Changes in Oceans Surrounding Canada

Sustained temperature observations in the oceans surrounding Canada began in the early 20th century, but these time series are limited to a few locations. In the Arctic Ocean, there have been very few continuous observations, and those that do exist are limited to the last several decades. Ocean temperature observations have evolved from the 19th century sampling of the ocean from ships to a more systematic and near-global coverage from satellites for surface waters (e.g., Larouche and Galbraith, 2016⁸⁶) and Argo floats (autonomous profilers that measure the temperature and salinity of the upper 2000 m of the ocean) for the deep ocean (Riser et al., 2016¹²⁵). Observations of subsurface ocean temperature on the continental shelves surrounding Canada continue to be acquired primarily through vertical profiles taken by research vessels, supplemented by continuous time series (typically with hourly sampling) from scattered moored instruments. This section will focus on long-term ocean temperature observations collected by Fisheries and Oceans Canada (DFO) monitoring programs, which draw on data from various sources of regular sampling initiated at some sites in the early 20th century. The site-specific time series presented in this section are representative of temperature over broader shelf and open-ocean regions (Ouellet et al., 2011¹¹¹,  Petrie and Dean-Moore, 1996¹¹⁹).

Because the heat capacity of water is much higher than that of air, anthropogenic ocean warming is expected to be somewhat less than that in the lower atmosphere over land, except possibly in some places where there are changes in ocean circulation (e.g., the warm Gulf Stream shifting northward). Projections from models in the fifth phase of the Coupled Model Intercomparison Project (CMIP5; see Chapter 3, Box 3.1) used in the Intergovernmental Panel on Climate Change’s (IPCC’s) Fifth Assessment Report (AR5) generally indicate widespread warming of the upper oceans around Canada during the 21st century, with greater warming for higher emission scenarios. Substantial variability is evident between seasons and from one region to another (Loder et al., 2015⁹⁴; Christian and Holmes, 2016¹⁷, Steiner et al., 2015¹³⁷; Christian and Foreman, 2013¹⁶). Projected changes in SST to mid-century (average for 2046–2065 relative to that for 1986–2005) for a high emission scenario (RCP8.5) have been computed as the ensemble mean of six of the CMIP5 models (Loder et al., 2015⁹⁴). Global emissions since 2005 (e.g., Peters et al., 2013¹¹⁷; 2017¹¹⁸), and climate-policy decisions (e.g., Sanford et al., 2014¹²⁹) have been closer to this scenario than to the low emission scenario (RCP2.6). The projected mid-century SST increases for the medium emission scenario (RCP4.5) are about 70% of those for RCP8.5 with similar spatial patterns, consistent with the scalability of projected air temperature changes discussed in Chapter 4 (also see Markovic et al., 2013¹⁰⁰). As a good approximation, these projected increases can be taken to apply until mid-century, assuming only limited further reductions in emissions.

In the Northeast Pacific off British Columbia, the projected SST increases to mid-century are roughly 2°C in winter and 3°C in summer, with a small and smooth increase with latitude (see Figure 7.8). In contrast, the projected increases in Canadian Arctic waters (including Hudson Bay) and the Northwest Atlantic have larger seasonal and spatial variations. The projected SST changes in the Arctic in winter are very small (because of the projected continued presence of winter sea ice), but those in summer are up to 4°C in areas such as the Beaufort Sea and Hudson Bay, where reduced sea ice cover is projected. The CMIP5 models do not have adequate spatial resolution and representations of sea ice and ocean physics in the complex Canadian Arctic Archipelago to reliably project the details of ocean temperature changes in summer and fall there, but substantial spatial structure in the ocean changes associated with changes in sea ice can be expected (e.g., Sou and Flato, 2009¹³⁵; Hu and Myers, 2014⁷⁰; Steiner et al., 2015¹³⁷). Reliable projections of ocean conditions in this region will probably require a combination of higher spatial resolution in global climate models and inclusion of both sea ice and ocean components in the regional climate models used in dynamical downscaling (see Chapter 3.5).

Air temperature increases are projected to be larger than the SST increases in most parts of the Northwest Atlantic (Loder et al., 2015⁹⁴), consistent with atmospheric warming being the primary driver of ocean warming (e.g., Collins et al., 2013²⁰; Hegerl et al., 2007⁶⁷). The latitudinal variation of future SST change in offshore waters will be different from the increase with latitude of air temperature over the land associated with Arctic amplification. In contrast to air temperature for Canada as a whole, the SST increase in the Northwest Atlantic is projected to be largest at mid-latitudes and smaller proceeding north into subpolar waters. Winter SST increases by mid-century of up to 3°C off the Maritime provinces, but only 1°C off Labrador, are projected for the high emission scenario (RCP8.5). Similarly, mid-century summer SST increases of up to 4°C are projected off the Maritime provinces, but increases are limited to 2°C off Labrador. The mid-latitude maximum in the SST increase is related to projected changes in large-scale ocean circulation and to a slight northward expansion of the subtropical gyre (and shift of the Gulf Stream), in particular.

In the North Atlantic south of Greenland, most models indicate that future warming will be more limited, with the Atlantic Meridional Overturning Circulation transporting less heat northward (Drijfhout et al., 2012³²). However, substantial uncertainty remains about the potential for significant reduction of this circulation in the future, due to the complexity of the atmosphere–ice–ocean system in the Northwest Atlantic and the limited capability of current climate models to simulate important processes in this complex system (Sgubin et al., 2017¹³¹).

As is the case for the Canadian Arctic Archipelago, the coarse horizontal resolution of the ocean in the CMIP5 Earth system models (of approximately 100 km) poses a challenge for modelling the ocean off Atlantic Canada, where the coastline and seafloor topography are complex. This results in a warm bias in SST due to a misrepresentation of the boundary between the subpolar and subtropical gyres; thus, existing climate change projections are based on a modelled regional ocean circulation that differs from current reality (Loder et al., 2015⁹⁴; Saba et al., 2016¹²⁷). This is important for Atlantic Canada, in particular, which is in a region of large spatial differences in ocean temperature (Figure 7.1). Regional climate downscaling has provided detailed information on the spatial structure of potential changes for Atlantic Canada (Long et al., 2016⁹⁶), but the overall magnitude of the changes is uncertain.

In summary, upper-ocean temperature has increased in the Northeast Pacific and most areas of the Northwest Atlantic over the last century, consistent with anthropogenic climate change (high confidence). This statement of confidence is based on high-quality in situ observations of sea surface and subsurface temperature, which are generally consistent with the regional variations in global interpolated SST datasets. The number of locations with long subsurface time series is limited and, although these data are expected to be representative of large areas, there is lower confidence in them. Natural decadal variability is comparable in magnitude to the long-term changes in ocean temperature; there is a region south of Greenland where there has been little or no warming over the last century. There are no long-term ocean temperature measurements for the Arctic Ocean, but warming is expected to have occurred in the summer and fall periods, based on observed increases in air temperature (see Chapter 4, Section 4.2.1.1) and declines in sea ice (see Chapter 5, Section 5.3.1) (medium confidence). This statement of confidence is based on a few short temperature time series and on expert judgment of the coupling of the atmosphere, cryosphere, and upper ocean.

Oceans surrounding Canada are projected to continue to warm over the 21st century in response to past and future emissions of greenhouse gases. The warming in summer will be greatest in the ice-free areas of the Arctic and off southern Atlantic Canada where subtropical water is projected to shift further north (medium confidence). During winter in the next few decades, the upper ocean surrounding Atlantic Canada will warm the most, the Northeast Pacific will experience intermediate warming rates, and the Arctic and eastern sub-Arctic ocean areas (including Hudson Bay and Labrador Sea) will warm the least (medium confidence). These statements of confidence are based on an analysis of six CMIP5 model projections of SST for the oceans surrounding Canada, which show an increase in SST in all seasons in the Northeast Pacific and Northwest Atlantic oceans. The statements are also based on physical understanding of the processes related to increasing surface air temperature, resulting in a positive heat transfer to the ocean surface waters. The level of confidence is medium rather than high owing to differences in the regional projections from the Earth system models.

Ocean salinity observations have been made since the late 19th century by research cruises. The coverage of these observations is, however, more sparse than observations of temperature, as salinity is more difficult to measure than temperature. Observations of ocean salinity on the continental shelves surrounding Canada are primarily acquired through vertical profiles taken by research vessels, supplemented by continuous time series from sparse moored instruments.

In the global context, the CMIP5 climate model projections suggest that subtropical regions with high sea surface salinity, dominated by net evaporation, will become more saline as the century progresses. High-latitude regions with lower sea surface salinity are projected to freshen over the coming century (Collins et al., 2013²⁰).

For the northeastern Pacific off Canada, future projections show significant freshening by mid-century (see Figure 7.12), with little change in the spatial structure under either a medium (RCP4.5) or high (RCP8.5) emission scenario (Christian and Foreman, 2013¹⁶).

Significant freshening by mid-century for the subpolar Northwest Atlantic Ocean is also projected under medium (RCP4.5) and high (RCP8.5) emission scenarios (see Figure 7.12; also Loder et al., 2015⁹⁴). On the other hand, increased salinity is projected in the subtropical gyre, thereby increasing the difference in salinity between the two gyres of the North Atlantic Ocean. The increased difference is important, because small changes in the boundary between the gyres will result in shifts in local salinity (and, potentially, stratification and circulation). The boundary for the shift from increasing to decreasing salinity trends generally lies around 40° north latitude (Loder et al., 2015⁹⁴) but there are significant differences among projections for this region from different CMIP5 models; therefore, confidence in the pattern of future projections of sea surface salinity is low. A high-resolution climate model projects significantly larger changes in salinity on the ocean bottom on the continental shelf in southern Atlantic Canada, (i.e., Scotian Shelf), suggesting that the CMIP5 climate change projections for the Northwest Atlantic shelf between Cape Hatteras and the Grand Banks may underestimate expected changes in salinity (Saba et al., 2016¹²⁷). The CMIP5 global models do not resolve the topography of the continental shelf or the spatial structure of the ocean overlying the shelf. CMIP5 global models also do not properly resolve the Gulf Stream separation off Cape Hatteras, North Carolina; therefore, the position of the Gulf Stream is too far north in the models’ simulations of past and present-day regional ocean climate.

Projected continued losses of sea ice will add fresh meltwater to the ocean (see Box 7.3) which, combined with projected increased precipitation (see Chapter 4, Section 4.3.1.3), will affect the freshwater input into the Arctic Ocean. The CMIP5 global model simulations project a fresher (decrease of approximately 2 by mid-century) near-surface ocean in the Beaufort Sea and area north of the Canadian Arctic Archipelago under the high emission scenario (RCP8.5) (see Figure 7.12). The spatial pattern of surface salinity shows increased freshening with distance north from the coast in the Beaufort Sea (Steiner et al., 2015¹³⁷). A high-resolution model simulation for the Canadian Arctic Archipelago projects strong decadal variability in surface salinity but no clear trend by mid-century (Hu and Myers, 2014⁷⁰). The southward transport of freshwater that is currently tied up in sea ice in the Arctic will be a contributor to the southward extent of low-salinity water off Atlantic Canada. With less seasonal sea ice, this transport mechanism is expected to weaken and, once there is no seasonal ice cover, eventually disappear.

In summary, there has been a slight long-term freshening of upper-ocean waters in most areas off Canada as a result of various factors related to anthropogenic climate change, in addition to natural decadal-scale variability (medium confidence). Salinity has increased below the surface in some mid-latitude areas, indicating a northward shift of saltier subtropical water (medium confidence). These statements of confidence are based on agreement among high-quality in situ observations of surface and subsurface salinity, available from DFO databases. The number of locations with long time series is more limited than those of ocean temperature, and this reduces confidence in the broader-scale representativeness of trends. Natural decadal variability is comparable in magnitude to the long-term changes in ocean salinity in most areas, which also reduces confidence in trends. Observations from the Arctic Ocean as a whole indicate freshening in most areas, but increased salinity in some others. Given the lack of data, no confidence statement has been made about climate change trends for ocean salinity in the Arctic.

Freshening of the ocean surface is projected to continue in most areas off Canada over the rest of this century under a range of emission scenarios, due to increases in precipitation and melting of land and sea ice (medium confidence). However, increases in salinity are expected in off-shelf waters south of Atlantic Canada due to the northward shift of subtropical water (medium confidence). The upper-ocean freshening and warming is expected to increase the vertical stratification of water density, which will affect ocean sequestration of greenhouse gases, dissolved oxygen levels, and marine ecosystems. These statements of confidence are based on the analysis of six CMIP5 model projections of sea surface salinity for the oceans surrounding Canada, and regional model studies. There are differences in the magnitude of salinity change among the CMIP5 model projections in the Northwest Atlantic, which means there is more uncertainty in projections for this region.

As is the case globally, the assessment of historical changes in winds and storms for the oceans surrounding Canada is hampered by limited evidence, in part related to sparse observations and challenges in integrating early marine observations, instrumental records, and satellite data. However, there is evidence of a slight northward shift of storm tracks of about 180 km over the North Atlantic Ocean (60° west to 10° east) and about 260 km for Canada as a whole (120° west to 70° west) for the 1982–2001 period relative to 1958–1977 (Wang et al., 2006¹⁵²). This trend is consistent with global assessments, which have observed a poleward shift of storm tracks and the jet stream since the 1970s (Wu et al., 2012¹⁵⁹; Hartmann et al., 2013⁶²), and is projected to continue through this century (Collins et al., 2013²⁰). The poleward shift results in a modest projected decrease in wind speed and wave heights over marine areas in Atlantic Canada (Casas-Prat et al., 2018¹²).

An increasing trend in the frequency of autumn (October–December) extreme storms (low-pressure systems of core pressure less than 980 hPa) over the 1958–2010 period has been observed over marine areas of Atlantic Canada, but there are no statistically significant trends for extreme storms in other seasons for the Atlantic and Pacific coasts of Canada (Wang et al., 2016¹⁵²). This is consistent with research that has demonstrated that human activities have contributed to an observed upward trend in North Atlantic hurricane activity since the 1970s (Kossin et al., 2017⁸⁵). Model projections of late-summer and autumn storms off Atlantic Canada suggest a slight northward shift of storm tracks and a modest reduction in intensities of storms, although extreme storms may have increased intensities (Jiang and Perrie, 2007⁸¹, 2008⁸²; Perrie et al., 2010¹¹⁶; Guo et al., 2015⁵⁴).

For the Arctic above 75° north latitude, an increasing trend in storm frequency and intensity has been seen in all long-term datasets that cover 1958–2010 or 1900–2010 periods (see Wang et al., 2016¹⁵²). This trend is independent of different storm identification and analysis methods and is consistent with the increasing trend in ocean surface wave heights in this region, as seen in satellite data (Francis et al., 2011⁴²; Liu et al., 2016⁹¹) and wave reanalysis data (Wang et al., 2015¹⁵⁵; see also Section 7.4.2). However, observations are sparse in the Arctic region, which lowers our confidence in storminess trends in this region. Increases in surface wind speed over Canadian sectors of the Arctic Ocean are projected, largely related to the projected declines in sea ice (Casas-Prat et al., 2018¹²).

Waves are an important physical feature of the ocean surface that affects fluxes of energy, heat, and gases between the atmosphere and ocean, as well as marine safety and transportation. Surface waves are generated by wind forcing, and “significant wave height” is a measure that is approximately equal to the average of the highest one-third of wave heights. Global and regional time series of wave characteristics are available from buoy data, voluntary observing ship reports, satellite measurements, and model wave reanalysis/hindcasts (i.e., simulations of past conditions using observations of other climate variables).

In the Arctic, over the 1970–2013 period, significant wave heights have increased over the Canadian Beaufort Sea westward to the northern Chukchi Sea in September, with the Beaufort–Chukchi–Siberian Seas regional mean significant wave height increasing at a rate of 3% to 8% per decade in July–September (Wang et al., 2015¹⁵⁵). These trends suggest that increasing wave energy could constitute a mechanism to break up sea ice and accelerate ice retreat (Thomson and Rogers, 2014¹⁴⁰; Wang et al., 2015¹⁵⁵); however, the rate of sea ice reduction could also be enhanced by wave mixing in the upper ocean, causing an added release of heat (Smith et al., 2018¹³⁴). For areas experiencing loss of sea ice (see Chapter 5, Section 5.3), significant seasonal increases in waves are projected for the future (Casas-Prat et al., 2018¹²). Reduced sea ice cover will result in greater distances of open water for waves to travel across and, with a southward mean wave direction for the Arctic Ocean, this will result in increased wave impacts on coastal infrastructure and communities in the Canadian Arctic.

For the waters off the Pacific coast, an analysis of buoy wave records revealed that wave heights in the region off British Columbia have decreased significantly over the past three to four decades in summer and increased slightly in winter, showing small decreasing annual trends (Gemmrich et al., 2011⁴⁷). The same trends and trend seasonality are evident in other studies (Wang and Swail, 2001¹⁵⁰) and are also projected to continue into the future (Wang et al., 2014¹⁵⁴; Casas-Prat et al., 2018¹²; Erikson et al., 2015³⁶). The wintertime increase in wave height in this region is also seen in observations from voluntary observing ships (VOS) for 1958–2002, but these results show much larger increases (Gulev and Griforieva, 2006⁵³). The reason for the difference between the VOS and other results is uncertain.

Over the past half-century, the large-scale pattern of North Atlantic wave heights is characterized by increases in the Northeast Atlantic, with decreases in the mid-latitude North Atlantic in winter (Wang and Swail, 2001¹⁵⁰, 2002¹⁵¹; Wang et al., 2012¹⁵³; Bromirski and Cayan, 2015⁸). For the waters off Atlantic Canada, small increases (around 2 cm per decade) in summertime wave heights and insignificant wintertime decreases were observed for the 1948–2008 period (Bromirski and Cayan, 2015⁸). Similar trends are also seen in other observational wave studies (Wang and Swail, 2001¹⁵⁰, 2002¹⁵¹). These results differ from VOS observations for 1958–2002, which show wintertime increases of around 0.1 m per decade for the waters off Atlantic Canada (Gulev and Griforieva, 2006⁵³), and the reason for this discrepancy is unclear. Modest decreases in wave height in the region off Atlantic Canada are projected over the coming century (Wang et al., 2014¹⁵⁴; Casas-Prat et al., 2018¹²). In the Gulf of St. Lawrence, downscaled projections suggest decreased mean significant wave heights in summer and increased wave heights in winter, with reduced seasonal sea ice playing an important role (Long et al., 2015⁹⁵; Perrie et al., 2015¹¹⁵; Wang et al., 2018¹⁴⁹)

In summary, consistent significant trends in winds, storminess, and waves have not been found for most of the waters off Canada, in part due to limited data and strong effects of natural variability. Long-term data are very limited, tend to have very coarse spatial resolution, and do not cover nearshore areas. A slight northward shift of storm tracks, with decreased wind speed and lower wave heights off Atlantic Canada, has been observed and is projected to continue (low confidence). Off the Pacific coast, wave heights have been observed to increase in winter and decrease in summer, and these trends are projected to continue in future (low confidence).These confidence statements reflect the limited amount of published literature specific to winds and waves in the marine regions off Canada, the lack of high-quality historical data, and discrepancies in trends derived from different datasets.

Surface wave heights and the duration of the wave season in the Canadian Arctic have increased since 1970 and are projected to continue to increase over this century as sea ice declines (high confidence). Off Canada’s east coast, areas that currently have seasonal sea ice are also anticipated to experience increased wave activity in the future, as seasonal ice duration decreases (medium confidence). This key message is based on limited wave time series in regions with seasonal ice coverage and a few published regional studies; however, there is strong evidence of past trends and future projections of declines in sea ice in both the Arctic and Atlantic Canada (see Chapter 5, Section 5.3). Increased wave activity resulting from sea ice decline is based on modelling results and expert judgment regarding the understanding of air–sea interaction processes.

Globally, for most of the 20th century (up to 1990), sea level rose at a mean rate slightly larger than 1 mm/year (average [90% uncertainty range]: 1.2 [1.0 to 1.4] mm per year [Hay et al., 2015⁶⁴]; 1.1 [0.5 to 1.7] mm per year [Dangendorf et al., 2017²⁵]). Recently, the rate of mean sea-level rise has increased, and the rate of global mean sea-level rise after 1993 is nearly three times as large (average [90% uncertainty range]: 3.0 [2.3 to 3.7] mm per year, 1993–2010 [Hay et al., 2015⁶⁴]; 3.1 [0.3 to 5.9] mm per year, 1993–2012 [Dangendorf et al., 2017²⁵]).

The long-term trends in relative sea level observed at tide gauges in Canada vary substantially from one location to another. Some of the variability is due to oceanographic factors affecting the absolute elevation of the sea surface, but a major determinant of relative sea-level change in Canada is vertical land motion. Land subsidence (sinking) increases relative sea-level, while land uplift does the opposite. Across much of Canada, land uplift or subsidence is mainly due to the delayed effects of the last continental glaciation (ice age), called glacial isostatic adjustment (GIA). GIA is still causing uplift of the North American continental crust in areas close to the centre of former ice sheets, such as Hudson Bay, and subsidence in regions that were on the edge of former ice sheets, such as the southern part of Atlantic Canada, as shown in Global Positioning System (GPS) data (see Figure 7.13). On the west coast, active tectonics, and, on the Fraser delta, sediment consolidation (Mazzotti et al., 2009¹⁰¹), contribute to vertical land motion.

In the Atlantic region, vertical measured land motion ranges from uplift rates of about 1–4.5 mm per year for Quebec sites to subsidence of up to about 2 mm per year at some locations in Nova Scotia (see Figure 7.13). On the west coast of Canada, vertical motion rates vary from negligible values near Vancouver to uplift of almost 4 mm per year in the middle part of Vancouver Island, and smaller rates of uplift further north. The largest variation in vertical land motion is observed in the Arctic. Hudson Bay coastlines are rising at a rate of 10 mm per year or more. Significant portions of the Canadian Arctic Archipelago coastline are uplifting at a rate of a few millimetres per year from a combination of GIA and the response of Earth’s crust to present-day changes in ice mass, whereas the Beaufort Sea coastline in the western Arctic is subsiding due to GIA at a rate of 1–2 mm per year.

The effects of vertical land motion are evident in tide gauge records (see Figure 7.14). Where the land is uplifting rapidly due to GIA, such as at Churchill, Manitoba (on Hudson Bay), sea level has been falling rapidly, at a rate of 9.3 mm per year. Where the land is sinking due to GIA, such as much of the Maritimes, southern Newfoundland, and along the Beaufort Sea coast of the Northwest Territories and Yukon, sea level is rising faster than the global average. At Halifax, sea level rose at a rate of about 3.3 mm per year during the 20th century.

Projections of relative sea-level changes for coastal Canada, based on the CMIP5 model projections used in IPCC AR5 (Church et al., 2013¹⁸), take into account projections of global sea-level change, vertical land motion, dynamic oceanographic changes, and redistribution of meltwater from glaciers, ice caps, and ice sheets in the oceans (James et al., 2014⁷⁸, 2015⁷⁹; Han et al., 2015b⁶⁰, 2015c⁶¹; Zhai et al., 2015¹⁶⁹; Lemmen et al., 2016⁸⁷). The following gives a brief description of the factors contributing to sea-level change.

Ocean-surface heights vary on timescales from seconds to hours to years, due to waves, tides, and atmospheric and ocean circulation. These fluctuations may arise from the large-scale modes of internal climate variability (ENSO, Pacific Decadal Oscillation, and North Atlantic Oscillation events; see Chapter 2, Box 2.5), seasonal warming and runoff, storms, and changes to ocean circulation. Extreme ENSO events can result in coastal sea-level changes of a few tens of centimetres (see Figure 7.14; see the high-water levels at the British Columbia sites at the end of 1997 and beginning of 1998). The ENSO cycle may intensify with global warming (Cai et al., 2014¹⁰), and this could generate larger peak water levels during El Niño events on Canada’s west coast. Together, these factors, superimposed on the tidal cycle, produce variability that causes peak water levels to vary substantially throughout the year, and from year to year.

One of the most serious consequences of sea-level rise is its effect on extreme high coastal water-level and flooding events. These events are typically associated with storm surges that coincide with high tides (see Box 7.5). Storm surges can have heights of 1 m or more above high tide levels (Bernier and Thompson, 2006⁵; Han et al., 2012⁵⁹; Ma et al., 2015⁹⁸; Manson and Solomon, 2007⁹⁹; Thomson et al., 2008¹⁴¹), with wave run-up further adding to the extent of flooding. Where relative sea level is projected to rise, extreme high water levels (combined tide and surge) are expected to be even higher and more frequent in the future.

The effect of sea-level rise on extreme water levels is illustrated for Halifax (see Figure 7.19). Sea-level has been rising at Halifax, and the number of water levels exceeding the 2.3 m flood level (red line in Figure 7.19) has increased through the 20th and early 21st century. The record shows that, for this particular flood level, 131 flooding events have occurred in the historical record (1901–2018), while for a 2.1 m flood level (magenta line) there have been 596 flooding events, which is more than four times as many. A 20 cm rise in mean sea level, which is projected to occur within two to three decades at Halifax for all emission scenarios (Figure 7.17), can therefore be anticipated to increase 2.3 m flooding events at this location by about the same factor of four. Generally, a projected rise in mean sea level is expected to increase the number of extreme water-level events at a given flood level as well as increase the largest flood height (Church et al., 2013¹⁸). For example, large, impactful events, such as the water level reached once every 50 years at Halifax in the past, may occur as frequently as every two years by mid-century under the relative sea-level rise caused by a high emission scenario (Atkinson et al., 2016²).

Increases in storm frequency or intensity would contribute to further increases in the occurrence of extreme high water-level events; however, projecting such increases is difficult because region-specific projections of storminess are not robust (Hartmann et al., 2013⁶²; see Section 7.4.1). While more thermal energy in a warmer atmosphere is projected to lead to increased storminess on a global scale, storminess may or may not increase in any given region, depending on storm-source regions and storm tracks. Projections of changes to wave height in the oceans surrounding Canada are also uncertain (see Section 7.4.2), but where winds and wind-driven waves increase, wave set-up and run-up (the maximum level waves reach) will also increase (see Box 7.5). Larger waves generally have greater erosive power and damage potential.

Reductions in sea ice cover (see Chapter 5, Section 5.3) also have important implications for wind-driven waves (see Section 7.4.2), storm surges, and extreme high water levels. Sea ice in the nearshore prevents waves from breaking directly onshore and reduces wave run-up (Forbes and Taylor, 1994⁴¹; Allard et al., 1998¹). Ice further offshore reflects waves and reduces their height before they reach the shoreline (Wadhams et al., 1988¹⁴⁸; Squire, 2007¹³⁶). A greater amount of open water leads to larger waves, even if the winds are unchanged (e.g., Lintern et al., 2011⁹⁰). Increased winds over open water and higher waves, arising from reduced sea ice that would otherwise diminish storm surges, lead to higher extreme water levels. Thus, in areas where sea ice is projected to continue to diminish, such as Atlantic Canada in winter and spring (Han et al., 2015a⁵⁸) and the Arctic in summer and fall, there is the potential for increased extreme high water levels due to stronger storm surges and wave run-up.

In summary, global mean sea level has risen globally and is projected to continue to rise by many tens of centimetres, possibly exceeding a metre, by 2100. This is primarily attributable to ocean thermal expansion and water delivered to the oceans from diminishing glaciers and ice sheets. Across Canada, however, relative sea level is projected to rise or fall, depending on the amount of global sea-level rise and local vertical land motion. Due to post-glacial land subsidence, parts of Atlantic Canada are projected to experience relative sea-level change higher than the global average during the upcoming century (high confidence). This statement of confidence is based on a strong mechanistic understanding of processes controlling global and relative sea levels. Uncertainty remains about the magnitude of some sources of global sea-level, especially the projected amount of water delivered by the Antarctic Ice Sheet. All emission scenarios are projected to result in global mean sea-level rise, with the magnitude of change diverging among scenarios in the latter half of the 21st century. Vertical land motion measurements are consistent over broad spatial scales, which contribute to the confidence in the relative sea-level projections. Adaptation actions need to be tailored to local projections of relative sea-level change.

Where relative sea level is projected to rise (most of the Atlantic and Pacific coasts and the Beaufort coast in the Arctic), the frequency and magnitude of extreme high water-level events will increase (high confidence). This will result in increased flooding, which is expected to lead to infrastructure and ecosystem damage as well as coastline erosion, putting communities at risk. The statement of confidence is based on long-term measurements of coastal sea level and mechanistic understanding of the processes controlling extreme water-level events. Where relative sea level is projected to rise, extreme high water levels (combined tide and surge) will be even higher and more frequent in the future. As yet, projections of regional storm intensity and frequency are not robust, so their potential contribution to changes in extreme water-level events is uncertain.

Extreme high water-level events are expected to become larger and occur more often in areas where, and in seasons when, there is increased open water along Canada’s Arctic and Atlantic coasts, as a result of declining sea ice cover, leading to increased wave action and larger storm surges (high confidence). The statement of confidence is based on a mechanistic understanding of the processes controlling extreme water-level events and expert judgment. This finding is supported by Chapter 5, which demonstrates significant declines in summer sea ice area observed across the Canadian Arctic, while winter sea ice area is decreasing in eastern Canada (e.g., Gulf of St. Lawrence). Sea ice is projected to continue to decline in the Canadian Arctic, and further reductions of seasonal sea ice are projected for eastern Canada (Chapter 5, Section 5.3.2).

An increasing concentration of CO₂ in the atmosphere is not only contributing to greenhouse warming of the global climate system, it is also affecting the ocean carbon cycle (see Box 7.6) and changing the fundamental chemistry of the ocean. The ocean has taken up more than a quarter of the CO₂ produced by human activities, mainly from fossil fuel burning, since the start of the Industrial Era (Sabine et al., 2004¹²⁸; Rhein et al., 2013¹²⁴; Jewett and Romanou, 2017⁸⁰). While this uptake has helped to slow the rate of anthropogenic climate change, it has also resulted in increased acidity of the ocean (referred to as ocean acidification).

Acidification takes place after CO₂ gas from the atmosphere is transferred into the surface of the ocean, where it dissolves and forms carbonic acid. This process causes a decrease in pH and in the concentration of carbonate ions (CO₃^2-), a building block of organisms with calcium carbonate (CaCO₃) shells and skeletons. This process also results in a decrease in the ocean’s saturation state (a measure of the thermodynamic potential for a particular mineral to form in a solid state or to be dissolved) with respect to CaCO₃. These changes can result in seawater having a corrosive effect on shells and skeletons, dissolving them, inhibiting their growth, or causing them to require more energy to grow. Ocean acidification may have many other harmful effects for marine organisms, including increased mortality of young, changes in behaviour, food web changes, reduction in suitable habitat for some species, and increases in harmful algal blooms (Haigh et al., 2015⁵⁵).

Observations confirm that pH, a measure of acidity, has a present-day range of 7.95–8.35 (mean 8.11) in the surface waters of the open ocean (Feely et al., 2009³⁸). Globally, the pH⁵of ocean surface waters has decreased by 0.1 since the beginning of the Industrial Era (Rhein et al., 2013¹²⁴). The largest reduction has occurred in the northern North Atlantic, and the smallest reduction, in the subtropical South Pacific (Sabine et al., 2004¹²⁸). Oceans have not experienced pH changes this rapid for at least the last 66 million years and possibly the last 300 million years (Hönisch et al., 2012⁶⁹). Some acidification events in Earth’s history have led to some species becoming extinct and others recovering slowly (Hönisch et al., 2012⁶⁹). This raises serious concerns about the resilience of marine ecosystems to increasing atmospheric CO₂.

Nearshore and coastal waters are affected by the same processes as open-ocean waters and are additionally affected by freshwater inputs from rivers, glacial meltwater, and sea ice melt that decrease the capacity of coastal waters to buffer CO₂, making them more vulnerable to acidification (Ianson et al., 2016⁷⁴; Moore-Maley et al., 2016¹⁰⁵; Azetsu-Scott et al., 2014⁴). Another factor in some coastal areas is nutrient inputs from human and industrial activities via rivers, and other runoff, which increase primary production in coastal waters. In turn, various forms of planktonic organisms and their decay products are consumed by bacteria that contribute to local ocean acidification and reduced oxygen concentrations (see Section 7.6.2) through bacterial respiration, which produces CO₂.

Each of Canada’s marine regions (Pacific, Arctic, and Atlantic) has distinct factors that affect the degree of ocean acidification, and these regions are interconnected through ocean circulation patterns (see Section 7.1). Levels of dissolved carbon in the Northeast Pacific are naturally high due to the global ocean’s meridional overturning circulation patterns (Feely et al. 2008³⁹). In this region, water below the winter mixed layer has been travelling through the ocean interior for years to decades (out of contact with the atmosphere), accumulating additional organic matter from sinking biological production that decomposes to nutrients and CO₂ (Feely et al., 2004⁴⁰). Summer upwelling there brings this nutrient- and CO₂-rich water to the surface and causes intermittent periods of exceptionally low pH (7.6) at ocean depths above 100 m (Ianson et al., 2003⁷³, 2009⁷⁵; Haigh et al., 2015⁵⁵). These processes make for a system with considerable variability over time and space. The key issue for ocean acidification in this region is that the upwelled waters will have increasingly more CO₂ and lower pH in the future (Feely et al., 2008³⁹).

The Canadian Arctic Archipelago (Chierici and Fransson, 2009¹⁵) and Canada Basin of the Arctic Ocean (Yamamoto-Kawai et al., 2009¹⁶⁰) are the first ocean regions off Canada to show low saturation state; that is, corrosive surface waters. The observed increase in acidity resulting from global CO₂ emissions has been augmented in the Arctic Ocean by rapid increases in freshwater input from accelerated ice melt and increased river input, which has reduced the CaCO₃ saturation state. In addition, in cold Arctic waters, CaCO₃ shells are even more soluble, making shelled organisms especially vulnerable to the effects of acidification. Rapid changes are projected to continue for the Arctic Ocean surrounding Canada, and this region is expected to be the first to experience undersaturation of surface waters (Feely et al., 2009³⁸).

In the central Labrador Sea, deep convection during wintertime transports anthropogenic CO₂ as deep as 2300 m (Azetsu-Scott et al., 2010³). Annual sampling of the Labrador Sea since 1996 has shown a steady decline of pH (of 0.029 per decade) in a layer 150–500 m below the ocean surface, which is ventilated every year by vertical mixing during winter (see Figure 7.21). Further south, the pH of the bottom waters in the Lower St. Lawrence Estuary (in the Gulf of St. Lawrence; see Figure 7.4), have decreased by 0.2 to 0.3 since the 1930s (rate of 0.021 per decade; see Figure 7.21), which is greater than can be attributed solely to the uptake of anthropogenic CO₂ (Mucci et al., 2011¹⁰⁷). The pH decrease has been accompanied by a decline in the CaCO₃ saturation state.

Waters of the Scotian Shelf have the lowest saturation states of the entire New England/Nova Scotia region (except for episodic nearshore events), due to cold water temperatures in the winter (Gledhill et al., 2015⁵⁰). As in other regions, the saturation state is modified by seasonal biological processes (Shadwick et al., 2011¹³²). A summary of the long-term trend from upper-ocean samples collected over the Scotian Shelf and Slope indicates that pH is declining at a rate of 0.026 per decade; however, there is a high level of uncertainty in data collected before the 1990s (before the establishment of international protocols and standards) (Dickson et al., 2007³⁰). For the 1995–2014 period, the trend on the Scotian Shelf is pH declining at a rate of 0.044 per decade (see Figure 7.21).

The circulation patterns connecting Canada’s ocean regions are also important for understanding differences in acidity (see Section 7.1). The naturally high levels of dissolved carbon in waters of the Northeast Pacific result in water entering the western Arctic Ocean (through Bering Strait) with low saturation states. The saturation state of the Pacific Ocean water is further decreased through the addition of sea ice meltwater and river input, as well as respiration of organic matter, in the Arctic Ocean. Outflow through the Canadian Arctic Archipelago in the eastern Arctic can be traced along western Baffin Bay and Davis Strait to the northwestern Atlantic Ocean. While local mixing in the northern Labrador Sea modifies this Arctic outflow, low saturation states can still be identified over the Labrador/Newfoundland Shelf (Azestu-Scott et al., 2010³; Yamamoto-Kawai, 2009¹⁶⁰). Within the Atlantic region, seasonally varying outflows from the Gulf of St. Lawrence and the Labrador/Newfoundland Shelf regions bring fresher and cooler water onto the Scotian Shelf and the Gulf of Maine, which seasonally increases ocean acidification.

Under all future emission scenarios for the 21st century, global ocean acidification is expected to continue to increase in the upper ocean, with pH expected to stabilize and remain above saturation under the low emission scenario (RCP2.6) (Bopp et al., 2013⁶). The high emission scenario (RCP8.5) would result in the Arctic surface waters becoming undersaturated by mid-century.

The oxygen content of the ocean is important because it constrains biological productivity, biodiversity, and biogeochemical cycles (Breitburg et al., 2018⁷). Waters with low levels of oxygen are described as “hypoxic” (dissolved oxygen concentration of less than 61 μmol/kg) while those that are devoid of oxygen are described as “anoxic” (zero dissolved oxygen concentration). As the global ocean warms under anthropogenic climate change, a loss of dissolved oxygen is expected (Gruber, 2011⁵²). The reason for this in the open ocean is twofold. First, as ocean temperature increases, the solubility of oxygen decreases, and therefore the ocean’s capacity to hold oxygen decreases. Second, increased upper-ocean stratification caused by warming and freshening of surface water (see Box 7.4) tends to reduce vertical mixing and ventilation of the main thermocline (an ocean layer where water temperature changes rapidly with depth), resulting in a decreased supply of oxygen from its surface to subsurface waters. The global ocean has lost about 2% of its oxygen since 1960, with large variations among ocean basins and at various depths (Schmidtko et al., 2017¹³⁰). For the upper ocean, over the 1958–2015 period, trends in oxygen concentration and ocean heat content are highly correlated (Ito et al., 2017⁷⁷).

There is only qualitative agreement between computer models and observations with regard to the amount of oxygen loss in the surface of the ocean. CMIP5 models consistently simulate a decline in the global dissolved oxygen inventory equal to only about half of the observation-based estimates and also predict different spatial patterns of oxygen change (Schmidtko et al., 2017¹³⁰; Bopp et al., 2013⁶; Oschlies et al., 2008¹¹⁰). This suggests that mechanisms of oxygen decline are not well represented in current ocean models.

Human activities can play a major role in changes in dissolved oxygen in coastal waters, which can be exacerbated by the impacts of anthropogenic climate change. Coastal and inland waters are particularly vulnerable to decreasing oxygen trends (Gilbert et al., 2010⁴⁸), because eutrophication (an increase in the rate of supply of organic matter to an ecosystem) is generally higher in these areas and because physical flushing may not be adequate to disperse oxygen-depleted waters. It can be difficult to separate the effects of nutrient enrichment and climate change in assessing changes in oxygen concentration in these waters. A general overview of oxygen status and trends in Canadian marine waters is provided in Figure 7.22.

Observations off both Pacific and Atlantic coasts of Canada indicate a general decline in the concentration of dissolved oxygen in subsurface (150–400 m) waters below the more continually ventilated surface layer (see Figure 7.23). In the Lower St. Lawrence Estuary, the oxygen decrease has been attributed mainly to an increased inflow of oxygen-poor waters from the North Atlantic’s subtropical gyre (see Box 7.2) entering the mouth of the Laurentian Channel at depth (Gilbert et al., 2005⁴⁹). However, excess nutrient loading from human activity likely plays a role as well (Hudon et al., 2017⁷²). The time series from the Labrador Sea indicates a rate of oxygen decline similar to that of the St. Lawrence Estuary, but the record extends back only to 1990 (Yashayaev et al., 2014¹⁶¹). While some estuaries in Prince Edward Island, New Brunswick, and Nova Scotia occasionally become hypoxic (Price et al. 2017¹²¹; Burt et al. 2013⁹), the role of ocean-climate change remains unclear.

Hypoxia and anoxia have occurred naturally for thousands of years in some inland fjords on the British Columbia coast (Zaikova et al. 2010¹⁶⁸). Measurements at Station P dating back to 1956 indicate that ocean oxygen concentrations in the offshore Northeast Pacific have been declining for several decades (Whitney et al. 2007¹⁵⁷; Crawford and Peña, 2016²²; see Figure 7.23). A combination of physical and biological drivers are likely responsible for the observed changes in oxygen concentration at Station P; however, the oxygen variability in the lower ventilated thermocline is a useful tracer of physical climate change (Deutsch et al., 2006²⁹). In contrast to the long-term dissolved oxygen decline at Station P, the subsurface waters adjacent to the continental slope of British Columbia demonstrate no clear trend from the 1950s to present (Crawford and Peña, 2016²²). This highlights that natural decadal variability in the Northeast Pacific Ocean needs to be considered in assessing long-term changes in dissolved oxygen in this region.

Long-term observations in the Arctic are limited, and dissolved oxygen trends are thus uncertain. The Arctic Ocean has shown little evidence of hypoxia and, in fact, primary production within the so-called subsurface temperature maximum (the layer where temperature is highest) raises oxygen levels in the underlying pycnocline (an ocean layer where water density increases rapidly with depth) to levels of supersaturation (Carmack et al., 2010¹¹).

Global models project that the total amount of dissolved oxygen loss (averaged over 200-600 m) will be a few percent by the end of the 21st century (Bopp et al., 2013⁶). However, differences in the spatial patterns of dissolved oxygen among models limit confidence in regional projections. Subsurface oxygen concentrations off Canada’s Atlantic and Pacific coasts are expected to continue to decline with increasing CO₂ and heat in the atmosphere and with increasing upper-ocean stratification in most areas (Collins et al., 2013²⁰).

Nutrients, the fundamental building blocks of life, are necessary to fuel the algal biomass (e.g., phytoplankton) that sustains marine food webs and the ocean’s production of harvestable resources. Algae growth relies on inputs of inorganic nitrogen, phosphorus, silicon, and other nutrients into the sunlit layer near the surface where photosynthesis takes place. These nutrient inputs reach this layer through vertical mixing and transport such as upwelling. Nitrogen is the primary growth-limiting nutrient in the oceans surrounding Canada and is affected by microbial processes that can result in a gain (nitrogen fixation) or loss (e.g., denitrification, nitrous oxide emission). These microbial processes are sensitive to the availability of dissolved oxygen (Gruber, 2011⁵²; see Section 7.6.2) and the level of ocean acidification (Das and Mangwani, 2015²⁶; see Section 7.6.1).

Climate changes, such as surface warming and decreasing surface salinity, affect nutrients by increasing vertical stratification in the oceans surrounding Canada (see Box 7.4). This increased stratification reduces the nutrient supply transported from deep waters to the surface layer. Such a reduction is important because it can result in chronically low nutrient concentrations in the sunlit layer during the biologically productive spring-to-fall stratification season. While long-term changes in nutrient concentrations can be an indicator of climate change and variability, there are additional factors in the coastal ocean resulting from human activities (e.g., agricultural runoff) that affect local trends.

Nutrient variations in the North Pacific Ocean reflect influences of the Pacific Decadal Oscillation and North Pacific Gyre Oscillation (Di Lorenzo et al., 2009³¹; see Chapter 2, Box 2.5). When the transient effects of these modes of climate variability are removed from the time series of upper-ocean nutrients available in the North Pacific (at less than 20 m depth), decreasing trends over the period 1961–2012 are evident for phosphate and silicate, while concentrations of nitrate remained stable (Yasunaka et al., 2016¹⁶⁴). This pattern is consistent with reduced vertical mixing as a result of increasing stratification and, for nitrate, a compensating nitrogen input via atmospheric deposition to the ocean (Duce et al., 2008³³; Kim et al., 2014⁸⁴). It is important to note that the linear trends in nutrient concentrations are only robust when averaged over the entire North Pacific Ocean, and regional trends are not statistically significant.

In the Northwest Atlantic Ocean adjacent to Canada, no consistent pattern of long-term trends in nutrient concentrations has been observed that could be attributed to climate change (Pepin et al., 2013¹¹⁴). While the eastern Labrador Sea and central Scotian Shelf show significant long-term declines in nitrate, silicate, and phosphate since the 1960s (Yeats et al., 2010¹⁶⁵; Pepin et al., 2013¹¹⁴; Hátún et al., 2017⁶³), trends in the western Labrador Sea have shown increased silicate concentrations coupled with notable declines in nitrate and, to a lesser extent, phosphate. The opposite pattern has been observed in most parts of the Gulf of Maine and Bay of Fundy (Pepin et al., 2013¹¹⁴). Most parts of the Gulf of St. Lawrence have had notable increases in nutrient concentrations since the early 1970s, but trends in this area are affected by inputs from human activities (see Section 7.6.2). Other areas of the ocean around Atlantic Canada generally have had weak trends that were variable among nutrients.

In the Arctic, there are no long-term records of nutrient concentrations. However, the surface ocean circulation patterns around Canada (see Section 7.1) result in a nutrient connectivity of the Pacific, Arctic, and Atlantic Oceans (Woodgate et al., 2012¹⁵⁸; Tremblay et al., 2015¹⁴³, 2018¹⁴⁵), and this may help future research to understand changes in nutrient inventory in the Arctic. There is some evidence that declining sea ice on the Canadian Beaufort Shelf (see Chapter 5, Section 5.3) has led to episodic increases in upward nutrient supply and biological production (Tremblay et al., 2011¹⁴⁴). A decline in the nutrient concentration in the central Beaufort Sea has been observed (Li et al., 2009⁸⁹) and modelled (Vancoppenolle et al., 2013¹⁴⁷), but evidence of long-term freshening or increased stratification is limited (Peralta-Ferriz and Woodgate, 2015¹¹³).

In summary, the available long-term time series of key chemical properties in the oceans surrounding Canada indicate trends that are consistent with global analyses. The increases observed in ocean acidification have been directly linked to human emissions of CO₂ and their subsequent transfer from the atmosphere to the upper ocean. Under all future emission scenarios, global ocean acidity is expected to continue to increase in the upper ocean, with pH stabilizing by 2100 only under a low emission scenario (RCP2.6). A high emission scenario (RCP8.5) would result in the Arctic surface waters becoming undersaturated by mid-century. Overall, high confidence has been assigned to the key message concerning ocean acidification because of the strong mechanistic understanding of the physical and chemical processes controlling these changes.

Deoxygenation of the subsurface oceans surrounding Canada is evident from high-quality time series over the last five decades in the Northeast Pacific (Station P) and Gulf of St. Lawrence. These trends are consistent with the expectations that surface warming, and in some cases freshening, will increase ocean stratification and thereby reduce vertical mixing and ventilation of the ocean interior. This conclusion has high confidence because of the consistency and quality of the oxygen time series in Canadian waters. In some high-population coastal areas, oxygen depletion is also affected by nutrients from runoff (e.g., agricultural activities). Deoxygenation of the global ocean is expected to continue; however, regional differences in model projections limit us to medium confidence in the expectation that these trends will continue in the subsurface oceans surrounding Canada.

The supply of nutrients to the upper ocean, where photosynthesis takes place, may also be affected by increased stratification as a result of climate change. Nutrient supply to the ocean-surface layer has generally decreased in the North Pacific Ocean, consistent with increasing upper-ocean stratification (medium confidence). No consistent pattern of nutrient change has been observed for the Northwest Atlantic Ocean off Canada. There are no long-term nutrient data available for the Canadian Arctic. The confidence statement in this finding reflects limited data availability, lack of statistical significance of regional trends, and, in some cases, differing trends in a region.