Abstract
As part of a larger project funded by the Danish Environmental Protection Agency (“Effects on the quality elements defined by the EU Water Framework Directive (WFD) of other pressure factors than excess nutrient load and climate change”) a number of environmental pressure factors other than excess nutrient loading and climate change have been identified as potential risks to all of the quality elements of the WFD (Petersen et al. 2018). When multiple pressure factors have been identified as being of importance to one or more of the WFD quality elements, the next task for environmental managers and their scientific support should be to cumulate their effects. This is easily done when the different anthropogenic pressures induce similar effects on the environment and can be cumulated by addition, e.g. resuspension of sediment by gravel extraction can be anticipated to have similar effects on water transparency as resuspension caused by deepening of shipping lanes or bottom trawling and can thus be added. However, pressures generated by e.g. invasive species and excess loading of nutrients have completely different effects on the marine environment and the WFD quality elements. Cumulating these pressures to calculate their overall impact is therefore not straight forward.
In this report, we critically review the methods that have been proposed for cumulating anthropogenic pressures. These include the Delphi method, mechanistic models, univariate and multivariate statistical methods, Bayesian models and machine learning. As the need for cumulating effects is primarily related to pressures whose effects are non-additive i.e. are synergistic, antagonistic or dominant, we have focussed on methods that are able to quantify such interactions. We use a number of criteria to assess the applicability of each method based on 1) the requirements of the method in terms of input data and process knowledge. Very data-intensive methods will have limited utility because, for a number of pressure factors, there is little data; 2) the procedure for selecting the relevant pressure factors and indicators; 3) the nature of the results given by the method. Does the method detect interactions between pressure factors and quantitatively cumulate their environmental impact; can the method quantify the contribution of individual pressure factors to the environmental impact; and can the method estimate the statistical significance of the relationships between the pressures and their cumulative effects on the quality elements; 4) the adequacy of the temporal and spatial resolution of the method; and 5) the applicability of the method in relation to the WFD.
Table 8.1 summarizes the analysis of the different methods with respect to data requirements and necessary process knowledge. In terms of applicability, it is concluded that, to the extent that observations of sufficient quality and knowledge on causal relationships between pressures and effects are available, statistical methods, Bayesian models, mechanistic models and machine learning all can be used to quantify the relationships between pressure factors and environmental conditions, also when non-additive cumulative effects are important. In contrast, the Delphi model and similar mapping tools based on super positioning layers of pressure factors on ecosystem elements using expert knowledge to weight their importance, can primarily be used as risk assessment tools because they do not quantify the cumulative effects of the pressures on the ecosystem elements, in casu the quality elements of the WFD. Hence, these tools cannot be used to quantify the relative importance of different pressures in a given area, identify thresholds of cumulative impacts, design measures of impact prevention, or determine to what extent such measures are required.
The analysis further shows that very few peer reviewed investigations of the cumulative effects of various pressure factors have been conducted in the marine environment, and that only a minority of these are relevant for the quality elements used by the Danish environmental authorities in relation to the WFD. The lack of specific analysis of cumulative effects by multiple pressures is likely to be caused by the focus on the excess load of nutrients, especially nitrogen, to Danish coastal marine environments, where it has been considered to constitute the most important pressure. Another explanation is that a large number of other identified pressures, when considered individually, are unlikely to have significant effects on the quality elements at the waterbody level (Petersen et al 2018). Finally, there is generally limited experimental knowledge on interactions between multiple pressure factors on specific quality elements, and a limited amount of data on the importance of the various pressure factors except for nutrient supply. These knowledge gaps, and a lack of empirical analyses of the cumulative effects of the different pressures, are mainly due to a lack of data and process knowledge, and seem not, or only to a very limited extent, to be caused by a lack of methods.
In this report, we critically review the methods that have been proposed for cumulating anthropogenic pressures. These include the Delphi method, mechanistic models, univariate and multivariate statistical methods, Bayesian models and machine learning. As the need for cumulating effects is primarily related to pressures whose effects are non-additive i.e. are synergistic, antagonistic or dominant, we have focussed on methods that are able to quantify such interactions. We use a number of criteria to assess the applicability of each method based on 1) the requirements of the method in terms of input data and process knowledge. Very data-intensive methods will have limited utility because, for a number of pressure factors, there is little data; 2) the procedure for selecting the relevant pressure factors and indicators; 3) the nature of the results given by the method. Does the method detect interactions between pressure factors and quantitatively cumulate their environmental impact; can the method quantify the contribution of individual pressure factors to the environmental impact; and can the method estimate the statistical significance of the relationships between the pressures and their cumulative effects on the quality elements; 4) the adequacy of the temporal and spatial resolution of the method; and 5) the applicability of the method in relation to the WFD.
Table 8.1 summarizes the analysis of the different methods with respect to data requirements and necessary process knowledge. In terms of applicability, it is concluded that, to the extent that observations of sufficient quality and knowledge on causal relationships between pressures and effects are available, statistical methods, Bayesian models, mechanistic models and machine learning all can be used to quantify the relationships between pressure factors and environmental conditions, also when non-additive cumulative effects are important. In contrast, the Delphi model and similar mapping tools based on super positioning layers of pressure factors on ecosystem elements using expert knowledge to weight their importance, can primarily be used as risk assessment tools because they do not quantify the cumulative effects of the pressures on the ecosystem elements, in casu the quality elements of the WFD. Hence, these tools cannot be used to quantify the relative importance of different pressures in a given area, identify thresholds of cumulative impacts, design measures of impact prevention, or determine to what extent such measures are required.
The analysis further shows that very few peer reviewed investigations of the cumulative effects of various pressure factors have been conducted in the marine environment, and that only a minority of these are relevant for the quality elements used by the Danish environmental authorities in relation to the WFD. The lack of specific analysis of cumulative effects by multiple pressures is likely to be caused by the focus on the excess load of nutrients, especially nitrogen, to Danish coastal marine environments, where it has been considered to constitute the most important pressure. Another explanation is that a large number of other identified pressures, when considered individually, are unlikely to have significant effects on the quality elements at the waterbody level (Petersen et al 2018). Finally, there is generally limited experimental knowledge on interactions between multiple pressure factors on specific quality elements, and a limited amount of data on the importance of the various pressure factors except for nutrient supply. These knowledge gaps, and a lack of empirical analyses of the cumulative effects of the different pressures, are mainly due to a lack of data and process knowledge, and seem not, or only to a very limited extent, to be caused by a lack of methods.
Original language | Danish |
---|
Publisher | DTU Aqua |
---|---|
Number of pages | 44 |
ISBN (Electronic) | 978-87-7481-282-1 |
Publication status | Published - 2020 |
Series | DTU Aqua-rapport |
---|---|
Number | 359-2020 |
ISSN | 1395-8216 |