1. Introduction
Environmental monitoring is essential to understanding human impacts on natural ecosystems, and continuity is a key property of successful monitoring programs. Unfortunately, this was not possible when the COVID-19 pandemic hit and impacted almost every aspect of human life, including some unexpected ecological changes. The spring of 2020 brought the “anthropause”, a time with drastically constrained human activities, especially regarding industrial activity and travel (
Rutz et al. 2020). Headlines that declared pandemic lockdowns were having a positive effect on the environment quickly emerged (
Zerefos et al. 2021). In Venice, Italy, the canal bottom was seen for the first time in decades, and several large cities had improvements in air quality (
Bherwani et al. 2020;
Clifford 2020;
He et al. 2020). Although early indications were that the pandemic allowed some environmental renewal, contributing to improved air and water quality, these anecdotal observations do not provide a robust understanding of the pandemic's full environmental impacts (
Berman and Ebisu 2020;
Hallema et al. 2020;
Cooke et al. 2021).
One unfortunate effect of the COVID-19 pandemic was the pausing or cancelling of environmental monitoring programs. In the spring of 2020, much uncertainty surrounding the pandemic and concerns about maintaining social distancing led to a reduction in environmental monitoring (
Cooke et al. 2021). The U.S. National Park Service, for example, issued just 37% of its normal amount of research permits (
Miller-Rushing et al. 2021), while the Canadian federal government paused all water quality monitoring programs (
Zingel 2020). Compared to other locations, Ontario had extended closures during the summer of 2020, and details of the provincial re-opening plan can be found in
Howarth et al. (2021). Within Ontario, the Lake Partner Program, a province-wide volunteer-based water quality monitoring program that covers over 550 inland lakes, paused for the summer of 2020 (
Dorset Environmental Science Centre 2020). Along with federal and provincial governments, many local conservation authorities could not run their regular monitoring programs as employees were told to work remotely to reduce transmission of COVID-19 (
Akinsorotan et al. 2021). This multi-level monitoring shut-down reflected a data gap where thousands of lakes and streams in Ontario were not monitored for water quality during the entire summer of 2020.
The pause in water quality monitoring left gaps in long-term monitoring programs and constrained the ability to measure the immediate impacts of the COVID-19 pandemic on Ontario's lakes. In Ontario, the pandemic led to multiple public health interventions that restricted travel and social interactions, as well as changes in individuals’ hygiene habits (i.e., increased hand washing) (
Park et al. 2010;
Nielsen 2021). These changes improved air quality in many places, and it has recently been found that water quality may have been affected as well (
He et al. 2020;
Tokatlı and Varol 2021). Although most field monitoring was cancelled in 2020, some researchers used remote sensing to examine turbidity and found it decreased in the Ganga River and Vembanad Lake in India during the lockdown period (
Garg et al. 2020;
Yunus et al. 2020). Alternatively, in the Meriç-Ergen River Basin in Turkey, turbidity did not change, but there were reductions in metal(loid) levels (
Tokatlı and Varol 2021). The reductions in water quality parameters were attributed to reduced effluent from industrial sources and reduced pollution from human activities, such as tourism, in the area. Water consumed by households was also higher, with people staying home and increasing hygienic behaviours, such as hand washing (
Kalbusch et al. 2020;
Abu-Bakar et al. 2021). In areas with household septic systems, there was the potential for these systems to become overloaded, resulting in poor treatment performance. Excessive septic seepage can lead to increased nutrient concentrations in nearby surface waters (
Reay 2004;
Oldfield et al. 2020).
One avenue to continue research through laboratory closures was the implementation/expansion of community science monitoring programs.
Crimmins et al. (2021) found an increase in community science participation on the most popular community science programs (iNaturalist and eBird) in the spring of 2020 in the United States. However, they found the trends in community science greatly varied by geographic location, with higher participation in urban areas (
Crimmins et al. 2021). Although these programs can be useful for biodiversity monitoring, water monitoring, which often requires access to a laboratory, was harder to conduct during the pandemic; the largest community science program in Ontario, the Lake Partner Program, was shut down for the entirety of the spring and summer of 2020 (
Dorset Environmental Science Centre 2020). Alternatively, a survey of U.S. and Canadian community science water monitoring program coordinators (
Stepenuck and Carr 2022) found that 72% of programs planned to continue through the 2020 field season, despite delays. highlighting the flexibility of community science as a water monitoring tool.
While pandemic restrictions inhibited routine water quality monitoring from occurring in Ontario, Canada, in 2020, we demonstrate that a community science approach could address questions pertaining to lake water quality during a disruption of regular monitoring. As such, we (1) evaluated lake-front residents’ behaviours during the early months of strict pandemic measures in 2020, (2) assessed spatial nutrient dynamics across 16 lakes in the Kawartha Lakes region during the most restrictive pandemic year (2020), and (3) compared pandemic nutrient conditions in four Kawartha Lakes during a pre-pandemic (2019) year and pandemic years (2020 and 2021). This study offered the unique opportunity to test if there was a detectable effect of the anthropause in an agriculturally dominated region with high shoreline development. Given that agricultural activities were not affected by pandemic restrictions and more people were residing at lake-front properties during the summer of 2020, we hypothesized that nutrient inputs during the pandemic years would be higher than pre-pandemic years. Overall, we demonstrate that community science monitoring is a flexible tool that has the potential to enhance and, in some cases, replace mandated governmental monitoring work.
2. Materials and methods
2.1. Study site
The Kawartha Lakes region, located in south-central Ontario, is within a 1–2 h drive from the densely populated Greater Toronto Area (GTA) (
Fig. 1). Their proximity to the GTA and natural surroundings make the Kawartha Lakes an extremely popular tourist and cottage destination in Ontario (
City of Kawartha Lakes 2020). The Kawartha Lakes are also part of the Trent-Severn Waterway (TSW), a National Historic Site of Canada, which connects Georgian Bay to Lake Ontario. Tourism is one of the biggest industries for the City of Kawartha Lakes, with recreational activities on the lakes a major draw for the area, helping to bring in over 1.6 million visitors annually (
City of Kawartha Lakes 2020). The Kawartha Lakes watershed is also part of “The Land Between”—a biodiverse ecotone that reflects a geological shift from limestone to granite (
Alley 2006). The watersheds of these lakes also have over 55 provincially significant watersheds that provide key ecosystem services (
Kawartha Conservation 2023).
As part of the TSW, the Kawartha Lakes have controlled flow between the lakes through locks and dams. In our study lakes, the flow begins at Balsam Lake, feeds Canal and Mitchell to the west and Cameron to the east, and continues east with Katchewanooka as the final downstream lake in our study area (
Fig. 1 and Table S1). Lake Scugog is a headwater lake that flows into Sturgeon, and Sandy Lake is the only study lake not hydrologically connected to the TSW.
2.2. Community science model
Our research group had been conducting research with the local conservation authority, Kawartha Conservation, using community science for several years before the pandemic, with monitoring on five lakes included in this study (Scugog, Balsam, Cameron, Sturgeon, and Pigeon). Through this partnership, we had established a network of lake associations and volunteers to tap into this study; additionally, community science was an ideal approach that could ensure physical distancing during the pandemic years. We contacted a local environmental group, the Kawartha Lake Stewards Association (KLSA), which helped recruit additional volunteers through email and virtual meetings. We also partnered with Curve Lake First Nation, whose traditional lands and waters encompass the Kawartha Lakes region. Five volunteers from Curve Lake First Nation selected sites to monitor across their reserve territory, located between Buckhorn and Chemong Lake. Through partnering with these organizations, we recruited 58 community science volunteers to sample from 60 sites across 16 lakes in the first year of the pandemic (2020). Due to strict physical distancing requirements at this time, volunteers were asked to collect water samples with the provided containers and store them in their freezer until the end of the study period in September 2020.
Volunteer training was conducted virtually to ensure the safety of all participants. A training video was created and shared on YouTube, with a direct link sent to all volunteers. Two live virtual follow-up sessions were booked about a week after the release of the recorded video to answer questions and share additional information. Volunteers were also sent a document with visual and written instructions for collecting and storing water samples and recording field observations.
Finally, a vital component of community science research is information dissemination. All volunteers were provided with a final report, which included an overview of the project, a site-specific summary of water quality findings, and descriptions of some waterfront property best management practices and resources. Additionally, Curve Lake First Nation was provided with the site-specific results for all five of their monitoring stations, and with possession of these data, they are in control of using them as they see fit.
2.3. Sample collection
One week before the first sample collection date in 2020, water sampling kits were distributed to volunteers at four pick-up locations that included local marinas and volunteers' homes. Sample kits included eight high-density polyethylene (HDPE) 200 mL specimen containers, gloves, collection instructions, and a field datasheet. Monthly water samples were collected in duplicate by community science volunteers from June to September 2020. Samples were collected on the last Tuesday of the month between 8 and 9 am, for consistency across sites. If volunteers were unable to take the sample at the requested time, they recorded the date and time of their sample collection. Samples were collected by volunteers from their docks at a location where the depth was ∼1 m. Volunteers were instructed to fill their two specimen containers to 1 cm from the lid with lake water from approximately 10 cm below the surface. Volunteers were also required to fill out a data collection sheet where they recorded the air and water temperatures at their site, along with any other observations about the weather or water conditions. Once samples were collected, the labelled specimen cups were placed in the volunteers’ freezer until sample pick-up in late September. Overall, 227 samples were returned, resulting in a participation rate of 95%.
In 2019, water samples were collected at sites on Scugog, Balsam, Cameron, Sturgeon, and Pigeon lakes, and all of the above lakes were sampled in 2021 at the same sites except for Scugog. In 2019, water sampling kits were delivered directly to the volunteers’ homes, and in 2021, the pick-up depots established in 2020 were used to deliver and collect water samples. Sample kits in 2019 and 2021 were distributed monthly and included two acid-washed 1 L HDPE Nalgene bottles, one 100 mL sterile specimen cup, gloves, collection instructions, and a field datasheet. Water samples were collected monthly from June to September, with Scugog, Balsam, and Cameron lake samples collected on a Tuesday and Sturgeon and Pigeon lake samples collected on the following Thursday. Samples were picked up by researchers on the same day they were collected and kept on ice until returned to the laboratory. Aliquots were poured for phosphorus and nitrogen analysis and frozen within 24 h of sample collection.
Precipitation data was retrieved from Kawartha Conservation's monitoring station at Ken Reid Conservation Area, Lindsay. These data were used for all sites as they are centrally located and there are not sufficient weather stations in the region to have coverage by lake or watershed. Total precipitation in the 4 days prior to sample collection was calculated, and the presence/absence of a storm event (>15 mm of precipitation in a 24 h period) was recorded for each sample event.
2.4. Volunteer survey deployment
With various pandemic-related lockdown measures in place in 2020 and 2021, we wanted to examine if there was an impact of these measures on homeowners’ habits. Of particular interest were changes to the number of people and time spent at the lake-front property, habits that impacted septic tank loads, and lake-front property maintenance. Researchers created an anonymous online survey to investigate changes in waterfront property-owner habits. The survey was approved by the Ontario Tech University Research Ethics Board (REB) on 29 May 2020 (Supplementary material, REB# 15910). The survey was sent to all volunteers in July 2020 and asked participants to compare their activities at their lake-front property to the previous year (2019), with follow-up questions asking respondents to specify their activities, such as gardening habits as well as demographic questions. These comparisons were used to analyze trends in lakefront property owner habits before and during the COVID-19 pandemic. When the survey closed in the fall of 2020, there were 45 responses.
2.5. Water sample processing
Frozen samples from 2020 were thawed for analysis upon return to the laboratory in September, and frozen samples from 2019 and 2021 were thawed and analyzed within a month of sample collection. Frozen samples were sent to the SGS Environmental Analytical Laboratory in Lakefield, Ontario (SGS) for nitrogen suite analysis (nitrite, nitrate, ammonia + ammonium, and total Kjeldahl nitrogen). SGS is accredited for environmental tests by the Canadian Association for Laboratory Accreditation Inc. (CALA). Samples were thawed for total phosphorus (TP) and chlorophyll-a (Chla) analysis in the laboratory. TP was determined spectrophotometrically based on a modified ascorbic acid method (
Murphy and Riley 1962). Due to the samples being previously frozen in 2020, an Aquafluor handheld fluorometer (Turner Designs, Sunnyvale, CA) was used to estimate relative Chla values. Chla samples collected in 2019 and 2021 were filtered within 24 h of collection and determined spectrophotometrically using a 90% acetone extraction method (
Kirkwood et al. 1999). Due to the different methodologies and units used for Chla in 2020 (relative units) and 2019 and 2021 (mg/L), Chla values cannot be directly compared across years.
2.6. Statistical analysis
Statistical analyses were completed using the R program (
R Core Team 2023) in RStudio (
RStudio Team 2020), and all figures were created using the
ggplot2 package (
Wickham 2011). An exploratory data analysis was conducted to identify outliers. Points that fell out of the 1.5 times the interquartile range were examined to determine if there was a sampling error. One site on Lake Scugog that has been monitored previously had extremely elevated TP and TN values that appeared to be due to sampling error and were thus removed from the analysis. TP and Chla were log-transformed due to the non-normal distribution of residuals. Spatial variation was analyzed with a nested analysis of variance (ANOVA) to determine if there were differences between lakes, with Tukey's post hoc test to determine specific differences between lakes. Permutational ANOVA (PERMANOVA) was conducted with
vegan (
Oksanen et al. 2020) based on Bray–Curtis dissimilarity to determine if differences between watersheds were significant. A pairwise post hoc test was performed using the
pairwise.adonis (
Martinez Arbizu 2020) R package. Temporal trends (2019–2021) were examined with a mixed-effect model, with month nested in year as fixed effects and lake as a random effect. Pairwise comparisons across years were conducted with the
emmeans package (
Lenth 2016). Welch's
t-test was conducted on survey data to determine if there were significant differences between property owners’ habits in 2019 and 2020. An unconstrained ordination was conducted on the raw water quality data (2019–2021). First, a detrended correspondence analysis (DCA) was conducted to determine whether the data followed a linear or unimodal response. The standard deviation of the first DCA axis was less than three, and such a principal component analysis (PCA) was conducted with the R package
vegan (
Oksanen et al. 2020).
Acknowledgements
The authors acknowledge the lands and people of the Mississaugas of Scugog Island First Nation, for whose traditional territory this study was conducted. We are thankful to be welcomed on these lands in friendship. These lands are covered under the Williams Treaties and are the traditional territory of the Mississauga, a branch of the great Anishinaabeg Nation, including Algonquin, Ojibway, Odawa, and Pottawatomi. Thanks to the Kawartha Lake Stewards Association for their help in recruiting volunteers and connecting researchers to the local community. Thanks to Gary Pritchard, who connected us to Curve Lake First Nation, and to all the community scientists for their contributions to this study. Finally, we would like to thank the Ontario Tech University Research Ethics Board (REB) for taking the time to review and provide feedback on the online survey, which they approved on 29 May 2020 (REB# 15910).