Impact assessment and risk mitigation

What is Impact Assessment?

There are many types of human actions and activities that may impact individuals or populations of animals in their natural environments. Examples include those that may result in chemical pollution (e.g. an oil spill), those that create physical barriers (e.g. construction), those that create a lot of noise (e.g. pile driving for marine renewables), those that may result in direct mortality (e.g. vessel or turbine blade collision), and those that may lead to climate change. These physical, chemical or biotic changes or inputs to an environment can be referred to as stressors as they have the potential to impact animals and their populations.

Stressors can impact individual animals, for example they might change their behaviour, change where they spend their time, and/or some aspect of their health may be affected. Changes at the individual level that reduce survival or reproduction can ultimately lead to changes at the population level, e.g. a decline in population size. When individuals or groups of animals change where they spend their time then this can lead to population-level shifts in distribution and/or density.

Prospective impact assessments are those that are carried out in advance of a proposed action or activity to predict its likely effects and are increasingly a legal requirement for the environmental regulation of human activities. As well as predicting potential impact in advance of an activity, we may also need to assess whether an impact occurred during and/or after an activity (retrospective assessment). This may take the form of planned monitoring as part of an offshore wind farm construction or an unplanned environmental disaster (for example if changes in population size, health or distribution become apparent).

For all impact assessments (prospective or retrospective), we need to know something about the occurrence of animals in an area (i.e. their distribution and density), their likely exposure to a stressor (e.g. the proportion of individuals within audible range of pile driving and how much time they spend there), the response to exposure (e.g. number of individuals displaced), and the possible consequences for the population. There is active research within CREEM in all four areas: occurrence, exposure, response, consequence.

CREEM research on assessing occurrence, exposure, response and consequence for the purposes of impact assessment.


Modelling spatial distributions of animals allows us to look at trends over time and identify areas of persistence prior to any activity commencing, which gives an indication of space use and the natural variability of this (e.g. through Distance sampling and spatial modelling). These baseline distributions are a key component to the Environmental Impact Statements submitted by industry for e.g., offshore renewables installations. We have developed guidance for the survey design and analysis of baseline and during/post-impact monitoring for energy generation sites, in particular the development of an R software package, MRSea (Marine Renewables Strategic Environmental Assessment), to enable implementation of methods. This package incorporates Distance sampling (via the mrds R package developed in CREEM) and spatial modelling (in the form of spatially adaptive smoothing and generalised additive models) and, although developed with offshore renewables in mind, has many applications. In particular, MRSea can be used with GPS tracking data to estimate home-ranges and overlaps between individuals to determine space/resource use.

As well as predicting the potential impacts of an activity/action, we also need to assess how we might detect these impacts through monitoring surveys conducted during and after an activity. We have developed guidance, methods, and R software (MRSeaPower) for quantifying the power to detect changes in abundance and/or distribution for a given impact scenario and survey design.


The exposure of an individual to a stressor is dependent on its’ overlap in space and time with an activity. One aspect of this is understanding occurrence in space and time, as described above. Other required inputs for assessing exposure relate to the activity itself and are generally provided to us, e.g. operational parameters, sound levels at distances radiating from an activity, and hearing sensitivities of species. We can then combine these inputs with information on response (see below) to predict both individual and population level consequences. An important application of this is for scenario planning, where we use tools such as individual-based simulation models to investigate the consequences of different exposure levels through simulating a range of scenarios and activity parameters. This approach can also be applied retrospectively to investigate scenarios that may have resulted in an observed change in a population.


It is also important to understand how animals may respond to different activities and at what level of exposure. Behavioural response studies (BRS) are used to study how animals respond to different sources of disturbance, and to quantify the relationship between exposure and probability of response. We have developed a suite of quantitative methods for determining whether an animal changed its behaviour in response to exposure or not, and for fitting exposure-response models to BRS data that can be used to predict responsiveness of individuals to different levels of disturbance exposure (MOCHA website). We are also extending these models to allow for exposure to multiple stressors.

As part of the consenting process for offshore renewables, post-construction monitoring surveys must be conducted, and models are fitted to determine whether there was a response to the activity that resulted in a change in the distribution and abundance of animals during an activity and how long the effect lasted into the post-activity period. The MRSea R package described earlier to quantify occurrence may also be used to assess whether there have been significant changes in distribution between the various construction phases and assess the cumulative impact of multiple activities.


As described under Exposure, we can use the information from our understanding of occurrence, exposure and response to build individual-based simulation models to generate predictions of the numbers of individuals that might be affected under different scenarios to aid decision making about e.g., the timing or location of an activity. The models are also used to investigate possible mitigation options if the predicted impact is deemed to be above an acceptable threshold.

Assessment of impact at the individual level can also feed through into models that evaluate the population consequences of disturbance (PCoD models) or population consequences of multiple stressors (PCoMS models). These models can be used to make predictions into the future of the impact of current and planned activities on populations and can thus inform appropriate management and conservation measures. They can also be used retrospectively to help understand the scenarios that may have led to observed changes in a population. For example, we have been involved in retrospectively assessing the impact of environmental disasters, such as the Deep Water Horizon oil spill, on animal health.

What species are these methods used for?

The majority of our research relates to assessing the impacts of human activities and actions in the marine environment on marine mammals and seabirds. Example human activities include the installation and operation of offshore renewable energy devices, shipping, military training activities involving sonar, seismic prospecting for oil and gas, and chemical pollution incidents. Below we provide our focal research areas and some example projects where further information is available via project websites:

  1. Marine mammals and military training activities (sonar)
    1. Mocha – Multi-study Ocean Acoustics Human Effects Analysis (
    2. DenMod – Working group for the advancement of marine species density surface modelling (
  2. Marine renewable energy, marine mammals and seabirds
    1. WoW project – Wildlife and Offshore Wind (Wildlife and Offshore Wind; research and risk assessment framework)
    2. MRSea/MRSeaPower
  3. Population consequences
  4. Individual-based simulation models for predicting impact

Who in CREEM works on these methods?

Please note that people have been listed according to their primary research area, but many are involved in multiple areas.

Marine mammals and military training activities (sonar)

Marine renewable energy, marine mammals and seabirds

Population consequences

Individual-based simulation models for predicting impact

A few relevant publications by CREEM staff

Marine mammals and military training activities (sonar)

Bouchet, P, Harris, CM & Thomas, L 2021, ‘Assessing the role of sampling uncertainty when predicting behavioral responses of tagged cetaceans exposed to naval sonar’, Frontiers in Marine Science, vol. 8, 674554.

Durbach, IN, Harris, CM, Martin, C, Helble, TA, Henderson, EE, Ierley, G, Thomas, L & Martin, SW 2021, ‘Changes in the movement and calling behavior of minke whales (Balaenoptera acutorostrata) in response to navy training’, Frontiers in Marine Science, vol. 8, 660122.

Harris, CM, Thomas, L, Falcone, E, Hildebrand, J, Houser, D, Kvadsheim, P, Lam, F-PA, Miller, P, Moretti, DJ, Read, A, Slabbekoorn, H, Southall, BL, Tyack, PL, Wartzok, D & Janik, VM 2018, ‘Marine mammals and sonar: dose-response studies, the risk-disturbance hypothesis and the role of exposure context’, Journal of Applied Ecology, vol. 55, no. 1, pp. 396-404.

Jacobson, EK, Henderson, EE, Miller, DL, Oedekoven, CS, Moretti, D & Thomas, L 2022, ‘Quantifying the response of Blainville’s beaked whales to U.S. naval sonar exercises in Hawaii’, Marine Mammal Science, vol. 38, no. 4, pp. 1549-1565.

Marine renewable energy, marine mammals and seabirds

Mackenzie, M. L, L. A. S. Scott-Hayward, C. S., Oedekoven, H., Skov, E., Humphreys, and E. Rexstad. 2013. “Statistical Modelling of Seabird and Cetacean Data: Guidance Document. University of St. Andrews Contract for Marine Scotland; SB9 (CR/2012/05).” University of St Andrews.

Russell, DJF, Hastie, GD, Thompson, D, Janik, VM, Hammond, PS, Scott-Hayward, LAS, Matthiopoulos, J, Jones, EL & McConnell, BJ 2016, ‘Avoidance of wind farms by harbour seals is limited to pile driving activities’, Journal of Applied Ecology, vol. 53, no. 6, pp. 1642-1652.

Population consequences

Pirotta, E, Thomas, L, Costa, D, Hall, AJ, Harris, CM, Harwood, J, Kraus, S, Miller, PJ, Moore, M, Photopoulou, T, Rolland, R, Schwacke, L, Simmons, S, Southall, B & Tyack, PL 2022, ‘Understanding the combined effects of multiple stressors: a new perspective on a longstanding challenge’, Science of the Total Environment, vol. 821, 153322.

Pirotta, E., R.S. Schick, P.K. Hamilton, C.M. Harris, J. Hewitt, A.R. Knowlton, S.D. Kraus, E. Meyer-Gutbrod, M.J. Moore, H.M. Pettis, T. Photopoulou, R.M. Rolland, P.L. Tyack and L. Thomas. (In press) Estimating the effects of stressors on the health, survival, and reproduction of a critically endangered, long-lived species. Oikos.

Schwacke, LH, Marques, TA, Thomas, L, Booth, C, Balmer, BC, Barratclough, A, Colegrove, K, De Guise, S, Garrison, LP, Gomez, FM, Morey, JS, Mullin, KD, Quigley, BM, Rosel, P, Rowles, TK, Takeshita, R, Townsend, FI, Speakman, TR, Wells, RS, Zolman, ES & Smith, CR 2022, ‘Modeling population effects of the Deepwater Horizon oil spill on a long-lived species’, Conservation Biology, vol. 36, no. 4, e13878.

Tyack, PL, Thomas, L, Costa, D, Hall, AJ, Harris, CM, Harwood, J, Krauss, S, Miller, PJ, Moore, M, Photopoulou, T, Pirotta, E, Rolland, R, Schwacke, L, Simmons, S & Southall, B 2022, ‘Managing the effects of multiple stressors on wildlife populations in their ecosystems: developing a cumulative risk approach’, Proceedings of the Royal Society of London Series B: Biological Sciences, vol. 289, no. 1987, 20222058.

Wilson, LJ, Harwood, J, Booth, CG, Joy, R & Harris, CM 2020, ‘A decision framework to identify populations that are most vulnerable to the population level effects of disturbance’, Conservation Science and Practice, vol. 2, no. 2, e149.

Individual-based simulation models for predicting impact

Donovan, CR, Harris, CM, Milazzo, L, Harwood, J, Marshall, L & Williams, R 2017, ‘A simulation approach to assessing environmental risk of sound exposure to marine mammals’, Ecology and Evolution, vol. 7, no. 7, pp. 2101-2111.

Joy, R, Schick, RS, Dowd, M, Margolina, T, Joseph, JE & Thomas, L 2022, ‘A fine-scale marine mammal movement model for assessing long-term aggregate noise exposure’, Ecological Modelling, vol. 464, 109798.

Some relevant links