Final Report Summary - SFHABILF (Suction feeding hydrodynamics and biomechanics in larval fishes)
Objectives: The focus of this project was to investigate the hydrodynamic basis of prey capture in larval fishes and its consequences to feeding performance. In particular, the study focused on the following questions: 1) Does transition into higher Reynolds numbers during larval ontogeny affect the temporal and spatial patterns of flows external to the larva’s mouth? 2) Does transition into higher Reynolds numbers affect feeding performance? 3) Does transition into higher Reynolds numbers facilitate prey capture? A combination of state-of-the-art methods of flow visualization (particle image velocimetry and computational fluid dynamics), experiments with live fishes, and a range of modeling approaches were used to address these goals and provide insights on suction feeding in fish larvae.
Results:
1) Reynolds numbers effects on suction flows pattern: A computational fluid dynamics (CFD) model of the mouth was constructed, modeling mouth expansion from the size of first-feeding larvae to adult fishes. This model revealed that variation in mouth size at that range has strong effects on suction flows, specifically the spatio-temporal flow patterns in front of the mouth. Peak flow speed and Reynolds numbers increased with increasing mouth size. In adult fish, flow decays rapidly outside the mouth, and suction flows have a negligible effect on particles movement at a distance of ~2 mouth widths. However in larval fish flow decayed much slower, and significant flows were observed at a distance of ~4 mouth widths. Moreover, the radial symmetry that characterizes suction flows in adult fishes dissipated as mouth length decreased. This difference indicates that viscous forces in low Reynolds number flows modify the spatial distribution flow speed in front of the mouth. Consequently, simulated predator-prey encounters showed that these ontogenetic changes in flow patterns have a strong effect on prey capture performance. Larval fish could capture inert prey from a greater distance compared to adults. If prey attempted to escape, however, larval fish performed poorly: simulations inferred capture success in only weakly escaping prey immediately in front of the mouth. These ontogenetic changes in Reynolds number, suction-induced flow field, and feeding performance thus explain a widespread ontogenetic diet shift from passive prey at early life stages to evasive prey as larvae mature. These results are summarized in publication [1].
Subsequent work with a more elaborate CFD model that includes the gills indicate that, for small mouths operating in low Reynolds numbers, flow reversal is likely to occur during mouth closure. Such flow reversals ensue from viscous effects of the water that exit the mouth through the gills, and may result in prey entering and leaving the mouth before its closure. This study is currently underway; results for 5 larval sizes have already been obtained and are currently being analysed. In addition, visualizations of suction flows with live larvae are conducted. Developing the system to this project was slow, however this this study is at its first stages.
2) Reynolds numbers effects on feeding performance: CFD simulations (above) indicate that the hydrodynamic regime has potential implications on their ability to capture prey. To verify these expectations, the feeding rate of Sparus aurata (gilt head sea bream) larvae ranging from first feeding until metamorphosis (8-23 Days Post Hatching [DPH]) was quantified in a series of laboratory experiments. The feeding rate of larvae increased by >10 fold from first feeding (3 mm length) to metamorphosis (9 mm). Hi-speed videos of feeding larvae revealed that the increase in feeding rates did not merely reflect increasing motivation or metabolic demands with age. Instead, we observed a sharp increase in prey capture success from <5% for 8 dph larvae to >60% two weeks later. These results are summarized in publication [2]
Subsequent work generated a dataset of >700 feeding events of 8-33 DPH larvae used to elucidate the mechanism of failure in prey capture throughout ontogeny. Failed strikes resulted from insufficient force exerted on the prey (prey not engulfed during mouth opening), flow reversals (prey entering the mouth and leaving during mouth closure) and active spitting of prey items (after they were already engulfed). All these mechanisms scaled with larval size/age to reduce overall feeding efficiency in small sizes. To mechanistically separate the effects of hydrodynamics (size) from other ontogenetic changes (vision, morphology, behavior etc) we used dynamic scaling experiments in which older larvae were filmed foraging in higher-viscosity solutions. These experiments recapitulated age-associated patterns in all three identified mechanisms, indicating that larval size (rather than age) mechanistically impedes suction feeding performance. A manuscript summarizing these results was submitted to ICB on Jan 30th 2015
3) Reynolds numbers effects on prey capture: Small larvae might be less effective feeders, however larvae (and prey) in small Re are also known to swim slower, and this effect can reduce encounter rate with the prey. We used observed swimming speeds of larvae and prey in all dynamic scaling experiments to construct an encounter rate model. This model indicated that the decay in feeding rate due to viscosity effects on swimming speeds did not explain the marked decrease in experimentally observed feeding rates at high viscosities. Slower swimming could account only for ~1/4 of the observed decline in feeding rates. Subsequent high-speed video analysis of feeding strikes in a dynamic scaling setup showed that successful strikes are characterized by fast mouth opening and strike initiation distance. Under high viscosity, mouth-opening speeds declined, and successful strikes were characterized by shorter strike initiation. These results are also summarized in publication [2].
Results:
1) Reynolds numbers effects on suction flows pattern: A computational fluid dynamics (CFD) model of the mouth was constructed, modeling mouth expansion from the size of first-feeding larvae to adult fishes. This model revealed that variation in mouth size at that range has strong effects on suction flows, specifically the spatio-temporal flow patterns in front of the mouth. Peak flow speed and Reynolds numbers increased with increasing mouth size. In adult fish, flow decays rapidly outside the mouth, and suction flows have a negligible effect on particles movement at a distance of ~2 mouth widths. However in larval fish flow decayed much slower, and significant flows were observed at a distance of ~4 mouth widths. Moreover, the radial symmetry that characterizes suction flows in adult fishes dissipated as mouth length decreased. This difference indicates that viscous forces in low Reynolds number flows modify the spatial distribution flow speed in front of the mouth. Consequently, simulated predator-prey encounters showed that these ontogenetic changes in flow patterns have a strong effect on prey capture performance. Larval fish could capture inert prey from a greater distance compared to adults. If prey attempted to escape, however, larval fish performed poorly: simulations inferred capture success in only weakly escaping prey immediately in front of the mouth. These ontogenetic changes in Reynolds number, suction-induced flow field, and feeding performance thus explain a widespread ontogenetic diet shift from passive prey at early life stages to evasive prey as larvae mature. These results are summarized in publication [1].
Subsequent work with a more elaborate CFD model that includes the gills indicate that, for small mouths operating in low Reynolds numbers, flow reversal is likely to occur during mouth closure. Such flow reversals ensue from viscous effects of the water that exit the mouth through the gills, and may result in prey entering and leaving the mouth before its closure. This study is currently underway; results for 5 larval sizes have already been obtained and are currently being analysed. In addition, visualizations of suction flows with live larvae are conducted. Developing the system to this project was slow, however this this study is at its first stages.
2) Reynolds numbers effects on feeding performance: CFD simulations (above) indicate that the hydrodynamic regime has potential implications on their ability to capture prey. To verify these expectations, the feeding rate of Sparus aurata (gilt head sea bream) larvae ranging from first feeding until metamorphosis (8-23 Days Post Hatching [DPH]) was quantified in a series of laboratory experiments. The feeding rate of larvae increased by >10 fold from first feeding (3 mm length) to metamorphosis (9 mm). Hi-speed videos of feeding larvae revealed that the increase in feeding rates did not merely reflect increasing motivation or metabolic demands with age. Instead, we observed a sharp increase in prey capture success from <5% for 8 dph larvae to >60% two weeks later. These results are summarized in publication [2]
Subsequent work generated a dataset of >700 feeding events of 8-33 DPH larvae used to elucidate the mechanism of failure in prey capture throughout ontogeny. Failed strikes resulted from insufficient force exerted on the prey (prey not engulfed during mouth opening), flow reversals (prey entering the mouth and leaving during mouth closure) and active spitting of prey items (after they were already engulfed). All these mechanisms scaled with larval size/age to reduce overall feeding efficiency in small sizes. To mechanistically separate the effects of hydrodynamics (size) from other ontogenetic changes (vision, morphology, behavior etc) we used dynamic scaling experiments in which older larvae were filmed foraging in higher-viscosity solutions. These experiments recapitulated age-associated patterns in all three identified mechanisms, indicating that larval size (rather than age) mechanistically impedes suction feeding performance. A manuscript summarizing these results was submitted to ICB on Jan 30th 2015
3) Reynolds numbers effects on prey capture: Small larvae might be less effective feeders, however larvae (and prey) in small Re are also known to swim slower, and this effect can reduce encounter rate with the prey. We used observed swimming speeds of larvae and prey in all dynamic scaling experiments to construct an encounter rate model. This model indicated that the decay in feeding rate due to viscosity effects on swimming speeds did not explain the marked decrease in experimentally observed feeding rates at high viscosities. Slower swimming could account only for ~1/4 of the observed decline in feeding rates. Subsequent high-speed video analysis of feeding strikes in a dynamic scaling setup showed that successful strikes are characterized by fast mouth opening and strike initiation distance. Under high viscosity, mouth-opening speeds declined, and successful strikes were characterized by shorter strike initiation. These results are also summarized in publication [2].