Keywords

1 Introduction

Climate across the Pacific varies substantially in space and time. This is largely due to two atmospheric bands of high rainfall (convergence zones) that naturally shift over timescales of several years. The convergence zones and their changing positions are associated with droughts, floods, cyclones, and sea level variation in different parts of the Pacific. Pacific Island communities have adapted to fluctuating conditions over many generations and are often highly resilient despite limited financial resources.

Increased greenhouse gas emissions caused by anthropogenic activities are causing global climate change with worrying consequences for PICTs. Temperatures will be warmer in the future, and heatwaves are expected to cause substantial human and environmental impacts for PICTs (IPCC 2021). In particular, marine heatwaves could threaten coral reefs and fisheries. At the same time, oceans are absorbing excess carbon dioxide from the atmosphere, resulting in ocean acidification that could further exacerbate the risk to marine life. Sea level rise beyond the range of natural variability is perhaps the most well-known potential consequence of climate change for PICTs and a failure to cut global emissions could see low-lying islands inundated (IPCC 2021). However, the dynamics of erosion and sediment deposition are complex and certain island geologies may prove surprisingly robust under modest sea level rise (Kench et al. 2018; Nurse et al. 2014).

The atmospheric drivers of climate in the Pacific are notoriously difficult to simulate or predict. Together with high baseline variability, this makes climate change impacts on rainfall difficult to foresee. Some research suggests that the behaviour of the convergence zones is becoming more extreme (Power et al. 2017; Lee and McPhaden 2010). Heavy rainfall events are likely to become more frequent and some areas may see more intense (but less frequent) cyclones (IPCC 2021). Drought risk is likely to increase for some PICTs but decrease for others (Australian Bureau of Meteorology and CSIRO 2014).

Altered weather patterns, warmer temperatures, rising sea levels, and more acidic oceans are fast becoming the ‘new normal’ in a world where atmospheric carbon dioxide levels exceed 400 ppm. The situation for PICTs, particularly small island nations with limited resources, is becoming more precarious by the decade. However, there are examples of island communities and environments showing remarkable resilience in the face of climate variability and change. There is still time to avoid the worst impacts through drastic emissions reduction on a global scale, while also using community-centred approaches informed by traditional knowledge to adapt to changes already underway. Countries with substantial resources and high carbon footprints have a responsibility to the communities on the frontlines of climate change and should commit to rapid emissions reduction along with fair distribution of adaptation funding.

2 Baseline Climate in the Pacific

2.1 Weather and Seasonality

Climatic conditions across different Pacific Island Countries and Territories (PICTs) are diverse, but temperatures are generally warm and annual rainfall high. Most of the countries are positioned close to the equator and experience fairly consistent incoming solar radiation (and relatedly temperatures) throughout the year. Ocean circulation patterns play an important role in regional temperature variation. Warm water is pushed from east to west by the trade winds, resulting in the West Pacific Warm Pool (Australian Bureau of Meteorology and CSIRO 2011) that further increases and stabilizes surrounding temperatures.

At large scales, rainfall across the Pacific is heavily influenced by convergence of the trade winds to create zones of high rainfall (Keener et al. 2013). The Intertropical Convergence Zone (ITCZ) lies north of the equator and drives rainfall in the Federated States of Micronesia, Kiribati, the Marshall Islands, Nauru, Palau, and Papua New Guinea (Australian Bureau of Meteorology and CSIRO 2011). The South Pacific Convergence Zone (SPCZ) approaches the equator diagonally from the southeast and affects rainfall patterns in the Cook Islands, Fiji, Nauru, Niue, Samoa, the Solomon Islands, Tonga, Tuvalu, and Vanuatu (Australian Bureau of Meteorology and CSIRO 2011). The ITCZ and SPCZ converge over the West Pacific Warm Pool. Countries in the western Pacific also receive rainfall associated with the West Pacific Monsoon (WPM), which is highly seasonal and most active over the southern hemisphere summer (Australian Bureau of Meteorology and CSIRO 2011). Changes in strength and position of the ITCZ and SPCZ cause distinct wet and dry seasons throughout the Pacific (Keener et al. 2013). Broadly, the wet season is May to October for northern hemisphere islands and November to April for southern hemisphere islands (CSIRO et al. 2015).

Locally, climate may be impacted by the terrain. High, mountainous islands can experience significant rainfall and temperature gradients. For example, Suva lies on the windward side of Fiji’s largest island and receives substantially more rainfall than Nadi on the leeward side (Australian Bureau of Meteorology and CSIRO 2011). Topography is an important driver of climate in the Cook Islands, Fiji, Papua New Guinea (PNG), Samoa, Tonga, and Vanuatu (Australian Bureau of Meteorology and CSIRO 2011), as well as in Australia and New Zealand (Wratt et al. 1996; Cai et al. 2011). Terry and Wotling (2011) showed that the rain-shadow effect was amplified in streamflow, with runoff ratios 40% higher on average for windward catchments across the La Grande Terre in New Caledonia. The result is significant spatial variability in water resources across some islands. There can also be climatic variation across countries that are spread over a large area (Australian Bureau of Meteorology and CSIRO 2011). For example, the SPCZ sometimes lies between the Northern and Southern Cook Islands, and its position can drive high and low rainfall anomalies in different parts of the country simultaneously (Rongo and Dyer 2014).

2.2 Interannual and Decadal Climate Variability

As discussed in Sect. 5.2.1, conditions across the PICTs are influenced by the positions of the ITCZ and SPCZ (Ludert et al. 2018), which are bands of cloudiness associated with the trade winds. The relative positions of these convergence zones vary at the interannual timescale, and this movement is described by the El Niño Southern Oscillation (ENSO). La Niña events are associated with drier conditions around the equator as the rain-generating ITCZ and SPCZ shift northward and southward respectively, while El Niño events produce drier conditions in the north-west and south-west Pacific as both convergence zones move towards the equator (Fig. 5.1). Extreme El Niño phases are associated with ‘zonal’ SPCZ events, where the convergence zone swings up to 10˚ north of its average position and becomes less diagonally orientated (Cai et al. 2012).

Fig. 5.1
A map of regions around the Pacific ocean. The international convergence zone and South Pacific convergence zone are marked around Palau, Federated States of Micronesia, Marshall Island, Kiribati, Cook Islands, Niue, Fiji, Vanuatu, and East Timor. The warm pool is marked in the center. The direction of trade winds represented by arrows around the area is pointed to the left.

Conceptual map of the ITCZ, SPCZ and Pacific Warm Pool, which determine climate conditions across the Pacific. Shifts of the ITCZ and SPCZ towards the equator are associated with the El Niño phase of ENSO, while poleward shifts are associated with the La Niña phase (Figure reproduced with permission from Australian Bureau of Meteorology and CSIRO 2011)

Full size image

The phases of ENSO are linked to natural disasters (Power et al. 2017) and there has been a strong research focus on better understanding the associated dynamics (Matthews 2012; Rojo Hernández et al. 2020; Wang et al. 2019; Wodzicki and Rapp 2016). However, ENSO is notoriously difficult to predict or even simulate accurately due to known climate model biases in the region (Grose et al. 2014; Samanta et al. 2019). While this creates a challenge for forecasting and preparing for ENSO shifts on the ground, many PICT communities have adapted to high interannual climate variability over generations and have developed remarkable resilience. Knowledge gaps left by uncertain weather and climate models are filled by local understanding derived directly from lived experience (Rongo and Dyer 2014). For example, some residents of the Cook Islands associate high yields of certain crops over summer with subsequent transition to El Niño, and hence heightened cyclone risk. Traditional farming techniques that are particularly robust to climatic variation have been developed in Vanuatu (Mael 2013).

Secondary drivers are thought to influence the climate across the Pacific at longer timescales. The Interdecadal Pacific Oscillation (IPO) and Pacific Decadal Oscillation (PDO) are connected to fluctuations in wind, heat flux, and ocean ‘memory’ over time periods of 15–30 years (Newman et al. 2016). The two phenomena are related, but the IPO encompasses a broader area than the PDO (Folland et al. 2002). These fluctuations influence the strength of ENSO events and the background rainfall across the Pacific (Ludert et al. 2018; Salinger et al. 2001). In fact, Folland et al. (2002) showed that IPO-driven decadal shifts in SPCZ position were comparable in magnitude to interannual shifts induced by ENSO. These longer-scale oscillations are not well understood, and recent research has suggested that they may be the result of external forcing rather than intrinsic low-frequency cycles (Mann et al. 2020).

2.3 Climate Extremes

Climate extremes including drought, extreme rainfall, tropical cyclones, and storm surge events can have devastating impacts on PICTs. Some countries, including PNG and Australia, are also highly vulnerable to wildfires. The risk of natural disasters across the Pacific is strongly linked to the phases of ENSO.

Drought in the Pacific can have damaging consequences as surface water stores are depleted and groundwater resources become strained, impacting water and food security. These challenges are especially acute for small atoll islands where the freshwater lens (the layer of groundwater that is not salty and hence usable by communities) is typically thin, but lost agricultural productivity, decreased electricity production, and other serious impacts have also been highlighted for high islands (McGree et al. 2016). Barkey and Bailey (2017) cite island size, soil permeability and local rainfall patterns as key factors influencing freshwater lens thickness and hence drought vulnerability. As noted in Sect. 5.2.2, La Niña and El Niño events can cause drought in different regions of the Pacific. In some cases, the strength of an ENSO event can impact the sign of the rainfall change; for example, rainfall over Nauru normally increases during the El Niño phase, but extreme El Niño events have caused severe drought (Brown et al. 2013a).

In alternate ENSO phases, PICTs are more likely to experience flooding rains, tropical cyclones, and storm surge. Heavy rainfall and wet antecedent conditions caused severe flooding in Vanua Levu (Fiji) in 2012, first in January and then again in March. Eight people died and thousands were displaced, with damages to crops and infrastructure estimated at USD 50 million (Kuleshov et al. 2014). The same island was hit by Tropical Cyclone Ami in 2003, resulting in both fluvial and storm surge flooding. The storm killed 17 people and left 30,000 without a safe water supply, causing damages totalling USD 30 million (Terry et al. 2004; OCHA 2003). Tropical cyclones are behind ~76% of natural disasters in the South West Pacific (Diamond et al. 2013), with most incubating within 10˚S of the SPCZ (Vincent et al. 2011). Around 80% of Western North Pacific cyclones develop in the ITCZ region (Cao et al. 2012). Even far-distant tropical cyclones can cause storm surge flooding on PICTs due to propagating waves (Nurse et al. 2014).

3 Effects of Anthropogenic Climate Change

As greenhouse gas concentrations in the atmosphere climb, an anthropogenically altered climate is fast becoming the ‘new normal’. The following sections detail the observed and projected climate change impacts on the Pacific, including implications for ENSO dynamics, rainfall, cyclones, temperatures, sea level rise, and ocean acidification.

3.1 Shifts in the ITCZ, the SPCZ and ENSO

Climate conditions in the Pacific are dependent on the ITCZ and SPCZ, so future changes in their behaviour will be extremely consequential. However, high underlying variability (in both space and time) makes it difficult to detect trends or attribute them to anthropogenic climate change. In many small PICTs, a lack of long-term, reliable records of temperature, rainfall, streamflow, wind, and water extraction poses an additional barrier. The result is relatively high uncertainty around the observed effects of climate change across the Pacific (McGree et al. 2016). The Coupled Model Intercomparison ProjectFootnote 1 5 (CMIP5) simulations indicate that ENSO events have already increased in frequency since pre-industrial times (Power et al. 2017). This conclusion is supported by the analysis of Lee and McPhaden (2010), who showed increasing frequency and intensity of El Niño events in the central-equatorial Pacific through analysis of satellite observations. Higher background warming has been observed in the western than eastern Pacific in recent decades, altering sea surface temperature gradients and potentially driving a shift towards more extreme El Niño events (Wang et al. 2019). However, these trends could be due to natural variability rather than climate change (Capotondi et al. 2015).

There are known problems with climate model simulations of the Pacific stemming from mischaracterization of underlying sea surface temperature patterns (Widlansky et al. 2013; Grose et al. 2014; IPCC 2021). The Pacific Ocean temperatures along the equator tend to be too low and too much rainfall is simulated over a region south of the equator (Samanta et al. 2019). Relatedly, the ITCZ is located too far north in most simulations and the diagonal slope of the SPCZ is typically underestimated (Brown et al. 2013b; Stanfield et al. 2016). These problems, along with uncertainty around the physical processes governing ITCZ and SPCZ behaviour (Byrne and Schneider 2016; Popp and Bony 2019), make projection of climate change impacts difficult. Individual models tend to show a wide range of shifts, often in contradictory directions (Brown et al. 2013b; Byrne et al. 2018). As a result, the ensemble average shifts are often close to zero, but this should not be interpreted as a robust projection of minimal change (Evans et al. 2016). Rather, it suggests that large shifts are plausible but current projections are highly uncertain.

3.2 Future Drought and Rainfall Projections

Because average rainfall across the Pacific is heavily influenced by the ITCZ and SPCZ, future changes in the convergence zones and ENSO will affect the hydrology of the region. Unfortunately, as discussed in Sect. 5.3.1, there is little consensus around climate change impacts on these large-scale atmospheric drivers so future rainfall shifts are also uncertain. The problem is complicated further for islands where the terrain (e.g., mountains) impacts rainfall, since GCMs are run across grids with cells in the order of 100 km2 and cannot simulate these effects (Nurse et al. 2014). In general, increased average rainfall is expected near the equator, with decreases more likely in the subtropics related to changes in atmospheric circulation (Australian Bureau of Meteorology and CSIRO 2014; CSIRO and BoM 2015; McGree et al. 2016).

Extended periods of low rainfall can lead to drought, impacting agriculture and water security in the Pacific (McGree et al. 2016). While there is a perception of increased drought risk over recent decades, McGree et al. (2016) found mostly nonsignificant and spatially variable trends. Deo (2011) showed statistically significant downward trends in precipitation across Fiji from 1949 to 2008, but these were mostly driven by very low rainfall between 1969 and 1988. McGree et al. (2014) reported statistically significant decreases in average rainfall from 1951 to 2011 in the South Pacific subtropics, likely due to changes in atmospheric circulation (McGree et al. 2016). Southern Australia has seen drying, with downward trends in rainfall and upward trends in atmospheric evaporative demand (Chiew et al. 2014; Stephens et al. 2018; Delworth and Zeng 2014) and this is projected to continue as the climate warms (Grose et al. 2015). Overall, enhanced drought risk is projected for parts of Australia (CSIRO and BoM 2015) and the northern Cook Islands under a scenario of continued greenhouse gas emissions (Australian Bureau of Meteorology and CSIRO 2014). Uncertain or reduced future drought risk is projected for most small PICTs, but often these assessments consider precipitation changes only (Australian Bureau of Meteorology and CSIRO 2014). If warmer temperatures drive increased evaporative demand in the future, water supplies could be negatively impacted. The sixth IPCC assessment projects drier conditions for the subtropical Southern and Eastern Pacific (IPCC 2021).

Extreme rainfall can be a driver of damaging floods in the Pacific, particularly when combined with wet antecedent conditions (Kuleshov et al. 2014). Extreme rainfall is less dependent on ENSO than average rainfall (McGree et al. 2014), allowing higher certainty around future changes. Because warmer air temperatures enhance the atmosphere’s moisture holding capacity, increases in extreme rainfall are expected globally (Westra et al. 2014). For the PICTs, climate modelling suggests that current 5% Annual Exceedance Probability events, i.e., storms that are currently only seen once every 20 years on average, could occur four times more often in the future (Australian Bureau of Meteorology and CSIRO 2014). Such a change could substantially increase flood risk, especially in areas with increased average rainfall that could simultaneously drive wetter pre-storm soil conditions (Australian Bureau of Meteorology and CSIRO 2011; Wasko and Nathan 2019). However, high natural variability and limited long-term records hamper efforts to detect recent trends in many PICTs. Analysing an unusually long-term streamflow record for Ba River in Fiji (122 years), McAneney et al. (2017) found no significant change in flood risk.

3.3 Tropical Cyclones

Tropical cyclones are behind many natural disasters in PICTs including damaging winds, storm surge, coastal inundation, erosion, and fluvial flooding (Magee et al. 2020). The Western North Pacific region is a global hotspot for cyclone formation due to monsoon trough activity, and tropical cyclones are also a major cause of natural disasters in the South Pacific (Diamond et al. 2013; Keener et al. 2013). A decrease in cyclone frequency, but a greater proportion of major events, has been identified in the South Pacific across several studies (Keener et al. 2013; Deo et al. 2011; Webster et al. 2005), which is broadly consistent with the latest IPCC findings of increased cyclone intensity globally (IPCC 2021). However, high natural variability, operational changes that affect detection rates, and a lack of quality observations complicate cyclone trend analysis (Landsea et al. 2006). Kuleshov et al. (2008) reported a significant increase in severe tropical cyclone activity in the South Pacific, but later found that the trends were actually driven by problems with data reliability (Kuleshov et al. 2010). There are contradictory reports on tropical cyclone trends for the Western North Pacific. Keener et al. (2013) report fewer, but stronger, cyclones while Deo et al. (2011) point to weakened cyclone activity. CSIRO et al. (2015) note that contradicting trends in cyclone activity in the Western North Pacific region have been found by researchers using different datasets. The sixth IPCC report notes a north-westward shift in tropical cyclone tracks in the western North Pacific since the 1980s, but there is low confidence that this was caused by anthropogenic climate change.

Theoretically, fewer cyclones are expected in a warmer future climate because the minimum sea surface temperatures required to drive deep convection will increase (Walsh et al. 2012). Analysis based on satellite observations suggests that this threshold is already rising (Johnson and Xie 2010). Climate simulations largely produce decreased cyclone frequency and increased average intensity across the globe (IPCC 2021), but with regional variation (Knutson et al. 2015; Peduzzi et al. 2012). Decreased cyclone frequency is projected across the Pacific, with higher confidence for the South East than the North West region (Australian Bureau of Meteorology and CSIRO 2011). Wide-ranging projections encompassing increases and decreases in average intensity have been reported for the South Pacific (Knutson et al. 2015; Walsh et al. 2012); an increase in average intensity is likely in the North West Pacific (Knutson et al. 2015).

3.4 Warming Trends and Heatwaves

Marine heatwaves can cause dramatic impacts on aquatic biodiversity (Oliver et al. 2018; Smale et al. 2019), including damage to coral reefs and the fisheries they support. Extreme warming in 2015/2016, driven by the combined effects of anthropogenic climate change and El Niño, caused severe coral bleaching around many small PICTs and off the eastern coast of Australia (Hughes et al. 2018). In a global study, Smale et al. (2019) noted particularly high ecosystem vulnerability in the southwest Pacific due to strong projected warming and prevalence of species living near their warm range limits. High sea surface temperatures can cause a lag in the larval supply of reef fish and negatively impact spawning, as well as driving seagrass die-off (Nurse et al. 2014). Small island communities could be severely impacted because they rely on these ecosystems for services such as coastal protection, fishing, and tourism (Nurse et al. 2014).

Mean sea surface temperatures in the western Pacific have increased at a rate of around 0.3 ˚C per decade since the 1980s, in line with some of the strongest trends globally, but relatively little change has been observed in the eastern Pacific (Oliver 2019). In fact, strengthening of the trade winds across the eastern Pacific Ocean and associated cool conditions have been identified as a cause of the well-known global warming hiatus (England et al. 2014), when global temperatures remained steady for about a decade from the early 2000s despite increased incoming longwave radiation. Similar patterns of strong warming in the western Pacific and little change in the eastern Pacific were found by Australian Bureau of Meteorology and CSIRO (2011) and Oliver et al. (2018), who also pointed to large increases in heatwave frequency in the western Pacific. Note that trends in marine heatwaves are generally defined with reference to pre-warming temperature extremes and are largely driven by changes in the mean temperature, as opposed to higher variability around the mean. In fact, variance in sea surface temperature has not changed substantially across the Pacific except in the far north (Oliver 2019). Because baseline variability in annual sea surface temperature is low in the tropics, particularly in the Western Pacific Warm Pool, small changes in mean temperature lead to large changes in the probability of exceeding an extreme temperature threshold (Frölicher et al. 2018). This concept is illustrated in Fig. 5.2. The latest IPCC report notes that the tropical Western Pacific region experienced warming in both maximum and minimum temperature extremes over 1951 to 2011, and projects further increases in heatwave frequency, intensity, and duration (IPCC 2021).

Fig. 5.2
Two waveforms for before increase and after increase in high variability and low variability follow a fluctuating trend. For high variability, the waves have higher amplitude, while the opposite is for low variability. A dotted horizontal line of extreme threshold is at the top. The wave for after increase in low variability has all the peaks above the extreme threshold.

High and low variability time series subject to the same increase in mean (blue = before increase, red = after increase). Both series cross the extreme threshold only briefly in the initial (blue) case. However, for the series with increased mean (red), the high variability series is above the threshold less frequently and for less time than low variability series

Full size image

As discussed in Sect. 5.3.1, complex large-scale dynamics present a challenge for climate modelling across the Pacific, and this leads to inter-model spread in future projections of Pacific Ocean temperature (Ying 2020). However, the direction of change (warming) is modelled consistently. Frölicher et al. (2018) project large increases in marine heatwave likelihood, especially in the western equatorial Pacific where temperatures exceeding the preindustrial 99th percentile could become 60 times more likely under 3.5 ˚C global warming. Oliver et al. (2019) use CMIP5 results to show that, under the same high emissions scenario, the tropical Pacific could experience permanent marine heatwaves (i.e., entire years where the temperature never drops below the pre-warming 90th percentile) within the next 20 years. Future ocean warming will drive serious ecological impacts and have consequences for tourism and fisheries. For example, Lehodey et al. (2013) project declining skipjack tuna populations in the western tropical Pacific starting in the mid-twenty-first century as conditions become too warm for spawning.

Land temperatures across small PICTs are strongly influenced by the surrounding oceans. Although parts of the Pacific are seeing some of the strongest ocean warming, continental land surfaces are expected to warm at a greater rate than oceans (Byrne and O’Gorman 2013). Therefore, future warming across small island nations is likely to be relatively modest compared to warming in other countries, even in the western equatorial Pacific where increases of 1.5–2 ˚C (low emissions scenario) or 2.5–3 ˚C (high emissions scenario) are projected (Australian Bureau of Meteorology and CSIRO 2011). However, due to the limited area of small islands, temperature-driven range shifts could drive endemic species to extinction (Nurse et al. 2014). Larger countries in the Pacific are likely to warm at a greater rate; under a high emissions scenario, warming of 4–5 ˚C is projected for most of Australia and 3 to 4 ˚C is projected for most of New Zealand and Papua (IPCC 2013). Warm season daily maximum temperatures could become more variable across Australia, especially in mid-latitude Western Australia and in eastern Australia inland from the Great Dividing Range (Meehl and Tebaldi 2004). These temperature changes could drive increased drought and wildfire risk. Instances of fauna death due to temperature extremes are already being reported (Saunders et al. 2011; Welbergen et al. 2008).

3.5 Sea Level Rise

In small PICTs, sea level rise is expected to drive some of the most detrimental climate change impacts. These could include storm surge flooding, saltwater intrusion in groundwater, salinization and degradation of soils, loss of beaches damage to coastal habitat, and even permanent displacement of communities (CSIRO et al. 2015). Often these problems will be caused by sea level rise together with confounding factors such as increased coastal development pressure, more intense storms, and/or erosion by human activities (Nurse et al. 2014; Storey and Hunter 2010). For example, displacement of a village on the Torres Islands (Vanuatu) in the early 2000s was driven by a combination of sea level rise and tectonic subsidence (Ballu et al. 2011). Residents on the Cook Islands have reported increased sediment transport, rougher ocean conditions, stronger ocean currents, and shorter low tide durations, which seem to be impacting the productivity of fisheries (Rongo and Dyer 2014). Water security of islands reliant on thin freshwater lenses may be particularly vulnerable to sea level rise, although these lenses have been shown to rise with sea level in places where the geology is favourable (Nurse et al. 2014).

Global sea level rise has been attributed to loss of ice sheets, melting glaciers, and ocean thermal expansion with unsustainable groundwater pumping also causing localized exacerbation (Cazenave and Llovel 2010). Prior to the industrial revolution, sea levels were relatively stable for millennia (Cazenave and Llovel 2010). However, in recent decades, they have been rising at an increasing rate, reaching approximately 3.7 mm/year between 2006 and 2018 (IPCC 2021). The rate in the Pacific has been much higher, with a rise of around 12 mm/year observed in the tropical western Pacific between 1993 and 2009 (Nurse et al. 2014). This faster rate is associated with ENSO shifts that can drive variations of 20–30 cm, meaning that it does not reflect the expected long-term trend (Becker et al. 2012), but natural variation is expected to be overwhelmed by warming-induced sea level rise in the future (IPCC 2021). The recent combination of natural and warming-induced sea level rise has already led to dramatic consequences. For example, salinization of water sources and soils caused significant human migration from some outer islands of Tuvalu to the main island, which now supports over 6000 people within a 2.8 km2 land area (Becker et al. 2012). Widespread inundation by extreme waves across Papua New Guinea and the Solomon Islands in 2008 was exacerbated by concurrent high sea levels (Hoeke et al. 2013).

Future sea level rise depends heavily on the emissions trajectory, highlighting the importance of global coordination in urgently cutting global greenhouse gas releases to prevent the worst impacts on PICT communities. By 2100, a high emissions scenario will likely drive global sea level rise of 0.63 to 1.02 m, while a low emissions scenario would probably limit the rise to between 0.28 and 0.55 m (IPCC 2021). Sea level rise projections for high emissions scenarios are uncertain beyond 2100 due to the impact of ice-sheet responses to warming, and it is possible that sea levels could rise up to 5 m by 2150 (IPCC 2021). Average and extreme sea levels are expected to rise variably across the world due to gravitational effects between ice sheets and ocean waters, differing wind dynamics, and shifts in wave climate (Walsh et al. 2012). Models project that differential warming patterns will strengthen the south-easterly trade winds while the trade winds closer to the equator weaken; this could drive slightly faster sea level rise around Vanuatu, New Caledonia, and Fiji, but slower sea level rise around Tuvalu, Kiribati, the Cook Islands, and the Solomon Islands, compared to the global average (Timmermann et al. 2010; Walsh et al. 2012).

Sea level rise clearly presents a threat to communities in the Pacific, especially on small islands and low-lying atolls (Mimura 1999) and in coastal areas vulnerable to storm surge (IPCC 2021). However, there are examples of communities and natural resources showing remarkable resilience. Rapid sea level rise around Tuvalu has coincided with many islands (counterintuitively) becoming larger, likely due to wave and sediment supply processes (Kench et al. 2015). As the ocean continues to rise, it is likely that small sand islands in Tuvalu will shrink but medium and larger islands may continue to expand (Kench et al. 2018). Similar variability in how islands respond to sea level rise has been found in French Polynesia and the Marshall Islands (Nurse et al. 2014). This highlights that islands are not static and, if global greenhouse gas emissions are appropriately reduced, island abandonment may be avoidable even for seemingly vulnerable low-lying atolls (Barnett 2017). Preserving and restoring natural barriers to erosion, including mangroves and coral reefs, will help to maintain islands in the face of sea level rise (Mimura 1999).

3.6 Ocean Acidification

Around 30–40% of anthropogenic carbon dioxide emissions over the last century have been absorbed by the oceans, leading to chemical changes that have increased acidity (Lebrec et al. 2019; Wei et al. 2009). This has negative implications for marine environments, especially coral reefs that are critical for fisheries, tourism, shoreline protection, and biodiversity in the tropical Pacific (Shaw et al. 2015; Doney et al. 2009). For example, tourism generates 40% of employment in Palau, and 86% of tourists visit to see coral reefs (Lebrec et al. 2019). Long-term reef monitoring on Rarotonga (Cook Islands) indicates that reef recovery after bleaching is becoming slower (Rongo and Dyer 2014). On Manihiki (Cook Islands), pearl farmers report that oyster shells are becoming thinner and they have noticed increasing pearl deformities (Rongo and Dyer 2014).

Ocean acidity is measured using pH, a logarithmic scale that decreases with increasing acidity. As pH declines, the concentration of calcium carbonate (CaCO3) in seawater also declines. CaCO3 may take the form of calcite or aragonite—both minerals are important for calcifying organisms such as corals and some krill (Albright et al. 2010; De’ath et al. 2009). CaCO3 is most abundant in shallow, tropical marine environments, so the worst effects of ocean acidification are likely to be felt in the higher latitudes first (Orr et al. 2005) and not in the equatorial Pacific. Lebrec et al. (2019) found that large-scale coral damage due to acidification has not yet been observed in the Marshall Islands, Palau, or American Samoa. In fact, some reefs in the tropical Pacific are adapted to unusually high and/or variable acidity; Nikko Bay in Palau has pH levels similar to those expected globally in 2100 but maintains high coral diversity. However, the impacts of higher marine water temperatures will be felt strongly in the tropics (see Sect. 5.3.4) and acidification may make corals more vulnerable to temperature-driven bleaching (Anthony et al. 2008). De’ath et al. (2009) attributed a 14% decline in calcification across 69 reefs in the Australian Great Barrier Reef to a combination of temperature stress and declining aragonite concentrations.

Recent changes in ocean acidity are uncertain due to a lack of direct long-term measurements (Shinjo et al. 2013; Wei et al. 2009). Bates et al. (2014) analysed datasets 15–30 years in length from seven stations around the world, including one off the east coast of New Zealand. They found statistically significant declines in ocean pH across the sites, but the rate of change varied between locations. Shinjo et al. (2013) used coral isotopes sourced from Guam to create a proxy for pH in the West Pacific Warm Pool, noting a declining trend of 0.05–0.08 pH units over 60 years. Another coral isotope study in the southwestern Pacific pointed to large natural variation in pH over 300 years connected to the IPO (Pelejero et al. 2005), suggesting that marine ecosystems could actually be quite robust to changing pH, but this remains an area of contention (Matear and McNeil 2006; Pelejero et al. 2006). Wei et al. (2009) used a similar method to infer ocean acidity in the central Australian Great Barrier Reef over 200 years. They also found substantial natural pH variability connected to large-scale Pacific climate drivers, but rapid oscillations in pH detected around 1998 coincided with a major bleaching event in the Great Barrier Reef (Wei et al. 2009). This indicates that large or quick changes in pH may overwhelm an ecosystem’s capacity to adapt. The current rate of acidification is believed to be unprecedented within the last 300 million years (Lebrec et al. 2019), and compound stressors like overfishing, rising temperatures, and polluted terrestrial runoff may further amplify the negative impacts of acidification (Halpern et al. 2015, 2019).

As atmospheric carbon dioxide levels rise over the coming decades, oceans are expected to acidify further, and this would greatly impact food security in the Pacific. If emissions continue unabated, surface ocean acidity will be 150% higher than pre-industrial levels by 2100 (Lebrec et al. 2019; Orr et al. 2005). Kleypas et al. (1999) suggest a related decrease in aragonite saturation of 30% by 2050, with serious implications for marine ecosystems. Eyre et al. (2018) project that reef sediments will experience faster dissolution than calcification globally, and therefore become net dissolving, from around 2050. Coral breeding could also be impacted by acidification via reduced fertilization and settlement success (Albright et al. 2010). Impacts are likely to vary across locations depending on the rate of acidification and other compounding stressors, as well as the natural variability to which different organisms have been exposed (Shaw et al. 2015; Anthony et al. 2008).

4 Conclusions and Outlook

Communities in the Pacific are on the climate change frontline, already experiencing enhanced temperature extremes, rising sea levels, and effects of ocean acidification. They face an uncertain future that will likely include heavier extreme rainfall and possibly more intense tropical cyclones. Some countries may also be at greater risk of drought in upcoming decades. Small island nations are particularly vulnerable to climate change due to their limited natural resources and space. Lack of financial resources may also limit their capacity to respond with engineering adaptation approaches such as infrastructure upgrades. Large countries and high emitters clearly have a responsibility to reduce greenhouse gas releases (ultimately to net zero) to prevent the worst impacts on island communities. The global food and energy sectors, which provide vital services but also generate emissions, have an important role to play.

As an anthropogenically affected climate becomes the ‘new normal’, nations and communities in the Pacific will need to adapt. However, adaptation efforts are hampered by especially large uncertainties around future climate due to known deficiencies in climate model simulations of the region. The common approach of trying to foresee and pre-emptively mitigate against climate change impacts could lead to wasted resources in the Pacific if projections turn out to be unreliable (Barnett 2001). Instead, a flexible approach should be considered in which grassroots initiatives are gradually scaled up and adjusted as information becomes available. PICT communities have adapted over generations to high year-to-year climate variability (Barnett 2001) and have built a wealth of valuable traditional knowledge (Mael 2013; Rongo and Dyer 2014), as discussed further in Chap. 11. This can be leveraged for understanding and adapting to climate change by placing local communities at the centre of future planning efforts. Recent research has highlighted that adaptation to high climate variability increases resilience to climate change (Nathan et al. 2019), so programs that fortify communities under current variability will also have future benefits. However, transfer of vital traditional knowledge is being interrupted by a shift towards more westernized lifestyles and targeted data collection must ensure that centuries of learnings are not lost (Rongo and Dyer 2014). Climate change awareness programs may help people place their knowledge in the context of adaptation; in a survey conducted by Rongo and Dyer (2014) in the Cook Islands, over 90% of respondents saw a need for such programs. While climate change undoubtedly threatens PICTs and their water, energy, and food security, urgent emissions reduction combined with context-specific adaptation programs can secure a positive future for local societies and the environment. Well-resourced countries with high historical carbon emissions should play a constructive role in supporting more vulnerable countries, particularly within their own region.