Rudd’s Lark

Even within its restricted range, this species is limited to a very specific habitat, determined by a range of biotic, altitudinal, topographic and climatic variables; its distribution is therefore naturally fragmented and highly localised, both temporally and spatially (Barnes 2000). This produces challenges for accurately assessing habitat availability. Formative work on the species in its historical core range, indicated a preference for sour grassland on flat or gently sloping plateaus and hilltops at altitudes of 1700–2200 m (mostly >1800 m; Maphisa et al. 2009) in areas with high rainfall, typically >600 mm per year (Tarboton et al. 1987). However, the detection of Rudd’s Lark at lower altitude sites in the Eastern Cape (e.g. Molteno, Cathcart, Ncora Dam; W Rossouw in litt 2024) in years with above average rainfall suggests local movements based on rainfall regimes and a reliance on a wider array of grassland habitat during drier periods. Rudd’s Larks also occur at the edges of pans and vleis (Tarboton et al. 1987), but avoid valley bottoms and lowlands, as well as dense cover, old croplands, ploughed fields, rocky areas, valley bottoms, lowlands, intensely modified grasslands, and slopes of >10° (Maphisa 2004, Gush 2019).

Based on Low and Rebelo’s 1996 vegetation map, Rudd’s Lark was virtually restricted to Sandy Highveld Grassland. Dominant grass species within its preferred habitat included Themeda triandra, Tristachya leucothrix, Trachypogon spicatus, Heteropogon contortus and Eragrostis curvula (Hockey et al. 1988). In optimal habitats, the cover of Eragrostis curvula/E. chloromelas was high relative to Themeda triandra (Hockey et al. 1988). Updates to the vegetation map of South Africa by Mucina and Rutherford (2006) show that the Sandy Highveld grassland historically occupied by Rudd’s Lark may be broken down into several sub-types including Low Escarpment Moist Grassland, Northern KZN Moist Grassland, East Free State Sandy Grassland, Amersfoort Highveld Clay Grassland, Wakkerstroom Montane Grassland and Lydenburg Montane Grassland.

Rudd’s Lark typically breeds in sites with moderately to heavily grazed grass with short tussocks (Hockey et al. 1988), selecting sites with shorter than average grass height early in the breeding season (Gush et al. 2019). Hockey et al. (1988) noted that the species is often most numerous in intensively grazed grasslands, usually with trampled or bare patches, which suggests that high grazing pressure does not necessarily affect the birds adversely, but Maphisa (2009) contended that few breeding attempts occur in these heavily grazed sites, possibly due to higher predation rates. Indeed, surveys in the Wakkerstroom area reported that birds were absent from the most heavily grazed sites, and that an increase in the extent of bare ground, possibly due to excessive grazing, may lead to territory desertion (Maphisa et al. 2009, Gush et al. 2019). Conversely, territories may also be abandoned if grass becomes too tall and dense. If some open ground for terrestrial foraging remains, birds may remain in grassland up to 40 cm in height (Maphisa 2004). An absence of forbs has been suggested as an important habitat requirement (Hockey et al. 1988), but pairs do occasionally nest in sites where forbs occur (Maphisa 2004); nevertheless, the presence of forbs (such as spinescent forb Acalypha spp.) may be a sign of deteriorating and overgrazed grasslands (Maphisa 2004). A recent study found the highest densities recorded in management units with mixed livestock (cattle and sheep) or sheep dominant grazing regimes, winter cattle grazing regimes, and/or stringent rotational grazing regimes (Colyn et al. 2024). No study units with persistent year-round cattle grazing and annually burnt grasslands yielded any breeding records. Exclusive cattle grazing for extended periods can lower grass cover and increase forb composition (Toth et al. 2016), thereby possibly reducing habitat suitability for Rudd’s Lark (Colyn et al. 2024). Conversely, sheep-dominant grazing regimes and/or mixed livestock grazing lower forb composition while increasing grass composition, percentage organic matter, and basal cover (Abaye et al. 1997) which in turn favours Rudd’s Lark habitat suitability (Colyn et al. 2024). Between 2007 and 2017, the South African government reported a 28% decrease in sheep farming, coupled with an 11% and 8% increase in cattle and mixed (crops and cattle) farming respectively (Census of Commercial Agriculture 2017). Initial surveys in the Wakkerstroom area in the late 1980s (Tarboton et al. 1987) reported sheep as the predominant livestock farmed, while recent surveys only recorded sheep on 5% of study sites in the area. Livestock type may indeed play a more important role in habitat suitability for the species than grazing intensity and requires further investigation (Colyn et al. 2024).

Suitable breeding habitat was thought to be maintained by annual burning and heavy winter grazing (Hockey et al. 1988); the latter was likely sustained in the pre-agricultural past by large herds of grazing ungulates, but today mostly by domesticated livestock (Maphisa et al. 2009). As nesting attempts appear to be clustered in the second half of the breeding season (with egg-laying occurring between October and April, with peaks in January to February), late burning of grasslands may shorten the breeding season, and could account for population decreases (Maphisa et al. 2009). Maphisa et al. (2009) suggested that this species does best in grassland burned before the start of the breeding season, and indeed adult birds have been found in recently burned sites (W Rossouw in litt 2024). However, the species has also been detected at sites that are never burned, and grass length is maintained instead by grazers (Colyn et al. 2024). Primary defoliation through the use of fire, as opposed to functional grazing regimes, decreases habitat suitability (He et al. 2019), which may have been the driver of the absences recorded in recent surveys (2017-2020) at numerous (>10) annually burnt local study sites in Wakkerstroom, Memel and Amersfoort.

Habitat suitability modelling from transect data collected across the species’ range suggests that larger intact grassland patches (>100 ha) that incorporate well managed rotational grazing regimes with lower fire prevalence and higher basal cover increased the probability of species presence (Colyn et al. 2024). This extensive habitat specialisation, combined with extensive (30%) historic habitat loss, has resulted in a highly fragmented range with <1000 km² (40%) of highland grassland within the species’ range being suitable (Colyn et al. 2024). Habitat use and suitability within the recently discovered sites and peripheral areas of the Eastern Cape require further study. The more recent surveys in the Dordrecht area yielded sightings of birds utilising a tame (artificially seeded) pasture, which is a first for the species. If grassland structure and composition requirements can be mimicked within tame pastures and meet the needs of the species, it can become a significant conservation/rehabilitation tool used to promote the persistence of Rudd’s Lark (and other species) by alleviating the primary threat, namely habitat loss and degradation.