Water Resources Management 14: 369–376, 2000. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.
369
Design Flood Estimation Using GIS Supported GIUH Approach S. K. JAIN, R. D. SINGH and S. M. SETH National Institute of Hydrology, Roorkee 247, 667 India, e-mail:
[email protected] (Received 4 April 2000; accepted 12 December 2000) Abstract. Quantitative understanding and prediction of the processes of runoff generation and its transmission to the outlet represent one of the most basic and challenging areas of hydrology. Traditional techniques for design flood estimation use historical rainfall-runoff data for unit hydrograph derivation. Such techniques have been widely applied for the estimation of design flood hydrograph at the sites of gauged catchment. For ungauged catchments, unit hydrograph may be derived using either regional unit hydrograph approach or alternatively Geomorhological Instantaneous Unit Hydrograph (GIUH) approach. The unit hydrograph thus derived may be used for the simulation of flood events for the ungauged catchments. In this study Gambhiri dam catchment located in Rajasthan, India is selected for applying this approach. Gambhiri river is a small tributary of the Berach/Banas river of the Chambal basin in Rajasthan, India. The objective of the present study is to apply Geographical Information System (GIS) supported GIUH approach for the estimation of design flood. A mathematical model has been developed at the National Institute of Hydrology, which enables the evaluation of the Clark Model parameters using geomorphological characteristics of the basin. This model has been applied for the present study. From this study it is observed that the peak characteristics of the design flood are more sensitive to the various storm pattern as well as method of critical sequencing followed for the computation of design storm patterns. Earlier estimates for the peak and time to peak hydrograph was 9143.74 cumec and 18 hrs. respectively. However, the estimates for the peak characteristics of design flood hydrograph obtained from the GIUH based approach are 11870.6 cumec and 19 hrs. respectively considering the same design storm pattern. Key words: GIUH, design flood, GIS, flood hydrograph
1. Introduction The nature of stream flow in a region is related to the rainfall characteristics and and watershed geomorphology. The rainfall characteristics are the temporal and spatial distribution of the rainfall quantity. The geomorphic characteristics are the channel network and surrounding landscape, which translate the rainfall input into an output hydrohraph at the outlet of the watershed. One of the simpler approaches to rainfall runoff modelling is application of the unit hydrograph. A significant advances in the unit hydrograph method for an ungauged area was the development of the Geomorphological Instantaneous Unit Hydrograph (GIUH) (Lee K. T., 1998). The GIUH approach was originated by Rodriguez-Iturbe and Valdes (1979), who rationally interpreted the runoff hydrograph in the framework of travel
370
S. K. JAIN ET AL.
time distribution explicitly accounting for geopmorphological structure of a basin. Many investigators have endeavored to relate the IUH parameters to the catchment geomporhology and thus obtain the GIUH. (Chowdhry et al., 1995; Bhaskar et al., 1991; Sorman, 1995; Jain et al., 1997; Lee, 1998). In the GIUH approach, rainfall excess is assumed to follow different paths on overland areas and in channels of different stream orders to reach the watershed outlet. According to this approach the basic equations are of a third-order basin. The equations for higher-order can be derived with exactly the same framework. However, for basins of any order the peak qpg and the time to peak tpg , which are the most important characteristics of the GIUH, are worked out from the derived functional relationship of the GIUH as given below (Rodriguez-Iturbe, 1993) qpg = 1.31
RL0.43
tpg = 0.44
RL−0.38
V Lω
RB RA
(1)
0.55
Lω V
(2)
RB , RL and RA represent the bifurcation ratio, the length ratio and the area ratio. In this equation L ω is length of the highest order stream measured in km, V flow velocity in m/s, qpg in h−1 and tpg in h. The above equations represent general relationships which allow the estimation of the peak and time to peak of the IUH for any watershed. On multiplying Eq.(1) and Eq.(2) we get a non-dimensional term qpg * tpg as: qpg ∗ tpg = 0.5764
RB RA
0.55 (RL )0.55
(3)
The term qpg ∗ tpg does not depend upon the velocity and thereby on the storm characteristics and hence it is a function of only the catchment characteristics. One advantage of applying the geomorphic instantaneous unit hydrograph (GIUH) approach is due to its potential of deriving the UH using the information derived from topographic maps or remote sensing, possibly linked with geographic information system (GIS) and digital elevation model (DEM). The input to a GIS may be remotely sensed data, digital models of the terrain, or point or aerial data compiled in the forms of maps, tables or reports. GIS provide a digital representation of watershed characterization used in hydrologic modelling. Hydrologic applications of GIS have ranged from synthesis and characterization of hydrologic tendencies to predict the response of hydrologic events (Tao and Kouwen, 1989). A GIS can provide the basis for hydrologic modelling of ungauged catchments and for studying the hydrologic impact of physical changes within a catchment. The integration of GIS into hydrologic models follows one of the two approaches
DESIGN FLOOD ESTIMATION USING GIS SUPPORTED GIUH APPROACH
371
Table I. Geomorphological characteristics of Gambhiri dam catchment Order
No. of streams
Average Length
Average area
Value of constants
1 2 3 4 5 6
1062 247 68 16 4 1
0.794 1.187 2.781 7.662 17.995 13.430
0.492 2.928 13.578 57.944 238.839 969.27
Rb =4.00 Rl =1.946 Ra =4.49
(i) to develop hydrologic models that operate within a GIS framework (Moore et al., 1987), (ii) to develop GIS techniques that partially parameterize existing hydrologic models (Deroo et al., 1989). In the present study second approach has been considered i.e. some of the parameters are evaluated through GIS and applied to the GIUH model separately.
2. Methodology and Results Gambhiri river is a small tributary of the Berach/Banas river of the Chambal basin. The river rises in Madhaya Pradesh, India and travelling for a length of about 83km in the North West direction joins the Berach river downstream of Chittorgarh town. The Gambhiri river is joined by its important tributary, Daru nadi near the Gambhiri dam site. The GIS software used in the present study is Integrated Land and Water Information System (ILWIS). The boundary of the catchment and all the streams have been mapped at a scale of 1:50,000 from Survey of India toposheets. Also a contour map on the same scale was prepared. Both these maps were then converted to digital form using digitization and stored in ILWIS. For stream order, Strahler’s ordering system, has been followed (Strahler, 1957). According to this ordering system, which is applied through ILWIS over the entire drainage network of the study area, it is found that it is a sixth order basin. In the system, length of each stream is stored in a table. Then after adding length of each stream for an order, the total stream lengths of each order are evaluated. The drainage network map of Gambhiri dam catchment is prepared and Digital Elevation Model (DEM) has been developed. For the catchment the elevation varies from 300 to 1100 m. Table I provides the details of these geomorphological characteristics for the Gambhiri dam catchment. From this table it is observed that the bifurcation ratio, length ratio and area ratio, which are non-dimensional characteristics are 4.00, 1.946 and 4.49 respectively for the Gambhiri dam catchment.
372
S. K. JAIN ET AL.
These values are within the limits, which have already been reported in literature (Rodriguez-Iturbe, 1993). The DEM data generated from ILWIS for Gambhiri dam catchment are utilized to develop isochronal map for the catchment wherein the isochrones are plotted at hourly interval. In the present study the time area diagram is prepared using a DEM file in the GIS environment. The distance from the most upstream point in the basin to the outflow location along the principal watercourse is measured. The time-area diagram is prepared considering the time of travel between any two points proportional to the distance and inversely proportional to the square root of the slope between them. An initial estimate of the time of concentration may be obtained using the Kirpich’s formula. tc = 0.06628L0.77 H −0.305
(4)
Where tc is concentration time in hours L is the length of main stream in km H is the average slope of the stream (m/km) Substituting the values of L and H in equation (4) the time of concentration tc may be computed. The time area diagram using GIS has been developed and is shown in Figure 2. Since the rainfall-runoff records for this catchment is not available, recourse is made to obtain the synthetic unit hydrograph based on the regionalised regression equations recommended by the Central Water Commission, India (CWC, 1983) for the region. The time base of this 1 hour unit hydrograph for the catchment is 24 hours. The duration of the design storm as per the recommended practice considered being equal to this 24 hour for this catchment. The India Meteorological Dept. has published the PMP atlas for the whole country. The point PMP value as obtained from this atlas is given elsewhere (CES report, 1994). The rainfall distribution was available at 2-hour intervals. 1-hour rainfall distribution is arrived at interpolating the rainfall values available at 2-hourly intervals. The computed PMP rainfall excess increments are to be brought to a critical chronological pattern found in the observed storm characteristic of the basin under study in order to produce critical flow rates. For the catchment, critical time sequencing is carried out considering four different cases. First case is the sequencing in one bell mode where in the highest rainfall is kept in the middle and other rainfall values are alternatively arranged left and right to the highest rainfall block. In twobell case the highest value of first 12 hours rainfall is arranged as for one bell case. Similarly the remaining 12-hour rainfall values are also arranged. In three bells case same procedure is adopted after dividing the entire duration in three equal parts. For four bell case entire duration was divided in to the equal interval of 6 hours each. Critical sequencing has been carried out considering each 6-hour interval as one bell.
DESIGN FLOOD ESTIMATION USING GIS SUPPORTED GIUH APPROACH
373
Figure 1. Drainage network map of the study area.
In this study, the peak characteristics of the IUH as obtained in Equations (1) and (2) are utilised for the evaluation of Clark model parameters. Once the parameters of the GIUH based Clark model are known, the complete IUH may be derived. The step-by-step explanation of the procedure for estimating the design flood using the proposed approach is given elsewhere (Jain et al., 1997). Each set of design storm data together with geomorphological characteristics, time area diagram and initial parameter values are supplied to the GIUH based model. The design flood hydrograph for the Gambhiri dam catchment are computed considering the design storm of one, two, three and four bell respectively and shown in Figure 3. The results obtained using GIUH approach is shown as solid line in the figure. The results obtained are also compared with the flood hydrograph using the other approach given in the report prepared by Consulting Engineering Services (CES report, 1994). From Figure 3, it is observed that difference in design
374
S. K. JAIN ET AL.
Figure 2. Time area diagram of the study area.
flood hydrograph peak corresponding to design storm of two bell system is maximum for the two cases and minimum for the design storm of one bell system of design storm. Unfortunately for this site no historical records were available and also the information regarding the hydraulic characteristics were inadequate for evaluating the velocity. Under the circumstances reasonable values of velocity have been assumed based on the indirect sources available for the basin (Jain et al., 1997). 3. Conclusions For estimation of geomorphological characteristics and time area diagram, GIS technique has been used. The estimation of these parameters can be handled easily and more accurately using GIS which otherwise is very tedious using manual methods. It is observed that the design flood is more sensitive to the design storm pattern and its time distribution. From the study it is observed that the GIUH and GIS based approach has potential application for the estimation of the design flood particularly for the ungauged catchment.
DESIGN FLOOD ESTIMATION USING GIS SUPPORTED GIUH APPROACH
375
Figure 3. Rainfall-runoff simulation for (i) one bell, (ii) two bell, (iii) three bell, (iv) four bell case.
376
S. K. JAIN ET AL.
However certain limitations are also there in the present study, one is due to nonavailability of data for velocity estimation. The other limitation is non-availability of historical rainfall-runoff records and due to which the methodology could not be validated for the observed events. References Bhaskar, N. R. and Devulapalli, R. S.: 1991, Run-off Modelling Geomorphological Instantaneous Unit Hydrograph and ARC/INFO Geographic Information System, Proc. Civil Engineering applications of remote sensing and GIS, edited by D. B. Stafford, published by ASCE. CES Report: 1994, ‘Report on safety evaluation and remedial works to enhance safety status of dams, design floods at Morel, Alnia, Gambhiri and Jawai dams’, Consulting Engg. Services (India) Private Limited, August 1994. Chowdhary, H.: 1995, Derivation of GIUH for small catchments of Upper Narmada and Tapi subzone (subzone 3C)- Part 1, NIH, Roorkee, India. CWC, 1983: Flood Estimation Report for Upper Narmada and Tapi Subzone (Subzone 3C), Directorate of CWC, New Delhi. DeRoo, A. P. J., Hazelhoff, L. and Burrough P. A.: 1989, Soil erosion modelling using ANSWERS and GIS, Earth Surface Processes and Land Forms 14, pp. 517–32. Gupta, V. K., Wymire, Ed and Wang, C. T.: 1980, A Representation of an Instantaneous Unit Hydrograph from Geomorphology, Water Resources Research 16(5), pp. 855–862. Jain, S. K., Chowdhry, Hemant, Seth, S. M. and Nema, R. K.: 1997, Flood estimation using a GIUH based on a conceptual rainfall-runoff model and GIS, ITC Journal 1997-1, pp. 20–25. Lee, K. T.: 1998, ‘Generating design hydrographs by DEM assissted geomorphic runoff simulation: a case study’, Journal of the American Water Resources Association 34, No. 2, April 1998. Renaldo, A. and Rodriguez-Iturbe, I.: 1996, ‘Geomorphological theory of the hydrological response’, Hydrological Process 10, 803–829. Rodriguez-Iturbe, I. and Valdez, J. B.: 1979, ‘The geomorphologic structure of hydrology response’, Water Resources Research 15(6), 1409–1420. Rodriguez-Iturbe, I.: 1993, The Geomorphological Unit Hydrograph, Chapter 3, Channel Network Hydrology, Edited by K. Beven and M. J. Kirby, John Wiley & Sons Ltd. Sorman, A. U.: 1995, ‘Estimation of peak discharge using GIUH model in Saudi Arbia’, Journal of Water Resources Planning and Management 121, No. 4, July-August, 1995. Strahler, A. N.: 1957, ‘Quantitative Analysis of Watershed Geomorphology’, Transaction of American Geophysicsical Union 38(6): 913–920. Tao, T. and Kouwen, N.: 1989, Remote sensing and fully distributed modelling, Journal of Water Resources Planning and Management 115, No. 6, pp. 809–23. Valdes, J. B., Fiallo, Y. and Rodriguez-Iturbe, I.: 1979, ‘A rainfall runoff analysis of the Geomorphologic IUH’, Water Resources Research 15(6), 1421–34.