Trends in chlorophyll-a concentrations in Europe's transitional, coastal and marine waters, for two periods: pre 2000 (1980-1999) and post 2000 (2000-2023)
Chart (interactive)
The bar charts show trends in mean summer concentrations of chlorophyll-a (µg/l) in the upper 10 meters of the water column during the assessment period 1980-2023, divided into two periods: pre-2000 and post-2000, by subregion and combined for Europe's seas. Green bars indicate time-series in grid cells with significant (p<0.05) decreasing trends; orange bars show time-series in grid cells with significant (p<0.05) increasing trends; and grey bars show time-series in grid cells with no significant trends.
European Environment AgencyNIVA Denmark water reaserch
Units
Number of time series (entire numbers)
Additional information
* Trends were calculated using the hybrid EEA assessment grid (100+20km). * Results reflect data availability, which may vary considerably across regions and over time. The temporal range specified is the maximum range of dates for raw observations but this range varies depending on the availability of observations at each location.
Methodology
* Average summer concentrations of chlorophyll-a (µg/l) and trends were calculated using data from observations in the upper 10m of the water column. Data with bad quality flags (bad value, probably bad value), uncertain (value in excess, uncertain value, missing value) or duplicates were discarded and the 99.9% percentile of the highest values were sorted out. * Data were aggregated into squares with sides of 1.375km for coastal stations within 20km from the coastline and 5.5km for open water stations more than 20km away from the coastline. A masking of 5km inwards from the coast was also applied. * The trends were calculated based on yearly averages for each grid cell. The averages were calculated using a combination of estimated marginal means (EMM) and arithmetic means, where the EMM were used for time series containing data from both different years and months and the arithmetic means were used for time-series with yearly observations. * Temporal trends were estimated with the non-parametric Mann-Kendall test. Only stations having data of at least 5 years were considered. A p-value <0.05 was used to consider a statistically significant trend. * Growing season months considered are June-September for stations within the Baltic Sea north of 59 N and May-September for all other stations.