Mining and Conservation of the Mpumalanga Lakes District
Professor Terence McCarthy 1 , Professor Bruce Cairncross 2 , Professor Jan-Marten Huizenga 2 and Allan Batchelor 3
1: School of Geosciences , University of the Witwatersrand, Johannesburg ; 2: Department of Geology, University of Johannesburg ; 3: Wetland Consulting Services (Pty) Ltd.
Backfilled and rehabilitated opencast coal mines in the eastern Highveld fill with water in between 5 and 10 years. The water becomes acidic and enriched in sulphates due to oxidation of iron sulphide in the waste rock. Once filled, the polluted water from the void begins to decant and discharges on surface, causing pollution of surface water resources. Pans are particularly vulnerable to this form of pollution, as there is no possibility of removal of the toxic materials by natural flushing. Proposed opencast mining within the catchments of pans in the Mpumalanga Lake District thus presents a very severe threat to this unique and pristine wetland ecosystem.
Pans are fairly widespread, although widely scattered, in a broad belt across the interior of southern Africa extending from the Northern Cape Province across the Northwest Province into Mpumalanga . The majority occur in the drier, western portions of South Africa , particularly in the Kenhardt – Brandvlei area. Most pans are ephemeral, and are characterized by saline deposits on their floors, and many have been mined for salt in the past. The eastern extremity of the pan belt is marked by an unusually dense cluster, centred on Lake Chrissie , the largest of the pans. A smaller cluster is developed to the northeast in the Arnot area (fig. 1).
Figure 1 . A LANDSAT image of the Lake Chrissie pan field.
The Lake Chrissie pan field differs radically from other pan fields in the country in several respects. Pans in the western fields tend to be large, dry, floodplain-like features, elongated along river courses, whereas those in the Lake Chrissie area are isolated, usually oval in shape, and are perennially flooded. The density of pans is also extremely high. Moreover, whilst drainage within the pan field is essentially closed, several major river systems arise around the fringes of the field, namely the Vaal River , the Komati River (via the Boesmanspruit), the uMpuluzi River and the Usutu River (fig. 1). Thus, the pan field represents a local plateau of elevated ground, amongst the highest in the Highveld region. These various features combine to make the Mpumalanga Lake District (MLD), as the pan field is known, a unique geomorphic entity in the South African landscape.
Geomorphological uniqueness frequently generates a corresponding biological uniqueness, and is often associated with endemism. In the case of the Lake Chrissie area, only the avian fauna have been explored in any detail (see appendix), and other aspects of the biodiversity of this unique province await exploration.
In this summary report, the origins of this unique province are explored, and the possible negative impacts of human activities in the area will be examined.
Geology of the MLD
The geomorphological characteristics of a region often arise from geological factors. In the MLD, this appears not to be the case. Most of the eastern Highveld is underlain by rocks of the Ecca Group of the Karoo Supergroup, locally intruded by dolerite dykes and sills. In common with much of the surrounding area, the immediate bedrock in the MLD is formed by the Vryheid Formation of the Ecca Group. These rocks consist mainly of sandstones, with lesser siltstone and shale. Several coal seams are developed, of which the Number 2 (or E) and 4 (or C) are economically the most important. The rock strata are essentially horizontal, possibly with a slight tilt towards the southwest. The sandstones are massive, with widely spaced, narrow joints. There are no known differences between the strata beneath the pan field and those that underlie the surrounding areas, and hence a geological origin for the unique MLD has to be ruled out.
Geomorphology of the MLD
Southern Africa is geomorphologically unusual in that very limited erosion has occurred since the Cretaceous Period which ended 65 million years ago. In fact, in several parts of the country, remnants of the Cretaceous land surface still occur. This ancient land surface, known as the African surface, formed in the aftermath of the break-up of the supercontinent Gondwana, when southern Africa became surrounded by warm seas. The sub-continent was levelled by an extensive river network that included rivers ancestral to the Vaal, Orange and Limpopo . Soils formed at that time were deep and highly leached, akin to tropical soils of today. Since that time, the sub-continent has experienced two major periods of uplift, one about 20 million years ago and the other about 5 million years ago. In both events, uplift was more pronounced in the east than the west, and the overall effect of these was to raise and tilt the subcontinent, forming a vast internal plateau sloping gently to the west. As a result of this uplift, much of southern Africa today lies at an elevation above 1000 m above sea level, in stark contrast to most other regions of the world with similar geology that typically lie at elevations below 400 m above sea level (e.g. northern Canada and western Australia). This global topographic anomaly in southern Africa has recently caught the attention of scientists, and has been named the African Super Swell.
Periods of uplift and tilting of the sub-continent had the effect of rejuvenating rivers in the interior, promoting down-cutting and headward erosion (back-cutting), and forming new, younger erosion surfaces. Two such periods are identified, known as the post African I and II surfaces, corresponding to the 20 million and 5 million year uplifts respectively.
Detailed studies of the topography of southern Africa by Professor Tim Partridge and Dr Rodney Maud have shown that the largest contiguous remnant of the ancient African land surface is situated in the Standerton – Ermelo – Belfast area of western Mpumalanga . Even here though, the effects of later erosion caused by uplift and tilting have been felt. Rivers from the surrounding younger terrains have cut back and removed the veneer of ancient soil. However, within this greater region we can still identify remnants of the old land surfaces, foremost amongst which is the MLD. This last remnant, probably of the post-African I surface, is being encroached by head-cutting rivers from all sides: in the east by headwaters of the uMpuluzi, in the south by the Usutu, in the north by the Komati and in the west by the Vaal . The actual age of this remnant land surface is difficult to establish, but it is probably between 10 and 20 million years old.
The pan field itself thus provides a glimpse of one of the most ancient land surfaces in southern Africa , and also provides clues as to the origins of the pans themselves. The pans are generally oval in shape, and typically have bedrock exposed on their western margins. The eastern margins are characterized by thick, unconsolidated sand deposits. A detailed study carried out by the late Professor John Wellington showed that many of the pans, particularly the larger ones, can be linked together and form an apparently eastward flowing drainage network (fig. 2). He hypothesised that this linked system once formed a tributary network to the ancestral uMpuluzi. Head-cutting by the Vaal River captured the western headwaters of this river system and diverted it to the Vaal , thus depriving the easterly-flowing rivers of their water and rendering the network moribund. Sand from the river beds and surrounding slopes was blown by the prevailing westerly wind into dunes which divided the drainage network into a series of isolated segments, which became the pans of today.
Figure 2. Map showing the major pans in the MLD and their possible former linkages. The pan field is surrounded by headwaters of the Vaal, Komati, uMpuluzi and Usutu Rivers .
This hypothesis for the origin of the pans is broadly compatible with their present distribution, but it predates much of the modern understanding of continental uplift and tilting, of the ages of land surfaces in southern Africa , and of climate change over the region during the past million years. For this reason, the MLD is currently the focus of re-evaluation. Perhaps other factors in addition to headward erosion by the Vaal contributed to the demise of the drainage network and the formation of the pans. Foremost candidates are back-tilting of the river courses, thus decreasing their gradients, and the periodic occurrence of exceptionally dry periods that coincided with the ice ages in the northern hemisphere. Such dry periods would have promoted the formation of wind-blown sand deposits along the courses of the rivers, because of reduced vegetation cover at these times. Another possible factor is regional warping of the sub-continent, as the pan belt broadly coincides with a continental-scale drainage divide that separates northerly from southerly flowing tributaries. At this stage, the contributions of these other factors can only be speculated on, but what is clearly evident is that the MLD offers a window into the character and history of one of the most ancient land surfaces in southern Africa .
Hydrological functioning of the pans
In contrast to the pans in the west, those of the MLD are in the main perennial. Moreover, the water in the pans is generally fresh. One of the most important factors contributing to the perennial nature of the pans is the favourable water balance (i.e. the difference between rainfall and evaporation) of the region. In the MLD, rainfall is 800 mm per annum, and evaporation is 1600 mm per annum. In contrast, in the Kenhard district, rainfall is less than 100 mm per annum, and evaporation is 2700 mm per annum. Little wonder that the pans in the west are mostly dry.
The pans in the MLD receive water in a number of different ways. Rain falling directly onto a pan surface adds water. Each pan is surrounded by its own watershed, and some of the rainwater falling within this catchment forms surface run-off and flows directly into the pan, whilst the remainder percolates into the ground to become groundwater. This groundwater regime forms an important part of pan hydrology. The sandstone around and below the pans consists of massive, low porosity rock, with few, widely spaced joints (fig. 3). They have very little storage capacity for water, and transmit water very slowly. Nevertheless, some groundwater migrates through the joint system into the pans, but this source is probably the least important. Most of the groundwater collects in the weathered rock and soil (the regolith ) overlying the bedrock. Typically the soil profiles in the region contain a horizon rich in iron oxide ( ferricrete ) that provides a local barrier to downward flow of groundwater. The thickness of the regolith layer is greatest on the higher ground surrounding the pans, and becomes thinner on the slopes towards the pan, where it may disappear completely, giving way to rock outcrop (fig. 3). Spaces between the regolith grains become filled with groundwater, which saturates the material below a certain depth, known as the water table . The groundwater gradually migrates down slope under the influence of gravity. Because of thinning of the regolith down slope, the zone of saturation eventually impinges on the land surface, and here the groundwater seeps out (fig. 3). This process gives rise to the hill-slope wetlands that surround the pans. Seeped water flows slowly across these wetlands into the pan.
Figure 3. Diagram illustrating the geology and hydrology of a typical pan in the MLD.
Layers of ferricrete in the soil profiles may give rise locally to perched groundwater, in which case isolated wetland patches may form high up on the hill slopes around the pans, as groundwater is forced to surface and re-enters the ground further down the slope.
It is likely that groundwater also occurs in the sands on the eastern margins of the pans, although the groundwater hydrology of these pan margins has not been studied. Seepage of ground water and thus wetland formation also occurs along these margins, but the up-slope extent of these wetlands is evidently less than on the slopes where bedrock is close to surface, a consequence of the greater ease with which water percolates through this material.
During very wet periods, water depth in the pans increases, and some may overflow into neighbouring pans along the old drainage lines identified by Wellington (see above). There may also be some seepage from one pan to another through the sand dunes that obstruct these former river valleys, although this has not been conclusively established. There is evidence from elsewhere (Prairie potholes) that some pans leak, which might account for the differences in salinity recorded in the MLD pans and elsewhere on the Mpumulanga Highveld. Notwithstanding this, the most important manner in which water leaves the pans is by evaporation off the pan surface and by transpiration from plants in the vicinity, because each pan is essentially a closed system as far as water is concerned. Indeed, in extreme droughts the pans may dry up completely.
Water quality is quite variable amongst different pans, and salinity ranges from about 200 to about 8000 parts per million total dissolved solids. To put this in perspective, the maximum allowable limit for water for human consumption set by the SABS is 1900 parts per million. In a survey of 10 pans carried out in 1998, seven were within this limit. It is likely that salinity in any individual pan is seasonally variable, but in addition there are real differences between pans. The reasons for these variations have not been investigated. The total salinity of the pans is remarkably low considering the fact that most of the water entering the pans is lost by evaporation. This normally leads to a progressive concentration of salts and the formation of saltpans, but this has not happened in the MLD.
Rainwater and surface runoff have very low dissolved solid content, and dissolved solids in natural water are normally acquired by slow reaction with the rocks and soils through which the water flows as groundwater. In the MLD the regolith is very sandy and highly leached, and contains relatively few minerals with which water can interact. As a consequence, the groundwater also contains little in the way of dissolved solids (typically about 130 parts per million), and groundwater quality is very good. Groundwater emerging through the hill slope wetlands and entering the pans therefore adds only small quantities of dissolved solids to the pans, but it is likely that most of the dissolved solids in the pans arise from this source. This is probably the reason for the low salinity of the pans in the region. Groundwater flowing through the bedrock is restricted by the massive nature of the rock and the lack of suitable fractures, and probably does not contribute much to the dissolved solid load in the pans. Nevertheless, these rocks have the potential to greatly increase the dissolved solid load, as is discussed below.
Human disturbance in the MLD
The MLD has been the site of human occupation for millennia. In the pre-historic era, the land was most likely used by hunter-gatherer communities, and indeed was home to San peoples until relatively recent times. Rock art abounds in the area, attesting to the long human occupation. In these times, herds of game would have been the most significant users of the land. Little changed with the advent of modern farming in the region, as most of the land was, and still is, used for grazing. In effect, indigenous game species were replaced by cattle and sheep, which became the main users of the land. The extent of cultivation in the pan field remains relatively small, and environmental impacts such as phosphate and nitrate pollution from this source are probably very small. For these reasons, the MLD is in an essentially pristine condition, except for the change in the large herbivore population.
The MLD is currently the focus of interest from coal mining companies. Their intention is to undertake opencast mining within the MLD, which will fall within the local catchments of many of the pans. The problems arising from opencast mining in the region have been comprehensively studied by Professor Frank Hodgson and his colleagues on behalf of the Water Research Commission. His study concentrated on the Olifants catchment, but many of the results can be more generally applied.
The main problem arises from the fact that the rocks associated with coal contain the mineral pyrite, an iron sulphide (FeS 2 ). When this mineral is exposed to water containing dissolved oxygen, it undergoes oxidation in two stages:
2FeS 2 + 7O 2 + 2H 2 O => 2FeSO 4 + 2H 2 SO 4 reaction 1
4FeSO 4 + O 2 + 4H 2 0 => 2Fe 2 O 3 + 4H 2 SO 4 reaction 2
Both of these chemical reactions produce sulphuric acid, and the second also produces an orange-red precipitate of iron oxide (rust). The phenomenon is well known in mining areas throughout the world, and has been termed Acid Mine Drainage. In South Africa , the Witwatersrand goldfields are severely affected, because of the presence of pyrite in the mining residues.
Under natural conditions, the pyrite is trapped within the rock mass, and water and air are excluded, so pyrite is stable. There is however, a natural, but very slow, production of sulphuric acid as the rock weathers. It is slow, because the rock is dense and does not allow the passage of water very easily. The chemical reactions are sufficiently slow that the acid produced is diluted and washed away by groundwater. However, when mining takes place, the rock mass is broken up, exposing large amounts of pyrite. After the coal has been removed, the pit is back-filled with the broken, pyrite-bearing waste rock and the surface is rehabilitated (fig. 4). Water flows readily into and through the mass of broken rock, and oxidation of pyrite occurs rapidly. Large amounts of suphuric acid are quickly produced and the water filling the voids in the backfill becomes acidic. Oxygen dissolved in the percolating water is quickly consumed, and only the first chemical reaction (1 above) occurs, producing groundwater enriched in iron sulphate and sulphuric acid. If this water emerges on surface where oxygen is plentiful, the second reaction (2 above) occurs, producing more acid and an orange-red precipitate of iron oxide. This results in the red coloured water typical of sites experiencing acid mine drainage (fig. 5). According to Hodgson, the acidic water in the mine void also leaches heavy metals from the broken rock, including manganese, copper, zinc and in rare instances nickel, cobalt and cadmium. It also becomes enriched in aluminium.
Figure 4. Diagram illustrating the hydrological functioning of a typical pan after mining and rehabilitation have been completed.
Figure 5. Red acidified water in a collapsed coal mine in the Witbank coalfield.
Some rocks have a natural ability to neutralize sulphuric acid because they contain the minerals calcite (CaCO 3 ) or dolomite (CaMg(CO 3 ) 2 ). The chemical reactions leading to neutralization are:
CaCO 3 + H 2 SO 4 => CaSO 4 + H 2 O + CO 2
CaMg(CO 3 ) 2 + 2H 2 SO 4 => CaSO 4 + MgSO 4 + 2H 2 O + 2CO 2
Although these reactions lead to acid neutralization, they nevertheless still result in high concentrations of calcium and magnesium sulphate in the water.
Of the ten mines Hodgson studied, in only one case was there sufficient calcite and dolomite in the broken rock to neutralize the acid produced, and seven of the mines had what he termed ‘severe acid producing tendencies'. This is a consequence of the high pyrite content and the low amount of calcite and dolomite minerals in the rocks associated with our coal. Consequently, the water in the void becomes acidic, and enriched in dissolved heavy metals.
Hodgson found that the acidic water from the mine void can flow into the surrounding, undisturbed sandstone, displacing fresh groundwater (fig. 4). However, movement of water through this rock mass is very slow, and cannot keep pace with the rate of infiltration of rainwater into the backfill. Thus, the void becomes filled with water to a depth equal to the lowest point on the rim of the former open pit. Hodgon's study showed that backfilled opencast coal mines to the west of the MLD fill with water in between 5 and 10 years after closure. Once the void is thus filled, further inflow forces void water to decant and this emerges on the land surface (fig. 4). The decant water is not only acidic, but typically contains between 2000 and 3000 parts per million dissolved sulphate and a variety of dissolved heavy metals, and is completely unpotable (the SABS maximum allowable limit for sulphate in drinking water is 200 parts per million). The emergence on surface of water from the mine void also results in further acid production (reaction 2 above), and the precipitation of red iron oxide. The pH of this water is typically around 2.5. Professor Fred Bell and his associates have found that most of the acid water in streams in the Witbank area is in fact emanating from backfilled opencast coal mines. Evaporation of the water around the seep produces white crusts of calcium sulphate and numerous other sulphate minerals, many of which are acid-generating in their own right. The soils around and downstream of such seeps become acidic and saline, and most of the plants growing there die.
It may be possible to slow down the filling of the mine void by the use of clay layers in the backfill but it will never be possible to seal them in perpetuity. The onset of surface leakage of water from the mine void may thus be delayed, but its appearance on surface will inevitably occur. Production of acid and sulphate salts in the water filling the void will continue until all of the pyrite has oxidized, a process that may take centuries to complete.
The area Hodgson studied lies in the Olifants River catchment, an area of through-going drainage. In such regions, contaminated water entering rivers and streams tends to become diluted by water derived from non-mining areas and by periodic heavy storms that produce large amounts of runoff that flushes away and dilutes contaminated water. This will not occur in the MLD. Here, the pans have no outflow (except for a few, and then only under extremely wet conditions) so all dissolved compounds entering a pan become trapped there, and over time will accumulate. The accumulation of sulphate salts in the pans will eventually destroy all aquatic life, converting the pans into virtually sterile, toxic pools.
In summary, there is a very real danger that opencast mining within the catchments of pans in the MLD will destroy the natural hydrological functioning of the pan systems, and moreover will result in the pollution of surrounding groundwater with sulphuric acid and various sulphate salts. The groundwater will become completely unpotable. In time, this polluted groundwater will leak into the pans and they too will become contaminated. Since the pans have no outlet, and hence cannot be flushed during heavy rains, salts will accumulate there, irreversibly polluting them. Although the process may be slowed down by careful backfilling, including the importation and use of suitable clays, this will only defer the generation of acid and will not stop it. In the long term, destruction of the pans will be inevitable if mining takes place.
The MLD is a totally unique region in southern Africa , perhaps even globally. It represents the last fragment of the one of the most ancient land surfaces in southern Africa . As a pan field it is also unique, and there is no other region with such density of perennial pans in southern Africa . Moreover, the pans are in an essentially pristine state. This physiographical uniqueness may well be reflected in biological uniqueness, and the chances of finding endemic species in the area are high. Unfortunately, the biota of the region have not been studied in great detail.
There is little doubt that opencast coal mining will disrupt the hydrology of the pans, and irreversibly pollute the water in the pans. If mining is permitted, the pans will, in time, evolve to become toxic pools, devoid of all but bacterial life forms. The nation will have lost one of its true gems.
There is an urgent need to ring-fence the MLD, and to exclude all exploration and mining from the area. Moreover, farming practices need to be reviewed, and any potentially polluting activities stopped. The MLD is a unique geomorphic province and biotic habitat, and must be conserved for future generations.
BIRDLIFE AT LAKE CHRISSIE
The birds frequenting the pans around Lake Chrissie were studied on a monthly basis over a period of eight years from 1996 to 2004. This was done by selecting samples of 15 pans and 3 lakes and studying these as being representative of the 300 pans in the area. The pans and lakes range in size from 1043ha in the case of Lake Chrissie to 2 or 3ha in the case of the smallest grass pans.
In dry years (<450mmpa) 80% of the pans dry up whereas in normal to wet years (>750mmpa) most of the pans retain their water throughout the year. In the flood years of 2000 and 2002 the Koolbank road (running along the eastern border of Lake Chrissie ) was flooded for 18 months causing a major inconvenience to some 12 farmers. The birds are counted by circumnavigating the pans in a 4 x 4 vehicle. However many of the study pans were inaccessible for months during the floods.
Two of the fifteen pans (in the study area) are reed covered which apart from providing cover for skulking birds such as night herons, crakes, rails and swamp hens, the reeds also provide roosts and nesting sites for cormorants, egrets, glossy ibises, spoonbills and swallows. The pans and lakes that are redlined are either open water, sedge lined largely grass covered or saline. To date 78 different species have been recorded in the study area. The monthly survey entailed counting over a million birds over the 8 year period.
In the dry years the predominating birds are flamingoes, Cape Teal , Avocets and migrant waders. In contrast in the wet years – the predominant birds are grebes, Egyptian and Spurwing Geese, Yellow-billed Duck, Pochard, Coot, Glossy Ibis and Stilts. The two geese are the most numerous birds on Lake Chrissie . It should be borne in mind they feed and breed on land and only moult and preen on the pan edges.
Lake Chrissie's panveld is unique in terms of occurring in a high rainfall area (743 mmpa average over 55 years) and in the highlands of the northern most extensions of the eastern Drakensburg giving run to the Usutu, Komati and Vaal rivers while the lakes have inlet and outlets and are considered to reflect a relict drainage system, the pans are generally grasslined, circular in shape and have no obvious inlets or outlets.
The most sought after birds are the Crowned Crane, Blue Korhaan and Chestnut-banded Plover – the latter is normally only found on the dry pans in the western Free State and northern Cape. The latest new records for the area are Yellow-billed Stork, Western Marsh Harrier, Asiatic Golden Plover, Lesser Sandplover, Redshank and Curlew.
According to Marius Wheeler of the Avian Demography Unit at the University of Cape Town, the MLD qualifies as a Global Important Bird Area.
Dr J de Villiers