Can benthic quality assessment be impaired by uncertain species sensitivities? Romualda Chuševė a, ⁎, Darius Daunys b a Marine Technology and Nature Science Faculty, Klaipėda University, H. Manto 84, LT 92294 Klaipėda, Lithuania b Marine Science and Technology Centre, Klaipėda University, H. Manto 84, LT 92294 Klaipėda, Lithuania abstract article info Article history: Received 20 September 2016 Received in revised form 5 January 2017 Accepted 7 January 2017 Available online xxxx This study tested robustness of mathematically defined species sensitivity with manipulation of disturbance gra- dient coverage by datasets on two different pressures and two depth zones of the central Baltic Sea. The results indicate large differences of sensitivity values for the same species when depth range changed. After addition of samples from impacted sites to the control datasets, estimated sensitivity values decreased for half of analysed species. Sensitivity of tolerant species appeared to be highly dependent on the coverage of disturbance gradient by the dataset, while sensitive species were highly robust to this effect. Although pressure type and coverage of disturbance gradient by the dataset was important, sensitivity changes were primarily linked to the altered en- vironmental conditions. It is suggested that sensitivity values based on natural variability of pre-selected commu- nities will better integrate the role of important environmental factors and increase robustness of discrimination between disturbed and undisturbed sites. © 2017 Published by Elsevier Ltd. Keywords: Macrozoobenthos BQI Species sensitivity Anthropogenic impact Baltic Sea 1. Introduction Healthy marine environments with diverse biological communities and good environmental status provide valuable ecosystem services and support a wide range of human activities (Elliott, 2011). Marine ecosystems are, however, exposed to an increasing number of distur- bances such as eutrophication, over-fishing, pollution, etc. (Lotze et al., 2006). In the face of these multiple pressures, an accurate assessment of environmental status is a prerequisite for establishing environmental targets and selecting specific management measure. Stationary and long-living benthic communities can amalgamate medium-to-long term information about environmental conditions and the importance of anthropogenic pressures, therefore structural pa- rameters of bottom macrofauna communities were synthesized in a number of multimetric indices to be used for environmental status as- sessment (Borja et al., 2013). Following guidance for recent assessments of environmental status, these indices integrate sensitivity, richness and abundance of macrofauna species (Water Framework Directive; 2000/ 60/EC; Marine Strategy Framework Directive; 2008/56/EC; HELCOM, 2013). While richness and abundance are directly assessed from benthic samples, defining species sensitivity is less straightforward. Generally, sensitivity is described as a product of a likelihood of damage due to pressure (termed intolerance or resistance) and the rate of, or time taken for recovery once the pressure has been removed (termed recoverability or resilience) (Laffoley et al., 2000; OSPAR, 2008). Typically, very sensitive species react fast to the impact of natural stressors or human activity (e.g. low resistance) and recover only after a prolonged period, if at all (e.g. low recoverability). Species grouping into sensitivity classes based on expert valuation is used in several indices (e.g. AMBI, Borja et al., 2000; BENTIX; Simboura and Zenetos, 2002; M-AMBI, Muxika et al., 2007, Leonardsson et al., 2009). Typically this approach enables one to take into account different sensitivity meanings (e.g. resilience, resistance, recoverability and vul- nerability) (Tyler-Walters et al., 2001), which otherwise are difficult to quantify using numerical algorithms. Additionally, expert-based sen- sitivity values may have a fixed range of variability, predefined number of sensitivity classes and fixed distance between sensitivity values for organisms belonging to different sensitivity classes. These features make species sensitivity estimates well-structured and simple to inter- pret in a context of overall status assessment. In contrast to expert opinion, standardized mathematical algorithm for assessing species sensitivity has an obvious advantage for screening sensitivity performance under different conditions. It also makes sensi- tivity values comparable between different areas and datasets leaving less freedom for low certainty assumptions. So far, the only standard- ized species sensitivity assessment algorithm was proposed by Rosenberg et al. (2004) and recently revised by Leonardsson et al. (2015). The overall concept relies on the assumption that lower sensi- tivity species will be more abundant and will occur more frequently in disturbed sites, characterized by lower species richness, while species of higher sensitivity will attain higher abundance at higher species rich- ness and less disturbed sites. This implies that sensitivity value is depen- dent on the diversity (i.e. disturbance) range covered by the dataset and may significantly influence the overall assessment. However, since its invention the only studies of exploring variability and reliability of Marine Pollution Bulletin xxx (2017) xxx–xxx ⁎ Corresponding author. E-mail address: romualda.chuseve@jmtc.ku.lt (R. Chuševė). MPB-08319; No of Pages 8 http://dx.doi.org/10.1016/j.marpolbul.2017.01.019 0025-326X/© 2017 Published by Elsevier Ltd. Contents lists available at ScienceDirect Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul Please cite this article as: Chuševė, R., Daunys, D., Can benthic quality assessment be impaired by uncertain species sensitivities?, Marine Pollution Bulletin (2017), http://dx.doi.org/10.1016/j.marpolbul.2017.01.019