaria construens are grouped together and named Fragilaria pinnata/construens. The diatom profile comprises fluctuations between periods dominated by the planktonic and facultative planktonic species Aulacoseira ambigua and Fragilaria pinnata/construens and periods of more diverse diatom assemblages dominated by aerophilic and littoral taxa, especially aerophilic Eunotia praerupta and Pinnularia divergentissima (Figure 7). Zone MP5, for which no pollen could be counted, demonstrates the highest proportion of Fragilaria pinnata/construens (Figure 7). Principal component 1 accounts for a statistically significant 68.1% (p < 0.0001) of the observed variance in diatom species distribution across the profile and, consistent with the observations made above, segregates at extremes Fragilaria pinnata/construens and Aulacoseira ambigua with strong negative scores (−1.9 and −1.6, respectively) from Eunotia praerupta, Eunotia bilunaris and Pinnularia borealis with the strongest positive scores (1.9, 1.0 and 1.0, respectively).

Figure 7.

Stratigraphic diagram of the diatom results from Mafadi Wetland.

Discussion

With a basal date of ~8140 cal. yr BP, this sequence does not span the entire Holocene but rather commences at a period coincident with the final disappearance of the major Northern Hemisphere ice sheets (Mayewski et al., 2004). The very slow sedimentation rate of 100 yr cm⁻¹ is attributed to the particularly low temperatures throughout the late Quaternary to present (Grab, 1996, 2004). Climate would have been an important control at high altitude, given the peat inclusion in the sediment record, which required relatively mild temperatures to facilitate plant growth and sufficient humidity for peat to develop under water-logged conditions (Meadows, 1988; Van Zinderen Bakker, 1955). The zones, based on the CONISS analysis on the pollen results, coincide with changes across diatoms and sediment properties, indicating clearly defined climate and environmental shifts at Mafadi Wetland. At present, grass and small shrub cover ceases mid-way up the slope between the wetland and Mafadi summit (at ~3400 m a.s.l.), suggesting that the site is located in close proximity to the terminal altitude for vegetation cover in eastern Lesotho (Grab, 1996, 1999). This being the highest altitude wetland in southern Africa, it provides an opportunity to discuss marginal communities and threshold conditions associated with climate change through the Holocene.

The distribution of plant species represented by the pollen record for Mafadi summit is largely consistent with the contemporary regional vegetation including the ubiquitous bogs, comprising grasslands, sedges and small herbs and shrubs. The few occurrences of Podocarpus pollen are wind-derived from lower altitude regions in South Africa (Neumann et al., 2014; Van Zinderen Bakker, 1955). Despite the isolation of Mafadi Wetland, no endemic diatom species were found; in fact they were rather cosmopolitan. Two communities stand out due to their presence in similar environments elsewhere in the world. The first are the Fragilaria species, which stand out due to their role as r-strategists and consequent tolerance to harsh environments and in particular to ice and snow cover. For the Mafadi site, the most harsh conditions are inferred as particularly cold periods, with considerable ice cover (Grab, 1996). Therefore, their presence supports inferences of colder temperatures in instances where they dominate the diatom profile. These are cosmopolitan species, which appear in a range of high-stress environments but are dominant in high alpine wetlands and lakes in a range of locations, including Uganda (McGlynn et al., 2010; Panizzo et al., 2008), Switzerland (Lotter and Bigler, 2000; Ohlendorf et al., 2000), Canada (Karst-Riddoch et al., 2005), Finland (Shala et al., 2014) and Siberia (Mackay et al., 2012; Westover et al., 2006). The second are planktonic Aulacoseira ambigua and aerophilic Hantzschia amphioxys, which indicate periods of fluctuating wet and dry conditions, at sites in South Africa (Finné et al., 2010), Mozambique (Sitoe et al., 2015), Australia (Gell and Little, 2007) and Canada (Hargan et al., 2015). All of the diatom species identified at Mafadi Wetland have been previously recorded in Lesotho (Schoeman, 1973) and South Africa (Harding and Taylor, 2011). While the pollen record is specific to the vegetation of southern Africa, the dominant Poaceae, Cyperaceae and Asteraceae are common to wetland conditions globally, with shifts between Poaceae and Cyperaceae representing a cosmopolitan indicator of shifts between dry and wet conditions within the local environment (Gasse and Van Campo, 1998).

MP5: ~8140–7580 cal. yr BP

Zone MP5 is marked by a complete absence of pollen (Figure 6), which may be due to a number of reasons, such as absence of plants at the site during this period, the prevention of pollen being deposited onto wetland sediments (perhaps during periods of extended snow and ice cover) or an alkaline pH that would compromise the preservation of pollen grains (cf. Grab et al., 2005; Gasse and Van Campo, 1998; Metwally et al., 2014). The diatom composition almost entirely comprises Aulacoseira ambigua and Fragilaria pinnata/construens (Figure 7), reflecting a period dominated by benthic r-strategy taxa which are able to tolerate particularly harsh conditions. The habitat of these species suggests the presence of a shallow water environment (Gasse and Van Campo, 1998; Sitoe et al., 2015), which does not support an inference of conditions too dry to support plants. Fragilaria pinnata/construens are common in alpine lakes of East Africa and survive under conditions of substantial ice cover and cold water temperatures (Ohlendorf et al., 2000; Schmidt et al., 2004; Wang et al., 2013). The diatom composition does not suggest pH conditions markedly different from the remainder of the core, and thus, it is unlikely that the pollen deteriorated under conditions of very high pH. It is thus inferred that this was a harshly cold but wet period, with temperatures too low to sustain terrestrial plants, resulting in a climatic barrier to plant growth at the site during this time. The percentage composition of carbonates and organic material in the sediments is low during this period (Figure 5), further indicative of low rates of primary production. The large percentage of silt-sized particles further confirms wetland presence in this zone (Figure 5). The period at the beginning of this zone, marked by the highest percentage of Fragilariod diatoms (Figure 7), suggesting a very cold, wet period and the persistence of a shallow lake, likely ice covered, notably occurs immediately after the ‘8.2 kyr’ cold event and final break-up of persistent ice sheets in the Northern Hemisphere (Alley and Ágústsdóttir, 2005; Chase et al., 2015; Mayewski et al., 2004).

MP4: ~7520–6680 cal. yr BP

Fossil pollen emerges at the start of this zone, dominated by Cyperaceae, suggesting at least localised wet patches (Figure 6). This is concurrent with a peak in Cheno-Am, Acanthaceae and the drought-resistant Crassula and Aizoaceae, with relatively high percentages of Poaceae and Asteraceae (Figure 6). This coincides with a decrease in the relative abundance of Aulacoseira ambigua and Fragilaria pinnata/construens as the zone commences, replaced by an increase in aerophilic diatoms (Eunotia praerupta, Pinnularia borealis and Pinnularia divergentissima, Figure 7) indicating a shift to shallower water conditions, most likely associated with a drier, possibly warmer climate. The beginning of the zone is marked by high Asteraceae:Poaceae ratio values, tentatively representing weak rainfall seasonality and a strengthening of the Westerlies, but shifts to <0.5 for the remainder of the profile, indicating the re-establishment of summer rainfall (Figure 6; Norström et al., 2009). This shift is concurrent with a brief decrease in Cyperaceae and Asteraceae, a decrease in the percentage organic content and an increase in the percentage of sand-sized sediment particles (Figures 6 and 7). Paired with a second abrupt decrease in the proportion of planktonic and facultative planktonic diatoms, this is indicative of a decline in lake levels to form a marshy environment, which progressively shifts to a drier environment, indicated by the decline in pollen PC1 scores and the dominance of aerophilic diatoms (Figures 7 and 8).

Figure 8.

Multi-proxy plot comparing the climatic reconstruction for Mafadi Wetland to regional moisture and temperature metrics.

MP3: ~6610–5700 cal. yr BP

The pollen record for zone MP3 is characterised by an overall increase in the relative abundance of Poaceae and decrease in the percentage of Asteraceae pollen (Figure 5) and is similar to zone MP4 reflecting a continuation of the drying process. The Asteraceae:Poaceae ratio suggests a shift to more exclusive austral summer precipitation (Norström et al., 2009), while the relative decrease in Cyperaceae indicates a progressively drier environment. The diatom record is marked by a larger proportion of epiphytic diatoms (Figure 6), supporting inferences for warmer wetland conditions and a reduction in ice cover than during the previous zone (~7520–6680). The proportion of benthic and planktonic, relative to aerophilic, diatoms fluctuates throughout the zone (Figure 6). The relatively low proportion of ice-tolerant Fragilaria species, and the increase in epiphytic diatoms, supports the inference of even warmer conditions than that represented in zone MP4. This warming period coincides with the Holocene Altithermal (Neumann et al., 2014), although due to the altitude of the site it must necessarily be inferred as a ‘milder’ event rather than an objectively ‘warm’ event. A peak in the percentage organic composition of sediments is noted for the middle of this period, with the concurrent peak in Cyperaceae pollen indicating the continued presence of small ponds or flarks during this relatively dry period (Figures 6 and 7). The period of slower sedimentation commences at ~6000 cal. yr BP and persists to the end of this zone, with a lithology comprising a mixture of dark peat and clay.

MP2: ~5600–1100 cal. yr BP

The period of slow sedimentation continues throughout much of this zone. The first half of zone MP2 is marked by abrupt changes in pollen, diatoms and sediment properties leading up to peaks at ~4500 cal. yr BP. Poaceae increased considerably to a peak at ~4500 cal. yr BP. This is coincident with the first of the Holocene neoglacial events (Jerardino, 1995). The remainder of the zone comprises a more consistent period relative to MP4 and MP3, with greater stability in all three proxies. This could be because of the lower temporal frequency of sampling, which does not facilitate the detection of more short-lived climate events. However, as the flora do not demonstrate notable changes for the samples spanning this period, the stability in the pollen and diatom PC1 sample scores suggests that, overall, the conditions were indeed relatively stable.

Poaceae dominates the pollen record for this zone with relative stability in abundance until ~2000 cal. yr BP (Figure 6). This results in a Poaceae:Asteraceae ratio of >0.5 (Figure 6), indicative of a more dominant summer rainfall regime likely driven by weakened Westerlies (Norström et al., 2009). The relative abundance of Cyperaceae and Asteraceae is very low (Figure 6). Compared with the contemporary environment, this would suggest a decrease in the surface water extent and a re-establishment of grassland conditions. However, the presence of large proportions of Aulacoseira ambigua suggests that a shallow lake, albeit probably smaller in size, persisted throughout this period (Gasse and Van Campo, 1998; Sitoe et al., 2015). The decrease in Cyperaceae and Asteraceae pollen, therefore, may be in response to colder temperatures, which may have exceeded their cold tolerance. This hypothesis is supported by the proportional increase in Fragilaria species relative to the preceding two zones, which would suggest climatic conditions were suitable only to the more cold water–tolerant, ice-tolerant and snow-tolerant species (Ohlendorf et al., 2000; Wang et al., 2013). Alternately, because of the bowl-shaped topography of the wetland, a substantial surface water extent of the wetland, as supported by the presence of Aulacoseira ambigua, may possibly have forced these terrestrial plants to less suitable areas upslope, which only grass species can tolerate. Notably, the highest relative abundance of Apiaceae in the profile is observed in this zone (Figure 6). This is most likely because this semi-aquatic species has a considerable cold tolerance, as evidenced by its common presence on Marion Island (Nyakatya and McGeoch, 2008). While the relative abundance of Fragilaria pinnata/construens is equivalent to zone MP5 (Figures 7 and 8), the presence of pollen suggests that conditions were not as cold as during MP5 but is consistent with a period of cooling following the Holocene Altithermal indicated for southern Africa until 2800 cal. yr BP (Marcott et al., 2013). The zone terminates with an increase in sedimentation rate and a short-lived peak in Poaceae pollen following the relatively consistent percentage composition for a period of ~2500 years (Figure 6). This is concurrent with the highest relative abundance of planktonic Aulacoseira ambigua in the profile (Figure 7).

MP1: ~1060 cal. yr BP to present

The pollen indicates more rapid moisture fluctuations in the first half of this zone, and fluctuations in the hydrology of the wetland, from one with shallow ponds to drier marshy conditions. During this period, aerophilic and benthic diatom species re-emerge (Figure 7), likely a result of low water levels caused by reduced rainfall. This period is also characterised by a shift in the Asteraceae:Poaceae ratio, indicating a decrease in rainfall seasonality and an increase in the strength of the Westerlies (Norström et al., 2009). The percentage organic matter increases during this zone, which, given no prominent shifts in pollen composition in the second half of the zone, rather suggests a change in sediment accumulation, possibly associated with more year-round rainfall distribution. At ~820 cal. yr BP, a brief peak in the Asteraceae:Poaceae ratio is reflected in the pollen record (Figure 6), potentially suggesting a short-lived increase in the strength of the Westerlies. A concurrent short-lived peak in Fragilaria species occurs (Figure 7), indicating cooler conditions which are tolerated more readily by these species which proliferate in ice and snow conditions, tentatively supporting an inference of more frequent mid-latitude cyclones resulting from strengthened Westerlies. Evidence for this cold event appears broadly contemporaneous with the beginning of the ‘Little Ice Age’ (Tyson et al., 2000). From ~750 cal. yr BP, fluctuations in the environmental proxies stabilise with relatively constant proportions of Asteraceae, and a slight change to Poaceae pollen from Cyperaceae, indicating an overall drying. A decrease in aerophilic diatoms indicates wetland conditions becoming more shallow. This is supported by a very low relative abundance of Fragilaria species from ~750 to ~150 cal. yr BP, followed by a slight increase in the uppermost samples (Figure 7).

Regional comparison

Comparison of regional temperature reconstructions

In addition to the proxy composition, it is valuable to explore the extent to which the climate and environmental reconstruction for Mafadi Wetland is consistent with those elsewhere in southern Africa and globally. This is of particular interest given the unique high-altitude alpine niche environment represented in eastern Lesotho, which provides a setting for testing palaeoenvironmental lapse rates and topographic boundaries to synoptic climate scenarios (Grab, 1997). The Mafadi Wetland record reflects two colder events, indicated by a predominance of ice-tolerant Fragilaria species, extending from ~8140 to 7850 cal. yr BP and ~1500 to 1050 cal. yr BP and notable temperature variability between 7000 and 6000 cal. yr BP and between 1500 and 500 cal. yr BP (Figure 8). The inverse of the percentage Fragilariod composition demonstrates broad similarity with the Marcott et al. (2013) regional temperature curves for 30°N–30°S and for 30°S–90°S (Figure 8). As the Eastern Lesotho Highlands are situated at the confluence of these two regional reconstructions, with the Mafadi site at 29.2°S, it is notable that for different temporal periods, the relative similarity with each of the two records shifts, particularly given inferences of shifts in the influence of the Westerlies in this region. For example, the fluctuations in Fragilariods between ~7000 and 6000 cal. yr BP are similar to fluctuations in the Marcott et al. (2013) 30°N–30°S curve, whereas the Fragilariod fluctuations between ~1500 and 500 cal. yr BP are more similar to those of the 30°S–90°S curve (Figure 8). The relative stability in Fragilariod composition in the diatom record from ~5500 to 1500 cal. yr BP is consistent with the Marcott et al. (2013) temperature curves for both regions (Figure 8).

A short-lived cold period driven by a large meltwater pulse in the northern Atlantic Ocean, often assumed to have influenced climates only in the Northern Hemisphere, is the ‘8.2 kyr’ event (Alley and Ágústsdóttir, 2005; Mayewski et al., 2004). The very cold conditions which characterise the commencement of the Mafadi Wetland record from ~8140 to 7850 cal. yr BP notably occur very soon after this event. Isotope records from archaeological settlements in western Lesotho provide further evidence for markedly cold conditions dated to coincide with the 8.2 kyr event (Smith et al., 2002), while phytolith records from Braamhoek indicate an increase in C₃ vegetation (interpreted as an indicator of cooling in southern African alpine environments) at 8000 cal. yr BP (Finné et al., 2010). The Makapansgat speleothem record provides evidence for a cool event at 8500 cal. yr BP (Holmgren et al., 2003), which may be synchronous with these events because of age uncertainties (Norström et al., 2009), and evidence for cooling attributed to the ‘8.2 kyr’ event has most recently been reported for the south-western Cape (Chase et al., 2015). It would thus seem possible that the cool conditions around 8000 cal. yr BP in southern Africa may be related to the final break-up of the Northern Hemisphere ice sheets. This may point towards ocean heat transport during this early Holocene deglaciation in the Northern Hemisphere impacting climate dynamics in the Southern Hemisphere.

The overall warming period associated with deglaciation continues until ‘optimal’ conditions at the Holocene Altithermal (Wanner et al., 2015). The timing of this event is unclear, with discrepancies for much of southern Africa, but it appears broadly to span the period ~7500–6500 cal. yr BP for the Southern Hemisphere (Holmgren et al., 2003; Neumann et al., 2014; Truc et al., 2013; Wanner et al., 2015). There is no clear warm signal coinciding with this event at Mafadi Wetland, although a decrease in the relative abundance of the more cold-tolerant Fragilaria species is observed. This may reflect temperatures at this high altitude not reaching particular temperature thresholds required for less cold-tolerant taxa to establish themselves, with upslope range shifts being limited by the extent of warming. It is notable, given this hypothesis, that there was no clear indication for Holocene Altithermal conditions at the lower altitude Braamhoek Wetland (Finné et al., 2010; Norström et al., 2009, 2014), yet pollen records from the similarly low-altitude Drakensberg site at Mahwaqa Mountain suggest a clearly defined Holocene Altithermal maximum at 6500 cal. yr BP (Neumann et al., 2014). Unfortunately, the isotope records from the archaeological shelter sites in western Lesotho do not span this period.

Considerable discussion has concerned itself with, and provided evidence for, the ‘Little Ice Age’ in southern Africa (cf. Herbert, 1987; Holmgren et al., 2003; Sundqvist et al., 2013; Talma and Vogel, 1992; Talma et al., 1974; Tyson and Lindesay, 1992; Zinke et al., 2014), a short-lived cold event that occurred between AD 1300 and 1800 (Tyson et al., 2000; Wanner et al., 2015). The second climate event during the last ~1000 years is the ‘Medieval Warm Period’ (or ‘Medieval Climatic Anomaly’), which preceded the ‘Little Ice Age’, with warmer than contemporary conditions between AD 1000 and 1300 (Wanner et al., 2015). The only evidence in southern Africa for a ‘Medieval Warm Period’ is derived from high-resolution speleothem isotope records (Holmgren et al., 2003; Tyson et al., 2000). It is therefore interesting that the relatively mild conditions are inferred from a particularly low relative abundance of Fragilaria species and increase in pollen taxa diversity at Mafadi Wetland during this period. However, the resolution of our data is too low neither to facilitate the accurate identification of particularly cold conditions at this site associated with the ‘Little Ice Age’ nor to provide concrete evidence for warmer conditions that are contemporary with the ‘Medieval Warm Period’. Comparisons with the temperature curves presented by Marcott et al. (2013) demonstrate far more robust similarities.

Comparison of regional moisture reconstructions

In terms of moisture, the reconstruction for Mafadi Wetland is in broad agreement with Marker’s (1994, 1995, 1998) sedimentologically derived results for eastern Lesotho (Figure 8). Notably, however, when comparing the moisture inferences made in this study to those from lower altitude sites at Braamhoek and Mahwaqa Mountain, delays in the onset of dry and wet periods are noted. In particular, dry conditions commence consistently earlier at Mafadi Wetland than at Mahwaqa Mountain, located to the east (Neumann et al., 2014). This most likely is due to the influence of the Great Escarpment which restricts the transfer of moist air from the Indian Ocean to the Lesotho Highlands. Under regionally drying conditions, moisture will precipitate out at Mahwaqa because of orographic uplift, with limited moisture remaining for transport into the Lesotho Highlands. By contrast, moist conditions persist longer at Mafadi Wetland than at Braamhoek Wetland, located to the north (Finné et al., 2010; Norström et al., 2009, 2014). This is most likely because of the predominant role of the Westerlies in transporting moisture from the south-western region of the country, particularly in winter (Mills et al., 2012; Norström et al., 2009, 2014). A northward shift in the extent of the Westerlies would increase the likelihood of moisture transfer to the eastern Lesotho Highlands (Mills et al., 2012). Because of the lower latitude of Braamhoek wetland, the Westerlies would have less frequently extended sufficiently far north to ensure moisture transport to this region (Norström et al., 2009).

The earliest part of our record indicates relatively moist conditions. This is in contrast to the dry conditions associated with the period 8400–8000 for much of Central and North Africa (Gasse, 2000). This may indicate an early shift into the moist conditions which follow (Gasse, 2000) because of the mid-latitude position of the site and the heightened influence of the Westerlies. Debate is still ongoing with regard to the existence of an African Humid Period in southern Africa (Burrough and Thomas, 2013; Chase et al., 2009). This event spans the period 14,500–5500 cal. yr BP and has been detected in a range of records from western and tropical Africa (Burrough and Thomas, 2013). Evidence from hyrax middens in Namibia for humid conditions during this period has raised hypotheses of such an event extending as far as 23°S (Chase et al., 2009), yet a review of palaeoclimatic records from southern Africa, with a particular focus on dryland conditions, is not in support of this view (Burrough and Thomas, 2013). The Mafadi Wetland record broadly suggests relatively wet conditions until ~6500 cal. yr BP. This period is, however, marked by pronounced dry events. Precipitation reconstructions for other sites in eastern Lesotho indicate more consistent marked dry periods throughout this time period (cf. Grab et al., 2005; Marker, 1994, 1998). Evidence from pollen, diatom and sediment records presented in this study is, therefore, insufficient to determine whether the prolonged wet periods inferred for eastern Lesotho during this period are coincident with, or more substantive indicators of, an African Humid Period.

For the ‘Medieval Climate Anomaly’, a period of greater stability in precipitation is noted for central southern Africa from ~2000 cal. yr BP to present (Burrough and Thomas, 2013). The results from Mafadi Wetland indicate rapid high-amplitude fluctuations in precipitation throughout the period, as supported at a global scale (Wanner et al., 2015) and more recently noted for southern Africa (Chase et al., 2013; Norström et al., 2014). Debate on the ‘Little Ice Age’ event in southern Africa focuses more on precipitation during this period, with a current understanding that dry conditions occurred in the summer rainfall zone (Ekblom et al., 2008; Holmgren et al., 1999; Lee-Thorp et al., 2001; Neumann et al., 2010) but wet conditions in the winter rainfall zone (Stager et al., 2012; Weldeab et al., 2013). Further research using records of a higher temporal resolution would potentially contribute to confirming these hypotheses because of the position of the eastern Lesotho Highlands at the boundary of the palaeo-extent of the Westerlies.

Conclusion

This study presents a multi-proxy palaeoenvironmental reconstruction for the highest wetland in southern Africa, located at 3390 m a.s.l. This setting induces a marginal environment for highly cold-resistent plant and diatom species. The severity of this cold environment is highlighted by the absence of pollen at the commencement of the sequence at ~8140 cal. yr BP and dominance of cold-tolerant Fragilaria pinnata/construens diatoms. The emergence of pollen in the MP4–MP1 subsamples suggests upslope succession in terrestrial vegetation from ~7520 cal. yr BP thereafter, indicative of a shift to warmer temperatures more typical to those today.

The pollen, diatom and sediment results indicate fluctuations between dry and wet conditions throughout the mid- to late-Holocene. Notable periods of wet conditions, inferred from an increased relative abundance of semi-aquatic pollen taxa and supported by the presence of an increased surface depth required to support planktonic and facultative planktonic diatoms, are detected between ~8140 and 7580 cal. yr BP and between ~5600 and 1100 cal. yr BP. Periods of notably dry conditions, inferred from an increase in pollen from succulents and grasses, are reconstructed for ~7520–6680 cal. yr BP, ~6160–5700 cal. yr BP and from ~1000 cal. yr BP to present, and corroborated by increased abundances of aerophilic diatoms. Pronounced cold conditions are identified at ~8140 cal. yr BP and are possibly indicative of enhanced and persistent cool conditions after the 8.2 kyr event (Mayewski et al., 2004; Smith et al., 2002). Cold conditions at ~150 cal. yr BP may provide local evidence for cooling, yet the temporal resolution is too coarse to confirm coincidence with the ‘Little Ice Age’ (Tyson et al., 2000; Wanner et al., 2008). The age of large relic periglacial sorted patterned ground and stone-banked lobes around Mafadi is unknown but is presumed to be of ‘late-Holocene Neoglacial age’ (Grab, 2000, 2002b). Continued multi-proxy palaeoclimatic research may facilitate attributing such geomorphic activity to defined cold periods during the Holocene.

The use of a multi-proxy approach, exploring changes in pollen, diatoms and sediment characteristics, facilitated the corroboration of palaeoenvironmental inferences. This is particularly important in a climatically marginal environment, for which changes may, and do, result in the occasional absence of certain proxies. There is therefore an awareness of the relative lack of information for interpretation. This study therefore provides a detailed palaeoenvironmental reconstruction for an under-researched region, which is of geographical and environmental significance to the broader southern African sub-continent.

Footnotes

Acknowledgements

Many thanks to Professor Louis Scott for his invaluable assistance in pollen grain identification. Thanks to Professor Jasper Knight for valuable input on the preparation and interpretation of sediment samples. Our thanks also go to the Petrus Chakane, Janet Hope and Tula Maxted for their assistance with the laboratory preparation of pollen and diatom samples.

Funding

This project was funded by a National Research Foundation Scholarship awarded to JF, Oppenheimer Memorial Trust funding awarded to JF and DST-NRF Centre of Excellence for Palaeoscience Funding for 2013 and 2014 awarded to SG and 2016 awarded to JF.

References

Alley

Ágústsdóttir

(2005) The 8k event: Cause and consequences of a major Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–1149.

Battarbee

Jones

Flower

. (2001) Diatom analysis. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments, Vol. 3: Terrestrial, Algal and Siliceous Indicators. Dordrecht: Kluwer Academic Press, pp. 155–202.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6(3): 457–474.

Boelhouwers

Meiklejohn

(2002) Quaternary periglacial and glacial geomorphology of southern Africa: Review and synthesis. South African Journal of Science 98: 47–55.

Burrough

Thomas

DSG

(2013) Central southern Africa at the time of the African Humid Period: A new analysis of Holocene palaeoenvironmental and palaeoclimate data. Quaternary Science Reviews 80: 29–46.

Cain

(2009) Cultural heritage survey of Lesotho for the Maloti-Drakensberg Transfrontier Project, 2005–2006: Palaeontology, archaeology, history and heritage management. South African Archaeological Bulletin 64(189): 33–44.

Camburn

Charles

(2000) Diatoms of Low-Alkalinity Lakes in the Northeastern United States. Philadelphia, PA: The Academy of the Natural Sciences of Philadelphia.

Carbutt

Edwards

(2004) The flora of the Drakensberg Alpine Centre. Edinburgh Journal of Botany 60(3): 581–607.

Carbutt

Edwards

(2006) The endemic and near-endemic angiosperms of the Drakensberg Alpine Centre. South African Journal of Botany 72: 105–132.

10.

Carter

(1976) The effects of climatic change on settlement in eastern Lesotho during the Middle and Later Stone Age. World Archaeology 8: 197–206.

11.

Chase

Meadows

(2007) Late Quaternary dynamics of southern Africa’s winter rainfall zone. Earth-Science Reviews 84: 103–138.

12.

Chase

Boom

Carr

. (2013) Holocene climate change in southernmost South Africa: Rock hyrax middens record shifts in southern westerlies. Quaternary Science Reviews 82: 199–205.

13.

Chase

Boom

Carr

. (2015) Evolving southwest African response to abrupt deglacial North Atlantic climate change events. Quaternary Science Reviews 121: 132–136.

14.

Chase

Meadows

Scott

. (2009) A record of rapid Holocene climate change preserved in hyrax middens from southwestern Africa. Geology 37: 703–706.

15.

Chase

Quick

Meadows

. (2011) Late glacial interhemispheric climate dynamics revealed in South African hyrax middens. Geology 39(1): 19–22.

16.

Coetzee

Vogel

(1967) Evidence for the Paudorf interstadial in Africa. Palaeoecology of Africa 2: 100–101.

17.

Ekblom

Gillson

Risberg

. (2008) Rainfall variability and vegetation dynamics of the lower Limpopo Valley, Southern Africa, 500 AD to present. Palaeogeography, Palaeoclimatology, Palaeoecology 363–364: 68–78.

18.

Esterhuysen

Mitchell

(1996) Palaeoenvironmental and archaeological implications of charcoal assemblages from Holocene sites in western Lesotho, Southern Africa. Palaeoecology of Africa 24: 203–232.

19.

Faegri

Iversen

Krzywinski

(1989) Textbook of Pollen Analysis. Caldwell, NJ: The Blackburn Press.

20.

Finné

Norström

Risberg

. (2010) Siliceous microfossils as late-Quaternary paleo-environmental indicators at Braamhoek wetlands, South Africa. The Holocene 20(5): 747–760.

21.

Fitchett

Grab

Bamford

. (2016) A multi-disciplinary review of late Quaternary palaeoclimates and environments for Lesotho. South African Journal of Science. doi:10.17159/sajs.2016/20160045.

22.

Gasse

(2000) Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Reviews 19: 189–211.

23.

Gasse

Van Campo

(1998) A 40,000-yr pollen and diatom record from Lake Tritrivakely, Madagascar, in the southern tropics. Quaternary Research 49: 299–311.

24.

Gell

Little

(2007) Water Quality History of Murrumbidgee River Floodplain Wetlands. Canberra: Land and Water Australia.

25.

Grab

(1996) A note on the morphology of miniature sorted stripes at Mafadi Summit, High Drakensberg. South African Geographical Journal 78(2): 59–63.

26.

Grab

(1997) Analysis and characteristics of high altitude air temperature data from northern Lesotho: Implications for cryogeomorphic occurrences. Geoöko Plus 4: 109–118.

27.

Grab

(1999) Block and debris deposits in the High Drakensberg, Lesotho, southern Africa: Implications for high altitude slope processes. Geografiska Annaler 81(1): 1–16.

28.

Grab

(2000) Stone-banked lobes and environmental implications, High Drakensberg, Southern Africa. Permafrost and Periglacial Processes 11: 177–187.

29.

Grab

(2002a) A note on needle-ice mound formation in the High Drakensberg, southern Africa. Permafrost and Periglacial Processes 13: 315–318.

30.

Grab

(2002b) Palaeoenvironmental significance of relict sorted patterned ground, Drakensberg Plateau, Southern Africa. Quaternary Science Reviews 21: 1729–1744.

31.

Grab

(2004) Thermal regime through a sorted circle and stone-banked lobe, Drakensberg, Southern Africa. Zeitschrift für Geomorphologie 48: 501–518.

32.

Grab

Linde

(in review) Some attributes of snow occurrence and snowmelt/sublimation rates in the Lesotho Highlands: Environmental implications. Water SA.

33.

Grab

Scott

Rossouw

. (2005) Holocene palaeoenvironments inferred from a sedimentary sequence in the Tsoaing River Basin, western Lesotho. Catena 61: 49–62.

34.

Grimm

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

35.

Hanvey

Marker

(1994) Sedimentary sequences in the Tlaeeng Pass area, Lesotho. South African Geographical Journal 76: 63–67.

36.

Harding

Taylor

(2011) The South African Diatom Index (SADI) – A Preliminary Index for Indicating Water Quality in Rivers and Streams in Southern Africa. Water Research Commission Report No. 1701/1/11. Pretoria: WRC.

37.

Hargan

Rühland

Paterson

. (2015) The influence of water-table depth and pH on the spatial distribution of diatom species in peatlands of the Boreal Shield and Hudson Plains, Canada. Botany 93: 57–74.

38.

Harper

(1969) Periglacial evidence in South Africa during the Pleistocene epoch. Palaeoecology of Africa 4: 71–91.

39.

Heiri

Lotter

Lemcke

(2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: Reproducibility and comparability of results. Journal of Paleolimnology 25: 101–110.

40.

Herbert

(1987) Late-Holocene Climatic Change: The Little Ice Age and El Niño from Planktonic Foraminifera in Sediments off Walvis Bay, South West Africa. Bulletin No. 18. Cape Town: Marine Geosciences Unit, Department of Geology, University of Cape Town.

41.

Hogg

Hua

Blackwell

. (2013) SHCal13 Southern Hemisphere calibration, 0–50,000 years cal BP. Radiocarbon 55(4): 1889–1903.

42.

Holmgren

Karlén

Lauritzen

. (1999) A 3000-year high-resolution stalagmite-based record of palaeoclimate for northeastern South Africa. The Holocene 9(3): 295–309.

43.

Holmgren

Lee-Thorp

Cooper

GRJ

. (2003) Persistent millennial-scale climatic variability over the past 25,000 years in Southern Africa. Quaternary Science Reviews 22: 2311–2326.

44.

Inouye

(2008) Effects of climate change on phenology, frost damage and floral abundance of montane wildflowers. Ecology 89(2): 353–362.

45.

Jerardino

(1995) Late-Holocene Neoglacial episodes in southern South America and southern Africa: A comparison. The Holocene 5: 361–368.

46.

Juggins

(2007) C2, Software for ecological and palaeoecological data analysis and visualisation, User guide Version 1.5. Available at: http://www.staff.ncl.ac.uk/stephen.juggins/software/code/C2.pdf (accessed 11 July 2014).

47.

Karst-Riddoch

Pisaric

MFJ

Smol

(2005) Diatom responses to 20th century climate-related environmental changes in high-elevation mountain lakes of the northern Canadian Cordillera. Journal of Paleolimnology 33: 265–282.

48.

Krammer

(2002) Diatoms of Europe, vols 1–6. Ruggel: A.R.G. Gantner Verlag K.G.

49.

Krammer

Lange-Bertalot

(1986) Süßwasserflora von Mitteleuropa, vols 1–5. Jena: VEB Gustav Fischer Verlag.

50.

Lee-Thorp

Holmgren

Lauritzen

. (2001) Rapid climate shifts in the southern African interior throughout the mid to late-Holocene. Geophysical Research Letters 28(23): 4507–4510.

51.

Lotter

Bigler

(2000) Do diatoms in the Swiss Alps reflect the length of ice-cover. Aquatic Sciences 62(2): 125–141.

52.

McGlynn

Mackay

Rose

. (2010) Palaeolimnological evidence of environmental change over the last 400 years in the Rwenzori Mountains of Uganda. Hydrobiologia 648(1): 109–122.

53.

Mackay

Bezrukova

Leng

. (2012) Aquatic ecosystem responses to Holocene climate change and biome development in boreal, central Asia. Quaternary Science Reviews 41: 119–131.

54.

Marcott

Shakun

Clark

. (2013) A reconstruction of regional and global temperature for the past 11,300 years. Science 339: 1198–1201.

55.

Marker

(1994) Sedimentary sequences at Sani Top, Lesotho highlands, Southern Africa. The Holocene 4: 406–412.

56.

Marker

(1995) Late Quaternary environmental implications from sedimentary sequences at two high altitude Lesotho sites. South African Journal of Science 91: 294–298.

57.

Marker

(1998) New radiocarbon dates from Lesotho. South African Journal of Science 94: 239–240.

58.

Masselink

Hughes

Knight

(2014) Introduction to Coastal Processes and Geomorphology. 2nd Edition. Oxon: Routledge.

59.

Matlala

Taylor

Harding

(2011) Development of a Diatom Index for Wetland Health. Water Research Commission Report Number KV 270/11. Gezina: Water Research Commission.

60.

Mayewski

Rohling

Stager

. (2004) Holocene climate variability. Quaternary Research 62(3): 243–255.

61.

Meadows

(1988) Late quaternary peat accumulation in Southern Africa. Catena 15: 459–472.

62.

Metwally

Scott

Neumann

. (2014) Holocene palynology and palaeoenvironments in the Savanna Biome at Tswaing Crater, central South Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 402: 125–135.

63.

Mills

Grab

(2005) Debris ridges along the southern Drakensberg escarpment as evidence for Quaternary glaciation in Southern Africa. Quaternary International 129(1): 61–73.

64.

Mills

Grab

Carr

(2009) Recognition and palaeoclimatic implications of late Quaternary niche glaciation in eastern Lesotho. Journal of Quaternary Science 24: 647–663.

65.

Mills

Grab

Rea

. (2012) Shifting westerlies and precipitation patterns during the Late Pleistocene in southern Africa determined using glacier reconstruction and mass balance modelling. Quaternary Science Reviews 55: 145–159.

66.

Mitchell

(1996) The Late Quaternary of the Lesotho highlands, Southern Africa: Preliminary results and future potential of ongoing research at Sehonghong shelter. Quaternary International 33: 35–43.

67.

Mitchell

Parkington

Wadley

(1998) A tale from three regions: The archaeology of the Pleistocene/Holocene transition in the Western Cape, the Caledon Valley and the Lesotho Highlands, Southern Africa. Quaternary International 50: 105–115.

68.

Mitchell

Parkington

Yates

(1994) Recent Holocene archaeology in western and southern Lesotho. South African Archaeological Bulletin 49: 33–52.

69.

Mitchell

Plug

Bailey

. (2011) Beyond the drip-line: A high-resolution open-air Holocene hunter-gatherer sequence from highland Lesotho. Antiquity 85(330): 1225–1242.

70.

Mucina

Rutherford

(2006) The Vegetation of South Africa, Lesotho and Swaziland. Cape Town: South African National Biodiversity Institute.

71.

Nel

Sumner

(2008) Rainfall and temperature attributes on the Lesotho-Drakensberg escarpment edge, Southern Africa. Geografiska Annaler 90(1): 97–108.

72.

Neumann

Botha

Scott

(2014) 18,000 years of grassland evolution in the summer rainfall region of South Africa: Evidence from Mahwaqa Mountain, KwaZulu-Natal. Vegetation History and Archaeobotany 23(6): 665–681.

73.

Neumann

Scott

Bousman

. (2010) A Holocene sequence of vegetation change at Lake Eteza, coastal KwaZulu-Natal, South Africa. Review of Palaeobotany and Palynology 162: 39–53.

74.

Norström

Neumann

Scott

. (2014) Late Quaternary vegetation dynamics and hydro-climate in the Drakensberg, South Africa. Quaternary Science Reviews 105: 48–65.

75.

Norström

Scott

Partridge

. (2009) Reconstruction of environmental and climate changes at Braamhoek wetland, eastern escarpment South Africa, during the last 16,000 years with emphasis on the Pleistocene-Holocene transition. Palaeogeography, Palaeoclimatology, Palaeoecology 271(3): 240–258.

76.

Nyakatya

McGeoch

(2008) Temperature variation across Marion Island associated with a keystone plant species (Azorella selago Hook. (Apiaceae)). Polar Biology 31: 139–151.

77.

Ohlendorf

Bigler

Goudsmit

G-H

. (2000) Causes and effects of long periods of ice cover on a remote high Alpine lake. Journal of Limnology 59(1): 65–80.

78.

Panizzo

Mackay

Ssemmanda

. (2008) A 140-year record of recent changes in aquatic productivity in a remote, tropical alpine lake in the Rwenzori Mountain National Park, Uganda. Journal of Paleolimnology 40(1): 325–338.

79.

Parker

Lee Thorp

Mitchell

(2011) Late-Holocene Neoglacial conditions from the Lesotho highlands, southern Africa: Phytolith and stable carbon isotope evidence from the archaeological site of Likoaeng. Proceedings of the Geologists’ Association 122: 201–211.

80.

Parmesan

Yohe

(2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421: 37–42.

81.

Patrick

Reimer

(1975) The Diatoms of the United States, Exclusive of Alaska and Hawaii (vols 1–2, Monographs of The Natural Academy of Natural Sciences of Philadelphia, Number 13). Philadelphia, PA: Natural Academy of Natural Sciences of Philadelphia.

82.

Plug

(1993) The macrofaunal and molluscan remains from Tloutle, a Later Stone Age site in Lesotho, Southern Africa. Field Archaeology 2: 44–48.

83.

Roberts

Lee-Thorp

Mitchell

. (2013) Stable carbon isotopic evidence for climate change across the late Pleistocene to early Holocene from Lesotho, southern Africa. Journal of Quaternary Science 28(4): 360–369.

84.

Saarinen

Petterson

(2006) Image analysis techniques. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments: Vol. 2: Physical and Geochemical Methods. Dordrecht: Springer, pp. 23–82.

85.

Schmidt

Kamenik

Lange-Bertalot

. (2004) Fragilaria and Staurosira (Bacillariophyceae) from sediment surfaces of 40 lakes in the Austrian Alps in relation to environmental variables, and their potential for palaeoclimatology. Journal of Limnology 63(2): 171–189.

86.

Schoeman

(1973) A Systematical and Ecological Study of the Diatom Flora of Lesotho with Special Reference to the Water Quality. Pretoria: V&R Printers.

87.

Schoeman

Archibald

REM

(1976) The Diatom Flora of Southern Africa, no. 1–6. Pretoria: Council of Scientific and Industrial Research.

88.

Schwabe

(1995) Alpine mires of the eastern highlands of Lesotho. In: Cowan

(ed.) Wetlands in Southern Africa. Pretoria: Department of Environmental Affairs and Tourism, pp. 33–40.

89.

Scott

Nyakale

(2002) Pollen indications of Holocene palaeoenvironments at Florisbad spring in the central Free State, South Africa. The Holocene 12(4): 497–503.

90.

Scott

Neumann

Brook

. (2012) Terrestrial fossil-pollen evidence of climate change during the last 26 thousand years in Southern Africa. Quaternary Science Reviews 32: 100–118.

91.

Sene

Jones

Meigh

. (1998) Rainfall and flow variations in the Lesotho Highlands. International Journal of Climatology 18: 329–345.

92.

Shala

Helmens

Jansson

. (2014) Palaeoenvironmental record of glacial lake evolution during the Holocene at Sokli, NE Finland. Boreas 43: 362–376.

93.

Sitoe

Risberg

Norström

. (2015) Paleo-environment and flooding of the Limpopo River-plain, Mozambique, between c. AD 1200–2000. Catena 126: 105–116.

94.

Smith

Lee-Thorp

Sealy

(2002) Stable carbon and oxygen isotopic evidence for late Pleistocene and early-middle Holocene climatic fluctuations in the Caledon River Valley, Southern Africa. Journal of Quaternary Science 17: 683–695.

95.

Snoeijs

Balashova

(eds) (1998) Intercalibration and Distribution of Diatom Species in the Baltic Sea, vol. 5. Uppsala: Opulus Press Uppsala.

96.

Stager

Mayewski

White

. (2012) Precipitation variability in the winter rainfall zone of South Africa during the last 1400 yr linked to the austral westerlies. Climate of the past 8: 877–887.

97.

Stewart

Dewar

Morley

. (2012) Afromontane foragers of the Late Pleistocene: Site formation, chronology and occupational pulsing at Melikane Rockshelter, Lesotho. Quaternary International 270: 40–60.

98.

Sundqvist

Holmgren

Fohlmeister

. (2013) Evidence of a large cooling between 1690 and 1740 AD in southern Africa. Scientific Reports 3: 1767.

99.

Talma

Vogel

(1992) Late Quaternary palaeotemperatures derived from a speleothem from Cango Caves, Cape Province, South Africa. Quaternary Research 37: 203–213.

100.

Talma

Vogel

Partridge

(1974) Isotopic contents of some Transvaal speleothems and their palaeoclimatic significance. South African Journal of Science 70: 135–140.

101.

Truc

Chevalier

Favier

. (2013) Quantification of climate change for the last 20,000 years from Wonderkrater, South Africa: Implications for the long-term dynamics of the Intertropical Convergence Zone. Palaeogeography, Palaeoclimatology, Palaeoecology 386: 575–587.

102.

Tyson

Lindesay

(1992) The climate of the last 2000 years in southern Africa. The Holocene 2: 271–278.

103.

Tyson

Karlén

Holmgren

. (2000) The Little Ice Age and medieval warming in South Africa. South African Journal of Science 96: 121–126.

104.

Van Zinderen

Bakker EM

(1955) A preliminary survey of the peat bogs of the Alpine Belt of Northern Basutoland. Acta Geographica 14: 413–422.

105.

Venables

Smith

(2015) An introduction to R, notes on R: A programming environment for data analysis and graphics version 3.1.3 (9 March 2015). Available at: https://www.cran.r-project.org/manuals/r-release/R-intro.pdf (accessed 14 December 2014).

106.

Wang

Mackay

Leng

. (2013) Influence of the ratio of planktonic to benthic diatoms on lacustrine organic matter δ¹³C from Erlongwan maar lake, northeast China. Organic Geochemistry 54: 62–68.

107.

Wanner

Beer

Bütikofer

. (2008) Mid- to Late-Holocene climate change: An overview. Quaternary Science Reviews 27: 1791–1828.

108.

Wanner

Mercolli

Grosjean

. (2015) Holocene climate variability and change: a data-based review. Journal of the Geological Society 172: 254–263.

109.

Weldeab

Stuut

JBW

Schneider

. (2013) Holocene climate variability in the winter rainfall zone of South Africa. Climate of the past 9: 2347–2364.

110.

Westover

Fritz

Blyakharchuk

. (2006) Diatom palaeolimnological record of Holocene climatic and environmental change in the Altai Mountains, Siberia. Journal of Paleolimnology 35: 519–541.

111.

Zinke

Loveday

Reason

CJC

. (2014) Madagascar corals track sea surface temperature variability in the Agulhas Current core region over the past 334 years. Scientific Reports 4: 4393–4401.