HA 37/97

Hydraulic Design of Road-edge

Surface Water Channels
THE HIGHWAYS AGENCY
THE SCOTTISH OFFICE DEVELOPMENT DEPARTMENT
THE WELSH OFFICE
Y SWYDDFA GYMREIG
THE DEPARTMENT OF THE ENVIRONMENT FOR
NORTHERN IRELAND
DESIGN MANUAL FOR ROADS AND BRIDGES
Summary: This Advice Note describes a method of determining the length of road between
outlets that can be drained by a given size of surface water channel constructed
along the edge of the road. The design method can be used for channels of
triangular, rectangular or trapezoidal cross section.
August 1997
DESIGN MANUAL FOR ROADS AND BRIDGES
VOLUME 4 GEOTECHNICS AND
DRAINAGE
SECTION 2 DRAINAGE
PART 4
HA 37/97
HYDRAULIC DESIGN OF ROAD-EDGE
SURFACE WATER CHANNELS
SUMMARY
This Advice Note describes a method of determining the
length of road between outlets that can be drained by a
given size of surface water channel constructed along
the edge of the road. The design method can be used for
channels of triangular, rectangular or trapezoidal cross
section.
INSTRUCTIONS FOR USE
1. Insert HA 37/97 into Volume 4, Section 2.
2. Remove HA37/88 and Amendment No 1 from
Volume 4, Section 2.
3. Archive this sheet as appropriate.
Note: A quarterly index with a full set of Volume
Contents Pages is available separately from the
Stationery Office Ltd.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
REGISTRATION OF AMENDMENTS
Amend Page No Signature & Date of Amend Page No Signature & Date of
No incorporation of No incorporation of
amendments amendments
Registration of Amendments
Volume 4 Section 2
Part 4 HA 37/97
August 1997
REGISTRATION OF AMENDMENTS
Amend Page No Signature & Date of Amend Page No Signature & Date of
No incorporation of No incorporation of
amendments amendments
Registration of Amendments
VOLUME 4 GEOTECHNICS AND
DRAINAGE
SECTION 2 DRAINAGE
PART 4
HA 37/97
HYDRAULIC DESIGN OF ROAD-EDGE
SURFACE WATER CHANNELS
Contents
Chapter
1. Introduction
2. Scope
3. Safety Aspects
4. Basis of Design Method
5. Description of Design Equations
6. Storm Return Period
7. Geographical Location
8. Channel Geometry
9. Channel Gradient
10. Drainage Length
11. Channel Roughness
12. Catchment Width
13. Surcharged Channels
14. By-Passing at Outlets
15. Construction Tolerances
16. Worked Examples
17. Glossary of Symbols
18. References
19. Enquiries
DESIGN MANUAL FOR ROADS AND BRIDGES
August 1997
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Annex A Rainfall data
Annex B Channel roughness
Annex C Run-off from cuttings
Volume 4 Section 2
Part 4 HA 37/97
August 1997
1. INTRODUCTION
1/1
Chapter 1
Introduction
1.1 In the United Kingdom there are three main
methods of dealing with surface run-off from rural trunk
roads: filter drains; kerbs and gullies connecting to pipes
below ground; and surface water channels along the
pavement edge. This document is concerned with the
hydraulic design of surface water channels.
1.2 The use of surface water channels allows larger
distances between outlets than conventional gully
systems, and the positions of the outlets can often be
chosen to suit the local topography and the occurrence of
natural watercourses. With this type of system, seepage
flows in the pavement construction are usually drained
by means of fin drains at the edge of the pavement.
Geometric and performance requirements for surface
water channels and fin drains are given in Volumes 1 to
3 of the Manual of Contract Documents for Highway
Works (MCHW 1, 2 & 3) and in HA 39, Edge of
Pavement Details (DMRB 4.2).
1.3 In order to design a surface water channel for a
given frequency of flooding, it is necessary to take
account of the time of travel of flow along the channel
and the variation of rainfall intensity with storm
duration. The design method described in this document
is obtained by combining results from kinematic wave
theory about time-varying flow conditions in channels
[Refs 3 and 4] with a newly-developed description of
UK rainfall characteristics. The validity of kinematic
wave theory for flows in shallow drainage channels was
checked using measurements from a trial on the M6
motorway [Ref 5].
1.4 Guidance on the hydraulic design of triangular
and rectangular surface water channels was previously
given in HA 37/88, Hydraulic Design of Road-edge
Surface Water Channels (DMRB 4.2) in the form of a
series of design curves. The present document uses the
same hydraulic principles as before but with a new
rainfall formula that is optimized for storm durations
between two minutes and twenty minutes; this range
covers the great majority of applications. The design
method has been generalized so that it can additionally
now be applied to channels of trapezoidal cross-section.
The results are presented in the form of equations
instead of graphs so as to allow a simpler calculation
procedure that is also suitable for use in computer-based
design packages.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 2
Scope
2/1
2. SCOPE
2.1 This Advice Note describes a method of
determining the length of road between outlets that can
be drained by a given size of surface water channel
constructed along the edge of the road. The design
method can be used for channels of triangular,
rectangular or trapezoidal cross-section (see Figure 1),
and is based on application of Equation (13) in
paragraph 5.3.
2.2 Safety considerations limit the depth and types of
cross-sectional shape that may be used for surface water
channels which are not segregated from traffic.
Permitted dimensions and cross-falls given in MCHW
and HA 39 (DMRB 4.2) are summarized in Chapter 3.
2.3 In the design method, the longitudinal gradient of
a channel may either be constant or vary with distance.
In the latter case, the gradient can be zero at the
upstream or downstream end of the channel, but at all
intermediate points there must be a positive fall towards
the outlet. An approximate procedure is also given for
checking the performance of channels when they are
surcharged.
2.4 The design equations are based on assumptions
that the cross-sectional properties of a channel do not
vary with distance or depth of flow and that the width of
road drained between two adjacent outlets is constant.
If these assumptions are not satisfied, approximate
results may be obtained using average values of road
width or cross-sectional shape.
2.5 The design of outlets for use with surface water
channels is described in HA 78, Design of Outfalls for
Surface Water Channels (DMRB 4.2).
Volume 4 Section 2
Part 4 HA 37/97
August 1997
2/2
Chapter 2
Scope
Figure 1 Cross-sectional shapes of channels
c. Trapezoidal
1:40
B
b
b
2
1
1
b
1
y
b. Rectangular
y
B
b
1:40
a. Triangular
B
y
1
b
1
b
2
1
C
L
C
L
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 3
Safety Aspects
3/1
3. SAFETY ASPECTS
3.1 When considering the use of surface drainage
channels, in particular those of rectangular cross-
section, safety aspects relating to their location should
be taken into account. Triangular or trapezoidal
channels will usually be sited adjacent to the hardstrip
or hardshoulder or at the edge of the carriageway and
in front of the safety fence, where one is provided;
layout details are given in the A and B Series of the
Highway Construction Details (HCD) (MCHW 3). In
these locations the maximum design depth of flow
(dimension y
1
in Figure 2) should be limited to 150mm.
In both verges and central reserves the channel slopes
(1:b
1
and 1:b
2
in Figure 2) should not normally be
steeper than 1:5 for triangular channels and 1:4.5 for
trapezoidal channels; in very exceptional cases slopes
of 1:4 are allowable for both types of channel.
3.2 Rectangular channels, or triangular channels of
depth greater than 150mm, should be used only when
safety fencing is provided between the channel and the
carriageway; such channels can, therefore, normally
only be justified when safety fencing is warranted by
other considerations. In addition, these channels should
not be located in the zone behind the safety fence into
which the fence might reasonably be expected to
deflect on vehicle impact. Shallower channels as
described in paragraph 3.1 may lie in this deflection
zone, or be crossed by the safety fence (usually at a
narrow angle) provided that the combined layout
complies with the requirement of TD 19, Safety Fences
and Barriers (DMRB 2.2), TD 32, Wire Rope Safety
Fence (DMRB 2.2.3) and the HCD (MCHW 3).
Further advice on such layouts should be sought from
the Overseeing Department.
3.3 Co-ordination of the layout of safety fences and
surface water channels must be arranged at an early
stage in design and not left to compromise at later
stages. Where safety fences are not immediately
deemed necessary, sufficient space should be provided
in the verge or central reserve to allow for their
possible installation. The combined layout must
comply with the requirements of TD 19, TD 32 and the
HCD in terms of set-back and clearance dimensions
and the mounting height of the safety fence.
3.4 The geometric constraints given in this document
should also be applied to channel outfall details. In
outfall design and at other channel terminations, slopes
exceeding 1:4 should not be used on any faces,
particularly those orthogonal to the direction of traffic,
unless such faces are behind a safety fence.
V
o
l
u
m
e

4


S
e
c
t
i
o
n

2
P
a
r
t

4


H
A

3
7
/
9
7
A
u
g
u
s
t

1
9
9
7
3
/
2
C
h
a
p
t
e
r

3
S
a
f
e
t
y

A
s
p
e
c
t
s
F
i
g
u
r
e

2

C
r
o
s
s
-
s
e
c
t
i
o
n

o
f

s
u
r
c
h
a
r
g
e
d

c
h
a
n
n
e
l
C
L
Channel
b
2
1 1
b
1
1:40
B
b
y
3
y
1
y
2
b
3
1
Carriageway, hardstrip or hardshoulder
1:40
B
b
Note: Vertical scale exaggerated
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 4
Basis of Design Method
4/1
4. BASIS OF DESIGN METHOD
4.1 The design method is based on the results of
kinematic wave theory [Refs 3, 4 and 5] which is able to
provide a mathematical description of the time-varying
flow conditions that can occur during a storm in a
channel receiving inflow along its length. It is assumed
that the main factors determining the depth and velocity
of the water at any point are the rate of flow, the
longitudinal slope of the channel and its hydraulic
resistance. The effect of other factors associated with
the inertia of the water and the slope of the water
surface along the channel are assumed to be small in
comparison and are neglected. These assumptions are
normally reasonable in long drainage channels where
changes in velocity occur relatively slowly with time
and distance.
4.2 In using kinematic wave theory to determine time-
varying flow conditions in a collecting channel, no
account has been taken of any flow attenuation due to
water temporarily held on the road surface during a
storm. Water depths on roads are normally only a few
millimetres and any depression storage is likely to have
been filled by rainfall occurring before the most intense
part of the storm that constitutes the design event. It is
therefore assumed that the hydrograph of the lateral
inflow to the channel is equal to the product of the
instantaneous rainfall intensity and the effective width of
the catchment but with a time shift corresponding to the
time of travel of the runoff across the road surface.
4.3 The relationship between the flow rate and water
depth at any point along the channel is obtained from the
Manning resistance equation:
Q =
AR S
n
2 3 1 2
(1)
where Q is the discharge (m
3
/s), A is the cross-sectional
area of flow (m
2
), S is the longitudinal gradient of the
channel (m/m) and n is the Manning roughness
coefficient; the hydraulic radius R (m) is given by:
R = A/ P
(2)
where P is the wetted perimeter of the flow (m).
4.4 For a channel of trapezoidal cross-section with a
base width B
b
, side-slopes of 1: b
1
and 1: b
2
(vertical:
horizontal) and flow depth y (see Figure 1c), the
Manning resistance equation becomes:
(3)
where the effective cross-fall b of the channel is defined
as:
b = b + b
1 2
(4)
and the hydraulic-radius factor r (= flow width/wetted
perimeter) is given by:

( )
( ) ( )
[ ]
r =
B + b + b y
B + 1 + b + 1 + b y
b 1 2
b 1
2
2
2
1 2 1 2
(5)
For a wide shallow channel, r tends towards unity.
4.5 For the special case of a triangular channel, the
sole width B
b
=0 so that Equations (3) and (5) can be
written as:
Q =
bS
n

r y
32
1 2 2 8
1 3

1
]
1
(6)
( ) ( )
r =
b + b
1 + b + 1 + b
1 2
1
2
1 2
2
2
1 2
(7)
A triangular channel with one side vertical can be
catered for by putting b
1
= 0.
4.6 For the special case of a rectangular channel of
width B
b
, the side-slopes b
1
= b
2
=0 so that:
[ ]
Q = B S r y
b
1 2 2 5
1 3
(8)
(9)
( )
( )
Q =
S
n
2B + by
B + by
r y
1 2
b
b
2 5
5
2
1 3
32

1
]
1
1
r =
B
B + 2y
b
b
Volume 4 Section 2
Part 4 HA 37/97
August 1997 4/2
4.7 Kinematic wave theory enables the peak depth of
flow at the downstream end of a channel to be
determined for a given intensity and duration of rainfall.
For design, it is then necessary to identify the critical
storm duration which will produce the maximum
possible depth of flow for a given frequency of
occurrence. This has been done by using the following
equation:
( )
( )
I = 32.7 N- 0.4
T- 0.4
T
(2minM5)
o
0.223
0.565
(10)
where I
0
is the mean rainfall intensity (mm/h) occurring
in a storm of duration T (minutes) with a return period
of N (years), such that a storm of this intensity will
occur on average once every N years. The quantity
2minM5 is the depth of rainfall (in mm) occurring at the
particular geographical location in a storm with T= 2
minutes and N = 5 years. Equation (10) was obtained
for the purpose of this document by curve-fitting data on
UK short-period rainfalls given by the Meteorological
Office in British Standard BS 6367: 1983 [Ref 6].
Further information is given in Annex A.
4.8 The intensity of rainfall normally varies with time
during a storm, and this affects the way in which a
channel responds to run-off. Typical profiles of rainfall
intensity in summer and winter storms are given in
Volume II of the Flood Studies Report [Ref 7]. For
impervious surfaces such as roads, the largest flows tend
to be produced by heavy short-period storms which
occur more frequently in summer than winter. The
design method therefore uses what is termed the 50%
summer profile, in which the peak intensity at the mid-
point of the storm is approximately 3.9 times the
average intensity; 50% of summer storms in the UK can
be expected to have lower ratios of peak to mean
intensities. Data from the M6 motorway [Ref 5] showed
that using the 50% summer profile in the design method
gave conservative results.
4.9 As explained in paragraphs 4.7 and 4.8, the
critical storm duration for a channel depends on the
geometric properties of the channel and the catchment
that it drains and also upon the local rainfall
characteristics and the type of storm profile. However,
in the design equations given in Chapter 5, the critical
storm duration has been eliminated as an independent
variable and does not need to be determined separately
when calculating the length of road that a channel can
drain.
Chapter 4
Basis of Design Method
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 5
Description of Design Equations
5/1
5. DESCRIPTION OF DESIGN EQUATIONS
5.1 For economic reasons, it will normally be
appropriate to use only one or two sizes of channel for a
particular road scheme and to vary the spacing between
outlets in order to allow for the effect of changes in
channel slope or drained width of road. The length of
road that a channel can drain may be determined directly
by use of Equation (13) in paragraph 5.3. A second
type of design problem may sometimes occur if it is
required to find the size of channel needed to drain a
given length of road. As explained in paragraphs 5.4
and 5.5, a direct solution is possible for triangular
channels but iterative procedures are necessary for
rectangular and trapezoidal channels.
5.2 The flow capacity of a channel depends upon its
shape as well as its cross-sectional area. The shape
characteristics can be expressed in terms of a parameter
m defined as:
m =
By
A
- 1
(11)
where B and A are, respectively, the width of flow and
the cross-sectional area of flow corresponding to the
design depth of flow in the channel. For a triangular
profile m = 1, and for a rectangular profile m = 0;
trapezoidal channels will have values of m between
these two limits. The design equations given in the next
clause can also be applied to dished channels in which
the flow width and flow depth are related by an equation
of the form:
B = cy
m
(12)
where c is a constant for a particular type of channel.
5.3 The length of road L that can be drained by a
surface water channel is obtained from the general
equation:
(13)
where the factor G
m
for channel shape is:
( ) G = 2.90 x 10 2.65- m
m
6
(14)
and W
e
is the effective width of the catchment drained
by the channel (see Section 12). Equation (13) is
dimensional and the following units must be used: L, y
and W
e
in m; S in m/m; N in years; A in m
2
; and
2minM5 in mm.
5.4 In the second type of design problem described in
paragraph 5.1 it is required to find the size of channel
needed to drain a given length of road. For the case of a
triangular channel, Equations (13) and (14) can be used
to obtain a direct solution for the flow depth:
where the effective cross-fall b and the hydraulic-radius
factor r are given by Equations (4) and (7) respectively.
For the situation of a rectangular channel of width B
b
,
the corresponding result is:
In this case, the unknown flow depth y appears on both
sides of the equation so a small amount of iteration is
needed to find the solution. The units for the quantities
in Equations (15) and (16) must be as specified in
paragraph 5.3.
5.5 No general equation for directly determining the
flow depth y in trapezoidal channels can be obtained
because different solutions are possible depending upon
the particular values of base width and side-slope
chosen. The following design procedure is therefore
recommended.
(a) Make an initial estimate of the size and shape of
channel needed.
(b) Calculate the flow area A and the hydraulic-
radius factor r (from Equation (5)).
(c) Determine the values of m from Equation (11)
and G
m
from Equation (14).
( ) ( )
( )
L = G
S
n
ry N- 0.4
A
W 2minM5
m
2 3 -0.362
e
1.62
1 2

1
]
1
(15)
(16)
( )
( )
y = 2.60 x 10
nL
S
r
. N - 0.4
W 2minM5
b
-2
1/ 2
-0.171
e

_
,

1
]
1
0 256
0 093
0 415
.
.
.
( )
( )
y = 9.75 x 10
nL
S
1 +
2y
B

. N - 0.4
W 2minM5
B
-4
1/ 2
0.437
b
0.292
e
b
0.708

_
,

_
,

1
]
1
0 158 .
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 5
Description of Design Equations
5/2
(d) Calculate from Equation (13) the length L of road
that can be drained and compare with the required
length.
(e) Revise the channel geometry and repeat steps (a)
to (d) until the required drainage length is
achieved.
5.6 If it is necessary to determine the critical storm
duration that produces the design flow conditions in a
channel, a method of estimating the duration is given in
Annex A.
5.7 Information on how to calculate or choose
suitable values for the various parameters in Equations
(13), (15) and (16) are given in Chapters 6 to 14.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
6. STORM RETURN PERIOD
6.1 The degree of security against flooding that
should be provided by a surface drainage channel needs
to be decided on a case-by-case basis.
6.2 On critical lengths of road, it may be necessary to
design channels for storms of higher return period than
normal in order to reduce the risk of overflowing.
Examples are lengths in which a change in super-
elevation from one side of the road to the other causes
the cross-fall to be locally zero; flooding at such a point
might spread across the full width of the road. A higher
standard of design may also be appropriate for sections
draining to longitudinal sag points where it is important
to prevent ponding.
6.3 In the absence of special factors of the type
described in paragraph 6.2, it will normally be
appropriate for a channel in the verge to be designed to
just flow full for a storm with a return period of N = 1
year. The HCD (MCHW 3) and HA 39 (DMRB 4.2)
permit the outer edge of a verge channel (farthest from
the carriageway) to be set higher than the inner edge, as
shown in Figure 2. This allows some surcharging to
occur on to the adjacent hardstrip or hardshoulder during
rarer storm events. The maximum permissible widths of
flooding are 1.0m for all-purpose roads and 1.5m for
motorways. As an example, a 1.0m width of flooding on
a road with a cross-fall of 1:40 can be allowed by
constructing the outer edge of the channel so that it is
25mm higher than the inner edge. Larger flows will
cause the channel to overflow into the verge rather than
to encroach farther on to the carriageway. Unless there
are special factors, it will normally be appropriate to
check that a surcharged channel is able to cater for a
storm with a return period of N = 5 years without
overflowing.
6.4 Channels in the central reserve may need to be
designed to a higher standard than those in the verge
because it is important to prevent water from
encroaching on to the adjacent fast lane or from
overflowing on to the opposite carriageway. HA 39
(DMRB 4.2) specifies certain geometric requirements in
order to prevent such occurrences. If no surcharging of
a channel in the central reserve is permissible, it should
normally be designed to flow just full for a storm with a
return period of N = 5 years (in the absence of other
special factors such as those described in paragraph
6.2).
If the cross-sectional profile in the central reserve does
safely permit some surcharging, then the normal design
requirements should be similar to those for verge
channels: ie, N = 1 year for the channel just flowing
full; and N = 5 years for the channel with the permitted
amount of surcharging.
6.5 The standard of performance appropriate for
critical sections of road (see paragraph 6.2) will depend
upon the particular circumstances but design storm
return periods of N = 10 years or N = 20 years may be
suitable choices.
Chapter 6
Storm Return Period
6/1
Volume 4 Section 2
Part 4 HA 37/97
August 1997
7. GEOGRAPHICAL LOCATION
7.1 The effect of geographical location on rainfall
characteristics is taken into account in Equation (10) by
means of the value of 2minM5 - the rainfall depth (in
mm) occurring in a storm event with a duration of T = 2
minutes and a return period of N = 5 years. The value
of 2minM5 appropriate for any particular road scheme
in the UK is obtained from Figure 3 (taken from BS
6367: 1983, Ref 6). This value is then used in the
design Equations (13), (15) or (16) as appropriate.
7.2 It should be noted from Figure 3 that the most
severe rainfall conditions are to be expected in East
Anglia and the South-East of England. Although these
areas have much lower values of average annual rainfall
than parts of Wales, Scotland and the North-West of
England, they experience heavier and more frequent
short-duration storms, of the kind typically associated
with summer thunderstorms.
Chapter 7
Geographical Location
7/1
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 7
Geographical Location
7/2
Figure 3 Values of 2min M5 rainfall depth for UK
(Reproduced from BS 6367: 1983 by permission British Standard Institution)
Statute miles
Kilometres
National grid
Irish grid
mm
mm
mm
mm
Volume 4 Section 2
Part 4 HA 37/97
August 1997
8. CHANNEL GEOMETRY
8.1 The design equations (13), (15) and (16) are valid
for all types of triangular, rectangular or trapezoidal
channel provided the cross-sectional shape factors b
1
, b
2
and B
b
(as appropriate) do not vary with the depth of
flow or with distance along the drainage length.
8.2 Limits on the permissible depth and side-slopes
for surface water channels that may be subject to traffic
are given in paragraph 3.1. A trapezoidal channel
provides a higher flow capacity than a triangular channel
with the same depth and side-slopes and will therefore
enable longer lengths of road to be drained between
outlets. If a channel is protected or removed far enough
from traffic, larger depths or steeper side-slopes can be
used; the occurrence of the flow depth y in Equation
(13) shows that a relatively deeper channel is more
efficient hydraulically than a shallower channel of the
same cross-sectional area.
8.3 The base sections of trapezoidal or rectangular
channels should be given a 1:40 cross-fall away from
the carriageway (see Figure 1) so as to provide self-
cleansing characteristics that are similar to those of
conventional kerbed channels. The effect of the 1:40
cross-fall on flow capacity is very small and can be
neglected; the value of y used in Equations (11), (13)
and (16) should be the flow depth measured from the
centreline of the channel invert.
8.4 Figure 2 defines the flow depths that may need to
be considered when designing a surface water channel.
The depth y
1
from the lower inner edge of the channel to
the centreline of the invert corresponds to the channel
full capacity and is termed the design flow depth. If
water is permitted to encroach on to the adjacent
hardstrip or hardshoulder during rarer storms, the
surcharged flow depth y
3
is the depth from the outer
edge of the channel to the centreline of the invert. There
should, if possible, be no step between the inner edge of
the channel and the top edge of the carriageway.
However, if one is formed, it may be necessary in the
calculations to take account of the depth y
2
between the
top edge of the carriageway and the centreline of the
channel invert. In cases where porous asphalt surfacing
is used in conjunction with road-edge channels, a step
will normally be necessary to allow water to drain from
the permeable layer; the vertical distance between the
top of the porous asphalt layer and the invert of the
channel should not exceed 150mm.
8.5 As described in Chapter 6, surface water channels
should be designed to provide specified degrees of
security against flooding. As an example, a verge
channel might typically be sized so that the design flow
depth y
1
provides sufficient flow capacity for storms
with a return period of N = 1 year. It might also be
specified that the channel should be able to cater for
storms of N = 5 years without exceeding the surcharged
flow depth y
3
. If the second criterion was not satisfied,
the size of the channel would need to be increased or the
distance between outlets reduced. If no surcharging was
permitted (eg, for a channel in the central reserve), it
might be specified that the design flow depth y
1
should
provide sufficient capacity for storm return periods of N
= 5 years. Further guidance on the choice of suitable
design criteria is given in Chapter 6.
8.6 If a channel is permitted to surcharge, the cross-
sectional shape of the flow area becomes more complex
and does not comply with the assumptions given in
paragraph 8.1. An approximate method of determining
the drainage capacity of surcharged channels is given in
Chapter 13.
8.7 In some situations it may be wished to increase
the size of a channel part way along a drainage length.
In the case of a triangular profile, it is possible to deepen
the channel without altering the cross-falls b
1
and b
2
, and
similarly a rectangular channel can be deepened while
keeping the width B
b
constant; examples of suitable
transitions are shown in Figure 4. The design equations
(13), (15) and (16) can be applied to such cases without
approximation. The flow capacity of the smaller,
upstream channel should be checked to ensure that it is
sufficient to drain the length of road upstream of the
transition point; the capacity of the larger, downstream
channel should be similarly checked using the overall
length of road draining to the outlet.
8.8 If a trapezoidal channel is enlarged part way
along a drainage length, it is not possible to keep all the
shape factors b
1
, b
2
and B
b
constant. The capacities of
the two parts of the channel should be checked in the
way described in paragraph 8.7, but the result for the
downstream length may be less accurate because the
assumption of constant shape is not satisfied. For this
reason, it is recommended that the downstream channel
should be designed assuming a drainage length 5%
greater than the actual length draining to the outlet.
Chapter 8
Channel Geometry
8/1
Volume 4 Section 2
Part 4 HA 37/97
August 1997
8.9 Transitions of the type shown in Figure 4 should
be gradual in order to minimize energy losses. If the
invert is lowered, the length of the transition should not
be less than 15 times the change in depth. Similarly, the
side of a channel should not diverge outwards from the
longitudinal centreline at an angle greater than 1:3 in
plan.
8.10 Fixed obstructions in a drainage channel, such as
a longitudinal line of posts for a safety barrier, will
reduce its flow capacity. In order to prevent excessive
local disturbance of the flow, no more than 15% of the
cross-sectional area of the flow should be blocked by an
obstruction if it is located within the downstream half of
a drainage length; in the upstream half the blockage
should not be more than 25% of the cross-sectional area
of the flow. The energy losses produced by a
longitudinal line of posts can be taken into account by
using a higher value of the Manning resistance
coefficient n, as described in paragraph 11.2.
Allowance need not normally be made for one or two
isolated posts in a particular drainage length.
8/2
Chapter 8
Channel Geometry
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 8
Channel Geometry
8/3
Figure 4 Transitions for channel enlargements
b
1
1
b
2
1
b
1
1
1
b
2
a. Triangular
b. Rectangular
Volume 4 Section 2
Part 4 HA 37/97
August 1997
9. CHANNEL GRADIENT
9.1 The longitudinal gradient S of a channel is
defined as the vertical fall per unit distance measured
along the channel. A channel will normally have the
same longitudinal gradient as the adjacent carriageway,
but this is not a condition for use of the design method.
9.2 If the gradient of a channel varies with distance,
an equivalent value of uniform slope S
e
can be
calculated for use in the design method described in
Chapter 5. To evaluate S
e
, values of the local gradient S
j
should be determined at eleven equally-spaced points (j
= 1 to 11) along the length L of channel draining to a
particular outlet: S
1
is the gradient at the upstream end
and S
11
the gradient at the outlet. The distance between
adjacent points is not limited to any particular value, but
should be equal to L/10. The equivalent value of
gradient is calculated from:
(17)
9.3 The design method can be used if the
longitudinally-varying gradient is locally zero (but not
adverse) at the upstream or downstream end of a
channel. In order to obtain a valid result for the
equivalent gradient S
e
, the zero value S
1
or S
11
should be
replaced in Equation (17) by:
S = S / 9
1 2
(18a)
or
S = S / 9
11 10
(18b)
9.4 The design method is not valid if the gradient
becomes zero at an intermediate point between the
upstream and downstream ends of a length of channel.
In such cases, it is necessary to place an outlet at the
intermediate point and design the channel as two
separate lengths.
Chapter 9
Channel Gradient
9/1
S = 400 S + S S
e 1
-
1
2
11
-
1
2
j=2
j=10
j
-
1
2
-2
+

1
]
1
2
Volume 4 Section 2
Part 4 HA 37/97
August 1997
10. DRAINAGE LENGTHS
10.1 The drainage length L is the distance along a
channel between two adjacent outlets on a continuous
slope, or the distance between a point of zero slope and
the downstream outlet.
10.2 Surface water channels can be used most
effectively if their layout is considered at an early stage
in the design of a road scheme. Where possible,
horizontal and vertical alignments should be chosen so
that suitable drainage lengths can be defined taking into
account the location of outlets discharging to natural
watercourses.
10.3 If the drainage length to a natural watercourse
requires too large a channel capacity, intermediate
outlets should be used to remove water from the road-
edge channel. The flow from the outlets may be
conveyed to a suitable discharge point by carrier pipes
or an open ditch.
10/1
Chapter 10
Drainage Lengths
Volume 4 Section 2
Part 4 HA 37/97
August 1997
11. CHANNEL ROUGHNESS
11.1 The hydraulic resistance of a road-edge channel
depends upon its surface texture, the standard of
construction, and the presence of deposited silt and grit.
Recommended values of the Manning roughness
coefficient for use in conjunction with the design
equations are given in Table 1.
Further information about the factors influencing the
hydraulic resistance is given in Annex B.
11.2 Posts located in a surface water channel will
disturb the flow of water and produce additional energy
losses. The effect of a line of posts on flow capacity can
be estimated by increasing the appropriate value of
roughness coefficient from paragraph 11.1 by an amount
n
p
given by:
n = 0.7
I
gL

A
A

ry
m+1
p
p
p
1/ 2
2/ 3

_
,

1
]
1
1

_
,

(19)
where g is the acceleration due to gravity (=9.81 m/s
2
),
L
p
is the average distance between successive posts (in
m) and A
p
is the wetted cross-sectional area (in m
2
) of a
post normal to the flow, when the cross-sectional area of
the flow is A (in m
2
). The derivation of Equation (19) is
explained in Annex B.
Channel Type Condition n
Concrete average 0.013
Concrete poor 0.016
Black top average 0.017
Black top poor 0.021
Table 1 Values of Mannings n
11/1
Chapter 11
Channel Roughness
Volume 4 Section 2
Part 4 HA 37/97
August 1997
12. CATCHMENT WIDTH
Soil Type
High permeability low 0.07
medium 0.11
high 0.13
Medium permeability low 0.11
medium 0.16
high 0.20
Low permeability low 0.14
medium 0.21
high 0.26
Table 2 Run-off coefficients for cuttings
12/1
Chapter 12
Catchment Width
12.1 The effective catchment width W
e
is equal to the
width of the carriageway drained, plus the width of the
channel and any other impermeable surface draining to
the channel, plus an allowance (if appropriate) for run-
off from a cutting.
12.2 The design method assumes that W
e
does not vary
along a particular drainage length. However, minor
local variations can be allowed for by using an average
width, calculated by dividing the total effective area
draining to an outlet by the drainage length L.
12.3 It is assumed that 100% run-off occurs from
concrete and black-top surfaces during design storms.
12.4 Surface channels may be used to collect run-off
from cuttings as well as from roads. The amount of run-
off from a cutting depends upon many factors including
its height, slope, soil type and antecedent wetness, and
also upon the quantity of rainfall and the direction of the
wind.
12.5 Information on the additional run-off contributed
by cuttings is very limited. In the absence of suitable
field data, the effective width W
e
of a catchment should
be calculated from the formula:
W = W + C
e
_

(20)
where W is the width of the impermeable part of the
catchment (m), is the run-off coefficient for the
cutting, and is the average plan width of the cutting (m)
drained by the length of channel being considered.
Details of the derivation of Equation (20) are given in
Annex C.
12.6 Suitable values of the run-off coefficient can be
estimated from Table 2. These figures contain some
allowance for the relative steepness of road cuttings,
which may result in more run-off than from equivalent
natural catchments. For design purposes it is assumed
that the antecedent wetness of the cutting is dependent
upon the average annual rainfall at the site. Appropriate
choices of antecedent wetness for Northern Ireland,
Scotland, Wales and English Counties are given in Table
3.
Antecedent
Wetness
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Chapter 12
Catchment Width
12/2
Low
Bedfordshire
Buckinghamshire
Cambridgeshire
Essex
Greater London
Hertfordshire
Norfolk
Rutland
Suffolk
Medium
Berkshire
Cleveland
Derbyshire
Durham
East Sussex
Hampshire
Hereford &
Worcester
Humberside
Isle of Wight
Kent
Leicestershire
Lincolnshire
North Yorkshire
Northamptonshire
Northumberland
Nottinghamshire
Oxfordshire
Shropshire
South Yorkshire
Staffordshire
Surrey
Tyne & Wear
Warwickshire
West Sussex
West Yorkshire
High
Northern Ireland
Scotland
Wales
Avon
Cheshire
Cornwall
Cumbria
Devon
Dorset
Gloucestershire
Greater Manchester
Lancashire
Merseyside
Somerset
Wiltshire
Table 3 Antecedent wetness categories
The basis of the data in Tables 2 and 3 is explained in Annex C.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
13. SURCHARGED CHANNELS
Also required is the hydraulic conveyance factor K for
the surcharged channel which is given by:
(23)
where n
c
is the Manning roughness coefficient of the
carriageway.
13.4 The next step is to determine the shape factor m,
see Equation (11), for an equivalent channel having the
same values of y
3
, A and K as the surcharged channel.
This is done by calculating the value of the quantity:
X =
K
y A
3
2/ 3 (24)
from which m can be determined using:

m =
1
2
X-1 + X +
14
3
X + 1
2
1/ 2

_
,

1
]
1 (25)
Values of m for the equivalent channels may sometimes
be slightly greater than unity but this does not affect the
design procedure.
13.5 Having determined values of y
3
, A, r and m for
the equivalent channel, the length of road that may be
drained is calculated using the equations given in
paragraph 5.3.
13/1
Chapter 13
Surcharged Channels
13.1 As described in Chapters 6 and 8, some
surcharging of surface drainage channels may often be
permissible. The type of channel geometry shown in
Figure 2 allows storms of higher return period to be
accommodated without overflowing and without causing
water to encroach beyond the edge of the hardstrip or
hardshoulder.
13.2 Surcharged channels have compound cross-
sectional shapes and do not comply with all the
assumptions given in paragraph 8.1 for the validity of
the design equations in Chapter 5. However, an
approximate estimate of the length of road that may be
drained by a surcharged channel can be obtained by
defining an equivalent channel that has the same values
of overall depth, cross-sectional area and flow capacity.
The following procedure refers to the general case of the
trapezoidal channel shown in Figure 2. However, the
equations can equally be applied to the case of a
triangular channel (with B
b
= 0 ) or of a rectangular
channel (withb
1
=0 = b
2
).
13.3 First calculate the cross-sectional area A of the
surcharged channel which is given by:
(21)
where b
3
is the transverse slope of the carriageway
adjacent to the channel. As explained in paragraph 8.4,
it is preferable if channels can be constructed without a
step between the inner edge of the channel and the top
edge of the carriageway; in such cases y
2
=y
1
in
Equation (21). The depths y
1
, y
2
and y
3
are all measured
relative to the centreline of the invert, see paragraph 8.4.
The corresponding hydraulic-radius factor r (= surface
width/wetted perimeter) is determined from:
( ) ( )
( )
A =
1
2
b + b y - b y - y
+ b y - y + 2B y
1 2 3
2
2 3 1
2
3 3 2
2
b 3

1
]
1
1
( ) ( )
( )
K
3
8

b + b y - b y - y
+
n
n
b y - y +
8
3
B y
1 2 3
8 3
2 3 2
8 3
c
3 3 2
8 3
b 3
5 3

_
,

1
]
1
1
1
( ) ( )
( ) ( ) ( ) ( )
( )
r =
b y + b y + b y - y + B + y - y
b + 1 y + b 1 y + b + 1 y - y + B + y - y
1 3 2 1 3 3 2 b 2 1
1
2
1 2
3 2
2
1 2
1 3
2
1 2
3 2 b 2
1
+
(22)
Volume 4 Section 2
Part 4 HA 37/97
August 1997
14. BY-PASSING AT OUTLETS
14.1 Recommendations on the types and sizes of outlet
to be used with surface water channels are contained in
HA 78 (DMBR 4.2). The flow rate approaching an
outlet is given by Equation (3), (6) or (8) depending
upon the particular cross-sectional shape of the channel.
If the longitudinal gradient of the channel varies with
distance, the effective slope S
e
(see paragraph 9.2)
should be substituted for S in these equations.
14.2 A more economic overall design of drainage
system may often be achieved by allowing a certain
degree of flow by-passing to occur at intermediate
outlets; HA 78 specifies that collection efficiencies
between 80% and 100% can be used in such cases. The
design equations given in Chapter 5 assume that there is
no inflow at the upstream end of a drainage length.
Further analysis of systems with and without by-passing
indicates that the additional inflow resulting from up to
20% by-passing will not normally produce any increase
in the peak design flow occurring at the downstream end
of a drainage length. This is because the by-pass flow
does not arrive at the next outlet until later in the storm
when the rainfall intensity has already started to
decrease. By-passing will therefore produce two peak
flows at the downstream end of a drainage length but the
second peak will normally be smaller than the first.
14.3 Although the general conclusion is that by-
passing will not normally increase peak design flows in
drainage channels, there may be some special cases
where its effect could be significant. One possible
example is by-passing from a long channel into a much
shorter one, thus producing an inflow that is large
compared with the amount of water collected by the
downstream channel. To provide a simple rule that
should be safe for almost all cases, it is recommended
that the following allowance should be made when
designing any channel subject to by-passing. If the
upstream and downstream channels have lengths of L
1
and L
2
respectively, the downstream channel should be
designed according to the recommendations in Chapter 5
so that it is able to drain a length:
14/1
Chapter 14
By-Passing at Outlets
( ) L = L +
1
2
L
2 1
1
(26)
where is the collection efficiency of the outlet draining
the upstream channel; note that an efficiency of 80%
corresponds to a value of = 0.80 in Equation (26).
Volume 4 Section 2
Part 4 HA 37/97
August 1997
15. CONSTRUCTION TOLERANCES
15.1 The effect of allowable construction tolerances on
the capacity of a channel should be considered in design.
As an approximate guide, flow capacity will be within
5% of the design capacity if the local channel gradient
is within 10% or the depth of the channel within 2%
of the required values. In general, the design capacity of
the channel should be determined for the combination of
tolerances that give minimum channel capacity.
Adherence to tolerances is most important at the
downstream ends of drainage lengths where the flows
are greatest.
15/1
Chapter 15
Construction Tolerances
Volume 4 Section 2
Part 4 HA 37/97
August 1997
16. WORKED EXAMPLES
16.1 Determine the length of road that can be drained
by a triangular surface water channel having the
following characteristics.
Symmetrical cross-falls: b
1
= b
2
= 5
Channel depths (see Fig 2): y
1
= 0.120m
y
3
= 0.145m
Longitudinal channel gradient:
S = 1/200 = 0.005
Mannings roughness
coefficient: n = 0.013
(concrete, average condition)
The value of the hydraulic-radius factor corresponding
to the design flow depth y
1
is calculated from Equation
(7) to be:
( )
r =
5 + 5
2 1 + 5
= 0.981
2
1/ 2
The width of flow corresponding to the design flow
depth y
1
is:
B = 10 x 0.120 = 1.200m
and the corresponding flow area is:
A = B y
1
= x 1.200 x 0.120 = 0.0720 m
2
The shape parameter m of the channel (see Equation
(11) and paragraph 5.2) has a value of:
m =
1.20 x 0.120
-1 = 1.00
0 0720 .
The width of the two-lane carriageway drained by the
channel is 9.300m (including two 1.000m wide
hardstrips). The overall width of the channel itself is:
B = b
1
y
3
+ b
2
y
1
= 5 x 0.145 + 5 x 0.120 = 1.325m
The road is on embankment and there is no run-off from
the verge into the channel. The effective catchment
width is therefore:
W
e
= 9.300 + 1.325 = 10.625m
The road is located near Coventry and from Figure 3 it
is found that the characteristic value of rainfall depth
occurring in 2 minutes with a return period of 5 years is:
2minM5 = 4.0mm
The channel is to be designed so that the design flow
depth y
1
is not exceeded by run-off from storms
occurring once every year on average; the design return
period is therefore:
N = 1 year
The length of road that can be drained by the channel is
calculated from Equation (13), in which the factor G
m
corresponding to the triangular shape of the channel is
obtained from Equation (14) as:
G
m
= 2.90 x 10
6
(2.65 - 1.00) = 4.79 x 10
6
The maximum drainage length is therefore:
The critical storm duration corresponding to the design
flow condition can be estimated from Equation (A.1) in
Annex A as:
( ) T = 0.085
0.013 x 244
0.981 x 0.120 = 15.9 minutes
c
-2/ 3
0 005
1 2
.
/

1
]
1
16.2 Consider the example in paragraph 16.1 for the
same road and channel but constructed in a cutting
which contributes run-off to the channel. The road is
located in Warwickshire so from Table 3 the antecedent
wetness is medium. The soil in the cutting is a fairly
heavy clay with a low permeability so from Table 2 the
run-off coefficient is = 0.21. The average width of the
cutting draining to the channel is = 15.0m. Compared
to the example in 16.1, the effective catchment width is
increased to the following value given by Equation (20):
W
e
= 10.625 + 0.21 x 15.0 = 13.775m
16/1
Chapter 16
Worked Examples
( )
( )
( )
L = 4.79 x 10
0.005
0.981 x 0.120
. 1.0 - 0.4
0.0720
10.625 x 4.0
= 244m
6
-0.362
1.62
1 2
2 3
0013 .

1
]
1
C
Volume 4 Section 2
Part 4 HA 37/97
August 1997
The width of four-lane motorway drained by the channel
is 17.900m (including a 3.300m hardshoulder). The road
is on embankment and there is no run-off from the verge
into the channel. The effective catchment width is
therefore:
W
e
= 17.900 + 1.925 = 19.825m
The section of motorway is located near Watford and
from Figure 3 the characteristic rainfall depth is:
2minM5 = 4.1mm
The channel is to be designed so that the design flow
depth y
1
is not exceeded by run-off from storms with a
return period of N = 1 year. The shape factor G
m
for the
channel is obtained from Equation (14) as:
G
m
= 2.90 x 10
6
(2.65 - 0.71) = 5.63 x 10
6
The maximum length of road that can be drained by the
channel is calculated using Equation (13):
16. 4 Along part of the motorway considered in
paragraph 16.3, superelevation causes run-off to drain
from one of the carriageways towards the central
reserve. Both sides of the central reserve are protected
by safety barriers so it is possible to use a channel of
rectangular cross-section in this location. The width of
the channel is chosen to be B
b
= 1.000m and it is
required to determine the design depth of flow when
draining a maximum distance of L = 300m. The
effective catchment width (carriageway + hardshoulder
+ channel) is:
W
e
= 17.900 + 1.000 = 18.900m
The values of n, S and 2minM5 are as given in
paragraph 16.3. Since the channel is not permitted to
surcharge on to the carriageway that it drains, it is
decided to determine the design depth of flow y of the
channel for storms with a return period of N = 5 years
(see paragraph 6.4). The value of y for a rectangular
channel is determined from Equation (16) but it should
be noted that y also appears on the right-hand side of the
equation. A short iterative procedure is therefore
necessary as illustrated by the following calculations.
Chapter 16
Worked Examples
16/2
All the other parameters in Equation (13) are unchanged
so the revised length of road that can be drained by the
channel is:
L = 244
10.625
13.775
= 160m
1.62

_
,

16.3 Determine the length of road that can be drained

by a trapezoidal channel having the following
characteristics.
Base width of channel: B
b
= 0.300m
Symmetrical cross-falls: b
1
= b
2
= 5
Channel depths (see Fig 2): y
1
= 0.150m
y
3
= 0.175m
Longitudinal channel gradient:
S = 1/200 = 0.005
Mannings roughness
coefficient: n = 0.013
(concrete, average condition)
The value of the hydraulic-radius factor given by
Equation (5) for the design flow depth y
1
is:
( )
( )
r =
0.300 + 5 + 5 x 0.150
0.300 + 2 1 + 5 x 0.150
0.984
2
1
2
The width of flow corresponding to the design flow
depth y
1
is:
B = 0.300 + 10 x 0.150 = 1.800m
and the corresponding flow area is:
The shape parameter of the channel has from Equation
(11) the value:
m =
1.800 x 0.150
0.158
- 1 = 0.71
( )
( )
A = B y +
1
2
b b y
= 0.300 x 0.150 +
1
2
5 + 5 x 0.150 = 0.158m
b 1 1 2 1
2
2 2
+
( )
( )
( )
L = 5.63 x 10
0.005
x 0.150
. 1.0 - 0.4
0.158
19.825 x 4.1
= 417m
6
-0.362
1.62
1
2
2
3
0 013
0 984
.
.

1
]
1
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Guess a likely value for the design flow depth, eg y =
0.150m, and substitute this on the right-hand side of
Equation (16) so that:
Substituting this calculated value of y on the right-hand
side of the equation gives a revised value of y = 0.169m;
one final iteration converges to the solution y = 0.170m,
which is the required design depth of flow in the 1.0m
wide rectangular channel.
16/3
Chapter 16
Worked Examples
m
( )
y = 9.75 x 10
0.013 x 300
0.005
1 +
2 x 0.150
. 5 - 0.4
18.900
x 4.1 = 0.168m
-4
1
2
0.437
0.292
0.158
0.708

_
,

_
,

1
]
1
1000
1000
.
.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
17. GLOSSARY OF SYMBOLS
17/1
Chapter 17
Glossary of Symbols
C
Units
A Cross-sectional area of flow m
2
A
p
Cross-sectional area of post m
2
B Surface width of flow m
B
b
Base width of channel m
b Effective cross-fall of channel (rate of increase of surface width
per unit depth)
b
1
Slope of side of channel remote from carriageway -
(1 unit vertical: b
1
units horizontal) -
b
2
Slope of side of channel adjacent to carriageway -
(1 unit vertical: b
2
units horizontal) -
b
3
Transverse slope of carriageway adjacent to channel
(1 unit vertical: b
3
units horizontal) -
Average plan width of cutting drained by channel m
c Coefficient in Equation (12) variable
C
d
Drag coefficient of post -
G
m
Coefficient defined by Equation (14) -
g Acceleration due to gravity m/s
2
I
o
Mean rainfall intensity mm/h
j Integer indicating section number -
K Hydraulic conveyance factor, Equation (23) m
8/3
L Drainage length of channel; distance between outlets m
L
p
Average distance between posts m
L
1
Value of L for upstream channel m
L
2
Value of L for downstream channel m
m Shape factor of channel defined by Equation (11) -
N Return period of storm years
n Manning roughness coefficient of channel s/m
1/3
n
c
Manning roughness coefficient of carriageway s/m
1/3
n
p
Additional roughness coefficient due to posts s/m
1/3
P Wetted perimeter of channel m
PIMP Percentage impermeable area of catchment %
PR Percentage run-off coefficient %
Q Flow rate at downstream end of channel m
3
/s
R Hydraulic radius of flow (= A/P) m
r Hydraulic-radius factor ( = B/P) -
S Longitudinal gradient of channel (vertical fall per unit m/m
distance along channel)
S
e
Effective value of S for channel with non-uniform slope m/m
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Units
S
j
Local value of S m/m
SOIL Soil index -
T Duration of storm minutes
T
c
Critical storm duration for channel minutes
UCWI Urban Catchment Wetness Index -
W Width of impermeable part of catchment m
W
e
Effective width of whole catchment m
X Factor defined by Equation (24) -
y Design depth of flow m
y
1
Depth of channel from lower edge of carriageway m
to centreline of invert (see Figure 2)
y
2
Depth of channel from top edge of carriageway m
to centreline of invert (see Figure 2)
y
3
Overall depth of surcharged channel to centreline of m
invert (see Figure 2)
Run-off coefficient for cutting -
Collection efficiency of outlet (= flow rate collected by -
outlet / flow rate approaching outlet)
2minM5 Rainfall depth occurring in 2 minutes with return mm
period of 5 years
Chapter 17
Glossary of Symbols
17/2
Volume 4 Section 2
Part 4 HA 37/97
August 1997
18. REFERENCES
1. Design Manual for Roads and Bridges (DMRB) (HMSO)
HA 37 Hydraulic Design of Road-edge Surface Water Channels (DMRB 4.2)
HA 39 Edge of Pavement Details (DMRB 4.2)
HA 78 Design of Outfalls for Surface Water Channels (DMRB 4.2)
TD 19 Safety Fences and Barriers (DMRB 2.2).
TD 32 Wire Rope Safety Fence (DMRB 2.2.3).
2. Manual of Contract Documents for Highway Works (MCHW) (HMSO)
Specification for Highway Works (MCHW 1)
Notes for Guidance on the Specification for Highway Works (MCHW 2)
Highway Construction Details (MCHW 3).
3. HR WALLINGFORD. Design of Highway Drainage Channels: Preliminary Analysis. Report DE 30,
1976.
4. MAY R W P. Design of Highway Drainage Channels. OECD Symposium on Road Drainage, Berne,
1978, pp 450-459.
5. HR WALLINGFORD. Motorway Drainage Trial on the M6 Motorway, Warwickshire. TRRL
Contractor Report 8, 1985.
6. BRITISH STANDARDS INSTITUTION. BS 6367 : 1983, Code of Practice for : Drainage of Roofs
and Paved Areas.
7. NATURAL ENVIRONMENTAL RESEARCH COUNCIL. Flood Studies Report, NERC, 1975.
8. IZZARD C F. Hydraulics of Runoff from Developed Surfaces. Highway Research Board (USA),
Proceedings, 26, 1946, pp129 - 150.
9. SWINNERTON C J. Hydrological Design for Motorway Stormwater Drainage Systems. PhD thesis,
Imperial College of Science and Technology, University of London, 1971.
10. HR WALLINGFORD. The Wallingford Procedure: Design and Analysis of Urban Storm Drainage.
1981.
Chapter 18
References
18/1
Volume 4 Section 2
Part 4 HA 37/97
August 1997 19/1
19. ENQUIRIES
The Divisional Director
Traffic Safety and Environment
St Christopher House
Southwark Street G CLARKE
London SE1 0TE Divisional Director
The Deputy Chief Engineer
The Scottish Office Development Department
National Roads Directorate
Victoria Quay J HOWISON
Edinburgh EH6 6QQ Deputy Chief Engineer
The Head of Roads Engineering (Construction) Division
Welsh Office
Y Swyddfa Gymreig
Crown Buildings
Cathays Park B H HAWKER
Cardiff CF1 3NQ Head of Roads Engineering
(Construction) Division
The Assistant Technical Director
Department of the Environment for
Northern Ireland
Roads Service
Clarence Court D O'HAGAN
10-18 Adelaide Street Assistant Technical Director
Chapter 19
Enquiries
Approval of this document for publication is given by the undersigned:
All technical enquiries or comments on this document should be sent in writing as appropriate to the above.
Belfast BT2 8GB
Volume 4 Section 2
Part 4 HA 37/97
August 1997
Annex A
Rainfall Data
A/1
ANNEX A RAINFALL DATA
A.1 The Meteorological office is able to provide individual data on rainfall characteristics for any location in the
UK. However, in order to produce a general design method for surface water channels, it is necessary to be able to
describe the rainfall characteristics by means of an equation relating mean intensity to the duration and return period
of the storm event.
A.2 Relevant information on short-period storms with durations between 1 minute and 10 minutes was provided by
the Met Office for use in British Standard BS 6367: 1983 [Ref 6]. The following general calculation procedure is
given in Annex A of BS 6367:
(a) Determine from a map of the UK (reproduced in this document as Figure 3) the value of rainfall depth
(2minM5) occurring in 2 minutes with a return period of 5 years at the chosen location.
(b) Calculate (using a table and the value of 2minM5) the rainfall depth for the required duration T but still with a
return period of 5 years (ie TminM5)
(c) Calculate (using a graph and the value of TminM5) the rainfall depth for the required duration T and return
period N (ie TminMN).
(d) Divide the rainfall depth TminMN by the duration to give the mean intensity I
0
.
A.3 Equation (10) for predicting I
0
was obtained for this document by curve-fitting the tabular and graphical data
corresponding to steps (b) to (d) above. The applicability of the equation to storm durations greater than 10 minutes
was checked using data in Volume II of the Flood Studies Report [Ref 7]. The equation is optimized for values of T
= 2-20 minutes and N = 1-20 years and tends to overestimate I
0
(ie err on the safe side) outside these ranges. The
recommended upper limits for use of Equation (10) are T = 30 minutes and N = 50 years.
A.4 The critical storm duration T
c
that gives rise to the design flow conditions in a particular channel can be
estimated from:
( ) T = 0.085
nL
S
ry
c 1
2
-
2
3

_
,

(A.1)
where T
c
is in minutes and the units for the other quantities are as defined in paragraph 5.3.
Volume 4 Section 2
Part 4 HA 37/97
August 1997
ANNEX B CHANNEL ROUGHNESS
B.1 The Manning resistance equation is appropriate when a flow is rough-turbulent (ie, with its resistance mainly
determined by the surface texture of the channel). This is likely to be the case in most road drainage channels, except
perhaps near the upstream end where the velocity or depth of flow is small and the flow may be smooth-turbulent (ie,
with its resistance mainly determined by the viscosity of the water). It is often assumed that the roughness coefficient
n depends only upon the surface texture of the channel, but experimental evidence indicates that it can vary with the
relative depth of flow, the cross-sectional shape of the channel and the intensity of any lateral inflow.
B.2 A modified version of Mannings equation for shallow triangular channels was developed by Izzard [Ref 8]
and is recommended by the US Federal Highway Administration. The equation has the form:
Q =
3
8

bS
n
y
1
2
8
3
(B.1)
and is obtained by applying Mannings equation to vertical elements in the cross-section and integrating the
discharge across the channel; no allowance is made for the resistance of a vertical kerb. Equation (B.1) gives
capacities that are approximately 20% higher than the conventional version in Equation (6). Values of n quoted in
the literature for triangular channels therefore depend upon which of the two formulae were used to analyse the
experimental data.
B.3 The conventional form of Mannings equation has been used in this document (ie, Equations (1), (3), (6) and
(8)) so as to provide a common basis for all shapes of channel. Values of n given in the literature vary typically from
0.010 to 0.017 for concrete channels and from 0.012 to 0.022 for asphalt channels. In Table 1 the recommended
figures for average condition correspond approximately to the mean of the published values; the figures for poor
condition are slightly less than the corresponding maximum values.
B.4 Factors which will tend to increase the resistance of a channel are lateral inflow from the road surface and the
presence of silt and grit in the invert. Data on these effects are not available, but it is probable that at the
downstream ends of channels (which are the most critical points) they are not very large in relation to the
uncertainties in the basic n values. An average value of n = 0.013 for a concrete channel would be appropriate if it
has a trowel-type finish, no sharp discontinuities in line or elevation, and is regularly cleaned.
B.5 An approximate procedure is given in Chapter 13 for applying the design method for simple cross-sectional
shapes to the case of surcharged compound channels. It was found that a direct solution for compound channels could
be obtained only when using a development of the modified Mannings equation in Equation (B.1). The relationships
in Equations (21) to (25), between the relevant hydraulic characteristics (flow capacity and storage capacity) for a
compound channel and an equivalent simple channel, are therefore based on the modified form of Mannings
equation.
B.6 Equation (19) in paragraph 11.2 for estimating the extra flow resistance produced by a longitudinal line of
posts is obtained by considering the drag force acting on each post. The factor of 0.7 in the equation corresponds to a
drag coefficient of C
d
= 1.2 with an allowance for the effect of varying water depth between the upstream and
downstream ends of a channel.
Annex B
Channel Roughness
B/1
Volume 4 Section 2
Part 4 HA 37/97
August 1997
ANNEX C RUN-OFF FROM CUTTINGS
C.1 Measurements of run-off from lengths of motorway in cutting were made in the UK by the Transport
Research Laboratory and reported by Swinnerton [Ref 9]. The published data do not enable the relative
contributions of the pervious and impervious areas to be assessed.
C.2 In the absence of other suitable data, it was decided to use an existing run-off equation to provide a method
for estimating run-off from road cuttings. The selected equation was developed during studies for the Wallingford
Procedure [Volume 1 of Ref 10] as a means of predicting surface run-off to urban storm sewers in the UK, and has
the form:
PR = 0.662 PIMP + 0.00219 (100 - PIMP). SOIL. UCWI (C.1)
where PR is the percentage run-off from the whole catchment, PIMP is the percentage of impervious area, SOIL is a
number related to the infiltration potential of the soil and UCWI is an index of the urban catchment wetness. [A
different run-off formula was finally adopted for the Wallingford Procedure, but Equation (C.1) is more suitable for
application to roads in cuttings].
C.3 The effective width W
e
of a road in cutting is defined as the equivalent width of road which will produce the
same total amount of run-off as a road of width W and a cutting of average width C. From Equation (C.1) it can be
shown that:
(C.2)
C.4 In the Flood Studies Report [Ref 7], soils in the UK were classified according to their infiltration potential
into five classes and assigned values of the SOIL parameter between 0.15 and 0.5; the lowest value would apply to a
well-drained sandy soil and the highest to a rocky soil on a fairly steep slope. Maps showing regional distributions
of SOIL classes are given in Volume 5 of Ref 7 and Volume 3 of Ref 10; the type of soil in a cutting should,
however, be assessed from a site survey since the maps do not identify small local variations. Cuttings for roads are
steeper than most natural catchments, and for a given soil type may produce relatively more run-off. Approximate
allowance can be made for this by classifying a soil in a category of lower infiltration potential than normal; values
of the SOIL parameter for cuttings might therefore be expected to be in the range 0.3 to 0.6.
C5 Design values of UCWI for use with Equation (C.1) are given in Volume 1 of Ref 10 as a function of the
standard average annual rainfall at a site. Representative values for UK regions were obtained by finding the
maximum range of UCWI within each region (excluding only peaks in highland areas), and adopting a figure one
third of this range below the maximum value. The regions were then grouped into categories of high, medium and
low catchment wetness, and an average representative value of UCWI calculated for each group.
C.6 The values of the run-off coefficient in Table 2 were obtained from Equation (C.2) using figures of SOIL =
0.3, 0.45 and 0.6 for soils of high, medium and low permeability, and representative values of UCWI = 71, 107, and
132 for low, medium, and high categories of antecedent wetness.
C.7 Site data should be collected where possible to improve the estimate of run-off from the cuttings.
Annex C
Run-off from Cuttings
C/1
W = W +
SOIL. UCWI
C
e
300

_
,