Keywords

1 Introduction

Knowledge of natural, preindustrial climate change and associated processes is essential for assessing current and projected environmental changes. Only when the interrelationships of all essential influencing factors are sufficiently understood will we be able to recognize environmental changes at an early stage, to assess them, and to react to them with sustainable (utilization) concepts. To achieve this goal, sediments as climate archives are of inestimable importance. The application of different methodological and analytical approaches is a prerequisite for the reliable interpretation of the multitude of proxies parameters and their information. The case studies presented here certainly do not cover the entire spectrum of comparable studies in southern Africa, but they represent successful model examples of innovative paleoclimate research. The reconstruction of past environments is critical to the objectives of several projects within the SPACES program. Strong links exist, for example, between the science presented in Chaps. 1 and 2 where a sound understanding of past environments is considered necessary to anticipate present and future response to climate change. Additionally, a detailed understanding of vegetation and hydrologic change during the Holocene can be considered as critical reference information for conservation and land use management practices (Chap. 12). Marine and terrestrial sediment cores are used to understand past environmental conditions from proxy data sources, including a large variety of physical, chemical, and biological indicators. These proxies are used to infer past fluctuations in vegetation, precipitation and weathering regimes, sediment fluxes, pathways, and depositional sites as well as changes in erosional runoff. Some of these indicators are measured on allochthonous sediment components, that is, components that were transported to the site by water or wind. At offshore coring sites located close to a river mouth or leeward of areas with high winds and sparse vegetation cover, sediments may be almost entirely of allochthonous nature. The interpretation of terrigenous (terrestrial) climatic signals measured on these components requires a thorough understanding of their source, or provenance, as well as of their transport pathways to final deposition. A complex factor in downcore analysis of proxies measured on terrestrial-derived material is that different components (such as pollen and spores, charred fragments, plant waxes, minerals, and rock fragments) may have different source regions, as well as transport pathways. Within a particular catchment, the provenance of inorganic and organic material is likely to be distinct, especially if there is a large variability in relief and vegetation cover. This has been shown in overview studies of riverine transport (Leithold et al. 2016) as well as in individual catchment studies (e.g., Bouchez et al. 2014; Häggi et al. 2016; Galy et al. 2011; Hahn et al. 2016; Hemingway et al. 2016; Herrmann et al. 2016; Zhao et al. 2015). Furthermore, the transport pathways and deposition of the various terrigenous fractions will be a function of the characteristics of source areas that govern the size, shape, density, and stability of sedimentary particles as shown by several on-shelf sediment distribution studies (e.g., Petschick et al. 1996; Rogers and Rau 2006). Sediment depocenters on the shelf are also in many ways controlled by sea-level history and ocean currents as they play a large role in fractionating grain size, grain shape, and composition (Green and Mackay 2016; Flemming and Hay 1988).

Source-to-sink studies shed light on the catchment processes and depositional conditions that influence proxy signatures. They facilitate reliable paleoclimatic interpretation of proxies measured on allochthonous sediment components. The diversity of the climatic settings, types of rivers, accompanying wetland systems, and depositional environments in southern Africa are immense and the work presented here has attempted to target different climatic and environmental zones within this (Fig. 28.1 shows simplified climatic and oceanic circulation systems in southern Africa). The case studies presented from southern Africa illustrate how source-to-sink studies can be applied in a variety of paleo-environmental and paleo-climatic contexts.

Fig. 28.1
A map of the southern part of Africa depicts the atmospheric and oceanic circulation. They mark Benguela current in the west, tropical easterlies in the east and Agulhas current along the south. It features the summer, year around, and winter rainfall zones.

Schematic overview of the main components of the oceanic (thick arrows) and atmospheric (thin arrows) circulation over southern Africa (modified from Truc et al. 2013). Southern Hemispheric Westerlies (SHW) and Intertropical Convergence Zone (ITC) are shown in their summer position. The numbers indicate the approximate location of the case studies in this chapter, more detailed maps are provided in the following paragraphs. Figure modified from Hahn et al. (2017)

Full size image

2 Depositional Studies/Stratigraphy of Marine Environments at the Sink

2.1 Description of Mapping Strategies

A core or grab sample provides a window of insight into that specific site location, but context is necessary to construct robust interpretations. Hydroacoustic surveys of the seafloor and coastal water bodies have been carried out in both ocean and coastal water body settings. The purpose of this work was to (1) provide suitable geological context around the deposition of sediments that were cored by allowing for a holistic “picture” of the sedimentary environment and (2) mapping was carried out before coring to inform an appropriate site for sampling. In all cases, these water bodies were mapped with subbottom profilers, and specifically, offshore voyage M123 traversed the continental margin from Walvis Bay to the Delagoa Bight and continuous coverage of multibeam echosounder data was acquired for depth, and PARASOUND for sub-seafloor context. In the coastal water bodies, boomer subbottom profiles were collected. Around each proposed site, numerous profiles were specifically collected to contextualize appropriate coring sites prior to sampling.

Boomers are low-frequency seismic profilers, characterized by a broadband frequency spectrum and high peak intensity. Seismic energy is derived from a bank of capacitors, discharged into an electromagnetic coil and providing a clear pulse and resolvability of seismic return. Boomer subbottom profilers are considered medium-penetrating seismic systems and allow an insight into upper continental shelf stratigraphy.

The PARASOUND method of subbottom profiling works on two frequencies (18 kHz and a secondary variable frequency of 20.5–23.5 kHz). The superimposition of these variable frequencies results in a “nonlinear parametric effect” that reduces the area of reflecting the surface compared to conventional subbottom profilers (Hempel et al. 1994).

Echosounders determine the travel time of an acoustic pulse by detecting the sharp leading edge of the return echo, providing a depth to seafloor (Mayer and Hughes-Clarke 1995). Acquisition of sonar swath bathymetry using a multibeam echosounder encompasses the principle of a three-dimensional fan shape of acoustic energy, subjected to pitch, roll, yaw, and vessel oscillation. The array transmits pulses triggered at known intervals to insonify an area of seafloor normal to the ship’s track. High-resolution motion reference units compensate for the inherent motion in the data acquired. The angular coverage sector of beam angles varies with depth, allowing accurate bathymetric determination from a relatively small number of passes across a deeper area.

2.2 Challenges and Limitations

For voyage M123 (Zabel 2016), areas on the seafloor (the continental shelf and shelf-edge environments) were located that preserve terrestrial deposits, such as fluvial, lacustrine, and mudbelt depo-centers. Reconnaissance sites were broadly defined using legacy data of Dingle et al. (1987) and Birch et al. (1986) and past and ongoing work of South African marine geologists. These maps and existing data allowed the sites to be constrained according to surficial textures and sediment characteristics, and proved particularly beneficial in identifying the loci of muddy sediments on the shelf.

2.3 Examples of Southern African Hydroacoustic Studies That Have Been Used for Source-to-Sink Understanding

2.3.1 Case Study 1: Southernmost South African Shelf

The inner to mid continental shelf of the Agulhas Bank, which forms part of the Paleo-Agulhas Plain when it was exposed as a terrestrial landscape in the past, is covered with Pleistocene deposits. The wide lateral extension of these remnant deposits is the expression of a flat underlying substrate, which differs markedly to the adjacent onshore area (Cawthra et al. 2018, 2020). Six sediment vibrocores were obtained across the Paleo-Agulhas Plain in submerged environments of estuaries, lakes, and river floodplains (GeoB18305-3, GeoB18306-1, GeoB18307-2, GeoB18308-1, GeoB20628-1 and GeoB20629-1) that revealed sediments preserved since the Last Glacial Maximum. A seismic stratigraphic framework was developed from the inner to mid-shelf between the Breede River in the West and Plettenberg Bay in the East, for the RAiN project, and this work elucidated 20 different Quaternary units within 2 depositional sequences (Fig. 28.2). Incised paleo-river channels were mapped and cored deposits also mapped seismically from estuarine, lacustrine, and fluvial systems are grouped to represent the lower floodplain (Fig. 28.3). The most pervasive stratigraphic pattern in these shelf deposits was left behind during the sea-level drop from 125,000 years ago to 20,000 years ago, or the penultimate cycle of warming into the Last Glacial Maximum.

Fig. 28.2
Two illustrations. At the top, It depicts sequence boundary 1 and 2, which mark modem seafloor, chaotic reflectors, bounding surfaces, and high amplitude reflector. At the bottom, is a table with column headers systems tract, base level, events, and surfaces.

(Top) Representative subbottom profiles to demonstrate how the bounding horizons, e.g., wave ravinement surface (“WRS”) and sequence boundaries, as well as seismic units and facies were delineated in this study. (Bottom) Schematic showing the basis for definition of systems tracts (–A): negative accommodation space

Full size image
Fig. 28.3
An illustration of sequence Boundaries. They feature sequence boundary 1, 2, incised channels 1 and 2, wave ravinement surface, bedrock, mud belt, shelf sands, dunes, lagoons, lower floodplain, and channel fill.

Subbottom profile intercepting the paleochannel of the Breede River near Cape Infanta. The site of core GeoB20628-1 is shown

Full size image

2.3.2 Case Study 2: Limpopo Shelf

The Limpopo shelf of the Delagoa Bight comprises the extensive submerged and asymmetrical wave-dominated delta of the combined Matola, Incomati, Lusutfu, and Limpopo rivers. Hydroacoustic studies in the region (e.g., Dyer et al. 2021) gave further background information on sediment depositional processes to correctly interpret core stratigraphy. The seismic stratigraphy and architecture of the shelf was revealed by PARASOUND subbottom profiling to comprise two submerged, wave-dominated deltas sandwiched between three associated paleo-shorelines, all draped by sediment that had accumulated after the postglacial transgression. Given the complexity of the underlying submerged delta stratigraphy, especially the convoluted delta top deposits that comprise most of the proximal areas of the shelf, the marine sediment cores were located in the most stratigraphically consistent areas of the postglacial cover sediment. These included at the shelf edge, amid high frequency and continuous well-layered reflections, and behind aeolianite pinnacles on the mid shelf, where post glacial transgressive erosion was minimized and the contemporary mud clinoform of the rivers could be preserved (Fig. 28.4).

Fig. 28.4
Two boundary diagrams and a map of southern Africa. The boundary diagram features, palaeo delta top, mud clinoform, muddy glacial sediment, delta phase 2, delta phase 1, older pre L G M acoustic treatment, and contorted reflections or soft sediment deformation.

Seismic stratigraphy of the transgressive deposits of the Limpopo shelf. Two phases of submerged deltas are evident, covered to landward by the prograding delta top. This is in turn draped by the modern mud clinoform of the area, which has dammed behind an aeolianite pinnacle (after Dyer et al. 2021)

Full size image

2.3.3 Case Study 3: Durban Shelf

Unlike the above-mentioned examples, the Durban shelf is a current-swept and mostly bedrock-controlled feature where Holocene sediments are typically only preserved as the infill of old river courses (Green et al. 2013a). Despite the scarcity of sediment, previous multibeam bathymetric surveying revealed a series of submerged lagoonal deposits on the seabed that appeared to comprise muddy infilling material in line with the contemporary lagoons of the southeast African margin (Green et al. 2013b, 2014). The detailed hydroacoustic surveying and seismic stratigraphy of these lagoonal features was highlighted by Pretorius et al. (2016), who identified multiple aeolianite and beachrock shorelines preserved on the seabed, behind which extensive Holocene sediment had accumulated (Fig. 28.5). Much of the deeper Holocene sediments on the shelf were considered to be relict portions of a detached shoreface, driven by abrupt rises in sea level (Pretorius et al. 2016). The core sites were chosen on the basis of sediment thickness and their position within a large incised valley that had back-flooded to form the lagoons during the Younger Dryas period; the aim was to retrieve muddy material suitable as a climate archive spanning that period.

Fig. 28.5
Four boundary maps and a map of southern Africa. The boundary maps feature contemporary shore face, tempestites, acoustic basement, aeoliantite beach rock, and reflect shoreface.

Seismic stratigraphy of the Durban continental shelf. (a) Core location GeoB18303-2, which retrieved a mixture of shoreface and wash-over deposits, sheltered behind an aeolianite barrier. These overlie the paleo-lagoon complex imaged by Green et al. (2013a). (b) Core location of GeoB18304-1, which intersected a series of tempestites (hummocky reflectors) that adjoin the modern shoreface

Full size image

The cores retrieved comprised a series of alternating gravels and sands; however, with the aid of the seismic stratigraphy and stratigraphic architecture of the subbottom, it became apparent that these reflected a series of sediments that had been deposited during periods of intense marine storminess. The longer core, GeoB18303-2, was associated with wash-over into the back-barrier lagoon system during a period of storminess that spanned ca. 11,400–11,200 years ago (Green et al. 2022), consistent with the post Younger Dryas submergence of the lagoon system (Pretorius et al. 2016). The shallower core was located in the more proximal incised valley of the system, the Holocene cover deposits of which comprised a series of intercalated marine- and terrestrial-derived sediments (Green et al. 2022). In the subbottom profiles, these had a characteristic hummocky seismic appearance and were consequently interpreted as classic shelf tempestites, their ages spanning the period ca. 7000–4800 years ago. Green et al. (2022) linked this period of storminess to an unprecedented series of southward tracking tropical cyclones that made landfall in Durban, in contrast to the contemporary situation where tropical cyclones do not reach as far south. These were linked to positive Indian Ocean Dipole anomalies, driven by a period of increased sea surface temperatures in the southwest Indian Ocean.

3 Catchment Studies at the Source

3.1 A Description of Sampling Strategies

During field campaigns, a representative sampling strategy was developed to best capture catchment conditions in a time-efficient manner. In general, three different sample types were collected:

  • Suspension load samples representing a current mean value of the entire catchment upstream of the site. Two techniques were used to sample riverine suspension load. Firstly, an approach of pumping river water from midstream into several canisters (total: 100l) that were subsequently centrifuged. Secondly, finding deposits where the fine fraction transported during higher flow conditions is retained (for examples, see Fig. 28.6).

  • River bank samples.

  • Paleoflood deposits were sampled to gain an idea of the temporal variability in the catchment and representing nonmodern analogue conditions.

Fig. 28.6
Four photographs labeled from a to d. a captures a large, flat stone with a straight edge on one side, b captures a barren rocky land, c captures a river with large flat rocks rising slightly above the water, and d captures a 3-pronged tree trunk.

Examples of sediments collected from dried out puddles (a,c) and flood deposits (b,d) along the Orange River. Copyright Ralph Kreutz, MARUM

Full size image

Sampling sites were chosen at localities of assumed minimum human impact. Samples were taken from as many confluences as possible, as well as at locations directly downstream of every confluence.

3.2 Challenges and Limitations of Catchment Studies

The catchments of most rivers are diverse. Large river systems may traverse several climatic and vegetation zones, as well as a suite of geological formations (examples from southern Africa: Orange and Limpopo River systems). Small catchments can be equally diverse due to, for example, microclimates, the effects of large altitudinal differences, or localized wetland occurrences near the river channels. The variability throughout a catchment increases further when seasonal changes are considered. Capturing the total diversity of entire catchment areas with the number of samples that can be collected by a research team within a limited time frame of a research project is close to impossible. Catchment areas situated in more rural areas (e.g., parts of the Eastern Cape, South Africa) are often inaccessible via the road network, limiting the availability of sampling sites. The above-mentioned sampling strategies (Sect. 28.3.1) are employed to obtain the best representation of the temporal and spatial diversity in a catchment. Although sampling sites with obvious signs of human or animal (e.g., bioturbation due to hippopotamus or cattle trampling) disturbance are avoided during sampling, contamination of samples by human activity can never be excluded, with some catchments being heavily polluted especially near urban areas. Sites that are easily accessible to sampling are often used for common recreational activities such as fishing or boating. Depositional processes might therefore be disturbed and contaminants (oils, plastics, sewage) may be found in the samples. The downcore application of proxy interpretations based on catchment studies may be further distorted by situations when nonanalogue conditions occurred. The interpretations of proxy indicators that can be deduced from modern catchment studies is only valid if the conditions in the catchment remain comparable with the modern environment throughout time. However, vegetation and erosion processes within a catchment may change with large climatic shifts, for example, during glacial-interglacial transitions. During no-modern-analogue scenarios, changes in sediment source area have to be accounted for, and downcore proxy interpretation may have to be adjusted.

Info boxes on sediment proxies:

Palynology is the analysis of organic-walled microfossils, which comprise pollen and spores as well as algal cysts, fungal remains, dinoflagellate cysts, scolecodonts, and even the inner linings of foraminifera. Fossil pollen in particular offers one of the most widely used tools for understanding past environmental conditions, because pollen grains are morphologically diverse, abundant and generally well dispersed, and extremely resistant to decay. By counting and identifying pollen and other microfossils preserved in sediments, it becomes possible to reconstruct past environments including vegetation composition, climate shifts, changes in soil properties and hydrology, and sometimes also human disturbances through farming and pastoralism.

The elemental composition of sediment is used to gain insights into geological mechanisms and earth processes such as climate change, local and regional events (e.g., floods, landslides, storms), and anthropogenic changes (e.g., changing land use, pollution). A large number of element proxies have been recognized as important indicators of climate, weathering, and erosion conditions as well as sediment provenance.

Plant-waxes as vegetation and hydrological indicators. All terrestrial plants protect their leaves but also other plant surfaces with a surficial layer of wax (Eglinton and Hamilton 1967). These waxy compounds are made to resist, so can persist in the environment (e.g., soils, sediments) for long times after plant decay. In particular, n-alkanes as nonfunctionalized lipids are very refractory, preserving their distributional and isotope signatures up to millions of years (Schimmelmann et al. 2006).The concentrations of waxes in sedimentary archives depend on the terrestrial input, either due to higher plant coverage in the source area or stronger (fluvial, aeolian) transport. Different plants, however, produce differential amounts of waxes, complicating a quantitative approach (Garcin et al. 2014). As resistant compounds, they get relatively enriched during organic matter degradation, providing information on the degradation status of total organic material. Other degradation information can be obtained from their internal distribution. Plant synthesize preferentially even- (fatty acids, n-alcohols) or odd-numbered (n-alkanes) waxes, a signal that diminishes with proceeding wax degradation (Bray and Evans 1961).Some additional information on the contributing vegetation might be obtained from the internal distributions of the waxes. Generally, grasses tend to produce longer-chain waxes than woody vegetation (Vogts et al. 2009). However, this information might be misleading as plants from specific environments can contribute very specific chain-lengths, such as sedges, for example, Cyperaceae in swamps and floodplains producing preferentially very long-chain waxes (Schefuß et al. 2011). In sedimentary archives, the terrestrial wax signal has to be disentangled from other, microbial and/or algal, contributions. Overall, wax distributions alone should thus only be interpreted with caution and such interpretations preferentially be based on calibration studies on wax distributions of regional plant types (Carr et al. 2014; Herrmann et al. 2016). On the other hand, the analysis of multiple wax components is able to provide information on the diversity of plant changes and environmental responses within an ecosystem (Hoetzel et al. 2013). More specific information on contributing photosynthetic plant types can be obtained from their compound-specific stable carbon (δ 13C) isotope compositions (Diefendorf and Freimuth 2017). A broad distinction can be made between waxes from C3 or C4 plants with waxes from C4 vegetation being isotopically enriched in 13C due to a more effective photosynthetic carbon (CO2) fixation (Collins et al. 2011). Most tropical grasses are of C4 type, while all trees and shrubs are C3 plants. In this sense, the occurrence of C4, that is, 13C-enriched, waxes has been attributed to contributions from drier, grassy environments. However, also plants from specific environments, such as the Cyperaceae from swamps or salt-tolerant plants in saline settings, can be of C4 type, complicating a straightforward interpretation of 13C compositions alone (Schefuß et al. 2011). An additional complication are CAM plants, being able to switch carbon fixation mechanisms and thus producing variable and intermediate 13C compositions, such as, for instance, in biomes along the west coast of south Africa (Boom et al. 2014). Plants also adjust their 13C compositions due to environmental stress, such as water shortage. Under drought conditions, plants increase their water use efficiency, leading to a 13C enrichment (Hou et al. 2007). Thus, 13C compositions of plant waxes should not be interpreted alone but in conjunction with other parameters. Direct hydrologic information can be obtained through compound-specific hydrogen isotope (δD) analyses of waxes. The δD composition of waxes depends on the isotope composition of their meteoric water source, ultimately related to the isotopic composition of rainfall with an overprint of isotopic enrichment due to evaporation and/or plant transpiration in dry settings. Such isotopic enrichment due to evapo-transpiration has been detected in dry environments of South Africa (Herrmann et al. 2017). Additionally, a secondary dependency of δ D composition of waxes is observed between different plant types (Sachse et al. 2012). Nonetheless, the δ D composition of waxes provides a direct link to atmospheric isotope hydrology. Rainfall isotope effects such as the amount-, temperature-, altitude- and source-effects are all observed in specific settings in South Africa (Burdanowitz et al. 2018; Hahn et al. 2017, 2018, 2021; Miller et al. 2019, 2020). On longer time-scales (i.e., pre-Holocene) the ice volume effect additionally leads to an enrichment of isotopes in the hydrological cycle. For some of these effects, their contribution to the δ Dwax signal can be accounted for (e.g., temperature, ice volume) or can be assumed to be constant (e.g., altitude) in specific settings. Others, that is, amount and source effect, are often complicated to disentangle, in particular, in southern Africa where several moisture sources overlay. A careful multiwax parameter and multiproxy (incl. pollen, inorganic geochemistry) analysis, however, can lead to meaningful interpretations and far-reaching insights into (paleo-)hydrological changes compared to single parameter analyses. Plant-wax analyses thus yield integrated information on broad-scale vegetation changes, organic matter degradation, and the eco-hydrological status of the entire ecosystem. Combined in a multiproxy approach with palynology, inorganic geochemistry, etc., plant-wax analyses are able to provide very specific and detailed insights on the (vegetation and hydrologic) status of modern ecosystems, transport pathways of terrestrial organic matter, climatic-driven vegetation changes, and atmospheric-driven hydrologic changes.

3.3 Examples of Southern African Source-to-Sink Studies That Have Been Used for Paleoclimatic Work

3.3.1 Case Study 4: Western South African Coast

The Orange River, the largest river system in southern Africa (catchment area of almost 106 km2, that is, equivalent to about 77% of the land area of South Africa), drains large parts of South Africa from the headwaters in the Maloti-Drakensberg Mountains in the east to the mouth in Namaqualand where it enters the Atlantic Ocean (Fig. 28.7). It transports about 106 × 106 m3 of sediment annually into the Atlantic Ocean (Birch 1977; Compton et al. 2010) and formed the ~70–120 m below sea level coastal-parallel Namaqualand mudbelt, which contains sedimentary records of the discharged material of the past ca. 11,000 years (Compton et al. 2010). For a thorough interpretation of the stored paleo-climatic and paleo-environmental signals, it is important to know where the material originates from and how it is altered along the way from the source to the Namaqualand mudbelt. Therefore, inorganic and organic geochemical analyses of soil samples, river suspended material, riverbed samples, marine surface sediments, and marine downcore sediments were carried out.

Fig. 28.7
A map of the southern part of Africa where biomes are marked. These include fynbos, succulent karoo, desert, and azonal vegetation. Surface sediments, river samples, and soil samples are indicated with colored dots. Geo B 8323-2, and Geo B 8331-4 are indicated with colored stars.

Map of all sampling sites relevant for case study 1 (soils, river, surface sediments, cores), biomes (after Mucina and Rutherford 2006). Figure after Burdanowitz et al. (2018)

Full size image
Fig. 28.8
Two box and whisker plots epsilon app versus biomes and delta 13 C wax versus biomes. They show up values in negative. The plots for orange, holgat, buffels, and olifants are highlighted in a different shade.

Box and whisker plots for (a) the apparent hydrogen isotope fractionation (εapp) of plant-wax-derived n-alkanes in soils from different biomes (dark grey) and rainfall zones (light grey) and (b) δ13Cwax of soils from different biomes (dark grey) and river samples (red). Boxes comprise the middle 50% of samples and the horizontal black line within the box represents the median. Black dots outside the whisker plots indicate the uppermost and lowermost 10%. Na Karoo and Sc Karoo indicate Nama Karoo and Succulent Karoo, respectively. Note that εapp is shown on an inverse axis. Figure after Herrmann et al. (2016)

Full size image

Hydrogen isotopic analyses of plant-wax derived n-alkanes (δDwax) in soils show a diverse picture across the different rainfall zones in South Africa (Herrmann et al. 2017) (Fig. 28.8). In the summer rainfall zone (SRZ), δDwax reflects the annual δD of precipitation (δDp), although affected by evapotranspiration, which is indicated by the apparent fractionation factor (εapp, Fig. 28.8a). Thus, δDwax is a suitable qualitative paleo-hydrological recorder in the SRZ. In the winter rainfall zone (WRZ), the situation is less straightforward. There is no relationship between δDwax and annual δDp, probably due to wide microclimatic variability, distinct vegetation communities, and diversity as well as potential influence of summer rain, especially in the eastern parts. In the WRZ, the processes overprinting the δDp signal in the δDwax remain unclear and require further research. The n-alkane distribution patterns and carbon isotopes of n-alkanes (δ13Cwax) of South African soils reflect the vegetation type in the surrounding area and are distinct in the distinct biomes.

As the Orange River is the main sediment contributor to the Namaqualand mudbelt offshore western South Africa, there is a temptation to regard the Orange River sediment as the basis of past climate change interpretation. However, several small ephemeral rivers located in the adjacent Succulent Karoo biome and the Berg and Olifants Rivers in the southern part of the mudbelt are additional possible sediment contributors (Benito et al. 2011; Burdanowitz et al. 2018; Granger et al. 2018; Hahn et al. 2016; Herrmann et al. 2016). In addition, wind-driven input (e.g., of material from the adjacent western coast biomes) may be also important adding to the sediment contribution in the middle and southern mudbelt (Birch 1977; Gray et al. 2000; Zhao et al. 2015).

The isotopic and geochemical composition of the sediment cores offshore the Orange River mouth varies with shifting rainfall patterns throughout the large Orange River catchment area (Herrmann et al. 2017; Hahn et al. 2016). Overall, the case study shows that proxy interpretation is not simple. Knowing the sources and overprinting processes of the used proxies is crucial to reconstructing past climatic changes.

3.3.2 Case Study 5: Southernmost African Coast

The Gouritz River catchment is located in southernmost South Africa (Fig. 28.9). Despite the relatively small size of the catchment (53,139 km2) compared to the Orange (973,000 km2), the altitudinal gradient is steep as the Swartberg Mountains rise abruptly above 2000 m a.s.l. within 100 km of the coast (Le Maitre et al. 2009) and the Gouritz River catchment is the largest on the Cape South Coast. The catchment is located mainly in the year-round rainfall zone (Fig. 28.1) and has a mean annual runoff of ca. 488 × 106 m3 (Le Maitre et al. 2009). Major floods, caused by extreme rainfall events, are characteristic of this area (Desmet and Cowling 1999). The geochemical signatures of the paleoflood deposits of the Gouritz river catchment were essential for deciphering the downcore signal in a marine sediment core near Mossel Bay, offshore the Gouritz River mouth (Hahn et al. 2017). In the source-to-sink study, terrestrial catchment samples were all taken at lowland locations. There was, however, a distinct difference between the plant wax δDC31 values of soil samples and plant wax δDC31 values of flood deposits (Fig. 28.9). Plant wax δDC31 values are directly related to the isotope composition of precipitation (Sessions et al. 1999). Therefore, the conditions under which plant waxes contained in soils and plant waxes derived from paleoflood deposits were synthesized must have been different. Rainfall δD signatures become deuterium-depleted with altitude (ca. 10–15‰ per 1000 m; Gonfiantini et al. 2001). In view of the extreme elevation difference in the Gouritz River catchment, it is most likely that the relatively deuterium-depleted paleoflood deposits contain a considerable amount of upper-catchment material. Southern Hemispheric Westerlies-related precipitation events in the otherwise arid winters have been described as the main source of precipitation in the upper Gouritz River catchment (Chase et al. 2013).

Fig. 28.9
Two illustrations. a, depicts the map of the southern part of Africa. b plots delta D n C 31 alkane versus cut bank sites and a picture of a huge rock. It gives scatter plots for soil horizon and flood deposit.

(a) Map of the Gouritz River catchment and sampling sites. The winter rainfall zone (WRZ), summer rainfall zone (SRZ), and year-round rainfall zone (YRZ) are indicated. (b) Variations in δD of the n-C31-alkane (‰ VSMOW) in the distinct horizons of paleoflood vs. soil formation horizons. (c) Gouritz River tributary cut bank is depicted as an example illustrating these alternating horizons. Figure modified after Hahn et al. (2017)

Full size image

For the interpretation of layers with deuterium-depleted plant waxes (Fig. 28.10), in the marine sediment core GeoB18308-1 in Mossel Bay offshore the Gouritz River mouth (Hahn et al. 2017), this information was particularly valuable. These layers could be linked to an increase in Southern Hemispheric Westerlies-related precipitation events in the upper Gouritz River catchment. In this manner, the Gouritz River catchment study led to a better understanding of variations in past regional climatic conditions.

Fig. 28.10
A line graph over an illustration of a log of wood depicts cal years B p. It features event deposit, M C A, lowland and rainfall.

Core log and deuterium isotopes of plant waxes in GeoB18308-1. The core log shows internal structures with recorded colors demonstrating the presence of an event deposit. The deuterium-enriched values during the medieval climate anomaly (MCA) indicate humid conditions during this interval

Full size image
Fig. 28.11
A map of the southern part of Africa marks the drainage systems. Below are seven line graphs and seven corresponding box and whisker plots.

(a) Study area of case study 3 in south-eastern Africa showing relevant drainage systems plotted with the sampling locations. The river catchments depicted on the map are (from

Full size image

3.3.3 Case Study 6: South-Eastern African Coast

A comparative study of catchment samples (riverbank sediments, flood deposits, suspension loads, and soils) and offshore deposits along the south-east African coast was used to investigate the source of various terrigenous sediment fractions and to delimit the movement of sediments along the shelf initiated by the Agulhas Current (Hahn et al. 2018). In the river catchments along the coast, subtle differences in sampling locations and associated environments (e.g., riverbank sediments, flood deposits, suspension loads, and soils) were reflected in the various proxy-indicators (Fig. 28.11). Large-scale trends, such as the climatic and environmental shift from the temperate winter rainfall zone in the southwest to the tropical summer rainfall zone in the northeast (Fig. 28.1), were also reflected in the data (Fig. 28.12). Plant wax δD signatures are indicative of increasing rainfall amounts toward the north concurring with a change from subtropical to tropical climate. Pollen assemblages and plant-wax δ13C signatures document a shift from a mixed C4/C3 signature farther north in the subtropical grasslands/savanna to Mediterranean and mountain shrublands vegetation (C3 dominated) in the southern Cape and Drakensberg regions, respectively. The petrographic, chemical, heavy-mineral, and bulk mineralogical composition of the riverine material reflects catchment geology, which consists of metamorphic and igneous rocks in the north (Kaapvaal Craton and Bushveld Igneous Complex), and mainly sandstones in the south (sedimentary rocks of the Cape and Karoo Supergroups). Offshore sample signatures were comparable to those of the adjacent river systems. It was therefore inferred that the influence of the Agulhas Current affects sediment deposition and distribution only seaward of the mid-shelf (Fig. 28.11). It was concluded that sediments on the inner shelf, especially from locations leeward of coastal protrusions, are protected from erosion and redistribution. These sediments thus constitute valuable archives of local climatic change.

Fig. 28.12
A map of the southwestern Africa has an arrow of the eastern coast, moving in a circular clockwise direction to indicate the elagoa bight eddy. b is a bar graph that plots E M versus four parameters. It gives values for Matola, Limpopo, Incomati, and Lusufu.

(a) Overview map of the Incomati, Limpopo, Lusutfu, and Matola catchments including sample locations from terrestrial sampling campaigns as well as the Delagoa Bight with GeoB surface sediment sites. The dashed line marks the shelf break. (b) Contribution of empirical end-members of the Limpopo, Incomati, Matola, and Lusutfu catchments to sea-floor GeoB surface samples from the Delagoa Bight. The pathway of the sediment drift in the Delagoa Bight is indicated by the red arrow marking the Delagoa Bight Eddy. Figure after Schüürman et al. (2019)

Full size image

3.3.4 Case Study 7: Delagoa Bight

Catchment samples from the river systems (e.g., Incomati, Matola, and Lusutfu rivers) that discharge into the Delagoa Bight were studied to determine the provenance of sediment deposits in the Delagoa Bight (Schüürman et al. 2019). The terrigenous material in sediments on the Delagoa Bight near Maputo, Mozambique, has several potential source areas including the coastal river catchments of the Incomati, Matola, and Lusutfu rivers and the large, interior catchment of the Limpopo River. A selection of trace element concentrations in river sediment samples was used to determine end-members for the four river catchments. All element concentrations were normalized by aluminum to avoid the dilution effects by the marine fraction. The entire dataset was scrutinized by multivariate analysis of variance (MANOVA) in order to identify the elements most useful for the regional end-member analysis. Relative end-member contributions to core top samples from the shelf were then used to trace inorganic sediment deposition in the Delagoa Bight (Fig. 28.12). The results show that the local cyclonic circulation induces a strong eastward sediment drift in part preventing sedimentation in the bight. This was essential for the paleoclimatic interpretation of several marine sediment cores located in the bight (Miller et al. 2020; Hahn et al. 2021). In particular, the low relative contribution of the Limpopo River was unexpected.

3.3.5 Case Study 8: Umzimbuvu River

A multiseason provenance study of the Umzimbuvu River (Fig. 28.13; (Frankland 2020; Sect. 28.3.1) was conducted to understand catchment dynamics and provide taphonomic basis to the interpretation of the marine sediment core collected offshore the river mouth (GeoB20623-1). The Umzimvubu is a large, undammed catchment (19,852 km2) on the east coast of South Africa comprising five major tributaries, which have their headwaters in the Drakensberg Mountains. The main stem of the catchment is the Umzimvubu River, which flows over ~400 km through deeply incised river valleys from the source through the coastal belt before discharging into the Indian Ocean at Port St. Johns. The catchment transitions between the grassland biome at higher altitudes to the savanna biome at intermediate altitudes and finally the Indian Ocean Coastal Belt biome at the coast.

Fig. 28.13
Four box and whisker plots along with a heat map of southern Africa.

Box and whisker plots illustrating the differences in sediment geochemistry and palynology between the upper and lower catchment areas of the Umzimvubu River. The minimum, maximum, median, upper, and lower quartile δ13C, δD, and Fe/K values of upper and lower catchment samples are shown. The geochemistry plots include river bed sediment sampled from each site during the dry (July), intermediate (September), and wet (November) seasons of 2017. Phragmites/Cyperaceae pollen ratio is based on 18 viable catchment samples across the three seasons. The upper and lower catchments are defined in the accompanying elevation map by the dotted line marking 500 m a.s.l. δ13C values of the C31 n-alkane are presented in ‰ VPDB and δD of the C31 n-alkane in ‰ VSMOW. Figure after Hahn et al. (2021)

Full size image

Eleven sampling sites were identified at the most accessible locations representing the full spatial extent of the Umzimvubu Catchment and capturing all five major tributaries (Mzintlava, Umzimvubu, Kinira, Thina, and iTsitsa). To capture temporal sediment dynamics within the catchment, three sampling campaigns were conducted to sample the dry (June), intermediate (September), and wet (November) months according to monthly rainfall for the region. Four types of samples were taken at each location to determine the provenance dynamics of sediment, hydrologic, and vegetation signals within the catchment, viz., suspension load, riverbed sediment, dried puddle sediments, and water isotope samples. Suspension load samples were collected using a portable water pump deployed off of a bridge at the center of the river. Riverbed sediment was collected using grab samples. Dried puddle sediments were sampled from dried flood deposits where there was evidence of potential past flood events. Water isotope samples were collected downstream of each sample site for δ18O and δ2H analysis. These were ultimately excluded from the overall dataset as they did not show a major relation to the inorganic and organic signal provenance, nor did they have a strong correlation to the other proxies. Collectively, the catchment-wide sampling of river bed and suspension load sample was used to distinguish upper from lower catchment sediment signatures, which could be applied downcore on a sediment archive offshore the river system (GeoB20623-1, Hahn et al. 2021).

In terms of geochemistry, the catchment study showed that there was a distinct difference in the sediment signatures of the upper and lower catchment. Upper catchment sediments are characterized by high δ13C values (due to the prevailing C4 grassland vegetation), low δD values (due to the altitude effect; Gonfiantini et al. 2001), and higher Fe/K value (due to the high chemical weathering in the arid upper catchment) (Fig. 28.13). This is relevant for the interpretation of proxy indicators in the sediment core, since a greater proportion of upper catchment material in the sediments transported into the ocean occurs during high rainfall events.

For the palynological aspects of the study, suspension load samples yielded poor pollen concentration; thus, river bed samples were selected as a more appropriate sample type for pollen analysis. These surface river bed samples are not indicative of the time of sampling, but rather reflect pollen accumulation over an extended period, likely spanning several months or even years depending on the site. Moreover, samples taken in the downstream catchment represent a greater allochthonous signal, incorporating pollen grains transported from higher up the catchment. A further caveat to catchment pollen interpretation is the fact that seasonal variations in the pollen signal, such as those observed for Zea mays (maize), are indicative of flowering season of the parent vegetation, rather than of any seasonal climatic variations influencing vegetation type. This leads to the broader consideration of pollen as a direct indicator for parent vegetation, but an indirect indicator of prevailing climate, with inherent lags in the timing of vegetation response to climatic change. The catchment pollen analysis samples were therefore grouped across seasons, since seasonal flowering patterns do not inform paleo-vegetation or paleo-climatic interpretation. The aim of the catchment pollen research was to determine whether there is a coherent pollen signal indicative of the dominant biomes (Mucina and Rutherford 2006) across the Umzimvubu River catchment, viz. grassland in the upper catchment area (n = 14), savanna at intermediate altitudes (n = 12), and Indian Ocean Coastal Belt biome along the coastline (n = 6). Perhaps unsurprisingly, the pollen signature was dominated by a few ubiquitous taxa, including grasses (Poaceae excluding Phragmites-type) and sedges (Cyperaceae), which tend to be overrepresented in South African pollen records from wetlands. This overrepresentation obscured any clear biome-related signals in the vegetation. The pollen signal also revealed the presence of pollen from exotic pine trees (Pinus spp.) in the upper catchment areas, reflecting the transformed nature of the catchment area, with parts of the natural grassland vegetation having been afforested by commercial plantations. A multiproxy comparison of upper and lower catchment samples showed clear altitudinal differences in organic and inorganic geochemical proxies, but also in the ratio of Phragmites/Cyperaceae pollen (Fig. 28.13; Hahn et al. 2021). In the upper (lower) catchment areas, Phragmites-type pollen is on average less (more) abundant than Cyperaceae. Phragmites australis is an aquatic/semiaquatic reed that grows in shallow water, whereas Cyperaceae (sedges) is a more ubiquitous taxon, with sedges growing in a range of habitats from wetlands to grasslands and forest fringes (Gaigher 1990; Archer 2000). The higher Phragmites/Cyperaceae pollen ratio in the lower altitude parts of the catchment likely reflects the prevalence of P. australis in the floodplain areas near the river mouth, and more humid climatic conditions than the higher altitude sites.

The paper by Hahn et al. (2021) was able to use these insights from modern catchment geochemistry and pollen sampling to apply the Phragmites/Cyperaceae pollen ratio, the isotope geochemistry, and the elemental composition downcore. The Phragmites/Cyperaceae pollen ratio is used as a moisture indicator to interpret the offshore marine fossil pollen record of core GeoB20623-1 recovered offshore the Umzimvubu River mouth, with ratio increases (decreases) suggesting more (less) humid conditions. Likewise, as a result of the catchment study, high δ13C values, low δD values, and higher Fe/K values in core GeoB20623-1 are associated with an increased input of upper catchment sediments. Since a greater proportion of upper catchment material in the sediments transported into the ocean occurs during high rainfall periods, intervals with high δ13C values, low δD values, and higher Fe/K values in core GeoB20623-1 are interpreted as humid phases.

4 Synthesis and Practical Recommendations

Summing up, we state that the suite of source-to-sink and depositional studies described in this chapter has been essential for the accurate interpretation of environmental conditions from sediment cores taken in and around southern Africa. Exploring depositional systems and the stratigraphy of marine environments at the sediment sink has proven indispensable for locating sediment bodies suitable for coring and therefore consequent analysis of past changes. This is particularly true for current swept shelves, such as the South African shelf, where great expanses of the seafloor are devoid of sediment. Furthermore, seismic stratigraphy can give insights into changing sedimentary conditions that will entirely change the interpretation of downcore proxies. Eminent examples of this are the lake sediment sequences dating from lower sea level conditions found in several Agulhas Bank/Paleo-Agulhas Plain sediment cores. In more practical terms, seismic stratigraphy studies can enable the choice of suitable coring equipment (gravity corer, vibrocorer, box corer, and multicorer) depending on the sediment grain size and texture. Possible penetration depth and thus barrel lengths for coring can also be estimated using seismic surveying. For the interpretation of the proxies measured on the retrieved sediment cores, we consider catchment studies a crucial. The examples outlined in these studies show how the suite of organic and inorganic indicators can vary with changes in the source area that can only be understood, and thus correctly interpreted, in a framework of a catchment area study. Considering source-to-sink study conduct, after this work, we recommend choosing sampling locations before and after tributaries and/or confluences and analyzing a suite of different samples (suspension load, river bed, flood deposits) from each site. Organic parameters can be excellent indicators of vegetation present at the site (river bank samples) as well as upstream of the confluence or tributary (suspension load samples). Compound-specific hydrogen isotope analyses can inform on present and past eco-hydrologic status. Inorganic parameters can give insights into changes in sediment provenance (geological formations) and/or shifts in weathering regimes at or upstream of the site. As best practice for future paleo-environmental studies, it is recommended that sediment cores are interpreted in the context of their depositional conditions explored using seismic surveying techniques and that catchment samples are analyzed in order to shed light on the processes that influence proxy parameters measured on allochthonous sedimentary components.

5 Conclusions

Using examples from southern Africa, we illustrate that catchment studies and hydroacoustic/stratigraphic studies are essential for reliable paleo-climatic and paleo-environmental interpretations of paleo-archives:

  • Paleoclimate studies are important for anticipating present and future responses to climate and climate change as they provide reference information on past climate changes and variability.

  • Often, several rivers discharge into the same embayment and/or strong oceanic currents may displace sediments. In both cases, it is vital to determine from which river systems sediments originate from prior to commencing environmental interpretations.

  • Rainfall is not always evenly distributed over a catchment area. By identifying source areas within a catchment, a more locally constrained climate reconstruction can be made.

  • Hydroacoustic surveying is vital for identifying suitable coring locations and contextualizing depositional processes.

  • The present is not always an analogue of past conditions. This work has shown the value in comparison/verification at different temporal scales.

The overall messages for land use management and environmental protection policies emerging from this chapter are as follows:

  • Reliable records of past climate and ecologic changes are indispensable for understanding present and future responses to climate and climate changes.

  • Vegetation patterns during the Holocene differed across southern Africa, as did the past climate. Management strategies thus need to be regionally specific and consider region-specific variabilities and sensitivities.