Bioluminescence as an ecological factor during high Arctic polar night

Bioluminescence commonly influences pelagic trophic interactions at mesopelagic depths. Here we characterize a vertical gradient in structure of a generally low species diversity bioluminescent community at shallower epipelagic depths during the polar night period in a high Arctic fjord with in situ bathyphotometric sampling. Bioluminescence potential of the community increased with depth to a peak at 80 m. Community composition changed over this range, with an ecotone at 20–40 m where a dinoflagellate-dominated community transitioned to dominance by the copepod Metridia longa. Coincident at this depth was bioluminescence exceeding atmospheric light in the ambient pelagic photon budget, which we term the bioluminescence compensation depth. Collectively, we show a winter bioluminescent community in the high Arctic with vertical structure linked to attenuation of atmospheric light, which has the potential to influence pelagic ecology during the light-limited polar night.


Example: How polar night bioluminescence distributions could influence ecological interactions
Here we present one example of how in situ data on bioluminescence produced by the planktonic community might be used to explore pelagic trophic interactions. We modeled the effect of bioluminescence on visual perception of an ecologically significant predator-prey pair.
Specifically, we considered the maximum range at which the krill Thysanoessa inermis could detect a predatory little auk. Thysanoessa inermis is both a dominant micronekton species in Kongsfjord during winter and other times of year [1][2][3][4][5] , and is a member of the bioluminescent community measured in the current study. Little auks are resident year-round on Svalbard and are active predators on T. inermis in Kongsfjord during winter, often feeding selectively on krill 3,5 . Mechanical stimulation of bioluminescence by diving little auks as they move through the water column would present a visual stimulus for krill. Accordingly, we modeled krill visual range for perception of a little auk diving through the bioluminescent community that we observed in Kongsfjord.
We used equations from Nilsson et al. 6 parameterized by our own observations of the environmental light field (downwelling radiance derived from HydroLight modelling and bioluminescent light derived from UBAT measurements) and krill visual perception (derived from both electrophysiological recording from T. inermis eyes and histology, e.g., Cohen et al. 7 ) to quantify visual range for the krill superposition compound eye detecting the silhouette of a little auk approaching from above and therefore appearing as an extended black target triggering bioluminescence against downwelling space-light. Visual range for a krill at a given depth viewing an oncoming diving little auk can be solved for with equation (1):  Fig. 3b), our parameterization of bioluminescence is likely an underestimation of the bioluminescent community. Opposing that, however, we assume an equal capacity for mechanically stimulating luminescence between the UBAT's high turbulence and the impact of a diving little auk. It is likely that birds are less efficient at evoking luminescence than the UBAT making our quantification of bioluminescence an overestimate in this respect.
Collectively, there is uncertainty in the value of bird-stimulated bioluminescence, but our parameterization of it here provides a useful starting point.
We calculated visual range for krill positioned at 1 m depth increments from 1 to 99 m depth using three different scenarios for the stimulated bioluminescent community in order to test whether variations in bioluminescent community composition altered visual performance.
These included scenarios of: (1) no bioluminescence throughout the entire water column; and (2) the average depth-stratified Kongsfjord luminescent community as measured by UBAT at 20 m intervals. In both visual models, the distance at which krill could not discriminate between the little auk and the background was set at twice the wingspan of a little auk, 0.76m, or the point at which it subtended more than 28° of the visual field 6 .
The Visual detection of a predator does not guarantee a krill's ability to behaviorally escape predation, and this may be particularly true when krill reside at shallower depths. There, even though predators may theoretically be visible against downwelling spacelight, the ability of krill to behaviorally respond and escape predation may be limited in optical conditions dominated by atmospheric light where predator foraging behavior is enhanced 11 , and in turn krill avoid them.
A major limitation of our analysis is that the visual model employed here is based on visual discrimination of the target against the background during one integration time, and therefore is best applied to longer viewing distances such that collections of bioluminescent point