Small Farm Dams: A TICKING TIME BOMB? - WRC

30 Farm dams

Small
farm dams:
a ticking time bomb?

Many of South Africa’s small farm dams have become mere sediment traps threatening the
country’s larger storage reservoirs, as research in the Karoo indicates, writes John Boardman,

Ian Foster, Kate Rowntree, Tim Mighall and Tony Parsons.

South Africa is heavily dependent with water deficits or prone to drought. shown by a 40-year-old record was likely
on storage reservoirs to maintain Their role has been one of harvesting to reflect two key factors: a reduction
reliable water supplies at times of runoff to provide water for stock or for in the amount of sediment available for
water stress. Recent estimates suggest irrigation. While major dam projects have erosion, and sediment trapping by farm
that the country already allocates over received much publicity and their his- ponds. The latter contention is
98% of its available water resources and torical significance is well documented, supported by work done by the authors,
that many storage reservoirs are accu- small farm dams have been neglected in as described below.
mulating sediment at a pace that may the literature and their wider role in the
make water provision at current rates landscape largely ignored. The data presented point to small farm
unsustainable. dams in badly eroded dryland areas
Sediment traps playing a significant role in trapping
Across South Africa a plethora of small sediment. Although the exact timing of
farm dams and reservoirs have also The longest monitored sediment yield sediment yield increases and the amount
played an important role in the eco- record in South Africa is from the lower of sediment deposited in these small
nomic development and sustenance Orange River. It is suggested that the reservoirs varies in relation to local fac-
of rural areas – especially those regions decline in sediment yields after 1950 tors, such as the presence or absence of

The Water Wheel Jul/Aug 2009

Farm dams 31

badlands, the extent of rainfed wheat Digging for evidence contained in these reservoirs provides
production, the role of fire, both natural an historical record of the physical proc-
and anthropogenic, and the intensity About ten years ago, as part of a project ess of catchment erosion and sediment
of grazing, the pattering of the authors’ on erosion, runoff and overgrazing in the yield. Where documentary evidence has
findings shows a consistent increase from uplands of the Karoo, the authors started provided the age of dam construction
the early twentieth century, followed by a to document small dams. Their interest or re-construction, by coring at several
declining but still high rate of sedimenta- in the history of erosion led them to look points in the reservoirs the project team
tion after the middle of the century. at several dams in some detail in order has been able to estimate the volume of
to reconstruct the depositional his- stored sediment. They have then been
Many of these dams are now full and tory of their sediments and identify the able to estimate sediment yields for
have lost their trapping capacity; others drivers of erosion. This work fitted into specific periods since dam building by
have breached and are beginning to lose the ongoing debate on ‘desertification’ using sediment characteristics to date
their sediment to the downstream sys- of the Karoo, but shifted the focus from the different layers.
tem. Do we now have a situation where vegetation change to broader land-
the myriad of small dams that protected scape change and, in particular, to the Dating the sedimentary layers has relied
South Africa’s larger storage reservoirs erosional history of badlands and gullies on 137Cs, first seen in 1958 and peaking in
are now providing many small stores of (dongas). 1965, on variations in unsupported 210Pb
sediment threatening those selfsame res- activities with depth in the sediment
ervoirs? This article provides evidence to While documentary evidence (farm column, and on fine gravel layers which
suggest that this is indeed the case and diaries, oral records etc) had provided can be association with known storms,
points to the need for more in-depth important information on the history of for example, four extremely wet days
research into this sediment ‘time bomb’. land management, the sediment from 1 to 4 March 1974 with 209 mm of

South Africa’s Van Ryneveldspas Dam on the Sunday’s River at Graaff Reinet has a catchment area of 12,382 km2. By 2009 it is esti-
mated that there will be around 49,1 million m3 of sediment in the reservoir, with a remaining storage capacity of only 29 million m3.

The Water Wheel Jul/Aug 2009

32 Farm dams

Above: Compassberg Dam breach.
Initial breaching occurred in 2000 and this
photo, taken in December 2003, shows the
establishment of a new gully that is eroding
previously stored sediment.

Right: By July 2006, a well established gully
was cutting back into the stored reservoir
sediments. The gully head has extended
around 25 m since 2003.

Compassberg Farm dam rain that brought widespread flooding to
the Craddock-Graaff-Reinet region.
Catchment area: 6,33 km2
Relative relief: 662 m Finding old dams has been a chal-
Average annual rainfall: 498 mm lenge and, to date, with the focus on
Land use: Sheep grazing and former cultivated land on valley bottom the Karoo, the oldest investigated is at
Reservoir area: 3,37 ha Cranemere (1843), located some 10 km
Dam age: ca. 1935 west of Pearston in the Eastern Cape and
Current status: Full of sediment; breached in 2000 made famous by Eve Palmer in her book
Average sediment yield: 490 t/km2/y1* The Plains of Camdeboo. The project
Total sediment volume: ca. 50 000 m3 team has concentrated on small catch-
ments where the history of land use
*Takes into account estimated trap efficiencies is known and where, because of their
small size, most eroded sediments are

The Water Wheel Jul/Aug 2009

Farm dams 33

not stored on footslopes or hillslopes Ganora farm dam
but reach the dam. Catchment area: 2,78 km2
Relative relief: 281 m
Site selection Average annual rainfall: 338 mm
Land use: Grazing; includes about 15% badland areas
This allows them to reconstruct the his- Reservoir area: 5,23 ha
tory of catchment sediment yields from Dam age: 1910
the sediment stored in the dam. Sites Current status: 74% full and unbreached
for detailed investigation were chosen Average sediment yield (1939-97): 662 t/km2/y1*
partly based on contrasting land uses Total sediment volume: 109 303 m3
within the catchments. All have been *Takes into account estimated trap efficiencies
subject to sheep farming in historical
times; some are overgrazed, others less of the twentieth century, led to severe areas – hillslopes and badlands – and
so; some include formerly cultivated erosion and badland formation. How- rivers and dams. Erosion of gullies in
areas, some not. Some include badland ever, badlands have also appeared in the the nineteenth century under intense
erosion (an intricate network of small catchments where there is no documen- grazing pressure seems most likely, and
gullies), but in all cases the channel tary evidence of cultivation. there is some documentary evidence to
network has been deepened by gully support this.
erosion to form the characteristic linear One of the more surprising conclu-
dongas of the Karoo. sions of the work is that the extensive Documentary evidence has, in some
gully systems appear to have been cases, provided the age of dam con-
Results indicate that, in extremis, bad- fully developed before the first aerial struction or re-construction. Estimates of
lands probably developed as a result of photographs were taken in 1945 and dam trap efficiency have continued to
overgrazing between 1850 and 1950. have changed little since, contributing be a challenge, especially at Cranemere
Once established they persist and even little sediment to the dams. At present, where the dam wall has been raised on
reduced stocking rates seem to have they act as efficient conduits of water at least three occasions since it was first
little direct effect in terms of re-coloni- and sediment between the main source built in the 1840s.
sation of grasses and shrubs. In parts of
the area, dryland wheat production on
valley floors, especially in the first half

Badland at Compassberg in
the Sneeuberg.

The Water Wheel Jul/Aug 2009

34 Farm dams

Sediment storage in ones often constructed with assistance For example, the Compassberg Dam
small dams of government grants in the mid-1950s wall was breached as a result of 118 mm
and one as recently as 1980. With the of rain on 24 March 2000; the dam was
While investigating a small number of exception of the latter, all stone-built already full of sediment. Since 2000, a
dams in detail the authors have become dams are full of sediment and currently gully has begun to eat back into the
aware of the number and variety of store no surface water. Breaching of sediments in the reservoir above the
these small dams. Mapping an area of stone-built dams is rare, but has hap- breach. Breaching is also occurring as
just over 80 km2 revealed 95 small dams. pened on occasion, resulting in stored a result of animal burrowing activity in
Their age and size vary. Some are still sediment being released directly into dam walls. Many dams are therefore
functional in terms of holding water, but the channel. providing a ready source of sediment to
many are now full of sediment, may be downstream sections of rivers and reser-
breached, and are no longer capable of Those dams constructed of earth may voirs. If the authors’ conservative esti-
water retention. have a stone-built spillway or a spillway mates of 27 000 t/km2 are correct, there
on bedrock which resists erosion. How- is a large volume of sediment waiting to
Estimating amounts of sediment stored in ever, during exceptional rainfall events be mobilised in a relatively short time.
the small reservoirs is difficult, but details the dams fill, the spillway cannot cope
of two typical ones at Compassberg and and overtopping occurs at the lowest Big dams
Ganora are provided in the boxes on point of the earth wall. This may result in
pages 32 and 33. Sizes vary, but 50 000 erosion of the wall and the development Amounts of sediment stored in big
m3 of sediment per dam would seem a of an incised channel through the dam dams are given in Roosemboom et al.
reasonable estimate. A very conservative sediment. (1992). Generally their figures apply to
estimate of sediment storage in small a period before the mid-1980s. To esti-
reservoirs would be 2 million m3 in an Of 95 small farm dams mapped in the mate current storage some extrapola-
area of 100 km2. Assuming a sediment Sneeuberg, 46 (48%) are full of sediment tion has to be done.
density of around 1,35 t/m3 this would be and incapable of water storage. Of the
equivalent to a store of 27 000 t/km2. 95, 28 of the dams are breached and The nearest big dam to the study area
therefore potential sites of sediment in the Sneeuberg is the Van Ryne­
Dam breaching and loss. Having acted for up to 170 years as veldspas at Graaff Reinet. This dam on
sediment loss sites of water and sediment storage the the Sundays River, has a catchment
reservoirs are now beginning to act as area of 13 382 km2 and surveys show
Most dams in the area investigated are sources of sediment in that gullies are that about 47 million m3 of storage
earth-built, with only 18 stone-built eating back into the sediment fill. capacity remained in 1978 with about

Badland in the Sneeuberg;
willows in the distance
The Water WoinhcefcielullepJdyustl/hmAeuasgliltd2e0ao0mf9t.he

Farm dams 35

31 million m3 sediment partly filling of this trend would lead to an increase Dates of breaching will be established
the reservoir. This suggests an average in breaching and more rapid sediment through interviews with local land
deposition rate of 584,906 m3/y and evacuation from dams. There is also owners and by an analysis of farm
that by this year there will be around evidence of changing land use. The diaries, historical photographs, aerial
49,1 million m3 of sediment in the res- effect of changing land use is difficult to photographs and satellite imagery.
ervoir with a remaining storage capac- predict since decreasing stocking rates Of particular concern in this project
ity of about 29 million m3. At present should be influencing runoff and ero- is the number of breached dams and
rates of sediment transport this sug- sion, but changes in farm practice from the rate at which sediment is being
gests a dam life of about 50 years. dominantly sheep to wildlife/conserva- recycled from these sites to impact
tion are introducing an unknown factor. on downstream river systems and,
Potential impact of more importantly, the rate at which
small dams Future work sediment is likely to be reaching large
reservoirs.
It is not known how many small farm The above analysis, based on evidence
reservoirs there are in South Africa, from a small number of farm reservoirs Give the present estimates of the
however, the project team was surprised in the Sneeuberg area of the Karoo, has amount of sediment stored in farm
by their density, and their sediment pointed to a potential problem of signifi- reservoirs, they may indeed be a tick-
storage capacity, in a small area of the cant magnitude; these reservoirs, small ing bomb that may leave South Africa
Sneeuberg. in size but large in number, provide a facing a water crisis much sooner than
ready source of sediment for the larger anticipated.
A simple and conservative extrapolation storage reservoirs downstream.
from mapping in the Sneeuberg John Boardman is with the
suggests that in a catchment of The project team aims to assess the Environmental Change Institute
13,382 km2 (the area of the catchment of actual number of small reservoirs in a at the Oxford University Centre for
the Van Ryneveldspas Dam) there could major catchment that has a storage the Environment; Ian Foster is with
be as many as 10 000 of these reservoirs. reservoir at its downstream limit. In the Department of Molecular and
If each is the size of the Compassberg addition, there is a need to know the Applied Biosciences at the University
reservoir, containing about 50 000 m3 of approximate volumes of stored sedi- of Westminster; Kate Rowntree is
sediment, then there is a potential for ment within these farm reservoirs and with the Department of Geography
around 500 million m3 of sediment the rate at which the sediment is being at Rhodes University; Tim Mighall is
storage in the catchment. released from dams that are breached. with the Department of Geography at
The research will involve construction of the University of Aberdeen; and Tony
The sediment stored in Sneeuberg dams a GIS database in which information for Parsons is with the Department of
is largely fine-grained – it is not coarse individual reservoirs will be stored and Geography at the University of
bedload. The potential for movement as field surveys to calculate sediment stor- Sheffield.
suspended load means that the material age and rates of sediment release.
could reach large dams quite quickly.

The increasing availability of small farm Sources
reservoirs which are full of sediment,
and perhaps the continuing failure to Beinart, W. (2003) The Rise of Conservation in South Africa. Oxford University Press, Oxford.
repair breaches, suggests that the sedi- Boardman, J. and Foster, I.D.L. (2008). Badland and gully erosion in the Karoo, South Africa Journal of
ment yields in rivers will increase and Soil and Water Conservation 63 (4), 121-125.
large reservoirs will fill quicker than in Foster, I.D.L., Boardman, J., and Keay-Bright, J. (2007) Sediment tracing and environmental history for
the past. However, available potential two small catchments, Karoo uplands, South Africa. Geomorphology 90, 126-143.
sediment for downstream transporta- NWRS (2004) National Water Resource Strategy. Pretoria, Department of Water Affairs and Forestry
tion exceeds the remaining capacity of (DWAF). Available online at: http://www.dwaf.gov.za/Documents/Policies/NWRS/Default.htm.
a reservoir such as the Van Ryneveldspas Rooseboom, A., Verster, E., Zeitsman, H.L. and Lotriet, H.H. (1992) The development of a new sediment
by an order of magnitude. yield map for southern Africa. Water Research Commission, Pretoria.
Rooseboom, A. (1992) Sediment transport in rivers and reservoirs: a South African perspective. Report
Such crude calculations also ignore to the Water Research Commission of South Africa, Sigma Beta Consulting Engineers, Stellenbosch.
changing climate – in the Sneeuberg
at least there is good evidence for
increased intensity of rainfall events
over the last 50 years, a continuation

The Water Wheel Jul/Aug 2009