Skip to main content

Maritimes Ecosystem Indicator Catalogue

7 Satellite Ocean Colour

Data Type: Gridded Data

Spatial Scope: Northwest Atlantic

Duration: 1998-2024

Source: SOPhyE group, DFO

Contact:

7.1 Introduction to Indicator

The Satellite Ocean Colour dataset is a product of the Satellite Ocean Color and Phytoplankton Ecology (SOPhyE) group at Bedford Institute of Oceanography. Satellite Ocean Colour data uses remotely sensed multispectral data capturing the reflectance properties of the ocean to proximally serve as an indicator of plankton biomass, chlorophyll, and productivity of marine environments. Given differences in absorption and reflectance between plankton species and non-algal particulates, multispecral ocean color measurements categorize the particulate content of water, and infer indators of algal abundnace, biomass, and productivity.

Here, satellite ocean color is used to infer five indicator variables across space:

  • Start Day of spring phytoplankton bloom
  • Duration of the spring phytoplankton bloom
  • Maximum concentration during the spring phytoplankton bloom (amplitude)
  • Total chlorophyll-a produced during the spring phytoplankton bloom (magnitude)
  • Average chlorophyll-a produced annually

Each of these indicators are expected to respond to oceanographic variability and long-term climate change, with variable responses across space or between algal species At mid-latitudes, climate change is expected to delay the date of spring blooms at a global scale (Henson et al. 2018), and changes in other oceanographic variables such as stratification can affect bloom magnitude (Song et al. 2011).

7.2 View Data 


# prepare sf shapes as coordinates
# NAFO regions
sf <- sf %>% st_transform(crs = terra::crs(t_duration))
# coastline
data("coastline")
# get timespan
timespan <- paste0(
  str_replace(years[1],"y.",""),
  "-",
  str_replace(years[length(years)],"y.","")
  )
# get bounding box 
bbox <- ext(t_start[[1]])


# plot means over timespan ------------------
means_baseplot <- function(df){
  
  # find mean of every cell, excluding NAs
  mean_df <- app(df, "mean", na.rm=T)
 
  # base plot
  p <- ggplot() +
    geom_sf(data = coastline) +
    geom_spatraster(data = mean_df,aes(fill = mean)) +
    geom_sf(data = sf, aes(color = I(color)),
            fill = "transparent", linewidth = .5)+
    coord_sf(xlim = c(bbox$xmin, bbox$xmax),
             ylim = c(bbox$ymin, bbox$ymax), expand=F) +
    scale_fill_viridis_c(na.value = "transparent") +
    theme_light(base_size = 9) +
    theme(legend.position = "bottom",
          legend.title.position = "top",
          legend.key.height = unit(6,"pt"),
          legend.key.width = unit(24,"pt"))
   return(p)
}

## plot all mean values
mean_start_p <- means_baseplot(t_start) + 
  scale_fill_viridis_c(breaks = c(60, 121),labels = c("Mar 1","May 1"),
                       na.value = "transparent") +
  labs(title = "Mean Bloom Start",
       subtitle = timespan,
       fill = "Day of Year")

mean_duration_p <- means_baseplot(t_duration) + 
  labs(title = "Mean Bloom Duration",
       subtitle = timespan,
       fill = "Number of Days")

mean_amplitude_p <- means_baseplot(amplitude) + 
  labs(title = "Mean Bloom Amplitude",
       subtitle = timespan,
       fill = "Max. Concentration (mg/m^3)")

mean_magnitude_p <- means_baseplot(magnitude) + 
  labs(title = "Mean Bloom Magnitude",
       subtitle = timespan,
       fill = "Total Chlorophyll (mg/m^3)")

mean_mean_p <- means_baseplot(mean) + 
  labs(title = "Mean Avg. Chl.",
       subtitle = timespan,
       fill = "Chlorophyll (mg/m^3)")


# second, slopes per cell
trends_baseplot <- function(df){
  
  # function to apply to extract slopes over time for every cell
  slope_fun <- function(x) {
    if (all(is.na(x))) return(NA)
    # exclude cells that are 50% or more NA
    if (sum(is.na(x)) > length(x)*.5) return(NA)
    lm(x ~ c(1:length(x)))$coefficients[2]
  }
  # find mean of every cell, excluding NAs
  slope_df <- app(df, slope_fun)
  names(slope_df) <- "trend"
 
  # base plot
  p <- ggplot() +
    geom_spatraster(data = slope_df,aes(fill = trend)) +
    geom_sf(data = coastline) +
    geom_sf(data = sf, aes(color = I(color)),
            fill = "transparent", linewidth = .5)+
    coord_sf(xlim = c(bbox$xmin, bbox$xmax),
             ylim = c(bbox$ymin, bbox$ymax), expand=F) +
    scale_fill_gradient2(low = "#A6611A", high = "#018571",
                         na.value = "grey75") +
    theme_light(base_size = 9) +
    theme(legend.position = "bottom",
          legend.title.position = "top",
          legend.key.height = unit(5,"pt"),
          legend.key.width = unit(20,"pt"))
   return(p)
}

## plot all slope values
trend_start_p <- trends_baseplot(t_start) + 
  labs(title = "Change in Bloom Start",
       subtitle = timespan,
       fill = "Days/Year")

trend_duration_p <- trends_baseplot(t_duration) + 
  labs(title = "Change in Bloom Duration",
       subtitle = timespan,
       fill = "Number of Days/Year")

trend_amplitude_p <- trends_baseplot(amplitude) + 
  labs(title = "Change in Amplitude",
       subtitle = timespan,
       fill = "Concentration/Year")

trend_magnitude_p <- trends_baseplot(magnitude) + 
  labs(title = "Change in Magnitude",
       subtitle = timespan,
       fill = "Total Chlorophyll/Year")

trend_mean_p <- trends_baseplot(mean) + 
  labs(title = "Change in Avg. Chl.",
       subtitle = timespan,
       fill = "Concentration Change/Year")



library(patchwork)
full_p <- 
  (
    ((mean_start_p | mean_duration_p | mean_amplitude_p | mean_magnitude_p |mean_mean_p )  +
       plot_layout(axes = "collect") & 
       theme(axis.text = element_blank(),
             legend.margin = margin(t = -4))
     )/
      (
        (trend_start_p | trend_duration_p | trend_amplitude_p | trend_magnitude_p |trend_mean_p ) +
          plot_layout(axes = "collect") &
          theme(axis.text = element_blank(),
                legend.margin = margin(t = -4))
        )
    )

full_p #+ ggview::canvas(10,7)
Summary values for Satellite Ocean Colour across space within the study region (Northwest Atlantic); 1998-2024. Top row shows mean values per raster cell of five ocean colour indicators. Bottom row shows slope values for variable values across time using linear models, with cells that contain 50% or more NA values excluded and shown in grey. Target NAFO regions for this report are shown in purple and orange. Note that many variables/year combinations have NA values, so more precise filtering may be required for more in-depth analyses.

Figure 7.1: Summary values for Satellite Ocean Colour across space within the study region (Northwest Atlantic); 1998-2024. Top row shows mean values per raster cell of five ocean colour indicators. Bottom row shows slope values for variable values across time using linear models, with cells that contain 50% or more NA values excluded and shown in grey. Target NAFO regions for this report are shown in purple and orange. Note that many variables/year combinations have NA values, so more precise filtering may be required for more in-depth analyses.

7.4 Relevance to Research and Stock Assessments

Phytoplankton form the base of the marine food web, so bloom dynamics are directly relevant to fisheries and marine ecosystems through bottom-up effects. Changes in algal bloom dynamics can cause “mismatches” in timing between consumers (e.g., marine larvae) and algal nutrient sources (Platt, Fuentes-Yaco, and Frank 2003), and bloom activity can also include algal species that are harmful to ecosystems or humans (Boivin-Rioux et al. 2022).

In Atlantic Canada, spring bloom timing anomalies are predictive of larval fish survival, where earlier blooms result in greater survival of key harvested fish species (Platt, Fuentes-Yaco, and Frank 2003).

7.5 Variable Definitions

Satellite Ocean Colour is stored as a stars object in marea, with five data attributes arranged over three dimensions (x, y, time). See attribute names and definitions in the table below (Table 7.2)

Table 7.2: Attribute names and definitions in the Satellite Ocean Colour stars dataset.
attribute description unit
t_start_value Bloom start date Day of year
t_duration_value Bloom duration Days
amplitude_real_value Maximum concentration during the spring phytoplankton bloom period mg/m3
magnitude_real_value Total chlorophyll-a produced during the spring phytoplankton bloom period days*mg/m3
annual_mean_value Average chlorophyll-a over the year mg/m3

7.6 Additional Data

Satellite Ocean Colour Data is presented as gridded data over the Northwest Atlantic region, with no regional groupings. Summary values displayed on this page were generated, and are not explicitly provided in the marea package.

7.7 Get the Data 

library(marea)
library(stars)
data('satellite_ocean_colour')
plot(satellite_ocean_colour, style = "fill")

References

Boivin-Rioux, Aude, Michel Starr, Joel Chasse, Michael Scarratt, William Perrie, Zhenxia Long, and Diane Lavoie. 2022. “Harmful Algae and Climate Change on the Canadian East Coast: Exploring Occurrence Predictions of Dinophysis Acuminata, d. Norvegica, and Pseudo-Nitzschia Seriata.” Harmful Algae 112: 102183.
Henson, Stephanie A, Harriet S Cole, Jason Hopkins, Adrian P Martin, and Andrew Yool. 2018. “Detection of Climate Change-Driven Trends in Phytoplankton Phenology.” Global Change Biology 24 (1): e101–11.
Platt, Trevor, Csar Fuentes-Yaco, and Kenneth T Frank. 2003. “Spring Algal Bloom and Larval Fish Survival.” Nature 423 (6938): 398–99.
Song, Hongjun, Rubao Ji, Charles Stock, Kelly Kearney, and Zongling Wang. 2011. “Interannual Variability in Phytoplankton Blooms and Plankton Productivity over the Nova Scotian Shelf and in the Gulf of Maine.” Marine Ecology Progress Series 426: 105–18.