Spatial distribution of South African demersal trawl activity, 2005-2018

The objective of these data layers was to map the South African demersal trawl intensity. All commercial trawl logbook records between 2005 and 2018 were obtained from DFFE. As demersal trawling was the focus, data were restricted to demersal trawls. Following this selection, the 2005-2018 dataset consisted of 673 037 trawl events. A sequence of data cleaning and correction followed: Records lacking spatial coordinates were removed and obvious errors such as negative longitudes or positive latitudes were corrected. Filters were then applied to remove spurious records that were recorded to 20’ grid cell corners (lacked more precise coordinates), were of an unrealistic length (either zero or > 30 NM), or had coordinates outside South Africa’s EEZ. Thereafter corrections were applied to coordinates of records affected by systematic biases, including: records that had been rounded down to 59’ coordinates were re-distributed between 59’ and 60’; coordinates that had been rounded up to integer degrees were re-distributed among 50’ to 59’ coordinates; addition of a 0-59 second ‘jitter’ to records that were rounded to minute coordinates. Following these corrections, two final cleaning steps involved 1) filtering by density in that start and end positions that occurred at < 4 records per km2 over the entire 2005-2018 period were removed; and 2) removal of records when their measured distance between start and end points far exceeded their expected maximum distance based on duration and a maximum speed threshold. Following the data preparation steps above, the swept area ratio (Amoroso et al. 2018) was then calculated at a 1 km2 resolution, both as an average for the 2005-2018 period and for each year separately. Lastly, the swept area ratio was up-scaled to account for the effort (records) removed during the data cleaning steps. The archive consists of 16 files, namely the 2005-2018 average swept area ratio (SAR; ‘SAR_average_2005-2018_upscaled_filterd4_1KM.tif’), a multi-banded raster, where bands 1-14 represent the years 2005-2018 respectively (‘SAR_2005to2018_upscaled_filterd4_1KM.tif’), and the 14 annual files (SAR_upscaled_filterd4_1KM_yyyy.tif, where ‘yyyy’ represents the relevant year). Further details of cleaning and correction procedures are captured in a report that can be requested from the lead author.