The site is a granite catena of third-order stream in the Southern Granite supersite (Smit et al. 2013) in the Kruger National Park, South Africa (Figure 1). The mean annual precipitation is 560 mm/a (Smit et al. 2013) and the geology comprises granites and gneissic migmatites of the Nelspruit suite (Dippenaar & Van Rooy 2014; Robb et al. 2006).

The crest to the midslope is associated with gentle gradients and comprises apedal soils frequented by Terminalia sericea (Riddell et al. 2014). In the lower midslope positions, there is an abrupt transition to sodic duplex soils. Woody vegetation dominates the crest (Combretum apiculatum and Combretum zeyheri), whilst Acacia nilotica dominates the lower slopes and Euclea divinorium is associated with the duplex sodic soils (Smit et al. 2013).

As part of a greater study, botanists laid out 49 plots of 10 m² from the crest of the hillslope to the valley bottom at 10-m intervals. Soil observations were made to refusal using a hand auger in the centre of these plots, effectively creating a 10-m-interval transect. The soils were classified according to the Soil Classification Working Group (SCWG 1991) and the International Union of Soil Sciences Working Group World Reference Base (WRB 2015). The soil properties and their spatial distribution were interpreted and related to their hydropedological significance and reclassified into hydrological soil types (Van Tol & Le Roux 2019).

Soil samples were collected from the top and subsoil horizons of 49 observations. Four modal profiles were identified based on terrain position and soil type. These modal profiles were located on the crest, upper footslope, lower footslope and valley bottom terrain positions (Figure 2). These profiles were sampled at 100mmdepth intervals.

The samples were dried, sieved and analysed according to standard methods (The Non-Affiliated Soil Analysis Work Committee 1990). Samples were analysed for particle distribution using the pipette method and pH (H₂O and KCl) with 1:2.5 solution. The soluble and exchangeable cations and cation exchange capacity were determined using NH₄OAc solution.

Selected laboratory analyses were presented as average values of the entire profiles for 49 observations as well as the vertical distribution of selected elements for the modal profiles. The S-value was used to express the total content of basic cations (expressed as equivalent values): S−value=Σ(Na, Ca, Mg, K)[Eqn. 1] and the exchangeable sodium percentage (ESP) was calculated as follows: ESP=NaS−value×100[Eqn. 2]

Soil hydraulic properties were determined on three profiles representing diagnostic horizons dominating the upslope, midslope and valley bottom positions. The saturated hydraulic conductivity of each diagnostic horizon in the modal profiles was measured in duplicate using the modified Bouwer and Rice (1976) double-ring falling-head method: KS=Lt×ln((h0+L)(h1+L))[Eqn. 3] where K_s is the saturated hydraulic conductivity (mm h⁻¹), L is the thickness of the horizon, t is the time until constant infiltration and h₀ and h₁ are the heads of water above the surface before and after t respectively.

Water retention characteristics were measured for each horizon of the modal profiles on undisturbed core samples. The samples were de-aired at −70 kPa at room temperature for 24 h. Water was then allowed to gradually flow out of the chambers. The saturated samples were mounted on a hanging water column setup in accordance with the procedure shown by Dirksen (1999). The suction levels (h) were set at 0 mm, 38 mm, 50 mm, 100 mm, 200 mm, 400 mm, 600 mm and 800 mm. The gravimetric water content was determined at each level. The samples were then oven-dried at 105 °C to determine the bulk density to calculate the volumetric water content (θ_v).

Ethical approval to conduct the study was obtained from the Interfaculty Animal Ethics Committee of the University of the Free State (reference No. UFS-AED2019/0121).

The soil types range from sandy and apedal soils on the crest to strongly structured soils on the footslope with alluvial deposits on the banks on the river (Figure 2). The soil forms and their associated hydropedological grouping are presented in Table 1. Clovelly soils dominate the crest position. There is a small sodic area, covered by Sterkspuit soils on the transition from the crest to the midslope position. Fernwood and Pinedene (Arenosols) soil forms are found on the midslope. These soils are relatively deep (>750 mm) when compared to the soils on the footslope. The upper part of the footslope is dominated by the Sterkspruit (Solonetz) soils. The change in soil type (from sandy to clayey duplex) is correlated with a change in vegetation, that is, from woody vegetation to small shrubs. Topsoil horizons of these duplex soils are shallow and often eroded completely (Table 1). The lower footslope position is characterised by the pedocutanic horizons of the Bonheim (Luvisol) having a strongly developed blocky structure. The textural discontinuity is, however, not as dominant in these soils as in the Sterkspruit soils. Dundee soil (Fluvisol) is found along the streambeds (valley bottom). These soils are characterised by alluvial layering.

At the crest and midslope, the soil morphology of the sandy Clovelly and Mispah forms indicates that soil weathering of granitic parent material is the dominant soil-forming process. Horizonisation is visible as yellowing of the B horizon and slight darkening of the orthic horizon because of accumulation of organic matter. The low clay contents indicate that silicate clay, which is formed during weathering of the granite, is eluviated (Figures 3b and 4a). The yellow colour of the Clovelly soils is related to the presence of iron oxides, especially Goethite, which are stable and do not occur in soils saturated for significant periods (Schwertmann 1985). Hydrologically, it implies that a perched water table is not formed in these soils, and deep drainage into the saprolite and fracture rock is the dominant process.

The occurence of a bleached (grey colours as defined in SCWG 1991) B2 horizon in the Pinedene soil form (Table 1 and Figure 2) in the lower-midslope indicates redoximorphic process in the deep subsoil. Bleaching is dependent on reduction associated with energy supplied to microbes in the form of organic matter and prolonged saturation (Vepraskas, Huffman & Kreiser 2006). In environments subjected to high evaporative demands, prolonged saturation requires that water is fed to the soil through a sustained flowpath of water. In this hillslope, this implies that flow at the soil/bedrock interface could occur in the saprolite horizon of the Clovelly soil or that the water which infiltrated into the fractured rock below the Clovelly and Mispah soils returns to the solum under the influence of gravitation to the Pinedene soil downslope. These indicators are supported by the occurrence of a Fernwood soil below the Pinedene soil. In simple terms, the bleached horizon at the soil/bedrock interface of the Pinedene soil becomes deeper downslope until the entire subsoil is bleached.

The lateral flowing water continues to accumulate until it reaches the prismatcutanic horizon in the Estcourt soil. Estcourt soils are known for the generation of lateral flow (Van Tol et al. 2013). The prismacutanic horizon acts as a clay plug and governs the lateral flowing water to return to the surface, forming a seepline. The grey colour of the E-horizon of the Estcourt soil further supports this lateral flowpath. The grey colour is the result of removal of iron oxide under reducing (saturated) conditions (Schwertmann 1985; Van Huyssteen et al. 2007). Topsoil erosion, high clay contents and the presence of prismacutanic horizons with strong structure signify that overland flow is the dominant hydrological flowpath in this part of the hillslope. Soil morphology indicates soil formation by clay enrichment due to either, or a combination of, physical (illuviation) or chemical (neo-formation) processes (Owliaie, Abtahi & Heck 2006).

The Bonheim and Milkwood soils of the lower-footslope indicate that weathering is more prominent than in the upper-footslope. Angular blocky structured pedocutanic horizons are in a more advanced stage of weathering than the primary prismatic and secondary angular blocky structures of the prismacutanic soils in the upper-footslopes. The dominant flowpath is still overland flow towards the valley bottom with limited vertical flow. This is expected because of the high clay contents and strong structure of the melanic and pedocutanic horizons. Dundee soils occurring in the valley bottom/riverbed are formed because of alluvial deposits from upstream. The coarse texture of the Dundee soils serves as an area of infiltration and recharge of overland flow from upslope positions as well as during streamflow episodes, that is, the stream is losing water to the groundwater.

The distribution of the soils in the catena indicates four distinctly different zones of soil formation related to different sets of hillslope flowpaths and storage mechanisms. Firstly, soils on the crest indicate soil formation by weathering of granitic parent material with a dominant vertical flowpath recharging the underlying fractured rock. Secondly, in the midslope position, soil morphology indicates subsoil saturation and associated reducing conditions. This results in gleysation and bleaching at the soil/bedrock interface because of removal of iron oxides. The dominant flowpaths are interflow at the soil/bedrock interface which returns to the surface to form a hillslope seep-line. Thirdly, on the footslope, soil morphology indicates soil degradation in the form of water erosion. Overland flow dominates these soils and leaching is subdominant. The strong prismatic structure of the soils of the upper-footslope is an indication that solonisation (accumulation of Na) dominates. In the lower-footslope, the weathering of the prismatic structure into angular blocky structure is because of slightly wetter moisture regimes. Lastly, in the valley bottom, recharge occurs in coarse materials that were deposited during flood events.

The change in subsoil chemistry from the crest to the valley bottom is correlated to changes in soil types (Figures 2 and 3). Low concentrations of base cations and low pH in the apedal soils at the crest and midslope positions (plots 1 to 7) indicate that eluviation is a dominant process. The almost constant low S-value (Figure 4c) and soluble Ca, Mg and Na cations (Figure 5) at different depths in the Clovelly soil support the deduction made from morphological properties, that is, the vertical flow through and out of these profiles is dominant. Bouwer et al. (2015) found similar trends in freely drained soils in the Weatherley research catchment, South Africa.

At plot 8, there was a marked increase in silt+clay contents and cations associated with the Sterkspruit soil form. The accumulation of cations, especially Ca, suggests that leaching is limited. Significant increase in the Ca:Cl ratio (Figure 3d) signifies the presence of 2:1 clay minerals formed by neo-formation (Wilson 1999).

Very high soluble concentrations of Ca, Mg and Na cations in the subsoil of plot 17 (Ss – Figure 5) suggest that a considerable amount of these cations accumulate in this ‘clay plug’. The interpretation is that lateral flow occurs at the soil/bedrock interface from plot 11 to 15, carrying dissolved cations which accumulate in the prismacutanic layer of the Sterkspruit soil form. This supports the morphological interpretation that plots 11 to 15 are interflow soils. Water presumably dams up behind the clay plug and, as water is lost to evapotranspiration (ET), cations accumulate in the soil resulting in an increase in the pH.

The change from apedal soils to structured soils (Figure 2) is marked by an increase in clay contents (Figure 3b) as well as an increase in ESP (Figure 3e). This marks the change from acidic to sodic soils. There is a change from leaching environment in the apedal soils to an accumulation in the structured soils, as clearly shown by the increase of base cations at the sodic site (note the increase in Na:Cl ratios as shown in Figure 3d).

Plot 36 marks the lower boundary of the sodic site; the soils change from Sterkspruit to Bonheim forms (Figures 2 and 3a). The relatively high pH and S-value of the surface horizon (50 mm) of the Bonheim (Bo) soil, compared to the 150 mm layer (Figure 4), suggest that surface accumulations of cations are dominant. This indicates that overland flow water, rich in dissolved cations, will accumulate at the break of slope. Below plot 36, prismacutanic horizons are degrading into pedocutanic horizons (Figure 2). Although the absolute concentrations of exchangeable cations are still relatively high, the cation to clay ratio decreases notably (Figure 3d). This implies that vertical drainage becomes slightly more important than in the sodic soils. The combination of more gentle slopes and blocky structures (compared to prismatic) could be attributed to enhanced infiltration (Kutílek 2004; Leeds-Harrison et al. 2007).

The variation in concentrations in plots 47–49 is associated with the deposition of different materials in the Dundee soil on the flood banks of the streams (Figure 3). Vertical variation in chemical properties is clearly depicted in Figures 4 and 5, which demonstrate the variation in deposited materials. The decreasing S-value with depth suggests that leaching is dominant in this profile, also supported by the morphological interpretation of soil properties.

The water contents at different suction levels and saturated hydraulic conductivity of the Clovelly soil support the morphological and chemical deductions that apedal horizons are a fast flowpath (Table 2). The high K_s value in the orthic horizon suggests that soluble cations could easily be leached from the horizon. The steep decrease in water retention suggests that macroporosity in this horizon is high (Leeds-Harrison et. al. 2007). This implies that conditions for complete saturation are hardly met because of the fast drainage of macropores.

The water retention data and K_s of the Sterkspruit soil are in accordance with morphological and chemical interpretations that water movement is very slow in the prismacutanic horizon (Table 2). Dispersion of clays, associated with the high Na content, could clog soil pores resulting in the low conductivity. The gradual decrease in water retention of the Sterkspruit indicates that macroporosity is relatively low under saturated conditions, that is, most of the water is held in the soil matrix.

The water retention data and K_s of the Bonheim soil support the morphological and chemical interpretations that water movement is faster than that observed in the Sterkspruit soil (Table 2). The decrease in water content with an accompanying increase in pressure is considerately less than in the Pinedene soil. The structure of the pedocutanic horizon promotes the presence of more macropores, as is evident by the water retention data. In spite of the greater number of macropores, compared to the Sterkspruit, the soil is dominated by smaller pores.

The conceptual hydrological response model of this hillslope is presented in Figure 6. This model was constructed from the interpretation of soil morphological properties and then refined and improved using measurements and interpretations of soil chemical and physical properties. The numbered arrows in Figure 6 reflect different dominant and subdominant hydrological flowpaths/processes as discussed below: