Life cycle and ecology of the loggerhead turtle (Caretta caretta, Linnaeus, 1758): development and application of the Dynamic Energy Budget model

Marn, N. 2016
Life cycle and ecology of the loggerhead turtle (Caretta caretta, Linnaeus, 1758): development and application of the Dynamic Energy Budget model.
PhD-thesis VU University Amsterdam and Zagreb University


My main motivation for starting this journey, which resulted in (but does not end with) a doctorate of science, was to find out how much damage we are doing by allowing the plastic waste to enter the marine ecosystem. It is common knowledge that plastic takes a very long time to degrade; after all, the durability of plastic items is what made plastic so useful, and consequently, so ubiquitous! So, where exactly does all that plastic go? Does it sink? Does it just float in the oceans, swirling around in the ocean currents forever? Does it get ingested by marine organisms? And if so, what happens then?

There are reports of almost three hundred different species of marine organisms interacting with our plastic waste. Getting entangled in it, eating it, or using it as a transportation device to arrive to new habitats, where they sometimes thrive so successfully that they `squeeze out' native species. And while being attached to a piece of floating plastic to find a new ecosystem does sound like a promise of a fresh new start, being entangled by a discarded fishing net, or starving to death after eating too much plastic, certainly does not sound so inviting. The second scenario is, however, more common, and is the one experienced by sea turtles.

Sea turtles are remarkable creatures. They have existed in the form we see today for over 150 million years - this means they have coexisted with, and then by far outlived the large dinosaurs! They have fascinated humans from the early civilizations, but the fascination did not help them - all seven species of sea turtles that currently exist are on the IUCN list of endangered species, and most populations are declining despite the protection. The most vulnerable species for plastic are the long-lived ones, because their potential the adapt to changing environments across generations is most limited. Sea turtles fall into this category.

Loggerhead sea turtles are present throughout the temporal zone of all the world oceans, evolving into multiple populations and local subpopulations. They live longer than 65 years, and their sex is determined during the last third of their 60-days long embryonic development by the temperature during incubation. During their life they increase in size 25 times: from a 4 cm long and 20 g `heavy' hatchlings that exit the nest, to adults weighing over 100 kg at a carapace length of 100-130 cm that come to lay eggs at the same beach where they hatched. These two moments (hatching and nesting) are also the only two moments in their long life when loggerhead turtles have contact with the land environment. Consequently, beach and offshore (coastal) sea were for a long period the only two habitats in which loggerhead turtles could be observed. The remaining period - be it 5 years, a decade, two, three or more - they referred to as `the lost year(s)' (a term coined by Archie Carr in 1986), and it remained a mystery for a long time.

The advancements in science during several decades that followed made it possible to study the sea turtles and uncover many of the mysteries. A lot of studies have been performed, and a lot of literature has been published, but the focus of a study has most often been specific life trait or a specific life stage. Life cycle of loggerhead turtles could roughly be divided into three life stages: embryo, juvenile, and adult, but now it was possible to include observations about their ecology and define life-stages in more detail. Within the juvenile stage, one can differentiate between the hatchling (individual that has just hatched and is moving towards the open sea), posthatchling (a slightly older individual, up to 15 cm carapace length), oceanic juvenile (individual larger than 15 cm carapace length that mostly resides in the oceanic habitat feeding on plankton and other pelagic organisms), and neritic juvenile (individual larger than 30-50 cm carapace length that mostly resides in the neritic habitat feeding on benthic organisms). The transition from the oceanic to neritic habitats (assumed rapid and called the recruitment to neritic habitat) occurs for most individuals when they reach a certain size or developmental stage, but sometimes the transition is longer, or doesn't happen at all, resulting in adults feeding in oceanic habitats.

Due to the (i) different use of habitat, (ii) different sampling (such as taking different measures of carapace, and then devising expressions to translate between them; calculating growth rates from capture-mark-recapture data or growth marks visible on bones) and (iii) different analytical techniques (such as studying the change in length, or the change in mass, and fitting different growth models), reported data were not only disjointed, but were often conflicting. Most conflicts related to growth rates and growth models reported for different populations and life stages, lack of agreement whether to use the minimum or the average carapace length of nesting females within a population as `length at puberty', and the estimates of age at puberty ranging from 6 years to 38 years. Furthermore, several author pointed to significant differences between populations, most obvious being the size difference between Mediterranean adults compared to adults from other populations, but also differences present within a population, such as different growth rates and different expressions used to convert one measure of carapace length into another.

For me to solve a mystery of the `lost plastic' and its effect on the loggerhead turtles, I needed to know a lot more about the biology and ecology of loggerhead turtles: How long does it exactly take for a loggerhead turtle to mature? Why do some loggerhead turtles grow faster than others? Why are loggerhead turtles in the North Atlantic larger than those in the Mediterranean sea? Do larger turtles also reproduce more? Are size and reproduction results of environmental or physiological characteristics? How will loggerhead turtles cope with the global environmental changes? Or, more precisely, how do available food, salinity, or temperature influence processes such as growth, maturation, maintenance, and reproduction? And - if I want to know the effect of ingested plastic on those relevant processes fueled from the energy budget - how much energy does a loggerhead turtle daily need for the processes, and how much energy can it obtain?

Defining and following the energy budget of a loggerhead turtle was the most logical approach to take, one that would provide answers to most, if not all of my questions, as any effect of plastic ingestion on a species must become visible as an effect on the energy budget and/or life span. I chose the Dynamic Energy Budget theory was as the path to my `Holy Grail': The DEB model of a Loggerhead Turtle. It had everything: observance of the laws of thermodynamics, several types of homeostasis that any system (from cells to individuals and ecosystems) tries to obtain and keep, the effects of food and temperature onthe energy budget, the interaction of the energy budget with processes such as growth, maintenance, maturation, and reproduction. Additionally, it was and is the most consistent theory currently available.

Generally, mass is more informative than length when defining energetics, but as the same curve was successfully fitted for the relationship of length and mass across the whole size span of individuals from several different populations (Wabnitz and Pauly, 2008.), I focused on length. The reported differences in expressions for converting carapace lengths were my chosen starting point, because differences in conversion expressions for the same two types of measurements suggest that the shape of individuals differs among life stages and possibly even populations. The differences can have important implications for modeling the energy budget, as the shape (structural) homeostasis is one of the assumptions of DEB theory. Change in shape (deviations from isomorphy) can easily be accounted for by modifying the shape coefficient (dM), but first their significance needs to be analyzed. Focusing on the North Atlantic population for which the (inconclusive) difference in conversion expressions was reported, I compared the data from two different regions (`north' and `south') of the North Atlantic, and three different life stages (`posthatchlings and oceanic juveniles', `neritic juveniles', and `adults'). The results suggested that there are no significant differences when same life stages of different regions are compared, but that one should be careful when extrapolating shape-dependent conclusions from the smallest (`posthatchlings and oceanic juveniles') to larger life stages, and vice versa. Still, the noted differences in shape were not significant enough to require additional shape coefficients for different life stages, as the deviation from an ideally isomorphic organism was only around 5%. This conclusion implied that I can use the standard (simplest) form of the DEB model.

Developing and formalizing a complete life cycle DEB model of a loggerhead turtle was the second step. The standard DEB model describes an individual by following the dynamics of three compartments: structure, reserve, maturity, and (after puberty) the reproduction buffer. The first two (structure and reserve) can be indirectly measured as length and/or mass of an individual, whereas the third one (maturity) is formally quantified as the cumulative investment of reserve for increasing in complexity. The dynamics of each compartment is unique, and fully specified by the parameters of the model which are estimated simultaneously. The starting hypothesis was that differences between populations (North Atlantic and Mediterranean), and effects of plastic ingestion on the energy budget will be visible as changes in parameter values or changes in predictions of the DEB model. These values must, therefore, be determined first. The procedure of parameter estimation uses all available life-history data (such as length and age at birth and puberty), and other type of data (growth curves, reproduction output etc.) at the same time to arrive at the most realistic set of parameter values of the DEB model. Due to large variation within a single population, analyzing more than one population simultaneously was not a viable option. While focusing on the North Atlantic population - the largest (and probably the best studied) population of loggerhead turtles in the world, I obtained the values of all primary parameters of the DEB North Atlantic loggerhead model. The model had a very good fit with the observed data used as input, ranging from prediction for incubation duration, length and weight growth rates, to length at puberty, and reproduction output. Furthermore, by obtaining the parameter values that specify the whole life cycle of loggerhead turtles, I was also able to study the daily energy budget of the same loggerhead turtles. The results suggested that while the young posthatchlings use most of their energy for maturation and growth, a fully grown adult uses almost 3/4 of the energy budget for (somatic and maturity) maintenance. In addition, I could explore effects of mothers' feeding conditions on the embryo's energy budget: while at the food level resulting in the maximum food intake the embryo needs to use less than half of the initial energy in an egg for development and growth, at 20% lower food level it needs to use more than half. This directly translates to the amount of reserves (yolk sac) left at hatching, and thus possibly the survival of embryos.

However, as it usually happens, the predictions that did not have an excellent fit with data were the more interesting ones. Namely, while the duration of incubation was predicted reasonably well, the size at hatching was overpredicted. The age at puberty was predicted to be at the low end of the range reported for loggerhead turtles (around 13 years), whereas most of the more recent studies point to the high end of the reported range (20 or more years). Are the loggerhead turtles allocating to reproduction much sooner than is currently thought? Or was the assumption that the loggerhead turtles adapt to their environment (resulting in more or less constant conditions throughout their life-cycle), and grow similar to what the von Bertalanffy model predicts, an over-simplification? An interesting result was also the reported growth of posthatchlings, which appears to be faster than the model was able to reproduce. Is it possible that the posthatchlings grow faster because they have a metabolism even faster than the standard model predicts? This pattern has been recognized as ˇ°waste to hurryˇ± in species that need to grow fast even at the expense of wastefully using their resources. Evolutionary it would make sense that small posthatchlings (hatching during summer) maximize their growth to use the available resources and arrive to a larger size less appealing to predators, but faster growth, as well as the other noted peculiarities, may have been related to biases in the data, such as a higher food quality of posthatchlings compared to that of adults. Studying another set of data for a different population would therefore help to confirm or dismiss the hypotheses.

Studying the Mediterranean population, much smaller in several ways (smaller number of smaller individuals living in a smaller habitat), was the third step in my work. To obtain a first insight into the extent and possible reasons for reported size differences, size data (length and weight) from each population were analyzed and compared at two most distinct moments of the loggerhead turtle life cycle: hatching and nesting. The size of eggs was taken into account as well, as it has been previously found to account for most of the size difference between hatchlings of different populations. The average size of hatchlings and nesting adults was indeed substantially different between populations. Surprisingly, the ratios of the average weight and cubed length (the condition index) wasn't. The condition index did, however, differ between different life stages. I discussed various possible reasons for the size difference, from incubating environment of embryos to food abundance experienced by juveniles and adults. None of those pressures could, however, result in such large differences in size at nesting and simultaneously supported the observed reproduction output. The answer to that puzzle was revealed only by estimating the parameters of a DEB Mediterranean loggerhead turtle model, and studying the implications. Whereas maturing earlier at a smaller size in an environment with less food wasn't intuitive, the model suggested that the main explanation for it was the lower level of maturity that the Mediterranean loggerheads need to obtain to reach puberty. This implied two things: (i) the Mediterranean loggerhead turtles need to cumulatively invest less energy to reach puberty, suggesting they can reach puberty earlier and at smaller size than the North Atlantic ones, and (ii) the Mediterranean loggerhead turtles need to maintain a lower level of maturity (via maturity maintenance), suggesting a larger part of their energy budget can be allocated to reproduction. The predicted properties (such as size and age at puberty being smaller than, but reproduction similar to that of the North Atlantic loggerheads) were consistent with the observations, and the underlying mechanistic explanation was consistent with the DEB theory. The size of the hatchlings was slightly overpredicted, and the predicted age at puberty was close to the lower end of the range estimated by other authors, as was the case with the North Atlantic loggerhead turtles. The growth of posthatchlings, analyzed now in more detail and simultaneously for both populations, indeed confirmed a metabolic acceleration during the observed period. The same analysis also emphasized potential problems in analyzing the observed growth rates, as the faster growth of Mediterranean posthatchlings (compared to that of the North Atlantic ones) was evident only after the growth rates were calculated for a reference temperature and food Contents 5 level. In addition, using the same DEB model, I simulated a substantial change in food availability during the life of loggerhead turtles, and explored effects of the changes on growth. The resulting growth curve suggested biphasic growth, similarly to that proposed by very few authors while the others were using classic (monophasic) growth models. Biphasic or even polyphasic growth would indeed result in a greater age at puberty, consistent with the estimations at the higher end of the reported range, and is a pattern worth further exploring. Arriving at such a distinct growth pattern was interesting, but I was not sure whether just food was responsible for the differences, or should also temperature be included? And what exactly are the effects of one or the other on the whole energy budget and its underlying processes?

The most recent part of my journey, and the last part of this thesis, explores first independently and then simultaneously, effects of food and temperature on the energy budget. Experimentally, it is very difficult, if not impossible, to keep conditions completely constant throughout the life of a loggerhead turtle (65 years), and it is even more difficult to do this for as many turtles as is needed to study all the combinations of food and temperature we desire to test, hoping that loggerhead turtles in our study are good representatives of the species. One of many strengths of using a mechanistic modeling approach is precisely an opportunity to test such scenarios. Focusing again first on the North Atlantic population, I simulated a realistic range of food densities and temperatures experienced by loggerhead turtles. The effects of food density differences were present on growth rates, but were the strongest on the ultimate size of adults. The effects of temperature were most evident in the growth and maturation rates. Both environmental factors substantially affected the reproduction output. The length at puberty was hardly affected by either of the tested environmental factors, corroborating the conclusion of some authors that, even though variability in length at puberty is present, compared to age and decrease in growth rates (also suggested as indicators of attained puberty), it is one of the least variable observable properties. The results also consolidated the conclusions of an intrinsic (physiological) difference that allows the Mediterranean loggerhead turtles to reach puberty at a smaller size. The model for the Mediterranean loggerhead turtles was then used to compare the responses of Mediterranean and North Atlantic loggerhead turtles to the conditions present in the Mediterranean environment, and explore to what extent can organisms with different physiology respond to similar environmental conditions. Individuals of both populations are often encountered in the Mediterranean, and recently their growth and maturation rates in the Mediterranean have been reported separately for individuals of different origin. Results obtained using the DEB models were in agreement with the published results and conclusions, successfully reproducing the faster growth and earlier maturation of Mediterranean loggerhead turtles. In addition, it became clear why loggerhead turtles of the North Atlantic origin are generally not observed nesting in the Mediterranean, as the model predicted their reproduction output would be extremely low in the simulated environment.

Lastly, the global environmental pressure that has set all the wheels in motion - the anthropogenic debris and the effects of its ingestion on the energy budget - were studied. The effect on the energy budget was modeled in the context of Synthesizing Units, or more precisely, assimilation units (AUs) that are normally responsible for converting the ingested food into reserves and provide energy for all required processes (growth, maintenance, and maturation or reproduction). Simply put, the AUs can either be busy with processing particles (extracting energy from them) or free to accept new particles. When an increasing proportion of food particles becomes replaced by plastic (or other inert debris) particles, an increasing proportion of the busy AUs are processing particles that have no energy gain. First I assumed that the processing time of plastic particles is the same as that of food, and I quantified long-term effects resulting from ingestion of reported quantities. The reported proportion of stomach volume taken up by plastic was on average 3% of the stomach contents (range from 0 to 25%), percentage probably higher when the whole digestive system is considered. Then, having in mind that the gut residence time of plastic debris has been reported as a couple to several times longer than that of food, I simulated a proportion of ingested plastic of 3%, requiring more processing time. Therefore, first I simulated a range of realistic values of ingested plastic with the same residence time as food, and then I simulated a range of different residence time of ingested plastic taking up 3% of gut volume. The effect of ingested plastic, to my scientific excitement and moral dismay, turned out to be substantial. The ingested plastic effectively had the same consequences as reduction of food intake, resulting in slower growth (i.e. higher predation risk), smaller ultimate size, and a smaller reproduction output. When equal residence times were assumed, already 16% of volume taken up by plastic made it impossible for loggerhead turtles to reach puberty and reproduce. When a residence time three or more times longer than that of food was assumed, the same effect occurred already at a 3% volume proportion. In nature, the proportion of ingested plastic is not constant, nor do all the ingested particles have the same residence time. Equally realistic scenarios are (i) loggerhead turtles can tolerate a short (acute) exposure to a load even higher than 16% and recover, and (ii) ingestion of even smaller amount of debris will result in death by starvation, as the individuals normally ingesting more food have grown to a larger size, requiring more energy for maintenance, which now cannot be paid due to insufficient energy being available.

Completion of the work carried out as part of this thesis has provided many valuable insights. To my satisfaction, many questions have been answered, but also more have arisen - and those I wish to pursue further. For example, why was the size of the hatchlings consistently overpredicted? Can the `waste to hurry' pattern explain faster growth of the posthatchling, and does it hold for all populations of loggerhead turtles, or even all species of sea turtles that share similar environmental pressures? Is the combined effect of pelagic environment with lower food density and lower temperature, and the neritic environment with higher food density and temperature, resulting in a biphasic growth curve for most juveniles? Can these two patterns, one driven by metabolism (`waste to hurry') and the other by the environment (changes in food density and temperature) simultaneously explain the hypothesized polyphasic growth and the mismatch between the reported (20-30 years) and predicted (13-15 years) age at maturity? Would studying the effect of plastic ingestion on those growth curves predict an even larger age at puberty and more grim scenarios for the future of loggerhead turtle populations? And finally, could this improved understanding of biology and ecology of this magnificent species, and the detrimental effects that plastic waste has on our environment, explain why some loggerhead turtle populations are still declining despite the protection, and motivate us to change our behaviour?

Full text in pdf format

Presentation as PhD thesis award nominee of the Nederlandse Vereniging voor Milieuchemie en Milieutoxicologie (KNCV + NVT)