Why are we so helpless in containing the spates of Brahmaputra and coping with its flood hazard? Why do our efforts go awry and all civil engineering measures end up in shambles? The answer, in my opinion, lies in our failure to recognize the reality of active faults and continuing crustal movements in this geodynamically restless region. Understandably, the flood coping measures have never been designed to accord with this recognition.

Girdled as it is, by the arms of the Eastern Himalayan Syntaxis – the knee-bend of the mountain ranges – the
Assam terrane is underthrusting northwards under the Arunachal Himalaya and, less energetically, eastwards beneath the Indo-Myanmarese ranges^2,3. There is, therefore, very severe deformation and attendant faulting and thrusting in the terrane caught between the Himalaya–PatkaiNaga ranges and the Meghalaya massif (Figure 1). The drastic reduction of the width of the alluvial plains from 350 to 300 km, respectively, in the flood plains of Sindhu and Ganga to less than 100 km in the Brahmaputra basin is not without significance. Coming of the Meghalaya–Mikir blocks of the Peninsular Indian Shield closer to the Himalaya explains this attenuation of the alluvial domain of the Quaternary foreland basins.

The severe deformation of the Assam region is eloquently expressed in its much faulted framework^4,5. The E–W trending faults (Dauki Fault, Brahmaputra-Mikir Fault), and the transverse tear faults (Kopili Lineament, Dhubri Fault, Dudhnoi Fault, Chidrang Fault, Um Nagot Lineament) and thrusts (Dapsi Thrust, Barapani Thrust) are among the many that dissect the terrane of the Meghalaya–Mikir blocks (Figure 2). The E–W Dauki and Brahmaputra–Mikir Faults roughly demarcate the southern and northern physiographic limits of the Meghalaya Plateau which is a horst of sorts. The plateau stands as a ~ 2000 m high physiographic eminence against the sunken 3–4 m high (above sea level) Sylhat Plains in Bangladesh.

Most of these faults are seismically quite active⁴ as borne out by the distribution of epicentres in the fault zones (Figure 2), such as the locations of 7 earthquakes of M ³  4.5 along the 26.5°

parallel – the line of the Brahmaputra–Mikir Fault; and in the NNW–SSE trending belt of the Kopili Lineament (Figure 2).

The catalogue of earthquakes of magnitude 4.5 and above in the period 1897–1992 shows the higher (87%) seismicity in the Meghalaya-Mikir blocks and adjoining Indo-Myanmarese range and northern Bangladesh plains (Figure 2), compared to near absence in upper Assam and lower seismicity in Arunachal Himalaya⁶. In the Meghalaya block, the western part (Tura region) riven with faults exhibits greater

seismicity as testified by major earthquakes of 1897, 1923, 1930 and 1943 (Figure 2).

Gowd et al.⁶, Molnar⁷ and Gahlaut and Chander⁸ attribute the higher seismicity of the Meghalaya–Mikir blocks to reactivation of what has been postulated as a gently dipping shallow thrust or midcrustal detachment. I believe that many E–W and transverse faults that dissect the block (Figure 2) are active. Compared to the N3° E orientation of ^SH_max in the Meghalaya–Mikir blocks, it is N26° E in upper Assam (and N23° E in Peninsular Indian Shield)⁶. The NNW–SSE trending Kopili Lineament of high seismicity thus constitutes the tectonic boundary between the two stress domains. Significantly, it was in the belt of this Kopili Lineament that earthquakes have occurred recurrently – one of M 7.2, three of 6.0 £  M < 7.0 and several of 4.5 £  M < 6. Doubtless, tectonic movements have been taking place time and again on the faults of the Kopili Lineament. This seismically active Kopili lineament cuts across the Brahmaputra about 35 km NW of Navagaon between Mikir Hills and the prolongation of Meghalaya massif (Figure 2).

Levelling observations⁹ made three times during 1910–197  indicated that the blocks of the Guwahati–Dergaon section have been consistently rising – the vertical velocity increasing progressively from 0.3 mm/yr to 4.5 to 31 mm/yr at Dergaon (approximately 30 km west of Jorhat) implying faster rise of the Mikir block. In the Guwahati–Goalpara sector, the eastern side moved in the NNE direction and the western part shifted SSW-ward during 1856–1938 (ref. 9). North of the Dauki Fault, the Meghalaya block has been rising at the rate of 0.3 to 0.4 mm/yr (ref. 10).

It is obvious that the block delimited by seismically active faults has been rising perceptibly though variably in space and time. The uplift of the Meghalaya block is evident in the gorge that the otherwise wide Brahmaputra has cut near Guwahati (Figure 3) – in the place where a N–S lineament traverses the river.

The uplift of the Meghalaya block – through the prolongation of which the Brahmaputra has cut its channel – must have caused ponding of the river. The continuing movements must be impeding its flow with resultant accumulation of sediments in the upstream channel (Figure 4). The deposition of great volumes of sediments – occurring as islands and bars – in the channel has phenomenally reduced the carrying capacity of the Brahmaputra. Flood waters are bound to spill over and spread far and wide. The 1897 earthquake of M 8.7 had caused not only the ponding of streams but also blockade of the Brahmaputra. According to Oldham¹¹, there was a deluge unrelated to rains as a result of a barrier formation across the

 Brahmaputra, downstream of Hathi- mura, accompanied possibly by subsidence of the floor under the river. The 15 August 1950 earthquake (M 8.7) had likewise drastically affected the gradient of this river, stopping the flow temporarily and bringing about flooding and rapid accumulation of enormous volume of sediments in the channel. The lowest water level rose by 3 m as a result of the earthquake, and near Dibrugarh and downstream the channel was silted up 2.5 to 3 m. The scarp¹³ that one sees north-east of Dibrugarh (Figure 5) is probably the surface expression of the fault that lifted up the downstream block and caused ponding of the Brahmaputra.

Imagine a similar development overtaking the Brahmaputra valley in the event of a major earthquake! Taking

into consideration the reality of the continuing tectonic movements on active faults, it should be evident that building of embankments and drainage channels in upper Assam would not bring lasting relief from flood hazards. The solution lies in channelizing discharge of the order of 48,160 m³/s (ref. 12) of the Brahmaputra through canals, aqueducts, tunnels, etc. across the northward prolongation of the Meghalaya block, particularly between Guwahati and Hathimura and other places (west of Mikir Hills) where identified active faults cross its channel. The canals would have to be deepened periodically in order to keep pace with the rate of uplift of the block. Absence of channelization of the Brahmaputra in spate would always cause ponding, leading to deposition of sediments, reduction in carrying capacity and

 inundation of the flood plain. In my opinion, this is what is happening to- day. It will therefore be advisable to embark on a comprehensive multidisci-

1. Goswami, S., Souvenir: National Workshop on Geodynamics of North-eastern India, Dibrugarh University, Dibrugarh, 1998, pp. 36–41.
2. Mukhopadhyay, M. and DasGupta, S., Technophysics, 1988, 149, 299–322.
3. Acharyya, S. K., Indian J. Geol., 1997, 69, 211–234.
4. Nandy, D. R. and DasGupta, S., Phys. Chem. Earth., 1991, 18, 1147–1163.
5. Kayal, J. R. and De, R., Bull. Seismol. Soc. Am., 1991, 81, 131–138.
6. Gowd, T. N., Srirama Rao, S. V. and Chary, K. B., Curr. Sci., 1998, 74, 75–80.
7. Molnar, P., J. Geol. Soc. India, 1987, 30, 13–27.
8. Gahalaut, V. K. and Chander, R., Tectonophysics, 1992, 204, 163–174.
9. Rajal, B. S. and Madhwal, H. B., Himalayan Geol., 1996, 17, 17–32.
10. Chugh, R. S. and Valdiya, K. S., Indian J. Geol., 1989, 61, 1–13.
11. Oldham, R. D., Memoir Geol. Surv. India, 1899, 29, 1–379.
12. Goswami, Dulal, C., in Flood Studies in India, Geol. Soc. India, Bangalore 1998, pp. 53–75.
13. Proc. Intern. Symp. Neotectonics of South Asia, Survey of India, Dehradun, 1986.