River flow analysis

The ultimate aim of many computational techniques in engineering hydrology is the derivation of river discharges, and it might appear that, once these are obtained, the hydrologist's work is done. However, whether they are gained indirectly from considerations of other hydrological variables (to be described in following chapters) or directly from river discharge measurements, the discharge data are only samples in time of the behaviour of the river. The hydrologist then must assess the utility of the data and their representativeness over the period for which the information is required, usually the expected life of a water engineering project.

The basic river flow data that are normally readily available from the responsible agencies1 are daily mean discharges and instantaneous peak discharges. These are usually derived from river stages (Section 7.3) and complementary velocity-area measurements (Section 7.4) to give a stage-discharge rating curve (Section 7.7), or from discharges obtained with a calibrated structure (Section 7.6). They will normally have been quality controlled, but for higher peak discharges may be associated with some uncertainty, because of the need to extrapolate the rating curve beyond the range for which measurements are available. For greater detail, the hydrologist may also be able to obtain hourly or 15 mm interval stage recordings and equivalent discharges.

Two examples of a year's record of daily mean discharges from rivers in the UK are shown in Fig. 11.1 from the gauging station at Sheepmount on the River Eden in the Environment Agency north-west region and the Panshanger Park gauging station on the River Mimram in the Thames region (note that the discharge scale is logarithmic). The daily mean discharges are for the year 2007 and are shown in relation to the extreme daily mean flows, the maximum and minimum discharges on corresponding days in the complete period of record. Although 2007 was a generally average year overall, the extremely high flows in June and July on the River Eden show quite clearly in relation to the long-term extremes (Fig. 11.1a). June and July in 2007 was the period of extensive flooding in Yorkshire and further south in the Severn and Thames catchments (see Section 9.1.4). The great irregularity shown by the sequence of daily mean flows during the wet months is indicative of a catchment responding rapidly to rainfall and the steep recessions in January/February and March/April emphasizes lack of storage in the drainage basin. A contrasting record is seen in Fig. 11.1b. The daily mean discharges for the Mimram, one of the headwaters of the River Lea north of London, for the same year of 2007, show a much more even pattern. The heavy rainfalls in the summer period have produced a few peaks on the hydrograph but the

Hydrograph River Mimram

Fig. 11.1 Mean daily discharge record for 2007 and daily extremes (darker grey shading) recorded in the period of record: (a) Eden at Sheepmount (076007), 1967-2008; and (b) Mimram at Panshanger Park (039019), 1952-2008. (Reproduced from the UK National River Flow Archive, Centre for Ecology and Hydrology.2 Copyright NERC CEH.)

Fig. 11.1 Mean daily discharge record for 2007 and daily extremes (darker grey shading) recorded in the period of record: (a) Eden at Sheepmount (076007), 1967-2008; and (b) Mimram at Panshanger Park (039019), 1952-2008. (Reproduced from the UK National River Flow Archive, Centre for Ecology and Hydrology.2 Copyright NERC CEH.)

flow is generally well within the long-term extreme values. Lack of variation in the daily flows is a notable feature of a catchment with a large storage. The Mimram drains a part of the Chiltern Hills, an area predominantly composed of chalk, which dampens the effects of minor irregularities in daily rainfalls, but a clay subcatchment responds quickly to rainfall and is largely responsible for the summer peaks.

The records that are shown here are gauged daily mean discharges calculated directly from the measurements made at a gauging station. However, some rivers may be

Fig. 11.2 Derwent at Yorkshire Bridge (28001). Mean daily discharges for 2006 and daily extremes (darker shades of grey) from period of record, 1933-2006. (Reproduced from the UK National River Flow Archive, Centre for Ecology and Hydrology, Wallingford. Copyright NERC CEH.)

affected by large abstractions upstream of the gauging station or the river flow may be controlled totally by regulated reservoir releases. The hydrologist must pay attention to potential modifications such as those illustrated in Fig. 11.2 by the hydrograph for 2006 for the River Derwent, a tributary of the River Trent, at Yorkshire Bridge, immediately downstream of series of reservoirs built for the water supply of Sheffield. The effect of the reservoirs, in both maintaining low flows and controlling high flows, is evident throughout the summer and in November/December when the reservoirs were refilling. For some types of application it is important to account for abstractions to give estimates of naturalized or gross flows (see also Section 17.6.1).

Abstractions from a river, although taken regularly each day for domestic or other water supplies, are usually quantified on a monthly basis. Hence a very useful statistic of river flow is the average of the daily mean flows over a month (the monthly mean flow). Table 11.1 shows the monthly mean gauged and naturalized discharges for the River Thames at Teddington for 1973. Teddington Weir is the tidal limit of the Thames, and thus the difference between the gauged and naturalized flows gives a measure of the demands of London and other towns abstracting water supplies from the river on the freshwater resources of the river.

Further monthly statistics for a selection of rivers are given in Table 11.2. The highest instantaneous peaks recorded are essential for assessing regulation requirements and the maximum and minimum daily mean discharges indicate the range of water availability. The value of these statistics is enhanced with each year of record: the longer the record at a gauging station, the more reliable can be the evaluation of water resources and the estimation of extreme events either of dangerous floods or of harmful droughts (see sections below).

Table 11.1 Monthly mean gauged and naturalized discharges (m3 s 1) River Thames at Teddington 1973 (drainage area 9870 km2)

Flow

J

F

M

A

M

J

J

A

S

O

N

D

Gauged

50.1

42.9

25.4

22.0

33.7

21.7

22.6

11.7

11.6

9.8

10.0

20.3

Naturalized

69.5

60.3

43.5

42.1

50.7

40.4

41.3

29.6

27.5

28.2

27.3

41.4

Difference

19.4

17.4

18.1

20.1

27.0

18.7

18.7

16.9

15.9

18.4

17.3

21.1

Figures rounded off to the first decimal place. (Reproduced from Department of the Environment (1978) Surface Water: United Kingdom 1971-73, by permission of the Controller, Her Majesty's Stationery Office. © Crown copyright.)

Table 11.2 Monthly statistics of gauged flows (m3 s-1; data from National River Flow Archive, Centre for Ecology and Hydrology). Monthly highest instantaneous peaks and extreme daily mean discharges in 1990

River Area Flow (m3 s 1)

Table 11.2 Monthly statistics of gauged flows (m3 s-1; data from National River Flow Archive, Centre for Ecology and Hydrology). Monthly highest instantaneous peaks and extreme daily mean discharges in 1990

River Area Flow (m3 s 1)

(km2)

J

F

M

A

M

J

J

A

S

O

N

D

Tay at

4587

Peaks

776

1746

1102

252

137

496

297

84

194

597

213

535

Ballathie

Max.

663

1647

1033

223

119

234

228

81

112

423

201

443

Min.

98

243

212

107

47

45

47

43

51

51

99

79

Derwent at

1586

Peaks

57.1

59.2

23.5

8.2

9.2

24.4

9.6

4.2

4.7

17.8

28.2

84.5

Buttercrambe

Max.

52.5

57.6

22.4

8.0

8.4

19.8

8.8

3.9

4.3

15.5

27.2

83.2

Min.

6.8

13.9

7.8

5.3

4.3

3.9

3.2

2.9

2.8

3.4

5.7

6.5

Little Ouse at

699

Peaks

5.6

15.2

9.1

4.0

3.2

2.2

2.1

1.4

2.0

-2.7

3.2

Abbey Heath

Max.

4.0

13.9

8.4

3.6

3.0

1.8

1.3

1.2

1.3

1.5

2.0

2.2

Min.

2.3

3.8

3.0

2.6

1.5

1.1

0.9

1.0

1.0

1.1

1.0

1.2

Usk at Chain

912

Peaks

404

627

80

17

16

19

18

7

41

99

74

286

Bridge

Max.

230

472

68

14

12

14

15

6

20

62

58

146

Min.

35

54

1 1

8

5

5

4

3

3

5

10

12

Eden at

2286

Peaks

485

705

160

42

103

46

117

21

64

310

192

457

Sheepmount

Max.

333

528

126

38

69

29

77

16

38

207

101

318

Min.

34

84

21

18

13

13

11

10

10

17

20

18

Example: Converting units of discharge (see also Appendix, Table A5). Taking the maximum daily flow for the River Eden at Sheepmount for January 1990 of 333 m3s-1 over a catchment area of 2286 km2 (Table 11.2), this can also be expressed as:

333 x 24 x 60 x 60 = 28 771200m3 d-1 = 28 771.2 MLday-1

333 x 24 x 60 x 60 x 1000/(2286 x 1000 x 1000) = 12.59mmday-1

11.1 Peak discharges

High river discharges are caused by various combinations of extreme conditions. Heavy rainfalls over short durations, deep snow cover melted by warming rain and moderate rainfalls on frozen ground or saturated soil, can all contribute to a rapid and large runoff. Flood conditions are of great concern, and notable events are always studied in detail (see the discussion of the Boscastle 2004, Carlisle 2005, UK Summer 2007 and Aigle 1991 events in Section 9.1). With the expansion of the river gauging network in the UK, many more floods are now being measured, but when the river has overtopped the gauging control and where a river is ungauged, only estimates of the peak discharges can be made by hydraulic calculations (see Section 7.7.2).

Peak discharges from major runoff events in the UK were first assembled in relation to reservoir practice (Institution of Civil Engineers, 1933). Later this was updated by the Institution of Civil Engineers (1960), the Flood Studies Report (Natural Environment Research Council; NERC, 1975) and the Flood Estimation Handbook (Institute of Hydrology, 1999). A database of historical flood events for over 1000 sites in England and Wales has been collated by the UK Environment Agency,3 while the University of Dundee hosts the British Hydrological Society database of historical floods in the UK4 (see Section 11.9). In the USA, the USGS National Water Information Service maintains a database of flood peaks for over 27 000 sites.5

From a combination of flood discharge measurements and post-event discharge calculations taken from the available records in the UK, a graph of maximum recorded peak discharge against catchment area can been drawn (Fig. 11.3a). Twenty peak discharges for areas ranging from 4 km2 to nearly 1800 km2 are plotted, and an envelope curve is drawn. Most of the points represented by catchments less than 25 km2 pertain to the tributary flows contributing to the disastrous Lynmouth flood in August 1952, where the peak flow was estimated at 650 m3 s-1 from 101 km2 (Dobbie and Wolf, 1953). The highest measured flow in the UK is the 2402 m3 s-1 on the River Findhorn at Forres in Scotland in 1969 from a catchment area of 782 km2. The highest measured in England and Wales is the 1516 m3 s-1 on the River Eden at Sheepmount in the 2005 Carlisle flood from a catchment area of 2286 km2. There are other historic records of flood levels on many rivers in the country,3 and some of these could possibly provide peak discharges above the curve in Fig. 11.3a. However, changes in the river profiles and cross-sections over the ensuing years makes the conversion of level data into discharges unreliable. A worldwide relationship between maximum floods and catchment area is shown in Fig. 11.3b from data provided by O'Connell and Costa (2004).

The US data come from drainage areas that range from 2 km2 to over 8000 km2, with corresponding peak discharges from 144 m3 s-1 to 21 000 m-3 s-1 (Costa and Jarrett, 2008). These extreme events in the USA come from a wide range of different climatic regions, some of which experience tropical rainfall intensities. From the enveloping curves of the two plots, a 5 km2 catchment somewhere in the USA may well produce a peak of 600 m3 s-1, but in the UK a peak of only 150 m3 s-1 may be expected from the same drainage area. (The Findhorn record in Scotland would not be worth plotting on the USA graph!) Thus, when making a general appraisal of peak discharges, such a relationship of peak discharge to catchment area should be made according to climatic region. For the Flood Studies Report in the UK, regional subdivisions were

(a) Area (Km2)

1000000

100000

10000

1000

100000

10000

1000

1000 Area (Km2)

Fig. 11.3 Some extreme peak discharges and upper envelope curves for drainage basins of different area for: (a) the UK (data added to those of Boorman et al., 1990); and (b) world (data from O'Connell and Costa, 2004). Note the different scales.

1000 Area (Km2)

10000

Fig. 11.3 Some extreme peak discharges and upper envelope curves for drainage basins of different area for: (a) the UK (data added to those of Boorman et al., 1990); and (b) world (data from O'Connell and Costa, 2004). Note the different scales.

Table 11.3 Peak discharge per unit area from some documented floods in the UK

River (site)

Area (km2)

Peak discharge m3 s-lkm-2

Year

Chulmleigh (Devon)

1.7

40.0

1982

Caldwell Beck (Dumfries)

5.8

32.6

1979

Claughton Beck (Lancs)

2.2

30.2

1967

Red-a-ven (Dartmoor)

4.0

27.6

1917

Hoaroak Water (Lynmouth)

8.1

18.4

1952

Tyne (East Lothian)

4.1

10.0

1948

Alphin Brook (Exeter)

7.2

8.3

I960

Valency (Boscastle)

23.1

6.09

2004

Wye (Pant Mawr)

27.2

6.40

I973

Divie and Dorback (Moray)

365.0

5.31

I970

Muick (Invermuick)

110.0

4.28

1981

Findhorn (Forres)

782.0

3.07

I969

Oykel (Easter Turnaig)

330.7

2.56

I978

Tywi (Dolau Hirion)

231.8

2.30

I979

Rawthey (Brigg Flatts)

200.0

2.24

I982

Dart (Austins Bridge)

247.6

2.22

I979

Tyne (Haydon Bridge)

751.1

1.24

2005

Eden (Sheepmount)

2286

0.66

2005

made, but not only for hydrological reasons. At the time, the regions corresponded to the administrative areas of the Regional Water Authorities. In the Water Act of 1995 parts of the responsibilities of the Water Authorities were subsumed into the national Environment Agency. This is reflected in the Flood Estimation Handbook that abandoned the regional analyses in favour of grouping catchments together according to their hydrological characteristics.

Some hydrologists prefer to convert the absolute peak discharges into relative values per unit area for plotting against catchment area, thus giving an inverse relationship and a declining curve with increased catchment area. Sample values from the UK are shown in Table 11.3. These have been taken fromBoorman et al. (1990). They demonstrate the contrast between the exceptional discharges from small areas (< 10 km2) evaluated after the events and the peak flows at formal river gauging stations.

Numerous empirical formulae can be found in the hydrological literature for the relationship between peak discharges and catchment areas, with the coefficients specifically determined for particular countries or climatic regions. In attempting to use such formulae to obtain peaks for ungauged catchments, the hydrologist must guard against applying them to inappropriate conditions and areas.

11.2 River regimes

In Table 11.1, the monthly mean discharges for the River Thames in 1973 exhibit a distinctive seasonal pattern, with the highest values occurring in the winter months. The expected pattern of river flow during a year is known as the river regime. Flow records for 20-30 years are required to provide a representative pattern, since there may be considerable variations in the seasonal discharges from year to year. The averages of the monthly mean discharges over the years of record calculated for each month, January to December, give the general or expected pattern: the regime of the river.

The river regime is the direct consequence of the climatic factors influencing the catchment runoff, and can be estimated from knowledge of the climate of a region. The eminent French geographer, Pardé, first identified and classified distinctive river regimes and this can be a great help to engineers faced with unfamiliar conditions and sparse data. The classification is based on an understanding of the role of the main climatic features, temperature and rainfall, in causing river runoff. An illustration of simple river regimes resulting from a single dominant factor is given in Fig. 11.4. For each river, the monthly mean discharges from January to December are represented as proportions of the mean of the 12 monthly values. In this way, the graphs are comparable and independent of the absolute values of the monthly mean discharges and catchment areas.

11.2.1 Temperature-dependent regimes

Rivers with a dominant single source of supply, initially in the solid state (snow or ice), produce a simple maximum and minimum in the pattern of monthly mean discharges according to the seasonal temperatures.

• Glacial. When the catchment area is over 25-30 per cent covered by ice, the river flow is dominated by the melting conditions. Such rivers are found in the high mountain areas of the temperate regions. There is little variation in the pattern from year to year, but in the main melting season, July and August, there are great diurnal variations in the melt water flows.

• Mountain snowmelt. The seasonal peak from snowmelt is lower and earlier than in a glacial stream, but the pattern is also regular each year providing there has been adequate winter snowfall. The low winter flows are caused by freezing conditions.

• Plains snowmelt. The regular winter snow cover of the interior regions of the large continents in temperate and sub-Polar latitudes melts quickly to give a short 3-month season of high river flows. The timing of the peak month depends on latitude, with the more southerly rivers (e.g. the River Don in Fig.11.4a) having rapid flood peaks in April, whereas further north there is a slower melt and the lower peak occurring in June is more prolonged (the River Ob, Fig. 11.4a).

11.2.2 Rainfall-dependent regimes

In the equatorial and tropical regions of the world with no high mountains, the seasonal rainfall variations are the direct cause of the river regimes. Temperature effects in these areas are mostly related to evaporation losses, but with these being dependent on rainfall, the overall effect of evaporation is of secondary importance in influencing the river flow pattern.

• Equatorial. Drainage basins wholly within the equatorial belt experience two rainfall seasons with the annual migration of the intertropical convergence zone, and these are reflected directly in the river regime (e.g. the River Lobe in Cameroun in Fig. 11.4a).

Temperature dependent Glacial

+1 0

Post a comment