Engineers, planners and scientists have long been troubled by the complex and sometimes catastrophic interplay between waterways and infrastructure during flood events. Aurecon Flood Modeller, Lee Williams, has conducted new research into hydraulic flood modelling sensitivity analysis of culvert blockage factors and blockage timing.
The research findings will assist the Australian water industry to understand flooding effects and improve industry-wide policies and design standards in the future, which is particularly important given the variability of climate change and the anticipation of more frequent and severe extreme weather events.
The research covers hydraulic modelling sensitivity analysis of culvert blockages. During a flood event, the large volume of rainfall falling across a catchment area is highly likely to collect debris material such as floating vegetation, erosion related sediment material and even urban debris. When this debris accumulates and is transported towards a culvert or other such hydraulic structure, the available area of the culvert may become blocked, reducing the capacity to pass the flood waters.
This can have devastating effects to regions upstream of the structure and significantly increases the risk to local inhabitants.
This research identified that culvert blockage factors, the time the blockage occurs, the magnitude of the flood event and the available upstream storage volume all influence the outcome of the water level flood profile, in addition to the effect on the downstream discharge on the culvert capacity and weir overflow conveyance relationship. In particular, the research found that:
Australian Rainfall and Runoff (ARR) is as a national guideline published by Engineers Australia and used for the estimation of design flood characteristics and flood risk. A significant revision of the document was released in 2016, superseding the previous publication from 1987. This document was prepared in part due to our industry’s enhanced theoretical understanding and significant development in computer software analysis over the past 30 years, among other factors.
Book 6 – Flood Hydraulics of the ARR 2016 publication, "Chapter 6: Blockage of Hydraulic Structures" and by extension ARR Project 11: Blockage of Hydraulic Structures Stage 2 Report were the source of motivation for preparing this research paper.
ARR 2016 introduced a Blockage Assessment Form to assist engineers in assessing the availability, mobility and transportability of debris based on site specific characteristics, which are used to determine appropriate blockage factors to apply when assessing hydraulic structures. Although representing current best practice, ARR identifies that advice “…is based heavily on limited observations and should be updated as further data becomes available.”
These observations largely consist of photographic evidence and other forms of noted observations taken by residents after a flood event has occurred. While these are a useful source of information, Rigby & Barthelmess (2012) identified the need for more quantifiable data:
“If we are to advance our understanding we will need continuous upstream and downstream stage and flow to be recorded together with associated catchment characteristics and characteristics of debris forming the blockage.”
The hydraulic flood modelling sensitivity analysis conducted within this research paper has been undertaken to assess the influence of blockage timing upon a hydraulic structure.
The precipitation falling within a catchment area forms surface runoff that accumulates within rivers and watercourses. As the rainfall event arrives and dissipates over time, so too will the discharge runoff with a rising limb, peak discharge and failing limb as illustrated below in Figure 1. During a flood event, a number of different debris types can become trapped within a hydraulic structure, such as floating (trees), non-floating (sediment) and urban (garbage bins, cars, etc.).
Figure 1: Rainfall to Discharge Hydrograph Diagram (Geography Revision, n.d.)
The arrival time of a blockage to the hydraulic structure during the discharge curve can make a significant difference to the water levels both upstream and downstream as well as the quantity of flow being conveyed through the structure.
As the planet experiences an increase in atmospheric temperature, the variability of heavy rainfall is expected to increase in intensity and frequency. Given the characteristics of the urban catchment researched within this analysis an increase in rainfall intensity during a short duration event may have a significant impact on the conveyance at the hydraulic structure and alter the risk of the flood event, highlighting the importance of climate variability moving forward.
The application of blockage timing remains a difficult concept to adopt in practice. Many modelling software packages lack the capability to model transient property changes, while the variability of the overall effect makes the impacts and risks difficult to quantify without reverting to a Monte-Carlo simulation, which in most cases is not a feasible option. A simplified approach has been undertaken within this research paper to bridge the gap in the understanding of blockage timing.
A TUFLOW model was prepared to assess the hydraulic blockage of a road crossing using a 1D culvert structure within an urban setting, as illustrated in Figure 2. Typical roughness materials were adopted across the model and a theoretical upstream inflow hydrograph was assumed.
Figure 2: TUFLOW Model Layout
In order to assess the influence of blockage timing, a sluice gate operational feature was adopted to apply the blockage at a fixed time during the discharge curve. The sluice gate operational feature reduces the height of the hydraulic structure in order to simulate the reduction of culvert capacity due to debris blockage. This assessment has taken into consideration an unblocked condition as a baseline for comparison in addition to 50 per cent, 75 per cent and 100 per cent reduction in hydraulic structure height from the top down.
This research paper has focused on a pulse or rapid blockage to simulate the sudden arrival of a large form of debris, investigating the influence of both the magnitude and timing of the debris blockage arrival. An alternative progressive blockage assessment is to be explored in the future. Flow arriving at the road can be conveyed either through the hydraulic structure (culvert) or by overtopping the roadway as weir flow.
Figure 3: Road Embankment Overtopping Long Profile Arrangement
The configuration investigated by the study provides a relatively large potential storage volume upstream of the hydraulic structure. However, the volume storage is relative to the catchment in question. For overtopping of the roadway to occur as weir flow, the upstream storage volume would need to fill with water.
The influence of a blockage occurring, reduces the time taken for the storage volume to fill and for overtopping to occur. For example, if the hydraulic blockage was applied during the rising limb of the discharge curve, the volume of water escaping from the upstream storage area would be reduced from the culvert, the upstream water level would rise and would overtop the roadway at a time much sooner than if the hydraulic blockage had not been applied.
With the TUFLOW model run, the theoretical inflow hydrograph illustrated below depicts the timings at which the sluice gate was applied during the simulation. Blockage has been applied during the rising limb, at the peak discharge and during the falling limb with all model conditions to be compared against the unblocked condition.
Figure 4 depicts the blockage in effect within the culvert structure with the time of the blockage applied in accordance with the theoretical inflow hydrograph. As most of the hydrograph volume has already passed through the culvert when the falling limb blockage is applied, the outcomes presented within this report will only focus on the rising limb and peak flow blockages.
Figure 4: Graphical representation of applied Blockage Timing
The graphical output presented below depicts the water level impacts caused by applying a blockage factor of 50 per cent and 75 per cent to a theoretical cross drainage hydraulic structure within the TUFLOW model for the rising limb and peak discharge sluice gate times compared to the unblocked condition.
Figure 5: Water Levels Blockage applied during different modelling conditions
The outcome of this research sensitivity analysis demonstrates in Figure 5 that:
Figure 6: Increased Magnitude Flood Profile with Overtopping Weir Flow
However, the outcomes listed above are conditional upon the unblocked condition not exceeding the road embankment. The graphical representation of Figure 6 depicts that if the unblocked condition exceeds the hydraulic structure’s capacity and overtops the road embankment producing weir flow an increase in hydraulic blockage factor will continue to increase the upstream water level, however by a lesser margin than if the unblocked condition does not overtop the road embankment.
For an Unblocked Condition that does not overtop the road embankment, a combined weir and hydraulic structure flow with an applied blockage factor at any time during the inflow hydrograph may not exceed the Unblocked Condition as shown in Figure 7.
Figure 7: Downstream Discharge Comparison for 50 per cent and 75 per cent Blockage
However, for an Unblocked Condition that experiences road embankment Overtopping Weir Flow, a blockage factor applied during the rising limb can match or exceed the Unblocked Condition, which may seem contradictory to the pre-conception that a culvert blockage should reduce downstream flows. This is because, although the culvert discharge is significantly reduced, this is offset by increased overflow over the weir, which has been demonstrated in Figure 8.
Figure 8: Increased Magnitude Downstream Discharge Comparison
As expected, reducing the culvert capacity results in an increase in upstream water level, with the increase a direct function of the amount of blockage applied. Timing of the blockage is also a factor, with blockage occurring later stage in the flood hydrograph having less impact given most of the flood volume may have already passed through the structure. Downstream impacts are a much more complicated result of the interaction of inflows, outflows and storage volume. Lastly, changes in water level downstream of the road can be related to the discharge from the structure when addressed in conjunction with the culvert and weir overtopping relationship.
The sensitivity analysis results presented within Figure 7 and Figure 8 above demonstrate the influence of applying a blockage factor to a hydraulic structure through a comparison of discharge curves against an unblocked condition. Therefore, when conducting a hydraulic analysis, in order to determine the maximum downstream conveyance for a cross drainage hydraulic structure, the nature of the unblocked condition in relation to the presence of overtopping weir flow has been identified within this sensitivity analysis as a characteristic that warrants further investigation.
The hydraulic modelling research sensitivity analysis of hydraulic structures within this cross-drainage structure aimed to evaluate the impact of different blockage factors, blockage timing, inflow magnitudes of discharge and the relationship between available upstream storage for an unblocked condition compared with road embankment overtopping weir flow. A TUFLOW model was utilised to assess the different hydraulic blockages using a sluice gate operational feature applying a blockage factor during the rising limb of the inflow hydrograph or at the time of the peak flow.
The outcomes of the analysis concluded that hydraulic blockage increases the upstream water level, however, the influence of the timing of the applied blockage will dictate how the downstream water level will behave. Furthermore, the application of a blockage factor and the relationship between the unblocked condition with the road embankment experiencing weir flow has a significant effect on the downstream hydrograph. Further investigation into the relationship between the culvert conveyance and the weir overflow relationship through further hydraulic modelling sensitivity analysis is recommended moving forward.
The modelling research from Lee Williams is only the start to more industry-specific research, investigation and modelling. There is more work to be done to understand flood behaviour and blockages at culverts, and how structural design can better respond to flood events and flow arrangements under blockage conditions.
It would be a wise investment to investigate alternative bottom-up or porous plug blockage configurations, as well as a progressive blockage arrangement over different durations to compare how the sensitivity of these different variables relates to the results from Williams’ research.
Furthermore, investigation into a range of hydrograph durations and temporal patterns across a wide range of flood magnitude and varying upstream storage volumes would open new and exciting opportunities for culvert blockage research based on hydraulic blockage, rather than the observation blockage datasets currently influencing national policy.
The next phase in this research could be a review of blockage factors for hydraulic design in addition to community flood risk management, preparedness for existing conditions, and disaster resilience warning/planning for evacuation strategies.
Imagine if the implementation of this blockage analysis methodology into a large-scale infrastructure project with numerous cross drainage culvert structures provided an improved understanding of the flooding impacts resulting in an increase in project safety through design.
Imagine if by analysing the blockage and determining that the impacts that are produced do not have a detrimental effect to the design or the wider community, the overall project drainage expense is significantly reduced rather than conservatively over-estimating the required culvert capacity.
Imagine if the consequence of a range of culvert blockage scenarios was known for an existing structure ahead of the next big flood. Given the variability of climate change and the anticipation of more frequent and severe extreme weather events to occur caused by a warmer atmosphere, this information could become invaluable for future events. This knowledge could be acted upon to implement floodplain management processes to ensure culverts highly sensitive to blockage factors are managed more appropriately, to improve the performance and longevity of the asset and reduce expensive upgrades that may require road or rail closures to be implemented.
Lee Williams is a Flood Modeller in Aurecon’s Newcastle office and this research was part of his postgraduate studies in a Master of Water Engineering at the University of Technology Sydney, mentored by Professor James Ball.
He holds industry leadership positions as Co-chair of the Engineers Australia Young Engineers network in Newcastle, and the 2020 Co-chair of Aurecon’s Emerging Professionals network.
 Geography Revision. (n.d.). Retrieved from River Discharge.