Flood Protection Performance Assessment: Nerang Catchment

Introduction and Background

One of the first signs of global warming will be a rise in water levels around the world. Flooding will increase in many areas already prone to flooding; in addition, new floodplains will be created, placing humans and property at risk. The escalating impact of climate change makes it necessary to reassess existing flood protection systems and to evaluate the need for new ones in areas not previously at risk for flooding. The purpose of this research is to assess the performance of the Nerang Floodplain Catchment system in the Gold Coast area of Australia.

This assessment will develop a 2D hydraulic simulation model for the Nerang Floodplain catchment with regard to present and future estimated strain from global warming. The Nerang Floodplain will serve as an example of how such an assessment can be accomplished in other at-risk areas. The study will result in the development of a model that can be used to evaluate other catchment systems elsewhere in the world.

The focus of the research is on methodology development, using the Nerang Floodplain as a case example. The analysis uses five primary variables to assess the catchment system, derived from previous research and a review of existing literature on floodplain catchment evaluation. The five variables are vulnerability (α), reliability (β), resiliency (γ) (Hashimoto, 1982), Flood Risk Index (FRI), and Flood Damage Index (FDI) (Zongxue, 1998). Measurements will illustrate the overall performance of the catchment system in a spatially distributed manner for the Nerang catchment.

The result of the project will be to evaluate the performance results to determine the actions needed to bring the Nerang Catchment up to the standards necessary to counteract the effects of rising coastal waters. Recommendations will address the modification, upgrade, or complete replacement of the existing system as determined by the performance assessment. This research will also contribute a consistent methodology for assessing not only flood protection systems, but any engineering infrastructure more broadly.

In order to fully understand the scope and application of the model developed in this research study, it is important to understand the basic parameters and concepts presented. This section explores the precise meanings of various terms used throughout the study, as well as an introduction to the targeted geographic area and its current catchment system.

A floodplain is defined as a flat area of land adjacent to a waterway that experiences periodic or occasional flooding. The morphology of a floodplain is highly related to the hydrogeology and sediment structure of the area. The floodplain can be described in terms of its floodway, which consists of a stream channel and the areas that carry flood flows. The floodway carries current during periods of flooding and typically extends beyond the normal stream or river channel. The flood fringe refers to an area covered by the flood that does not experience strong currents during flooding periods.

Floodplains are created as the banks of a stream erode. Periods of flooding accelerate riverbank erosion, leaving the underlying bedrock exposed and allowing layers of sediment to build up into the characteristic flat landform. Floodplains generally consist of unconsolidated sediments of heterogeneous materials, including layers of sand, gravel, loam, silt, or clay. These layers form a natural filtration system as water filters into local aquifers. The composition of these layers determines the nature of flooding in the area. As flooding continues, stream terraces form on the sides, representing the hydrogeologic history of the area.

Floodplains are an important part of the ecosystem, supporting riparian-zone plants. These areas are continually wet and support large quantities of microorganisms that help break down organic material. As a result, the soil in a floodplain is often fertile and well suited for growing many crops. Floodplains are often tempting areas for human development, offering flat building sites, highly fertile soil, and access to nearby water. However, these areas are risky depending on the return period of floods, and human occupation of floodplain areas forms the major concern driving the need for adequate catchment systems.

The Gold Coast corner of Queensland, Australia is a highly populated area situated among subtropical rainforests. The area's subtropical climate makes it an important tourist destination that contributes significantly to the economic prosperity of Australia. The geography of the area once consisted of wetlands between the Nerang River and the coastal strip; these have since been converted into man-made waterways and artificial islands supporting upscale homes. The most highly developed and populated area is situated atop a narrow barrier sandbar lying between the waterways and the sea. A sand bypass system was installed to help stabilize sand movement due to tides and flooding in the area (Boswood & Murray, 2001).

Global Climate Change and Southeast Queensland

The level of development and economic prosperity of the Gold Coast area make it a high priority for evaluation of its floodplain catchment systems. Due to its local geology and coastal location, the Gold Coast is an area of high concern as waters rise due to global warming. This research will play an important role in the ability to sustain this highly valued asset to the Australian people.

Current models indicate that the risks to humans in coastal areas are changing due to climate shifts and social trends (Stansby, Walkden, & Dawson et al., 2009). These changes are driven by rising water levels from the melting of glaciers and polar ice caps. Increased water results in increased coastal erosion, which creates conditions favorable to the development of floodplains in areas not previously at risk for coastal flooding (Stansby, Walkden, & Dawson et al., 2009).

Coastal dynamics result in sedimentary exchange between different geographic locations. Older models of coastal flooding considered populated areas in isolation from surrounding regions (Stansby, Walkden, & Dawson et al., 2009). However, improved modeling of the interdependency of coastal areas is rapidly changing this philosophy. The new paradigm in coastal floodplain modeling considers the entire coastal zone as a dynamic system, rather than as isolated patches (Stansby, Walkden, & Dawson et al., 2009). This is an important point for the development of flood catchment systems in the Gold Coast area: anything affecting the Gold Coast will influence the surrounding wilderness areas as well, and vice versa. These dynamics are quickly reshaping long-term management approaches for coastal areas.

It is not surprising that the Gold Coast is taking a proactive approach to protecting its local resources from the impact of flooding due to climate change. The Nerang River Flood Mitigation Strategy, for instance, imposes requirements on new development, mandating that any new construction take into consideration its impact on flooding to neighboring properties, whether downstream or upstream (Gold Coast Council, 2006). This research will serve as an important tool in the development of short-, mid-range, and long-term flood planning in the area.

The Gold Coast has been rated as one of the most vulnerable areas to flooding as a result of climate change (Mirfenderesk, 2009). Since 1924, the area has experienced over 45 floods, triggered by cyclones and periods of excessive rainfall (Mirfenderesk, 2009). The Gold Coast comprises seven major catchment areas: Tallebudgera, Currumbin, Nerang River, Coomera River, Pimpama River, South Moreton Bay, Sandy Creek, and the Broadwater area (Mirfenderesk, 2009). The Nerang River catchment is adjacent to the Tallebudgera catchment to the south, bordered by the Broadwater and Coomera River area to the north, and adjacent to the Pacific beach area (Mirfenderesk, 2009).

Catchment areas have different levels of tolerance before water-to-sediment concentration reaches saturation levels. The Tallebudgera, Currumbin, and Broadwater areas have time concentrations of approximately three hours, creating conditions favorable to short-duration local flooding (Mirfenderesk, 2009). The Nerang River and Coomera catchments have time-of-concentration values ranging from 3 to 92 hours, making them susceptible to regional-scale flooding of longer duration as well as short-duration local flooding (Mirfenderesk, 2009). Local flooding results from heavy rainfall over a short period in a confined area and typically drains quickly, while regional flooding covers a large geographic area, may trigger local flooding, and can take several days to subside (Mirfenderesk, 2009).

CSIRO Sustainable Ecosystems developed a matrix examining the effect of climate change on infrastructure across four scenarios: buildings in coastal settlements, electricity distribution and transmission, water supply infrastructure in major cities, and port infrastructure and operations (CSIRO Sustainable Ecosystems, 2006). Impacts were evaluated according to estimated economic shock, using projections across three time periods: 2007–2030, 2031–2070, and 2071–2100, assessed under seven different climate change scenarios.

Climate Change Predictions: Rainfall, Sea Level, and Storms

According to the CSIRO model, increased storm surge height and land penetration would have a negative impact on current flood protection systems for communities within 50 km of the Queensland coastline (CSIRO Sustainable Ecosystems, 2006). The magnitude of potential damage is currently considered low to moderate but is expected to increase to moderate to high during 2031–2070, with some areas experiencing extreme potential for severe flood damage (CSIRO Sustainable Ecosystems, 2006). These findings align with other modeling scenarios, all agreeing that the potential for damage from coastal flooding is expected to increase significantly from 2030 to 2070, making the evaluation and improvement of existing flood protection systems essential to the sustainable development of Queensland coastal communities.

Climate change is expected to increase the severity and frequency of floods in the Gold Coast area, compounding an already flood-prone situation (Mirfenderesk, 2009). From a flood protection planning perspective, design must account for the worst-case scenario. Estimates range from conservative to alarmist; according to one media release, by 2060 to 2070, tides are expected to be 40–50 cm higher than current average springtime tides (Coulter, 2009). The most relevant factors for Gold Coast flood prediction are average rainfall, sea level, and the frequency and intensity of storms.

Mean annual rainfall totals for Southeast Queensland average 1,354 mm per year, with some of the highest rainfall occurring in the Nerang River catchment (Gold Coast City Council, n.d.). Extreme rainfall events are most significant along Australia's eastern coastline, which also corresponds to the highest population centers (Abbs, n.d.). Trend analysis on rainfall in Southeastern Queensland and Northern New South Wales found that rainfall has been steadily increasing over the past decade; however, when larger regions are considered, the averaging effect reduces the apparent magnitude of the increase (Abbs, n.d.). Regression analysis of a specific selected area over time provides the most accurate assessment of local rainfall trends.

Rainfall simulation for a selected area of Southeastern Queensland and Northern New South Wales projected current rainfall and compared it to predicted rainfall in 2040. The model used the Regional Atmospheric Modeling System (RAMS) to produce a compressible non-hydrostatic model using data from CSIRO Mark 3 GCM, with terrain interpolated from the USGS 30-second data set (Abbs, n.d.). The system was tested against data from 1960–1999. Though the model underestimated the intensity of extreme rainfall events, predicted and actual estimates were sufficiently close to be considered reliable for future projections. The 2040 prediction found that rainfall tended to increase over mountainous regions but decrease elsewhere, with some increases near coastal areas such as north of Brisbane and south of Cape Byron (Abbs, n.d.).

Long-term rainfall predictions are subject to considerable error. Trend analysis from 1950–2004 examining summer and winter rainfall in the Southeastern Queensland coastal area demonstrated that as global temperatures increased, the amount of rainfall in the area decreased (Whetton, McInnes, & Jones et al., 2005). Temperatures are expected to warm significantly between now and 2070, while rainfall is expected to decrease regionally (Whetton, McInnes, & Jones et al., 2005). Rainfall predictions are highly sensitive to temperature projections and to the model selected for the analysis, and a majority of models agree on the prediction of drier conditions in the area.

Current predictions call for an 80 cm rise in sea levels in the Queensland area (Frasier, 2009). The Intergovernmental Panel on Climate Change indicated that this rise is expected to be close to 30 cm by 2050, with the Gold Coast, Yeppoon, Mackay, and Cairns expected to feel the greatest impact (Frasier, 2009). Several factors contribute to this rise: melting ice from glaciers and polar regions, thermal expansion of the oceans as temperatures increase, and storm surges, which will add to coastal flooding as tropical cyclone frequency or intensity changes. Storm surge height is a function of wind speed and tidal phase when the surge arrives (McInnes, Walsh, & Pittock, 2000), and wide, shallow coastal ocean floors tend to produce the highest surges.

The area of Southeastern Queensland and Northern New South Wales is prone to severe tropical and subtropical storms and is classified as a transition area by the Commonwealth Bureau of Meteorology (Abbs, n.d.). Tropical cyclones, while more severe, are also relatively rare; however, the region frequently experiences subtropical depressions (Abbs, n.d.). Tropical cyclones are measured as the radius from the centre to the region of maximum winds, with an average radius of 30 km, though larger and smaller storms occur (McInnes, Walsh, & Pittock, 2000).

Predictions indicate that several factors may increase the risk of flooding from storm surges and tides. As cyclone intensity and frequency change, sea levels will respond in kind (McInnes, Walsh, & Pittock, 2000). Increasing cyclone intensity produces larger waves. The impact of changes in cyclone frequency due to global warming is a topic of considerable debate; it is not known whether cyclones will increase or decrease overall, as much depends on how global warming affects differential ocean temperatures (McInnes, Walsh, & Pittock, 2000).

Performance Assessment of the Nerang Catchment

Excess runoff from Worongary, Bonogin, and Mudgeerba Creeks flows into the Merrimac floodplain, and excess runoff from the upper catchment causes the Nerang River to break its banks and flow into the floodplain at Carrara (Gold Coast City Council, n.d.). The most significant floods of this type occurred in 1931, 1947, 1954, 1967, and 1974, with cyclones, major low-pressure systems, and severe thunderstorms serving as the main triggers (Bureau of Meteorology, 2009). A majority of flooding events in the Nerang catchment are minor, with floodwater rises of only 6 meters or less, representing localized flooding limited to a few roads or square kilometers of residential areas. Major regional flooding events have occurred only four times since 1920, with the last in 1974 (Bureau of Meteorology, 2009).

The flood protection system of the Nerang catchment has historically depended on early warning for its residents. The flood warning system is based on predicted rainfall amounts and river height observations, and relies largely on a volunteer organization that uses telephone notification when river heights exceed a certain level at monitoring stations (Bureau of Meteorology, 2009). Flood warnings and bulletins are issued to residents in at-risk locations along the Nerang River.

Flood protection systems must take into account the flood risk of the area. Flood risk is a function of hazard, exposure, and vulnerability (Mirfenderesk, 2009). Traditional flood risk assessment uses the 100-year Average Recurrence Interval (ARI), which measures the average occurrence of a flood of a particular size — not the interval between floods. For instance, during the 1890s, the Brisbane River experienced three 100-year ARI floods within five years (Mirfenderesk, 2009). With the unknown factors resulting from global warming and the potential for increased severity and frequency of flooding, Gold Coast City Council has begun to focus on flood mitigation, with an early proposal to raise the Hinze Dam for flood mitigation purposes (Gold Coast City Council, 2003).

The Gold Coast City Council (n.d.) has developed a detailed map of the Nerang Catchment identifying high-risk flooding areas. The Nerang River Catchment area includes portions of Main Beach, Broadbeach, and Burleigh Heads, covering approximately 490 square kilometers — approximately one-third of Gold Coast City's area — and encompasses 72,000 properties (Gold Coast City Council, n.d.). The river enters the ocean near the Gold Coast Bridge. The catchment can be divided into the Upper Catchment (west of the Pacific Highway) and Lower Catchment (east of the Pacific Highway), and consists of the Nerang River and a number of associated creeks and local catchments.

During a flood, water from the Nerang River and its associated creeks accumulates in the lower portions of the river. The Nerang River is partially controlled by two dams: a small dam on the Little Nerang Catchment, and the Hinze Dam, which has a catchment area of 210 square kilometers and controls 43% of the overall catchment. Six additional catchments within the system account for nearly 57% of the Nerang Catchment area and flow into the floodplain to the south of the Nerang River (Gold Coast City Council, n.d.). Storm surges increase the flooding potential of the area as they enter the Broadwater from the ocean.

In flood-prone areas such as the Nerang catchment, having an adequate flood protection system in place is essential to the lives and livelihoods of residents and businesses. The Nerang River catchment is more prone to coastal flooding from cyclones and severe storms than to inland flooding, making coastal flooding the key risk to residents and property in Gold Coast City. In the face of increased flood potential due to global warming, it is essential that the current system of structural flood controls in the Nerang River catchment be examined to ensure they offer maximum protection for predicted future flood scenarios.

Water supply performance is measured by the ability of a reservoir to supply a chosen community with its target water resources. The ability to supply the needed target release is referred to as reliability (Cohon & ReVelle, 1986). Performance can also be measured as a function of the maximum shortfall from the targeted release, referred to as vulnerability (Cohon & ReVelle, 1986). The number of consecutive periods of deficit is referred to as system resilience (Cohon & ReVelle, 1986). These three parameters constitute the most common performance measures of a reservoir system.

Using mixed-integer linear programming, a tradeoff was found among reliability, vulnerability, and resilience. As reliability increases or the maximum length of consecutive shortfalls decreases, the vulnerability of the system to greater deficits in supply increases (Cohon & ReVelle, 1986). Designing a failure-free system is next to impossible, and even the best-designed system can be inundated by extremes from nature. Therefore, reliability, resiliency, and vulnerability must be considered as design elements of any flood protection system.

The reliability of a system can be described in terms of the frequency of a system's satisfactory output state — expressed as the probability that a system is operating in a satisfactory state at any given time (Hashimoto, 1982). This definition does not provide information about the severity of failure; for that, one must also examine resiliency and vulnerability. The reliability of any engineered system can be expressed as a function of its load (demand) and resistance (capacity). In the context of flood deterrence, the amount of water the system must resist represents the load, and the capacity of the system represents the maximum resistance before failure occurs.

Maier, Lence, & Tolson et al. (2001) introduced the First-Order Reliability Method (FORM), which has been applied to many engineering projects including surface water applications and the prediction of rainfall-runoff in the Vermilion watershed in Illinois, USA. FORM performed well and demonstrated good agreement between storm magnitude and storm types.

Resiliency describes how quickly a system is likely to recover once a failure occurs, expressed in terms of time from the point of failure to the satisfactory operational recovery of the system (Hashimoto, 1982). Applied to a flood deterrence system, resiliency refers to the time from when flooding occurs to when the system — whether natural or engineered — allows water to recede and community life to return to normalcy. Several factors affect resiliency after a flood, including the type of sediment and its ability to channel or absorb excess water, the presence of diversion channels or ditches, and man-made pumping and canal systems that remove excess water from the area.

Vulnerability refers to the magnitude of damage once a system has failed (Hashimoto, 1982). In a flood scenario, vulnerability depends on a combination of natural and engineered factors: the strength of the storm, the ability of the catchment area to absorb or divert excess water, and the effectiveness of flood deterrence structures such as dykes, channels, floodwalls, and pumping systems.

The failure of the levee system in New Orleans during Hurricane Katrina is a sobering reminder of the importance of minimizing vulnerability in the design of flood deterrence systems. Hashimoto (1982) notes that building a fully fail-safe water supply system is not possible from a logistical and cost perspective, and the same principle applies to flood deterrence systems. Therefore, safety in the event of failure must be a key design consideration, and vulnerability must be a core component of future system modeling.

Conclusion

Global climate change represents one of the greatest challenges that humans currently face. Flooding and changes in water conditions around the world are one of the first manifestations of the warming trend. The Gold Coast is one of the most important coastal cities in Queensland and has historically been known as an area prone to flooding, particularly during coastal storms.