Journal of Applied Hydrology Vol. VI, Nos. 1 to 4, 1993 pp. 4 0 - 6 6

FLOOD GEOMORPHOLOGY OF THE INDIAN PENINSULAR RIVERS Vishwas

S. Kale, Sheila 'Mishra,** Yehoucla Enzel," Lisa Ely* S.N. Rajaguru*' a n d V.R. B a k e f

Department of Geography, University of Poona, Pune (India) 'Department of Geosciences, University of Arizona, Tucson (USA) **Department of Archaeology, Deccan College, Pune (India) '

ABSTRACT Recently geomorphological techniques are employed to enhance the geological evidence in extending the flood records. Palaeoflood studies provide information about the upper limits of the largest floods and increase the reliability of risk analysis. Extreme floods form an integral element of the hydrologic system in the Deccan Peninsular rivers. Maximum flood 'discharges from a series of rivers in the Deccan Peninsula are used to prepare an envelope curve encompassing gauged peak discharges to define the physical or hydrometeorological limit of the magnitudes for maximum flood that can be expected. Minimum streampowervalues estimated for some rivers indicated high tiansport and erosional capabilities. . The need for comprehensive and complete study to establish Palaeoflood chronologies of the Indian rivers in general and the peninsular rivers in particular is stressed. Introduction Flood Geomorphology deals with the study of the geomorphic aspects and chronology of flood events, using geologic and sedimentologic information. In the 1950s and 1960s the emphasis in fluvial geomorphology was on the morphogenetic impacts of common discharges with a recurrence interval of 1 to 2 years. However, in the last two decades the interest in the geomorphic effects of rare and catastrophic floods has increased and several studies have been undertaken in USA, Australia, Israel and China (Baker et al, 1988). Conventionally, two approaches are adopted in estimating the magnitudes of floods 1. Flood Frequency Analysis (FFA) of gauged data by using annual maximum series and the partial duration series of floods. Different probability distributions are fitted to estimate the exceedencc probibility of high magnitude floods.

2. Application of hydrological medels for determining the Probable Maximum Floods (PMF). The modelling of the response of a river basin, to design or to hypothetical extreme storms are undertaken to estimate the PMF. Both approaches have certain limitations. Large, and. rare floods are often underrepresented or unrepresented or may even be overrepresented in the short-term gauged records and therefore FFA fails to adequately describe the flood hydrology of the rivers. The second approach is more rheoretical and often overestimates the PMF values (Enzel et al, 1993). Therefore, geomorphologists have used the geological evidence to extend the records of floods. The geomobhic studies provide information about the historic and pre-historic floods and complement

Vishwas S. Kale e t a1

short-term flow records (Baker et al, 1988). Palaeoflood studies provide information about the upper llimits of the largest floods (Enzel et al, 1993) and thus increase the reliability of risk analyses and improve the understanding of the flood hydrometeorolopy.

Methodology The methodology adopted in the palaeoflood studies includes (Baker et al, 1988, Janet 1991) : 1. Geomorphic and sedimentologic studies : Several geomorphic features, known as PalaeoStage Indicators (PSI) such as scour lines and silt lines, occurring at different elevations, are used to estimate the high flood levels. Flood deposited sediments, known as Slack Water .Deposits (SWD), demarcate peak stages and provide accurate information about the flood levels. 2. Flood modeling by generating water surface profiles with a step-backwater hydraulic model (for example HEC-2). The variables required for modeling include surveyed cross-sections to characterize channel geometry.

3. Comparison of the elevation of rhe surveyed SED-PSI with the n~odelledwater surface profiles and estimation of the magnitude of largest floods. 4. Palaeodischarges are also estimated from the force necessa j to transport large boulders. Such an approach known as threshold-velocity and stream power approach uses theoretical and enlpirical equations. I

5. Dating the SWD to reconstruct the flood chronology. Flood Hydrology of Indian Peninsular Rivers Indian Peninsular rivers such as. Godavari, Krishna, Cauvery, Narrnada and Tapi (Fig. 1) expericnce serni-arid and dry subhumid typeof' climate and the annual precipitation varies between 550

Flood Geomorphology

and > 1000 mm. The floods are dependent upon the amount and spatial distribution of rainfall process, in the monsoon season. Extreme floods form an integral element of the hydrologic system in the Deccan peninsular rivers. Table I gives the list of major floods recorded in this century. Table I : Major flood events in this Century - -

p -

Rivers

Date

Discharge (cms)

Synop. conditions

-

1. Cauvery 2. Cauvery 3. Cauvery 4. Cauvery 5. Narmada 6. Cauvery 7. Gadilam 8. Narmada 9. Godavari 10. Tapi 11. Tapi 12. Tapi

13. 14. 15. 16. . 17. 18, 19. 20. 21. 22. 23. 24. 25. 26.

27. 28. 29. 30.

Tapi Krishna Godavgri

~ahnada Godavari Godavari Tapi Narmada Krishna Narmada Nmada Krishna Tapi Narmada Narmada Cauvery Narmada Godavari Godavari Nannada

31. 32. 33. Tapi

LPS Shallow Dep.

1923,9 July 1923,~ e c . 1924,July 25 1925,Nov., 9 1926,Sept. 23 1930,Oct. 23 1933,Dec. 16 1937,July 13,26 1937,July 22 1937,July U,26. 1941,July 1, 11 1942,Aug. 6 1944,Aug. 24 1949,Sept. 24 1953,Aug. 11 1954,Sept. 24 1958,Aug. 30 1959,Sept. 14 1959,Sept. 17 1959,Sept. 15 1961,July 1961,Sept. 17 1964,Aug. l3 1964,Sept. 28 1968,Aug. 6 1968,Aug. 6 1970,Sept. 6 1972,Dec. 5 1973,Aug. 30 1976 Aug. 22 1983,Aug. 14 1984,Aug. 20 1991,July

Source : IDWR,Rarnaswamy (1987) and other reports ; LD Pressure System

Cyclone . Cyclonic St. Cyclonic St Storm Depression Cyclone 2 LD 2 Depressions 3 Depressions 2 Depressions Depression 2 Cy. Storms 5 Cy. Storms Depression Depression Depression Depression

-

4 Depressions 4 Depressions Depression Depression Depression Land Dep. Cyclonic St. Depression Cyclonic St. Cyclonic St. Cyclonic St . Depression =

Land Depressions ; LPS = Low

'According to Dhar et a1 (1986) and others, the meteorological situatiyns which are responsible for high floods in the Deccan Peninsular region are as follows : (a) Incidence of heavy rainfall associated with monsoon depressions and cyclones originating in the Bay. of Bengal and the Arabian Sea.

-

Vishwas S. Kale e t a1

(b) Prolonged active to vigorous monsoon conditions ; and (c) Monsoon breaks. Most flood-generating monsoon conditions over. the Peninsular India can be grouped under four synoptic situations, viz., Low Pressure Systems, Cyclonic Storms, Bay Depressions and Land Depressions. In addition, in recent years floods have been caused by dam breaks for example Panshet (Mutha-1961), Morvi (Machhu-1979), Chandora (Tapi-1991), etc. Gauge data indicate that the Peninsular rwers have experienced devastating floods and the peak discharges range from 10,000 crns to >60,000 crns. The majority of the peak floods are recorded in August and September for all the Peninsular rivers, 'except Cauvery. Available data reveal the following exponential relationship between the catchment area (A in km2) and maximum peak discharge (bmax in crns) : Qmax = 10

(r2 = 0.91)

In comparison with the Dicken's Formula (Mutrej 1986), which is commonly used in central and northern India, the exponent is comparable but the base constant is lower. The available gauge and historical records of floods, indicate that this century has experienced relatively high flood activity. This can be attributed to either the natural increase in the flood frequency or to the increasing anthropogenic effects in the past 100 years or simply to better instrumental gauged records. The maximum flood discharges from a series of rivers in the Deccan Peninsula have been used to prepare an envelope curve encompassing gauged peak discharge to define the.physica1 or hydroclimatological limit of the magnitude for maximum flood that can be expected (Fig. 2). The curve, broadly agrees with she envelope curves for Indian rivers reported in CWC (1972) report. The curve indicates that most Southern rivers have lower upper limits than the Northern and Central Indian rivers. The minimum stream power values estimated for some rivers indicate that the peak floods, recorded in this century, are characterized by hgh transport and erosional capabilities (0 = 1 X 103t05 X 105w/m). These values are comparable with the estimated 0 for Amazon, Mississippi, Katherine Gorge and Chang Jiang (0 = 1.5 x 104t05 x 106w/m) (Baker and Costa, 1987). The collation suggests that the Peninsular rivers can accomplish avariety of sedimentologic and hydraulic phenomena, including the erosion of bedrock and transportation of coarser sediments in suspension, during the monsoon floods.

Palaeoflood Hydrology The discharge generated over most of Peninsular India is funnelled through bedrock gorges. These gorges are particularly suitable for the accumulation and preservation of SWD. Scour lines, boulder berms, potholes and SWD have been observed at several places, reflecting high floods in recent times. Preliminary palaeoflood investigations were carried out in parts of Upper Knshna and Godavari Rivers in Maharashtra, Choral Narmada and Tapi Rivers in Madhya Pradesh (Fig. 1). Geomorphic, sedimentologic and archaeologic studies reveal that records of 7 to 30 floods have been documented in the SWDs. Archaeological material and a few C-14 dates suggest that most of the flood deposits are not older than 300 years. Scour lines in the Narmada (near Punasa) and flood deposits in the Choral (Barjar) suggest that they are associated with flood discharges of >60,000'cms and 24500 cms respectively. Step-backwater modeling has yielded flood discharges between 1500 and 2000 crns for the Pushpavati river. fi

Flood Geomorphology

PENINSULAR RIVERS Qmax IN C M S

- Envelope Curvs

CATCHMENT AREA IN 8Q. KM.

Fig.2

SLACK WATER DEPOSITS CHORAL RIVER

- BARJAR

SAND WIT11 RUBBLE

Fig. 3 53

r'

Vishwns S . Kale et a1

I

Huge doulders, measuring 2 to 3 meters in diameter in the Pushpavati Gorge (Upper Krishna) Kadakwasala (Mutha - ca 10 m), Dhanora (Tapi) and at Dardi Falls (Narmada) (Fig. 1) indicatc that the minimum unit stream power (w) associated with the modern, historic and pre-historic flood. is 800 - 10,000 watts/m. In addition to these, d,eep canyons, inner channels, deep potholes, boulde berms and flood-scoured gorge walls, etc., which cannot be dated directly, provide evidence of higl flood power, associated with high magnitude floods. In the Pushpavati Gorge (Fig. I), 18th centuty pottety has been discovered in the slackwater deposits Similarly, C-I4 dates of hearth from Annapur (770 f 80 yrs BP) from Ghod valley, shells from Bib (180 f 90 yrs BY) in Bhima Valley represent catastrophic flood events in the Krishna Basin. A Manjn, on Krishna, human skeletons have been discovered below flood deposits about 6 m thick an( about 4,000 yrs BP in age. Sedimentological evidence of high magnitude floods in the Chalcolithil period ( c a 3000 yrs BP) have been preserved at Daimabad in Godavari Valley (Rajaguru 1986) However, due to unconfined nature of the channels, at the above mentioned sites, it is difficult tc determine the flood magnitude. The oldest dated material from .Barjar (Choral river) indicate tha the flood deposits were emplaced by floods that occurred about 5100 yrs BP (Fig. 3). \

The rivers under review thus contain abundant evidence of late Holocene floods in general and lato flistorical in particular. The abundance of confined bedrock canyons and the monsoonal climate art conducive to the preservation of the geologic evidence of palaeofloods. Nevertheless, the fact tha rhe late Historical records are relatively well represeked in most Peninsular rivers suggest, that eithe the prehistoric records have been destroyed by flood erosion or that the late historical flood event are rhe largest. Similar inferences have been drawn from studeis in the south-western USA (Ely an( Baker 1990) and Australia (Pickup 1989). I n view of the significance of such studies for the reconstruction of palaeo-climates and the bettel unclcrstanding of the flood hydrology of the rivers, it is necessary to undertake a comprehensivt ;md complete study to establish the palaeoflood chronologies of the Indian rivers in general and the I'cninsular rivers in particular.

Acknowledgement VSK snd SM are grateful lo the Department of Science and Technology, Ncw Delhi for providing l'u~idsi'or field work and data collection. Additional support, particularly for Yc, LE and VRB was provicled by thc US National Science Foundation, Engineering Directorate, Natural and Manmade Hxzards Mitigation Program.

References Baker, Y.R. and Costa, J.E. (1987) : Flood Power ; In Catastrophic Flooding, (Eds.) L. Mayer and D. Nash, Allen and Unwin, London, pp. 1-24. Baker, V.R., Kochel, R.C. and Patton, P C . (1988) : Flood Geomorphology (Eds.), John Wiley and Sons, New York. CWC (1972) : Estimation of Design Flood Recommended Procedures ; Central Water Commission. New Delhi.

. Dhar, O.N., Mulye, S.S., and Manda, B.N. (1987) : The highest ever floods in the Major rivers - P brief appraisal ; Transactions of Inst. Ind. Geographers, Val. 8, pp. 13-26.

Flood Geornorphology

Ely, L.L. and Baker, V.R. (1990) : Large Floods and Climate Change in Southewestern United States. In : Hydraulic/Hydrology of Arid Land ; Ed. French, R.H. : Anlerican Soc. of Civil Engineers, N.Y., pp. 361-366. Enzel Y., Ely, L.L., House, P.K., Baker, V.R. and Webb, R.H. (1993) Paleoflood evidence for a Natural Upper bound to flood Magnitudes in the Colorado river basin ; Water Resources Research (in press). Janett, R.D. (1991) : Palaeohydrologyanditsvalue in Analyzing Floodsand Droughts, U.S. Geological Survey Water Supply Paper 2375, pp. 105-115. Mutreja, KN.(1986) : Applied Hydrology ; Tata McGraw-Hill Pub. Co. Ltd., New Delhi, pp. 674478.

Pickup, G. (1989) : Palaeoflood Hydrology and Estimation of the Magnitude, Frequency and Aerial extent of Extreme floods - An Australian Perspective, Civ. Engg. Trans. I.E. Aust. vo ; CE 31, pp. 19-29. Rajaguru, S.N. (1986) : Geoarchaeology of Daimabad. In : Sail SA. (Ed.), Daimabad 1976-79, Archaeological.Su~eyof India, New Delhi, pp. 580-87.