Go back to "Publications"

Aggregating trend data for environmental reporting

Trend analyses performed on many sites are regularly aggregated by water quality variable and summarised in tabular, graphical or map format as part of environmental reporting. The intention of aggregated site summaries (eg, proportion improving and degrading, by variable) is to provide an overview of recent water quality changes over a domain of interest (eg, the entire country, a region, an environment class). The purpose of this study was to develop new methods for aggregating site trends representing a spatial domain of interest that incorporate all available information and that quantifies the uncertainty of the aggregate statistic. The new trend aggregation methods reduce the risk of misinterpreting insignificant or insufficient data trends and make maximal use of the available information. This report describes the new aggregation methods and provides a case study of their application to a national dataset of river water data that was first reported by Larned et al. (2015).

Trend analyses performed on many sites are regularly aggregated by water quality variable and summarised in tabular, graphical or map format as part of environmental reporting. The intention of aggregated site summaries (eg, proportion improving and degrading, by variable) is to provide an overview of recent water quality changes over a domain of interest (eg, the entire country, a region, an environment class). The purpose of this study was to develop new methods for aggregating site trends representing a spatial domain of interest that incorporate all available information and that quantifies the uncertainty of the aggregate statistic. The new trend aggregation methods reduce the risk of misinterpreting insignificant or insufficient data trends and make maximal use of the available information. This report describes the new aggregation methods and provides a case study of their application to a national dataset of river water data that was first reported by Larned et al. (2015).