Continuous observation of animal behaviour could be quite a challenge in the field, and the situation becomes even more complicated with aquatic species mostly active at night. In such cases, biologging techniques are real game changers in ecology, behavioural ecology or eco-physiology. An accelerating number of methodological applications of these tools in natural condition are thus published each year [1]. Biologging is not limited to movement ecology. For instance, fine grain information about energy expenditure can be inferred from body acceleration [2], and accelerometers has already proven useful in monitoring reproductive costs in some fish species [3,4]. The first part of the study by Tentelier et al. [5] is in line with this growing literature. It describes measurements of energy expenditure during reproduction in a fish species, Allis shad (Alosa Alosa), based on tail beat frequency and occurrence of spawning acts. The study has been convincingly conducted, and the results are important for fish biologists. But this is not the whole story: the authors added to this otherwise classical study a very original and insightful analysis which deserves closer interest.
Tentelier et al. propose to use static accelerometer to monitor change in body roundness through the reproductive season. These semelparous fish first mature and built up reserves in the Atlantic Ocean and migrate into fresh water to reproduce. Contrary to iteroparous species, female shads do not have to strategically preserve energy for future reproduction. The females die few days after spawning having exhausted their energetic reserves: they typically lose almost half of their body mass during the spawning season. The beautiful idea in this study was to track down information about this dramatic slimming in the accelerometer data. Indeed, the accelerometer was attached on the side of the fish (close to the dorsal fin). A change in its angle with the vertical plane could be correlated with the change in roundness, the angle declining with the female thinning. Accelerometers have already been used to record body posture [6] but, in the present study, the novelty was to monitor the change in body shape.
Unfortunately, the data by Tentelier et al. are inconclusive so far. Broadly speaking, the accelerometer angle recorded declined through the spawning season, indicating an average slimming of the females, but there was no correlation between the change in angle and the mass loss at the individual level. This was partly due to the fact that the dorsal position of the accelerometer was not optimized to measures egg laying whose effects are mostly observable on ventral side.
Yet, this nice idea deserves more scrutiny. The method seems to be sensitive enough to detect inflation of swim bladder, the gas-filled organ helping the fish to control their position in the water column, as the accelerometer angle increased when the fish stayed close to the water surface. Additional works and proper calibration are certainly needed to validate the use of accelerometer angle as a proxy for body roundness. The actual data were not strong enough to justify a standalone publication on the subject, but it would have been shame to lose traces of such analysis and keep it in the file drawer. This is why I strongly support its report as a side question in a broader study. Science progresses not only with neat conclusive studies but also when unexpected (apparently anecdotal) observations stimulate new researches.

References

[1] Börger L, Bijleveld AI, Fayet AL, Machovsky‐Capuska GE, Patrick SC, Street GM and Vander Wal E. (2020) Biologging special feature. J. Anim. Ecol. 89, 6–15. 10.1111/1365-2656.13163
[2] Wilson RP et al. (2020) Estimates for energy expenditure in free‐living animals using acceleration proxies: A reappraisal. J. Anim. Ecol. 89, 161–172. 10.1111/1365-2656.13040
[3] Tsuda Y, Kawabe R, Tanaka H, Mitsunaga Y, Hiraishi T, Yamamoto K and Nashimoto K. (2006) Monitoring the spawning behaviour of chum salmon with an acceleration data logger. Ecol. Freshw. Fish 15, 264–274. 10.1111/j.1600-0633.2006.00147.x
[4] Sakaji H, Hamada K, Naito Y. 2018 Identifying spawning events of greater amberjack using accelerometers. Mar. Biol. Res. 14, 637–641. 10.1080/17451000.2018.1492140
[5] Tentelier C, Bouchard C, Bernardin A, Tauzin A, Aymes J-C, Lange F, Récapet C, Rives J (2020) The dynamics of spawning acts by a semelparous fish and its associated energetic costs. bioRxiv, 436295. doi: 10.1101/436295 ver. 7 peer-reviewed and recommended by PCI Ecology. 10.1101/436295
[6] Brown DD, Kays R, Wikelski M, Wilson R, Klimley AP. 2013 Observing the unwatchable through acceleration logging of animal behavior. Anim. Biotelemetry 1, 20. 10.1186/2050-3385-1-20

In the field of ecology, there is a growing interest in machine (including deep) learning for processing and automatizing repetitive analyses on large amounts of images collected from camera traps, drones and smartphones, among others. These analyses include species or individual recognition and classification, counting or tracking individuals, detecting and classifying behavior. By saving countless times of manual work and tapping into massive amounts of data that keep accumulating with technological advances, machine learning is becoming an essential tool for ecologists. We refer to recent papers for more details on machine learning for ecology and evolution (Besson et al. 2022, Borowiec et al. 2022, Christin et al. 2019, Goodwin et al. 2022, Lamba et al. 2019, Nazir & Kaleem 2021, Perry et al. 2022, Picher & Hartig 2023, Tuia et al. 2022, Wäldchen & Mäder 2018).

In their paper, Nakagawa et al. (2023) conducted a systematic review of the literature on machine learning for wildlife imagery. Interestingly, the authors used a method unfamiliar to ecologists but well-established in medicine called rapid review, which has the advantage of being quickly completed compared to a fully comprehensive systematic review while being representative (Lagisz et al., 2022). Through a rigorous examination of more than 200 articles, the authors identified trends and gaps, and provided suggestions for future work. Listing all their findings would be counterproductive (you’d better read the paper), and I will focus on a few results that I have found striking, fully assuming a biased reading of the paper. First, Nakagawa et al. (2023) found that most articles used neural networks to analyze images, in general through collaboration with computer scientists. A challenge here is probably to think of teaching computer vision to the generations of ecologists to come (Cole et al. 2023). Second, the images were dominantly collected from camera traps, with an increase in the use of aerial images from drones/aircrafts that raise specific challenges. Third, the species concerned were mostly mammals and birds, suggesting that future applications should aim to mitigate this taxonomic bias, by including, e.g., invertebrate species. Fourth, most papers were written by authors affiliated with three countries (Australia, China, and the USA) while India and African countries provided lots of images, likely an example of scientific colonialism which should be tackled by e.g., capacity building and the involvement of local collaborators. Last, few studies shared their code and data, which obviously impedes reproducibility. Hopefully, with the journals’ policy of mandatory sharing of codes and data, this trend will be reversed. 

REFERENCES

Besson M, Alison J, Bjerge K, Gorochowski TE, Høye TT, Jucker T, Mann HMR, Clements CF (2022) Towards the fully automated monitoring of ecological communities. Ecology Letters, 25, 2753–2775. https://doi.org/10.1111/ele.14123

Borowiec ML, Dikow RB, Frandsen PB, McKeeken A, Valentini G, White AE (2022) Deep learning as a tool for ecology and evolution. Methods in Ecology and Evolution, 13, 1640–1660. https://doi.org/10.1111/2041-210X.13901

Christin S, Hervet É, Lecomte N (2019) Applications for deep learning in ecology. Methods in Ecology and Evolution, 10, 1632–1644. https://doi.org/10.1111/2041-210X.13256

Cole E, Stathatos S, Lütjens B, Sharma T, Kay J, Parham J, Kellenberger B, Beery S (2023) Teaching Computer Vision for Ecology. https://doi.org/10.48550/arXiv.2301.02211

Goodwin M, Halvorsen KT, Jiao L, Knausgård KM, Martin AH, Moyano M, Oomen RA, Rasmussen JH, Sørdalen TK, Thorbjørnsen SH (2022) Unlocking the potential of deep learning for marine ecology: overview, applications, and outlook†. ICES Journal of Marine Science, 79, 319–336. https://doi.org/10.1093/icesjms/fsab255

Lagisz M, Vasilakopoulou K, Bridge C, Santamouris M, Nakagawa S (2022) Rapid systematic reviews for synthesizing research on built environment. Environmental Development, 43, 100730. https://doi.org/10.1016/j.envdev.2022.100730

Lamba A, Cassey P, Segaran RR, Koh LP (2019) Deep learning for environmental conservation. Current Biology, 29, R977–R982. https://doi.org/10.1016/j.cub.2019.08.016

Nakagawa S, Lagisz M, Francis R, Tam J, Li X, Elphinstone A, Jordan N, O’Brien J, Pitcher B, Sluys MV, Sowmya A, Kingsford R (2023) Rapid literature mapping on the recent use of machine learning for wildlife imagery. EcoEvoRxiv, ver. 4 peer-reviewed and recommended by Peer Community in Ecology.  https://doi.org/10.32942/X2H59D

Nazir S, Kaleem M (2021) Advances in image acquisition and processing technologies transforming animal ecological studies. Ecological Informatics, 61, 101212. https://doi.org/10.1016/j.ecoinf.2021.101212

Perry GLW, Seidl R, Bellvé AM, Rammer W (2022) An Outlook for Deep Learning in Ecosystem Science. Ecosystems, 25, 1700–1718. https://doi.org/10.1007/s10021-022-00789-y

Pichler M, Hartig F Machine learning and deep learning—A review for ecologists. Methods in Ecology and Evolution, n/a. https://doi.org/10.1111/2041-210X.14061

Tuia D, Kellenberger B, Beery S, Costelloe BR, Zuffi S, Risse B, Mathis A, Mathis MW, van Langevelde F, Burghardt T, Kays R, Klinck H, Wikelski M, Couzin ID, van Horn G, Crofoot MC, Stewart CV, Berger-Wolf T (2022) Perspectives in machine learning for wildlife conservation. Nature Communications, 13, 792. https://doi.org/10.1038/s41467-022-27980-y

Wäldchen J, Mäder P (2018) Machine learning for image-based species identification. Methods in Ecology and Evolution, 9, 2216–2225. https://doi.org/10.1111/2041-210X.13075

The processes that trigger community assembly are still in the centre of ecological interest. While prior work mostly focused on spatial patterns of co-occurrence within a meta-community framework [reviewed in 1, 2] recent studies also include temporal patterns of community composition [e.g. 3, 4, 5, 6]. In this preprint [7], Franck Jabot and co-workers extend they prior approaches to quasi neutral community assembly [8, 9, 10] and develop an analytical framework of spatial and temporal diversity turnover. A simple and heuristic path model for beta diversity and an extended ecological drift model serve as starting points. The model can be seen as a counterpart to Ulrich et al. [5]. These authors implemented competitive hierarchies into their neutral meta-community model while the present paper focuses on environmental filtering. Most important, the model and parameterization of four empirical data sets on aquatic plant and animal meta-communities used by Jabot et al. returned a consistent high influence of environmental stochasticity on species turnover. Of course, this major result does not come to a surprise. As typical for this kind of models it depends also to a good deal on the initial model settings. It nevertheless makes a strong conceptual point for the importance of environmental variability over dispersal and richness effects. One interesting side effect regards the impact of richness differences (ΔS). Jabot et al. interpret this as a ‘nuisance variable’ as they do not have a stringent explanation. Of course, it might be a pure statistical bias introduced by the Soerensen metric of turnover that is normalized by richness. However, I suspect that there is more behind the ΔS effect. Richness differences are generally associated with respective differences in total abundances and introduce source – sink dynamics that inevitably shape subsequent colonization – extinction processes. It would be interesting to see whether ΔS alone is able to trigger observed patterns of community assembly and community composition. Such an analysis would require partitioning of species turnover into richness and nestedness effects [11]. I encourage Jabot et al. to undertake such an effort.
The present paper is also another call to include temporal population variability into metapopulation models for a better understanding of the dynamics and triggering of community assembly. In a next step, competitive interactions should be included into the model to infer the relative importance of both factors.

References

[1] Götzenberger, L. et al. (2012). Ecological assembly rules in plant communities—approaches, patterns and prospects. Biological reviews, 87(1), 111-127. doi: 10.1111/j.1469-185X.2011.00187.x
[2] Ulrich, W., & Gotelli, N. J. (2013). Pattern detection in null model analysis. Oikos, 122(1), 2-18. doi: 10.1111/j.1600-0706.2012.20325.x
[3] Grilli, J., Barabás, G., Michalska-Smith, M. J., & Allesina, S. (2017). Higher-order interactions stabilize dynamics in competitive network models. Nature, 548(7666), 210. doi: 10.1038/nature23273
[4] Nuvoloni, F. M., Feres, R. J. F., & Gilbert, B. (2016). Species turnover through time: colonization and extinction dynamics across metacommunities. The American Naturalist, 187(6), 786-796. doi: 10.1086/686150
[5] Ulrich, W., Jabot, F., & Gotelli, N. J. (2017). Competitive interactions change the pattern of species co‐occurrences under neutral dispersal. Oikos, 126(1), 91-100. doi: 10.1111/oik.03392
[6] Dobramysl, U., Mobilia, M., Pleimling, M., & Täuber, U. C. (2018). Stochastic population dynamics in spatially extended predator–prey systems. Journal of Physics A: Mathematical and Theoretical, 51(6), 063001. doi: 10.1088/1751-8121/aa95c7
[7] Jabot, F., Laroche, F., Massol, F., Arthaud, F., Crabot, J., Dubart, M., Blanchet, S., Munoz, F., David, P., and Datry, T. (2019). Assessing metacommunity processes through signatures in spatiotemporal turnover of community composition. bioRxiv, 480335, ver. 3 peer-reviewed and recommended by PCI Ecology. doi: 10.1101/480335
[8] Jabot, F., & Chave, J. (2011). Analyzing tropical forest tree species abundance distributions using a nonneutral model and through approximate Bayesian inference. The American Naturalist, 178(2), E37-E47. doi: 10.1086/660829
[9] Jabot, F., & Lohier, T. (2016). Non‐random correlation of species dynamics in tropical tree communities. Oikos, 125(12), 1733-1742. doi: 10.1111/oik.03103
[10] Datry, T., Bonada, N., & Heino, J. (2016). Towards understanding the organisation of metacommunities in highly dynamic ecological systems. Oikos, 125(2), 149-159. doi: 10.1111/oik.02922
[11] Baselga, A. (2010). Partitioning the turnover and nestedness components of beta diversity. Global ecology and biogeography, 19(1), 134-143. doi: 10.1111/j.1466-8238.2009.00490.x

Today, understanding spatio-temporal dynamics of pathogens is pivotal to understand their transmission and controlling them. First, understanding this dynamics can reveal the ecology of their transmission [1]. Indeed, such knowledge, based on data that are quite easy to access, can shed light on transmission modes, which could rely on different animal species that can be spatially distributed in a non-uniform way [2]. This is especially true for pathogens with complex life-cycles, despite that investigating such dynamics is very challenging and rely mostly on mathematical models.
Moreover, this knowledge can also highlight some weak points in a complex web of transmission and therefore allowing us to envision new innovative control strategies. This has been first proposed on human pathogens, where connectivity among populations can be analyzed to identify which connections need to be targeted to stop or slow down an epidemics [3]. However, this idea is increasingly recognized as a promising new approach for pathogens involving vector populations, especially regarding the complexity to decrease on a long-term the abundance of these vector populations [4].
In "Environmental heterogeneity drives tsetse fly population dynamics and control" [5], Cecilia and co-authors have developed a sophisticated spatio-temporal mechanistic model to figure out how local environment, involved within landscape of different complexities, can impact the population dynamics of tsetse flies, an invertebrate species that can serve as a vector for many pathogens of animal and human importance. They found that spatial patches with the lowest temperature mean and the lowest environmental fluctuations can act as refuge for this species, representing therefore preferential targets for disease control.
The reviewers and I agree that the mathematical framework developed address very well an important topic for both ecological and public health literature. More importantly, it shows how fundamental ecological knowledge can drive pathogen control strategies, opening an interesting avenue for cross-disciplinary research on vector-borne diseases.

References

[1] Grenfell, B. T., Bjørnstad, O. N., & Kappey, J. (2001). Travelling waves and spatial hierarchies in measles epidemics. Nature, 414(6865), 716-723. doi: 10.1038/414716a
[2] Perkins, S. E., Cattadori, I. M., Tagliapietra, V., Rizzoli, A. P., & Hudson, P. J. (2003). Empirical evidence for key hosts in persistence of a tick-borne disease. International journal for parasitology, 33(9), 909-917. doi: 10.1016/S0020-7519(03)00128-0
[3] Colizza, V., Barrat, A., Barthélemy, M., & Vespignani, A. (2006). The role of the airline transportation network in the prediction and predictability of global epidemics. Proceedings of the National Academy of Sciences, 103(7), 2015-2020. doi: 10.1073/pnas.0510525103
[4] Pepin, K. M., Leach, C. B., Marques-Toledo, C., Laass, K. H., Paixao, K. S., et al. (2015) Utility of mosquito surveillance data for spatial prioritization of vector control against dengue viruses in three Brazilian cities. Parasites & Vectors 8, 1–15. doi: 10.1186/s13071-015-0659-y
[5] Cecilia, H., Arnoux, S., Picault, S., Dicko, A., Seck, M. T., Sall, B., Bassène, M., Vreysen, M., Pagabeleguem, S., Bancé, A., Bouyer, J. and Ezanno, P.(2019). Environmental heterogeneity drives tsetse fly population dynamics and control. bioRxiv 493650, ver. 3 peer-reviewed and recommended by PCI Ecology. doi: 10.1101/493650

Identifying the factors which favour the establishment and spread of non-native species in novel environments is one of the keys to predict - and hence prevent or control - biological invasions. This includes biological factors (i.e. factors associated with the invasive species themselves), and one of the prevailing hypotheses is that some species traits may explain their impressive success to establish and spread in novel environments [1]. In animals, most research studies have focused on traits associated with fecundity, age at maturity, level of affiliation to humans or dispersal ability for instance. The “composite picture” of the perfect (i.e. successful) invader that has gradually emerged is a small-bodied animal strongly affiliated to human activities with high fecundity, high dispersal ability and a super high level of plasticity. Of course, the story is not that simple, and actually a perfect invader sometimes – if not often- takes another form… Carrying on to identify what makes a species a successful invader or not is hence still an important research axis with major implications.
In this manuscript, Charbonnel and collaborators [2] provide an interesting opportunity to gain novel insights into our understanding of (the) traits underlying invasion success. They nicely combine the power of Next-Generation Sequencing (NGS) with a clever comparative approach of two closely-related invasive rodents (the house mouse Mus musculus and the black rat Rattus rattus) in a common environment. They use this experimental design to test the appealing hypothesis that pathogens may be actors of the story, and may indirectly explain why some non-native species are so successful in invading novel habitats.
It is generally assumed that the community of pathogens encountered by non-native species in novel environments is different from that of their native area. On the one hand (the enemy-release hypothesis), it can be hypothesized that non-native species, when they arrive into a novel environment, will be relaxed from the pressure imposed by their native pathogens because local pathogens are not adapted (and hence do not infect) to this novel host. Because immune defence against pathogens is highly costly, non-native species establishing into a novel environment could hence reallocate these costs to other functions such as fecundity or dispersal apparatus. This scenario has been termed the “evolution of increased competitive ability” (EICA) hypothesis [3]. On the other hand (the EICA-refined hypothesis [4]), one can assume that invaders will encounter new pathogens in newly established areas, and will allocate energy toward cost-effective immune pathways to permit allocating a non-negligible amount of energy toward other functions. Finally, a last hypothesis (the “immune protection” hypothesis) assumes major changes in pathogen composition between native and invaded areas, which should lead to an overall increase in immune investment by the native species to successfully invade novel environments [4]. This last hypothesis suggests that only non-native species being able to take up the associated costs of immunity will be successful invaders.
The role of immunity in invasion success has yet been poorly investigated, mainly because of the difficulty to simultaneously analyse multiple immune pathways [4]. Charbonnel and collaborators [2] overpass this difficulty by screening all genes expressed (using a whole RNA sequencing approach) in an immune tissue: the spleen. They do so along the invasion routes of two sympatric invasive rodents in Africa and compare anciently and newly invaded areas (respectively). For one of the two species (the house mouse), they found a high number of immune-related genes to be up-regulated in newly invaded areas compared to anciently invaded areas. All categories of immune pathways (costly and cost-effective) were up-regulated, suggesting an overall increase in immune investment in the mouse, which corroborates the “immune protection” hypothesis. For the black rat, patterns of gene expression were somewhat different, with much less pronounced differentiation in gene expression between newly and anciently invaded areas. Among the few differentiated genes, a few were associated to immune responses and some of theses genes were even down-regulated in the newly invaded areas. This pattern may actually corroborate the EICA hypothesis, although it could alternatively suggest that stochastic processes (drift) associated to recent decrease in population size (which is expected during a colonisation event) are more important than selection imposed by pathogens in shaping patterns of immune gene expression.
Overall, this study [2] suggests (i) that immune-related traits are important in predicting invasion success and (ii) that two successful species with a similar invasion history and living in similar environments can use different life-history strategies to reach the same success. This later finding is particularly relevant and intriguing as it suggests that the traits and strategies deployed by species to colonise new habitats might actually be idiosyncratic, and that, if general trends actually emerge in regards of traits predicting the success of invaders, the devil might actually be into the details. Comparative studies are extremely important to identify the general rules and the specificities sustaining actual patterns, but these approaches are yet poorly used in biological invasions (at least empirically). The work presented by Charbonnel and colleagues [2] calls for future comparative studies performed at multiple spatial scales (native vs. non-native areas, anciently vs. recently invaded areas), multiple taxonomic resolutions and across multiple traits (to search for trade-offs), so that the success of invasive species can be properly understood and predicted.

References

[1] Jeschke, J. M., & Strayer, D. L. (2006). Determinants of vertebrate invasion success in Europe and North America. Global Change Biology, 12(9), 1608-1619. doi: 10.1111/j.1365-2486.2006.01213.x
[2] Blossey, B., & Notzold, R. (1995). Evolution of increased competitive ability in invasive nonindigenous plants: a hypothesis. Journal of Ecology, 83(5), 887-889. doi: 10.2307/2261425
[3] Charbonnel, N., Galan, M., Tatard, C., Loiseau, A., Diagne, C. A., Dalecky, A., Parrinello, H., Rialle, S., Severac, D., & Brouat, C. (2019). Differential immune gene expression associated with contemporary range expansion of two invasive rodents in Senegal. bioRxiv, 442160, ver. 5 peer-reviewed and recommended by PCI Ecology. doi: 10.1101/442160
[4] Lee, K. A., & Klasing, K. C. (2004). A role for immunology in invasion biology. Trends in Ecology & Evolution, 19(10), 523-529. doi: 10.1016/j.tree.2004.07.012