OBTAINING SCS SYNTHETIC UNIT HYDROGRAPH BY GIS TECHNIQUES For determining design discharges for hydraulic structures, it is necessary to determine unit hydrographs for the corresponding basins. In Turkey, suitable data to determine the unit hydrograph of a basin are not easy to find, therefore unit hydrographs are usually determined synthetically. Coefficients used in the synthetic unit hydrograph determination are taken from studies made for the specific areas in the world. In this study, synthetic unit hydrographs are obtained for a basin where rainfall-runoff data were available to determine actual unit hydrographs also. The goals of the study were, first, to compare the unit hydrographs obtained synthetically and from observed rainfall-runoff data, and if the results were acceptable, then the method used to determine synthetic unit hydrograph could be used for ungaged basins also, and second, to use GIS techniques in determining synthetic unit hydrograph and show its effectiveness. To satisfy the second goal, SCS synthetic unit hydrograph determination method was chosen since the method was very suitable for the application of GIS techniques.
1. INTRODUCTION In Turkey, rainfall and runoff data are seldom adequate to determine unit hydrographs of drainage basins. When it is necessary to determine a unit hydrograph for a basin, therefore, one of the synthetic unit hydrograph determination methods is used. The most commonly used methods are the Snyder, the Mockus, the State Hydraulic Works (DSI) synthetic and the U.S. Soil Conservation Service (SCS) methods, [4]. Most of these synthetic methods require some coefficients for the basin under study. Since such coefficients do not exist for a large part of Turkey, they are generally taken from the studies made for the specific regions of the world. In this study, unit hydrographs are obtained for a basin in Northwest part of Turkey from observed rainfall and runoff data for a number of storms. Then using SCS synthetic unit hydrograph method and Geographic Information Systems (GIS) techniques, synthetic unit hydrographs are determined for the same storm durations for this same basin.
2. STUDY AREA The study area, Kumdere Basin, is located in North-Western region of Turkey, near the city Edirne, Fig. 1. It is a subbasin on a very upstream branch of Meric River which flows to Aegean Sea making a boundary between Turkey and Greece. This basin is a research basin belonging to one of the Research Institutes of General Directorate of Rural Services (GDRS).
Kirklareli Atatürk Research Institute started collecting hydrometeorologic data in this basin in 1985. Kumdere Basin has an area of 4.40 km 2 and there are one runoff and three precipitation gaging stations as shown in Fig. 2. All of these gages are recording type. The characteristics of the basin are obtained from a 1:25000 scale topographic map manually and then checked by GIS techniques as will be explained later. The maximum and minimum elevations of the basin are 154 m and 115 m respectively. As it can be seen from the hypsometric curve, Fig. 3, the median elevation is 141 m. This basin has medium slope and consequently medium erosion range. From on site investigations it was found out that brown color soil with poor drainage condition covers the surface of the basin.
Kumdere Basin has continental climate with cold and wet winters, and hot and dry summers, observed highest and lowest temperatures are 41.5°C and -22.2°C respectively. Total annual rainfall of the basin was found as 530.5 mm, which is close to the average value for Turkey, and corresponding total annual runoff as 8.8 mm. Seasonal precipitation distribution is 37.6% in fall, 24.5% in winter, 26.8% in spring and 11.1% in summer. The highest monthly rainfall occurs in November with an average value of 92.4 mm and the lowest in July as 17.4 mm obtained for the record period. The highest runoff is also observed in November as an average value of 2.5 mm, but the lowest in April as zero, for the same period. Hydrometeorological characteristics of the basin are given in Table 1, and monthly total rainfall and runoff histograms are given in Fig. 4, [1]. As it is clear from this figure also, the runoff coefficient of this basin is very low, the annual average being only 1.66 %, which is the case for most of the upstream subbasins in Turkey.
Table 1: Hydrometeorologic Characteristics of Kumdere Basin MONTHS X
XI
XII
I
II
III
IV
V
VI
VII VIII
IX
Annual
AVE. TEMP. (°C)
2.2
3.9
7.2
12.7
17.8
21.9
24.4
23.9
19.6
14.1
9.1
4.5
13.4
TOTAL PRECIP. (mm)
52.9
92.4
54.0
30.4
43.7
56.1
51.4
46.1
44.6
17.4
21.6
19.9
530.5
AVE. REL. HUM. (%)
81
77
73
68
67
62
57
57
63
72
80
83
70
AVE. WIND VEL. (m/hr)
1.5
1.6
1.9
2.0
2.2
2.2
1.9
1.6
1.6
1.6
1.6
1.5
1.8
TOTAL RUNOFF (mm)
0.4
2.5
1.4
0.1
1.6
1.5
0.0
0.3
0.1
0.1
0.6
0.2
8.8
3. UNIT HYDROGRAPH FROM OBSERVED DATA Unit Hydrograph is the hydrograph of the surface runoff which is the response of a basin to a uniformly distributed rainfall producing 1 cm effective depth over the basin. Unit hydrograph can be determined in gaged basins by measuring the concurrent rainfall and runoff amounts for the storms. Since one of the main assumptions in unit hydrograph theory is the linearity principle, when a unit hydrograph is determined for a basin, then its response to any other storm can be obtained very easily and this is very important for the design of hydraulic or any other structure which would be affected from floods in this basin. Observed storm hyetographs and their corresponding total runoff hydrographs are studied in detail to derive unit hydrograph for Kumdere Basin. In the record period of 11 years there were a number of storm data with concurrent rainfall and runoff values, but among them only 8 single storms were found to be suitable for unit hydrograph derivation. Using well known methods that could be found in engineering hydrology text books, one 5 min., two 10 min., two 30 min. and three 60 min. unit hydrographs were obtained for these eight storms. These unit
hydrographs, referred to as observed unit hydrograph from now on, are given in Figure 5, and their characteristics in Table 2. Table 2 : Parameters of Observed unit hydrographs tr
tp
Tb
Qp
(min)
(h:m)
(h:m)
(lt/s/mm)
May. 5, 93
5
0:25
4:20
800
Dec. 11, 88
10
2:30
6:50
814
Jan. 27, 95
10
1:00
5:55
682
10 min. ave.
10
1:45
6:22
748
May. 30, 95
30
2:02
11:00
432
Oct. 17, 91
30
0:30
4:25
1304
30 min. ave.
30
1:00
6:18
868
Dec. 9, 87
60
3:30
9:00
434
Aug. 7, 89
60
2:20
7:40
507
Nov. 11, 92
60
0:30
4:20
1039
60 min. ave.
60
2:06
7:00
660
Storm Date
4. SYNTHETIC UNIT HYDROGRAPH
As mentioned before, actual or observed unit hydrographs can not be determined for all the basins since there are not available rainfall and runoff data everywhere. Therefore for such basins unit hydrographs are determined synthetically, to be used in the design of hydraulic structures. Synthetic unit hydrographs are developed along two main lines; 1) each watershed has a unique unit hydrograph, and 2) all unit hydrographs can be represented by a single family of curves or a single equation. There are a number of synthetic unit hydrograph derivation methods proposed by Bernard, McCharty, Snyder, Commons, Williams, Mitchell, SCS and Gray, following one or the other of the main lines, [5]. A number of parameters is important in determining the shape of the unit hydrograph for a watershed. The discharge parameter which is mostly used is the peak discharge (Q p). Lag time (t L), time to peak (t p), time of concentration (t c) and base time (T b) are often used as the time parameters. Watershed parameters of most concern, influencing the shape of the outflow hydrograph, include area (A in sq. mi.) and its shape, main stream length (L in ft), length to watershed centroid from the outlet (L c in ft) and average slope of basin (y in %), [2]. In this study SCS method is used, since it requires geographical parameters which can very easily be obtained by using GIS techniques. As mensioned before, the study area is one of the research basins of Kirklareli Ataturk Research Institute of GDRS. SCS method was developed by U.S. Soil Conservation Service in 1957. It is based on dimensionless unit hydrograph which is developed from a large number of unit hydrographs obtained from basins ranging in size and from different geographic locations. In this method, the hydrograph is represented as a simple triangle (Fig. 6) with peak flow Qp (cfs), time to peak t p (hr), and time of fall B (hr) for a rainfall duration t r (hr), 2.
From the review of a large number of hydrographs, it was found that for one inch of rainfall excess;
From Fig. 6 it can be seen that;
Lag time, t L, is the time from the centroid of rainfall to the occurrence of peak flow and it is estimated from the following equation :
where; the parameters are as defined above in the text, with the followings, S = 1000/CN - 10 CN = Curve number for various soil/land use combination.
The determination of the curve number (CN), which is a function of soil and land use characteristics of the basin, is essential for this method. In addition to these two characteristics, some other hydrologic conditions of the basin such as the vegetal cover and antecedent moisture situation in previous five days, are important factors in estimation of CN. In order to determine the representative CN value of a basin, basin area has to be divided into sub-areas which have same land use and soil type characteristics. Then CN values for every sub-area are determined using appropriate tables. After determination of CNs for particular sub-areas, weighted average of the CN values with respect to their areas will give the representative CN value for the whole basin. In addition to the information necessary for the determination of curve number, other information like area, main stream length and average slope of the basin are also required for SCS method. The use of GIS techniques comes into picture in this part of the study.
5. USE OF GIS TECHNIQUES GIS techniques are applied to derive unit hydrographs of the basin using SCS Curve Number method. In this study PC version of ArcInfo was used as the GIS software, [3]. Determination of areas with the same characteristics in a basin is very easy with overlay analysis of GIS, if the coverages for these different characteristics of the basin are available. Similarly the areas, lengths and slopes can easily be obtained from these coverages. Therefore as the first step for GIS part of the study, necessary coverages were determined, and five different coverages were obtained by digitizing the corresponding maps which were provided by GDRS, using ArcInfo software as described below: KUCONT : Coverage of contour lines; a line coverage from 1:25000 scale topographic map. KULUS : Coverage of land use; a polygon coverage from 1:25000 scale land use map prepared by GDRS personnel. KUSOIL : Coverage of soil type; a polygon coverage from 1:25000 scale soil type map prepared by GDRS personnel. KUNET : Coverage of stream network; a line coverage from 1:25000 scale topographic map. Digitization for stream network was done separately for each different order of branches and then they are joined to get the whole network.
KUBND : Coverage of basin boundary; a polygon coverage from manually drawn basin boundary on 1:25000 scale topographic map, since it was impossible to obtain boundary in PC version of the software. Following the digitization step, necessary corrections were made and elevation of the contour lines, land use and soil type codes, and stream orders are added to the corresponding attribute tables of the coverages. After these preparations, the parameter values such as basin area, main stream length, average basin slope and basin perimeter were obtained easily. Determination of the curve number for the whole basin was achieved by using overlay analysis. In overlay analysis a union of two or more coverages is obtained to get the areas with the same characteristics. Two examples for coverages of Kumdere Basin are given in Fig.'s 7 and 8, and the union coverage for CN determination is in Fig. 9.
6. CONCLUSION The results of the study are given as a summary in Table 3. The characteristics used for the comparison are peak discharge (Q p), time to peak (t p) and base time (Tb) for different effective rainfall durations (t r ). ). In this table observed hydrograph parameters are given as average values for different duration storms. Beneath
the average values, the same parameters which are obtained from SCS synthetic unit hydrograph method are given. As it is seen in the table, average values of these characteristics of observed data are in harmony with the synthetic ones in the basin. Observed and synthetic values match reasonably well especially in peak discharges. Table 3 : Observed and Synthetic Unit Hydrograph Characteristics
tr
tp
Tb
Qp
diff. in Qp
(min)
(h:m)
(h:m)
(lt/s/mm)
(%)
OBSERVED
5
0:25
4:20
800
SYNTHETIC
5
1:10
3:08
806
OBSERVED
10
1:45
6:22
748
SYNTHETIC
10
1:13
3:14
778
OBSERVED
30
1:00
6:18
868
SYNTHETIC
30
1:22
3:41
684
OBSERVED
60
2:06
7:00
660
SYNTHETIC
60
1:38
4:21
580
0.75 4.01 -21.20 -12.12
The only exception to this is 30 minute storm. This inconsistency could be generated from errors in observed data and their processing as well as from the errors in the application of synthetic method. For example, models that consider the spatial and temporal distribution of CN instead of using CN as a lump parameter would produce more precise values for the estimated parameters. This study has taken into consideration a very limited application of GIS in hydrology compared to many other possible applications. With these digital data at hand for Kumdere Basin, any further study and updating any information on the basin will be very fast and easy using GIS techniques. In addition, using macro language of ArcInfo software would make the repetitive analysis for different basins in an automated way and save time. On the other hand using GRID module, some coverages which were used in the study could be generated instead of digitizing. One of these coverages is the boundary coverage and the other is stream network coverage. These characteristics must be obtained with different grid sizes to get the appropriate detail in the coverage. GRID module can also help to use synthetic models in a distributed way. Parameters can be determined for parts of the area instead of the whole basin. These parts can of course be as small as the grid size and much precise studies can be performed. Remote Sensing technology could be coupled with the method to update land use information, where possible.
In this study, application of GIS techniques to determine synthetic unot hydrograph for a basin was demonstrated with reasonable results. The authors are thankful to the Kirklareli Atatürk Research Institute for providing necessary data and maps.
7. REFERENCES 1. Akbay, S., Bakanogullari, F., 1996, Rainfall and Runoff Characteristics of Edirne - Kumdere Basin. (Report of 1985 - 1994), Kirklareli Atatürk Research Institute of GDRS, Kirklareli. (in Turkish) 2. Bedient, P., Huber, W., 1948, Hydrology and Flood Plain Analysis, AddisonWesley Publishing Company. 3. Understanding GIS, The ArcInfo Method, 1990, Environmental Systems Research Institute Inc., Redlands. 4. Usul, N., Tezcan, B., 1995, Determining Synthetic Unit Hydrographs and Parameters for four Turkish Basins, Journal of Soil and Water Conservation, Vol. 50, p:170-173. 5. U.S. Department of Agriculture, Linear Theory of Hydrologic Systems, Technical Bulletin No.1468, Agricultural Research Service, 1973, Washington D.C.
AUTHORS INFORMATION: Nurünnisa USUL, Assoc.Prof. Middle East Technical University Civil Engineering Department Water Resources Laboratory Inonu Bulv. Ankara / TURKEY Tel : (312) 210 5448 Fax : (312) 210 1262 e-mail :
[email protected]
Okan KÜPÇÜ, Res. Assit. Middle East Technical University Civil Engineering Department Water Resources Laboratory Inonu Bulv. Ankara / TURKEY Tel : (312) 210 2485 Fax : (312) 210 1262 e-mail :
[email protected]
