Additional fish species occurrence records for Uganda

Study Extent

The data that were mobilized to develop this dataset were retrieved from scattered electronic datasets, stored as Microsoft Excel files at NaFIRRI. The data accumulated from fish biodiversity surveys conducted by scientists between 1971 and 2015. The aquatic ecosystems covered included major lakes, including Lakes Victoria, Albert, Kyoga, Edward and George (including Kazinga channel). The data sets also covered other small water bodies and the River Nile. Some of the water bodies were sampled on multiple events during the period.

Sampling

During the surveys, fish was captured using fishing gears including gillnets, hooks, and beach seines on specified sites (localities) in the water bodies. The nets were usually set in locations at specified distance from the shoreline including near shore, mid shore or offshore stations. The distances differ by water body but near shore sites are normally 50m from the shoreline, mid shore 100-500m and off shore sites ≥500m. The gillnets included multiple sets of mesh sizes 1 to 8 inches. The dimensions varied but typical gillnets were 90 m long and 26-52 meshes deep. In each set, the smaller nets of (1 to 5.5 inch) were graded at half-inch intervals while the larger meshed nets (6 to 8 inch) were at intervals of one-inch. The offshore nets were set at least three kilometers from the shoreline. The nets were normally set in the evening and retrieved the following morning. Beach seines were not common in the surveys but where used, were in areas suitable for seining such as sandy shorelines close to river mouths. Habitats including the shoreline, open waters, rocky, sandy, vegetated areas, river mouths, laggons, bays etc. were targeted. A description of a typical fish sampling procedure used in the surveys is provided in Wandera & Balirwa, 2010. Once captured, fish were identified to the lowest taxonomic level where possible using taxonomic keys (Greenwood, 1966; Greenwood, 1981).

Quality Control

Because most of the mobilized/rescued data have no associated physical specimens deposited in the museum, occurrence records were verified by fisheries experts within NaFIRRI based on the expert’s species knowledge, photographs (if available), and known species list of locations. Records that were outside their known range and observations, including the species’ possible extent of occurrence given in FishBase, and for which photographs and actual specimens could not be traced, were left out of this dataset to guard against suspicious species that could have been a result of misidentification. A thorough taxonomic standardization was done by relating the given taxon names within the original dataset with those of Fishbase (Froese & Paully, 2018). Fishbase is one of the most authoritative and comprehensive databases with list of names of marine and freshwater fishes, including information on synonymy, and was therefore considered the best choice for taxonomic standard quality check for the species names. By cross-referencing the given taxon names to the widely accepted FishBase taxonomic standard, it was possible to rule out spelling variations and synonyms, but more importantly, invalid names that were possibly regarded as valid at the first time of identification and data capture. All the location coordinates were verified by plotting them first in Google Earth to ensure that they agree with the mentioned sampled site name.

Method steps

1. Selection of datasets: We selected data sets stored on computers previously held by scientists who led fish biodiversity surveys. On the computers, the data carried separate folders for each of the lakes or group of related lakes with several files. We selected electronic Microsoft Excel files with the original data as the basis for generating the occurrences. Some occurrences were digitized from published literature and for these, applicable references are provided.
2. Developing the datasets into Darwin core template for occurrence records: The information in the datasets was used to develop the fish species occurrences in the Darwin core format using an appropriate template. The datasets had basic information useful for developing occurrences such as water body, name of fish species, collection event date, locality, and GPS coordinates (especially for records after 2001). These were entered in appropriate fields within the occurrence template. Data on individual count was obtained by summing the number of individuals recorded for each record (species) for each event date in a given locality.
3. Geo-referencing and coordinate conversions: The data sets, especially before 2001, did not have GPS coordinates but had information on the actual locations or sites where the fish were captured. These were geo-referenced where possible by verifying the locations and obtaining their GPS coordinates in Google Earth. The verification of the locations benefited from local knowledge and from field staff who participated in the surveys and are still at NaFIRRI. Where GPS coordinates were available, they were used. Coordinates that were not in decimal format were converted using online coordinate conversion tools including canadensys (http://data.canadensys.net/tools/coordinates) and PGC Coordinate Converter (https://www.pgc.umn.edu/apps/convert/) of the University of Minnesota.
4. Confirming fish species names: Species names were confirmed using FishBase (Froese & Paully, 2018). FishBase is the most widely used reference guide for fish species. For the species that were allocated to the genus Barbus in the original survey data, the currently accepted nomenclature, Enteromius for the small diploid species and Labeobarbus for the large hexaploid species is followed in this dataset (Van Ginneken et al., 2017; Vreven et al. 2016). The two genera are both given for records where the highest taxonimic level allocated at the time of observation was Barbus. In addition, all haplochromine cichlids are allocated to genus Haplochromis which is accepted in Fishbase. However, where the original record name is valid in the Catalog of Fishes (Eschmeyer et al. 2016), it is also mentioned in the appropriate field (with the name at first identification).
5. Data cleaning: The GPS coordinates were checked for errors by visualizing in Google Earth or QGIS. This step enabled us to correct errors in GPS coordinates.

The dataset covers only fish species in Uganda. It covers classes Actinopterygii and Sarcopterygii that are the only two taxonomic classes of fish in Uganda.

1. Actinopterygii
   common name: Ray-finned fishes rank: class
2. Sarcopterygii
   common name: Lobe-finned fish rank: class

The dataset covers only Uganda. Uganda is a beautiful country in East Africa