Notes: WW, wet weight; P _SW, time spent in surface waters; P _MW, time spent in mesopelagic waters;T _SW, mean temperature in surface waters;T _MW, mean temperature in mesopelagic waters;R _SW, daily food ration in surface waters;R _MW, daily food ration in mesopelagic waters;FC , food carbon; BC , body carbon;K_e , body carbon release rate; DOC, dissolved organic carbon; CO₂, carbon dioxide; PC, particulate carbon. Data in brackets are ranges corresponding to the above mean values.

a, The total WW of the mesopelagic fish in the open ocean between 40°N and 40°S was assumed to be 10⁹–10¹⁰t, and the WW of the mesopelagic fish in other regions between 40°N–70°N and 40°S–70°S was assumed to be 0.3×10⁹–10¹⁰ t (Lam & Pauly, 2005; Irigoien et al., 2014). The biomass of DVM and NM mesopelagic fishes were calculated by assuming 30%–50% of the mesopelagic fish undergo diel vertical migration (Davison et al., 2015; Klevjer et al., 2016).

b, Davison et al., 2015; Davison et al., 2013

c, Davison et al., 2013; Irigoien et al., 2014

d, Kaeriyama & Ikeda, 2004; Max et al., 2012

e, Davison et al., 2013

f, Harris et al., 2000

g, Childress & Nygaard, 1973

4 Discussion

Existing knowledge regarding the food carbon allocation and body carbon release (/turnover) of marine fish is limited. The carbon AE of marine fish has seldom been reported.

The proportion (7%–8%) of PC released from the food carbon of marine medaka was in the range (0.8%–9.7%) of those of seven carnivorous marine fishes fed fish (Ammodytes personatus ) pieces (Tang et al., 2003) and was comparable to that (8.2%–9.7%) of the detritivorous fish Liza haematocheila (Kang et al., 2007, 2010).

The K _e (0.053–0.12 d^-1) of marine medaka measured in this study was within the range (0.0044–0.14 d^-1) of reported carbon turnover rates for fish muscle tissue (Weidel et al., 2011). Our results showed that most (40%–45%) of the replaced and released body carbon was released in the form of CO₂, indicating that respiration is the largest loss route for the released body carbon. The proportion was at the lower end of the reported values (44.3%–79.4%) for carnivorous marine fishes (Tang et al., 2003). The measured CO₂ from body carbon in our study may be only part of the total carbon used for respiration because some catabolized carbon (in the form of bicarbonate) is excreted in fish intestines, forms precipitated carbonates, and is finally released as fecal pellets (Salter et al., 2017; Wilson et al., 2009). Providing that the entire measured PC released from the body carbon was precipitated carbonates in fish feces, we could estimate that up to 58%–61% (59.5% on average) of the released body carbon was used for respiration. In fact, according to our conceptual model, the carbon used for marine medaka respiration was from not only replaced (and released) fish body carbon but also from ingested food. In other words, the daily respiration rate of mesopelagic fish could be derived from the daily release rate of CO₂from food and that from released body carbon (Text S6).

The mesopelagic fish respiration rates derived from the model fish marine medaka are consistent with recent understanding about the power-law relationship of the mesopelagic fish respiration rate to the fish wet mass and habitat temperature (Text S6; Figure S3). Using 74 data points (each of which includes the respiration rate of myctophids, one of the most biomass dominant groups of mesopelagic fish, the temperature and the fish WW) from five studies, a power-law equation was developed to calculate the WW-specific respiration rate from fish WW and ambient temperature (Belcher et al., 2019). By using this equation, we calculated the daily respiration rates of 0.5-g mesopelagic fish at the different ambient temperatures (3, 8, 9, and 25°C) used in our estimation. The daily respiration rates calculated by using the equation were not different from the daily CO₂ release rates derived from the model fish (paired t-test, p > 0.1; Figure S3).In addition, significant power-law relationships exist between the calculated daily respiration rates, and the model fish-derived daily CO₂ release rates (Figure S3).This consistency justifies our use of carbon release parameters derived from marine medaka to extrapolate the carbon release of wild mesopelagic fish.

Our results showed that substantial proportions of the ingested food carbon and the lost body carbon of the model fish are released as DOC. To our knowledge, the DOC excreted by marine fish has not yet been directly measured in previous studies.

In addition, the release of DOC from fish feces has not yet been reported. The contribution of fecal leakage to the measured DO¹⁴C was not examined in the present study. Contradictory results have been reported about the contribution of fecal pellets to DOC released from zooplankton. Substantial DOC released from fecal pellets of zooplankton has been reported (Thor et al., 2003; Urban-Rich, 1999), but some studies show that the leaching of DOC from fecal pellets of zooplankton was insignificant compared to the DOC released through excretion (Steinberg et al., 2000). We do not think that the leakage of fish feces (if it exists) would contribute much to the DO¹⁴C measured in the present study; PC in feces only accounted for a small proportion (7%–8%) of the food carbon release (Figure 5), and the release rates of the DOC did not peak during the first 2–4 h of the depuration, when the release rates of feces peaked (Figure 4d, e). However, why a substantial proportion (46%–49%) of the released DOC was COC (Figure S2) is still open for discussion.

Assuming a mean global primary production of 59.2 Pg C/y (41–77 Pg C/y) as the scaling basis (del Giorgio and Duarte 2002), our estimation shows that the DOC (1.34–15.2Pg C/y), CO₂ (0.95–10.8 Pg C/y), and PC (0.35–3.97 Pg C/y) released by mesopelagic fish in the open ocean were 2.3%–25.7%, 1.6%–18.2%, and 0.6%–6.7% of the global primary production, respectively. Our estimation of the global CO₂ released by mesopelagic fish was comparable to an estimation that the carbon consumed in by mesopelagic fish respiration was approximately 10.5% ± 7.8% of the primary production along a global investigation transect (Irigoien et al., 2014).The upper limit of the estimated CO₂ released by mesopelagic fish was comparable to the amount of carbon consumed by mesozooplankton respiration (13.0 Pg C/y) in global oceans (Hernández-León & Ikeda, 2005). The amount of PC in fecal pellets released by mesopelagic fish was approximately 1/20 to 1/2 of the amount of fecal carbon (with upper limits of 6.2–6.8 Pg C/y) released by mesozooplankton in epipelagic oceans (Steinberg & Landry, 2017). By assuming that half of the daily food carbon release of DVM mesopelagic fish occurred in deep waters, the active carbon export mediated by the DVM mesopelagic fish (0.54–8.86 Pg C/y) was estimated to be comparable to or even higher than the carbon export mediated by DVM zooplankton (1.04 ± 0.26 Pg C/y) (Archibald et al., 2019). The contribution of DVM micronekton (mainly dominated by fish) to respiratory flux has been reported to be similar to that of DVM zooplankton in the northeastern Atlantic Ocean (Ariza et al., 2015). The high biomass of mesopelagic fish, which is comparable to or even higher than the biomass of mesozooplankton in global oceans, may explain the high carbon release from mesopelagic fish. The global mesopelagic fish biomass (1.3–13 Pg WW or 0.11–1.1 Pg C) used in our estimation is comparable to or even higher than the global mesozooplankton biomass (0.26 Pg C) in the epipelagic ocean (0–200 m in depth) (Hernández-León & Ikeda, 2005), where most zooplankton are distributed. In fact, mesopelagic fish biomass is likely higher than mesozooplankton biomass at low latitudes. For example, recent studies show that the biomass of mesopelagic fish is approximately 1.51–29.38 g C/m² in the open oceans between 40°N and 40°S, and 2.09–3.10 g C/m² in the southern California current ecosystem, whereas the epipelagic mesozooplankton biomass is only 0.15–1.3 g C/m² at the same latitudes (Davison et al., 2015; Hernández-León & Ikeda, 2005; Irigoien et al., 2014).

Our estimation of the active export of DOC by mesopelagic fish (0.28–4.59 Pg C/y) was comparable to the estimates of the global export of DOC below 74 m by mixing (2.31 ± 0.60 Pg C/y) (Roshan & DeVries, 2018) and indicates that DVM mesopelagic fish are an important source of not only ammonium (Bianchi et al. 2014) but also DOC for the mesopelagic layer. This may explain the dissolved organic matter anomalies concurrent with migrating animals (mainly fish) in the mesopelagic layer (Boyd et al., 2019) and support the argument that the supply of significant amounts of labile DOC from mesopelagic fish sustains a microbial growth efficiency in the mesopelagic layer that is twice as high as that at the surface of the Red Sea (Calleja et al. 2018).

The present estimation of the active export of PC by mesopelagic fish (0.07–1.14 Pg C/y) is lower than a recent estimation of the global magnitude of carbon export by the mesopelagic migrant pump (0.9–3.6 Pg C/y), which is mediated mainly by mesopelagic fish (Boyd et al., 2019), but is comparable to an estimation of the active carbon export by vertically migrating marine fish (0.19 Pg C/year) (Aumont et al. 2018).

Our estimation of the carbon released from mesopelagic fish is undeniably still far from precise. This is an opportunity to thoughtfully examine the values coming from lab experiments and the other inputs to the extrapolation analysis and to offer specific advice on future research topics. Uncertainties may come from the use of parameters derived from marine medaka and from the literature, including the estimation of mesopelagic fish biomass, the vertical migration behavior of mesopelagic fish, the use of only a single food for the model fish, the allocation of ingested food carbon to release, and the exclusion of varying metabolic rates and K_e s of fish during different activities. Better information about these factors will help to improve the estimations.

Our estimation of the carbon released from mesopelagic fish is strongly dependent on the mesopelagic fish biomass. The substantially varying estimates of global mesopelagic fish biomass lead to much uncertainty. The mesopelagic fish biomass used in our estimation (1.3–13 Pg) is similar to a recent estimate of global mesopelagic fish biomass of 1.8–16 Pg (Proud et al., 2019). The large variation in the estimates of mesopelagic fish biomass could be the most important factor accounting for the large ranges (covering one order of magnitude) in the estimated DOC, CO₂, and PC released by mesopelagic fish in the present study (Figure 8). In fact, the estimates of mesopelagic fish biomass used in our estimation and in Proud et al. (2019) are based on the acoustic method. Methodological uncertainties from the acoustic method (e.g., interference from siphonophores) are the major cause of variation in the estimation of mesopelagic fish biomass (Proud et al. 2019). In addition, the lack of acoustic data about mesopelagic fish at high latitudes may further undermine the estimation of mesopelagic fish biomass in the global open ocean. Early studies based on trawling document that the density of mesopelagic fish at high latitudes could be several fold higher than that at low latitudes (Lam & Pauly, 2005). A recent study based on global observations from a satellite-mounted lidar also shows that the total DVM animal biomass is higher in the more-productive high-latitude oceans (Behrenfeld et al., 2019). Therefore, constraining the uncertainty of the acoustic method and performing more surveys based on multiple methods and covering a broader area, especially those at high latitudes, are needed to more precisely determine the biomass of global mesopelagic fish.

The ratio of DVM mesopelagic fish to total mesopelagic fish was based on two recent studies at low latitudes, from 40 °N to 40 °S, and in the southern California current system (Davison et al., 2015; Klevjer et al., 2016). Little is known about the proportion of DVM mesopelagic fish at high latitudes. It may be reasonable to assume that the DVM mesopelagic fish at low latitudes spend 12 h of diurnal time in the upper ocean and another 12 h of nocturnal time at mesopelagic depths. However, diurnal and nocturnal times at high latitudes vary significantly in different seasons. Little is known about seasonal variations in vertical-migration behavior or about mesopelagic fish biomass at high latitudes. More observations at high latitudes are needed, and recent progress in satellite observations of DVM animal biomass may facilitate such work (Proud et al. 2019).

The use of parameters derived from the model zooplanktivorous fish marine medaka to extrapolate the allocation of ingested food and the released body carbon of mesopelagic fish may lead to uncertainties. However, the extrapolation is reasonable, at least for the moment. First, the model fish marine medaka ecologically resembles mesopelagic fish, especially because both the model fish and most mesopelagic fish live on zooplankton, and they have similar body sizes (in centimeters). Second, it is still a technical challenge to catch and rear living mesopelagic fish in the lab (Belcher et al., 2019), making it difficult (if not impossible) to measure the “actual” carbon release parameters of mesopelagic fish. Third, very little (if any) data on the carbon release of marine zooplanktivorous fish (let alone mesopelagic fish) were available before our study, making it difficult to find appropriate data in the literature to fit our model. As discussed above, the consistency of the model fish-derived respiration rates of mesopelagic fish with those from the literature provides strong support for our extrapolation (Text S6; Figure S3). Undoubtedly, more experimental work is needed to examine the carbon release parameters of other zooplanktivorous fish, especially mesopelagic fish, if possible.

The use of only rotifers as the food for the fish may lead to the underestimation of the fraction of ingested food allocated to fish feces. Only rotifers were used as living zooplankton food for the marine medaka, and no other zooplankton, such as copepods, the main natural food for mesopelagic fish, was used to feed the fish. One of the main reasons for this is that it is still a challenge to rear enough living copepods or other zooplankton to feed fish and complete experiments. The fish may allocate the carbon of ingested food differently depending on the zooplankton types in their diet. For example, the exoskeletons of copepods may lead to an increased fraction of feces, as the chitin in the exoskeletons cannot be digested by most fish (Durbin & Durbin, 1981, Pinnegar & Polunin, 2006). Therefore, more lab experiments that feed zooplanktivorous fish (including marine medaka) copepods and other zooplankton are needed to examine the carbon release parameters of zooplanktivorous fish.

Although our model fish-derived respiration rates for mesopelagic fish are consistent with those from the literature (Text S6; Figure S3), the estimated allocation of ingested food carbon and lost body carbon to respiration and release as CO₂ by mesopelagic fish might be underestimated because metabolism during feeding and diel swimming between the upper ocean and mesopelagic depths were not considered in the present estimation. According to our experimental designs, CO₂ release from food was counted only after the feeding; no intense swimming occurred after the feeding because the individual fish had been placed in small beakers for the depuration. The active, feeding metabolic rate can be four times the standard metabolic rate of resting, inactive fish (Davison et al., 2013; Smith & Laver, 1981); therefore, our estimation, which did not consider the fish metabolism during feeding and swimming, may underestimate the CO₂ released from the ingested food of the mesopelagic fish. For the same reason, the derived K_e s and related CO₂ release rates of mesopelagic fish might also be underestimated. Thus, future work is needed to examine the release rates for the model fish during different activities, such as swimming.

In contrast, the decrease in K _e with the increase in the daily food ration indicates that the K _e of marine medaka used for the estimation might be overestimated. According to our pilot studies, 1000 and 2000 rotifers were enough to fill the experimental fish stomachs 38% and 75% full, respectively. However, as a daily food ration for a marine medaka with WW of approximately 0.08 g and living at 25°C, 1000 rotifers is below the maintenance level for fish growth, and the high K _e (0.12 d^-1) indicates that substantial body carbon was used for catabolism. The decreased K _e (0.053 d^-1) following the doubling of the daily food ration indicates that, as the supply of food increased, the body carbon used for metabolic turnover decreased. We expect thatK _e would continue to decrease if we further increased the daily food ration. From this perspective, theK _e of 0.053 d^-1 used to extrapolate theK _es of mesopelagic fish might overestimate the body carbon release from the mesopelagic fish. However, the consistency of our model fish-derived respiration rates of mesopelagic fish with those from the literature (Text S6; Figure S3) indicates that the uncertainties from the two factors discussed above may offset each other.

Other factors, such as the simplification of the mean temperature of seawater in the upper and mesopelagic depths, the use of a temperature-dependent daily food ration, the exclusion of fish mortality and reproduction, and the assumption that all mesopelagic fish to be zooplanktivorous, may also have led to uncertainties in the present estimation. Further efforts to minimize the negative influences of the factors discussed above are needed to improve the accuracy of the estimates of carbon releases from mesopelagic fish in the global open ocean.

By providing the first quantitative estimates of DOC, CO₂, and PC released by mesopelagic fish in the global open ocean, our results strengthen the argument that mesopelagic fish may play important roles in the ocean carbon cycle by mediating carbon export in the ocean. First, our results show that the DOC released by mesopelagic fish could be an important organic carbon source for heterotrophic biota in the ocean. Substantial amounts of released DOC, as well as of CO₂ and PC, may be actively transported to mesopelagic depths through the vertical migration of mesopelagic fish. The DOC influx to mesopelagic oceans through DVM mesopelagic fish may narrow the carbon imbalance between the estimated organic carbon influxes and the measured heterotrophic carbon consumption, which is significantly higher than the former in mesopelagic oceans (Burd et al., 2010; Giering et al., 2014; Steinberg et al., 2008).

Second, the PC in the fecal pellets produced by mesopelagic fish could contribute greatly to carbon export through the biological carbon pump. As noted above, the amount of PC in fecal pellets (0.35–3.97 Pg C/y) released by mesopelagic fish is nonnegligible, even substantial. The contribution of mesopelagic fish to carbon export may be even more important, if we consider that the sinking rates of fish fecal pellets are much (even one order of magnitude) greater than those of zooplankton fecal pellets (Saba & Steinberg, 2012).

5 Conclusions

We propose a carbon release model that divides fish-released carbon into two parts, i.e., food carbon release and body carbon release (on the basis of the source: ingested food or the fish body, respectively), and three forms, DOC, CO₂, and PC, which enable the quantification of the release of carbon by fish. By using¹⁴C-labeled living zooplankton to feed a model marine zooplanktivorous fish, this study provided a detailed methodology for precisely quantifying the carbon budget and carbon release of marine fish. By using the carbon release model and parameters derived from the model fish and the literature, we estimated the DOC, CO₂, and PC released by mesopelagic fish in the global open ocean. Our results demonstrated that marine zooplanktivorous fish such as marine medaka can convert substantial fractions of their daily ingested food carbon (26%–42%) and released (/replaced) body carbon (39%–42%) into seawater as DOC. Mesopelagic fish in the global open ocean were estimated to produce 1.34–15.2, 0.95–10.8, and 0.35–3.97 Pg C/y of DOC, CO₂, and PC, respectively. The conceptual model, the laboratory experiments with model fish, and the extrapolation to mesopelagic fish generated a complete solution for estimating the carbon released by fish, especially by global mesopelagic fish. Our estimation is undeniably still far from precise, and factors bringing about uncertainties were discussed. More experimental work is needed to examine the carbon release parameters of marine zooplanktivorous fish, and further observations based on multiple methods are suggested to cover broader areas, especially those at high latitudes, to more precisely determine the mesopelagic fish biomass and their vertical migration behaviors at different latitudes. Our study indicates that mesopelagic fish could be an important source of DOC in the ocean and play critical roles in the biological pump by producing substantial amounts of DOC and fast-sinking fecal pellets and by the active export of DOC, CO₂, and PC into deep waters through their diel vertical migration.

Acknowledgments and Data Statement

We thank Jie Xu, Yehui Tan, and Xingyu Song for their technical supports for conducting the experiments and measuring the radiocarbon. This work is supported by National Key Research and Development Project of China (2017YFC0506302), the Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0405; GML2019ZD0402), the National Natural Science Foundation of China (41506150), the Guangdong Basic and Applied Basic Research Foundation (2015A030310169; 2017A030313217; 2019A1515011645), the Science and Technology Planning Project of Guangdong Province, China (2017B030314052). All the data supporting the conclusions can be obtained from the tables and figures in the main manuscript and in a related data file in the repository of ResearchGate (http://dx.doi.org/ 10.13140/RG.2.2.15051.85281).

Author contributions

Q. Liu played a key role in designing and implementing the experiments, analyzing results and preparing the manuscript. L. Zhou played a key role in forging the model scenarios, interpreting results, preparing the manuscript, and helped the implementation of the experiments. Y. Wu contributed to the data analysis and paper writing. X. He contributed to the data analysis and modeling, and paper writing. N. Gao contributed to implementation of the experiments. Professor L. Zhang supervised the whole research design, experimental process, data interpretation, and the manuscript composition

References

Archibald, K. M., Siegel, D. A., & Doney, S. C. (2019). Modeling the impact of zooplankton diel vertical migration on the carbon export flux of the biological pump. Global Biogeochemical Cycles, 33 (2), 181-199. doi:10.1029/2018GB005983

Ariza, A., Garijo, J. C., Landeira, J. M., Bordes, F., & Hernández-León, S. (2015). Migrant biomass and respiratory carbon flux by zooplankton and micronekton in the subtropical northeast Atlantic Ocean (Canary Islands). Progress In Oceanography, 134 , 330-342. doi:10.1016/j.pocean.2015.03.003

Aumont, O., Maury, O., Lefort, S., & Bopp, L. (2018). Evaluating the potential impacts of the diurnal vertical migration by marine organisms on marine biogeochemistry. Global Biogeochemical Cycles, 32 , 1622-1643. doi:10.1029/2018GB005886

Bachiller, E., Utne, K. R., Jansen, T., & Huse, G. (2018). Bioenergetics modeling of the annual consumption of zooplankton by pelagic fish feeding in the Northeast Atlantic. PLoS ONE 13 , e0190345. doi:10.1371/journal.pone.0190345

Bianchi, D., Babbin, A. R., & Galbraith, E. D. (2014). Enhancement of anammox by the excretion of diel vertical migrators. Proceedings of the National Academy of Sciences, 111 , 15653. doi:10.1073/pnas.1410790111

Belcher, A., Saunders, R. A., & Tarling, G. A. (2019). Respiration rates and active carbon flux of mesopelagic fishes (Family Myctophidae) in the Scotia Sea, Southern Ocean. Marine Ecology Progress Series, 610 , 149-162.

Bo, J., Cai, L., Xu, J. H., Wang, K. J., & Au, D. W. T. (2011). The marine medaka Oryzias melastigma - A potential marine fish model for innate immune study. Marine Pollution Bulletin 63 , 267-276. doi:10.1016/j.marpolbul.2011.05.014

Boyd, P. W., Claustre, H., Levy, M., Siegel, D. A., & Weber, T. (2019). Multi-faceted particle pumps drive carbon sequestration in the ocean.Nature, 568 , 327-335. doi:10.1038/s41586-019-1098-2

Behrenfeld, M. J., Gaube, P., Della Penna, A., O’Malley, R. T., Burt, W. J., Hu, Y., . . . Doney, S. C. (2019). Global satellite-observed daily vertical migrations of ocean animals. Nature. doi:10.1038/s41586-019-1796-9

Burd, A. B., Hansell, D. A., Steinberg, D. K., Anderson, T. R., Arístegui, J., Baltar, F., . . . Tanaka, T. (2010). Assessing the apparent imbalance between geochemical and biochemical indicators of meso- and bathypelagic biological activity: What is wrong with present calculations of carbon budgets? Deep Sea Research Part II, 57 (16), 1557-1571. doi:10.1016/j.dsr2.2010.02.022

Calleja, M. L., Ansari, M. I., Røstad, A., Silva, L., Kaartvedt, S., Irigoien, X., & Morán, X. A. G. (2018). The mesopelagic scattering layer: A hotspot for heterotrophic prokaryotes in the red sea twilight zone. Frontiers in Marine Science, 5 (259). doi:10.3389/fmars.2018.00259

Childress, J. J., & Nygaard, M. H. (1973). The chemical composition of midwater fishes as a function of depth of occurence off southern California. Deep Sea Research and Oceanographic Abstracts 20 , 1093-1109. doi: 10.1016/0011-7471(73)90023-5

Cho, C. Y., Slinger, S. J., & Bayley, H. S. (1982). Bioenergetics of salmonid fishes: Energy intake, expenditure and productivity.Comparative Biochemistry and Physiology Part B: Comparative Biochemistry, 73 (1), 25-41. doi:10.1016/0305-0491(82)90198-5

Davison, P., Lara-Lopez, A., & Koslow, J. A. (2015). Mesopelagic fish biomass in the southern California current ecosystem. Deep-Sea Research II 112 , 129-142. doi:10.1016/j.dsr2.2014.10.007

Davison, P. C., Checkley, D. M., Koslow, J. A., & Barlow, J. (2013). Carbon export mediated by mesopelagic fishes in the northeast Pacific Ocean.Progress in Oceanography 116 , 14-30. doi:10.1016/j.pocean.2013.05.013

del Giorgio, P. A., & Duarte, C. M. (2002). Respiration in the open ocean.Nature 420 , 379-384. doi:10.1038/nature01165

Durbin, E. G., & Durbin, A. G. (1981). Assimilation efficiency and nitrogen excretion of a filter-feeding planktivore, the Atlantic Menhaden, Brevootia tyrannus (Pisces: Clupeidae). Fishery Bulletin 79 , 601-616.

Giering, S. L. C., Sanders, R., Lampitt, R. S., Anderson, T. R., Tamburini, C., Boutrif, M., . . . Mayor, D. J. (2014). Reconciliation of the carbon budget in the ocean’s twilight zone. Nature, 507 , 480–483. doi:10.1038/nature13123

Gillooly, J. F., Brown, J. H., West, G. B., Savage, V. M., & Charnov, E. L. (2001). Effects of size and temperature on metabolic rate.Science 293 , 2248-2251. doi:10.1126/science.1061967

Hansen, B., Wernberg-Møller, T., & Wittrup, L. (1997). Particle grazing efficiency and specific growth efficiency of the rotiferBrachionus plicatilis (Muller). Journal of Experimental Marine Biology and Ecology 215 , 217-233. doi: 10.1016/S0022-0981(97)00053-1

Harris, R. P., Wiebe, P. H., Lenz, J., Skjldal, H. R., & Huntley, M. (Eds.). (2000). Zooplankton Methodology Manual . San Diego, California: Academic Press.

He, X., & Wang, W.-X. (2006). Relative importance of inefficient feeding and consumer excretion to organic carbon flux from Daphnia.Freshwater Biology 51 , 1911-1923. doi:10.1111/j.1365-2427.2006.01631.x

Hernández-León, S., & Ikeda, T. (2005). A global assessment of mesozooplankton respiration in the ocean. Journal of Plankton Research 27 , 153-158. doi:10.1093/plankt/fbh166

Irigoien, X., Klevjer, T. A., Røstad, A., Martinez, U., Boyra, G., Acuña, J. L., . . . Kaartvedt, S. (2014). Large mesopelagic fishes biomass and trophic efficiency in the open ocean. Nature Communications 5 , 3271. doi:10.1038/ncomms4271

Kaeriyama, H., & Ikeda, T. (2004). Metabolism and chemical composition of mesopelagic ostracods in the western North Pacific Ocean. ICES Journal of Marine Science 61 , 535-541. doi:10.1016/j.icesjms.2004.03.009

Kang, B., Xian, W.-w., & Wu, Y.-f. (2007). Carbon budget of redlip mullet (Liza haematocheila T. & S.) under different feeding rates. Periodical of Ocean Universtiy of China 37 , 247-250. doi:10.3969/j.issn.1672-5174.2007.02.013

Kang, B., Xian, W.-w., & Wu, Y.-f. (2010). Carbon budgets of redlip mullet (Liza haematocheila ) fed on two diets. Marine Sciences 34 , 11-14.

Klevjer, T. A., Irigoien, X., Røstad, A., Fraile-Nuez, E., Benítez-Barrios, V. M., & Kaartvedt, S. (2016). Large scale patterns in vertical distribution and behaviour of mesopelagic scattering layers.Scientific Reports 6 , 11. doi:10.1038/srep19873

Kong, R. Y. C., Giesy, J. P., Wu, R. S. S., Chen, E. X. H., Chiang, M. W. L., Lim, P. L., . . . Au, D. W. T. (2008). Development of a marine fish model for studying in vivo molecular responses in ecotoxicology.Aquatic Toxicology 86 , 131-141. doi:10.1016/j.aquatox.2007.10.011

Lam, V. W. Y., & Pauly, D. (2005). Mapping the global biomass of mesopelagic fishes. Sea Around Us Project Newsletter 30 , 4.

Lee, B.-G., & Fisher, N. S. (1992). Degradation and elemental release rates from phytoplankton debris and their geochemical implications.Limnology and Oceanography 37 , 1345-1360. doi:10.2307/2837955

Martin, A., et al. (2020), The oceans’ twilight zone must be studied now, before it is too late, Nature , 580 (7801), 26-28.

Max, L. et al.. (2012). Sea surface temperature variability and sea-ice extent in the subarctic northwest Pacific during the past 15,000 years.Paleoceanography 27 , PA3213. doi:10.1029/2012PA002292

Mu, J., Jin, F., Wang, J., Zheng, N., & Cong, Y. (2015). Effects of CO₂-driven ocean acidification on early life stages of marine medaka (Oryzias melastigma ). Biogeosciences 12 , 3861-3868. doi:10.5194/bg-12-3861-2015

Ney, J. J. (1993). Bioenergetics modeling today: growing pains on the cutting edge. Transactions of the American Fisheries Society 122 , 736-748. doi:10.1577/1548-8659(1993)122<0736:BMTGPO>2.3.CO;2

Øie, G., Makridis, P., Reitan, K. I., & Olsen, Y. (1997). Protein and carbon utilization of rotifers (Brachionus plicatilis ) in first feeding of turbot larvae (Scophthalmus maximus L. ).Aquaculture 153 , 103-122. doi: 10.1016/S0044-8486(96)01514-1

Pershing, A. J., Christensen, L. B., Record, N. R., Sherwood, G. D., & Stetson, P. B. (2010). The impact of whaling on the ocean carbon cycle: why bigger was better. PLoS ONE 5 , e12444. doi:10.1371/journal.pone.0012444

Pinnegar, J. K., & Polunin, N. V. C. (2006). Planktivorous damselfish support significant nitrogen and phosphorus fluxes to Mediterranean reefs. Marine Biology 148 , 1089-1099. doi:10.1007/s00227-005-0141-z

Proud, R., Handegard, N. O., Kloser, R. J., Cox, M. J., & Brierley, A. S. (2018). From siphonophores to deep scattering layers: uncertainty ranges for the estimation of global mesopelagic fish biomass. ICES Journal of Marine Science, 76 (3), 718-733. doi:10.1093/icesjms/fsy037

Roshan, S., & DeVries, T. (2018). Efficient dissolved organic carbon production and export in the oligotrophic ocean. Nature Communications, 8 , 2036. doi:10.1038/s41467-017-02227-3

Saba, G. K., & Steinberg, D. K. (2012). Abundance, composition, and sinking rates of fish fecal pellets in the Santa Barbara Channel.Scientific Reports 2, 716. doi:10.1038/srep00716

Salter, M. A., Harborne, A. R., Perry, C. T., & Wilson, R. W. (2017). Phase heterogeneity in carbonate production by marine fish influences their roles in sediment generation and the inorganic carbon cycle.Scientific Reports 7 , 15. doi:10.1038/s41598-017-00787-4

Smith, K. L., & Laver, M. B. (1981). Respiration of the bathypelagic fish Cyclothone acclinidens . Marine Biology 61 , 261-266. doi:10.1007/BF00401564

Steinberg, D. K., Carlson, C. A., Bates, N. R., Goldthwait, S. A., Madin, L. P., & Michaels, A. F. (2000). Zooplankton vertical migration and the active transport of dissolved organic and inorganic carbon in the Sargasso Sea. Deep-Sea Research I 47 , 137-158. doi:10.1016/s0967-0637(99)00052-7

Steinberg, D. K., & Landry, M. R. (2017). Zooplankton and the ocean carbon cycle. Annual Review of Marine Science 9 , 413-444. doi:10.1146/annurev-marine-010814-015924

Steinberg, D. K., Van Mooy, B. A. S., Buesseler, K. O., Boyd, P. W., Kobari, T., & Karl, D. M. (2008). Bacterial vs. zooplankton control of sinking particle flux in the ocean’s twilight zone. Limnology and Oceanography, 53 (4), 1327-1338. doi:10.4319/lo.2008.53.4.1327

Tang, Q., Sun, Y., & Zhang, B. (2003). Bioenergetics models for seven species of marine fish. Journal of Fisheries of China 27 , 443-449. doi: 10.3321/j.issn:1000-0615.2003.05.010

Thor, P., Dam, H. G., & Rogers, D. R. (2003). Fate of organic carbon released from decomposing copepod fecal pellets in relation to bacterial production and ectoenzymatic activity. Aquatic Microbial Ecology 33 , 279-288. doi:10.3354/ame033279

Trueman, C. N., Johnston, G., Hea, B., & MacKenzie, K. M. (2014). Trophic interactions of fish communities at midwater depths enhance long-term carbon storage and benthic production on continental slopes.Proceedings of the Royal Society B: Biological Sciences 281 , 20140669. doi:10.1098/rspb.2014.0669

Urban-Rich, J. (1999). Release of dissolved organic carbon from copepod fecal pellets in the Greenland Sea. Journal of Experimental Marine Biology and Ecology 232 , 107-124. doi: 10.1016/S0022-0981(98)00104-X

Warren, C. E., & Davis, G. E. (1967). Laboratory studies on the feeding, bioenergetics and growth of fish. In S. D. Gerking (Ed.), The Biological Basis of Freshwater Fish Production (pp. 175-214). Oxford: Blackwell Scientific Publication.

Weidel, B. C., Carpenter, S. R., Kitchell, J. F., & Vander Zanden, M. J. (2011). Rates and components of carbon turnover in fish muscle: insights from bioenergetics models and a whole-lake ¹³C addition. Canadian Journal of Fisheries and Aquatic Sciences 68 , 387-399. doi:10.1139/f10-158

Wilson, R. W., Millero, F. J., Taylor, J. R., Walsh, P. J., Christensen, V., Jennings, S., & Grosell, M. (2009). Contribution of fish to the marine inorganic carbon cycle. Science 323 , 359-362. doi:10.1126/science.1157972

Zhang, W., & Wang, W.-X. (2004). Colloidal organic carbon and trace metal (Cd, Fe and Zn) released by diatom exudation and copepod grazing.Journal of Experimental Marine Biology and Ecology 307 , 17-34. doi:10.1016/j.jembe.2004.01.016