Melting ice and rising seas – connecting projected change in Antarctica’s ice sheets to communities in Aotearoa New Zealand

ABSTRACT

Changes in global mean sea level are a clear indicator of a warming climate, but local factors including land subsidence or uplift, cause changes in relative sea level that drive shoreline shifts. These local changes and their impact on coastal hazards matter to coastal communities. NZ SeaRise produced relative sea level projections for Aotearoa to include the latest global climate and Antarctic Ice Sheet research and estimates of vertical land movement at high spatial resolution. Research-informed communication to the public and planners included a web-based projections tool supplemented by written and visual narratives, and a media engagement plan. This communication, and analysis of media impact, provided a case study for audience-relevant information on sea-level rise. Information regarding shoreline change and evolving hazards, required for risk assessment, was not included in the NZ SeaRise projections. New research is needed to reduce uncertainty in future Antarctic Ice Sheet contribution to sea level, link changes in sea surface height to our dynamic land surface and enhance communication approaches. Several examples of the required research are presented here but ongoing efforts must refine the timing and magnitude of coastline change, better define coastal hazards and risks, and develop appropriate adaptation strategies for unavoidable climate change impacts.

KEYWORDS:

• Sea-level rise, Antarctica, ice sheets, coastal adaptation, science communication

Introduction and background

Sea-level rise is an important indicator of a warming climate (Milne et al. 2009; Church et al. 2011; Stammer et al. 2013) with significant consequences for our future (Hallegatte et al. 2013; Hinkel et al. 2014; Nurse et al. 2014; Hinkel et al. 2018; Oppenheimer et al. 2019; Bachner et al. 2022). Global mean sea level has risen ∼0.2 m since 1900 (Frederikse et al. 2020; Fox-Kemper et al. 2021) and this increase in sea level has already had a significant impact on humanity with attributable increases in flood frequency (Lin et al. 2016) and cost of damage (Strauss et al. 2021). The observed increase in average ocean height over the twentieth century is primarily due to thermal expansion of the ocean as it warms, and meltwater input as glaciers and ice sheets lose mass. Globally, major research efforts have focused on projecting future increases in sea level and recent estimates show it is ‘likely’ (17%–83%) to rise between 0.29 and 1.1 m above a late twentieth-century baseline by 2100, depending on greenhouse gas emissions pathway (Oppenheimer et al. 2019; Fox-Kemper et al. 2021; Garner et al. 2021).

An imperative to understand the impact of sea-level rise across coastal regions in New Zealand (Parliamentary Commissioner for the Environment 2015; Ministry for the Environment 2017, 2022a) is driving our community to better connect knowledge of Antarctic ice sheet behaviour to local factors in Aotearoa. Changes in land-based ice mass and thermal expansion of the world’s oceans are the primary causes of global mean sea-level rise. However, it is local relative sea level change that affects coastal hazards and risk and matters most to coastal communities. Vertical movement of the land surface can have a significant effect on local relative sea level; land subsidence can accelerate climate driven sea-level rise whereas uplift can slow it down. These changes in land surface height can expose coastal inhabitants to rates of sea-level rise up to four times faster than the global mean (Nicholls et al. 2021).

The New Zealand Sea Rise Programme – Te Tai Pari o Aotearoa

New Zealand is a high user of sea level projections in planning processes (Hirschfeld et al. 2023). This reflects the availability of four national scale SLR scenarios and clearly articulated guidance for practitioner use (Ministry for the Environment 2017; Lawrence et al. 2018). However, until recently, these national scenarios did not include the influence of local vertical land movement (VLM). The New Zealand Sea Rise Programme (NZ SeaRise) – Te Tai Pari o Aotearoa was designed to produce relative sea level projections every 2 km for the over 15,000 km of New Zealand’s coastline (Levy et al. 2020). These projections incorporate the latest regional sea level projections data from the Intergovernmental Panel for Climate Change Assessment Report 6 (Fox-Kemper et al. 2021; Kopp et al. 2023) but also include highly variable rates of VLM identified from satellite-based observation systems (Hamling et al. 2022). These new projections improve the ‘one size fits all’ approach to sea level projections previously available and have been promulgated to local/regional authorities through an interim coastal guidance (Ministry for the Environment 2022b).

While the NZ SeaRise projections connect our Antarctic Ice Sheet research to our coastline, they don’t automatically connect our science to New Zealanders. Communicating with publics about sea-level rise presents challenges involving ‘complex science, uncertainty, invisibility, and politicization’ (Covi and Kain 2016). Public engagement associated with the launch of the NZ SeaRise rise projections was informed by academic literature on sea-level rise communication (Akerlof et al. 2017) and a national survey of New Zealanders into what they knew and understood about sea-level rise (Priestley et al. 2021). In assessing sea-level rise to 2100, nearly 75% of respondents in this study selected options that were in line with scientifically plausible projections, selecting ‘up to 0.4 m’ (28.6%), ‘up to 1 m’ (30.9%), and ‘up to 2 m’ (14.9%). But a large group of respondents (18.9%) overestimated sea-level rise projections to 2100, selecting ‘up to 5 m’ (10.7%) or ‘more than 5 m’ (8.2%). Respondents also judged global sea-level rise as an issue that was of more concern, and likely to be higher, than the sea-level rise that would affect the coastline of Aotearoa New Zealand.

As the public face of NZ SeaRise, a web-based GIS tool (referred to here as ‘the map’) with enough versatility to engage public and stakeholder users (https://www.searise.nz/maps-2), was developed by Auckland-based data management and analytics platform Takiwā. It was built around a map of Aotearoa New Zealand and developed through an iterative process that involved feedback from NZ SeaRise researchers and invited stakeholders. Variations in VLM rate are visualised by graduated coloured points distributed every ∼2 km around the coastline of Aotearoa New Zealand (Figure 1A). Detailed projections for each location are accessed by clicking on the relevant data point, which brings up a graph showing sea-level rise with medium confidence to 2150 and low confidence to 2300 under different Shared Socio-economic Pathways (SSPs) (Meinshausen et al. 2020) with and without VLM (Figure 1B). To avoid focus on unlikely worst-case and best-case scenarios (e.g. SSP5-8.5 and SSP1-1.9) (Huard et al. 2022; Jehn et al. 2022) NZ SeaRise graphs depicting medium confidence sea-level rise projections to 2150 were set at a default option showing projections for a ‘middle-of-the-road’ scenario (SSP2-4.5) (Meinshausen et al. 2020) with and without VLM.

Figure 1. A, Vertical land movement was illustrated via data points every 2 km around the New Zealand coastline, from dark red, representing 9 mm/year uplift, with a gradient through pink, white, light blue to dark blue, representing 9 mm/year subsidence. This map shows the bottom of the North Island, including the harbour coastline of Wellington city. B, Sea-level rise projections for a location in Evans Bay, Wellington Harbour (site 2506) where the VLM rate is −2.8 ± 0.9 mm/year.

The map is accessed at https://www.searise.nz/maps-2. Users have the option of either clicking on ‘For the Public’ or ‘For Planners’ buttons. Both provide access to the projections, background information and video-based information. The planners button leads to a version of the map where key data sets including site details (latitude and longitude, VLM rates with error, and a data quality estimate) and sea level projections data can be downloaded. Data are licensed under a Creative Commons Attribution 4.0 International. A ‘readme’ file describing the data set can be downloaded and links to relevant government guidance documents and related academic papers can be accessed via a technical information tab.

Previous research into communicating sea-level rise found that visual communication tools were ‘often designed by experts for experts, with ‘the public’ as a decidedly untested upon and absent secondary audience’ (Stephens and Richards 2020). Without contextualisation such visuals are unlikely to increase engagement or willingness to act. Stephens and Richards (2020) proposed that one way to help publics connect with mapped information was to add qualitative information – to combine data visualisations with communications that include ‘people and their stories’ (Stephens and Richards 2020). Informed by this research, we added qualitative information to the NZ SeaRise public map, mostly in the form of narrative context, to complement the quantitative information conveyed in the VLM points and graphs. A ‘stories’ button leads to locally specific illustrations, stories from historic and recent newspapers, case studies, and other media items telling stories about sea-level rise. A ‘Sea-level rise’ button leads to information about sea-level rise, VLM, and the IPCC shared socio-economic pathways (Meinshausen et al. 2020). A ‘Context’ button gives public users the opportunity to add additional context to the map, such as marae, council boundaries, bridges, and runways. The map was supplemented by information on the NZ SeaRise website (http://searise.nz), including videos, case studies, frequently asked questions, and fact sheets.

The need for further work to reduce uncertainty and enhance uptake

Like New Zealand’s dynamic coastline, knowledge is never static, and uncertainties regarding future sea-level rise and its impact remain. New research in three key areas is required to improve SLR projections and their application to coastal studies and use in adaptation planning:

First, there remains a critical need to enhance our understanding of the physical processes that drive ice sheet retreat (and advance) to improve our climate and ice sheet models and Antarctic Ice Sheet projections (Colleoni et al. 2022). In this paper, we summarise the current state of numerical climate and ice sheet modelling and future projections. We analyse results from Ice Sheet Model Intercomparison Project 6 (ISMIP6) and highlight new work to couple ice sheet and ocean models, which represents a step forward in our ability to project the future response of Antarctica’s ice shelves and consequences for grounded ice.

Second, while NZ SeaRise research provided location specific relative sea level projections for every 2 km along our coastline, detailed estimates of changes in local elevation and sea level datums are required by researchers, planners, and coastal communities to accurately assess the impact of relative sea-level rise or fall. Here we provide an example from Wellington Harbour that connects offshore bathymetry, onshore land elevation data, and local hydrological information including mean sea level.

Finally, we need to create effective mechanisms to engage audiences with our sea level science and the likely impacts of SLR on coastal hazards (e.g. erosion, storm-surge-driven flooding/overtopping, tsunami, groundwater inundation, and fluvial and pluvial flooding). This is particularly important amid public ambiguity and confusion regarding some aspects of climate change and SLR (Priestley et al. 2021), the rise of disinformation (Allgaier 2019; Hannah et al. 2022), and the widening uncertainty in future projections. A range of approaches have been developed to help people grapple with the complex challenge of climate change in New Zealand, including serious games (Flood et al. 2018; Blackett et al. 2022) (https://mycoastalfutures.niwa.co.nz/), dynamic adaptive pathways planning and hybrid approaches (Lawrence et al. 2019), and other coastal climate services (Lawrence et al. 2021). Here we outline the public launch of the sea-level rise projections as a case study of how we generated research-informed and community-relevant communication and use simple tools for a preliminary assessment of its impact.

Antarctic ice sheet projections: processes and uncertainties

Both the Greenland and Antarctic Ice Sheets have contributed to global mean sea-level rise over the past decades (Rignot et al. 2019; Shepherd et al. 2020). Satellite-based observations show mass loss from Earth’s polar ice sheets is accelerating (Velicogna et al. 2014; Nerem et al. 2018; Shepherd et al. 2018; Khan et al. 2022) increasing meltwater flux to the world’s oceans. A growing number of studies suggest that as temperatures rise above 1.5°C, the Greenland and West Antarctic ice sheets may reach a tipping point (Armstrong McKay et al. 2022 and refs within). At least 0.27 m of sea level equivalent from Greenland may be unavoidable and could occur as soon as 2100 (Box et al. 2022). Portions of the West Antarctic Ice Sheet are already retreating at a pace that may not slow or stop due to internal ice dynamics that drive runaway retreat (Rott et al. 2002; Scambos et al. 2004; Favier et al. 2014; Joughin et al. 2014; Rignot et al. 2014; Christie et al. 2016; Khazendar et al. 2016; Scheuchl et al. 2016).

The response of the Antarctic Ice Sheet to warming and the rate and magnitude of future meltwater contribution from the ice sheet to global sea-level rise is important to know but difficult to constrain and model. Work to reduce uncertainty has been a focus for the ice sheet research community with significant effort aimed at improving process understanding and representation in models (DeConto and Pollard 2016; DeConto et al. 2021). However, despite progress in using observations to guide numerical representations of physical processes, such as grounding line migration and calving (Levermann et al. 2012; Cornford et al. 2013), projections of the Antarctic Ice Sheet remain uncertain (Edwards et al. 2021; Lowry et al. 2021). Sources of uncertainty include the initial states of the models (Seroussi et al. 2019), and the parameterisation of small-scale or complex processes (Bulthuis et al. 2019; Edwards et al. 2019). It is also uncertain how polar climates will evolve under a given emissions scenario (Barthel et al. 2020), and whether the interactions between the ice sheets, ocean, and atmosphere will lead to any system feedbacks that exacerbate or mitigate future ice mass loss (Palerme et al. 2017; Golledge et al. 2019).

The Ice Sheet Model Intercomparison Project

The Ice Sheet Model Intercomparison Project 6 (ISMIP6) is a community effort to quantify some of the main sources of uncertainty in ice sheet projections of the twenty-first century (Nowicki et al. 2016). Because ISMIP6 was conducted concurrently with the Coupled Model Intercomparison Project 6 (CMIP6), the majority of climate forcings used in the ice sheet model experiments were from earlier iterations of climate models (CMIP5) under Representative Concentration Pathway (RCP) scenarios (i.e. RCP2.6 and RCP8.5 for low and high emissions, respectively). The CMIP5 models were selected based on analysis of how they compared to observations of Arctic and Antarctic climatology while also considering a wide-range of end-of-century oceanic and atmospheric conditions (Barthel et al. 2020). Under RCP8.5, ISMIP6-Greenland experiments showed higher uncertainty related to ice sheet model differences than climate forcing (Goelzer et al. 2020), but the overall model spread was much lower than that of ISMIP6-Antarctica (Seroussi et al. 2020).

Here we build on previous ISMIP6-Antarctica results by comparing modelled ice sheet mass balance with satellite observations from the Gravity Recovery and Climate Experiment (GRACE) and its successor GRACE-Follow On (GRACE-FO) (Figure 2A). The GRACE programme consists of twin satellites that measure changes in Earth’s gravitational field and are used to observe changes in ice sheet mass (https://grace.jpl.nasa.gov/). GRACE and GRACE-FO data are publicly available and can be downloaded from the National Aeronautics and Space Administration. We focus our analysis on the period from January 2015 until present as the start date of the ISMIP6 model projections is 2015. For each ISMIP6 simulation, we calculated the change in ice mass above floatation relative to simulation start date. We did not subtract the ice mass change of a control simulation, as is done in Seroussi et al. (2020), so that we could include the committed ice sheet response to past warming in our analysis, but in doing so we may retain model drift of some of the participating models.

Figure 2. A, Antarctic ice above flotation mass change (Gt) in ISMIP6 models and in observations (i.e. GRACE and GRACE-FO data). The model ensemble averages of RCP2.6 (rcp26ea) and RCP8.5 (rcp85ea) are shown in dark blue and red, respectively. B, Antarctic ice above flotation mass change (Gt) in rcp26ea, rcp85ea, VUW PISM, and UCIJPL ISSM. The shaded coloured boxes show the quartile 1 to quartile 3 range of the model output, with the orange line indicating the median. The whisker bar shows the 1.5x range of quartile 1 to quartile 3, with model outliers marked by the open circles. The blue boxes indicate RCP2.6 scenario for VUW PISM and UCIJPL ISSM; there is no range shown because this refers to a single model simulation.

Modelled Antarctic ice sheet mass change during the period from January 2015 to present ranges from – 4000 to + 4000 Gigatons (Figure 2A). This wide range is largely due to differences in model initialisation. Many of the ice sheet models also either under – or overestimate surface mass balance (SMB) and basal mass balance (BMB) under ice shelves (Lenaerts et al. 2012; Depoorter et al. 2013; Rignot et al. 2013). Only two ice sheet models produced SMB and BMB within the range of observation error in control simulations: the Victoria University of Wellington contribution using the Parallel Ice Sheet Model (VUW PISM) and the UC Irvine – Jet Propulsion Lab contribution using the Ice Sheet System Model (UCIJPL ISSM). These two models and the overall ensemble averages of the RCP2.6 and RCP8.5 scenarios track with the GRACE and GRACE-FO satellite observations. Importantly, however, little difference is observed with respect to the emissions scenario by the year 2100 (Figure 2B).

In some models, forcing under RCP8.5 results in less mass loss than under other RCP’s because the warming atmosphere produces more snowfall over the ice sheet, counterbalancing ice discharge resulting from oceanic warming (Seroussi et al. 2020). This relationship is also observed in ice sheet simulations forced by CMIP6 models under SSP1-2.6 and SSP5-8.5 (Payne et al. 2021). To further explore this process, we analysed output from the CMIP6 model IPSL-CM6A-LR and performed ice sheet model experiments using VUW PISM (Figure 3). We compare two experiments: (1) SMB remains constant, but ocean forcing (OF) warms under an SSP5-8.5 scenario to the year 2100; and (2) SMB increases under an SSP5-8.5 scenario to the year 2300, and OF increases only to 2100 (SMBOF). Both experiments show an inflection in sea level contribution in 2094 due to ice shelf collapse resulting from negative BMB. As Southern Ocean winter sea ice declines, the substantial SMB increase of the relatively dry continental interior of the ice sheet has a mitigating effect on the ice sheet sea level contribution (Figure 3B). However, ice shelf collapse from twenty-first century oceanic warming under high emissions counteracts this effect, suggesting that increased SMB is not sufficient to prevent future sea-level rise from Antarctica over centennial timescales.

Figure 3. A, Zonal average Surface Mass Balance (SMB) change (%) from present in two polar latitudinal bands versus Southern Ocean winter sea ice area (m2) in the IPSL-CM6A-LR SSP5-8.5 extended run to 2300. B, Ice sheet sea level contribution (m) in PISM simulations: (1) SSP5-8.5 ocean forcing (OF), but constant SMB; SSP5-8.5 ocean forcing and SSP5-8.5 SMB forcing (SMBOF). C, Ice thickness change (2300–2015) of OF. Blue (red) indicates an increase (decrease) in ice thickness (m). D, Ice thickness between SMBOF and OF. Note the increase in ice thickness of most of the ice sheet; the main exceptions are the Ross and Ronne-Filchner ice shelves (bright red areas), which show increased ice discharge resulting from higher SMB.

Improved models of ice-ocean interactions

Given the importance of ice shelf stability in Antarctica, ice-ocean interactions have become a primary focus of the ice sheet modelling community (Asay-Davis et al. 2017; Pattyn et al. 2017). It should be noted that the ISMIP6 simulations are simplified as the CMIP5 models do not resolve ocean cavities under ice shelves, and there is no coupling between the two systems. Rather than extrapolating ocean temperatures and salinities at the margins of ice shelves to determine basal melt rates, we are now coupling ice sheet models to regional ocean models that do resolve the ocean circulation under ice shelves (Gladstone et al. 2021; Naughten et al. 2021). This is an important advance because these sub-ice shelf oceanic currents strongly influence ice sheet dynamics (Nakayama et al. 2019; Wåhlin et al. 2021). Here, we show an example using PISM and the Regional Ocean Modelling System (ROMS) for RCP2.6 and RCP8.5 (Figure 4). In these coupled simulations to the year 2080, ocean temperatures averaged over depth are relatively warmer along the West Antarctic Ice Sheet in RCP8.5, a region of the ice sheet that is potentially prone to runaway retreat (Mercer 1978; Favier et al. 2014; Joughin et al. 2014).

Figure 4. A, Schematic of ice sheet-ocean model coupling using ROMS and PISM. Atmospheric forcing and sea ice is provided by the NorESM1-M CMIP5 model. PISM and ROMS exchange updated ocean forcing of ice shelf cavities and ice sheet topographic changes at 5-year intervals. B, Depth-averaged Southern Ocean temperature and basal ice shelf temperature (K) in PISM at the year 2080 for RCP2.6, RCP8.5, and the scenario difference.

Continually improving and validating ice sheet models is a critical component of coastal planning and adaptation. Model intercomparisons like ISMIP6 offer an opportunity to assess uncertainty related to ice sheet, oceanic and atmospheric processes in model projections. Despite uncertainties, emissions pathways of today will dictate the scale of the committed ice sheet contribution to sea level for hundreds to thousands of years, affecting many future generations. Models indicate that a low emissions pathway (e.g. SSP1-2.6) is the most likely to ensure future ice shelf stability (Golledge et al. 2015; DeConto et al. 2021; Lowry et al. 2021). Because these floating ice shelves buttress the land-based ice sheets, coupled ice sheet-regional ocean models are the best tools available for determining how ice-ocean interactions and feedbacks will ultimately affect Antarctica’s future sea level contribution.

Locally relevant sea level projections for New Zealand

NZ SeaRise projections indicate how much sea level will likely rise but do not show how shorelines will shift and coastal hazards will evolve. Better estimates of shoreline change e.g. (Ford and Kench 2015; Bamunawala et al. 2021) and shifts in coastal hazard frequency, exposure, and risk (Kirezci et al. 2020; Taherkhani et al. 2020; Almar et al. 2021; Paulik et al. 2021; Tebaldi et al. 2021; Paulik et al. 2020) are needed to guide future coastal adaptation.

To investigate and estimate how the dynamic coastline of Aotearoa will change over the coming decades and centuries requires additional effort to: (1) improve regional and local sea level projections (Slangen et al. 2014; Jackson and Jevrejeva 2016; Jevrejeva et al. 2016; Hamlington, Gardner, et al. 2020; Hieronymus and Kalén 2020), (2) better estimate land deformation at high resolution across the coastal zone (Hamling et al. 2022), (3) generate high-resolution elevation models that improve shallow water bathymetry (Li et al. 2019; Pereira et al. 2019; Al Najar et al. 2021) and connect land to sea (Bergsma et al. 2021) and (4) improve projections of coastal erosion and sediment transport and accumulation (Dickson et al. 2007; van Maanen et al. 2013; Goldstein et al. 2019). In addition, improved projections of changing total water level at points along evolving shorelines are needed to assess the emergence (or reduction) of future coastal hazards (Paulik et al. 2023). To achieve this, time varying estimates of a range of critical water level information including mean sea level, astronomical tides, storm surges, and waves must be determined relative to a common elevation datum (Gregory et al. 2019).

Development of these data is the aim of future work (e.g. Future Coast Aotearoa and Te Ao Hurihuri: Te Ao Hou – Our Changing Coast research programmes funded by MBIE and the Joining Land and Sea Project at Toitū Te Whenua Land Information New Zealand). However, here we provide an example from Wellington Harbour to show how new offshore and onshore land elevation data can be combined with accurate water elevation data and NZ SeaRise projections to provide location-specific estimates of future water levels.

Joining land and sea

Global sea level projections (Church et al. 2013; Oppenheimer et al. 2019; Fox-Kemper et al. 2021) employ a baseline period to determine time averaged (global mean) sea level from which future sea level changes are modelled under various emissions scenarios. IPCC AR5 and SROCC employ a baseline period of 1986–2005, which was updated in AR6 to a baseline period of 1995–2014. NZ SeaRise projections follow the IPCC AR6 protocol and project changes in sea level from ‘zero’ starting in 2005 (baseline mid-point). These projections offer an estimate of the magnitude of change in sea level through time but must be tied to local elevation datum to evaluate the impact on the surrounding environment (Ministry for the Environment 2022b). We can then assess how key sea surface elevations, (including Mean Sea Level, Lowest Astronomic Tide, and Highest Astronomic Tide), will change relative to the land surface. These data allow examination of the evolution of coastal hazards including changes in the frequency of nuisance flooding and coastal overtopping due to storm surge.

In 2016, Toitū Te Whenua Land Information New Zealand (LINZ), released a new national vertical datum to be used when referencing heights across the country. The New Zealand Vertical Datum 2016 (NZVD2016) allows for the consistent collection and seamless exchange of heighted data within the landmass (including offshore islands) and open waters of Aotearoa New Zealand. NZVD2016 is a quasi-geoid-based datum, where the reference is related to an equipotential gravity surface and global mean sea level, rather than a single point or reference tide gauge.

We intend to improve usability and accessibility of NZVD2016 by determining the relationship between the datum and key sea level surfaces around the New Zealand coastline. This work is underway at LINZ under a project known as Joining Land and Sea (JLAS). JLAS comprises three components (1) determination of Tidal Surfaces extracted from an improved New Zealand EEZ Tidal model (produced by NIWA), (2) connection of existing tide gauges to NZVD2016 and month-long deployment of temporary tide gauges, and (3) a JLAS algorithm based on interpolation using tide gauge data and the tidal model. The JLAS algorithm uses the tide gauge sites to connect the tidal surfaces to the corresponding level in terms of NZVD2016. Therefore, JLAS will allow us to determine absolute values for present and future mean sea level relative to NZVD2016 and correlate these data with surrounding features including topography of the wider coastal environment.

Here, we use preliminary data from JLAS to integrate NZ SeaRise projections with sea surface data in Seaview and Lowry Bay, Wellington. This initial study is an example of the detailed location specific data that are required for more accurate coastal hazard assessment. The study region was selected based on optimal availability of abutting elevation data from both terrestrial LiDAR (Wellington – Hutt City LiDAR (2021) 1 m Digital Elevation Model; LINZ) and bathymetry surveys (Wellington Harbour Bathymetry; NIWA). The gridded data were merged using QGIS opensource tools and voids between datasets were filled using a low pass filter resampling method to produce a smooth and gradational transition. Sea surface elevation data, including mean sea level, mean low and high water, and lowest and highest astronomical tides, were generated for two locations (Sites 2491 and 2493 in the NZ SeaRise data set) relative to NZVD2016. Topographic profiles (cross sections) that pass near the SeaRise sites were generated from the integrated LIDAR and bathymetry data (Figure 5A). Sea level projections data for the SSP2-4.5 emissions scenario were then used to estimate changes in the height of the key sea surface datums.

Today, mean sea level (MSL, as defined using the data above) at Seaview and Lowry Bay is approximately 0.23 m below the NZVD2016 zero elevation and the highest astronomical tide (HAT) reaches 0.69 m above the zero datum. These surfaces are below the current elevation of the protection infrastructure along the coastal margin at Seaview (∼2.05 m) and road that circumvents Lowry Bay (∼1.75 m). Assuming global greenhouse gas emissions follow a ‘middle-of-the-road’ shared socio-economic pathway (SSP2-4.5), MSL at Lowry Bay and Seaview is projected to be between 0.11 and 0.25 m and 0.09 and 0.25 m (respectively) above the current zero datum by 2050 (66% probability – likely range). An increase between ∼0.31 and 0.48 m since 2005. MSL is projected to reach an elevation between 0.53 and 0.94 m and 0.56 and 0.94 m above the modern zero datum by 2100 (66% probability – likely range). An increase between ∼0.76 and 1.07 m above the average elevation in 2005.

The upper bound (83rd percentile) of the likely range for the elevation of HAT at Lowry Bay and Seaview is projected to reach an elevation of 1.85 m by 2100 under SSP2-4.5. This projection places the surface close to the top of coastal protection infrastructure around Seaview and above the road around Lowry Bay. This height exceeds the average elevation across much of the Seaview cross section (∼1.7 m between the coastal margin and mid-point of the cross section in Figure 5). Projected elevation changes in local relative MSL and HAT due to climate-change-driven sea-level rise and local land subsidence clearly increase the chance of coastal flooding due to inundation and overtopping on fair weather days. Elevation estimates shown here do not include storm surge or wave run up. Previous studies have shown that low frequency (rare) inundation events in Wellington Harbour over past decades become significantly more frequent with relatively small increases in sea level. If sea level rises by 0.3 m above the 2005 baseline, a coastal flood that was previously only likely to occur during a large storm-tide that coincides with a very high tide (once every 100 years), is likely to occur on average every year under more frequent ‘mild’ conditions (Parliamentary Commissioner for the Environment 2015; Paulik et al. 2020).

Figure 5. A, Digital elevation model for the Seaview-Lowry Bay study region from terrestrial LiDAR (Wellington – Hutt City LiDAR (2021) 1 m Digital Elevation Model; LINZ) and bathymetry surveys. Location of topographic profiles W-X and Y-Z shown in red. B, Detailed topographic profile across Seaview with key sea surfaces including Mean Sea Level (solid line) and Highest Astronomical Tide (dashed line) shown for ‘today’ (blue lines) and in 2050 (yellow lines) and 2100 (orange lines) under an SSP2-4.5 emissions scenario including vertical land movement. The 66% probability (likely) range for each projection is indicated by the pale yellow and orange bands that encompass the mean value (solid and dashed lines). C, Same as B but for Lowry Bay. Elevation data in metres are listed at bottom of figure and are relative to NZVD 2016. Note these sea level elevation estimates do not include storm surge, wave setup, or wave runup. These dynamic surfaces must be added to the elevation datums when evaluating coastal flooding hazard.

Consistent and coherent national scale elevation information can enhance local adaptation planning as sea levels change. Improvement of the New Zealand tidal model (and evaluation of the various Tidal Surfaces) and collection of the calibrated tide gauge sites and development of the JLAS algorithm are all currently under development. A soft launch of this service is expected in late 2023. Detailed and connected coastal elevation information at the land-sea interface, as presented in the Seaview and Lowry Bay case study (Figure 5), are uncommon around the New Zealand coastline as this zone is difficult to observe. Terrestrial LiDAR is unable to penetrate water and bathymetry data are seldom able to be collected in waters less than 5 m deep and gaps in data across the transition between offshore and onshore are common. In addition, we find that source data were collected for varying purposes, at different times, and to different vertical reference frames. Even where abutting datasets exist, the resolution and accuracy of any modelling can be highly variable. Significant work is required to improve these elevation data sets at a national scale. New satellite technologies (Li et al. 2019; Pereira et al. 2019; Al Najar et al. 2021; Bergsma et al. 2021) and airborne electromagnetic data (Billy et al. 2022) offer promise and are beginning to be used in the New Zealand context (Costa et al. 2023). Filling the ‘coastal gap’ offers an opportunity for future investment as we work to quantify and adapt to our ever-changing coastline.

Public engagement and uptake of sea level projections

Increasingly detailed datasets and projections are only useful if planners and publics have access to useable information. The NZ SeaRise team developed the projections map in consultation with a group of stakeholders from government and industry and began working with select media up to three months before the launch of the map. In their study of public response to sea-level rise communications, Covi and Kain (2016) found that ‘localised information appears to raise people’s interest in the risk and possible response strategies’. Consequently, the media plan involved finding location specific angles for each media outlet’s audience. NZ SeaRise researchers also worked with data graphics experts and science journalists at New Zealand Geographic (Meduna 2022) and Stuff.co.nz (Gibson and Rodriguez 2022) to ensure the richness of the data available was conveyed in engaging, visually striking and accurate ways.

On Sunday 1 May 2022 Newshub ran a television news piece as the first and second items on its flagship 6pm news programme. This broke a news embargo and resulted in many other news stories going live early on Sunday night, before the map was made available at 5am on Monday 2 May 2022. The political sensitivity of the information and the popularity of the map led to a presumed Denial of Service (DOS) attack. In the first few hours of the map’s public availability, it was receiving upwards of 1,200 requests per second. This resulted in the map becoming unavailable for much of the day of its launch until Takiwā allocated more servers to the database to cope with the influx of requests. One consequence of the map’s lack of availability was that the suspected DOS attack became a media story (Fuller 2022; Quinlivan 2022), extending the public’s interest in the data.

With the launch of the interactive map, sea-level rise became a focus of New Zealand media coverage and public attention. This contrasts with the findings of Akerlof et al. (2017), whose review of sea-level rise communication concluded that sea-level rise suffers from low media attention and salience as a public issue. During launch month, Google searches show that public interest in ‘sea-level rise’ was more than 400 percent higher than at any other time in the last decade (Figure 6A). On 2 May 2022, ‘sea-level rise map nz' was New Zealand’s seventh most searched term on Google (Data source: Google Trends https://www.google.com/trends).

Figure 6. A, New Zealand interest in the term ‘sea-level rise’ for the five years between 10 December 2017 and 9 December 2022. Numbers represent search interest relative to the highest point on the chart – the peak search time being May 2022 when the NZ SeaRise findings were launched (100 = peak popularity, 50 = half as popular, 0 = not enough data for the term). B, Global interest in the term ‘sea-level rise’ over the same time period. Data source: Google Trends (https://www.google.com/trends).

We analysed the impact of the NZ SeaRise launch by enlisting media monitoring company iSentia to search New Zealand media for ‘sea-level rise’ OR ‘sea rise’ OR ‘searise’ for 1–14 May 2022. The search, which initially captured 878 stories from television, radio, print and online media, was manually cleaned to remove duplicate items and letters to the editor; correct spelling mistakes from transcriptions of spoken content; and remove headline or summary content that was not about sea-level rise. The following analysis focuses solely on the cleaned headlines and summaries of 235 unique news stories, which resulted in a corpus of more than 30,000 words. We used NVIVO to conduct word frequency analysis on this corpus and to create a series of word clouds (Figure 7) to gain an overview of some aspects of the media coverage.

Figure 7. Word clouds were created in NVIVO from a word frequency analysis on a corpus of more than 30,000 words captured from the headlines and summaries of 235 unique sea-level rise news stories in New Zealand media from 1–14 May 2022. A, Illustrates word frequency of the full corpus after basic words and proper nouns (except for New Zealand and Aotearoa) were added to a stop list. B, Results when all words except sentiment words (words that conveyed opinions, appraisals, emotions, attitudes) were put on the stop list before word frequency analysis. C, Results when all words except place names were put on stop list before word frequency analysis.

In the two-week period following the launch, New Zealand media coverage of sea-level rise tended to avoid worst case scenarios in favour of likely scenarios. And while sentiments were still heightened (Figure 7C) – with frequent use of words such as vulnerable (25 occurrences), alarming (11), important (12) – words such as catastrophic (5), doomsday (1), and apocalypse (0), which have featured in some local and international media coverage of climate change and sea-level rise (Edwards 2017; Weise 2019; Unknown 2021) in the past, were less or not apparent. Overall, there was a nuanced and complex telling of the sea-level rise story. Rather than focusing on how much sea level would rise by a certain date, stories focused on the impacts of sea-level rise. And rather than focus on one or two very vulnerable locations (e.g. South Dunedin has had significant sea-level rise media coverage in the past) (Unknown 2014; Telfer 2017; Unknown 2017) the media coverage covered the nation. Local media used the map and interviews to provide media coverage on the impact of sea-level rise on their coastal settlements and infrastructure (Figure 7C).

More in-depth analysis of the impact of the project’s public engagement could involve a longitudinal study of media coverage of sea-level rise, and surveying or interviewing users of the map, but this is beyond the scope of this multidisciplinary paper.

Future actions

Covi and Kain (2016) note that ‘effective risk communication requires more than conveying accurate information … it should raise awareness, increase understanding, and move audiences to action’. We experienced considerable media and public interest in the NZ SeaRise projections, but we have yet to gauge how people have used and acted on the data. An interim guidance for use of the NZ SeaRise Projections, including local relative sea-level rise, has been released (Ministry for the Environment 2022b) and our National Adaptation Plan (Ministry for the Environment 2022a) recommends that decision makers use the interim guidance projections to stress test adaptation strategies and land-use plans. Work is underway to fold the interim guidance into a revision of the existing coastal guidance. Insurance companies, consultants, and councils have downloaded the data to use for local and regional hazard assessment and team members are regularly asked to present to councils and community groups. However, a robust assessment of the range of community outreach activities and uptake of the SLR information and data has yet to be performed.

Climate anxiety is gaining attention around the world (Sangervo et al. 2022). Research indicates the framing and amount of information about climate change in the media can contribute to the occurrence of climate related anxiety (Pihkala 2019; Clayton 2020). However, the effect of climate anxiety on mental health is debated. For example, a heightened and enduring level of climate anxiety may lead to mental health issues (Clayton 2020), whereas a moderate level of anxiety could engender feelings of virtue, encouraging people to rethink actions with negative ecological impacts (Maran and Begotti 2021). Better evidence from further research into climate change and mental health is needed to define appropriate strategies for climate change adaptation and mitigation (Every-Palmer et al. 2016). There are clear opportunities to use insight from NZ SeaRise to guide future communication efforts that enhance positive environmental and mental health outcomes.

A focus of Te Ao Hurihuri: Te Ao Hou—Our Changing Coast, is to work with coastal communities to facilitate effective communication and reduce climate anxiety, while working towards a fair and equitable response to unavoidable change across our coastal environments.

Summary

Local sea-level projections for Aotearoa New Zealand integrate state-of-the-art climate, glacier, and ice sheet research from the IPCC AR6 with local estimates of vertical land movement. Accessible to New Zealanders via a web-based platform, they connect science on Antarctic Ice Sheet change with our coastline. These projections and our visualisation platform ‘localise’ sea-level change and improve the set of four sea level scenarios previously available to researchers, planners, and communities. Our research-informed communication to the public and planners that accompanied the release of the projections provides a case study of how we generated audience-relevant information on sea-level rise. However, knowledge used to inform and communicate the science behind the sea-level projections is not static and can be improved.

In this study we built on ISMIP6-Antarctica model simulations by comparing outputs from ice sheet mass balance experiments with gravity-based observations of ice mass change obtained from GRACE Programme satellites. Results suggest that simulations produced by VUW PISM and UCIJPL ISSM best match the gravity measurements through the period from January 2015 to today. In addition, new model experiments show that ice shelf collapse due to ocean warming under high CO₂ emissions counteracts increases in surface mass balance towards the end of the century. These results suggest that increased precipitation over Antarctica is not sufficient to prevent future sea-level rise from the Antarctic Ice Sheet over centennial timescales.

NZ SeaRise provided local sea level projections every ∼2 km along the coast of Aotearoa, but the projections were not connected to any specific elevation datum. Elevation data that seamlessly connect our land and sea are needed to improve estimates of the future position of key sea level datums (e.g. mean sea level, highest astronomical tide, total water level). These accurate projections of relative sea level are required to evaluate future coastline location and coastal flooding hazards. In this study we provided an example from Wellington Harbour where we connected local sea level elevation datums to the New Zealand Vertical Datum 2016. NZ SeaRise projections data were then used to project changes in the elevation of these key datums out to 2100 under SSP-2 4.5. Results suggest that plans to adapt to rising sea level will need to be developed and implemented well before the end of this century.

This study offers examples of research that connects projected change in Antarctica’s ice sheets to communities in Aotearoa New Zealand. But more work is needed to reduce uncertainty in future Antarctic Ice Sheet contribution to sea level, link changes in sea surface height to our dynamic land surface and enhance communication approaches so that sea level science is more accessible to all. These efforts will refine the timing and magnitude of change along New Zealand’s coastline, better define hazards and risks, and help us develop appropriate adaptation strategies for unavoidable climate change impacts and are the focus of new research programmes including Te Ao Hurihuri: Te Ao Hou—Our Changing Coast.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the New Zealand Ministry of Business, Innovation and Employment through the NZ Sea Rise Programme [Contract RTVU1705] and Our Changing Coast Programme [RTVU2206]. Aspects of DL, NG, AA-B, RD, SJ, KO, NT, and RL's contributions were also funded through the New Zealand Antarctic Science Platform [Contract ANTA1801].