The Working Group on Nephrops Surveys (WGNEPS) is the international coordination group for Nephrops underwater television and trawl surveys within ICES. This report summarizes the national contributions on the results of the surveys conducted in 2019 together with time series covering all survey years, problems encountered, data quality checks and technological improve-ments as well as the planned for survey activities for 2020. In total, 19 surveys covering 25 func-tional units (FU’s) in the ICES area and 1 geographical subarea (GSA) in the Adriatic Sea were discussed and further improvements in respect to survey design and data analysis, standardiza-tion and the use of most recent technology were reviewed.
A new survey summary template by FU/GSA has been developed and adopted for future re-ports, which shall allow the data end users to extract the most relevant information on the survey results in a more easy way.
Necessary actions and reviewer comments were addressed on the draft version of the Series of ICES Survey Protocols (SISP). Similarly, the working group reviewed the specifications for a Nephrops underwater TV database to be established at the ICES data centre and agreed on fur-ther action on this issue.
First results from field studies on behaviour aspects of burrow emergence using bottom cages monitored by an automated camera system and on short-range migration using acoustic tracking are now available.
Comparison of standard definition (SD) and high definition (HD) indicates the change to HD system mounted with a different camera angle may affect the detection rate and may thus require a revision of bias correction factors. New image reviewing software allows an easier way of an-notation of burrows than previous mosaicking methods, which has further advantages for inter-preting the results from different counters and for providing quality assured material for deep learning methods. The WG members agreed to collect information on burrow diameter size us-ing HD images and burrow annotation or mosaicking software because a change in the burrow size distribution may indicate recruitment events and the size of the burrow has an effect on bias correction factors in general.
Automatic burrow detection based on deep learning methods applied to a test data set with an-notated burrow counts from a HD camera system showed promising results. The WG members were encouraged to provide more material with annotated burrow counts for further develop-ment of machine learning tools.