Quantifying the Effects of Sea Level Rise on Estuarine Drainage Systems
Abstract
Much of the development of the low elevation coastal zone has involved the reclamation of low-lying floodplains and wetlands through the construction of flood mitigation and drainage systems. These systems function throughout the tidal range, protecting from high tides while draining excess catchment flows to the low tide. However, drainage can only be achieved under gravity when water levels in the catchment drains are higher than those in the estuary. Changes to the tidal range and to the duration of the rising and falling tides that occur throughout estuarine waters will result in dynamic variations in the window of opportunity for gravity discharge within and between different catchments and under sea level rise (SLR). Existing concerns regarding SLR impacts have focused on the acute effects of higher water levels, but SLR will affect the full tidal range, and drainage systems will be particularly vulnerable to changes in the low tide. This study introduces the concept of the drainage window to address this limitation by assessing how the present-day and future SLR tidal regimes may influence the drainage of different estuarine floodplains. Applying the drainage window to two different estuaries indicated that SLR may substantially reduce the opportunity for discharging many estuarine floodplain drainage systems. Reduced drainage creates a host of chronic problems that may necessitate changes to existing land uses. A holistic assessment of future changes to all water levels (including low tide levels and extended flood recession periods) is required to inform strategic land use planning and estuarine management.
Key Points
-
The drainage window is conceptualized and applied to two estuaries to quantify the effects of sea level rise on tidal drainage systems
-
Areas that are protected from intermittent flooding may be vulnerable to chronic waterlogging due to impeded drainage
-
Loss of function and amenity due to impeded drainage should be considered in future land use planning
Plain Language Summary
Estuaries are the tidal waters located where rivers meet the sea. The floodplains adjacent to estuaries are some of the most heavily developed areas in the world. Much of this development relies on integrated flood management and drainage schemes that use one-way valves (floodgates) to protect the floodplains from inundation by high tides and floods, while allowing the floodplain drains to discharge when the water level in the estuary is lower than the water level in the drains. Tidal levels can vary along an estuary and may change under accelerating sea level rise (SLR). This study introduces the concept of the drainage window to quantify how much time is available to drain different floodplain catchments within an estuary and to identify how that window of opportunity may be affected by SLR. The drainage window was analyzed for two estuaries, with the results indicating that SLR may substantially reduce the time available to drain each system. Areas with less time to drain are more susceptible to chronic problems associated with prolonged inundation and waterlogging that may necessitate changes to existing land uses. These results could therefore be used to inform strategic land use planning and management in estuaries worldwide.
1 Introduction
As the nexus between land and sea, coasts and estuaries have been a focal point for human settlement, with their abundance of natural resources and ecosystem services attracting extensive and ongoing development (Martínez et al., 2007; Neumann et al., 2015). Worldwide, over one billion people reside less than 10 m above current high tide levels (Domingues et al., 2018; Kulp & Strauss, 2019). Two-thirds of the world's megacities are situated on coasts and estuaries (Oliver-Smith, 2009) and approximately 14% of the world's gross domestic product is generated in the low elevation coastal zone (Magnan et al., 2019).
Much of the development of the low elevation coastal zone has been facilitated by the anthropogenic drainage of floodplains and wetlands, predominantly for agriculture, but also for urban, maritime, and industrial use (Church et al., 2010; J. G. Titus et al., 1987; Tulau, 2011). Channels, pipes and culverts have been installed throughout estuarine catchments to efficiently remove excess surface and groundwater from backswamps, wetlands, and floodplains (J. G. Titus et al., 2009). Frequently, natural levees are augmented and dykes and seawalls are constructed to protect these lands from tidal inundation or high fluvial water levels (Kroon & Ansell, 2006; Lugo & Snedaker, 1974; Poulter et al., 2008). The reclaimed areas created by these works are variously referred to as polders, koogs, or wei. Intermittently operated tidal gates, such as the Thames Barrier in London (Horner, 1979) or the Lake Borgne Surge Barrier in New Orleans (Huntsman, 2011) may be used to prevent storm surges from progressing upstream along an estuary. To protect low-lying floodplains and reclaimed lands from regular inundation by high tides however, one-way valves (herein referred to as floodgates, but also known inter alia as tidal flaps, flap gates, non-return or reflux valves) are often installed where tributaries or drainage channels discharge to the main estuary (Johnston et al., 2005; Ota, 2018; Ruprecht et al., 2018). Tidal floodgates are widely implemented throughout estuaries along the east coast of Australia (Boys et al., 2012; Williams & Watford, 1997), and can also be seen in Europe (Díez-Minguito et al., 2012; Solomon, 2010), Asia (Award, 1995; Choi, 2014; Warner et al., 2018; Zhao et al., 2020), and North America (Giannico & Souder, 2004; Rillahan et al., 2021). These floodgates operate throughout the full tidal range, providing protection against high tide inundation while facilitating drainage to the low tide level. Their continued operation will therefore be vulnerable to future sea level rise (SLR).
Globally, the impacts of SLR are already being experienced in the low elevation coastal zone (Magnan et al., 2019), with the largest changes in tidal dynamics observed in estuaries and tidal rivers (Talke & Jay, 2020). According to the latest report from the Intergovernmental Panel on Climate Change (IPCC), the average global mean sea level is predicted to increase by between 0.28 and 1.01 m by 2100, relative to 1995–2014 average (Masson-Delmotte et al., 2021). A growing body of literature indicates that, within estuaries, the impact of SLR will vary throughout the full tidal range, with diverse effects from high to low tide levels (Haigh et al., 2020; Khojasteh, Glamore, et al., 2021; Talke & Jay, 2020). Each estuary, including tributaries and different reaches within an estuary, may respond differently to SLR (Du et al., 2018; Khojasteh, Chen, et al., 2021).
While the potential for reduced drainage due to higher low tide levels has been recognised (Khojasteh, Glamore, et al., 2021), the implications of changing tidal dynamics on floodplain drainage have received limited attention and are yet to be quantified. Despite the experience of low-lying areas in the Netherlands (Hoeksema, 2007) and Indonesia (Wahyudi et al., 2019), for example, where excess surface and groundwater must be pumped to the receiving waters, inundation due to SLR has been regarded as a consequence of higher peak sea levels (Holleman & Stacey, 2014) rather than a lack of drainage. Consequently, the majority of research has focused on the potential impacts of SLR on the extent and frequency of extreme coastal storms and flooding (Bosello & De Cian, 2014; Vitousek et al., 2017), groundwater emergence (Hoover et al., 2017; Manda et al., 2017; Wake et al., 2019) and increased nuisance (“sunny day”) flooding (B. S. Hague et al., 2020; Hanslow et al., 2019; Karegar et al., 2017). Yet drainage infrastructure is crucial for the effective management of these intermittent events. Indeed, in urban and industrial environments, constructed drainage systems are primarily designed to mitigate flood risk (ASCE, 1992), although they play a critical role in maintaining public health and amenity (Barbosa et al., 2012; Gaffield et al., 2003; Vlotman et al., 2020) and optimizing agricultural productivity (Cavazza & Pisa, 1988; Hurst et al., 2004).
A typical floodplain drainage scheme consists of a series of interconnected open channels or piped culverts which allow surface and groundwater to drain under gravity and ultimately discharge into the adjacent waterway. Floodgates preclude the flow of tidal waters from the estuary to the floodplain, only permitting discharge from the floodplain catchments when sufficient positive hydraulic head is provided, that is, when water levels in the catchment drains are higher than those in the estuary (Giannico & Souder, 2004). At any point within an estuary, the availability of a positive hydraulic head is influenced by the catchment runoff and hydraulic characteristics of the drainage system upstream of the floodgates and the tidal water elevation downstream of the floodgates. The tidal water levels are characterized by the amplitude (tidal range) and shape (tidal duration asymmetry) of the tidal wave, which may be distorted by the effects of friction, convergence, reflection and inertia (van Rijn, 2011) and is subject to changes to the geometry of an estuary. Additionally, SLR has the potential to modify the water depth, width convergence, floodplain connectivity or entrance conditions of an estuary, which, in turn, can affect the propagation of tidal waves along an estuary (Haigh et al., 2020; Khojasteh, Glamore, et al., 2021; Talke & Jay, 2020). Any changes to the tidal water levels and/or duration can influence the time available for drainage of the estuarine catchments, which may have significant impacts on the estuarine environment, including current land use and management.
To assess how the tidal regime may influence the drainage of estuarine floodplains, and particularly the potential impact of changing tidal regimes under SLR, this study introduces the concept of the drainage window. The drainage window describes the relationship between hydraulic head and the time available for the gravity discharge of floodplain catchments based on local tide characteristics. The drainage window is calculated and applied at two different estuaries in south-east Australia to highlight how SLR may affect floodplain drainage. The influence of present-day and future SLR hydrodynamic regimes on the drainage window is discussed in relation to reduced catchment drainage and potential impacts on existing land management practices.
2 Methodology
2.1 Defining the Drainage Window
Within coastal and estuarine drainage systems, the drainage window is the portion of the tidal cycle when a positive hydraulic head is available to facilitate gravity discharge to the receiving waters at a selected elevation (Figure 1). This describes the temporal period provided both by the tide (at a nominated water level) and to the drainage catchment (at the same topographic level). Under wet weather and flood conditions, the drainage window will vary dynamically, with differential water levels between the estuary and floodplain drainage system subject to local and regional rainfall distribution and the diverse hydrologic and hydraulic responses of catchments throughout both the estuary and upstream river system. Conversely, during non-flood, or dry weather periods, and in the absence of any significant catchment or river inflows that may otherwise affect the hydraulic gradient between the drainage channels and estuary, the drainage window at a given site is primarily controlled by daily tidal conditions. Floodplain drainage systems typically include numerous minor and high-level outlets located above or within the upper portion of the tidal range, in which case the drainage window will also be affected by the invert level of the outlet. However, throughout the lower lying floodplain areas, the primary floodgates servicing the main drainage systems are located at or below the lowest tidal levels to maximize the opportunity for discharge (Ruprecht et al., 2018). Thus, as indicated in Figure 1a, the drainage window would be restricted to the falling tide, with floodgates precluding discharge as a negative hydraulic gradient develops during the rising tide. In these circumstances, the drainage window can be simply defined as the height above the low tide. Thus, assessing the drainage window at the low-lying floodgates under dry weather conditions provides a benchmark to identify the relative opportunity for flows from different floodplain catchments to discharge to the low tide, with increased potential for waterlogging and prolonged inundation to develop when drainage is persistently limited.
When calculated incrementally over the full tidal range, the results can be graphed as a drainage window duration-elevation curve (conceptualized in Figure 1c) representing drainage conditions specific to each floodgate. The elevation axis of the drainage window duration-elevation curve can then be compared to topographic levels (using a hypsometric curve and mapping as described in Section 2.5) to identify areas vulnerable to reduced drainage. This technique may be equally applied to identify critical levels and to assess storage capacity within drainage infrastructure.
The drainage window duration-elevation curve may vary throughout an estuary as the hydraulic head would be affected by any changes in the tidal range and the time available for discharge is dependent on the duration of the falling tide (tidal duration asymmetry). Comparing the drainage window duration-elevation curve for various catchments can provide an indication of the relative drainage risk throughout an estuary. Additionally, estimating the drainage window under varying hydrodynamic or catchment conditions can provide insights into how natural and anthropogenic changes may impact the drainage of coastal and estuarine floodplains. This study uses SLR as an example, as it is expected that changing water levels would affect the drainage window throughout the full tidal range, with the extent of these changes reflecting varying hydrodynamic characteristics of the estuary. For instance, where a rise in water level results in dampening of the tidal range (Figure 2a), the reduction in the drainage window would be greater at elevations below, and less at elevations above, the mid-tide level. Under either existing or future conditions, dampening of the tidal range would enhance the opportunity for discharge to levels above the mid-tide height compared to tidal amplification (Figure 2b), while reducing the drainage window available at lower levels. Dampening of the tidal range is typically experienced where the effects of friction dominate the tidal wave energy or there is an expansion in the area of flow, for example, where low-lying land is inundated or channel banks diverge (Khojasteh, Glamore, et al., 2021; Talke & Jay, 2020). Conversely, tides may be amplified by a gradual contraction in the width or depth of an estuary (respectively termed funneling and shoaling) or by reflection or resonance of the tidal wave (Khojasteh, Glamore, et al., 2021; Talke & Jay, 2020). The net effect on estuarine water levels will depend on the relative impact of each of these influences (Friedrichs, 2010).
Where one-way floodgates have been installed at or below the low tide level, drainage during the rising tide would be precluded during dry weather conditions. A reduction in the duration of the rising tide (positive tidal duration asymmetry as indicated in Figure 2c) would therefore increase the drainage window over all water levels compared to areas experiencing shorter falling tides (negative tidal duration asymmetry as indicated in Figure 2d). Within estuaries, water level or tidal duration asymmetry is often associated with compound overtides caused by changes in energy associated with friction and channel convergence (L. Guo et al., 2015). Tidal duration asymmetry does not necessarily align with tidal current asymmetry (Gallo & Vinzon, 2005), although many of the forcing mechanisms can be similar as a shorter rising tide may lead to stronger flood currents in the absence of significant fluvial discharges (L. Guo et al., 2019). Positive tidal duration asymmetry (flood dominance) is typically encountered when friction is reduced as the depth of flow increases (Friedrichs & Aubrey, 1994), often where estuaries feature shallow inlet systems or large inter-tidal storage (W. Zhang et al., 2018). It may also result from increasing water levels due to SLR. Conversely, where new or existing inter-tidal or overbank areas are activated by the rising tide, or by SLR for example, increasing friction may reduce the available drainage window by inducing negative tidal duration asymmetry.
2.2 Study Sites
The estuaries of the Hastings and Clarence Rivers were selected to test the drainage window concept across multiple catchments within different estuaries, as these two systems provide insights into how the drainage window will respond to varying tidal range, tidal duration and estuary geometry under different SLR scenarios. The Hastings (Figure 3) and Clarence (Figure 4) Rivers are located in north-east NSW, Australia. Each river has a shallow estuary with a trained entrance that has been stabilized to prevent migration and permit regular exchange of the semi-diurnal tide. The average offshore mean tidal range increases from 1.95 m at the Hastings River, north along the coast to 2.04 m at the Clarence River (Couriel et al., 2012). The Clarence River is the largest coastal river in NSW, with the estuarine section reaching 110 km inland and incorporating extensive intertidal areas. These features provide an opportunity to examine the effect of varying tidal characteristics on the drainage window compared to the Hastings River estuary, where the main arm is only 36 km long and the variation in the tidal range is 37% of that experienced in the Clarence River estuary (Couriel et al., 2012). Characteristics of the studied estuaries are presented in Table 1.
Estuary | Catchment areaa (km2) | Estuary areaa (km2) | Estuary volumea (ML) | Estuary lengtha (km) | Average deptha (m) | Tidal rangeb (m) |
---|---|---|---|---|---|---|
Hastings River | 3,659 | 30 | 52,690 | 36c | 1.9 | 1.157–1.668 |
(Wauchope to Port Macquarie) | (0.8%) | |||||
Clarence River | 22,055 | 132 | 283,000 | 110d | 2.2 | 0.477–1.822 |
(Grafton to Yamba) | (0.6%) |
- a Environment NSW (2020).
- b Spatial variation in tidal plane range calculated by subtracting the Indian Spring Low Water from the High High Water Solstice Springs (HHWSS–ISLW) (Couriel et al., 2012).
- c Tidal limit surveyed in 1998 (Allsop, 2006).
- d Tide stopped by rocky rapids at Copmanhurst, surveyed 1997 (Allsop, 2006).
2.3 Hydrodynamic Modeling and Water Level Data
Detailed hydrodynamic models of the estuarine sections of the Clarence and Hastings Rivers were developed using the RMA-2 suite of models (developed by Resource Modeling Associates) to generate long-term, continuous water level data that can be used to determine the statistical distribution of the drainage window. RMA-2 has been widely used to represent tidal estuaries (Elmoustafa, 2017; Hottinger, 2019; Proudfoot et al., 2018). The model solves depth-averaged, shallow water wave equations using the Reynolds' form of the Navier-Stokes equation for turbulent flows to calculate water levels and flow velocities at each node of a flexible, two-dimensional mesh (King, 2015).
The model development, calibration, and verification are detailed by (Harrison et al., 2022a) for the Clarence River and (Harrison et al., 2022b) for the Hastings River. In summary, the RMA-2 finite element mesh was varied to represent the irregular configuration of each estuary, providing higher resolution at locations with more complex energy transitions such as lagoon entrances, junctions, and bends, as indicated in Figures 3 and 4. The channel cross-sections become more regular in the upper reaches of each estuary, which were modeled using one-dimensional elements. Model bathymetry was obtained from detailed spatial surveys undertaken between 2014 and 2020, with the most recent data given preference. All levels are relative to the Australian Height Datum (AHD), with 0.0 m AHD representative of the average oceanic mean sea level around the Australian coast.
Upstream and downstream boundary conditions were defined by gauged catchment inflows and oceanic tide levels respectively. Long-term water level and short-term flow gauges throughout each catchment were also used for model calibration, which was undertaken by adjusting the Manning's “n” roughness coefficients, with adopted values varying from 0.020 to 0.023 in the main channels, up to 0.045 in tributaries. The models' ability to represent a range of tidal conditions throughout each estuary was then verified by simulating both “wet” (2013) and “dry” (2019) rainfall years, the selection of which was based on historic rainfall records. The water level and flow gauge locations used for boundary conditions, calibration, and verification of the model are indicated in Figures 3 and 4, with the historical variability in the recorded flow data at the upstream boundaries presented in the Supporting Information and summarized in Table 2. This data indicates median catchment inflows for the representative “dry” year of 2019 were no more than 20% of the long-term median flow rates (January 2000 to December 2019). The total annual rainfall for 2019 was 481 mm in the Hastings catchment (represented by gauge 207004, Figure 3) and 398 mm in the Clarence catchment (gauge 204007, Figure 4), compared to long-term averages of 1,124 mm and 1,293 mm respectively (January 2000 to December 2019). The boundary conditions defined by the representative “dry” year were adopted for the drainage window analysis to mitigate the impact of catchment hydrology and to isolate, as far as possible, the effect of SLR on the drainage window during non-flood periods. Under these conditions, the mean annual low tide (modeled) lies within the range of operation (invert level to obvert level) of the floodgates servicing the major drainage systems. Full details of the floodgate levels, culvert dimensions and low tide ranges for each of the major drainage systems within the studied catchments are provided in the Supporting Information.
Exceedance | |||||
---|---|---|---|---|---|
River flow (ML/d)a | 0% | 25% | 50% | 75% | 100% |
Clarence River (gauge 204007) | |||||
2000–2019 | 10 | 775 | 1,782 | 4,513 | 1,132,954 |
2019 (dry year) | 10 | 198 | 372 | 657 | 2005 |
Hastings River (gauge 207004) | |||||
2000–2019 | 0 | 163 | 385 | 971 | 239,828 |
2019 (dry year) | 0 | 17 | 51 | 94 | 349 |
Wilson River (gauge 207014)b | |||||
2000–2019 | 0 | 23 | 81 | 241 | 98,041 |
2019 (dry year) | 0 | 0 | 3 | 11 | 73 |
- a Sourced from WaterNSW for 1 January 2000 to 31 December 2019. Refer to Supporting Information S1.
- b Wilson River is a tributary of the Hastings River, refer to Figure 3.
Varying downstream boundary conditions were applied to reflect the present-day, near- and far-future SLR scenarios (refer to Section 2.4). No changes were made to the catchment inflows. Water levels were extracted at hourly timesteps at the main drainage discharge locations for each catchment throughout both estuaries (located between 5 and 29 km along the Hastings River, and between 5 and 70 km along the Clarence River, with major floodgate locations in the Supporting Information) and used to determine the drainage window (the difference in time between each nominated level and the subsequent low tide) at 0.1 m increments in elevation.
The results were first assessed on a catchment basis to identify local drainage conditions and how they may be impacted by SLR. As chronic conditions are established by persistent rather than intermittent exposure, an analysis of the mean annual drainage window was undertaken to identify the underlying drainage conditions at each discharge location. Thus the maximum drainage window represents the maximum (mean) annual duration of the falling tide, with the minimum corresponding water level identified as the critical elevation below which the catchment would experience a persistent reduction in the available drainage window. Similarly, the zero value for the mean annual drainage window is a conservative representation of the lowest annual low tide, as the mean annual low tide (to which the catchment can consistently drain) would normally be higher. Progressive longitudinal changes in the mean annual drainage window at key drainage outlets along each estuary were then compared to changes in the tide duration asymmetry and the tidal range to identify the extent to which these factors may contribute to drainage risk.
For the purposes of this study, the calculation of tidal duration asymmetry and tidal range has been based on the same modeled water levels used to determine the drainage windows. Tidal duration asymmetry was calculated as the duration of the falling tide compared to the total tidal cycle averaged over the annual time series. This is an annual mean interpretation of tidal skewness as presented by (Nidzieko, 2010), whereby a positive tidal asymmetry (positive skewness) is indicative of a longer falling tide (Song et al., 2011; W. Zhang et al., 2018). The tidal range was represented by the difference between the maximum and minimum water levels generated by the model over the annual time series.
2.4 Sea Level Rise Scenarios
The impact of SLR on the estuarine water levels was modeled by adjusting the downstream tidal boundary condition to reflect near-future (NF) and far-future (FF) sea levels. Locally adopted SLR benchmarks of +0.4 m by 2050 and +0.9 m by 2100, relative to the mean sea level (MSL) of 1996, were applied (Glamore et al., 2016). These values represent the median for the representative concentration pathway (RCP) 8.5 scenario (Pachauri et al., 2014) and are the most up to date values specific to the NSW coastline consistent with the Shared Socioeconomic Pathway, SSP5 (Masson-Delmotte et al., 2021). To account for SLR that has occurred between 1996 and 2020, downstream tidal water levels were increased by +4.5 mm/year, as per White et al. (2014), so that all water levels applied in the hydrodynamic models are relative to 2020, nominally the present-day. Values for mean sea level applied to the downstream boundaries of each hydrodynamic model for the near- and far-future cases, relative to the present, are presented in Table 3.
NF (2050) | FF (2100) | |
---|---|---|
RCP 8.5—median SLR relative to MSL 1996 | +0.27 m | +0.78 m |
SLR from 1996 to 2020 @ 4.5 mm/year | +0.11 m | +0.11 m |
Adopted SLR relative to PD (2020) | +0.16 m | +0.67 m |
- SLR, sea level rise.
2.5 Topographic Data
By adopting the same vertical datum for both topographic and water levels, the drainage window analysis can be used to provide an indication of the vulnerability of floodplain catchments to reduced drainage. To this end, one-metre resolution digital elevation models (DEMs) were sourced from the National Elevation Data Framework spatial data set (Geoscience Australia, 2020) to represent the catchment topography for each estuary. The data is reported to have an accuracy of 0.3 m in the vertical direction and 0.8 m horizontal.
The QGIS geographic information system was used to process the DEMs. Discrete catchment areas for each of the major drainage systems were defined based on the floodplain topography, including consideration of the connectivity of watercourses, drains and major floodplain infrastructure.
The hypsometric curve function in QGIS was used to plot the cumulative area against 0.1 m increments in elevation for each catchment to enable a direct comparison between the local topography and critical drainage levels. This 0.1 m increment is representative of the uncertainty in the calibration of water levels in the hydrodynamic models and must be considered in addition to the accuracy of the DEM. The topographic extent to which changes in the drainage window may impact the floodplain catchments, as mapped in Section 3, should therefore only be considered indicative.
QGIS was also used to map the topographic levels corresponding to the water levels that represent the zero and maximum mean drainage window durations, as well as that representing a 50% reduction in the maximum. During dry conditions, flows within the catchment drainage channels would be isolated from the tidal perturbations of the main estuary by the floodgates and low flow velocities would not incur significant head losses, so while the assumption of a static transfer of water levels from the estuary to the drainage catchment is a simplified approach, it is considered suitable to provide an indication of the extent of the area that would be affected by reduced drainage within each catchment.
3 Results
3.1 Drainage Window Analysis of Exemplar Catchment
An analysis of the drainage window under present-day, near- and far-future scenarios is illustrated in Figure 5 for the western side of Woodford Island. Woodford Island is situated between 34 and 55 km upstream of the mouth of the Clarence River (Figure 4), east of a large intertidal lagoon known as The Broadwater. Natural watercourses have been adopted, modified, and supplemented by constructed channels to improve floodplain drainage (Figure 5a) and enable discharge to the present-day lowest tide at −0.3 m AHD (mean zero drainage window for simulated dry year). Floodgates and levees around the island perimeter (Figures 5a and 5b) would protect against the highest annual water levels predicted under both present-day (maximum predicted water level of 0.84 m AHD) and far-future (maximum 1.61 m AHD) dry year scenarios. Currently, less than 5 km2 of the 37 km2 catchment would be unable to drain over the maximum drainage window of 6.5 hr (Figures 5c and 5d). The capacity of the catchment drainage channels is indicated by the level of the drainage window at the top of bank in Figures 5e and 5f). In the far-future scenario, the low tide (zero drainage window) would have at a minimum level of +0.3 m AHD. This would render 23% of the existing drainage channels ineffective, with a standing water level at the top of bank (100% reduction in drainage window). The drainage window for 59% of the channels (all drainage infrastructure below 0.7 m AHD) would be reduced between 50% and 100% (Figure 5g). Thus, despite an apparently strong degree of protection from inundation by high water levels, the area affected by a reduced drainage window has the potential to cause extensive waterlogging throughout the catchment in the far-future.
The floodplain extent that would be directly affected by a reduced drainage window is presented in Table 4 and shown in Figure 6 for the Hastings River estuary and in Figure 7 for the Clarence River estuaries under present-day, near- and far -future scenarios. Comparing these results with Figure 5 indicates that there would be extensive waterlogging due to reduced drainage throughout each estuary. Currently, as indicated in Table 4, the Hastings River, and all but 2 ha of the Clarence River's estuarine floodplains, discharge freely to the low tide at some stage of the tidal range. However, under the far-future scenario, SLR would increase the area of impeded drainage by over 70% in both estuaries. Unless a pumped discharge scheme was implemented, 2,499 ha of the Clarence River estuarine floodplain would be unable to drain, with low-lying backswamp and lagoon foreshore areas identified as being particularly susceptible to reduced drainage.
Area (ha) with drainage window limiteda by: | ||||
---|---|---|---|---|
Estuary | Scenario | ≥100% (no drainage window) | 50% | 0% |
Clarence | Present-day | 2 | 896 | 20,100 |
Near-future | 6 | 2,948 | 23,635 | |
Far-future | 2,499 | 15,202 | 34,474 | |
Hastings | Present-day | 0 | 132 | 8,480 |
Near-future | 0 | 124 | 10,898 | |
Far-future | 124 | 3,371 | 15,913 |
- a when compared to drainage window achieved over full duration of the falling tide.
3.2 Variation in the Drainage Response Along an Estuary and Under SLR
Under both present-day (Figures 8a and 8b)) and far-future (Figures 8c and 8d)) conditions, the drainage window varies in response to changes in the tidal range and the tidal duration asymmetry as the tide propagates along the Clarence and Hastings Rivers. In the lower reaches of both estuaries, a deltaic network of anabranches and shoals create shallow water conditions that enhance energy dissipation, contributing to tidal dampening and an increase in the duration of the falling tide. The effects of large flow bifurcations are noticeable at the Maria River (km 9.3) in the Hastings River estuary and at The Broadwater (km 29.2) in the Clarence River estuary. At these junctions, increasing hydraulic losses at higher water levels slow the propagation of the rising tide and reduce the duration of the falling tide, resulting in a corresponding reduction in the available drainage window. This effect is particularly pronounced around The Broadwater, where low-lying wetlands provide extensive intertidal storage capacity. Continuing upstream, convergence effects tend to amplify the tidal range in the upper reaches of each estuary, where the tidal wave is confined within the main channel and, as shown in Figures 8e and 8f), both estuaries exhibit a tendency for a progressive extension in the duration of the falling tide.
In the main arm of the Hastings River estuary, changes to the tidal characteristics are presently limited, with the effects of channel convergence approximately balanced by frictional losses. For any nominated water level, the drainage window does not vary by more than 0.5 hr throughout the estuary. Similarly, there is minimal variation in the tidal range, falling tide duration or drainage window under the future SLR scenarios (Figure 8c). The results are similar to those that would be achieved by the static addition of +0.67 m SLR to present-day water levels, although in the far-future scenario, a lengthening falling tide duration (Figure 8e) coupled with minor tidal range amplification (Figure 8g), would slightly reduce the influence of SLR on the drainage window. Throughout the estuary, gravity discharge would currently be available to a minimum level of −0.6 m AHD (Figure 8a), increasing to 0.0 m AHD in the far-future (Figure 8c).
Both the tidal duration asymmetry (Figure 8f) and tidal range (Figure 8h) are more varied along the length of the Clarence River estuary. This can be largely attributed to energy losses associated with a complex network of anabranches, channels and shoals, and the diversion of flows into extensive shallow lagoon areas. In contrast to the relatively homogeneity of the response in the Hastings River (Figure 9a), comparing the drainage window (Figure 9b) of a catchment near the mouth of the Clarence River (at km 4.8, as indicated in Figure 4, this is most representative of undistorted tidal conditions) with one near the point of maximum tidal distortion (km 29.2) reveals the longer falling tide would increase the upstream drainage window by up to 0.5 hr in the present-day and 0.3 hr in the far-future scenario. Under current conditions, this would be augmented by the effects of tidal dampening above the mid tide level of 0.2 m AHD. Below the mid tide level, the drainage window would be reduced by up to 1.9 hr at a level of −0.1 m AHD. Higher water levels under future SLR scenarios would reduce the degree of tidal dampening and the maximum reduction in the drainage window would be limited to 1.5 hr at a level of 0.5 m AHD. However, the most substantial change in both estuaries is the reduction in the drainage window resulting from an elevated tidal range under SLR.
4 Discussion
To date, studies regarding the potential impacts of SLR on low-lying floodplains have primarily focused on the increased risk of intermittent flood and storm inundation associated with altered high tide levels. In contrast, the reduction in the drainage window predicted in this study highlights the chronic pressures likely to affect floodplain drainage systems. The actual impact realised by a reduced drainage window will depend on the local drainage efficiency, the volume of storage available within the catchment and how much water needs to be discharged, whether this is from excess irrigation, wastewater, intercepted groundwater, or rainfall runoff. While this study has focused on chronic drainage conditions by investigating the impact of SLR during a period of relatively dry weather, reduced drainage during wet weather and flood conditions may have an even greater impact on coastal land management. The effects of river discharge on tidal propagation are highly dynamic (Cai et al., 2019; L. Guo et al., 2019). Increasing river discharge can not only raise water levels, but also increase tidal wave deformations (Godin, 1985; L. Guo et al., 2015), putting added pressure on the drainage window. Typically, higher river flow will dampen the tidal range (Díez-Minguito et al., 2012; L. Guo et al., 2015) although at various flow rates this effect may simultaneously amplify it in lower reaches of the river (Dykstra et al., 2022). Flood studies typically isolate extreme events, limiting their assessments to a number of days, or even hours, before and after the flood peak (Helaire et al., 2020; Hsiao et al., 2021; P. M. Orton et al., 2020). However, where the flood recession is impacted by tidal conditions, a reduced drainage window is likely to substantially prolong the recession period. The results of this study highlight the need for further investigation into the potential for the extended flood recession period from a rainfall event to coincide with the onset of a subsequent event(s), leading to extensive prolonged inundation and profound implications for existing land uses.
It has been suggested that the flood hazard characteristics of many estuarine systems could be aggregated into coastal, transitional, and fluvial regions, with different sensitivities to changing climate conditions in each (Helaire et al., 2020). Similarly, many estuaries are likely to exhibit zones of varying drainage hazard, as exemplified by the results for the Clarence River herein. High risk drainage areas (those with short drainage windows) are associated with tidal dampening or positive water level asymmetry. Tidal dampening is commonly associated with longer estuaries, estuaries which are prismatic or weakly converging, or those with restricted entrances (Khojasteh, Chen, et al., 2021). Areas with extensive intertidal flats are also susceptible to tidal dampening (Du et al., 2018; Lee et al., 2017). In these areas, typified by shallow coastal lagoons and backswamps, the reduction in the drainage window due to tidal dampening may be exacerbated by a reduction in the duration of the falling tide. Variations in the drainage window throughout the Clarence Estuary are more strongly affected by the impact of changes to the tidal range than tidal duration asymmetry, which is reflected in the fact that the principal astronomic constituents (M2, S2, K1, and O1) account for over 95% of the annual average tidal range measured throughout the estuary (Couriel et al., 2012). Conversely, the impact of tidal duration asymmetry is more substantial in the Hastings Estuary, where the overtides contribute to over 10% of the total tidal amplitude (Couriel et al., 2012). These impacts remain relatively minor as the Hastings Estuary was found to display a comparatively static response to the tide. This comparison, however, highlights significant potential to identify areas that are particularly susceptible to a reduced drainage window as a result of tidal duration asymmetry by examining the generation of overtides and compound tides.
The modeling undertaken for this study does not address uncertainties around anthropogenic, geomorphic, or vegetative adaptations to SLR. As flood and drainage conditions worsen, it is highly likely that there will be trade-offs between protecting reclaimed lands and retreating from them that will further impact the hydrodynamic response of the estuary. Additionally, a linear increase in oceanic sea level at the downstream boundary of the models has been assumed, with present-day catchment inflows used at the upstream boundary for all model scenarios. All of these variables are likely to result in complex feedback loops with dynamic impacts on the tidal range. However, despite these limitations, the modeling highlights that SLR is likely to result in prolonged periods of reduced drainage that are likely to lead to higher groundwater levels, soil waterlogging, and the permanent inundation of low-lying areas.
Varying responses to changes in water levels may redefine which areas within the estuary are more adversely affected by limited drainage conditions. For example, in many highly developed areas, such as the San Francisco Bay (Holleman & Stacey, 2014) and the Chesapeake and Delaware Bays (Lee et al., 2017), shoreline protection works have channelized tidal flows, leading to an amplification of the tidal range. Holleman and Stacey (2014) note that concerns have been raised that further reinforcement of the shoreline for flood protection from rising sea levels may increase tidal amplification, with higher peak water levels increasing the associated flood risk to adjacent areas, as has occurred following tidal-flat reclamation along the Shanghai coast of China (M. Zhang et al., 2021). Conversely, numerous studies have highlighted the role of energy attenuation for storm protection, examining opportunities to reduce channel depths and increase shallow wetland areas in Jamaica Bay, New York (Philip M. Orton et al., 2015), or install artificial sandbanks in the Elbe River Estuary (Ohle et al., 2017; von Storch et al., 2008) for example, Dampening of the tidal range by facilitating the inundation of low-lying areas has been postulated as an alternative mitigation strategy for future high tide inundation in the Chesapeake and Delaware Bays (Lee et al., 2017) and the use of hybrid flood defense systems incorporating restored tidal marshes is gaining traction (Smolders et al., 2020; Stark et al., 2016). The results presented in this study indicate that tidal attenuation strategies such as these may impede drainage and increase chronic inundation and waterlogging from rising sea levels, s highlighting that consideration of the drainage window may help to provide a holistic assessment of the impacts of changes to water levels through the whole tidal range. These changes are not limited to SLR and include natural and anthropogenic activities such as changes to river flow (Jalón-Rojas et al., 2018), sedimentation (Talke & Jay, 2020), dredging (Chant et al., 2018), channel realignment or armoring (W. Guo et al., 2018) as well as land reclamation or wetland restoration (Holleman & Stacey, 2014).
As drainage decreases, numerous floodplain catchments will be faced with economic pressures to protect or preserve existing land use. Historically, the response to these pressures has involved the construction of hard engineering structures such as levees, dykes, seawalls, pumps, and diversion channels to defend vulnerable areas from flooding and/or promote drainage (Day & Templet, 1989). However, the construction, operation, and maintenance of this infrastructure is only viable if it is offset by societal and/or economic returns, such as in the Netherlands (Xu & Blussé, 2019). Consequently, pumped systems are more typically implemented where periodic usage can augment gravity discharge, for example, in parts of Australia (Yang, 2008), the USA (Lang et al., 2010) and Asia (Marfai & King, 2008). The future expansion of pumped discharge systems would, however, only be economically justifiable where there are adequate commercial returns and may be complicated by environmental issues such as land subsidence (Nicholls, 2015; Talke & Jay, 2020) or acid sulfate soils (Dawson et al., 2010).
Where gravity systems remain the preferred option for drainage management, additional attenuating storage may be required to offset the reduction in drainage capacity. The relationship between the local topography and drainage window for a given catchment can be used to identify areas with sufficient capacity within the existing landscape to provide effective attenuation. Examining variations in the drainage window throughout an estuary and comparing it to catchment topography provides a means of identifying floodplain areas at risk from reduced drainage. As such, the drainage window analysis may complement topographic studies when considering future land use and management options and is particularly beneficial in examining future SLR scenarios. Comparing the hypsometric curve to the anticipated change in water levels resulting from SLR may indicate if (and when) a catchment is likely to experience a rapid increase in vulnerability to inundation (Kane et al., 2015). Extending this analysis to encompass the change in drainage window, as indicated in Figures 5c and 5d, would also indicate the susceptibility of a local catchment to drainage risks. In high-risk drainage areas, there may be substantial merit in considering alternative nature-based solutions, including wetland restoration projects which have considerable co-benefits, including improved water quality and ecological values as well as significant potential for carbon sequestration (Gulliver et al., 2020; Raw et al., 2021; Sheehan et al., 2019). In some circumstances, the removal of tidal barriers to low-lying estuarine floodplains may be used as a sacrificial measure to increase flood protection elsewhere in the estuary while creating highly valued coastal and estuarine ecosystems using nature-based solutions to accommodate SLR. This prospect is particularly relevant with the emergence of a global blue carbon market that may incentivize tidal inundation of poorly drained land over other low return agricultural production measures.
5 Conclusion
This study has introduced a “drainage window” concept to quantify and compare the time available for the effective drainage of estuarine catchments under present-day and future SLR conditions. As a proof of concept, hydrodynamic models of the Hastings and Clarence Rivers' estuaries in Australia were used to simulate tidal responses to varying oceanic water levels under current and future SLR scenarios. Modeling results indicate that the drainage window responds dynamically to changes in tidal characteristics as the tide propagates within an estuary. Tidal dampening and flood dominant tidal asymmetry were highlighted as key contributors to a reduced drainage window. Understanding the interactions between tidal range and tidal asymmetry within an estuary may help quantify potential reductions in the drainage window. This may be particularly important in long prismatic or weakly converging estuaries as they may become increasingly vulnerable to reduced drainage following SLR (Khojasteh, Chen, et al., 2021).
While previous studies have examined the impact of SLR on acute flooding events associated with higher high tides (Ben S. Hague & Taylor, 2021; Hino et al., 2019; Moftakhari et al., 2018), this research highlights chronic impacts that occur across the full tidal range. In direct contrast to flooding risks, which will be exacerbated by increased tidal amplification, reduced drainage capacity is likely to be more pronounced in areas subject to increased tidal dampening. A thorough assessment of the risks posed by SLR at all water levels is therefore required as the reduction in the drainage window could result in changes to land use and broader management policy. This may provide opportunities for adaptation using nature-based solutions given that shallow coastal lagoon and backswamp areas are particularly susceptible to reduced drainage.
Acknowledgments
This paper has been extended and improved with thanks to considered and generous input from Steven Dykstra, two anonymous reviewers and the editorial staff at AGU publishing. The development of the drainage window concept has been aided by discussions with many staff and students at the UNSW Water Research Laboratory, with particular credit to Grantley Smith, Alice Harrison, Jamie Ruprecht, Priom Rahman and Toby Tucker. The authors would also like to extend our gratitude to Priom Rahman for generating the water level data and to Anna Blacka from UNSW Sydney for her assistance with the preparation of figures. Katrina Waddington is supported by an Australian Government Research Training Program Scholarshiop. Danial Khojasteh is supported by a UNSW Scientia PhD Scholarship. Open access publishing facilitated by University of New South Wales, as part of the Wiley - University of New South Wales agreement via the Council of Australian University Librarians.
Conflict of Interest
The authors declare no conflicts of interest relevant to this study.
Open Research
Data Availability Statement
Hourly water level data generated by the RMA-2 model and used in the drainage window analysis for the Hastings and Clarence Rivers is available from Researchgate at https://doi.org/10.13140/RG.2.2.28047.87208/1 (Waddington et al., 2022). Rainfall, water flow and water level data were downloaded from the WaterNSW Water Information Hub Real-time water data (waternsw.com.au), with the data used in this study, as presented in the Supporting Information, sourced for Station 207004 from https://realtimedata.waternsw.com.au/?ppbm=207004&rs&1&rscf_org (WaterNSW, 2020), for Station 204007 from https://realtimedata.waternsw.com.au/?ppbm=204007&rs&1&rscf_org (WaterNSW, 2020), and for Station 207014 from https://realtimedata.waternsw.com.au/?ppbm=207014&rs&1&rscf_org (WaterNSW, 2020). QGIS software can be freely downloaded from Discover QGIS. Digital elevation data was obtained from the National Elevation Data Framework spatial data set Elvis (fsdf.org.au) managed by Geoscience Australia Digital Elevation Data | Geoscience Australia (ga.gov.au).