Fig 1.
Schematic illustrations of (a) the vertically divided fishing gear showing the upper and lower codends and the sizes of their square meshes in addition to the two frames used to keep the lower codend open, and (b) the standard codend showing the size of the diamond meshes. Photos of (c) the upper and lower codends just before the catch of the upper codend is emptied into the holding bin of the vessel, and (d) the standard codend as the catch is hauled up on deck.
Fig 2.
An overview of the data collection and assessments during the different steps of the value chain.
Table 1.
Description of the assessed damages found in fish and the scoring scheme according to the severity of the damage.
Table 2.
Operational conditions and catch weights.
Fig 3.
Photo of the test codend showing the frame with vertical bars mounted at the entrance of the lower codend and the geometry at the entrance of the upper codend during fishery.
Table 3.
Mean scores of the quality parameters for four fish species during three steps in the value chain.
Statistical comparisons were made using a proportional odds model, and significance level was calculated using the Wald test.
Fig 4.
Scale loss in cod (left), plaice (middle), and witch flounder (right).
Fig 5.
Odds with confidence intervals for obtaining better quality (lower scores) compared with the standard codend for scale loss (○) for newly caught fish from the (a) upper and (b) lower codends of the test gear and discolorations (◇) for newly caught fish from the (c) upper and (d) lower codends of the test gear.
The test and standard codends were equal for odds = 1 (dotted line). The results were statistically significant at *p < 0.05 and **p < 0.01 when the confidence intervals did not include 1.
Fig 6.
The odds for less scale loss cod (○), saithe (◇), plaice (□), and witch flounder (△) with increasing proportion of Norway lobster relative to fish in the catch of the lower (a) and standard codends (b).
When the odds were < 1, the scale loss was higher than that of the starting point (---), whereas for odds > 1 the scale loss was lower.
Fig 7.
Discolourations: blood spots in plaice (left); bruises on the head of cod (middle); bruises along the body of witch flounder (right).
Fig 8.
Odds with confidence intervals for obtaining lower quality scores compared with the standard codend for the quality parameters of landed fish caught in the (a) upper and (b) lower codends of the test codend.
The test and standard codends were equal for odds = 1 (dotted line). The results were statistically significant at *p < 0.05 and **p < 0.01 when the confidence intervals did not include 1.
Fig 9.
Blood spots (left) and severe degree of bruising (right) in fillets from plaice.
Fig 10.
Odds with confidence intervals for obtaining lower quality scores compared with the standard codend for the quality parameters of fillets from fish caught in the (a) upper and (b) lower codends of the vertically divided fishing gear.
The test and standard codends were equal for odds = 1 (dotted line). The results were statistically significant at *p < 0.05 when the confidence intervals did not include 1.
Fig 11.
Gaping in cod (left), saithe (middle), and witch flounder (right) fillets.
Table 4.
Conditions for different studies for which the proportion of individuals with bruises was compared.