The interactive relationship between soil and water (hydropedology) was exploited to determine the hydrological behavior of 52 hillslopes in South Africa. These hillslopes were qualitatively grouped into six hillslope classes based on their dominant hydrological response derived from interpretations of their hydropedology. Applications of the classification system are presented.

Abstract

Soil has an interactive relationship with hydrology. It is a product of water related processes (physical and chemical) and a first order control of the destiny of rainwater. It is mapable with transfer functionality. These properties make it an appropriate entity for classification of hillslope hydrological responses. Hillslopes from all over South Africa were surveyed and hydropedologically interpreted. Soils were classified and based on the interpretation of the dominant hydrological pathway grouped into five hydrological soil types. The type and position of a hydrological soil types in a hillslope served as basis for the hillslope classification. Each of the hillslopes surveyed were assigned to one of six hillslopes classes. A flow diagram of the hydrology is presented. Arrows indicate the dominant flowpaths, and a hydrograph shows the anticipated impact on streamflow. The results made an impact on distributed modeling and land-use decisions, including land-use change to forestry and selection of on-site sanitation limiting water pollution. The composition and distribution of hydrological hillslope classes can serve as a basis for classification of catchments.

The need for a globally agreed-on catchment classification system has received a great deal of attention in the last decade (McDonnell and Woods, 2004; Wagener et al., 2007, 2008; Bouma et al., 2011; Sawicz et al., 2011), largely motivated by the challenges of hydrological predictions in ungauged basins and uncertainty of the growing awareness of the impact of climate change and land-use change on hydrology and of environmental pollution. Indeed, classification of central entities of interest is essential in many scientific disciplines. A catchment, being a landscape element that integrates all aspects of the hydrological cycle, is considered to be the central entity in hydrological studies (Wagener et al., 2007). However, Sawicz et al. (2011) maintains that although the catchment provides a sensible unit for classification, it is not necessarily the only unit/scale driving dominant hydrological processes. To be functional as a classification system of a spatial element of nature, it must have easily identifiable taxons and the criteria defining it must be visible (easily detectable) or correlated with one, or a combination, of detectable elements of nature to facilitate mapping of space with a unique response.

The hillslope is generally accepted as a fundamental landscape unit (Weiler and McDonnell, 2004; Lin et al., 2006). It exhibits a common form of organization and symmetry needed to construct a classification system. Functional classification units can be grouped on a higher level for improving communication of general information and divided on several lower levels for reduced variation and improved application. The interaction between topography, soils, climate, and vegetation results in patterns or laws that contain valuable information on the way they function (Sivapalan, 2003a). These elements play a significant role in controlling hydrology, and their relationships with water distribution are valuable for serving as indicators of hydrological response (Le Roux et al., 2011; van Tol et al., 2010a, 2010b; Kuenene et al., 2011). The hillslope is therefore an important building block for understanding and simulating hydrological processes (Tromp-van Meerveld and Weiler, 2008). The hydrological response of catchments is determined by the combined hydrological response of hillslopes in the particular catchment (Sivapalan, 2003b).

Similar to catchment classification, there is no general consensus on methods to characterize and classify hillslopes (Weiler and McDonnell, 2004; McDonnell et al., 2007; Tromp-van Meerveld and Weiler, 2008). Some researchers argue that every hillslope is unique (Beven, 2001). Experiments in the past focused on documentation of the unconventional behavior of new hillslopes instead of the systematic examination of first order controls of hillslope hydrological behavior, without intercomparisons to obtain common process behaviors (Weiler and McDonnell, 2004). The transference of these studies is limited. McDonnell et al. (2007, p. 5) therefore argued that any mapping and characterization should be driven by the desire to generalize and extrapolate from one place to another over multiple scales by using an interdisciplinary approach, without relying on calibration “…but rather a systematic learning from observed data and an increased understanding and search for new hydrologic theories through embracing new organizing principles behind watershed behaviour that are derived from our sister disciplines.”

Soils integrate the influences of parent material, topography, vegetation/land use, and climate, and may act as a first order control on the partitioning of hydrological flowpaths, residence time distributions, and water storage (Park et al., 2001; Soulsby et al., 2006). Soils science is therefore clearly one of the “sister disciplines” that should be considered in any hydrologic classification system (Bouma et al., 2011). The influence of soil on hydrological processes is due to the ability of soil to transmit, store and react with water (Park et al., 2001). These influences are primarily controlled by the physical and hydraulic properties of soils. Hydrologists agree that the spatial variation of soil properties significantly influence hydrological processes, but they lack the skill to gather and interpret soil information at the scale required for hydrological response estimation (Lilly et al., 1998). The relationship between soil and hydrology is interactive. Water is a primary agent in soil genesis, resulting in the formation of soil properties containing unique signatures of the way they formed (Ticehurst et al., 2007; van Tol et al., 2010a, 2010b).

Almost every hydrological process of interest to hydrologists (e.g., evapotranspiration, infiltration, and subsurface flow) is difficult to observe and measure because these processes are dynamic in nature with strong temporal and spatial variation (Sivapalan, 2003a). The dominant hydrological processes are invisible. Spatial and temporal variation requires long-term measurements to develop an understanding of the patterns of response. On the other hand, soil properties are in the short term not dynamic in nature and their spatial variation is not random (Webster, 2000). The spatial distribution of soil properties exhibit a common form of organization and symmetry, including vertical horizonation typical of soils and lateral topographic related distribution of soils in a hillslope. This typical distribution of soils in the landscape has first been referred to as a catena by Milne (1936), but it was modified later by Bushnell (1942) to replace the catena concept with the toposequence. This concept of the association of soil properties with topography and hydrological processes is also captured in the terms pedosequence or hydrosequence in relation to hillslopes (Chappell and Ternan, 1992; Flügel, 1997; Sivapalan, 2003a; Weiler and McDonnell, 2004). The correct interpretation of spatially varying soil properties associated with the interactive relationship between soil and hydrology can serve as indicators of the dominant hydrological processes (Ticehurst et al., 2007; van Tol, 2010a, 2010b) and improve the understanding of hillslope hydrology (Lin et al., 2006). Some soil properties play an interactive role with hydrology, while others (e.g., color) do not control hydrology but serve as indicators of soil water regime (Van Huyssteen et al., 2005; van Tol et al., 2010a, 2010b; Kuenene et al., 2011).

We believe that hydrological behavior of hillslopes should the first step toward classification of hillslopes. The hypothesis is that soil properties controlling current (and future) hydrology and soil properties indicative of ancient hydrological behavior genetically related to the interactive relationship between soil and water, are scientifically sound, and can serve as criteria to define functional units for hydrological hillslopes. The control of hillslope hydrology by parent material, both lithology and weathering patterns, and climate is framed in and therefore effectively represented by the distribution of soils in the landscape. The aim of this study was first to interpret the hydropedological behavior of a range of hillslopes in South Africa and group them into hillslope classes that will have similar perceptual hydrological behavior. This classification was largely based on qualitative interpretations of the hydrological flowpaths associated with hydrologically important soil properties, grouped in horizons of the profiles. Second, the application of this classification in distributed hydrological modeling for predicting the influence of land-use change and water quality is highlighted.

Materials and Methods

Study Areas

A total of 52 hillslopes located in 20 catchments were surveyed in this study. The hillslopes cover a range of geographical, geological, and climate regions in South Africa (Fig. 1 and Table 1). The level of investigative detail varied considerably between the different hillslopes (Table 1). Some hillslopes formed part of research catchments (e.g., Weatherley, Cathedral Peak VI and Two Streams) with detailed pedological and soil physical measurements and long-term hydrometric data available, whereas other hillslopes (e.g., Schmitsdrift and Bloemfontein) were surveyed morphologically with varied soil physical support measurements. Although the hillslopes in research catchments were integral to theory development and improved understanding of the interactive relationship between soil and hydrology (see, inter alia, Riddell et al. (2010), van Tol et al. (2010a, 2010b, 2011), and Kuenene et al. (2011)], the classification of hillslopes in this study was based solely on results of hydropedological surveys of the hillslopes.

However, all the hillslopes were surveyed with a hydropedological survey technique presented in Le Roux et al. (2011). This technique involves the identification of representative hillslopes in a study area, augering observations along transects perpendicular to the slope, detailed descriptions, identification of horizons, taxonomic classification of the soil profiles, and recording of all soil features related to hydrology. The soil information gathered during the survey phase was interpreted and related to associate hydrological behavior.

Hydrological Soil Types

Hydropedological classification transforms pedogenetic knowledge of geochemical and hydrological relationships, embedded in soil properties, to hydrological information that is useful for classifying soils. Soils with the same pedogenetic classification might have significantly different hydrological functions. On the other hand, soils classified into different pedogenetic classes might have similar hydrological functions (Bouma et al., 2011). Pedogenetically classified soils were successfully regrouped into hydrological functional units based on their hydrological responses (van Tol et al., 2011; Kuenene et al., 2011). This classification is similar to the hydrology of soil types (HOST) classification system of the United Kingdom (Boorman et al., 1995) and the process decision schemes to determine the dominant runoff process on a soil profile (Scherrer and Naef, 2003; Schmocker-Fackel et al., 2007), although the latter was based on extensive measurements with little reference to the pedology. In HOST, soils are divided into 29 classes based on their expected hydrological responses. In a catchment or area of interest, each soil class is expected to have a unique influence on the hydrology; therefore, the HOST classes are hydrological response units (HRUs). Pedological differences are credited with a high value even if the hydrological responses are expected to be similar.

Soil types are not randomly distributed; therefore, hydrological soil types typically occupy specific positions in the hillslope and by implication can play more of a releasing or receiving role related to hillslope position, altering its role in hillslope hydrology. By implication, pedogenetically different soils may be grouped in the same hydrological functional class. The hillslope can therefore be used as an HRU, and the division of soils into different hydrological response classes as building blocks of the spatial systematic variation in a hillslope. Five different hydrological types have been assigned to the soils in the study of hillslopes. A brief description of the properties of these soils is presented in Table 2. These soils could be subdivided further, but further subdivision at this level is seen as inappropriate for the hydrological classification of hillslopes.

Framework of the Classification System

The classification system proposed was based on intercomparisons of the perceptual hydrological behavior of hillslopes. Flowpaths of water through the hillslope and into the stream were the fundamental aspect considered in the intercomparisons. Because of the complexity of the hydrological system and strong time dependencies of hydrological processes, only the hydrological response during the peak rainy season was considered in this study, that is, not the wetting-up and drying-out phases typical of seasonal climates.

The classification further strongly relied on Jenny’s algorithm of the factors of soil formation (Jenny, 1941), where soil properties are the result of the impact of parent material, climate, topography, organisms, and time. The latter were therefore not considered in the classification as their influence is revealed in the soil properties and their spatial distribution; for example, steep slopes (topography) with high intensity rain storms (climate), low vegetative growth (organisms), and impermeable bedrock (geology) will favor erosion and eventually (time) result in shallow soils. For this reason, the hydrological soil types in the hillslopes are presented as two-dimensional bars without inclination or differences in slope length. Arrows indicate dominant flowpaths in the hillslope (Fig. 2), and a hydrograph indicates the anticipated outflow from the hillslope (Fig. 4–9). The length of the bars is relative to the fraction occupied by a hydrological soil type in a specific hillslope.

The hillslope classes are presented as two-dimensional shaded block diagrams, where the different shades refer to the hydrological soil types (Table 2). The left-hand side of the block diagram represents the crest with the stream at the right-hand side. The numbered arrows in Fig. 2 refer to dominant hydrological pathways; arrow 1 refers to vertical flow through and out of the soil profile into fractured rocks or other material with higher permeability, arrow 2 refers to overland flow, arrow 3 refers to fractured rock to the soil return flow, arrow 4 represents infiltration and lateral flow on the soil/bedrock interface, arrow 5 represents shallow lateral flow that eventually returns to the surface to contribute to overland flow, and arrow 6 indicates slow lateral discharge to the stream through responsive soils. The magnitude of the different arrows gives some indication of the dominance of the various pathways.

The perceptual schematic representations of the 52 hillslopes were compared with one another. Four aspects were considered in the intercomparisons: (i) occurrence of different hydrological soil response types in a hillslope; (ii) the sequence in which different hydrological soil types occur in a hillslope; (iii) the fraction of the hillslope covered by different hydrological soil types, and (iv) most importantly, how water will reach the stream, as expressed by the hillslope hydrograph. On the basis of these aspects, the 52 hillslopes were grouped into classes with similar anticipated hydrological responses.

The conceptual hydrographs accompanying different hillslope classes are representations of the hydrological responses of the specific hillslopes during typical rain events and are only provided to improve the conceptual understanding through graphical representation. No attempt was made to accommodate climate, antecedent moisture conditions, land use, etc.

Results and Discussion

The 52 hillslopes and their classes are presented in Fig. 3. Hillslopes from the same geographical area may fall into different hillslope classes (e.g., Weatherley 1–5 and Mokolo 1–5). Hillslopes occurring on different geological formations with different climates may, however, fall into the same hillslope class (Fig. 3). The properties and processes of the different hillslope classes are presented in Fig. 4 to 9.

Hydrological Hillslope Classes

Class 1—Interflow (Soil/Bedrock Interface)

This class may have the full range of recharge, interflow, and responsive soils. The presence of a soil/bedrock interflow region above the wetland (responsive soils) is a key attribute of this hillslope class (Fig. 4). The soil/bedrock interflow soils may constitute a narrow portion of the hillslope.

In this class it is important that bedrock properties largely control redistribution of water in the hillslope. Bedrock permeability controls soil-to-bedrock flow. Bedrock layering, controlling variation in permeability, controls bedrock-to-soil return flow. The resistance of layers to permeability and weathering influences the topography. Topography supports the soil-to-surface return flow. In the valley bottom an impermeable rock layer controls the formation of a wetland and responsive soils. Upslope the bedrock is permeable and ET (evapotranspiration) excess water leaves the soil and fills up the pore space in the fractured rock. The storativity of the bedrock and depth to an impermeable layer control return flow to the subsoil. The flowpath is large and long enough to leave signatures of additional water, that is, increased chemical weathering and reduced or redox morphology. Colluvial action could have contributed to deep weathered saprolite. This class has been identified in semiarid and subhumid climates.

During the wetter parts of the rainy season, responsive character expands upslope, including the soil interflow zone. This type of hillslope accommodates a large variation in residence times. The hydrograph typically has peak, shoulder, and base flow elements. A second peak is possible where the soil, compared to surface, contribution to streamflow is delayed.

Class 2—Shallow Responsive

Figure 5 represents the perceptual hydrological flowpaths and anticipated hydrograph response of hillslope class 2. This hillslope is dominated by shallow responsive soils (>50% of the hillslope). A sharp transition to impermeable rock controls the hydrological response in this hillslope class. During the rainy season, the small water holding capacity fills up and overland flow by saturation excess occurs. This class has been identified in arid and semiarid climates. A clear distinction between pedogenesis and hydrology is important in this hydrological class. Pedogenetically it is important that precipitates of lime, gypsum, and salt are distributed downslope in drier climates. In wetter climates, surface horizons are bleached. Variation in this class is limited to the presence of small areas of other hydrological soil types related to alluvium/colluvium and variation in the permeability of the parent rock. Deeper alluvial/colluvial soils next to the stream (Fig. 5a2) may have a small impact on the hydrology. The impermeability of the underlying rock promotes slow and limited discharge to the stream on the soil/bedrock interface because it competes with ET extraction of soil water. High peak flow is typical in this class because of the prominence of overland flow and streams flow for short periods after rain events (Fig. 5b). Base flow is absent, and where present, it is expected to be low and related to groundwater aquifers. Groundwater recharge is localized, low, and slow (Hughes and Sami, 1993).

Class 3—Recharge to Groundwater (Not Connected)

Hillslopes where recharge is dominant are represented in Fig. 6. These hillslopes are dominated by high chroma soils with redoximorphic properties limited to the valley bottom, indicating that the underlying bedrock is permeable. On these hillslopes, the infiltration and vertical redistribution rate is generally higher than the precipitation rate, thereby promoting sustained oxidized soil chemistry. Interflow and responsive soils are scarce. This class is commonly associated with sand deposits (aeolian deposits, coastal plains, and quartzitic sandstones) and karst landscapes, among others. Although vertical flow through and out of the profile is the dominant flow direction, lateral flow can occur (even in vertical isotropic soil with regards to hydraulic conductivity), resulting in redoximorphic signatures in soils next to the stream (Fig. 6a2 and 6a3). When shallow responsive soils occur above the recharge soils (Fig. 6a3), overland flow generated on the upper portions will not directly contribute to streamflow but rather infiltrate the recharge soils.

The direct contribution of this hillslope class to streamflow will be minimal (Fig. 6b). Groundwater recharge is high, groundwater levels are typically not connected to the stream (losing stream), and a net loss of water to deeper aquifers is anticipated. By implication, the groundwater distribution pattern may not be hillslope or catchment related.

Class 4—Recharge to Wetland

This hillslope class is dominated by recharge soils and stable wetlands with indications of long periods of saturation (Fig. 7a1–7a3). A key attribute of this class is that recharge soils occur directly above responsive soils, that is, lateral flow at the soil/bedrock interface toward responsive soils is absent. Water exiting the soil in the recharge areas flows through fractured bedrock and feeds the soils next to the stream, resulting in waterlogged conditions. Since the soils next to the stream are saturated, additional precipitation cannot infiltrate and saturation excess overland flow will be generated. Mountainous fractured rocky areas with peat wetlands are typical. The recharge area may have coarse, shallow soils of high permeability. For the formation of peat wetlands, the ratio of recharge to wetland area is related to climate (Marneweck et al., 2001).

The presence of shallow responsive and interflow (A/B interface) soils above the “recharge zone” does not affect the dominant character of the hillslope and can contribute to the recharge of the wetland (Fig. 7a2 and 7a3). Peak flows associated with the overland flow generated on responsive soils can be expected with some lateral contribution from the wetland soils (Fig. 7b). A stable base flow component is characteristic of the hillslope hydrology of this class.

Class 5—Recharge to Midslope

The hydrology of this hillslope class is controlled by permeable fractured rock at higher elevations and impermeable layers deeper in the rock forcing fractured rock return flow in the midslope and lower slopes. It is dominated by recharge soils in the upper regions with indications of saturation on the soil/bedrock interface in lower midslope and footslope positions (Fig. 8a). As with hillslope class 4, the recharge soils typically feed lower lying soils via a bedrock flowpath. Lateral flow on the soil bedrock interface is the dominant contributor to streamflow. As interflow is generally considered to be fairly slow, this hillslope class will have a delayed and prolonged hydrograph response (Fig. 8b). Wetlands typically form in the lower part of the interflow zone during high rainfall years. This phenomenon is periodic.

Class 6—Quick Interflow

This hillslope class is marked by the presence of soils with indications of lateral flow at the A/B horizon interface (Fig. 9a1 and 9a2). Lateral flow is typically generated by textural discontinuities upslope and downslope due to saturation in the B horizon. In our study, hillslope flow at the A/B interface was generated on duplex soils with saturation in the B (gleyed) horizon; therefore, the two flow generation mechanisms were dealt with in one hillslope class. This flow may result in a second peak in the hydrograph (Fig. 9b). Indications are that the hydrology of the duplex soils is controlled by the soil body according to the luviation model. Rainwater infiltrates and drains vertically at a high rate in the A horizon and is then retarded by the clay horizon, in which saturation and interflow results. The rate of interflow on the A/B interface mainly depends on the slope (van Tol et al., 2013). Hillslope contribution of water (fractured rock soil return flow) plays a role in the subsoil of footslope soils by creating two flowpaths in the soil body (Jennings et al., 2008). Base flow largely depends on recharge of the fractured rock as leakage from the perched soil water table is probably negligible.

Applications

Distributed Modeling

When soil characteristics are lumped in catchment scale modeling, results may, at best, reflect an averaged catchment wide soil water balance. Even when mapped soil properties are disaggregated into HRU and used in distributed modeling, inaccuracies remain, often through neglect of topographical controls and storage mechanisms (Dunn and Lilly, 2001; Rodgers et al., 2005). Indeed, Dunn and Lilly (2001) advise that it is the way in which soils are distributed in a catchment, which is important in deriving catchment scale model parameters. It seems reasonable, therefore, that the information inherent in hydropedologically derived hillslope classes would yield valuable catchment modeling capabilities.

Differences in hydrological controls of hillslope classes are reflected in the large variability of soil and hillslope contributions to streamflow. Typically, throughflow from soils and hillslopes has travel times of months to a few years, and groundwater sources may have travel times ranging between 5 and 20 yr (Hoeg et al., 2000; Uhlenbrook et al., 2002; Asano et al., 2002; Rodgers et al., 2005; McGuire et al., 2005; McGuire and McDonnell, 2006), although catchment response times may be considerably shorter. Nevertheless, distinctly different catchment responses have been directly related to different hillslope processes (Uhlenbrook et al., 2008; Graeff et al., 2009; Bachmair et al., 2011).

In some hydrological models, the lack of knowledge of how these hillslope and groundwater mechanisms yield water has required the use of simplified transfer functions, which simulate the systematic release of a temporarily stored volume. The partitioning of precipitation into this runoff volume and the infiltrated water volume is driven by the antecedent soil moisture deficit, mostly though Soil Conservation techniques [ACRU (Schulze, 1995) and ATHYS (Harader et al., 2012)] or a power function [HBV (Bergström, 1995; Uhlenbrook et al., 2004)]. These concepts also require knowledge of rainfall intensity/soil infiltration rate relationships, release of the stored runoff volume, and the evapotranspiration and redistribution of the infiltrated volume, and then proceed in the model. This separation of the less known surface and subsurface runoff yield processes and the better known and well-studied soil water-plant-atmosphere processes has already been defined as first order (hillslope controlled) and second order (near surface controlled) water (Uhlenbrook et al., 2008). Hydropedological hillslopes could be effectively applied to yield the control parameters for both the detailed near surface processes as well as for the runoff response mechanisms.

Maximum use of the hydropedological sequence descriptions and characteristics will be realized where model developments preserve the detail of soil water-plant-atmosphere interactions and also capture the dominant storage and delivery mechanisms of the contributing hillslopes.

A modified version of the ACRU agrohydrological model (Schulze, 1995), ACRU-Int, was used to demonstrate the value of the hydropedological sequences derived for the catchment. Two modifications to the model were directed at providing a physical basis to the delivery of the stored runoff volume. In the first modification, an intermediate layer was introduced to the existing two-layer structure to simulate the threshold responses of the interflow profile type. In this intermediate layer, the water was assumed to be distributed close to an equilibrium state so that a critical volume could be defined at which positive soil water pressures are induced at the soil/bedrock interface (Fig. 10). This critical volume was easily derived from the water retention characteristics of the intermediate layer. Additional input to this layer triggered lateral discharge and percolation into the fractured bedrock.

The second modification allowed for the linking of sequential land segments to mimic the hydropedological hillslope (Fig. 11). Inherent in the model structure is the ability to define a priori links between any upslope layer with any downslope layer. The transfer functions linking these layers can take the form of linear, exponential, or advection-dispersion equations.

Application of the model in the Weatherley research catchment was based on the hydropedology of hillslopes 1 to 4 (Lorentz et al., 2008; Wenninger et al., 2008). Since the aim was to evaluate the applicability of the hillslope classification, the model was not calibrated against measured data. [For details of model parameters and comparisons between simulations with different levels of input data, see Lorentz et al. (2008) and Le Roux et al. (2011)]. Simulated and observed streamflow values of the highly responsive catchment were in acceptable agreement (R2 = 0.74) [Fig. 12 (Lorentz et al., 2008)].

Land-Use Change Impact

Hydropedological sequence descriptions would contribute to effective land management. The hydropedological descriptions in the Weatherley catchment were used to demarcate afforestation boundaries, not only in toe slope wetlands but also at accumulation zones in the hillslopes. Also, the hydropedology of hillslopes in the Kruger National Park (Skukuza 1—4 and Letaba 1—4) is suitable for and will be applied to improved ecosystem management, where water distribution is critical to vegetation and animal interactions.

Water Quality

Hydropedology mapping and associated water distribution dynamics could also be used directly to locate suitable development sites during land-use change (afforestation, housing development, on-site sanitation) and in assessing subsurface solute transport for limiting and remediation of pollution. The Taylor’s Halt hillslope hydropedology has been used to determine the propensity for near surface lateral flows to intercept pit latrine contents in a rural development. In addition, the Wartburg hydropedology descriptions have been used to estimate near surface discharge of nutrients in sugarcane fields.

Conclusions

The application of pedology in environmental science is overdue. Pedology can be researched to serve hydrology to the benefit of both (Bouma et al., 2011). We examined the hydropedology of 52 hillslopes covering a range of geographical areas, geological layers, and climate zones in South Africa. The level of experimental detail varied between the hillslopes, but a hydropedological survey was conducted on all of them. On the basis of the interpretation of soil morphology, the soils of the hillslopes were divided into five hydrological soil types. The area covered by the different soil types and the sequence of the soil types from the crest to the stream were used to group the 52 hillslopes into six different hillslope classes. These classes give a qualitative description of how water will reach the stream.

Hillslopes from different geographical areas, on different geologies, and in different climate zones might have the same hydrological behavior, for example, Cathedral Peak 1 [basalt with aridity index (AI) > 0.6] and Loeriesfontein (shales with AI = 0.2). Hillslopes from the same area (geology and climate) might have distinct hydrological behavior, for example, Weatherley 1 (class 4) and Weatherley 5 (class 6).

The hydrological classification of hillslopes addresses a core issue in hydrological modeling, namely the conceptual hydrological response model, and therefore an important element of reliable hydrological predictions, especially in ungauged basins. An example of how the hillslope classification can be used to structure a distributed model is demonstrated through a modeling exercise in a small research catchment (Weatherley). The hydrological classification is also applicable in evaluating the influence of land-use change and water quality assessments as well as for identification of strategic positions to install monitoring devices for feature quantitative characterization of hillslopes.

This classification is a first approximation and will be refined as more soils and hillslopes are studied. An improved understanding of the hydrology of soils (control mechanisms of flowpaths in soils, e.g., A/B interflow, subsoil interflow, water conducting macroporosity, near surface macropore interflow, bio-pores, structural pores, etc.) and hydropedology (interactive soil/fractured rock/hillslope processes) and signatures of these processes is vital for improved accuracy and applicability of the classification.

The hydrological classification of the hillslopes needs to be validated in relation to quantitative and/or semiquantitative measurements of flow processes. Virtual experiments (Weiler and McDonnell, 2004) using mechanistic models (e.g., Hydrus 2D) can aid in the validation of the classification. This will involve subjection of the studied hillslopes to a set of simulations under different environmental conditions by using the actual soil properties and soil distribution of the hillslopes. The validation of the classification might lead to new insights of the hydrological behavior of the hillslopes and can result in redefining the current classification or even more (or fewer) hillslope classes.

The authors gratefully acknowledge the Water Research Commission for funding this research under project K5/2021. The contributions of anonymous reviewers to the quality of this paper is also highly appreciated.

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Open Access