Seismo-tectonic modelling used for seismic hazard and risk assessments requires a complete comprehension of the geology, tectonics, palaeo-seismology, regional geophysical anomalies, historical and instrumental seismicity, and other neo-tectonic phenomena like relative plate motions and current tectonic plate stresses. This paper shows progress made in compiling available geospatial data relevant for this kind of modelling at the best resolution available and elaborates their importance in understanding the seismic load for the region.
Development in the KwaZulu-Natal (KZN) coastal regions is fast paced. The area is attracting multi-billion rand investments from both national and international investors for major infrastructure development projects. In Durban central, there is Transnet’s proposed port infrastructure upgrade and eThekwini’s upgrading of the Point harbour area; in La Mercy there is upgrading of infrastructure at the Dube Trade port in the vicinity of King Shaka International Airport and the infrastructure of the airport itself, and in Cornubia there is an extensive new housing and infrastructure development project that will extend from the already developed Gateway precinct to as far as the new airport. These civil projects make the region probably one of the fastest growing areas currently in South Africa.
Fig. 1: Isoseismal maps of significant earthquakes that have occurred within the province .
With infrastructure development, supporting power and water facilities and population density increases, comes an increase in vulnerability to rare natural disasters like earthquakes – which if understood properly could have their disastrous effects minimised. Recent earthquakes in Japan, New Zealand and Haiti have shown the need for proper infrastructure designto avoid the physical and economic diamage they create.
New infrastructure should incorporate seismic loads in their design, hence seismic hazard and risk assessments should be performed on a local scale. These assessments are critical for the construction of state of the art critical facilities like hazardous and high rise structures. A seismo-tectonic model is a critical building block for these evaluations which comprises, as a start, a multidisciplinary geo-database for understanding the root cause of earthquakes and classifying those regions that have similar properties for earthquake occurrence.
Fig. 2: Historical and instrumental seismic events that occurred within the province.
KZN coastal areas have experienced several potentially damaging earthquakes historically, albeit these earthquakes occurred at a time when it was less populated and built-up – hence there is very little memory or record of earthquake damage. Fig. 1 shows the distribution of isoseismal maps (curves delineating areas with different seismic intensities from each other), in Modified Mercalli (MM) scale of Richter , where potentially damaging intensities to infrastructure start from MM Intensity IV upwards. Historical reports of damage to the KZN coastal areas have been reported since 1932. These earthquakes are reported to originate from different sources, even from neighbouring countries like Mozambique. Note that the damage reported was significantly less compared to that if there were to be a repeat of the same earthquake today, mainly due to the rapid changes in infrastructure and development.
While seismic hazard maps have been calculated for South Africa [3 – 5], the methodology used for the calculations of these maps were based on “seismo-tectonic zone-free” methods (mainly because the seismo-tectonic model for the region was poorly understood). Regulations for seismic design of critical structures such as nuclear power plants, bridges, tunnels, LNG terminals, requires that seismo-tectonic models in some form be incorporated in the seismic hazard assessments (SHA) (see IAEA Seismic Safety Series; US Nuclear Regulatory Commission Seismic Regulatory Guides; Europeans building code Eurocode; National Nuclear Regulator’s guidelines for seismic design).
Fig. 3: Lithological units of the KZN province (from the 1:1 000 000 Regional Geology Map of the Council for Geoscience). The seismic event locations are superimposed on the lithological units.
While there are several agreed methodologies for deriving the seismo-tectonic model [6 – 12] and there are several methods for SHA [13 – 20], in all approaches, similar kinds of “geo-data” are required – for which its completeness, availability and resolution around the world differ. This work aims to put together the available data from literature and relevant organisations responsible for the data collection. Its status in terms of completeness, resolution and value to the seismo-tectonic modelling will be assessed in follow-up work. This work forms the first step for the seismo-tectonic modelling process.
The first publication for South Africa that built up a comprehensive multi-disciplinary geodatabase of this nature was that of Andreoli et al  and thereafter Singh et al . Several gaps were identified in the database if it were to be used for SHA for critical structures. Singh et al  proceeded to build a regional model despite the incompleteness in the datasets. Further Madi and Zhao  assimilated several other important datasets in order to investigate the potential for groundwater on a regional scale in South Africa. Other important work is that of Singh , with assimilated geo-data for use in the development of a landslide susceptibility map for the province.
Fig. 4: Mapped faults in the KZN province (from the 1:1 000 000 Regional Geology Map of the Council for Geoscience). The seismic event location is superimposed on the faults.
Wong et al  summarises the essential parameters required for a seismic source model that can be used in SHA. Two types of earthquake sources are required – fault sources and areal sources. Fault sources are typically modelled as three dimensional fault surfaces and details of their behaviour are incorporated into the source characterisation, while areal sources are regions where earthquakes are assumed to occur randomly with no clear association with any of the faults that might be included in the seismic source model. For fault sources to be included in the analysis, basic geometric source parameters that are required include the fault location, dip, and thickness of the seismo-genic zone. For the recurrence parameters one needs to include the recurrence model, recurrence rate (which can be obtained from the slip rate or average recurrence interval for the maximum event), slope of the recurrence curve (b-value), and maximum magnitude.
While the inputs to obtain fault and areal sources are complex, the risk implication varies considerably, considering the investment required for increasing seismic safety for design. Therefore it is imperative in the modelling to move from a more qualitative to a more quantitative methodology. There is a need for one to be able to propagate errors, where expert judgement is supported by data and is verifiable. All datasets should have room for updating at a later stage. The seismo-tectonic modelling analyst should be accountable and his hypothesis should be traceable or data-supported.
Therefore the style of this review is designed to start with raw data and to track the interpretative outputs in order to understand completeness and uncertainties. Ideally one should also be able to document multiple interpretations. For modeling that requires data from a multi-disciplinary field this is often impossible, unless the data has been acquired strictly for this modeling purpose, and all processes are documented.
In this study, the following datasets are presented: the seismic history of the region, the regional geology, and the regional geophysics. Several other datasets are still in the process of being digitised.
The Council for Geoscience provided both a historical (pre-1970) and instrumental catalogue of seismic events for the KZN province (see Fig. 2). Additional historical events have been found in literature but are presently being validated. A project is also underway to validate all seismic events in the CGS catalogue in terms of location and magnitude accuracy. Presently the seismic stations are sparsely distributed regionally, making the threshold of earthquake sensitivity equivalent to all magnitudes greater than and equal to 3.
There are a total of 177 earthquakes on record with local magnitudes (ML) within range of 2 – 4 from 1970 to 2012. In the historical catalogue there are 62 earthquakes on record with local magnitudes (ML) within range of 1,5 – 6,3 from 1906 to 1969.
The largest earthquake occurred in 1932 of ML 6,3 in the St Lucia area. Its effects are well documented in Krige and Venter .
Fig. 5: Regional aero-magnetics of the KZN province together with the seismic event locations.
This earthquake was located in the sea, offshore the Zululand coast. Shocks were reported in Port Shepstone, Kokstad, Koster, and Johannesburg (some 500 km away). The nearest point on land to the epicentre was Cape St. Lucia, where Modified Mercalli Intensity (MMI) of IX was assigned on the evidence of sand boils and cracks in the surface, but the damage in this area was small, possibly because it was uninhabited. In the severely shaken areas, poor-quality houses (built of unburnt or half-burnt bricks or other low-quality materials) were severely damaged. In well-built houses, small cracks were occasionally seen but the structures did not suffer major damage. The phenomenon of site-effects was clearly displayed in the observations of the after effects of this event. Structures built on thick sand were undamaged, while those built on alluvial sands suffered severe damage. Changing rock types in the area also had a strong influence on the attenuation of the seismic wave. From evidence of its effects, Krige and Venter  argue that this earthquake was probably caused by slip along a fault in the sea striking in a SSW-NNE direction parallel to the coast.
There is no obvious pattern when looking at the spatial distribution of the earthquakes except two very broad patterns as indicated. Towards the north, clustering of events occur in the north-west from blasting in the coal fields (containing events from the instrumental period only). An almost linear band of about 100 km extent of random occurrences of both historical and instrumental earthquakes can be located trending in a NE-SW coast parallel direction.
Geology and faults
The lithological units outcropping in the province is shown in Fig. 3. Whitmore et al  provided a generalised summary of the rock units outcropping into the province. The rocks in the region are of Achaean (3800 – 2700 Ma) to Cenozoic (65 Ma – recent) in age. The older rock units outcrop towards the northern extent of the province, towards the Swaziland border. On the eastern coastline, the Cenozoic Zululand Group silt, sandstone and unconsolidated sediments can be found. Most of the central part of the province consists of the Paleozoic to Mesozoic (300 – 180Ma) Karoo Supergroup (KS) sediments. The younger Proterozoic rocks (2500 – 490 Ma) of the Natal Group Sandstone (NGS) and the Natal Metamorphic Province (NMP) can be found outcropping in a NE band between Durban and Pietermaritzburg. From Pietermaritzburg up to Lesotho in a westerly direction one would find rock outcrops of the KS.
Fig. 6: Regional gravity of the KZN province together with the seismic event locations.
On first observation the only correlation with the north-easterly trend of the seismic locations identified earlier, is the linear outcropping trend of the sediments of the NGS and the NMP. However, this NE seismic trend extends into the sediments of the KS. A more detailed classification scheme is required than that of Whitmore et al  to identify the different rock units and their relative spatial and age relationships. Hence higher resolution geology data has been collected and is being assimilated [KZN regional geology 1: 250 000 scale, Durban 1: 50 000 scale, St Lucia and Richards Bay, Verulam and Maputuland and offshore bathymetrics].
The incipient break-up of the Gondwana supercontinent 183 Ma formed fractures and planes of weaknesses that acted as conduits for lava and formation of dolerite dykes and sills. After the separation of Africa and Antarctica at about 140 Ma, marine sediments of the Cretaceous Zululand Group were deposited in the newly opened Indian Ocean.
Major faults mapped in the province are superimposed on the seismic catalogue as seen in Fig. 4. Hughes, Von Veh and Maud [29 – 31] described consistent trends in the fault directions in the province, essentially coast parallel NE-SW faults and coast perpendicular NW-SE faults. These distinct groups occur at the interface with the sediments of the KS and the Pongola Supergroup (NW-SE faults) and at the interface between sediments of the KS and the sediments of the NGS and the NMP. The E-W fault line is also a prominent feature and occurs within the KS.
Hughes  looked specifically at the Durban area and found relative displacement of faults of a few centimetres to tens of metres. King and Maud  described faults in detail that had displacements of 300 m in places. Normal faults are dominant in the Durban area and are consistent with extensional palaeostresses acting at the time of faulting.
Interestingly, there are several correlations with the seismicity and the mapped faults in the area. Until the seismic event locations uncertainties are assessed, accurate correlations cannot be made. Work in progress includes understanding and classifying the faults such as mechanisms, stress orientations, age of faulting, displacement amounts and evidence for reactivation during the Quaternary period.
Geophysical investigations (e.g. gravity, magnetics, and seismic anisotropy) provide a better understanding of the subsurface geology of the Earth. Regional aeromagnetic and gravimetric maps of the province are shown in Figs. 6 and 7. The seismicity data are superimposed on these maps.
Gravity data provide information about densities of rocks underground. Generally, gravity highs indicate the presence of relatively dense rocks, and magnetic anomalies are caused by rocks with abundant magnetite in them. Very high-intensity anomalies (more than 50 milligals or more than 200 gammas) typify major changes in rock type, usually (but not always) in basement rocks. Most sedimentary rocks (with the exception of banded ironstones) contain little magnetite, so generally we are dealing with igneous and metamorphic rocks .
From the geophysics data, the regional geological provinces can be determined . The southern extent of the Achaean KC is indicated on the figures and the lower province is called the Namaqua-Natal Mobile Belt (NNMB). This boundary also falls in the same vicinity of the EW trending fault mapped through the province. The sediments of the KS overlay these provinces in most cases. Note that in both these figures the patterns formed by the geophysical signatures correlate to a large extent with the geological provinces beneath the Karoo sediments.
The rocks in the NNMB provide locally high magnetic signatures. The regional gravity map also maps the high density rock types along the coastline. No other direct correlations can be made with the seismicity and the geophysics at this stage.
Higher resolution datasets are available and is being assimilated for further analysis.
Other databases and work in progress
Other work that is being assimilated include: structural domains for the province which provide a tectonic basis for seismo-tectonic modelling ; regional geomorphological provinces which provide a geodynamic basis
for seismo-tectonic modelling , GPS measurements to characterise regional and local plate ; naturally occurring springs like Lalani, Shushu and Natal Spa that occur usually at deforming ground areas [37, 38], and local evidence of neo-tectonic deformation/faulting.
The region has experienced a significant level of seismic activity historically. Visual correlations of the seismic catalogue locations show NE-SW coast parallel patterns of seismic occurrences. This pattern correlates with the geology of the NMP and the NGS outcropping in a similar coast parallel orientation. The mapped faults have two distinct orientations: NE-SW coast parallel patterns in the south-west, and NW-SE coast perpendicular orientations in the north. Several correlations can be made with the mapped faults and the seismic locations. From the regional geophysics datasets, the boundaries of the basement geological provinces of the Achaean KC and the NNMB can be located. Other important studies on the geomorphological provinces, regional plate motion, natural springs, seismic tomography and structural provinces are still being compiled.
Preliminary evidence has shown that the region has significant levels of seismicity and strong evidence of locally tectonic active regions. Hence ground motion parameters should be incorporated to reduce economic and population vulnerability due to earthquakes.
This work is funded by the University of KwaZulu-Natal and the National Research Foundation. The CGS provided the regional geology, regional geophysics, historical and instrumental seismic data. Ian Saunders of the Council for Geoscience provided technical assistance with the catalogue. Kerissa Naidoo, Maruska Ramdeyal and Naomi Kunju of UKZN provided technical assistance on the reference database.
The paper was presented at SASGI 2013 and is republished here with permission.
 C.F. Richter: Elementary Seismology, Freeman W.H., San Francisco, 768, 1958.
 M. Singh and E. Hattingh: “Collection of Isoseismal Maps for South Africa”, Natural Hazards, Vol.50, 403-408, 2009.
 L. M. Fernandez and A. du Plessis: Seismic Hazard Maps for Southern Africa, Council for Geoscience, 1992.
 A. Kijko, G. Graham, M Bejaichund, D.L. Roblin and M.B.C. Brandt: Probabilistic Seismic-Hazard Maps for South Africa. Council for Geoscience, Report no 2003-0053, 2003.
 D. Giardini and P. Basham: “The Global Seismic Hazard Assessment Program (GSHAP) “, Annali di Geofisica, Vol.36, 3-13, 1993.
 D.V. Gonzalez and B.O. Skipp: “Hazard mapping in risk evaluation for engineering structures”, Bull Int As Eng Geol, Vol. 21,118–121, 1980.
 P. Gasperini, F. Bernardini, G. Valensise and E. Boschi: “Defining seismogenic sources from historical earthquake felt reports”. Bull Seismol Soc Am, Vol. 89(1), 94–110, 1998.
 W.K. Mohanty, M.Y. Walling, “First order seismic microzonation of Haldia, Bengal Basin (India) using a GIS platform. ” Pure Appl Geophys, 165:1325–1350, 2008.
 C. Meletti, E. Patacca and P. Scandone: “Construction of a seismotectonic model: the case of Italy”, Pure Appl Geophys, Vol. 157, 11–35, 2000.
 E.C. Hicks, Bungum and C.D. Linholm: “Seismic activity, inferred crustal stresses and seismotectonics in the Rana region, Northern Norway”. Quat Sci Rev , Vol.19, 1423–1436, 2000.
 B. Le Goff, D. Fitzenz and C. Beauval: “Towards a bayesian seismotectonic zoning for use in Probabilistic Seismic Hazard Assessment (PSHA) AIP Conf. Proc. Volume 1305, pp. 242-249, 2011.
 M. Terrier, J.L. Bl`es, P. Godefroy P(†), M. Dominique, M. Bour and C. Martin: “Zonation of Metropolitan France for the application of earthquake-resistant building regulations to critical facilities Part 1: Seismotectonic zonation”, Journal of Seismology, 4, 215–230, 2000.
 C.A. Cornell, “Engineering seismic risk analysis”, Bull. Seism. Soc. Am., Vol. 58, 1583–1606, 1968.
 D. Veneziano, C.A. Cornell and T. O’Hara: “Historic method of seismic hazard analysis”, Elect. Power Res. Inst., Rep. NP-3438, Palo Alto, CA, 1984.
 B.K. Bender D.M. Perkins: “Treatment of Parameter Uncertainty and Variability for Single Seismic Hazard Map”, Earthquake Spectra, 9, 165–195, 1993
 R.K. McGuire: “Computations of seismic hazard”, Ann. Geofis. 36, 181–200, 1993.
 R.K. McGuire and G.R. Toro: “Site-specific seismic hazard analysis”, 14th World Conference on Earthquake Engineering October 12-17, Beijing, China, 2008.
 A. Kijko and G. Graham G: ‘‘Parametric-Historic’’ Procedure for Probabilistic Seismic Hazard Analysis. Part I: Assessment of Maximum Regional Magnitude mmax, Pure Appl. Geophys. 152, 413–442, 1998.
 A. Kijko and M.A. Sellevoll: “Estimation of Earthquake Hazard Parameters from Incomplete Data Files. Part I, Utilization of Extreme and Complete Catalogues with Different Threshold Magnitudes”, Bull. Seismol. Soc. Am. 79, 645–654, 1989.
 A. Kijko and M.A. Sellevoll: “Estimation of earthquake hazard parameters from incomplete data files. Part II. Incorporation of magnitude heterogeneity. Bull. Seism. Soc. Am.,82, 120‑134, 1992.
 M.A.G Andreoli, M. Doucoure, J. van Bever Donker, D. Brandt, and N.J.B. Andersen: “Neotectonics of southern Africa: A review. ” Africa Geoscience Review, 3(1), 1–16, 1996.
 M. Singh, A. Kijko and R. Durrheim: “Seismotectonic Models for South Africa. Synthesis of Geoscientific Information, Problems and Way Forward”, Seismological Research Letters, Vol. 80, 65-33, 2009.
 M. Singh, A Kijko, and R. Durrheim: “First-order regional seismo-tectonic models for South Africa”, Natural Hazards, 59,1, 383-400, 2011.
 K. Madi and B. Zhao: “Neotectonic belts, remote sensing and groundwater potentials in the Eastern Cape Province, South Africa Vol. 5(6), pp. 332-350, June 2013
 R. Singh: “Landslide Classification, Characterisation and Susceptibily Modelling in KwaZulu Natal”, University of Witwatersrans, Johannesburg, MSc, 2009.
 I.G. Wong, P.A. Thomas and N. Abrahamson, N. “The PEER-Lifelines validation of software used in probabilistic seismic hazard analysis”, in 2004 Geotechnical Engineering for Transportation Projects, Proceedings, M. Yegian and E. Kavazanjian (Editor). American Society of Civil Engineers, Geotechnical Special Publication No. 126, 1, 807-815, 2004.
 L.J. Krige and F.A. Venter, “The Zululand earthquake of the 31st December 1932″. Transactions of the Geological Society of South Africa 36, 101–112, 1993.
 G. Whitmore, R Uken, D. Meth: KwaZulu-Natal: 3500 million years of Geological, History, University of Natal, 1999.
 S. Hughes: “The structural geology of central coastal KwaZulu-Natal: Analysis of faulting between Fields Hill and the Umgeni River mouth”, School of Geological Sciences, University of KwaZulu-Natal, 2008.
 M. von Veh: 1994 The structural geological analysis of lineaments and fault and dyke traces in the KwaZulu-Natal Province, School of Geological Sciences, University of KwaZulu-Natal, 1994.
 R.R. Maud: “A preliminary review of the structure of coastal Natal. ” Trans. Geol. Soc. S. Afr., Vol 64, 247-256, 1961.
 L.C. King, R.M. Maud, R. M.: The geology of Durban and environs. Geological survey, Government printer, Pretoria, 49 pp, 1964.
 D. Gibson: Primer on Gravity and Magnetics; http://www.gravmag. com/gmprimr.html, 2007.
 T. Nguuri, J. Gore, D. E. James, S. J. Webb, C. Wright, T. G. Zengeni, O. Gwavava, J. A. Snoke, and Kaapvaal Seismic Group: “Crustal structure beneath southern Africa and its implications for the formation and evolution of the Kaapvaal and Zimbabwe cratons. “Geophysical Research Letters 28 (13), 2,501–2,504, 2001.
 T. C. Partridge, E. Dollar, J. Moolman and L.H. Dollar: “The geomorphic provinces ofSouth Africa, Lesotho and Swaziland: A physiographic subdivision for earth and environmental scientists”, Transactions of the Royal Society of South Africa, 65: 1, 1 — 47, 2010.
 R. Malservisi, U. Hugentobler, M. Hackl and R. Wonnacott: “Constraining Nubia plate rigidity from Continuous GPS observations”, IUGG 2011.
 T.W. Gevers and D Phil1941, “Carbon Dioxide Springs and Exhalation in Northern Pondoland and Alfred County, Natal”, Transactions of the Geological Society of South Africa Publication, 1941.
 L.E. Kent: “The Thermal Springs of Southern-Eastern Transvaal and Northern Natal”, Geological Surveying, South Africa, p.51-67, 1981.
Contact Mayshree Singh, University of KwaZulu-Natal, Tel 031 260-1153, firstname.lastname@example.org
The post Geo-database compilation for seismo-tectonic investigations appeared first on EE Publishers.