Future frequencies of coastal floods in Australia: a seamless approach and dataset for visualising local impacts and informing adaptation
Ben S. Hague
A * , Dörte Jakob A , Ebru Kirezci B , David A. Jones A , Ilana L. Cherny B and Scott A. Stephens C
A
B
C
Abstract
The rise of pathways-based approaches to coastal adaptation in Australia has changed user requirements for coastal flood hazard information to support decision-making. This study identifies and addresses three aspects not considered in the existing Australia-specific scientific guidance for planning adaptation to sea-level rise. First, changes in the frequency of present-day extreme sea levels are compared between locations. Second, extreme sea levels are related to impact-based thresholds associated with past flood events. Third, the potential for chronic flooding emerging is assessed. This complements global studies that provide some Australian results on these topics. We survey these to identify the methods most suitable for our application and apply the chosen methods to the reference dataset for monitoring Australian coastal sea-level change. This yields a water-level frequency dataset covering daily to centennial water levels for 37 Australian tide gauges. We analyse the dataset to provide a national picture of how sea-level rise is expected to influence the future frequencies of coastal floods in Australia. For example, 85% of Australian locations expect present-day centennial extremes to occur 30 days per year with less than 1-m sea-level rise. The locations with the largest increases in the future frequency of these extremes have the smallest present-day sea-level extreme magnitudes relative to mean sea level, and lower flood thresholds relative to these extremes. We demonstrate three further potential applications of our dataset and methods using local case studies: impact-based forecasting, climate risk services and identifying the required sea-level rise for adaptation triggers and thresholds to be reached.
Keywords: adaptation pathways, climate risk, coastal flooding, extreme value analysis, hazard assessment, impact-based forecasting, sea-level extremes, sea-level rise.
1.Introduction
The Intergovernmental Panel on Climate Change (IPCC) found that chronic flooding at high tide is ‘the most urgent adaptation challenge’ facing coastal communities worldwide (Cooley et al. 2022). In high-income countries like Australia, such economic impacts of sea-level rise (SLR) are expected to dwarf those of other climate change-related risks, but can be mostly avoided with optimal adaptation (van der Wijst et al. 2023). Future climate risk is a function of hazard, exposure and vulnerability, and how they are affected by adaptive responses (Simpson et al. 2021; O’Neill et al. 2022). Future SLR is expected to further increase both the severity and frequency of coastal floods worldwide (Fox-Kemper et al. 2021), hence affecting the future risk estimations. Information on projected changes of both frequency and severity of floods is therefore needed to inform effective adaptation.
The most recent peer-reviewed coastal flood hazard information specifically produced for Australian decision-makers to support adaptation planning was published nearly a decade ago (McInnes et al. 2015). The focus of this paper was presenting sea level allowances, which are the ‘amount by which something, such as the height of coastal infrastructure, needs to be altered to cope with climate change’ (Hunter 2012). Sea level allowances are used to inform sea level benchmarks that set the SLR being planned for in a region (McInnes et al. 2015). Benchmarks ultimately differ between jurisdictions and from the computed allowances because of other considerations, including risk tolerance (Hirschfeld et al. 2023).
The questions being asked of scientific guidance have evolved as different climate change adaptation planning frameworks have developed from predict-then-act approaches to pathways approaches that enable decision-making under deep uncertainty (Walker et al. 2013; Lempert et al. 2024); whereas predict-then-act focuses on developing a consensus prediction of future SLR, pathways approaches encourage a planner to create a values-based vision of the future and commit to short-term actions within a longer-term decision-making framework (Haasnoot et al. 2013; Lempert et al. 2024). Pathways approaches have become mainstream in Australia and are embedded in state-level adaptation policy and guidelines in Victoria (Department of Energy Environment and Climate Action 2023), South Australia (Local Government Association of South Australia 2020), Western Australia (Department of Planning Lands and Heritage and Western Australian Planning Commission 2019) and Queensland (The Local Government Association of Queensland and Department of Environment and Heritage Protection 2016). To develop and enact these plans, decision-makers need to identify adaptation thresholds, triggers and signals, when these will be reached, and over what timeframes different adaptation measures reduce risk to a tolerable level (Hino et al. 2017; Stephens et al. 2018). These thresholds, triggers and signals are often expressed as a specific local impact occurring at a specific frequency (Department of Energy Environment and Climate Action 2023). For example, community engagement found that inundation of the main highway 5 days per year was an adaptation trigger for Lakes Entrance in southeast Australia (Barnett et al. 2014).
Scientific guidance is needed to identify the SLR increments that lead to possible thresholds, triggers and signals being reached on national (e.g. Paulik et al. 2021), regional (e.g. Hanslow et al. 2018) and local scales (e.g. Hanslow et al. 2023). Projections of SLR are needed to identify possible timings of reaching SLR increments that are associated with specific adaptation measures and decision points (Slangen et al. 2022). These approaches recognise that the most appropriate adaptation strategy may differ between coastal settlements and change through time. For example, Department of Energy Environment and Climate Action (2023) provides a strategic order of consideration for these options (most- to least-preferred): non-intervention, avoid, nature-based methods, accommodate, retreat and protect. Similar strategic orders of consideration are also provided in local-level coastal adaptation plans (Cairns Regional Council 2022). Therefore, scientific guidance is required to assess the relative merits of each measure and how these develop through time at federal, regional and local levels to avoid maladaptation (Schipper 2020; Haasnoot et al. 2021).
This study provides further information to support the evolving needs of Australian adaptation decision-makers, in addition to the existing published sea-level allowances of McInnes et al. (2015). First, we quantify changes in the frequency in extreme sea levels. To date, Australian results for this have only been provided by global studies (Hunter 2012; Wahl et al. 2017; Fox-Kemper et al. 2021; Tebaldi et al. 2021; Rasmussen et al. 2022; Hermans et al. 2023). Second, extreme sea levels are related to impact-based thresholds associated with past flood events using the common flood days metric (e.g. Sweet and Park 2014; Fox-Kemper et al. 2021). Third, we show that our coastal hazard assessment framework can also assess the potential of chronic flood hazards emerging (Moftakhari et al. 2017; Thompson et al. 2019). This is achieved through coupling a ‘mixed distribution’ approach, which seamlessly determines the frequency (and changes thereof) of any water level (Stephens et al. 2018; Ghanbari et al. 2019), with established impact-based threshold methods (e.g. Hague et al. 2019; Moore and Obradovich 2020). This approach delivers seamless insights into changes in how often floods of different severities occur – from nuisance flooding of streets and paths (Hague et al. 2019, 2022) to property damage and deaths (Callaghan and Power 2014). Relevant case studies for climate adaptation, operational climate risk services and impact-based forecasting are explored.
2.A sea-level frequency dataset for Australian tide gauge locations
2.1. Summary of methods used to produce the dataset
We conducted a survey of peer-reviewed methods that have been used over the last decade to quantify changes in extreme sea-level heights and frequencies. We examine their assumptions and limitations to identify those most suitable for our aims: to develop scientific guidance that enables coastal adaptation planning. Section S1 in the Supplementary material provides detailed discussion on this, with reference to McInnes et al. (2015), as the most recent national assessment providing information for policymakers adapting to SLR. As all the methods and data used in this study have been thoroughly described in previous peer-reviewed publications, we only provide a brief synopsis of our approach here. For completeness, further details are provided in section S2 of the Supplementary material.
Tide gauge data are obtained from the Australian National Collection of Homogenised Observations of Relative Sea-level (ANCHORS) dataset, including recently released data to 2022 (Hague et al. 2021) (section S2.1 of the Supplementary material). The ANCHORS dataset is used to estimate the heights of different sea-level frequencies using mixed empirical–Generalised Pareto (GP) distributions for each location (Stephens et al. 2018) (section S2.3 of the Supplementary material). The heights of more frequent water levels are computed directly by percentiles of the daily maxima of the detrended observed sea level timeseries. The heights of rarer water levels (with less than a 20% chance of occurring annually) are estimated by fitting a GP distribution to all daily maxima that exceeded the 99.7th percentile using the fevd function of R package extRemes (ver. 2.0-11, see https://cran.r-project.org/package=extRemes; Gilleland and Katz 2016), with default options and 3-day de-clustering (section S2.3 of the Supplementary material). This GP distribution fits the observed sea levels across Australian tide gauges best. A comparative analysis between different extreme value distributions reveals a systematic bias in GEV and Gumbel distributions’ ability to represent sea levels with more than 10% probability of occurring annually due to interannual variability in the height of annual maxima (Fig. S1 and S2).
Sea-level heights are expressed relative to the tide gauge zero (TGZ) for each site. A link to the offsets between the Australian height datum (AHD) and the TGZ datum used in the latest version of the ANCHORS dataset is provided in section S2.1 of the Supplementary material. Sea-level frequencies are estimated based on a 2022 climate, using exceedances per year (EY) if more frequent than once every 5 years, otherwise annual exceedance probabilities (AEPs). This is consistent with Australian standard practice for tide gauge data (Hague et al. 2021) and riverine and estuarine flood frequency analysis (Ball et al. 2019; Jolly and Green 2019). Conversions between AEPs, EYs and return periods (RPs, also termed average recurrence intervals, ARIs) are provided in section S2.1 of the Supplementary material (Table S1). Both the RPs and TGZ-AHD offsets aid future work and applications (e.g. section 3.3) by users more familiar with AHD and RPs.
2.2. Results
The locations with highest AEP and EY levels are those with large tidal ranges (Hague et al. 2022). These are located predominantly in northern Australia, central Bass Strait and South Australia (Fig. 1a–e). This agrees with previous Australian studies (McInnes et al. 2009; Haigh et al. 2014a, 2014b). This is true for all AEPs and EYs, indicating the dominance of tides on water-level variability at all timescales (Merrifield et al. 2013; Rueda et al. 2017; Woodworth et al. 2019; Ritman et al. 2022). Locations with small tidal ranges have smaller offsets between extreme sea levels and AHD, which is approximately equal to 1966–1968 mean sea level (refer to section S2.1 of the Supplementary material for further discussion on AHD). A smaller increase in SLR at these locations will lead to present-day extremes becoming the mean state (refer to section S2.1 of the Supplementary material for definition of ‘present day’).
Sea levels in metres above Australian height datum (m AHD) corresponding to (a) 1% AEP and (b) 10% AEP, and (c) 1, (d) 10 and (e) 100 EY levels, as well as (f) the EY of the minor flood level (based on levels from Hague et al. 2022). Lighter shading denotes (a–e) extreme sea levels closer to AHD (i.e. approximate mean sea level) or (f) more frequent minor flooding. Locations without defined flood thresholds shown as having 0 flood days in (f).
Fig. 1f shows that minor flooding occurs multiple days per year in some locations while it is much rarer in other locations (e.g. 1% AEP). This reinforces the findings of previous studies that have demonstrated regional variability in the frequency and severity of coastal flooding (Hanslow et al. 2018; Hague et al. 2019, 2022). Minor flooding is generally most frequent along Australia’s east coast, where it occurs at least several days per year. Hence, considering water levels with EYs of more than one (or equivalently, AEPs more than 63%) is required to characterise flood hazards nationally. This finding supports our use of mixed distributions, as these can reconcile infrequent, frequent and very frequent water levels.
3.Assessing and visualising future flood hazards
In this section, we describe how our dataset can be used to estimate changes in the frequency of present-day sea levels and their associated flood impacts under SLR. Global and Australian sea-level studies have typically relied on two assumptions to assess flood hazards under SLR: stationarity of variance and stationarity of impact threshold. A consequence of stationarity of variance and impact threshold being valid is that the height differences between a pair of thresholds (impact- or frequency-based) indicate the SLR that leads to the corresponding change in how often the threshold is exceeded. For example, a 0.1-m difference between the 10 and 1 EY levels translates into a present-day once-per-year level occurring 10 days per year under 0.1-m SLR. Similarly, if there is a 0.2-m difference between the 1% AEP and the minor flood level, then 0.2-m SLR will see the present-day 1% AEP level occur as often as minor flooding does at present. Hence, these offsets are a metric for the potential for chronic flooding under SLR (Taherkhani et al. 2020; Hermans et al. 2023). Section S3 of the Supplementary material provides further information on this approach. Chronic flooding is generally considered to be what occurs at least 30 days per year (e.g. Thompson et al. 2019; Habel et al. 2020; De Leo et al. 2022; Li et al. 2023; Hague and Talke 2024). We provide applications of this approach for national coastal hazard assessments (section 3.1), supporting impact-based forecasting and climate risk information (section 3.2), and enabling climate adaptation pathways planning (section 3.3).
3.1. A national assessment of chronic flood potential under SLR
Fig. 2a–d considers the height differences between 1% (i.e. EY = ~0.01) and 10% (i.e. EY = ~0.10) AEP and 1, 10 and 100 EY levels. These all approximately correspond to 10-fold increases in frequency of these thresholds being exceeded. The offsets between the 1 and 10% AEP levels vary spatially – from 0.03 m at Portland to 0.86 m at Carnarvon (Fig. 2a). The offsets between 1% and 10% AEP levels average 0.19 m across all sites, but the average offsets between 10% AEP and 1 EY, 1 and 10 EY and 10 and 100 EY are 0.16, 0.21 and 0.37 m respectively (Fig. 2b–d).
Differences between (a–e) specified AEP and EY levels and (f) minor flood threshold and 30 EY level, implying required SLR for change in frequency from higher to lower level.
Our analysis shows that present-day flood hotspots may not be the same as future flood hotspots. This can occur either because of differences in the rate at which the frequencies of present-day extremes increase (Fig. 2e) or where flood thresholds (see section S2.2 of the Supplementary material) sit relative to these extremes (Fig. 2f). Fig. 2e shows that the SLR that leads to chronic breaches of the present-day 1% AEP level differs between sites. For example, Ballina, Gold Coast and Port Kembla all expect their present-day 1% AEP water levels to occur 30 days per year with less than 0.4-m SLR. By contrast, 1.2-m SLR at Weipa, Carnarvon and Port Pirie does not lead to present-day 1% AEP events occurring 30 days per year. Overall, at 85% of locations the 1% AEP is likely to be exceeded 30 days per year with less than 1-m SLR. Fig. 2f shows that locations in New South Wales and southeast Queensland expect minor flooding 100 days per year with considerably less SLR than other locations.
The rankings of locations ‘at risk’ of chronic flooding based on the metric in Fig. 2f, differ from those based on the metric in Fig. 2e. For example, Portland ranks 4th of 37 in the smallest offset between 1% AEP and 30 EY levels, but 14th of 24 in the smallest offset between minor flood level and 100 EY. This demonstrates the importance of defining metrics that are relevant to decision-making. Metrics should relate to adaptation thresholds, triggers and signals if being used to inform adaptation to the impacts of SLR – refer to section 3.3 for more discussion on this. Such examples highlight the value of our dataset in its ability to assess changes in the frequency of all present-day water levels of interest and relating these levels to the impacts of past floods.
3.2. Applications for impact-based forecasting and climate risk services
In this subsection, we demonstrate the use of the sea-level frequency dataset together with impact-based thresholds to support impact-based coastal flood warning services. We also discuss how the dataset can be used to provide rapid assessment of changes in likelihood due to observed or projected SLR. This could assist with the delivery of operational climate risk services. We illustrate this using the 8 July 2023 storm surge event (middle photo, Supplementary Fig. S3).
For several days before the 8 July 2023 storm surge event, The Bureau of Meteorology’s coastal water-level forecast system was predicting water levels well above the Highest Astronomical Tide alert level (Fig. 3). This system aggregates astronomical tide predictions, storm surge from an ocean model, and inverse barometric pressure from an atmospheric model with bias correction to predict water levels at tide gauge locations (Taylor and Brassington 2017). The forecast peak of 1.45 m sat between the lowest water level associated with impacts (1.3 m, Hague et al. 2022) and the record water level of 1.64-m TGZ set in 2014 (Supplementary Fig. S3). Hence the likely impacts associated with the forecast water level were somewhat uncertain.
Screenshot from Bureau of Meteorology’s ‘Aggregate Sea Level Viewer’ (Taylor and Brassington 2017) indicating forecast water-levels-based aggregation of components of numerical weather and ocean models for June 2023 storm surge event. Forecast water levels based on 4, 5 and 6 July 12 UTC model runs are shown in light, medium and dark blue respectively, along with astronomical tides (green). The water levels are expressed relative to tide gauge zero (TGZ), with the highest astronomical tide (HAT, upper pink line) used as an alert threshold. This screenshot has been modified for clarity by re-labelling axes and the legend and presenting data with respect to TGZ.
To provide this information to forecasters and emergency management agencies, a DEM-based flood visualisation website Coastal Risk Australia (coastalrisk.com.au) was used to identify areas that sat no more than 1 m above mean sea level, as this was approximately equivalent to the forecast level. The vertical errors associated with the underlying DEM are of order 1 m (Gallant et al. 2011). However, locations inundated in previous coastal flood events (Hague et al. 2022, their table S1) were correctly shown to be flooded at the corresponding height in the visualisation. This validation gave analysts greater confidence in the DEM’s ability to predict the inundation for the imminent storm surge event (Habel et al. 2020; Prakash et al. 2020).
This impact-based assessment provided important context for operational meteorologists and guided the forecast and messaging strategy towards emergency services, media, the public and other stakeholders. This included a joint press release by the Bureau of Meteorology and the Victorian State Emergency Service (VICSES) outlining the expected severe weather (including elevated sea levels), an alerting phone call to water management agencies, and an increased emphasis on coastal flooding during routine and ad hoc radio crosses. Following this, VICSES issued a Storm Tide Advice that was broadcast over social media and the official Victorian Emergency mobile phone app. Storm Tide Advice had been issued for single Victorian locations by VICSES for three past storm surge events. However, this was the first time that multiple locations had been alerted, including the busy Southbank precinct. Impact-based information is more meaningful and relatable to stakeholders than estimated sea levels and frequencies alone. Without impact-based information it is unlikely that operational meteorologists would have had the confidence to pursue this strategy based on forecast or modelled water levels alone. Further, having a threshold that encapsulates impact information makes it easier for forecasters to monitor when coastal flooding may be likely to occur.
Volunteer observers took photographs of coastal flooding in the Melbourne metropolitan area (Fig. 4), which showed that coastal flooding affects more locations in Melbourne than previously thought, even compared to the higher 2014 event where impact reports were more ad hoc. No rainfall occurred at this time and flooding was confined to tidally influenced areas. This gives confidence that impacts were primarily due to high sea level rather than heavy rainfall (pluvial) or high streamflow (fluvial) flood drivers. This information has enabled improvements in delivery of impact-based warning services when similar water levels are forecast in future, by establishing a Coastal Hazard Warning service for Victoria.1 This case study also highlights the value of easy-to-use and publicly available flood mapping tools in conjunction with repositories of past flood impacts to ground-truth the mapped inundation extents (Habel et al. 2020). There would be merit in reinvigorating the Witness King Tides citizen science program for coastal flood monitoring in Australian bays, harbours and estuaries (Watson and Frazer 2009). This would accompany newer initiatives for monitoring beach erosion and shoreline change on open coasts (Harley et al. 2019; Pucino et al. 2021).
Images of flooding in the Port Phillip Bay region during the 8 July 2023 storm surge event, with locations numbered on map. Map background from Open Street Map. Images by authors and Elise Chandler, Stephanie Jacobs, Brad Murphy, Anita Pyne and Thomas Ramage. Reproduced with permission of the photographers.
Changes in flood event likelihood attributable to SLR can be identified in real-time using the principles described in section S3 of the Supplementary material. For example, 0.08-m SLR since 1966 (Hague et al. 2022) means that the 8 July 2023 storm surge reached a level 0.08 m higher than it would have if SLR had not occurred. In other words, instead of reaching a level of 1.45 m (current EY = 0.8), it would have only reached a level of 1.37 m (current EY = 2.5). The ratio of the EYs tells us that SLR has made this event 3.1 times more likely. Similarly, we can conclude that 0.15-m SLR would lead to the impacts observed on 8 July 2023 occurring as frequently as minor impacts do today. This follows from the minor flood level being 0.15 m below the level reached on 8 July 2023. With 0.46-m SLR, these impacts are expected 100 days per year. This follows from the minor flood level being 0.46 m higher than the 100 EY level. Refer to Supplementary Fig. S3 for further examples. These insights could be used to deliver climate and hydrometeorological risk services in context, guiding stakeholders in their approaches for managing flood hazards using familiar events (Harrison et al. 2021; Rasmussen et al. 2022).
3.3. Applications for climate adaptation pathways planning
Our dataset, and more broadly the methods described here, can directly assist coastal managers implementing pathways-based climate adaptation plans. A key feature of modern adaptation planning is the identification of triggers of when a chosen adaptation measure needs to be implemented, and signals to provide advance warning as these decision points approach (Stephens et al. 2018). The Victorian Government guidance for SLR planning in the state indicates that possible coastal flood-related adaptation triggers are infrastructure renewal, failure of nature-based solutions and inundation reaching a specific frequency (Department of Energy Environment and Climate Action 2023). In this context, our dataset and methods can quantify the required SLR (Hermans et al. 2023) for frequency-based triggers, which is needed to develop actionable adaptation plans (e.g. Allison et al. 2023). We demonstrate this for Lakes Entrance, in eastern Victoria, where specific frequency-related adaptation triggers have been proposed based on community consultation (Barnett et al. 2014). Two of these were ‘inundation of the Esplanade more than 5 days per year [that prevents traffic flow]’ and ‘two 1.8 m [AHD] floods in a year.’ It was proposed that reaching the first trigger would lead to increased development controls and preparation for managed relocation, and the retreat to higher ground would have occurred by the second. Both triggers relate to events occurring very frequently, so the SLR required for these triggers to be reached can be determined without extreme value analysis.
The 2 EY level is equivalent to the 99.5th percentile (of daily maxima), which is currently 0.63 m above AHD. This means that the trigger for retreat will be reached with 1.17-m SLR, as that is the difference between the 1.8-m AHD level and the level that currently occurs twice per year (Fig. 5). Once water levels reach the moderate flood level of 1.1-m AHD, the two east–west thoroughfares at Lakes Entrance (including the Esplanade) are closed, causing substantial disruption to business and recreational activities (State Emergency Service 2012). The 98.6th percentile is exceeded on average 5 days per year and is 0.55-m AHD. The offset between the moderate flood level (1.1 m) and the present-day 98.6th percentile is 0.55 m and the SLR that leads to this trigger being reached (Fig. 5). Based on Victorian Government guidelines, it is recommended that decision makers consider 0.5-m SLR being reached by 2070 and 1.2-m SLR at or after 2100 (Department of Energy Environment and Climate Action 2023). This gives the Lakes Entrance community ~50 years to prepare future actions, adapt to present-day and shorter-term increases in flood risk, and at least another 30 years to implement full-scale managed relocation, based on these triggers (Barnett et al. 2014; Stephens et al. 2018). These are best estimates based on average conditions over 2009–2022, following Hague et al. (2023) and the assumptions described in section S3 of the Supplementary material. This shorter more recent period is used as substantial increases in tidal range have occurred, leading to higher high-water levels over recent times at Lakes Entrance than in pre-2009 data (Hague et al. 2023). Future increases in tidal range with SLR, which is expected in many estuaries (Khojasteh et al. 2023), can reduce the SLR that leads to specific changes in flood frequency (Hague and Talke 2024).
Example of how to estimate the SLR leading to frequency-based adaptation triggers at Lakes Entrance (from Barnett et al. 2014), by taking the difference between two water levels (here presented as AHD) with specific frequencies or associated impacts.
This example also demonstrates how this approach can be applied to unprecedented events. The 1.8-m AHD water level considered by Barnett et al. (2014) is higher than the highest-on-record flood at Lakes Entrance of 1.69 m in 1952 (State Emergency Service 2012). It is more than 0.7 m higher than the 1-in-100-year level sea-level estimated by a previous study that quantified coastal extremes at Lakes Entrance (McInnes et al. 2009). Our approach only requires a knowledge of the flood level and its future frequency to determine the SLR that leads to it occurring at this frequency. No estimate of its present frequency is required. Estimates of return periods greater than four times the length of the underlying dataset are generally considered to be of lower confidence (Hunter 2010). The record lengths of quality Australian tide gauge data (e.g. Hague et al. 2021) means it is not possible to estimate events with AEPs less than 0.5% based on this principle. Our method does not suffer from such limitations, making it ideal for assessing the future probabilities of currently unprecedented sea-level extremes.
4.Conclusion
This study has developed a water-level frequency dataset covering daily to centennial water levels for 37 Australian tide gauges. This dataset can be used to address aspects of flood hazards not considered in the existing national SLR guidance for adaptation policymakers. Recent advances in sea-level frequency analysis have been incorporated in the methodology and chronic flood potential has been robustly assessed for the first time. Impact and adaptation trigger information can be easily incorporated, so that changes in extreme sea-level frequencies can be related to the flood hazards they pose. These advances are achieved by combining different elements of recent studies – impact-based thresholds, required SLR and mixed extreme value-empirical distributions. We used mixed distributions to estimate heights of sea levels that span daily, annual and centennial frequencies. We use the vertical distances between these sea levels to estimate the required SLR for specific frequency changes. We relate sea levels to their impacts to inform adaptation planning and impact-based forecasting.
The locations with highest AEP and EY levels are those with large tidal ranges. However, locations with smaller AEP and EY levels tend to see larger increases in frequency of their present-day extremes under SLR. Flooding, even of minor severity, is typically not frequent outside of eastern Australia. However, our analysis shows that present-day flood hotspots may not be the same as future flood hotspots. This change in hotspots can occur either because of differences in the rate at which the frequency of present-day extremes increases, or where flood thresholds sit relative to these extremes. Linking changes in frequency of extremes to adaptation triggers is needed for scientific guidance to enable adaptation pathways planning. It can also support coastal hazard forecasts and services that use impact-based criteria. The SLR required for emergence of frequency-based adaptation triggers can be estimated by taking the difference between two water levels with specific frequencies or associated impacts. This simple principle allows decision-makers to contextualise past, present and future unprecedented events based on lived experience and planning guidelines.
Data availability
The sea-level frequency dataset developed in this paper is provided in the in ES23024_TS2.CSV file and described in section S2 of the Supplementary material. The ANCHORS tide gauge dataset is described in Hague et al. (2021) and was downloaded from https://dx.doi.org/10.25914/6142dff37250b.
Declaration of funding
This work was funded by the Australian Climate Service. The Australian Climate Service provides improved information, data, intelligence and expert advice on climate risks and impacts to support and inform decision-making. The Australian Climate Service is a partnership between The Bureau of Meteorology, CSIRO, the Australian Bureau of Statistics and Geoscience Australia.
Acknowledgements
The authors thank several anonymous reviewers as well as Mandi Thran and Paul Fox-Hughes (Bureau of Meteorology) for their comments on earlier versions of the manuscript. The authors thank Kate Bongiovanni, Elise Chandler, Richard Hammond, Stephanie Jacobs, Brad Murphy, Anita Pyne, Steven Pyne, Thomas Ramage, Andrew Watkins and Allan McRae for providing impact reports of the July 2023 Melbourne region storm surge event. The authors also thank Melanie Gill of the Victorian State Emergency Service for providing details on past issues of Storm Tide Advice for Victoria.
References
Allison A, Stephens S, Blackett P, et al. (2023) Simulating the impacts of an applied dynamic adaptive pathways plan using an agent-based model: a Tauranga City, New Zealand, case study. Journal of Marine Science and Engineering 11(2), 343.
| Crossref | Google Scholar |
Ball J, Babister MK, Nathan R, et al. (2019) Australian rainfall and runoff – a guide to flood estimation. Version 4.2. (Geoscience Australia, Commonwealth of Australia) Available at https://www.arr-software.org/pdfs/ARR_190514_V4.2.pdf
Barnett J, Graham S, Mortreux C, et al. (2014) A local coastal adaptation pathway. Nature Climate Change 4(12), 1103-1108.
| Crossref | Google Scholar |
Cairns Regional Council (2022) Draft strategy for consultation. Our Cairns Coast: Adapting for the Future. (CRC: Cairns, Qld, Australia) Available at https://www.cairns.qld.gov.au/__data/assets/pdf_file/0005/472406/CHAS_Strategy_DRAFT2021.pdf
Callaghan, J, Power S (2014) Major coastal flooding in southeastern Australia 1860–2012, associated deaths and weather systems. Australian Meteorological and Oceanographic Journal 64, 183-213.
| Crossref | Google Scholar |
Cooley S, Schoeman D, Bopp L, et al. (2022) Oceans and coastal ecosystems and their services. In ‘Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change’. (Eds H-O Pörtner, DC Roberts, M Tignor, ES Poloczanska, K Mintenbeck, A Alegría, M Craig, S Langsdorf, S Löschke, V Möller, A Okem, B Rama) pp. 379–550. (Cambridge University Press: Cambridge, UK, and New York, NY, USA) 10.1017/9781009325844.005.379
De Leo F, Talke SA, Orton PM, et al. (2022) The effect of harbor developments on future high-tide flooding in Miami, Florida. Journal of Geophysical Research: Oceans 127(7), e2022JC018496.
| Crossref | Google Scholar |
Department of Energy Environment and Climate Action (2023) Victoria’s Resilient Coast: Adapting for 2100+. Framework and guidelines: a strategic approach to coastal hazard risk management and adaptation. (The State of Victoria) Available at https://www.marineandcoasts.vic.gov.au/__data/assets/pdf_file/0022/662503/Victorias-Resilient-Coast-Guidelines-.pdf
Department of Planning Lands and Heritage and Western Australian Planning Commission (2019) Geraldton Coastal Hazard Risk Management and Adaptation Planning Report. Coastal hazard risk management and adaptation planning guidelines. (City of Greater Geraldton: Geraldton, WA, Australia) Available at https://www.cgg.wa.gov.au/live/my-neighbourhood/geraldton-coastal-hazard-risk-management-and-adaptation-planning-report.aspx
Fox-Kemper B, Hewitt HT, Xiao C, Aðalgeirsdóttir G, et al. (2021) Ocean, cryosphere and sea level change. In ‘Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change’. (Eds V Masson-Delmotte, P Zhai, A Pirani, SL Connors, C Péan, S Berger, N Caud, Y Chen, L Goldfarb, MI Gomis, M Huang, K Leitzell, E Lonnoy, JBR Matthews, TK Maycock, T Waterfield, O Yelekçi, R Yu, B Zhou) pp. 1211–1362. (Cambridge University Press: Cambridge, UK, and New York, NY, USA) 10.1017/9781009157896.011.1212
Gallant JC, Dowling TI, Read AM, et al. (2011) 1 second SRTM derived products user guide. (Geoscience Australia) Available at https://elevation-direct-downloads.s3-ap-southeast-2.amazonaws.com/1sec-dem/1secSRTM_Derived_DEMs_UserGuide_v1.0.4.pdf
Ghanbari M, Arabi M, Obeysekera J, Sweet W (2019) A coherent statistical model for coastal flood frequency analysis under nonstationary sea level conditions. Earth’s Future 7(2), 162-177.
| Crossref | Google Scholar |
Gilleland E, Katz RW (2016) extRemes 2.0: an extreme value analysis package in R. Journal of Statistical Software 72(8), 1-39.
| Crossref | Google Scholar |
Haasnoot M, Kwakkel JH, Walker WE, et al. (2013) Dynamic adaptive policy pathways: a method for crafting robust decisions for a deeply uncertain world. Global Environmental Change 23(2), 485-498.
| Crossref | Google Scholar |
Haasnoot M, Lawrence J, Magnan AK (2021) Pathways to coastal retreat. Science 372(6548), 1287-1290.
| Crossref | Google Scholar | PubMed |
Habel S, Fletcher CH, Anderson TR, et al. (2020) Sea-level rise induced multi-mechanism flooding and contribution to urban infrastructure failure. Scientific Reports 10(1), 3796.
| Crossref | Google Scholar | PubMed |
Hague BS, Talke SA (2024) The influence of future changes in tidal range, storm surge, and mean sea level on the emergence of chronic flooding. Earth’s Future 12(2), e2023EF003993.
| Crossref | Google Scholar |
Hague BS, Murphy BF, Jones DA, et al. (2019) Developing impact-based thresholds for coastal inundation from tide gauge observations. Journal of Southern Hemisphere Earth Systems Science 69(1), 252-272.
| Crossref | Google Scholar |
Hague BS, Jones DA, Trewin B, et al. (2021) ANCHORS: a multi-decadal tide gauge dataset to monitor Australian relative sea level changes. Geoscience Data Journal 9(2), 256-272.
| Crossref | Google Scholar |
Hague BS, Jones DA, Jakob D, et al. (2022) Australian coastal flooding trends and forcing factors. Earth’s Future 10(2), e2021EF002483.
| Crossref | Google Scholar |
Hague BS, Grayson RB, Talke SA, et al. (2023) The effect of tidal range and mean sea-level changes on coastal flood hazards at Lakes Entrance, south-east Australia. Journal of Southern Hemisphere Earth Systems Science 73(2), 116-130.
| Crossref | Google Scholar |
Haigh ID, Wijeratne EMS, MacPherson LR, et al. (2014a) Estimating present day extreme water level exceedance probabilities around the coastline of Australia: tides, extra-tropical storm surges and mean sea level. Climate Dynamics 42(1–2), 121-138.
| Crossref | Google Scholar |
Haigh ID, MacPherson LR, Mason MS, et al. (2014b) Estimating present day extreme water level exceedance probabilities around the coastline of Australia: tropical cyclone-induced storm surges. Climate Dynamics 42(1–2), 139-157.
| Crossref | Google Scholar |
Hanslow DJ, Morris BD, Foulsham E, et al. (2018) A regional scale approach to assessing current and potential future exposure to tidal inundation in different types of estuaries. Scientific Reports 8(1), 7065.
| Crossref | Google Scholar |
Hanslow DJ, Fitzhenry MG, Hughes MG, et al. (2023) Sea level rise and the increasing frequency of inundation in Australia’s most exposed estuary. Regional Environmental Change 23(4), 146.
| Crossref | Google Scholar |
Harley MD, Kinsela MA, Sánchez-García E, et al. (2019) Shoreline change mapping using crowd-sourced smartphone images. Coastal Engineering 150(March), 175-189.
| Crossref | Google Scholar |
Harrison SE, Potter SH, Prasanna R, et al. (2021) ‘Where oh where is the data?’: identifying data sources for hydrometeorological impact forecasts and warnings in Aotearoa New Zealand. International Journal of Disaster Risk Reduction 66, 102619.
| Crossref | Google Scholar |
Hermans THJ, Malagón-Santos V, Katsman CA, et al. (2023) The timing of decreasing coastal flood protection due to sea-level rise. Nature Climate Change 13(April), 359-366.
| Crossref | Google Scholar |
Hino M, Field CB, Mach KJ (2017) Managed retreat as a response to natural hazard risk. Nature Climate Change 7(5), 364-370.
| Crossref | Google Scholar |
Hirschfeld D, Behar D, Nicholls RJ, et al. (2023) Global survey shows planners use widely varying sea-level rise projections for coastal adaptation. Communications Earth & Environment 4(102), 102.
| Crossref | Google Scholar | PubMed |
Hunter JR (2010) Estimating sea-level extremes under conditions of uncertain sea-level rise. Climatic Change 99(3), 331-350.
| Crossref | Google Scholar |
Hunter JR (2012) A simple technique for estimating an allowance for uncertain sea-level rise. Climatic Change 113(2), 239-252.
| Crossref | Google Scholar |
Jolly C, Green J (2019) What is the chance of an extreme event happening again next year? Australian Journal of Emergency Management 34(4), 18-19.
| Google Scholar |
Khojasteh D, Felder S, Heimhuber V, et al. (2023) A global assessment of estuarine tidal response to sea level rise. Science of The Total Environment 894(February), 165011.
| Crossref | Google Scholar | PubMed |
Lempert RJ, Lawrence J, Kopp RE, et al. (2024) The use of decision making under deep uncertainty in the IPCC. Frontiers in Climate 6(June), 1380054.
| Crossref | Google Scholar |
Li S, Wahl T, Piecuch C, et al. (2023) Compounding of sea-level processes during high-tide flooding along the US coastline. Journal of Geophysical Research: Oceans: Oceans 128, e2023JC019885.
| Crossref | Google Scholar |
Local Government Association of South Australia (2020) Coastal adaptation guidelines. November 2020. (LGASA) Available at https://www.lga.sa.gov.au/__data/assets/pdf_file/0034/837565/guidelines-coastal-adaptation.pdf
McInnes KL, Macadam I, Hubbert GD, et al. (2009) A modelling approach for estimating the frequency of sea level extremes and the impact of climate change in southeast Australia. Natural Hazards 51(1), 115-137.
| Crossref | Google Scholar |
McInnes KL, Church J, Monselesan D, et al. (2015) Information for Australian impact and adaptation planning in response to sea-level rise. Australian Meteorological and Oceanographic Journal 65(1), 127-149.
| Crossref | Google Scholar |
Merrifield MA, Genz AS, Kontoes CP, et al. (2013) Annual maximum water levels from tide gauges: contributing factors and geographic patterns. Journal of Geophysical Research: Oceans 118(5), 2535-2546.
| Crossref | Google Scholar |
Moftakhari HR, AghaKouchak A, Sanders BF, et al. (2017) Cumulative hazard: The case of nuisance flooding. Earth’s Future 5, 214-223.
| Crossref | Google Scholar |
Moore FC, Obradovich N (2020) Using remarkability to define coastal flooding thresholds. Nature Communications 11(1), 530.
| Crossref | Google Scholar |
O’Neill B, van Aalst M, Zaiton Ibrahim Z, et al. (2022) Key risks across sectors and regions. In ‘Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change’. (Eds H-O Pörtner, DC Roberts, M Tignor, ES Poloczanska, K Mintenbeck, A Alegría, M Craig, S Langsdorf, S Löschke, V Möller, A Okem, B Rama) pp. 2411–2538 (Cambridge University Press, Cambridge: UK, and New York, NY, USA) 10.1017/9781009325844.025
Paulik R, Stephens S, Wild A, et al. (2021) Cumulative building exposure to extreme sea level flooding in coastal urban areas. International Journal of Disaster Risk Reduction 66, 102612.
| Crossref | Google Scholar |
Prakash M, Cohen R, Hilton J, et al. (2020) An evidence based approach to evaluating flood adaptation effectiveness including climate change considerations for coastal cities: City of Port Phillip, Victoria, Australia. Journal of Flood Risk Management 13(S1), e12556.
| Crossref | Google Scholar |
Pucino N, Kennedy DM, Carvalho RC, et al. (2021) Citizen science for monitoring seasonal-scale beach erosion and behaviour with aerial drones. Scientific Reports 11(1), 3935.
| Crossref | Google Scholar | PubMed |
Rasmussen DJ, Kulp S, Kopp RE, et al. (2022) Popular extreme sea level metrics can better communicate impacts. Climatic Change 170(3–4), 30.
| Crossref | Google Scholar | PubMed |
Ritman M, Hague B, Katea T, et al. (2022) Past and future coastal flooding in Pacific small-island nations: insights from the Pacific Sea Level and Geodetic Monitoring (PSLGM) project tide gauges. Journal of Southern Hemisphere Earth Systems Science 72(3), 202-217.
| Crossref | Google Scholar |
Rueda A, Vitousek S, Camus P, et al. (2017) A global classification of coastal flood hazard climates associated with large-scale oceanographic forcing. Scientific Reports 7(1), 5038.
| Crossref | Google Scholar |
Schipper ELF (2020) Maladaptation: when adaptation to climate change goes very wrong. One Earth 3(4), 409-414.
| Crossref | Google Scholar |
Simpson NP, Mach KJ, Constable A, et al. (2021) A framework for complex climate change risk assessment. One Earth 4(4), 489-501.
| Crossref | Google Scholar |
Slangen ABA, Haasnoot M, Winter G (2022) Rethinking sea-level projections using families and timing differences. Earth’s Future 10(4), e2021EF002576.
| Crossref | Google Scholar |
State Emergency Service (2012) East Gippsland Shire flood emergency plan – version 1.1. Attachment 05. (SES, Victoria) Available at https://www.ses.vic.gov.au/documents/8655930/9320058/East+Gippsland+Municipal+Flood+Emergency+Plan+-+Gippsland+Lakes.pdf/0f808192-d354-58ba-5b76-1385eb1bef10
Stephens SA, Bell RG, Lawrence J (2018) Developing signals to trigger adaptation to sea-level rise. Environmental Research Letters 13(10), 104004.
| Crossref | Google Scholar |
Sweet WV, Park J (2014) From the extreme to the mean: acceleration and tipping points of coastal inundation from sea level rise. Earth’s Future 2(12), 579-600.
| Crossref | Google Scholar |
Taherkhani M, Vitousek S, Barnard PL, et al. (2020) Sea-level rise exponentially increases coastal flood frequency. Scientific Reports 10(1), 6466.
| Crossref | Google Scholar | PubMed |
Taylor A, Brassington GB (2017) Sea level forecasts aggregated from established operational systems. Journal of Marine Science and Engineering 5(3), 1-16.
| Crossref | Google Scholar |
Tebaldi C, Ranasinghe R, Vousdoukas M, et al. (2021) Extreme sea levels at different global warming levels. Nature Climate Change 11(9), 746-751.
| Crossref | Google Scholar |
The Local Government Association of Queensland and Department of Environment and Heritage Protection (2016) Developing a coastal hazard adaptation strategy: minimum standards and guideline for Queensland Local Governments. (State of Queensland) Available at https://www.qcoast2100.com.au/downloads/file/55/minimum-standards-and-guideline
Thompson PR, Widlansky MJ, Merrifield MA, et al. (2019) A statistical model for frequency of coastal flooding in Honolulu, Hawaii, during the 21st century. Journal of Geophysical Research: Oceans 124(4), 2787-2802.
| Crossref | Google Scholar |
van der Wijst KI, Bosello F, Dasgupta S, et al. (2023) New damage curves and multimodel analysis suggest lower optimal temperature. Nature Climate Change 13(5), 434-441.
| Crossref | Google Scholar |
Wahl T, Haigh ID, Nicholls RJ, et al. (2017) Understanding extreme sea levels for broad-scale coastal impact and adaptation analysis. Nature Communications 8(May), 16075.
| Crossref | Google Scholar | PubMed |
Walker WE, Haasnoot M, Kwakkel JH (2013) Adapt or perish: a review of planning approaches for adaptation under deep uncertainty. Sustainability (Switzerland) 5(3), 955-979.
| Crossref | Google Scholar |
Watson PJ, Frazer A (2009) A Snapshot of Future Sea Levels: Photographing the King Tide. 12 January 2009. DECCW 2009/722. (Department of Environment, Climate Change and Water NSW: Sydney, NSW, Australia) Available at https://www.climatechange.environment.nsw.gov.au/sites/default/files/2021-06/A%20Snapshot%20of%20Future%20Sea%20Levels.pdf
Woodworth PL, Melet A, Marcos M, et al. (2019) Forcing factors affecting sea level changes at the coast. Surveys in Geophysics 40(6), 1351-1397.
| Crossref | Google Scholar |
