Palustrine wetlands, which cover about 6% of the earth’s land surface, are among the most ecologically sensitive ecosystems and very important globally because of their unique role in biogeochemical cycles (Junk et al. 2013; Mitsch & Gosselink 2015). Because they support azonal vegetation that is distinct from the surrounding vegetation (Mucina & Rutherford 2006; Sieben et al. 2016), wetlands are ecological ‘islands’ within terrestrial environments in different landscapes across the globe. The distinction results from the prolonged water logging that causes oxygen deficiency (hypoxia) or its total absence (anoxia) in the wetland soil, with subsequent chemical changes in soil characteristics (Gopal 2015; Mitsch & Gosselink 2015). Mucina and Rutherford (2006) observed that the presence of water, whether seasonal or permanent, is the primary factor in creating wetland habitats and associated vegetation. Nonetheless, it is not wetness per se that primarily influences the geochemistry and morphology of wetland soils, but rather the anaerobic conditions that result from prolonged soil saturation or flooding (Collins 2005; Kotze et al. 1996).

Montane palustrine wetlands are a special sub-division of freshwater palustrine wetlands that are embedded within terrestrial ecosystems in mountain areas (Mucina & Rutherford 2006). These wetlands are often rich in endemics because many species remain isolated in high-altitude habitats (Junk et al. 2013). By providing a wide spectrum of ecosystem services, these ecosystems are of high ecological and socio-economic importance (Chatterjee et al. 2010). High-altitude wetlands provide essential ecosystem services which help to sustain human life, conserve biodiversity, preserve the hydrological regime and combat the impacts of climate change and are thereby key to water, food and ecological security (Singha 2011). Montane wetlands are located in the headwaters of major river basins, providing water for many trans-boundary rivers and play an important role in the ecology and hydrology of the local environment and downstream systems. They also play a fundamental role in the overall global and regional water cycles (Russi et al. 2013; Sieben, Mtshali & Janks 2014).

The montane wetlands of Lesotho are important for the supply of ecosystem services, including biodiversity conservation, water resources, livestock grazing, harvestable plant products and environmental regulation. Lesotho forms an important hydrological reservoir and watershed for several countries in Southern Africa (Nüsser & Grab 2002). For example, the montane wetlands in the country form part of the headwaters of one of the most important international watercourses in Southern Africa, the Senqu-Orange River (ORASECOM 2015), which is not only important for Lesotho, but also for South Africa and Namibia. Lesotho falls entirely within the catchment of this river system (ORASECOM 2015) and the country is drained by the river and its tributaries. Moreover, the high-altitude wetlands of Lesotho are critical in maintaining the water levels not only in dams supporting the Lesotho Highlands Water Project (LHWP), which transfers water to South Africa (Grab & Deschamps 2004), but also in those supplying the local population. Directly and indirectly, wetlands in Lesotho are estimated to contribute about 22% of the country’s gross domestic products (GDPs) and 30% of the total employment in the country (Department of Environment 2014). Furthermore, by providing freshwater that is transferred to the most densely populated and industrialised area of Gauteng, the wetlands in Lesotho also play a key role in the South African economy.

The most ubiquitous and conspicuous feature of the wetland environment, which also plays a critical role in wetland ecosystem functioning, is the vegetation (Cronk & Fennessy 2001; Gopal 2016). As a result, the description of wetland habitats based on vegetation and the associated classifications are useful in strategic planning for the conservation of wetlands (Mitsch & Gosselink 2015). Plants are such a critical component of wetlands that they are the most widely supported biotic indicator for assessing wetland condition (Collins 2005). Accordingly, vegetation is the most suitable feature to consider over a broad range of wetland habitats (Sieben et al. 2014). Within a single wetland, a large diversity of habitats may occur, and as a result, large differences in the types of vegetation may be observed in the same wetland (Mitsch & Gosselink 2000). Therefore, wetland habitats can be classified based on the plant communities in the system (Collins 2005; Sieben, Kotze & Morris 2010a).

Plant species occurring in a wetland can be used as indicators of environmental conditions and ecological changes taking place in the system (Mitsch & Gosselink 2015) because wetland vegetation is azonal (Brand, Du Preez & Brown 2013) and responds quickly to local environmental changes (Cronk & Fennessy 2001). Because the plant community gives a characterisation of habitat units within a wetland and also serves as habitat for associated animals, a detailed wetland vegetation description for a given wetland can be used as a proxy for the system’s biodiversity in general, highlighting the conservation value of the wetland (Sieben et al. 2016). Accordingly, the determination of wetland plants that are useful as biological indicators has become increasingly important for use in monitoring the integrity of specific wetlands (Sieben et al. 2014). Wetland typology provides useful information for water resource management and biodiversity conservation (Sieben et al. 2014).

Given the importance of wetlands, recent studies (Moor et al. 2017; Sieben et al. 2016) have emphasised the need for more information about their species composition, ecology and distribution. Although many studies have been carried out in the wetlands of surrounding South Africa (e.g. Brand et al. 2013; Sieben et al. 2014, 2016), no recent detailed countrywide survey has been conducted on the wetland vegetation of Lesotho. The current study is the first to provide a characterisation of the wetland vegetation in detail, covering many wetlands in the country.

The aims of this study for the palustrine wetlands of Lesotho are:

Overall, 150 vegetation plots from 30 palustrine wetlands in Lesotho were analysed in this study (Figure 1). A total of 312 plant species belonging to 51 families occurred in the wetlands (Online Appendix 1 – Table 1-A1[a] and [b]). Of the 312 species, 276 were encountered in the 150 plots and the remaining 36 species occurred within the wetlands but outside the plots (Online Appendix 1 – Table 1-A1[a]; Online Appendix 2 – Table 1-A2). Five most dominant families, accounting for most of the species occurring in the wetlands, were Poaceae (20.19%), Asteraceae (19.23%), Cyperaceae (14.10%), Scrophulariaceae (4.17%) and Polygonaceae (3.85%) (Online Appendix 1 – Table 1-A1 [b]). From these results, it can be observed that three (Poaceae, Asteraceae and Cyperaceae) out of the 51 families account for 53.52% of all the species occurring in these wetlands. While the species richness of whole wetland units ranged between 10 and 53, the number of species per 3 × 3 m plot ranged from 1 (monospecific communities, e.g. community 19) to 27 (highly diverse communities, e.g. communities 3 and 4), with a median species richness of about 14. The highest median number of species for whole wetland units (41.5) was recorded in the Highlands and the lowest (26) in the Lowlands. The five most frequently occurring species in the wetlands all represent different families (Ranunculaceae, Leguminosae, Asteraceae, Poaceae and Cyperaceae) (Online Appendix 1 – Table 1-A1[a] and [b]). The type of metabolism (C₃/C₄) is also indicated in Online Appendix 1 – Table 1-A1[a] for some of the species. The species that were collected during fieldwork were deposited in the ROML at the National University of Lesotho.

The earliest studies on the palustrine wetlands of Lesotho were carried out in the early 1960s and 1970s (Guillarmod 1962; Van Zinderen Bakker & Werger 1974). Since then, other studies conducted in the country (Du Preez & Brown 2011; Grab & Deschamps 2004; Meakins & Duckett 1993) were mainly carried out on small areas, focusing on specific or a few wetlands. The current study has provided the most recent and more comprehensive assessment of the palustrine wetlands in the country by providing a classification and description of the wetland vegetation. The wetland vegetation has been classified into 22 communities, which are influenced mainly by altitude, longitude, slope, soil parent material, landscape, inundation, peat, soil potassium, clay, total organic carbon, nitrogen, sulphur, electrical conductivity, calcium, soil depth, wetness, magnesium, aspect and latitude. The wetlands were found to be occurring mainly in those areas where topography allows wetland conditions to develop (mostly in flat areas); mainly on the summit plateau and in the Lowlands. Nonetheless, it is noteworthy that there may be more communities in the country than the number reported here because only 150 plots from 30 wetlands were analysed in this study, although this forms an important starting point for documenting and classifying the wetland vegetation for Lesotho.

Despite Lesotho being entirely Afromontane, the study found a clear distinction between wetlands that are found in the Highlands and those in the Lowlands, in terms of diversity, species making up the communities and structure of the communities (Table 3). While the highland wetlands (types 1–4) are mainly south-facing, those in the Lowlands (type 5) are mainly north-facing. Differences exist between the north- and south-facing slopes because of distinct solar radiation patterns. The resulting ecological differences that include snow cover duration and moisture conditions between the moister and colder south-facing slopes and the drier and warmer north-facing slopes contribute to the observed differences in vegetation types (Nüsser 2002). Indicators of wetland type 5 are typical lowland wetland plants. This wetland type is comparable to the most widespread and common type of wetlands in South Africa, the temperate grassy wetlands (Sieben et al. 2014). One of the highland wetland types (type 4) comprises wetlands that are found in the small high-altitude area underlain by sandstone and limited to the eastern edge of Lesotho (Sehlabathebe National Park). These wetlands are also located on very steep slopes. However, this wetland type was under-sampled in this study as it was represented by only two wetlands and nine vegetation plots. Despite the country’s high altitude (1388–3482 m asl) and rugged terrain, qualifying it to be entirely Afromontane (Carbutt & Edwards 2015), some of the communities in the Lowlands also fit into the temperate grassy wetland vegetation of Sieben et al. (2017a) and the eastern temperate freshwater wetlands of Mucina and Rutherford (2006).

All the wetlands surveyed in the current study can broadly be classified as the freshwater wetland vegetation type of Mucina and Rutherford (2006), which is further divided into eastern temperate freshwater wetlands (AZf 3), Drakensberg wetlands (AZf 4) and Lesotho mires (AZf 5). The dominance of Poaceae, Asteraceae, Cyperaceae and Scrophulariaceae families observed in the current study has also been reported in the entire Maloti-Drakensberg region and the surrounding areas, although Asteraceae was the most dominant in the latter (Chatanga et al. 2019). The current study also concurs with Sieben, Glen and Muasya (2017b) who report that Poaceae is the most common plant family in the South African wetlands based on species richness. However, of the five most dominant plant families recorded in the current study, two (Asteraceae and Scrophulariaceae) have higher than average levels of endemism in the Maloti-Drakensberg region (Cowling & Hilton-Taylor 1994). High levels of endemism in the high-altitude environments such as the Maloti-Drakensberg are attributed to the many small-scale microhabitats and a wide range of edaphic conditions that develop as a result of slope and aspect (Brand, Scott-Shaw & O’Connor 2019).

The five most common species in the study all represent different families. This implies that the wetland vegetation in the country is phylogenetically diverse and this is unlike the situation in South Africa where the five most common wetland plant species are all grasses (Poaceae) (Sieben et al. 2014). The high altitude that characterises the greater part of Lesotho could account for this high diversity, and montane wetlands have been identified as some of the most species-rich in South Africa (Sieben et al. 2014). While the high-altitude montane wetland communities in Lesotho are more diverse and are virtually never a monoculture, the lowland wetland communities are generally less diverse and sometimes monospecific communities. In lowland wetlands, it is common for just one or two species to dominate the entire wetland plant community (Boutin & Keddy 1993; Sieben et al. 2010a). This trend, also exhibited in the current study, is contrary to the decline in species richness with altitude, which has widely been recognised as a general law of ecology (Rosenzweig 1995). However, the large number of species and communities recorded in this study generally reflects the high diversity of wetland habitats in the country, although the highest diversity is associated with higher altitudes. It is noteworthy that exceptions such as Kniphofia caulescens and Carex cognata communities usually exhibit lower diversity than other communities at such higher elevations. This suggests that the dominant species in these communities are so adapted to the harsh high-altitude conditions that they have developed a very large competitive effect on local resources and act as a habitat filter by excluding the less competitive species (Maire et al. 2012). It is also noteworthy that the Lowlands is the area subjected to more anthropogenic pressures that include cultivation, urbanisation and conversion to other forms of land use and this threatens the wetlands in this part of the country.

The higher species richness and diversity in the Highlands highlights that these wetland habitats are more diverse than those in the Lowlands. This could be attributable, in part, to the fact that most of the dominant species in the high-altitude montane wetlands of the country are non-clonal (Table 3). Clonal plants are species that can reproduce vegetatively by sprouting from stolons or rhizomes, forming stands of individuals of that species (Song & Dong 2002). Typical clonal plants include Phragmites and Typha spp. that form dense rhizome mats and outcompete other non-clonal species. These findings corroborate Sieben et al. (2010b, 2017b) who suggest that, unlike lowland wetlands, high-altitude wetlands are unusual in that they are richer in species and particularly non-clonal species. Moreover, because the usually dominant wetland plants cannot cope well with the low temperatures characterising high-altitude environments, they cannot be as dominant as usual, leaving many vacant niches that then become available for colonisation by other plants (Sieben et al. 2010b). Furthermore, in wetland environments, the abundance of clonal plants has been reported to be negatively associated with the overall plant species diversity, as well as with altitude (Song & Dong 2002). The high diversity can also be attributed to the steep gradients in the landscape and harsh climatic conditions that create unique habitats in the mountains of Lesotho (Pooley 2003). The high diversity can further be ascribed to the underlying basaltic parent material. Mucina and Rutherford (2006) observe that the seepage water in these high-altitude wetlands is eutrophic as the underlying basalt is rich in nutrients. This also influences the floristic composition and community structure of the wetlands. These results are consistent with observations by Sieben et al. (2010a) who acknowledge the significant number of wetland community types in the montane areas of Lesotho. The high floristic diversity observed in the current study is also consistent with findings from the high-altitude montane wetlands of Alborz Mountains, Iran (Kamrani et al. 2011; Naqinezhad et al. 2009).

The ordination of the vegetation data for all plots reveals that the explanatory environmental variables supplied could explain only about 18.11% of the total variation. This highlights that the remaining variation could be explained by the environmental factors which were not included in the analysis, such as precipitation (assessed only indirectly through longitude) and wetland water quality. It may also be that plants colonise wetland habitats by chance (Chesson 2000). However, the degree of variation explained in the ordination is comparable to similar studies in the wetlands of South Africa (Sieben et al. 2016, 2017a). Moreover, Brand et al. (2013) highlight that substrate and hydrogeological conditions play a bigger role in influencing the floristic composition, structure and dynamics in high-altitude montane wetlands than microclimate. The inclusion of soil variables in the analysis increased the proportion of the total variation explained by the supplied variables from 18.11% to 65.40%. Thus, the inclusion of soil data in the analysis significantly improves wetland vegetation–environment assessments. The degree of variation explained in the ordination with soil data is significantly higher than that reported for the South African wetlands (Sieben et al. 2016, 2017a). Nevertheless, the amount of variation explained after the inclusion of soil data could also have improved because the data set was smaller (30 vegetation plots).

While the second axis is positively associated with slope, it is negatively correlated with the wetness and soil depth. This implies that communities on the lower part of the ordination diagram are associated with wetter habitats with deeper soils, while those on the upper part are associated with steeper slopes. The former include communities such as C. cognata–Juncus effusus, while the latter include communities such as Merxmuellera macowanii, which are often associated with hillslope seepages. The correlation between wetness and soil depth, which are both negatively correlated with slope, could be attributed to the soil deposition and water accumulation that is often consistent with fairly flat habitats. Altitude, longitude, slope, wetness, soil parent material, landscape, peat, inundation, aspect, soil depth and wetness have been observed to be very important factors explaining the distribution and variation of the wetland vegetation in this study.

Abundance of peat is also strongly correlated with altitude and longitude. Because most of these Afromontane wetlands are often located on a slope at high altitudes, they are unique (Mucina & Rutherford 2006; Sieben et al. 2014). A temperature drop of 1 °C has been estimated for every 125 m gain in altitude in Lesotho and the Maloti-Drakensberg region (Pomela et al. 2000). Such steep environmental gradients over short distances (Körner, Paulsen & Spehn 2011) in Afromontane areas are associated with huge spatial variation in physical features and this results in remarkable variation in terms of species diversity and distribution (Kotze & O’Connor 2000). As a result, increasing altitude corresponds with a decrease in temperature and an increase in rainfall (Mucina & Rutherford 2006) and the greater part of Lesotho is generally much higher and colder than the surrounding areas (Sieben et al. 2014). The hypoxia or anoxia in the wetlands, coupled with the low temperature and pH, often associated with these high-altitude wetlands, reduces the rate of decomposition and favours the accumulation of organic matter and peat formation (Chatterjee et al. 2010; Gopal 2016). Therefore, lower temperatures, higher rainfall and other environmental conditions associated with high altitudes also create habitats that can harbour unique vegetation (Sieben et al. 2014).

Through its influence on temperature and rainfall, altitude is a suitable surrogate measure for climate in Lesotho and the Maloti-Drakensberg region, which represents an indirect gradient (Sieben et al. 2010a). The strong correlation between longitude and altitude can therefore be explained by the fact that altitude generally increases on moving from west to east in Lesotho and rainfall also increases with altitude and with increasing longitude (Cowling & Hilton-Taylor 1994; Mucina & Rutherford 2006). The high altitudes of Lesotho are mainly associated with abundant orographic rainfall, which results in many springs and seepage zones (Mucina & Rutherford 2006; Sieben et al. 2014). The influence of altitude and wetness on high-altitude wetland vegetation has also been reported in South Africa (Kotze & O’Connor 2000; Mucina & Rutherford 2006; Sieben et al. 2010a, 2010b), southern Brazil (Rolon & Maltchik 2006) and in Cumbria, UK (Jones, Li & Maberly 2003). The importance of both altitude and slope gradients on the floristic composition of high-altitude montane wetlands has also been reported in Bulgaria, south-eastern Europe (Hájková, Hájek & Apostolova 2006) and in the Alborz Mountains, India (Kamrani et al. 2011; Naqinezhad et al. 2009).

Although a negative correlation between altitude and wetness was observed in this study, altitude operates at a larger scale, whereas wetness operates on a local scale. It is nevertheless noteworthy that, while some communities are restricted to either the Highlands or Lowlands, others seem to have a wide ecological amplitude. These include Eragrostis plana–Pennisetum sphacelatum, Eleocharis dregeana and Eragrostis planiculmis communities. Furthermore, some species, such as Ranunculus meyeri Harv., have been reported to occur at low abundances at lower altitudes but achieve greater cover at higher altitudes as they gain a competitive advantage because of the lower temperatures that reduce the vigour of the usually more competitive species (Sieben et al. 2010b). Because some of the plant communities recorded in this study are restricted to the highest altitudes, occurring at the summit plateaus, they are likely to disappear in the face of climate change as they cannot migrate any further up (Bentley, Robertson & Barker 2019; Lee et al. 2015).

While high-altitude communities (e.g. Limosella grandiflora–Haplocarpha nervosa, Agrostis bergiana and Cotula paludosa– R. meyeri communities) were also associated with soils that are high in total organic carbon, nitrogen and sulphur, the habitat soils for the low-altitude communities (e.g. Cynodon incompletus, Eleocharis limosa, S. paludicola and E. dregeana communities) were high in potassium and clay percentage. However, the negative correlation of potassium with altitude is contrary to Mucina and Rutherford (2006) who report that the high-altitude areas are associated with high potassium levels. In the current study, communities such as C. cognata, Gunnera perpensa and E. plana–P. sphacelatum were also found to be associated with high soil magnesium, calcium and electrical conductivity. The importance of potassium, electrical conductivity and clay percentage in influencing high-altitude montane wetland vegetation is consistent with findings from a similar study on the wetlands of the Alborz Mountains, India (Kamrani et al. 2011). Because of the low temperature and basaltic parent material, the soils in the mountains of Lesotho are high in organic matter and generally nutrient-rich, with a high water-retention capacity (Mucina & Rutherford 2006). However, the soils found in the Lowlands are considered to be generally nutrient-poor (Maro 2011).

Given their important role in water resources, livestock grazing and harbouring rare and endemic species, as well as unique biodiversity, the wetlands described in the current study are of high conservation value in Southern Africa, particularly in the face of increased water scarcity, climate change and unprecedented loss of biodiversity. Accordingly, the wetlands of Lesotho play a critical role in providing ecosystem services at the local, national and regional scales. As a result, the government of Lesotho and the Orange-Senqu River Commission (ORASECOM) have a responsibility in terms of scaling-up the efforts towards the protection of these wetlands for the sustainable provision of the benefits obtained from them. Because the wetland vegetation of a particular wetland can be used as a proxy for biodiversity of the wetland, understanding the wetland vegetation–environment patterns is important for the successful conservation planning of these ecosystems. Therefore, the current study has provided baseline information, which can be useful for monitoring the wetland vegetation and concomitantly the wetlands, which are vital for the water resources of Lesotho, South Africa and Namibia.