Report of the Workshop on guidance on how pressure maps of fishing intensity contribute to an assessment of the state of seabed habitats (WKFBI)

The Workshop on guidance on how pressure maps of fishing intensity contribute to an assessment of the state of seabed habitats (WKFBI), chaired by Adriaan Rijnsdorp (the Netherlands), met at ICES Headquarters on 31 May1 June 2016. The Workshop was attended by 28 participants from 10 countries, including representatives from various ICES Working Groups, two representatives from the fishing industry and one from DG Environment. The task of the workshop was to evaluate the information that is re-quired to assess the state of seabed habitats (high resolution data on the trawling intensity by métier, maps of seabed habitats, information on the sensitivity of seabed habitats for bottom-trawling pressure) and prepare a guidance document on how fish-ing pressure can be used to develop indicators of the state of seabed habitats.
The workshop was prepared by a group of experts and chairs from WGDEC Working Group on Deep-water Ecology, BEWG Benthos Ecology Working Group, WGMHM Working Group on Marine Habitat Mapping, WGSFD Working Group on Spatial Fisheries Data, and supported by an ICES professional officer to organize the required building blocks and carry out the subsequent impact analysis required for WKFBI. In addition to evaluating material prepared by ICES Working Groups, the WKFBI compared a selection of similar approaches developed within European-funded projects (BENTHIS) and regional seas conventions (BH3).
Maps of trawling intensities (surface and subsurface abrasion), based on VMS and log-book data taking account of the differences in the footprint across métiers, are pre-sented for surface and subsurface abrasion from all bottom-trawl métiers, as well as for the main fishing gears separately (otter trawl, demersal seine, beam trawl, dredge). Maps cover the European seas ranging from the Iberian peninsula in the south to the Norwegian Sea in the north and the Baltic Sea in the east at a resolution of 0.5o by 0.5o (c-square). A unified habitat map for the entire study area was generated based on the 2016 interim EMODNET maps.
Habitat sensitivity was estimated using the categorical approach developed in the UK (MB0102). Sensitivity depends on the resistance of the receptor (species or habitat fea-ture) and the ability of the receptor to recover (resilience). For each habitat resistance and resilience was estimated of a selection of key and characterizing species based on scientific evidence by experts. Because the sensitivity scoring used the MB0102 benchmark of medium physical pressure which is not related to a specific trawling intensity, trawling intensity classes were arbitrarily set. The sensitivity scoring for the shelf hab-itats (0-200m) was carried out by BEWG, the scoring for the deep-sea habitats was done by WGDEC. Because of the lack of data on deep-sea habitats, in particular the occurrence of biogenic habitats, WGDEC adopted a precautionary approach and classified all deep-sea habitats as highly sensitive. For the shelf habitats, habitat sensitivity in-creased from low to medium for surface abrasion to medium to high for subsurface abrasion. Pilot maps with the surface and subsurface abrasion were generated based on the categorical approach and compared to maps of trawling impact estimated using the mechaniztic approach developed by FP7-project BENTHIS providing an impact score on a continuous scale.
All results of the different analyses should be considered preliminary. The purpose of the impact analysis exercise for WKFBI was to go through all stages of the process in order to detect potential problems and evaluate and compare the strength and weak-nesses of the various approaches. One inherent challenge with the expert judgement approach considered is that it is difficult to interpret the differences in trawling impact in quantitative terms as both sensitivity and trawling intensity are categorical. Methods such as developed in BENTHIS provide a quantitative estimate of the impact on a continuous scale. These methods, however, are still under development and the uncer-tainty of the impact estimates have not been determined nor has the sensitivity of the methods been investigated. The quantitative methods do provide a promising ap-proach to derive the scientific basis to assess the impact of trawling on the state of the seabed. The first results presented in this report are already useful to relate the class-boundaries used in the categorical method with the estimated impact based on the ap-proaches relating longevity and the population dynamics to the respective trawling intensity.
All methods explored in this report provide estimates of the trawling impact on the benthic community at the level of the grid cell. In addition to pressure indicators such as the trawling footprint (surface area or proportion of a management unit or habitat) trawled, or the indicator for the degree of aggregation (such as the proportion of the footprint where 90% of the total fishing effort occurs), the estimates of trawling im-pact can be aggregated to a metric that reflects the trawling impact or seabed integrity at the level of the habitat or management area.
By applying a methodology that develops matrices of habitat sensitivity in relation with trawling pressure allows a consistent assessment of the relative impact of bot-tom trawling across different habitats taking account of the estimated trawling intensities of the surface and subsurface seabed. Applying the same classification criteria over time means that changes in trawling impact can be assessed. If a benchmark has been set, for instance in terms of the surface area of a particular habitat or management area that is impacted less than a predefined level, changes in trawling impact can be compared to the benchmark (both in space and time). Scientific effort is needed to further investigate possible benchmark and threshold settings for an appropriate assessment.
An inherent challenge is how to deal with impact across consecutive years. If an area is trawled its subsequent sensitivity will change, i.e. if a previously disturbed site is trawled again it may be impacted less (vulnerable species/habitats have not yet recovered) and can therefore be viewed as more resilient to additional trawling disturbance. Similarly a challenge will be to interpret the implications of the heterogeneity in trawl-ing and its effect on habitat fragmentation on the recovery, which will depend on the amount of undisturbed communities in neighbouring areas acting as sources of new recruitment to disturbed areas.
In the context of MSFD purposes there is first a need to assess whether there is impact (from the pressure, Article 8 assessment). Only later, following a decision to reduce that impact, will there be a need to consider possible management options such as intensity of trawling, productivity of the area and habitat recovery times.