SEA predictions are often affected by high levels of uncertainty, for instance about how the strategic action will be translated into actions on the ground, the likely future state of the environment, future technologies, and the effects of other strategic actions.
For near-term future activities and those for which good information is available, reasonably certain, detailed impact prediction should be possible. As one gets into the more distant future, only a broad-brush picture may be possible.
In many cases, a 'good enough' result for decision-making can be achieved despite uncertainties, so that uncertainties do not need to be dealt with specifically. For instance, future social conditions may clearly be better than current conditions; one alternative may be clearly better than another one; the economic benefits of a strategic action may clearly outweigh any environmental costs.
Where this is not the case and uncertainty needs to be dealt with specifically, the techniques shown in the table below can be used to address uncertainty. Uncertainty can also be documented in an SEA report by using ranges rather than precise figures (i.e. 'between 1 and 3 accidents per year') or probability of occurrence of impacts.
||UK South East Regional Planning Guidance, policy on provision of new land for housing: "It must be assumed that current commitments [of land for housing]... have been made with a view to minimising the basic environmental impacts"
||The SEA of the Dutch waste management plan 1992-2002 was based on two scenarios for future waste production: the 'policy scenario' assumed that national objectives regarding waste prevention, reduction etc. would be fully achieved. The 'headwind scenario' assumed that they would not be fully achieved, and that therefore more waste would have to be dealt with.
|Sensitivity analysis: tests the robustness of predictions under different assumptions/scenarios
||Importation of European honeybees to Canada: "the Varroa mite is a parasite that can infest and destroy entire honeybee colonies. [It could be imported with the European bees so strict protocols would be needed]... There is currently a mite infestation problem in many parts of Canada... If the current populations of mites continue infesting Canadian bee colonies and cannot be controlled, the issue of importing mite-free honeybees may well become a non-issue. If Canada is infested, the strict protocols may not be necessary."
|Using worst case scenarios based on the precautionary principle
||Leasing of the strategic petroleum reserve at St. James Terminal, Louisiana: "There is an estimated probability of 0.1 spills per vessel per year under the reasonable use scenario, and 3.3 spills per year under the maximum use scenario, with each spill having an estimated average volume of 22 bbl."
|Showing points of view
||A questionnaire about a proposed fisheries management programme in the US was sent to a range of affected stakeholders. The commercial fishery expressed concern that "incidental catch needs to be monitored and documented--there is a concern about incidental catch of salmon and steelhead"
||"Good monitoring and enforcement programs should deal with incidental catch of game and salmon species. BPA is proposing to comply with existing sport fishery regulations and monitor incidental catch for these species."
"Future proofing" means making the strategic action able to cope with possible future changes and uncertainties. It involves imagining future scenarios for e.g. climate change, petrol prices, an older population, more (or less) immigration etc; and planning so as to minimise the risks of negative impacts under these scenarios.
A small-scale example is building new houses with wheelchair access: it may cost slightly more and the residents may not need it now, but it could avoid costly retrofits later. On a larger scale, plans for "managed realignment" of coastlines that are likely to be eroded or flooded in the future as a result of climate change is a more "future proof" way of planning than allowing development in those areas that may later need to be moved.
The impact of a strategic action will often depend on how it is implemented: for instance, new housing may improve conditions for wildlife if it replaces an agricultural 'wildlife desert' with wildlife-friendly gardens, or it could have much paving and minimal plantings and so have no positive effect. The role of the SEA is to highlight this, and to suggest how things can be improved at the implementation stage, e.g.:
"all existing hedges will be maintained, a wildlife corridor will be provided through the estate, gardens will be planted with..."
And finally, remember that a broad-brush picture of impacts is much more useful than no picture at all!