Is present seismic map correct?

IN THE aftermath of the Gujarat earthquake, there is great interest even among the lay people to know about earthquakes. This is understandable given the sequence of aftershocks in the Kutch region and the occurrence of tremors in Karnataka and Kerala. There is a statement that earthquakes don't kill people but buildings kill people. In a sense, this is true but earthquake is the primary cause for buildings to collapse.

After a major earthquake it is difficult to be disppassionate about the reasons for collapse of civil engineering structures. It is natural for an affected house owner to feel that his house should have collapsed due to poor quality of construction, more so when neighbouring buildings continue to stand. However this damage variation has to be viewed in the right perspective. We hear often statements by influential persons including administrators, engineers and academics attributing the damage to poor design and faulty construction.

In this context a widespread opinion exists that in Bhuj and Ahmedabad the engineering design codes (also called building codes) had been flouted in the case of high-rise buildings leading to heavy damage. As a corollary to this belief the code on earthquake resistant design of structures IS-1893 of Bureau of Indian Standards (BIS) has been described as the final word in guaranteeing safety against building collapses during eathquakes. In fact several architects and engineers in Bangalore have gone to the press with the statements that the buildings they have designed and constructed are safe because they have followed the IS code. How correct is this belief?

Design code

There are three main aspects that a design code has to address, viz. delineation of seismic hazard, specification of the forces due to the hazard and methods of structural design for the estimated forces. The last step is perhaps the easiest since structural engineering design is a very well developed subject. Code or no code every civil engineer is trained from his undergraduate days in the design of load carrying elements. It is for the first two aspects namely hazard estimation and force computation the common designer is entirely dependent on the code and hence believes the code as sacrosanct.

Unfortunately, the track record of IS-1893 on these two aspects has been rather poor. The seismic zonation map of BIS has been proved to be wrong after the Koyna (1967) and Khillari (1993) earthquakes. Officially as of now the seismic map still has five zones, zone I, which includes Bangalore, Hyderabad, and Latur being the least hazardous. However a draft revision is in circulation since 1996 in which all parts of India which were previously in zone I have been shown as being in zone II. The best BIS does is to react to an earthquake by changing the zone boundaries. What level of confidence the public should place in such a map, and what sort of accountability the IS-code owes to its users? It is true that natural hazards are uncertain and hence mapping the quantified hazard is a tough proposition. But this doesn't permit the code makers to absolve their responsibility by putting up a deterministic zonal map with an attached sense of finality to change it after the next earthquake hits. For earthquakes peak ground acceleration (PGA) as a descriptor is used extensively all over the world. The return period maps of PGA for the country can be computed with the present knowledge of seismo-tectonics of the subcontinent.

Such statistical hazard description is not new to engineers. In fact BIS in its code on design against wind (IS-875) has adopted a statistical approach for describing the wind hazard. It is not clear why handling uncertainty in statistical terms is being avoided as far as earthquakes are concerned. The draft revision of 1996 puts up new jargon such as maximum credible earthquake and effective peak acceleration, without realising that these are statistical terms which need careful description before incorporation into the codes. In its present form the Indian code juggles with factors and multipliers to arbitrarily fix that in the most seismically active part of the country, for the most important structure, the peak acceleration shall not exceed 0.18g (g=981 cm/s/s). This limit has remained constant for nearly three decades. How credible this figure may be for bridges and dams can be seen by looking at the corresponding value for a seismically less hazardous country like the U.K., which uses a value of 0.375g.

It was mentioned earlier that the second use of the code is to find the seismic forces on buildings and other structures. Ground motion recorded on firm soil has energy in the range of periods 0.1 to 3 seconds, which is also the range in which civil engineering structures have their natural periods. The shape of the spectrum generally varies but by and large on alluvial soils the spectrum peak occurs in the range of 2 to 3 hz. However on hard rock much higher frequencies are possible.

Obvious mistake

The IS code till 1984 had given a response spectrum shape which was physically and mathematically incorrect. A very stiff and rigid structure with a high natural frequency, say a rock boulder or a concrete block, would move with the ground and not relative to the ground. The code had ignored this fact even though this has been common knowledge in structural dynamics. It is rumoured that this changed only after a professor from the U.S., who was visiting India as a World Bank expert, took exception to the obvious mistake. For the first time the spectra put up by BIS in 1984 were based on data recorded on Indian soil. This shape still reflects the situation reasonably well for the northern part of the country.

However, the standard spectrum given by the code deviates considerably from the spectra recorded in the peninsular region of India. Records from Khillari and Koyna regions show energy content at 10hz (0.1 sec) and beyond. Thus the proposal of the IS code to use the same spectral shape all over India doesn't reflect the seismological variation correctly. The use of the spectrum given in the IS code will lead to underestimation of forces on short high frequency structures whereas forces on tall, slender structures in the peninsular region will be overestimated.

Naive and unjustifiable

The IS code in its present form is again not of much help for cities that have evolved on existing faults. During a strong earthquake the epicentral tracts can show severe topographic modifications such as permanent sinking, uplift and ground cracking. The code addresses only ground shaking or vibration and not possible ground deformation. Cities such as Delhi are susceptible to both the above hazards. Ground deformation and shaking arise due to local faults whereas distant events will induce ground vibration as evidenced during the Uttarkashi (1991) and the Chamoli (1999) earthquakes. Cities founded on river banks have unconsolidated soil deposits of varying depths over which human settlement exists. Tank beds dried up naturally or otherwise have also been developed by governmental agencies as real estate.

Thus to view a whole city as being in zone IV or V and to attribute all structural damage to non-compliance of IS code regulation would be naive and unjustifiable. The earthquake resistant design code of BIS in its present form is not sophisticated enough to handle issues related to liquefaction, soil amplification, variation in building materials and construction practices, which largely dictate the eventual seismic risk. Micro-zonation of risk is the need for cities such as Delhi which have buried faults and deep alluvial deposits.

Earthquakes are low probability but high risk events. It is time BIS adopted a strictly objective attitude towards Indian seismological data and presented the hazard maps and response spectrum along scientific lines.

R. N. IYENGAR

Send this article to Friends by E-Mail

Section  : Features
Previous : The power of analysing
Next     : Probity in public life