::rn: F I tJ 0 d Ge.o M 0 ,... ph r> / 0 cl cds. V.7<. c. r\ochel/-P.C.?a..tton 'Wi ie-y 4- S 07'oS / N. Y. 25 GEOMORPHOLOGIC CONTRIBUTIONS TO FLOOD CONTROL PLANNING THOMAS DUNNE Department of Geological Sciences. University of Washington. Seattle. Washington flOOD CONTROL STRATEGIES The enormity of damage caused by floods and the cost of mitigating their effects have stimulated engineers, geo- morphologists, and planners to develop a wide range of flood control-strategies appropriate to different physical and cultural settings. The methods most commonly employed for the reduction of flood damage are: 1. Flood warning and emergency action 2. Impoundments 3. Channel alteration and stabilization 4. Diversion and storage of floodwaters above and below ground 5. Land management for soil and water conservation 6. Control of land use on floodplains Geomorphology can contribute to the choice' and design of these strategies. The possibilities range from the cor- relation of flood potential with simple descriptive indices developed in the early days of quantitative geomorphology to more sophisticated analyses of flow and sediment transport processes. Most applications of geomorphology arise be- cause of the interaction among flooding, sedimentation, and channel behavior. This chapter reviews some of the uses of geomorphology in flood control planning and sug- gests that inclusion of geomorphologic studies leads to more successful and stable schemes. Flood Warning and Emergency Action The primary responsibility for flood prediction, forecasting, and response planning rests with meteorologists, hydrol- ogists, and planners. However, there are geomorphologic aspects of the flood prediction problem (Reich, 1971; Ors- born, 1976). Physiographic regions have different flood potentials because of variations in elevation and associated climate, drainage density, channel gradients, and width of valley floor. Therefore, in the prediction of flood discharges it is useful to stratify large drainage basins into homogeneous subdivisions, and to design hydrometeorologic networks, and to monitor and route floods from each region, combining them downstream. These regional differences are already taken into account in some hydrologic procedures such as the use of regional flood frequency curves (Dalrymple, 1960; Wiard, 1962) and synthetic unit hydrographs (Snyder, 1938). However, with widespread availability of satellite imagery for mapping the large-scale geomorphic features of even remote areas, and recent developments in automatic monitoring and telemetry, there is potential for improving flood prediction and warning through systems designed partly on a geomorphic base. At a slightly greater level of complexity. Kirkby (1976) and Valdes et aI. (1979) have demonstrated how the structure of a drainage network influences flood hydrographs and can be used for their prediction. Much research remains to be done by geomorphologists and hydrologists to follow these suggestions. On a smaller scale, within flood-prone areas there is often considerable uncertainty about the consequences of a predicted discharge. Subtle changes in channel charac- teristics can alter the timing and magnitudes of flood waves. Campbell and others (1972) have made some illustrative computations of the effect of channel straightening on flood hydrographs. Burkham (1976) documented how a series of large storms at the head of Safford Valley, Arizona, 421