Data

Data Tables

The following data table illustrates the variables measured in the analysis of climate21 and cyanobacteria data20 (Table 2). The observed predictor variables used in this study are maximum temperature (°C) (continuous), minimum temperature (°C) (continuous), precipitation (mm) (continuous), and SWE (cm) (continuous). The response variables are Microcystin LR-equivalent concentration (µg/L), total cyanobacterial cell count (cells/mL), Microcystis mcyE (copies/mL), Anabaena mcyE (copies/mL), and Planktothrix mcyE (copies/mL). Design variables include township location, ecoregion, date, year, and waterbody name. 

Data was then aggregated by ecoregion and year to allow the annual comparison between the ecoregions (Table 3). Both minimum and maximum temperatures (°C) are the average of the respective temperatures recorded from May to October in a particular year within the ecoregion the samples were taken from. Precipitation is calculated as the total precipitation (mm) received from May to October in a particular year within the township and averaged across the ecoregion. SWE (cm) is the maximum equivalent quantities of water received in that particular calendar year averaged across townships in a particular ecoregion. Microcystin LR-equivalent concentration (µg/L), total cyanobacterial cell count (cells/mL), Microcystis mcyE (copies/mL), Anabaena mcyE (copies/mL), and Planktothrix mcyE (copies/mL) are the totals quantities seen from May to October in a particular year within the specified ecoregion, with measurements of each described in Table 1. 

Exploratory Graphics

To visually assess the data, RStudio24 was used to create graphics. The package ggplot225 was used to make the following line plots, separating the data by ecoregion. The following RStudio packages were also used in the graphical creations: ggsci26 and ggpubr27. 

This visual analysis was used to assess outliers in the climate data, as there should be a similar value for the annual means across the townships within the same ecoregion. There is expected to be some variation, such as the higher precipitation value seen in one of the Boreal townships, due to the geographic spread of the townships across the entire province and the chance of extreme events. The precipitation and temperature data are consistent across all townships within each ecoregion and across the province (Fig. 3-6). Notable outliers include one Boreal ecoregion township precipitation due to a more northern location compared to the majority of this ecoregion, a sub-selection of the Foothill ecoregions townships due to the geographic spread and influence of the topography of the area on precipitation events, and one Foothills ecoregion townships' minimum temperature due to it being more southern compared to other townships in this ecoregion. These data locations were kept in the analysis as they represent the variability seen within an ecoregion. 

This also analysis helped illustrate the data which was only collected in certain years (Microcystis mcyE: 2012-2016; Anabaena mcyE: 2013-2016; Planktothrix mcyE: 2013-2016) and any null values measured (Fig. 7 and 8). Figure 7 further allowed for the bi-annual sampling design to be seen, which led to the removal of all no data and zero values from the data set. The average total cyanobacterial cell count was then taken, which is the average cell count seen within an ecoregion in a specified year. 

Figures 3 to 8 illustrate the respective climate or cyanobacterial measurements taken over the year considered, for each township and are coloured by ecoregion. 

Data was then plotted by month across all years for the total cyanobacterial cell count to determine the months in which there are higher historical records of blooms, split by ecoregion (Fig. 9). All zeros were removed from the data before this was completed. 

Statistical Tests

Monthly climate data was then taken for each ecoregion across each year, taking the average minimum temperature across the months of May to October, the average maximum temperature from May to October, the average total precipitation received from May to October, and the maximum SWE received from January to August. The total cyanobacterial cell count was taken as the average total quantities measured for the year, with any ecoregions and years not containing any observations being removed from the base data. 

Linear models were used to calculate the data plotted in Figures 14 to 18, with the numerical data illustrated in Table 4.