Water Retention Landscapes in Fynbos and Renosterveld Restoration

A web-based search of academic and non-peer-reviewed sources found no work that applies or evaluates a Water Retention Landscape practice as part of restoring fynbos or renosterveld in the Cape Floristic Region (CFR). Claims about such a combination currently rest on adjacent CFR studies and on analogues from other Mediterranean-type climates.

Overview

A Water Retention Landscape (WRL) is a landscape restored so that rainfall is captured, held and released slowly. The practice comes from the Kravčík, Holzer, Jehne and Ruddick tradition and works at catchment scale [MS]. In situ rainwater harvesting (ISRWH) is the term South African agricultural literature usually applies to field-scale techniques such as tied ridges, basin tillage, mulching and contour bunds, mostly studied for crop production in semi-arid areas [TBV]. The two terms belong to different literatures, so the search covered a family of practices: WRL, ISRWH, keyline design, swales, contour bunds, level sills, leaky weirs, gully plugs, erosion-control structures, wetland rehabilitation and restoration hydrology.

The search found no source that pairs any practice in this family with fynbos or renosterveld restoration as its stated objective. The nearest evidence sits in four groups of CFR work: hydrology and erosion studies in renosterveld, erosion control and watercourse rehabilitation, alien plant clearing and fynbos catchment hydrology, and provincial ecological infrastructure policy. Three international analogues, from Portugal, Chile and Spain, add practice-level detail from Mediterranean-type climates. None of this work tests a water-slowing earthwork at a CFR restoration site.

The finding describes what a web search reached on 12 October 2026. The research run did not query Scopus or Web of Science directly, several sources were paywalled, and Afrikaans coverage was thin (three search strings). This page therefore holds the absence of a direct pairing as a working hypothesis. A search through the databases themselves, and inquiries to the authors and organizations cited below, would test it.

Scope and method

The search covered fynbos and renosterveld in all their named vegetation types, including old fields and transitional states. It accepted every source type: peer-reviewed articles, theses, conference papers, agency and NGO reports, consultancy reports, practitioner case studies and project websites. Findings were sorted into three tiers.

• Tier 1: the work applies or evaluates a WRL or ISRWH-family practice as part of restoring fynbos or renosterveld, or describes such a project.
• Tier 2: adjacent work in the CFR, such as alien plant clearing and streamflow, erosion and gully control, soil moisture and infiltration studies, and water-slowing earthworks on CFR farmland without a restoration outcome.
• Tier 3: analogues from other Mediterranean-type ecosystems, kept separate and assessed for transferability.

After the run, each citation was checked against an online record on 12 October 2026 where one could be found. A figure, quotation or attribution that could not be traced to its cited source carries a [TBV] tag, and the Provenance section lists what changed.

Tier 1: direct pairings

None found. The run searched every direct combination of “water retention landscape”, “in situ rainwater harvesting”, “keyline design”, “contour bunds”, “swales” and “gully rehabilitation” with “fynbos” and “renosterveld”, in English, with three further strings in Afrikaans. Each combination returned nothing relevant.

Tier 2: adjacent work in the CFR

Renosterveld hydrology and erosion

De Waal (2023) is a Stellenbosch University doctoral thesis, supervised by Adriaan van Niekerk and J. Miller, on the Overberg renosterveld landscape and the Bot River catchment. It combines a hydrological model with a sediment delivery component and a one-year river monitoring program that used ion chromatography and stable isotopes [IC]. Modeled soil erosion in the catchment rose from 22 t/km²/year under natural vegetation [IC]; the run reported 490 t/km²/year under 2018 land use [TBV]. Nutrient loading in the river is linked to agricultural land use, and nitrate and nitrite nitrogen periodically exceeds guidelines for aquatic ecosystems [IC]. Renosterveld fragments remediate polluted agricultural runoff [IC]. The run also reported higher soil carbon in fragments than in cultivated fields [TBV].

The thesis monitors and models; it tests no intervention. It establishes that agricultural transformation has changed hydrology and water quality in renosterveld and that remnant vegetation provides a measurable service. It leaves open whether any water-slowing or infiltration-enhancing intervention could reverse the damage.

Vermeulen (2010), a Stellenbosch MSc thesis found during verification, compared a mature renosterveld stand with a nearby wheat field in the Berg River catchment and reported large differences in groundwater levels and salinity between the two [IC]. It addresses the water balance consequences of land use change in renosterveld, and tests no intervention.

Erosion control and watercourse rehabilitation

The Overberg Renosterveld Conservation Trust (2018) reports on a three-year watercourse restoration project funded by the WWF Nedbank Green Trust, which began in September 2015 [MS]. The report describes mulching to stabilize livestock pathways and portable electric fencing to keep livestock off recovering areas [MS]. The run also listed gabions, soil blankets, drift fences, drainage hollows, alien plant clearing and ecological burns [TBV]; these were not found in the text retrieved during verification. The Trust states that passive restoration of the watercourses is insufficient, and that watercourse restoration is extremely beneficial to renosterveld [MS]. The run found no reported infiltration, runoff or soil moisture measurements [TBV].

The report shows that a CFR conservation organization has installed erosion-control measures in renosterveld watercourses and judges them beneficial. Whether the measures change infiltration, runoff or soil moisture, and whether they count as a WRL or ISRWH intervention, remains open. The benefit claim comes from the implementing party.

The run also cited an undated South African National Biodiversity Institute (SANBI) page on the Working for Wetlands program, which rehabilitates wetlands with erosion-control structures and alien plant clearing, including at sites in the Kromme and Sand River catchments [TBV]. This pass did not open that page. The run located no published hydrological outcomes for CFR wetland rehabilitation [TBV].

A 2022 review titled “Intersecting views of gully erosion in South Africa” reports that human activity is a major contemporary driver of gullying, that gullying is more prominent in central and eastern South Africa, and that medium- and long-term studies are needed [IC]. The run attributed the review to Panter and Ruwanza and credited it with guidance on rehabilitation techniques and the use of spekboom for revegetation. Neither the authorship nor those details could be confirmed from the abstract record [TBV]. Spekboom is a subtropical thicket species, so its relevance to CFR gullies is doubtful [inference].

Alien plant clearing and fynbos catchment hydrology

Van Wilgen and Wannenburgh (2016) review the Working for Water program, which clears invasive alien plants to conserve water and create employment [IC]. The program began in the mid-1990s because alien trees use more water than the vegetation they replace [IC]. The run attributed several figures to the review: about 12,000 ha cleared in the Berg River Dam catchment, modeled gains of 9.0 to 12.7 million m³/year, more than one million hectares cleared overall, and about 20,000 people employed annually. None could be traced to the article. A secondary record gives a cleared area above 2.5 million hectares, which differs from the run’s figure [TBV].

The mechanism is reduced transpiration after plants are removed. It adds no structure that slows, spreads or sinks water. The review establishes that fynbos catchment hydrology responds to vegetation management and that the CFR has an institutional apparatus for catchment-scale restoration. Whether earthworks would add to the gains from clearing remains open. This wiki holds source pages on the related streamflow and invasion literature (see Related).

Scott and Van Wyk (1990) examined the first-year hydrological response to a high-intensity wildfire in the Jonkershoek research catchment, about half fynbos and half Pinus radiata plantation [IC]. The run reported findings on soil wettability and infiltration in fynbos soils; those could not be traced to the paper [TBV]. A 1994 University of Natal thesis record found during verification reports that neither of two fynbos catchments showed a change in storm flow after fire, while two plantation catchments showed large increases in storm flow and sediment yield [IC]. These are process studies with no intervention.

The Greater Cape Town Water Fund (GCTWF) was established by The Nature Conservancy with the City of Cape Town and partners to restore the catchments that supply the city [IC]. Its business case frames ecological infrastructure restoration around seven priority sub-catchments and states that restoration supplies water at roughly one-tenth the unit cost of alternatives [MS]. The practice funded is invasive alien plant clearing. The Cape Water Performance-Based Bond ties investor returns in part to independently verified outcomes from that clearing [MS]. The run reported that the Fund’s water gains are modeled and not measured at catchment outlets, and gave program figures from a 2026 press release; neither could be confirmed [TBV].

Ecological infrastructure policy

The run cited a 2025 Western Cape Government and CapeNature biodiversity spatial plan as recognizing ecological infrastructure as a response to water risk. This pass could not confirm the document’s title, date or content [TBV]. SANBI’s own page on the GCTWF business case does use the ecological infrastructure frame for the Cape Town catchments [IC]. Neither source specifies or evaluates a WRL practice.

Tier 3: international analogues

Tamera, southern Portugal

Tamera, at Monte do Cerro in the Alentejo, has implemented a Water Retention Landscape since 2006, using keyline design, terracing, swales, decentralized lakes and ponds, reforestation, mixed groundcover and holistic grazing [MS]. From 2006 to 2015 the project created 29 lakes and retention spaces and raised its open water area from 0.62 ha to about 8.32 ha [MS]. Records give the site as 133 ha (a 2012 figure) and as 154 ha, so the area depends on the source. Both case studies come from European Union platforms that summarize the project’s own documentation, so the outcome claims of water self-sufficiency, stabilized groundwater and increased biodiversity are the project’s own [MS]. The run reported eutrophic lakes in 2012 to 2014 monitoring and no peer-reviewed hydrological measurements of infiltration, runoff or recharge [TBV].

The vocabulary is exactly WRL. Transferability to the CFR is partial. The winter-rain Mediterranean climate and the dryland setting are shared. The run described the vegetation as montado or dehesa (oak savanna), and the soils as schist and granite [TBV]. Fynbos and renosterveld are fire-driven shrublands on different soils. The run’s view that lake-based retention would suit few CFR catchments is an inference without a cited source [inference].

Keyline design in central Chile

Toro-Mujica (2023) is a conference paper from the XXV International Grassland Congress. It reports a trial on three farms in the semi-arid O’Higgins region, where keyline design influenced soil moisture and temperature and the crude protein, neutral detergent fiber and acid detergent fiber of the grasses [IC]. The run reported gains of 44 percent in forage height and 25 percent in yield; the abstract does not give these figures [TBV]. The vegetation is naturalized annual grassland, and the measures are soil and forage variables, so hydrological fluxes were not measured. Transferability is partial: the Mediterranean rainfall regime is shared, while fynbos and renosterveld are perennial, fire-dependent shrublands and many CFR soils are shallow and rocky [inference].

Shrub cover and erosion in Spain

Andreu and colleagues (1998) monitored 56 erosive rain events between 1988 and 1995 on experimental plots, comparing three planted shrub species (Medicago arborea, Atriplex nummularia and Psoralea bituminosa) with natural matorral vegetation [IC]. Medicago arborea gave the best protection and cut sediment production by more than 58 percent against bare soil. Natural vegetation cut it by almost 75 percent [IC]. The study shows that intact shrub cover protects soil. It tests planting, with no earthworks involved. The mechanism probably transfers to fynbos and renosterveld, where the protective role of shrub cover is already established [inference].

Evidence table

Counts are the sources located in this search. A zero records that no source was located.

Practice	Fynbos	Renosterveld	CFR-wide or national	International analogues
Water Retention Landscapes	0	0	0	1 (Tamera) [MS]
In situ rainwater harvesting	0	0	0	0
Keyline design	0	0	0	1 (Chile, conference paper) [IC]
Swales	0	0	0	1 (Tamera, one component of a package) [MS]
Contour bunds and berms	0	0	0	0
Gabions, check structures and gully rehabilitation	0	1 (Trust report) [MS]	2 (Working for Wetlands page, national gully review), details unconfirmed [TBV]	0
Vegetation as erosion and runoff control	0	0	0	1 (Spain) [IC]
Restoration hydrology (monitoring or modeling of clearing or land use change)	2 (Working for Water review, GCTWF) [IC, MS]	2 (De Waal, Vermeulen) [IC]	0	0
Infiltration and soil wettability baselines	1 (Scott and Van Wyk, fire; a 1994 thesis record) [IC]	1 (Vermeulen, soil water and salinity) [IC]	0	0
Ecological infrastructure policy	0	0	2 (SANBI page [IC], provincial plan [TBV])	0

For renosterveld, the De Waal thesis is the most substantial source. For fynbos catchment hydrology, the Working for Water literature is the deepest, and it concerns alien clearing. For WRL practice, Tamera is the best documented case this search found, and its outcome claims come from the project and the platforms that summarize it.

Gap statement

What no located study has done:

• Tested whether water-slowing earthworks (swales, keyline patterns, contour bunds, leaky weirs, gully plugs) change infiltration, soil moisture or baseflow at fynbos or renosterveld restoration sites.
• Measured the downstream hydrological effect of erosion-control structures such as those installed in renosterveld watercourses.
• Compared WRL or ISRWH practices with standard CFR restoration approaches: alien clearing, fire management and grazing exclusion.
• Assessed how earthworks interact with the fire regimes of fynbos and renosterveld, which burn periodically. The run flagged this as a likely interaction, and it remains an inference.

What is asserted without measurement: the Overberg Renosterveld Conservation Trust’s benefit claim for watercourse restoration; Tamera’s claim of stabilized groundwater and water self-sufficiency; and, per the run, the modeled water gains from alien clearing at some sites [TBV].

What a first study would measure (design proposed by the research run). A paired-catchment or before-after-control-impact design along a degradation gradient in renosterveld or fynbos, recording:

• infiltration rate, by double-ring or tension infiltrometer;
• soil moisture at several depths, by capacitance or time-domain reflectometry sensors;
• runoff volume and peak flow, by flumes or weirs;
• sediment yield, by traps;
• groundwater level, by piezometers;
• streamflow or baseflow at the catchment outlet;
• vegetation cover, species richness and indicator species;
• cost against alien clearing and fire management.

The run proposed at least three years of monitoring to capture rainfall variability, and one fire cycle where fire is part of management.

What this means for BioHub practice

A BioHub in a fynbos or renosterveld BioRegion can cite CFR evidence for the water gains from alien plant clearing, and for the hydrological damage that agricultural transformation does to renosterveld. A claim that water-slowing earthworks add measurable water yield in these vegetation types has no CFR study behind it. Until one exists, a service built on that claim fits the speculative pending research category of the Tenderable Services Portfolio readiness diagnostic [inference]. The finding also marks an opening: a measured first study would be new evidence for the field.

Related

• Water Retention Landscapes
• Water Retention Landscape (glossary)
• Integrated Landscape Management
• Performance-Based Water Bonds
• Bankable Service Alignment Template

Sources

Related works that have source pages on this wiki:

• Van Wyk (1987), afforestation and streamflow in the Western Cape
• Le Maitre et al. (2019), invasive Australian acacias and the fynbos biome
• Van Wilgen et al. (2008), the economic consequences of alien plant invasions
• Mills, Cowling et al. (2013), restoring transformed renosterveld

Works cited in this brief. Source pages for these have not yet been created.

• Andreu, V., Rubio, J.L., Gimeno-García, E., & Llinares, J.V. (1998). Testing three Mediterranean shrub species in runoff reduction and sediment transport. Soil and Tillage Research, 45(3-4), 441-454. doi:10.1016/S0933-3630(97)00040-8
• De Waal, J. (2023). Adverse impacts of agricultural expansion on hydrological and nutrient dynamics in a Renosterveld landscape: can natural vegetation offer solutions. PhD thesis, Stellenbosch University. https://scholar.sun.ac.za/handle/10019.1/617
• Greater Cape Town Water Fund business case, The Nature Conservancy (publication year differs between records).
• Intersecting views of gully erosion in South Africa (online 8 December 2022). Authors and journal details to be confirmed.
• Overberg Renosterveld Conservation Trust (2018). Three years later: here’s how watercourses helped protect Renosterveld. https://overbergrenosterveld.org.za/three-years-later-heres-how-watercourses-helped-protect-renosterveld/
• Natural Water Retention Measures platform. Water retention spaces, reforestation and grazing management in southern Portugal. https://www.nwrm.eu/node/3561; and European Climate Adaptation Platform, Tamera water retention landscape to restore the water cycle and reduce vulnerability to droughts.
• SANBI. (undated). Working for Wetlands. [TBV]
• SANBI. A Greater Cape Town Water Fund for Ecological Infrastructure. https://www.sanbi.org/cape/a-greater-cape-town-water-fund-for-ecological-infrastructure/
• Scott, D.F., & Van Wyk, D.B. (1990). The effects of wildfire on soil wettability and hydrological behaviour of an afforested catchment. Journal of Hydrology, 121(1-4), 239-256.
• The hydrological effects of fire in South African catchments (1994). Thesis record, University of Natal repository: https://ukzn-dspace.ukzn.ac.za/handle/10413/9404 (author not shown in the record viewed).
• The Nature Conservancy (2026). Press release on the Cape Water Performance-Based Bond, as cited by the run. [TBV]
• Toro-Mujica, P. (2023). Effect of the hydrological keyline design on the growth and quality of the naturalized grassland in the dry central zone of Chile. XXV International Grassland Congress. https://uknowledge.uky.edu/igc/XXV_IGC_2023/Utilization/3
• Van Wilgen, B.W., & Wannenburgh, A. (2016). Co-facilitating invasive species control, water conservation and poverty relief: achievements and challenges in South Africa’s Working for Water programme. Current Opinion in Environmental Sustainability, 19, 7-17. doi:10.1016/j.cosust.2015.08.012
• Vermeulen, T. (2010). Plant water relations of Elytropappus rhinocerotis with specific reference to soil restrictions on growth. MScAgric thesis, University of Stellenbosch. https://scholar.sun.ac.za/handle/10019.1/5416
• Western Cape Government and CapeNature (2025). Biodiversity spatial plan and protected area expansion strategy, as cited by the run. [TBV]

Provenance

Extracted from research-brief-wrl-renosterveld.md, the output of a Claude Research run on 12 October 2026 that followed a prompt written for this brief. Editorial decisions:

• Spelling converted from British to American, with the network’s voice conversions applied.
• epistemic_status set to working-hypothesis. The central finding is an absence reported from a web-based search, and the run described its null results as evidence of absence. That framing was removed.
• Citations confirmed against online records on 12 October 2026: De Waal (2023), Van Wilgen and Wannenburgh (2016), Scott and Van Wyk (1990), Toro-Mujica (2023), Andreu et al. (1998), the Overberg Renosterveld Conservation Trust page and the Tamera case studies.
• Corrections: De Waal is the sole author of the thesis, and Van Niekerk and Miller are its supervisors, not co-authors. The run’s DOI for Van Wilgen and Wannenburgh was wrong. The run’s DOI and findings for Andreu et al. were replaced with those from the published abstract. Toro-Mujica (2023) is a conference paper from the International Grassland Congress, not an institutional report. The Tamera site area is reported as 133 ha in one record and 154 ha in another.
• Downgraded to [TBV]: the run’s figures and quotations for the Working for Water review, the Scott and Van Wyk findings, the gabion and drift-fence details and quotation from the Trust report, the Toro-Mujica yield figures, the gully review’s authorship and rehabilitation guidance, and every item drawn from the Working for Wetlands page, the provincial plan and the 2026 press release.
• Added from verification and absent from the run: Vermeulen (2010), the 1994 University of Natal thesis record, and SANBI’s page on the GCTWF business case.
• Evidence table recounted. The run’s table placed De Waal in a fynbos cell, counted Spanish shrub planting as gully work, and omitted swales at Tamera.
• Omitted: the run’s names list, which carried contact details, social-media attribution for a practitioner and unconfirmed project-lead names. Outreach leads sit in the workbench, outside the published site.
• The run’s closing cross-reference list was dropped.

The research run did not query Scopus or Web of Science directly. A database search would test the central finding.