Table 2.

Recommendations related to data quality and methodologies to improve the robustness of functional analyses based on global fisheries datasets.

Stage	Recommendation
Functional analysis	1. Data: Highlight patterns in trait coverage, such as variation in space. Enrich trait information available from FishBase where possible with other databases such as the Ocean Biogeographic Information System (http://www.iobis.org/) and the Global Biodiversity Information Facility (http://www.gbif.org/).
	2. Methods: Careful choice of functional metric. In EEZs where there is high proportion of unreported data and low confidence in catch reconstruction, functional dispersion may be less affected by missing data than functional evenness.
	3. Methods: Comparison of patterns in functional diversity using fisheries-dependent and independent data to understand the links between exploitation patterns and ecosystem function.
Certainty	4. Data: Increased comprehensiveness of certainty data in catch database.
	5. Methods: Assessment of differences in functional diversity estimates among low and high certainty data.
Reconstruction of missing data	6. Data: For each EEZ more detail is needed in background documents relating to each step of the reconstruction process, such as data sources used and assumptions made.
	7. Data: Need to enrich data with fleet attributes such as gear type so can assess selectivity effects on functional diversity.
	8. Methods: Simulation testing exploring the effect of different reconstruction assumptions and approaches versus gear selectivity effects on functional diversity estimates.
Reporting	9. Data: Improved taxonomic resolution of FAO catch reporting is needed.
	10. Methods: Assessment of the impact of temporal changes in taxonomic resolution in FAO dataset on functional diversity estimates.

Shading corresponds to Fig. 1a, black shading represents steps beyond those shown in Fig. 1a.