Seismicity and seismic monitoring of Canada’s volcanic zones

The west coast of North America is a region of active plate tectonics, where the Pacific plate meets the North American plate off the coast of British Columbia, and where the much smaller Juan de Fuca plate system, which lies between these two large plates, is actively subducting underneath Vancouver Island, Washington, and Oregon. Volcanoes are a result of active subduction, tectonic plate spreading, and mantle hot spots. British Columbia and the Yukon encompass these active plate margins and consequently have substantial Quaternary and recent volcanoes and volcanic deposits.

All of Canada’s most recently active volcanoes are located in the western part of the country, extending from southwestern British Columbia to the Yukon (Fig. 1). Of the ∼348 Pleistocene or younger volcanic vents, ∼54 have been active in the Holocene (Hickson 1994; Hickson and Edwards 2001; Stasiuk et al. 2003; Wilson and Kelman 2021; Kelman and Wilson 2023). Although there has not been a volcanic eruption in Canada in more than 150 years, studies of past eruptions indicate that Canada has an annual probability of eruption greater than 1/200 (0.5%) and a recent risk evaluation indicates that both Mount Meager and Mount Garibaldi have a “Very High” threat level (Wilson and Kelman 2021; Kelman and Wilson 2023), comparable to Mount St. Helens in Washington State.

Canada’s volcanoes and volcanic deposits have been divided into five key volcanic belts (e.g., Russell et al. 2023 and Kelman and Wilson 2023), briefly described below and shown in Fig. 1. These five zones represent all Pleistocene to recent volcanism in Canada. Detailed information is provided in companion articles contained within this special issue, and in Wilson and Kelman (2021) and references therein. There are varying definitions of what constitutes an active volcano within the volcanic community. For this paper, active volcanoes are defined as last erupting within the Holocene. Not surprisingly, western Canada is also the most seismically active region (Fig. 2) of the country, with the largest and most frequent earthquakes occurring on or near the active plate boundaries. This article focuses on the seismicity and seismic monitoring of Canada’s active volcanic zones in western Canada.

In this section, we provide a very brief overview of the five volcanic zones in Canada. For details, readers are referred to other articles in this volume (e.g., Russell et al. 2023), and Wilson and Kelman (2021) and references therein.

The Garibaldi Volcanic Belt (GVB, Fig. 1), the youngest of the five volcanic regions described in this paper, is the northern extension of the Cascade Volcanic Arc in the United States and extends from the Canada/US border for nearly 500 km to the northwest through the Coast Mountains. This combined volcanic belt, which is a chain of stratovolcanoes running inland and parallel to the coast from northern California to the British Columbia mainland east of northern Vancouver Island (see fig. 1 of Hyndman and Wang 1995), is result of active subduction of the Juan de Fuca plate beneath the edge of the North America plate; subduction commenced at approximately 3 Ma (Barr and Chase 1974; Green et al. 1988) with the change in direction of the Juan de Fuca plate subduction (Riddihough 1977, 1984). Volcanism is dominantly characterized by explosive-style eruptions (e.g., Mount St. Helens in 1980, and Mount Mazama (now Crater Lake) and Mount Baker). A few examples describing the volcanic hazard and risk associated with Cascade volcanoes are provided in Hildreth (2007) and Ewert et al. (2018). In Canada, these Cascade Arc volcanoes include Mount Meager, Mount Garibaldi, and Mount Cayley; they are located ∼60–150 km to the north of Vancouver (Fig. 1). Mount Baker, another one of the Cascade Arc volcanoes, is just across the border in the United States (∼110 km from downtown Vancouver and 45 km from Abbotsford). Of these stratovolcanoes, Mount Baker is believed to be the most recently active (a minor eruption in 1843), followed by Mount Meager, with an explosive eruption ∼2400 years ago (Hickson et al. 1999). The stratovolcanoes in this volcanic belt are rated as the highest volcanic risk in Canada due to their proximity to major population centres and infrastructure, as well as the lahars, landslides, and other potential impacts (Wilson and Kelman 2021; Kelman and Wilson 2023). It is believed that, like Mount St. Helens, an eruption of any of these stratovolcanoes would be preceded by weeks, or likely months of microseismicity as magma moves towards the surface (e.g., Main 1987; Malone 1990).

The Anahim Volcanic Belt (AVB) extends across central British Columbia for ∼330 km, from the coast near Bella Bella and Bella Coola eastward to the Quesnel region. It consists of more than 40 volcanoes and hundreds of smaller cinder cones that show Alkaline basaltic (Hawaiian-style) volcanism (Bevier et al. 1979; Russell et al. 2023). This is interpreted to result from the North American plate sliding westward over a mantle hotspot (Bevier et al. 1979; Bevier 1989; Kuehn 2014; Kuehn et al. 2015). It has also been interpreted as an edge effect of the subducted Juan de Fuca plate in the mantle (Stacey 1974; Thorkelson et al. 2011). Nazko Cone (erupted 7300 BP) is the most easterly and youngest volcano in this belt (Souther et al. 1987). Recent seismic activity in the vicinity of the Nazko Cone and a 1940s sequence of felt earthquakes at the western end of this zone are documented later in this paper.

The Wells Gray–Clearwater Volcanic Field (WGCVF) is a small (∼200 km long) zone in eastern British Columbia. We note that this is labelled as the “Clearwater Quesnel Volcanic Province” (CQVP) in Kelman and Wilson (2023) but represents the same region. It is a collection of monogenetic alkaline basaltic volcanoes that erupted from ∼3.5 Ma to 400 BP (Hickson 1986; Metcalfe 1987; Hickson et al. 1995; Hickson and Vigouroux 2014; Russell et al. 2023). The cause of WGCVF volcanism is a topic of ongoing debate; one theory being that it may be associated with asthenospheric upwelling due to extensional crustal displacement along the Nootka Fault (Madsen et al. 2006).

The Chilcotin Group Basalts (CB) represents a large, approximately N–S-trending igneous province in south-central British Columbia, including a volcanic plateau that is >50 000 km² (Bevier 1983; Mathews 1989; Souther 1991). These basalts began erupting at ca. 31 Ma (Mathews 1989) and continued into the Pleistocene (2.58 Ma to 11 700 years ago) (Lambert 1963; Bevier 1983; Mathews 1989; Sun et al. 1991; Sluggett 2008; Dohaney 2009). This is consistent with the subducted Juan de Fuca plate slab window driving volcanism during this time.

The Northern Cordilleran Volcanic Province (NCVP) covers a large area of northwestern British Columbia and southwestern Yukon. It is the most volcanically active region of Canada and includes a broad group of alkaline basaltic to felsic, Miocene to Holocene eruptive centers (Edwards and Russell 1999). This volcanism is caused by trans-tensional extension of the North American plate in response to the northward movement of the Pacific plate along the Queen Charlotte Fault (Rogers 1983a). This zone includes many recent eruptive centers, including Tseax Cone (last eruption in mid-1700s; Brown 1969; Williams-Jones et al. 2020), Lava Fork (1800s; Hauksdóttir 1992), and Mt. Edziza (at least 29 Holocene eruptions; Souther 1992).

Each of these volcanic zones described above is different in terms of history, frequency and type of eruptions, exposure and potential impacts, and location relative to tectonic earthquakes. These factors have played a key role in seismic monitoring of these volcanic zones. Those volcanoes which are closest to major population centres and tectonic earthquake activity and that pose the greatest risk (e.g., the GVB) have had better seismic coverage than that of the other volcanic regions in Canada, particularly over the past few decades. The details of seismic monitoring are provided below.

In Canada, the earthquake history goes back to 1661 with a magnitude 5.7 earthquake near Montreal, QC. However, instrumental monitoring of earthquakes within Canada did not commence until 1897 in Toronto (Smith 1967). On 3 September 1898, the first seismic station in Western Canada (and one of the first in the world) was installed in Victoria, BC (Victoria Seismograph Notebook 1898). The Milne seismograph used a candle to reflect light via a mirror onto a chart strip recorder, which was driven by a clock to provide a constant speed for the chart strip recorder and accurate timing. The Victoria station underwent instrument additions and changes over the years; from a variety of lower gain, long period instruments (Milne, Wiechert, and Milne–Shaw) to higher gain, short-period instruments (Benioff) commencing in 1948. While the Victoria area seismograph was operated almost continuously since 1898, it moved from its original location in Victoria Harbour (VBH, 1989) to the Gonzales Observatory (VGZ, 1914) to the Dominion Astrophysical Observatory (VIC, 1939), and finally to the Pacific Geoscience Centre (PGC, 1978). The Victoria area station remained the sole site in western Canada until the 1950s, which saw seven more short-period (Willmore–Sharpe, Willmore) instruments deployed across southern BC to Banff, AB in the Rocky Mountains. In these early years, the seismic stations at Seattle, WA, Hungry Horse Dam, MT, Butte, MT, and Grand Coulee Dam, WA contributed greatly to the location of Canadian earthquakes (Milne 1955). Figure 3 shows an overview of the seismic monitoring network from 1955 to 2023.

In 1958, the Government of Canada undertook a project to expand the seismic network so that no point in the country would be more than about 480 km from a first-class seismograph station (Smith 1967). This was to primarily aid in the detection of distant earthquakes and nuclear explosions. The 1960s saw eight more analogue stations deployed in British Columbia, Alberta, and the Northwest Territories. Commencing in 1972, Canadian seismograph stations were classified as either “standard” or “regional” stations. A standard station consists, at minimum, of three orthogonal short-period and three orthogonal long-period seismographs, each producing a photographic record. Regional stations consist of a single-component, short-period seismograph only, using a vertical instrument and usually with visual recording and electronic amplification. The paper records produced from these stations were mailed back to the office, either in Victoria, BC, or Ottawa, ON.

Commencing on 1 September 1975, the first four digital stations were installed in BC (Victoria, Port Alberni, Haney, and Pender Island) as the Western Canadian Telemetered Network (WCTN) and were occupied by short-period digital seismometers. Digital data were telemetered into each of the two Geological Survey of Canada offices in Victoria and Ottawa. The onset of near real-time digital data transmission allowed data to be stored as segmented event data (a collection of all the acquired network data around the time period of a seismic event), which later, when disk space became less expensive, morphed into storage of continuous data streams. The network expansion continued steadily on the order of 10 stations per decade until the 1980s, which saw a boom of more than 37 new stations added to the western seismic network. The early 1990s saw the advent of the broadband seismometer that expanded the frequency range of the signals recorded by a single instrument. The advent of digital data allowed the routine sharing of data across the borders with Alaska, Washington State, and Montana networks. In addition, there have been a significant number of temporary station deployments (both analogue and digital), which have augmented the Canadian seismic network and provided windows of opportunity to adopt stations to expand the network. There were a series of Portable Observatories for Lithospheric Analysis and Research Investigating Seismicity (POLARIS)—a consortium of government, academia, and industry (see Eaton et al. 2005) deployments from 2001 to 2014 to study crustal structure. From 2004 to 2021 the US Transportable Array deployments rolled across the US from west to east and then into Alaska, with extra stations being deployed in the Yukon and northern BC. Transportable Array sites were occupied for approximately 2-year intervals although this time frame was extended up to 6 years at the end of deployment period, which allowed some of the northern stations in Alaska, BC, and the Yukon to be in place for this extended time frame.

There are difficulties with seismic volcanic monitoring. Earthquake detection is dependent on instrumentation, distribution of seismic stations, and proximity to source. The early instrumentation commonly had free periods of ∼15–20 s, which is good for detecting large distant earthquakes but is very poor at picking up the high frequency signals of small, local (nearby) earthquakes.

The seismic network in Canada was primarily designed to, first, detect distant earthquakes and nuclear explosions and, second, in later years (after the 1946 earthquake in Courtenay, BC, M7.3), to provide a backbone network so that earthquakes in Canada greater than magnitude three could be located throughout the country. Station density was increased to lower the detection threshold in areas of greater risk (more earthquakes and more population). This provides minimal monitoring capability of volcanic activity in any of the volcanic regions; however, it does not support detailed study of volcanic activity at any of the recently active volcanoes (those less than 10 000 years old) or provide reliable depth estimates of seismicity at these volcanoes.

There are also many practical challenges to monitoring seismicity in volcanic zones. In Canada, these areas are remote with steep terrain and complicated areas to work in. They are high-noise environments, often with thick layers of tephra and other loose material that contribute to high levels of seismic “noise”. In addition, volcanic belts in Canada commonly have both hot springs and glaciers that often have microseismicity associated with the circulation of fluids (and not related to magma), as documented by Rogers (1976).

A volcano typically has one or more magma conduits from the mantle to a magma storage chamber at shallower depth (possibly topping out around 10 km depth) and another one or more feeding conduits to the summit crater or a flank cone (Zobin 2003). There are several different types of seismicity associated with volcanoes (e.g., Matoza and Roman 2022). Generally volcanic earthquakes can be classified into four types (Minakami 1960, 1974). There are volcano-tectonic (VT) earthquakes, which are those generated by stresses that arise during magma migration, caused by shear or tensile failure through conduits, dikes, and sills, producing rupture at the tip of the magma body and within adjacent fault systems. These are high-frequency “tectonic-looking” events with sharp P- and S-wave arrivals. Distal VT events are often observed; these are events that are distant (e.g., 10–20 km) from a volcanic vent. VT earthquakes are further classified into A and B types. A-type volcanic earthquakes generally occur at depth of 1–20 km and take place before and during eruptive activity. They cannot be distinguished from those of general (tectonic) earthquakes. B-type volcanic earthquakes are generally limited to within a kilometer radius of an active crater, are very small magnitude, and occur in swarms. This type of earthquake is very important for the forecasting of volcanic eruptions. There are explosion earthquakes that accompany individual explosive eruptions. Another common seismic signal associated with volcanic activity is volcanic tremors. These are lower frequency signals that lack clear P and S waves, have an emergent onset, and long duration (e.g., minutes to hours). These types of events are considered to be directly or indirectly generated by magmatic activity within the volcanic conduit, volcanic explosions, or associated with lava dome growth and the migration of volcanic fluids underground (Zobin 2003).

In Canada, at the present time, we are primarily concerned with VT earthquakes, although there are a handful of sites that have the potential to support the study of volcanic eruptions. All types of volcanic earthquakes require a tight control on earthquake location and depth within the vicinity of the volcano, as well as a history (a decade or more) of tightly controlled event locations to provide a background of activity on which to observe changes. At a minimum, it requires one seismic station within ∼10 km of the volcano to provide some level of depth control. To provide tightly controlled event locations and depth, an array of a minimum of five seismometers within a 20 km radius of the volcano would be required. At the present, neither of these situations exist at any recent volcanic edifice in Canada, so only general observations can be made in each of the defined volcanic regions.

The following describes, in more detail, the seismic monitoring of volcanic zones over time in western Canada.

The majority of seismicity in western Canada (Fig. 2) is primarily related to the plate tectonic interactions off the west coast of Canada (Milner et al. 1978; Rogers 1983a) with some microseismicity likely associated with glacially induced effects (e.g., Sauber et al. 2021). Several limited studies have been conducted in a number of the volcanic zones, but specific long-term monitoring or long-term studies of the volcanic zones have not been conducted. Nevertheless, a base level of seismicity is available for each of the volcanic zones.

The detection magnitude thresholds described here (and provided in Tables 1–6) are the level to which events can be reliably detected. This does not mean that all events of that magnitude can be located. In addition, there will occasionally be events less than the threshold that may be located during perfect conditions (such as low background noise levels, rupture mechanisms well oriented to the direction of existing stations, and appropriate instrumentation in the vicinity of the event).

Completeness thresholds are the level at which all of the events of a given magnitude (and larger) have been located. There will be earthquakes smaller than this magnitude level in the earthquake catalogue but not all of those will have been located due to varying conditions, as mentioned above.

The GVB has been monitored at about the M2.5 level from 1951 to 1960 and since 1975 by the Canadian National Seismic Network (Fig. 4). On average, about five events per year were detected but not associated with any particular volcano. In 1974–1975, 122 days of high-gain seismic monitoring in the vicinity of Mt. Meager (Rogers 1975, 1983b) detected only one small event, too small to be located (M < 1.5). This is the only volcanic belt to have real-time monitoring (transmission of real-time digital data available for rapid earthquake locations) since the mid–late 1970s (Fig. 5), when the WCTN was deployed. Improvement to seismic monitoring of volcanoes in Canada began in December 1981, when a short-period seismometer was deployed near Whistler—permitting detection of all M2+ events in the central portion of the GVB, and slightly higher detection levels near Mount Meager, Mount Cayley, and Mount Garibaldi.

Delahaye (2002) conducted a study to assess potential unlocated events in the GVB by looking for single station detections on seismic stations (Pemberton, Watts Point, and Sechelt) located within the vicinity of the GVB during 1996–1998. The detections from the stations appeared to cluster in the area near Mount Meager, with ML < 1.4. The study compared past seismicity rates from 1985 to 2001 at Mount Meager to a standard b value of 1 and found that there was no anomalous amount of lower level seismicity within the GVB that was not being detected by the seismic network during 1996–1998. During this time, Delahaye found the earthquake location completeness level around Mount Meager to be between magnitude 1.3 and 1.4.

The AVB has historically had poor seismic coverage (Fig. 6). The first seismic stations in the area were deployed in 1965 (Fort St. James, located ∼150 km to the north of the AVB), and Bella Bella in the west, deployed in 1993. This very limited monitoring was further complicated (prior to the 1980s) in that the majority of the analogue (paper records) were reviewed by the local operator and then mailed to the data analysis centre in Sidney on a weekly basis, introducing significant delays in locating regional earthquakes.

In 2007, prior to the Nazko earthquake swarm, seven broadband seismographs were deployed (for 2–3 years) in the Nechako Basin area of the AVB as a part of a POLARIS project for mapping crustal structure. This significantly improved monitoring capability in the eastern portion of the Anahim Belt (Fig. 7).

The Wells Gray-Clearwater Volcanic Field (WGCVF) has historically also had poor seismic coverage (Fig. 8). In 1972, BC Hydro funded the McNaughton Lake array (Ellis et al. 1976; Ellis and Chandra 1981), recorded on paper records, to monitor the reservoir. McNaughton lake is located to the east of the WGCVF. In 1977, BC Hydro funded four permanent seismic stations, all to the east of the WGCVF: Blue River (BRBC), Mount Dainard (MNB), Sale Mountain (SALE), and Downey Slide (DOWB). These stations operated until 2018, with the exception of DOWB that was closed in 2012. Over the timespan of their deployment, they were initially analogue paper records, and then later, an analogue signal converted to digital. Since late 1972, there has been an average of two events per year detected in the WGCVF, which has been monitored at a magnitude 2 level (Fig. 9).

The Chilcotin Basalts (CB) were recently identified as a volcanic region (Wilson and Kelman 2021; Kelman and Wilson 2023). All the identified volcanic outcrops are older than 10 000 years and have not specifically been considered as a region to monitor for volcanic hazard in the past. Situated between the GVB to the west, the AVB to the north, and the WGCVF to the east, the CB have two currently operating seismic stations in the region, Lillooet (installed in 1957) to the north and Penticton (installed in 1960) to the south (see Fig. 10). To the west, the nearest seismic stations are at Whistler and Hope, both of which are important stations for monitoring the GVB that currently has a completeness level of about M2 (Fig. 11).

The Northern Cordilleran Volcanic Province (NCVP) has traditionally had few seismic stations operating (Fig. 12). The longest-running seismic stations are in the northern portion of this belt and include Dawson City, Whitehorse, Pleasant Camp, BC, and Dease Lake. From 1968–1972, 579 days of high-gain seismic recording (Rogers 1976, 1983a) was conducted within 30 km of Mount Edziza (most prominent volcano in the belt). Very little seismicity was detected in the volcanic belt, and none on the volcano itself. The located earthquakes were in the vicinity of the Mess Creek graben and the largest earthquake detected was M1.3; the event rate was approximately 1/day. The other two volcanoes of interest in the NCVP, both sites of volcanic activity within the last 250 years, Tseax and Lava Fork, have not had any monitoring associated with them due to their remote locations. Seismicity plotted as a function of time for the NCVP is shown in Figs. 13 and 14.

Since 1898, when seismic monitoring began in western Canada, no Canadian volcanoes have erupted and no significant volcanic seismicity has been recorded. Nonetheless, volcanic seismicity has, without a doubt, occurred, including the sequence of nearly 1000 small earthquakes near Nazko Cone in 2007. As described earlier, detecting volcanic seismicity is challenging, as it is often microseismicity that is associated with volcanoes. These tiny (often M < 1) events are difficult to detect given the relative lack of station coverage in remote areas. Here, we summarize what is known of the seismicity in Canada’s volcanic zones.

This is one of the most seismically active of the volcanic zones in Canada—with most earthquakes here being associated with crustal stresses from the Cascadia subduction zone. With the exception of Mount Meager (described below), there are no obvious significant clusters of seismicity associated with volcanic peaks in the GVB. However, the seismic detection level here is such that microseismicity (M < 1) is difficult to detect.

Seismicity rates since 1985 in the Mount Meager Volcanic Complex region ranged from 1 to 10 events per year until 2013, when the rates began to climb to a maximum of more than 50 events in 2014, then dying back down to background levels in 2016. The majority of the increased activity clusters between July 2014 and October 2015, with the events occurring both during the day and night. The events were all less than or equal to magnitude 2.5 and the majority of events are shallow, although depth determination of those events is poor as the closest seismic station at that time was located in Whistler, BC, ∼100 km away. Most of the events were to the south of Meager Mountain. In September 2016, a new seismic station (MGMB) was installed across the valley to the east of Mount Meager to better monitor the seismicity near the volcano and to improve earthquake depth determinations. Unfortunately, there were challenges as the signal from the site was dominated by high-frequency noise from the nearby Boulder Creek, drowning out the high-frequency signal from small local earthquakes. However, the site did record large distant earthquakes well. There is no evidence to indicate that the small magnitude events from the earthquake clusters in 2014–2015 were due to magma movement. We note that this was a substantially lower seismicity rate compared to (as one example) the pre-eruption period of Mount St. Helens (Malone 1990), which had hundreds of earthquakes per day, and included earthquakes of much larger magnitude. The 2013–2016 Mount Meager area seismicity may be related to the three known hot springs in the area and the associated hydrothermal fluid circulation, or perhaps associated with the Upper Lillooet Hydroelectric Project in the area that involved blasting for tunnels (Mulder 2017).

The AVB extending across central British Columbia is a seismically quiescent region. Very few earthquakes have occurred in central British Columbia since 1965 when local monitoring began with a seismic station at Fort St. James (180 km to the north of Nazko).

On 9 October 2007, an unusual sequence of earthquakes began in central British Columbia about 20 km west of Nazko Cone—the most recent (circa 7200 years) volcanic centre in the AVB. Within 24 h, eight earthquakes of magnitude 2–2.9 occurred in a region where no earthquakes had previously been recorded. During the next 3 weeks, more than 800 microearthquakes were located (and many more detected), most at a depth of 25–31 km within a radius of about 5 km of the first earthquake of the sequence. Precise relocation of the seismicity using double-difference suggested a horizontal migration at the rate of about 0.5 km/day, with almost all events within the lowermost crust (Cassidy et al. 2011). Neither harmonic tremor nor long-period events were observed; however, some “spasmodic bursts” were recorded and determined to be co-located with the earthquake hypocentres. These observations are all very similar to a deep earthquake sequence recorded beneath Lake Tahoe, California in 2003–2004 (von Seggern et al. 2008). Based on these remarkable similarities, the Nazko sequence was interpreted as an indication of an injection of magma into the lower crust beneath the AVB (Cassidy et al. 2011). This magma injection fractures rock, producing high-frequency VT earthquakes and spasmodic bursts.

The western end of the AVB, in the vicinity of Bella Coola, is the most seismically active portion of this volcanic belt, with 15–20 small earthquakes per year, on average over the past 30 years. A remarkable swarm of more than 40 felt earthquakes occurred near Bella Coola between September 1940 and August 1943 (Milne 1956). Many of these were felt “fairly strongly” in Bella Coola, but none caused damage (Milne 1956). It is assumed that these were shallow; however, due to the lack of seismic stations at this time, there is no information on focal mechanisms, magnitudes, or focal depths for these earthquakes. These events were not detected at the closest seismic station (Victoria, BC), roughly 500 km away, and so they are estimated to be less than about magnitude 4.5. Given the lack of seismic data, there is no definitive evidence that these earthquakes were associated with any volcanic activity; however, it is noteworthy that they were shallow, swarm-like, and occurred at the western end of the AVB.

During July–October 2016, two of the largest earthquakes (M3.4 and 3.5) in this area since the 1940s swarm occurred and were felt in the Bella Coola region. These events, a part of a sequence of small earthquakes from 2015–2017, serve as a reminder of ongoing seismicity at the western end of the AVB. A detailed analysis of western AVB seismicity is presented in this volume by Littel and Bostock (2023). Applying accurate earthquake location techniques, they were able to identify several linear strands of seismicity subparallel to the Coast Shear Zone. They determined both extensional and strike–slip focal mechanisms in this region and hypothesized that seismicity here (including earthquake swarms) results from strain-localization and high heat flow resulting from a weakened lithosphere associated with the interaction of the AVB and the Coast Shear Zone.

The WGCVF is a seismically quiescent region of British Columbia. Only 66 earthquakes have been located in this region over the past 40 years. The largest earthquake recorded here was M3.9 in 1983.

The CB region of British Columbia also experiences relatively few earthquakes. Over the past 40 years, ∼444 small earthquakes (the largest being M4.3 in 1962) have been located through this region, with most concentrated between Cache Creek and Merritt.

The NCVP is a very active seismic zone, but almost all of the earthquakes are associated with large faults (e.g., the Denali and Duke River Fault zones) and crustal stresses through this plate boundary region (Mazzotti et al. 2008; Hyndman et al. 2005). There are no obvious clusters of seismicity associated with volcanic cones in this region. For example, near the sites of two of the most recent eruptions in Canada—Lava Fork (1800s, Russell and Hauksdottir 2001) and Tseax (1700s, Williams-Jones et al. 2020), we currently see very little seismicity. Within 50 km of Lava Fork, only 26 small (M1–4.2) earthquakes have been located during the past 40 years. At Tseax, only seven small earthquakes (M1.5–3.2) have been recorded over the past 40 years. We point out though, that monitoring limitations (as described earlier in this article) cannot rule out undetected microseismicity.

Some of the most important research into volcanic risks in Canada has taken place in recent years and (or) is currently underway. This includes, in particular, the risk threat assessment that was recently completed (Wilson and Kelman 2021; Kelman and Wilson 2023) and the ongoing research and modelling of volcanic hazards funded by the Canadian Safety and Security Program (Kelman et al. 2023). This latter project will also provide an operational InSAR monitoring system that will benefit seismic hazard studies in the coming years and will contribute to prioritization of research activities at Canada’s volcanoes.

One recent study (Klaasen et al. 2021) applied distributed acoustic sensing (DAS) technology at Mount Meager in the GVB. By measuring the distributed strain along a fiber optic sensing cable, they found intense, low-magnitude seismicity, including long-lasting, intermediate-frequency (0.01–1 Hz) tremor, and tiny high-frequency (5–45 Hz) earthquakes. Between tens and several hundred high-frequency events were detected each day and showed distinct spatial clusters. Klaasen et al. (2021) conclude that both the tremor and the high-frequency earthquakes were likely related to fluid movement within Mount Meager’s geothermal reservoir. Their work illustrates that DAS carries the potential to reveal previously undiscovered seismicity in a challenging volcanic environment, where dense arrays of conventional seismometers are difficult to install.

Another recent study of seismicity in the vicinity of Mount Meager is documented in Lu and Bostock (2022). They identified, using nearby broadband seismic stations of the Canadian National Seismograph Network, the first deep long-period earthquakes (DLPs) in the GVB. Globally, DLPs are generally associated with the movement of fluids associated with magmatic processes. Lu and Bostock (2022) found a total of 42 small (ML∼0) events (26 of which could be located) satisfying the spectral criteria of DLPs that were detected at station PMB between 1993 and 1998. An additional six events were located using data from station MGMB between 2016 and 2019. The majority of the hypocentres were clustered at ∼45 km east northeast of Mount Meager and fall within the upper-mid crust. They lie in closer proximity and may be related to Quaternary volcanism associated with the Bridge River Cones.

These two examples demonstrate how volcanic seismology is rapidly evolving in Canada, especially the applications of new technologies such as fiber optic cables and improved waveform analysis methods that yield precise locations and depths (e.g., using waveform cross-correlations and double-difference location methods) and focal mechanisms.

Canada’s volcanic fields and complexes, all in British Columbia and Yukon, have not erupted in more than 150 years and little is known about many of these volcanoes. However, studies of past eruptions indicate that Canada has an annual probability of eruption greater than 1/200 (0.5%), and a recent risk evaluation indicates that both Mount Meager and Mount Garibaldi have a “Very High” threat level (Wilson and Kelman 2021; Kelman and Wilson 2023), comparable to Mount St. Helens in Washington State.

No eruptions have occurred in Canada since the deployment of the earliest seismographs in 1898 (Victoria, BC) and 1904 (Sitka, Alaska), but seismic detection levels have decreased from M∼7 in 1900 to M∼0–1 (in many regions) today, with ∼20 seismic stations operating in the volcanic zones of BC and YT. Some of the most recent significant seismic activity attributed to volcanic zones in Canada is the 2007 Nazko Cone earthquake swarm. Nearly 1000 tiny (M < 3) earthquakes occurred here over the span of about 2 months. These were all deep earthquakes (∼30 km) near the base of the crust and showed the patterns expected from an injection of magma deep into the crust. At the western end of the AVB, more than 40 felt earthquakes occurred from 1940 to 1943.

Recent advances in technology and data processing are providing new opportunities to better study seismicity in Canada’s volcanic zones. Two examples from the GVB include

1. The use of DAS technology to identify intense, low-magnitude seismicity, including long-lasting, intermediate-frequency tremor, and hundreds of tiny high-frequency earthquakes (Klaasen et al. 2021) likely related to fluid movement within Mount Meager’s geothermal reservoir.

2. The use of advanced waveform processing techniques and event locations to identify dozens of tiny (ML∼0) DLPs in the GVB. Globally, DLPs are generally associated with the movement of fluids associated with volcanic processes.

These recent studies, combined with ongoing development of InSAR monitoring, increased seismic instrumentation, and targeted temporary deployments will allow for better seismic detection and understanding of volcanic potential in the coming years. Finally, research into volcanic seismicity in Canada would greatly benefit by having permanent small seismic arrays deployed at key volcanic centres to better understand the active processes operating in these areas.

We thank C. Brillon, P. Einarsson, an anonymous reviewer, and Guest Editor, G. Williams-Jones for their thorough reviews and thoughtful comments and suggestions that have improved this manuscript. We are grateful to Dr. Garry Rogers for the information and knowledge of seismic monitoring and volcanic seismicity that he has shared with us over many years. This is NRCan contribution number 20230087.

Data generated or analyzed during this study are available from the corresponding author upon reasonable request. Seismic waveform data, earthquake epicentre information, and station information available from Earthquakes Canada: https://earthquakescanada.nrcan.gc.ca/.

Conceptualization: JFC

Visualization: TLM

Writing – original draft: JFC, TLM

This work was entirely funded by Natural Resources Canada, Geological Survey of Canada.