Impacts of Extratropical Cyclone Fiona on a sensitive coastal lagoon ecosystem

Oceanic storms can strongly disturb the physical and biogeochemical conditions of transitional coastal waters. Impacts of extreme oceanic storms on coastal ecosystems have received limited attention worldwide, with no studies at higher latitudes (> 45°) where tropical cyclones have usually abated. This study investigates the combined impacts from marine and atmospheric forcing on a coastal lagoon in Prince Edward Island, Canada, during and after Extratropical Cyclone Fiona in September 2022. Physical (water levels and temperature) and biogeochemical (dissolved oxygen [DO], electrical conductivity, pH, nitrate–nitrogen concentrations, total suspended solids [TSS]) datasets from the lagoon and the watershed's tributaries, groundwater springs, and piezometers were used to assess ecosystem disturbance and recovery timelines following the storm. Fiona resulted in a 1.6 m storm surge into the lagoon that elevated water temperatures by up to 6°C, disturbed the density‐dependent stratification of salinity and temperature, and reduced the diel amplitude of DO, indicating a reduction in plant respiration due to ecosystem disturbance. The freshwater tributaries revealed sharp changes in flow (30‐fold increase), nitrate–nitrogen (NO3‐N) concentrations and loading (70‐fold increase), and TSS loading (40‐fold increase) to the lagoon during and immediately following the storm. The lagoon rapidly recovered (hours) from the hydraulic disturbance of the storm surge, but elevated nutrient levels persisted for months. The intensity and frequency of extratropical cyclones is projected to increase in the Northwest Atlantic, making field‐based studies of cyclone impacts on coastal waters critical for understanding future coastal ecosystem disturbance and recovery periods relative to the timing of future storms.

) that provide a range of valuable ecosystem services (Newton et al. 2018).However, these critical transitional water bodies are threatened under anthropogenic impacts of invasive species, rising water temperatures, and eutrophication (Tummon Flynn et al. 2018;Chac on Abarca et al. 2021;Gonz alez-De Zayas et al. 2021).These aquatic stressors in coastal lagoons can be exacerbated by coastal development and by increasingly common and intense coastal storms (Mulligan et al. 2015;Gonz alez-De Zayas et al. 2021).These compounding threats provide the impetus for studies investigating how tropical cyclones and other coastal storms can disturb coastal ecosystems.
The impacts of cyclones on coastal lagoons are multifaceted (Greening et al. 2006).High precipitation rates and associated elevated loading of terrestrial contaminants (Pavlovskii et al. 2023) can change the physicochemical conditions throughout a lagoon ecosystem.For example, elevated freshwater discharge to transitional coastal waters can decrease salinity, deplete dissolved oxygen (DO), dilute or enhance nutrient levels, increase suspended sediment and organic matter, and mix formerly stratified zones (Steichen et al. 2020;Gonz alez-De Zayas et al. 2021;Walker et al. 2021).On the marine side, seawater intrusion can be triggered during coastal storms due to storm surges, landward Ekman transport, and wind-sheared salt flux (Li et al. 2022).The intruded seawater increases the salinity and alters coastal water temperatures through heat advection (Kurylyk et al. 2023).Similarly, landward marine sediment transport and deposition in lagoons can arise from waves overtopping coastal barriers (Tanaka et al. 2003).Coastal ecosystem recovery periods following coastal storms can last from days to months (Gonz alez-De Zayas et al. 2021;Walker et al. 2021).
Climate models project increases in the intensity, frequency, and latitudinal extent of tropical cyclones (Collins et al. 2019), yet relatively few studies have assessed the cumulative impacts of extreme coastal storms on coastal lagoons (e.g., Gonz alez-De Zayas et al. 2021;Patrick et al. 2022).Also, to our knowledge, no studies have holistically investigated the impacts of oceanic and atmospheric forcing from a tropical or extratropical cyclone on a lagoon from both the marine and terrestrial (watershed) boundaries.Accordingly, the overall goal of this study is to assess how oceanic and atmospheric conditions during an extratropical cyclone impacted a sensitive coastal lagoon and the connected aquatic systems (tributaries, salt marsh, and bedrock aquifer).We investigate both physical and biogeochemical changes through analysis of a rich, comprehensive dataset collected in a coastal lagoon and the surrounding watershed in Atlantic Canada during and following Extratropical Cyclone Fiona (September 2022).As extreme coastal storms are less common in northern countries like Canada, their impacts on coastal ecosystems at such latitudes are relatively unstudied, and further investigations are critically required in an age of warming, rising seas and intensifying coastal storms.

Study area and event description
The Basin Head lagoon is located on the eastern shore of Prince Edward Island, Canada (Fig. 1).The lagoon was established as a federal Marine Protected Area in 2005 to protect the endemic Giant Irish Moss (Chondrus crispus), which declined in biomass by over 99% from 1999 to 2012 due to high water temperatures, invasive species, and eutrophication linked to nitrate loading (Tummon Flynn et al. 2018;DFO 2020).Recent monitoring and restoration activities have involved artificially propagating the Giant Irish Moss, replanting and restoring eelgrass (used as an ecological indicator), and removing the invasive green crabs (DFO 2020).The lagoon is monitored as part of ongoing studies on longer-scale physical and biochemical processes in the coastal ecosystem.
The main basin of the lagoon is approximately 500 m in diameter with a maximum depth of 3 m at high tide.A 3-km arm drains to the basin from the northeast (i.e., northeast arm, Fig. 1c).The historic, natural lagoon outlet to the ocean located towards the northeast tip of the arm is now closed, and the present outlet towards the south (Fig. 1) requires frequent dredging due to sediment deposition.The total lagoon watershed area is 14.6 km 2 , which includes agriculture (41%), forest (31%), and wetland (12%) (KarisAllen et al. 2022).Six groundwater-dominated tributaries discharge into the arm and basin (Fig. 1), accounting for over 60% of the total terrestrial inflow on a year-round basis (KarisAllen et al. 2022).The remaining freshwater mostly comes from over 30 springs and seeps that discharge from the fractured sandstone aquifer along the northwestern shoreline (KarisAllen et al. 2022).The lagoon experiences a mixed tidal regime with a typical tidal range of approximately 0.8 m.
Like the rest of Atlantic Canada, Basin Head experiences extratropical cyclones every few years including Arthur in 2014, Dorian in 2019, Teddy in 2020, and Fiona in 2022 (Ollerhead et al. 2022).The most recent of these, Fiona, which made landfall during the morning of 24 September 2022, was the most intense storm to hit Prince Edward Island in three decades (Ollerhead et al. 2022).Because the storm's major wind direction was from the north, the northern shore of the island was the worst hit, with maximum sustained winds of 20.3 m s À1 and significant dune erosion (Mulligan et al. 2023).The minimum barometric pressure (95.8 kPa) measured near the Charlottetown airport was the lowest on Canadian record (Ollerhead et al. 2022).Basin Head experienced among the highest simulated surge along the coast of Prince Edward Island during Fiona (Mulligan et al. 2023).The storm caused an estimated $800 million in insurance payouts in Canada, not counting uninsurable assets (Weltman 2023).To our knowledge, no studies have yet addressed the ecosystem impacts of Fiona in Canada.

Methods
Field data were collected throughout the lagoon, tributaries, springs, salt marsh, and atmosphere with sensor locations shown in Fig. 1c.These sensors are part of an ongoing study on the Basin Head lagoon (e.g., KarisAllen et al. 2022).Herein, we only present data for the period immediately before, during, and after Fiona.

Meteorological monitoring
Air temperature, humidity, and wind speed/direction were collected at a meteorological station (HOBO, Onset, Massachusetts) near the outlet of the northeast arm (Fig. 1c).A pressure transducer (HOBO Water Level Logger, Onset) was also affixed to a nearby tree to record barometric pressure fluctuations to correct water levels for atmospheric pressure variations.

Lagoon monitoring
The main basin and the northeast arm were monitored to gauge the immediate effects of the storm on the coastal ecosystem, and to quantify the magnitude of marine forcing.Vertical sensor arrays (i.e., sensors at the top and bottom of the water column) were situated near the lagoon entrance on a rebar/buoy setup and in the main lagoon and northeast arm using a large weight affixed to a rope and buoy (Fig. 1c).The top and bottom sensors for each array were installed in white, slotted PVC casing attached to the buoy and bottom weight, respectively.All six sensor locations in the lagoon (i.e., top and bottom of the three arrays) measured water temperature (HOBO Tidbits, Onset except for the entrance which used two Leveloggers 5 LTC, Solinst), and five recorded electrical conductivity (HOBO Electrical Conductivity loggers, Onset).The lagoon DO (HOBO DO logger, Onset) was only recorded in the main bed due to sensor failure in another array.Diel variation in DO (range) is an indicator of ecosystem respiration rates and primary productivity, and thus we considered these diel signals before and after the storm as a first-order assessment of ecosystem disturbance and recovery.The lagoon water level was obtained from the water pressure at the stationary, bottom sensor in the entrance array by compensating for barometric pressure fluctuations.The storm surge was estimated by deconvoluting the tidal signal from the water level signal (e.g., Parker 2007) using a fast Fourier transform analysis on a longer lagoon tidal record.

Tributary monitoring
The two tributaries with the highest flows into the lagoon (Fig. 1c) were monitored to capture elevated flows and associated nutrient and sediment concentrations and loading into the lagoon due to the heavy precipitation.Water levels were measured using pressure transducers (HOBO Water Level Loggers, Onset) installed in stilling wells and corrected for atmospheric pressure.Rating curves for the tributaries were derived from flow measurements over several years (e.g., KarisAllen et al. 2022).However, the high water levels and flows during Fiona exceeded the rating curve range.Thus, high flows with levels above the rating curve were estimated through proration to the nearby Bear River discharge station (Water Survey of Canada station 01CD005) given the geographical proximity (22 km) and similar hydrology and geology.Concurrent flow measurements in the Basin Head tributaries and Bear River collected over 4 yr were used to develop the power exponent used in the proration (Ward and Robinson 2000).
The two Basin Head tributaries had autosamplers (ISCO6712 Autosamplers, Teledyne ISCO) installed to collect daily composite samples from 21 September 2022 to 15 October 2022.These samples were later analyzed for total suspended solids (TSS) using Standard Method 2450D (APHA 2022).A water quality sonde (EXO, YSI) was installed in tributary 1 from 20 September to 07 November 2022 to measure electrical conductivity, nitrate-nitrogen concentration (NO 3 -N), and pH.A further grab sample was collected from this tributary on 22 February 2023 to assess the longterm recovery in nitrate concentrations.

Groundwater monitoring
The springs and salt marsh groundwater system discharging to the Basin Head lagoon were monitored to analyze Fiona's subsurface impacts, as the subsurface response to storms and recovery is expected to be delayed (Cantelon et al. 2022) but still exert influence on subsequent surface dynamics (Douglas et al. 2022;Pavlovskii et al. 2023).Three pressure transducers (Levelogger 5, Solinst) were installed in shallow piezometers in the salt marsh and dune system between the ocean and the lagoon's northeast arm to monitor saltmarsh groundwater level and temperature (Fig. 1c).A further three water temperature loggers (HOBO MX2203, Onset) were installed on rebar pounded into the apertures of three groundwater springs (KarisAllen and Kurylyk 2021) along the lagoon's landward margin (Fig. 1c).

Impacts on meteorological conditions
A total of 109 mm of rainfall was measured at the Basin Head weather station between 23 September 2022 and 24 September 2022, with 89 mm falling over a 24-h period (Fig. 2a).According to the intensity-duration-frequency curves for nearby East Point, Prince Edward Island (Environment and Climate Change Canada, 2022), this represents a 1-in-30yr storm for historic climate conditions, but this return period would likely decrease under projected climate change (e.g., Lin et al. 2016).The minimum atmospheric pressure at Basin Head was 95.1 kPa on 24 September (Fig. 2b), which was the minimum air pressure recorded at the weather station during our 4 yr (2019)(2020)(2021)(2022)(2023) of environmental monitoring at this site (e.g., KarisAllen et al. 2022).Wind gusts reached 24.3 m s À1 after the eye of the storm had passed through, which were comparable to maximum wind gusts (23.2 m s À1 ) measured at our Basin Head climate station during Extratropical Cyclone Dorian in September 2019.

Impacts on the lagoon
The storm produced a surge level at the lagoon entrance logger (Fig. 1) of approximately 1.6 m after removing the tides from the water level (Fig. 3a).The peak water level of 2.86 m (relative to CGVD2013) was not nearly sufficient to overtop the $ 6 m dune barrier between the lagoon and ocean (Fig. 1).Water temperature at the entrance and in the northeast arm initially spiked during the storm, indicating an influx of warm seawater, followed by a decline due to the inflow of cooler freshwater draining from the watershed (Fig. 3c).The water temperature in the bottom temperature logger for the vertical array in the northeast arm rose above the 20 C stress threshold for the Giant Irish Moss but stayed below the extreme Basin Head water temperatures attained during summer heat waves (30 C, KarisAllen et al. 2022).
To investigate disturbances to the lagoon stratification, we examined changes to the thermal offset, which was calculated as the difference between the temperatures recorded at 15-min intervals in the array loggers at the top and bottom of the water column.The lagoon exhibits a thermal offset due to salinitydriven stratification (e.g., Kurylyk and Smith 2023) and due to thermally driven stratification effects from temperature differences between the springs, tributaries, and ocean/lagoon water.The thermal offsets were dominated by tidal variability (pink series, Fig. 3c,d), but storm impacts were still evident.For example, the thermal offset was up to 0.6 C in the entrance array before Fiona but decreased to $ 0 C during the first half of Fiona, before increasing to a magnitude of almost 1 C toward the end of Fiona as discussed later (Fig. 3c).The dominant direction of the offset also changed, making the water surface colder than the bottom.The northeast arm thermal offset was mostly negative (i.e., the bottom was warmer) both before and after Fiona, but its magnitude also decreased to around 0 C during Fiona (Fig. 3d).Thermal offsets returned to the pre-Fiona ranges within about 6 d of the storm in the northeast arm, suggesting the stratification was re-established.
We also analyzed the specific conductance from the sensor arrays at the lagoon entrance and the northeast arm (Fig. 1c) to investigate saltwater intrusion and disturbances to the lagoon stratification (Fig. 3e).Similar to temperature, the specific conductance offset between the surface and bed loggers was investigated to assess stratification dynamics.The northeast arm array generally had lower specific conductance than the entrance array, which was expected as it is much further from the outlet and closer to the mouths of freshwater tributaries and springs.The specific conductance offset in the northeast arm increased during the storm passing (Fig. 3f), maintaining this pattern for the remainder of the record length, whereas the opposite occurred at the outlet (Fig. 3e) as discussed later.
The lagoon DO measurements, particularly any patterns in the diel amplitudes, can be used to assess changes to ecosystem dynamics.From 01 September 2022 to 16 September 2022, DO levels exhibited an average diel range of 7.6 mg L À1 (Fig. 3b).On 16 September, the DO logger began reading 0 mg L À1 , presumably because it was buried in sediment.This persisted until 24 September when the logger was fortuitously unburied by sediment mobilized during the storm.The average diurnal range following the storm was significantly lower (4.3 mg L À1 , Fig. 3b, black) than before the storm.This reduction in DO range is indicative of pronounced ecosystem disturbance from Fiona, which caused a cumulative reduction in water column primary productivity and respiration (Bamforth 1962), similar to observations in inland watersheds (Mulholland et al. 2005).To ensure DO changes were not primarily an artifact of temperature changes, we also investigated diel signals in DO as percent saturation, which yielded the same general patterns as DO as mg L À1 (see figure in Supplementary Information).

Impacts on the tributaries and freshwater loading to the lagoon
Peak flows in the two monitored tributaries (Fig. 1) occurred around 08:00 h local time on 24 September and were 0.43 and 0.22 m 3 s À1 for tributaries 1 (3.0 km 2 area watershed) and 2 (2.6 km 2 ), respectively, which increased the pre-storm stream flows by over one order of magnitude (Fig. 4a).The water levels and thus flows in these tributaries have reached higher levels during spring rain and snowmelt periods in our 4-yr monitoring record.Tributary 1 had a greater peak flow to area ratio (i.e., specific discharge) as the watershed has more developed, impervious ground, while tributary 2 has more forested area (Latifovic 2019).
The water temperature data from the two tributaries indicate that the temperature logger in tributary 2 was placed in a reach with much greater tidal influence than the location of the tributary 1 logger (Fig. 1b).The semidiurnal temperature signals before and after the storm are evident, along with a large spike during the event from the mass and heat inflow associated with the surge (Fig. 4b).In contrast, the tributary 1 water temperature data record only reflects freshwater temperatures and thus is not significantly influenced by tides or the storm.The temperature signal in tributary 1 may also have been slightly obscured by accreted sediment that occasionally reaches depths < 0.02 m above our logger based on our past experiences at this location.
The pH and specific conductance recorded in tributary 1 both experienced a sharp decrease during the storm, followed by a nonlinear increase back to pre-storm levels around 01 October (Fig. 4c).These water quality data suggest that the stream was transformed from its normal baseflow dominance to runoff-dominated conditions during the storm.As a lower proportion of the water entered the stream via high-pH, high-specific conductance groundwater, the mixing of the two waters produced a sharp decrease in pH and specific conductance during runoff dominance and then returned to increased pH and specific conductance as the baseflow dominance returned.
The nitrate-nitrogen concentrations (NO 3 -N), which were only measured in tributary 1, initially decreased during the  passage of Fiona as the fresh rainfall and associated runoff diluted the baseflow-derived nitrate in the stream (Fig. 4d).However, the concentrations then spiked to 6 mg L À1 , approximately doubling the pre-event concentrations of 3 mg L À1 , after a few hours given the delayed response of the aquifer and baseflow to the rainstorm.These concentrations exceed the Canadian Council of Ministers of the Environment guideline for the protection of aquatic life in freshwater (3 mg L À1 , CCME 2012).The small diel variations indicate limited temperature sensitivity.The nitrate-nitrogen levels did not return to pre-storm values and instead generally continued to increase in the weeks following the storm up to the end of the sonde monitoring record.A grab sample collected from tributary 1 on 22 February 2023, yielded a measured concentration of 1.1 mg L À1 , which is within the range of the pre-storm levels.Collectively, the nitrate-nitrogen measurements from the discrete samples and autonomous sonde indicated that it took between 2 and 5 months to return to pre-event conditions.
The discharges in tributaries 1 and 2 (Fig. 4a) were multiplied by the nitrate-nitrogen concentrations (Fig. 4d) and TSS (concentrations not shown) to calculate the TSS and nitratenitrogen loading to the lagoon (Fig. 5 for TSS).The peak nitrate-nitrogen loading in tributary 1 reached 0.0013 kg s À1 , which represented a 70-fold increase from pre-event loading.The TSS readings in both streams spiked during the storm on 24 September, reaching a maximum recorded loading of 0.011 kg s À1 for tributary 1 and 0.0061 kg s À1 for tributary 2 (Fig. 5).The respective TSS concentrations reached 65 and 120 mg L À1 , respectively.However, since the autosampler only collected composite daily samples, the peak instantaneous TSS loading in both tributaries may have been significantly higher.When integrated over the event (23 September 2022 to 25 September 2022), the total TSS loads from the watersheds for tributaries 1 and 2 due to Fiona were 7.3 Â 10 À4 and 5.3 Â 10 À4 kg m À2 , respectively.Given that the average annual soil loss in Prince Edward Island is greater than 2 Â 10 À2 kg m À2 (Coote et al. 1981), this storm did not contribute a substantial portion of the annual TSS load to the lagoon.The relatively low contribution of the annual TSS load is likely partially due to the fact that the event occurred when there was still a full canopy cover on the agricultural fields, which would have mitigated soil erosion.

Impacts on the groundwater systems
The tributary water quality data revealed how the groundwater response to an extreme precipitation event is delayed and sustained in comparison to surface water dynamics.To further elucidate the groundwater response, data from the salt marsh and dune piezometers (Fig. 1) were explored.All piezometers displayed different responses to the storm based on their location (Fig. 6a).The groundwater level response in the marsh and dune loggers to the tidal and surge forcing was as expected based on their relative cross-shore distances from the lagoon (Trglavcnik et al. 2018).Piezometer 1, closest to the lagoon system, displayed an evident pre-storm, semi-diurnal tidal amplitude and the greatest groundwater level increase ($ 1.5 m) during the surge.Further into the salt marsh, tidal signals in piezometer 2 were less evident and the groundwater level increase due to the surge was only $ 1 m.Piezometer 3, the furthest from lagoon and in the transition to the sand dunes, was not generally tidally influenced and displayed a slower water level response and recovery to the storm (Fig. 6a).
The groundwater temperatures recorded in the coastal piezometers can also provide insights on the timing of seawater infiltration (Kurylyk et al. 2023).Surprisingly, piezometer 1, closest to the lagoon, experienced no abrupt temperature change over the monitoring period (Fig. 6b), which may be due to regular pre-Fiona mass and thermal mixing with lagoon water.However, temperature fluctuations are evident in the other two piezometers, which change in opposite directions given their different pre-Fiona temperatures relative to the intruded seawater (Fig. 6b).The temperature of recharging rainwater in the dune piezometer (P3) may also play a role.The monitored groundwater springs across the lagoon (Fig. 1) also both exhibited high temperature spikes of over 5 C during the storm (Fig. 6c), indicating that the natural hydraulic gradient was overcome by the storm surge and that warm lagoon water was pushed up into the spring fractures.

Spatial patterns in changing lagoon stratification and vertical mixing
The thermal and specific conductance offsets yield interesting insight into the lagoon stratification and vertical mixing dynamics.In particular, the relative locations of the entrance and northeast arm arrays cause them to experience different dominance in marine vs. freshwater forcing.For example, before Fiona, the entrance array had a generally positive thermal offset when tidal effects are averaged (Fig. 3c), while the northeast arm array thermal offset was generally negative (Fig. 3d).We postulate that this negative offset (cooler water on top) in the northeast arm is due to the cool freshwater from the springs and tributaries "floating" above the denser, warmer brackish lagoon water.In contrast, at the entrance, the brackish lagoon water lies above the denser, cooler saltwater from the ocean, causing a positive thermal offset.
During the first half of Fiona, the decrease in the thermal offsets at both arrays to $ 0 C (Fig. 3c,d) revealed vertical mixing from the wind, surge, waves, and current.The increase in the thermal offset at the entrance array during the latter half of Fiona (Fig. 3c) is likely due to the delayed response of cool, fresh water draining from the watershed following heavy precipitation.After Fiona, the maximum daily thermal offset stayed lower than pre-Fiona conditions in the entrance array (Fig. 3c) but increased after September 30 in the northeast arm Fig. 5. TSS loading (discharge times concentration) from the tributaries (see Fig. 1c for autosampler locations).
(Fig. 3d).These results respectively illustrate decreased (entrance) and increased (northeast arm) stratification after Fiona compared to pre-storm conditions, as also revealed by the specific conductance offsets discussed below.While these thermal offsets are all relatively low, the summertime thermal offset in the Basin head lagoon can be up to 5 C, and thus the timing of a coastal storm could strongly influence the associated thermal disturbance due to destratification.
The responses of the specific conductance offsets to Fiona at the entrance and northeast arm arrays also illustrate how the entrance array was more impacted by ocean dynamics, while the northeast arm array was more impacted by freshwater forcing.The stratification (specific conductance offset) decreased at the entrance array following Fiona due to the mixing from the surge and waves (pink series, Fig. 3e), while the stratification at the northeast arm array increased after Fiona due to high freshwater inputs in accordance with the estuarine Richardson number (Fischer et al. 1979).The increase in fresh groundwater discharge from the springs in the arm, which is comparable to streamflow during summer (KarisAllen et al. 2022), could also help maintain a more stratified system in the northeast arm following the storm.These contrasting changes to the stratification and mixing dynamics in the main basin vs. northeast arm of the lagoon could have profound implications for biota sensitive to changes in water temperatures and/or salinity.

Timeline of nitrate dynamics
The delayed response in NO 3 -N concentrations (Fig. 4d) arises from the time required for the recharged meteoric water to mobilize root-zone nitrate into the rivers via interflow and shallow groundwater, and to push out deeper "old" nitraterich water from the aquifer (Paradis et al. 2018).The sustained loading, which is due to the transport through the groundwater system, indicates that extratropical cyclones can trigger elevated nutrient loading that persists much longer than the surface effects of high flows, as has been recently reported elsewhere (Diego-Feliu et al. 2022;Pavlovskii et al. 2023), with the timing likely related to the shallow and deep nitrate transport pathways.Given that nitrate loading is the primary driver of eutrophication in Basin Head and in many other coastal waters worldwide (Gonz alez-De Zayas et al. 2021), such sustained nutrient loading may trigger sustained coastal ecosystem eutrophication, at least when storms occur in summer or early fall when the water is warm.

Comparison to other studies
The study results are qualitatively consistent with the impacts of tropical cyclones on coastal ecosystems observed in other, lower-latitude locations.Similarities include significant sediment movement, enhanced nutrient transport, and mixing of formerly stratified zones (Tanaka et al. 2003;Galv an et al. 2012;Barik et al. 2017).The recovery time (or lack thereof) of specific conductance, pH, and nutrient levels in the Basin Head lagoon generally aligns with findings from both Gonz alez-De Zayas et al. ( 2021) and Walker et al. (2021) who respectively observed similar dynamics in the Laguna Larga coastal lagoon in Cayo Coco, Cuba and in the Guadeloupe Estuary leading to a coastal lagoon in Texas, USA after hurricanes Harvey (2017) and Irma (2017).Gonz alez-De Zayas et al. ( 2021) reported a storm surge of 0.85 m, with a minimum pressure of 93 kPa and wind gusts of 53.9 m s À1 .This storm had much higher windspeeds than Fiona with slightly lower pressure (2 kPa), but the storm surge was approximately half the magnitude.Walker et al. (2021) observed peak windspeeds of 27.2 m s À1 , which are more comparable to Fiona.They observed streams in the area increasing in flow from < 1 to > 3 orders of magnitude, with the more pronounced changes greater than we observed in the Basin Head tributaries.Barik et al. (2017) studied a much stronger 2013 category 5 cyclone storm on a brackish lagoon in eastern India with windspeeds gusting to 61.1 m s À1 and observed a decrease in nutrient concentrations attributed to dilution from the influx of freshwater, similar to our observations during the passage of Fiona.They also noted a significant drop in lagoon salinity that had not fully recovered 4 yr later.Likewise, the Basin Head lagoon's specific conductance had not recovered to pre-storm levels by the end of our, admittedly shorter, monitoring period.
Despite these past studies that investigated the impacts of extreme coastal storms on lagoons, to our knowledge, the present study was the first to instrument the connected atmospheric and aquatic lagoon system and consider changes to the atmospheric conditions, freshwater inflows and solute loading, groundwater system, ocean forcing, and the coastal ecosystem response during a tropical (or extratropical) cyclone.The results, which were reported at higher latitudes than past studies, reveal profound changes to the physicochemical environment and ecosystem dynamics.Unlike previous studies (Kennish and Paerl 2010;Gonz alez-De Zayas et al. 2021;Walker et al. 2021), we did not observe anoxic conditions, but the integrated ecosystem response to the storm disturbance was still evident in reduced diel DO amplitudes.

Climate change implications
The impacts of hurricanes on coastal ecosystems at higher latitude locations like Atlantic Canada warrant far more investigation as such storm tracks are projected to become more frequent and intense (Collins et al. 2019) and may occur during warmer periods of the year when coastal ecosystems are subject to thermal or nutrient stress.In general, the North Atlantic is expected to experience more extreme surges, waves, and event-based precipitation due to cyclones maintaining their intensity further north (Horton and Liu 2014).Increases in ocean temperature are already causing earlier and longer hurricane seasons in the North Atlantic (Truchelut et al. 2022), and ocean warming has contributed to more extreme precipitation associated with hurricanes in this region (Reed et al. 2022).Also, the future impacts of extreme coastal storms on high water levels will be superimposed on the impacts of relative sea-level rise, which is projected to be particularly high in Atlantic Canada (Manson et al. 2019).
Associated ecosystem changes due to these long-and shortterm forcings will also occur at disparate timelines.For example, from an event perspective, sustained, elevated nitrate loading due to extreme precipitation associated with tropical cyclones could become more of an issue in Prince Edward Island (Pavlovskii et al. 2023) and in other nutrient-sensitive coastal waters worldwide (Maúre et al. 2021).Coastal lagoons like Basin Head will also experience multi-decadal bathymetry, salinity, and thermal changes that will in turn influence the way the ecosystem responds to future coastal storms.These multi-frequency dynamics due to climate change are complex and the subject of ongoing research using field data to constrain coupled hydrodynamic and hydrologic models of the Basin Head lagoon and watershed.Given that high water temperatures, invasive species, and nutrient levels have been identified as the primary stressors for the Giant Irish Moss in Basin Head, the impacts of extreme events that are already being experienced and are projected to intensify in the future should be considered in the ecosystem management plans for this federal Marine Protected Area.

Conclusions
This study investigated the holistic impacts of Extratropical Cyclone Fiona (September 2022) on a highly vulnerable coastal lagoon ecosystem in Atlantic Canada.The dense sensor network in the lower atmosphere, freshwater tributaries, lagoon, springs, and salt marsh revealed the following: • Sharp increases in freshwater discharge (up to 30-fold increase), nitrate-nitrogen concentration (doubling of concentration, sustained for months), nitrate-nitrogen loading (70-fold increase), and suspended sediment loading (40-fold increase) to the lagoon due to extreme rainfall.• A 1.6 m storm surge in the lagoon that disturbed the previously stratified lagoon and altered salinity and water temperature patterns throughout the lagoon.• Sudden changes in groundwater levels and temperatures as monitored in shallow piezometers and groundwater springs due to intruded seawater and associated pulses of advected heat that reveal pronounced groundwater-surface water mixing.• Decreases in DO diel ranges following Fiona, which suggests that the physicochemical changes noted above disrupted ecosystem processes.
The recovery period for the integrated aquatic system ranged from hours for storm surge levels and water temperatures to months for nutrient levels in the freshwater tributaries, with the difference in timescale depending primarily on the different pathways (i.e., atmosphere, ocean, surface water, or groundwater) for the disturbances.As storms like these become more frequent, the recovery period may begin to approach event interarrival periods, which could cause catastrophic and permanent loss of threatened biota.The study findings emphasize the vulnerability of coastal ecosystems to extreme events and the need for proactive measures to mitigate their impact and ensure the survival of important coastal species.

Fig. 1 .
Fig. 1.Map of the Basin Head lagoon and the two main tributary watersheds (c) and location in Canada (a) and Prince Edward Island (PEI) (b).Sensor locations associated with weather instruments, lagoon sensor arrays, salt marsh piezometers, springs, stream stilling wells, and sondes/autosamplers are indicated.Symbol shapes indicate sensor types (see the legend), while colors correspond to sensor locations as presented in other figures.Fiona tracks in (a) and (b) are from the National Hurricane Centre (2023), and symbols indicate Fiona's position every 6 h.

Fig. 2 .
Fig. 2. Hourly precipitation and windspeed (a) and air pressure (b) at the Basin Head meteorological station (Fig. 1c) during the passage of Extratropical Cyclone Fiona.

Fig. 3 .
Fig. 3. Water level (a), temperature and thermal offset (c, d), and specific conductance (e, f) data from the entrance and northeast arm lagoon arrays (see Fig. 1c for locations and colors), DO data from the northeast arm array (green) and the diurnal DO range (black, b).The storm duration from midnight on 23 September to midnight on 25 September is shaded in gray.

Fig. 6 .
Fig. 6.Groundwater depth above sensor (a) and groundwater temperatures (b) in the three piezometers and water temperature from loggers installed on rebar driven into the aperture of the two springs (c).Colors correspond to locations on Fig. 1c.