Chanson states that in a flow motion the longitudinal flow profile is controlled by the downstream flow conditions: for example, an obstacle, a structure, or a change of cross-section [10] . Any downstream control structure can result in a backwater effect. Figure 4 shows how an obstacle/structure can affect the natural profile. The constriction of the flow results in localized areas of turbulence and flow separation which reduces the volume of throughput at the structure. Thus, to maintain the continuity of flow, there must be an acceleration of the fluid through the structure.

Manning’s n is dependent on the open channel type, the size of the materials that make the bed and banks of the channel and the shape of the channel [12] . Different Manning’s n values have an impact on the flow of a river in various ways and allowance for variation in the various values must be considered within Manning’s assessment. To compute an overall Manning’s n requires consideration of several aspects of an open channel the base, surface irregularities, variations in shape and size of the channel cross-section, obstructions and vegetation along with a correction factor for the meandering of the channel. Compound channels where there is a main channel for normal flow with flood plains on one or both sides of the main channel also require computing of an overall n value.

The Oxford Dictionary defines obstruction as the action of trying to prevent something/somebody from making progress [13] . In open channel flow, obstructions can be logs, debris or bridges/culverts. Lujan et Peck [12] state that obstructions disturb the flow pattern in the channel. Figure 5 shows how obstructions affect

the flows in an open channel. When the flow hits the obstruction, the flows rise before the obstruction thereby creating a backwater curve and after the obstruction a hydraulic jump forms due to the changes in the flows regime as shown in Figure 6.

The Engineering Library states that head loss is the reduction in the total head of a fluid that moves through a fluid system [15] . CIRIA C786 gives the head loss through a culvert as the difference between the headwater elevation and the tailwater elevation [16] . Head loss is caused by friction within the open channel or culvert. It is important that head loss is at equilibrium throughout the flow of a channel/pipe. Cobbe and McDermott when discussing pipe work reports that larger diameter pipes can deliver water at the same rate as smaller diameter pipes but at much reduced velocities, leading to reduced total head loss [17] . Therefore, it is important that the head loss within the open channel is not impacted thereby retaining efficient flow movement. Increases in head loss result in a reduction of velocity and more turbulent flow.

The impacts of restrictions/utility crossings in open channel flow are that it impacts the water flow in the river. There is no specific research available that analyses the impact of utilities at bridge or culvert crossing in open channel flow in terms of flood risk, however, each scenario can be surveyed and modeled. The Scottish Environmental Protection Agency (SEPA) states the impacts of poorly designed river crossings which include barriers to fish movements, prevent sediment transfer, prevents natural river migration, and increase flood risk [18] . SEPA advises that if a utility needs to cross a watercourse it should be done so under the watercourse, underneath the riverbed [18] .

Flooding has an impact socially, economically and environmentally. Socially it can have an impact on people’s health and can cause deaths. Economically it is not only the costs of the repairs it is also the losses due to businesses being closed, health care costs, rescues and temporary accommodations [19] . It also has an impact on the environment with pollution and damaging habitats and ecosystems.

The most common cause of flooding of culverts is blockages. Lewis, A. [26] states

that during Storm Dennis some culverts had inadequate protection. Three culverts were identified as sources of flooding and their capacities were reduced by debris. The Environment Agency has maintenance programmers and they state that it plays an essential part in reducing flooding. Culvert and bridge blockages are the highest risk for flooding in urban areas. Over 5 years the Environment Agency has spent £35 - 55 Million a year on channel maintenance [19] .

CIRCA has a section on obstructions in the barrel of the culvert and states that obstructions are the most common cause of blockages, and it states that older culverts may have service crossings that cause an obstruction in the waterway [16] . It also notes that these situations should be avoided but it does not state that it is unacceptable. There are no current reports from the Water Research Council (WRC) on how utility crossing impacts below the bridge or culvert soffit open channel flows.

There are no documented reports in the UK specifying how the hydraulics of open channel flow have impacted flooding as well as no reports on utility crossings causing flooding. However, in the US there have been reports of flooding due to the Yazoo Backwater Project. The structures included levees, canals and drainage structures that created a backwater curve which resulted in flooding [27] . Backwater flooding is defined by Finish the Pumps [28] as upstream flooding caused by downstream conditions, these conditions include channel restriction.

The research found a myriad of guidance on culvert design. This research is unique in that there are no flooding event case studies where there has been research on how an obstruction or utility crossing in a culvert affects river flow. CIRIA C786 stated that utility crossings are a problem [16] , therefore, there is a gap in this research area.

The results from the survey provided the information shown in Table 5 and Table 6.

From Manning’s equation the flow for the open channel and culvert can be calculated:

Q = A n R 2 / 3 R 1 / 3

where Q is the discharge; S is the hydraulic gradient, in this case the channel bed slope or culvert floor slab slope; A is the area of the pipe; P is the wetted perimeter; R is the hydraulic radius (A/P); and n is Manning’s roughness coefficient.

Eliminating the utility crossing restriction and using the information in Table 5 measured results from the survey and Table 6 estimated Manning’s n values for the friction losses, the culvert can deal with a flow of 46.179 m³/s as shown in Table 7. This is over 2.5 times the flow in the upstream open channel at a depth of 2.186 m as shown in Table 8. While the height of 2.186 m is the level that the flooding of the Hegarty Property occurred, the culvert can deal with a flow depth in the open channel of over 3.0 m. From the flood water depth in the dwelling house, it is judged that the flow in the stream was at a depth of circa 2.700 m during the flooding event. From Table 8 it is estimated that the flow in the open channel upstream of the culvert was 23.909 m³/s or 51.77% of the culvert flow capacity during the flooding event. This estimate is made when considering the friction losses-controlled capacity of the culvert only. Therefore, for the given scenario, the utility crossing coupled with flood debris reduced the capacity of the culvert flow by circa 50%. The 50% value is not a definitive entry loss coefficient as there is insufficient data available relating to the flood debris that the utility crossing may have contributed to stopping temporarily at the culvert upstream entrance.

The Manning’s n values in Table 6 are estimated based on published data and therefore are not computed for the specific channel. The computing of Manning’s n involves consideration of n values for: the base, surface irregularities, variations in shape and size of the channel cross section, obstructions and vegetation along with a correction factor for the meandering of the channel. The calculations in Table 7 and Table 8 are derived judging that the culvert is losses controlled due to internal surface friction as opposed to being inlet or outlet controlled. An entrance loss coefficient was deliberately not established for these calculations to establish whether the culvert is inlet controlled or losses controlled, however the calculations in Table 7 and Table 8 suggest the culvert is inlet controlled due to the utility crossing. The culvert flow regimes including the impact of the utility obstruction will be investigated as a coursework exercise and self-study exercises by students at Ulster University completing Civil Engineering degree courses.