Schoener 1957 and Bailey at el. 2019
Blog Author: Annie Madsen
SCHOENER 1957
Main question: What is the currency of feeding strategies (what is maximized/minimized)?
4 strategies to optimize foraging:diet, space, period, group size
Optimal diet (M), a set of one or more items, can be predicted with a simple equation:
E = net energy
T = net feeding time
E/T = (potential energy-pursuit costs-handling and eating costs)/(pursuit time + handling and eating time)
This case does not incorporate search time (ambush predators or opportunistic foraging).
When considering search time, E/T is linear if: resources are equally distributed and replenish before predator returns, predator has complete knowledge of foraging grounds, and there is no prior investment required (travel/building a web).
Energy gain at different life history stages is relevant because energy allocation differs at those stages, especially when more/less energy intake during reproductive periods alters fitness. Animals must be able to gain energy above the baseline existence energy (the amount of energy required for basic cellular maintenance) to be able to allocate energy to storage, growth, or reproduction. Based on the literature, Schoener reviews how to calculate maintenance energy using metabolic rate and caloric content of food. He then discusses how animals can manage intake and allocation of energy through behavior and physiology. Schoener also mentions animals have to spend time to find mates and avoid predation, which compete with time spent foraging.
Using this background, Schoener presents a model (appropriately named the "general model") for optimal foraging that maximizes reproductive output. There are two special cases for this model: time minimizers and energy maximizers.
Optimal Diet
Schoener discusses the parameters that are usually incorporated into optimal diet models, reviewing equations from Laing, Holling, and others. The initially simple equation presented above is expanded to incorporate the specifics behind pursuit time (speed, distance, and fatigue), search and pursuit energy (metabolism and activity), handling and eating time (size of feeder, size of food, move to different location), pursuit and capture success (ratio of success:attempts), potential caloric content, and relative abundance of food items. The models vary from mechanistic representations of muscle movement to the different types of prey items. MacArthur and Pianka's equations regarding competition are also considered here. Two polarizing types of foragers are discussed: specialists vs. generalists (generalists favored during fluctuating prey abundance) and large vs. small prey. The feeding rates section addresses the functional response and maximizing fitness based on density.
Optimal Foraging Space
This section considers space and time as the relevant parameters, including home range, path of movement, and patch preference. Schoener first discusses how to estimate home range, which is a complicated enough problem that needs to consider food density, food preference, and metabolism. The path of movement is essentially the maximum distance a predator will move to eat a prey item and the patch preference perspective considers patch quality rather than item type. Here, Schoener mentions (without using this term) area-restricted search (ARS). He also mentions competition here in the form of territoriality.
Optimal Foraging Period
He says very little here, basically saying that no theory has been developed and no significant empirical work has been done.
Optimal Foraging Group Size
Foraging efficiency may decrease with group size if intraspecific competition increases beyond a threshold or if small costs of competition (or benefits of dilution effects) results in even larger group size. Conversely, foraging efficiency may not change with group size because animals distribute themselves equally across the prey distribution. Group size can increase efficiency through behaviors like flushing and pack hunting. Territoriality may again come into play here if the group cooperatively defends a patch. Community foraging (e.g., mixed species flocks) is also beneficial if there is little overlap in foraging preference among group members.
BAILEY ET AL. 2019
Bailey et al. present evidence for a specific case in optimal foraging literature: ARS. This foraging strategy is a non-random search pattern that involves biasing time spent in foraging patches based on encounter rates with prey items. It allows predators to forage more efficiently by spending more time in high density food patches. The authors tested whether two species of dolphins exhibited ARS using two hypotheses: Feeding increases feeding (positive feedback), foraging time increased in a patch after first encounter. They also looked for interspecific differences.
Using recording devices to detect increased echolocation use by dolphins (a reliable measure representing prey encounters), the authors defined separate encounters and used a Gaussian Mixture Model (GMM) to distinguish foraging clicks from other types of clicks. This type of model basically uses machine learning techniques to identify the foraging behavior instead of using an arbitrary cut off interval or eyeballing a spectrogram (this technique is also used to define foraging bouts in bird feeder studies; if anyone is interested, look up Wytham Woods studies). The "behavioral state transition" analysis tested whether feeding behavior led to more feeding behavior with any significance by comparing different behavioral states. The encounter duration was tested by defining a foraging patch, determining when the dolphins started foraging in said patch, and when they left the patch.
The behavioral state transitions were not significant. Dolphins were more likely to end a foraging bout (and therefore leave the area) if the foraging activity greatly increased between the first and last halves of the encounter.
The authors keep toting the novelty of their marine species study, which is a cool example, but I was hoping for other broader impacts for optimal foraging theory. I wish they had incorporated more social/spatial memory into the analysis, especially because they mentioned the impact of memory on ARS in the intro. However, I liked the use of localized interactions to model ARS because foraging is generally a mix of individual encounters, and the incorporation of pairwise associations was a welcome representation of the social interactions of a very social species. The authors go on again in the discussion to suggest that there is evidence to support ARS because of the likelihood to spend more time foraging in area after the first prey encounter. ARS is expected to occur when prey is not equally distributed across a patch (one assumption of classic optimal foraging models), which is consistent with their primary prey's schooling behavior. There may have also been seasonal variation in prey density that could have altered results.
Comments
I thought Schoener's approach to energetics was approachable because he eased us into the calculations with a well set-up background that was both relevant and easily digestable. I enjoyed how he integrated life history into energetics, which was consistent with how energetics has been presented to me in past lectures. However, the ending was a bit abrupt for me. I wanted a quick wrap-up at the end with some future directions and broader impacts. I was overall disappointed in the Bailey et al. paper because the intro set the story up to be really amazing, but I was a bit let down with the results and discussion that did not seem to broaden these results out to any other species or system. It left me wondering how much effect the charismatic species had on the publication process.
Schoener’s work builds on MacArthur and Pianaka’s foraging theory presented in their 1966 paper. While MacArthur and Pianka’s work was general, Schoener presented more fine-tuned equations and often noted species that were exceptions to the models in the paper. The discussion of “specialists versus generalists” in Schoener’s paper was pretty unsatisfying. All three of the provided definitions for a generalist relied on the term “greater,” which is a subjective and relative term, so what defines a “generalist” depends on what species is used for comparison. The definitions also get a bit fuzzy when Schoener compared the differences in diets between the young and the adult members of a species.
ReplyDeleteThe Bailey paper was interesting because of the focus on a marine system, which is a system that most of the classic papers that we’ve read in class tend to ignore (although this is understandable given the limitations to studying this system in the past). When accounting for the area-restricted search in the ocean, the depth of the ocean is another factor to consider because in terrestrial systems (excluding predators that hunt prey in the air), the area that is searched is typically defined by the geometric area (length x width). I am not used to working in field-based systems, but I was surprised at how little data the researchers were actually able to collect. Between the four sites sampled between November 2014 and April 2017, only 995 minutes of foraging behavior was recorded (Bailey did acknowledge that this value was likely an underestimate). When considering that there were over a million minutes of data recorded during the time frame, the 995 minutes of foraging behavior seems incredibly small. Does this small sample size (in comparison to what was collected), introduce a bias in how the results of the study were interpreted?
The Schoener (1957) paper was a dense one, but very information. The paper did something we all seem to like which is refer to copious amounts of case studies. I personally liked how Schoener built upon the theories and equations of others (such as MacArthur and Prianka). I was left with some questions, though. Firstly, I am a little confused on the meaning of some variables, especially "L". I am not quite sure what L is supposed to represent (though I know it has something to do with reproductive fitness). Also, was it my misunderstanding or were some variables used twice to represent slightly different things (i.e M and P)? Additionally, the beginning of the paper seemed to downplay the role searching has in optimal foraging theory (grouping it with other non-feeding functions) but later it seemed to play a larger role. Lastly, when Schoener stated that larger animals have the potential to have more variety of food, is this an actuality?
ReplyDeleteThe Bailey et al. (2019) paper, as Annie says, started out strong but never seemed to live up to its own introduction. Firstly, (and I hope this is not just me being nit-picky), I would have liked to them report their p-value when reporting the results. Perhaps I missed it, but simply stating they cannot reject the null hypothesis isn't very informative (I think they put it in their Supplementary Info. but still). I am also a little confused about ARS after reading this paper. It seems the dolphins were returning the feed within minutes of capturing/ encountering prey. Would this be considered an ARS pattern or simply something else based on the nature of the prey they are capturing? This is something I would need clarified in class.
I thought that Schoener laid out a very thorough and interesting framework for thinking about the physiological demands that underlie feeding strategies for different organisms. Further studies have continued to debate the relationship between increased food consumption/intake and increased reproductive output that he hypothesizes. Is this change merely opportunistic and behavioral, as organisms find themselves in more better environments with more optimal food sources that allow them to increase their caloric intake and reproductive capacity? Or is this a form of microevolution – selection for larger individuals that have a higher food consumption and reproductive ability (allowing for larger litters, eggs/offspring size, and more litters per season)? I was able to follow most of Schoener’s arguments, but I had one question that I was curious about other people’s thoughts on – why, as he mentions in the section on optimal diet, would a decreased search speed (i.e., slower searching) be associated with a high food density? Wouldn’t an increased food density make foraging more efficient/less time intensive?
ReplyDeleteThe Bailey et al. paper aimed to address some very important and novel questions, but the methods the authors described seemed somewhat limiting, or at least poorly defined. The C-PODs they used to monitor the two dolphin species’ echolocation clicks might be very common in studies of marine mammal communication, but I have no idea how precise they are (and the website provided is also very vague other than mentioning they have “high precision”). For a large-scale study, is the deployment of only 4 C-PODS enough? They mention the C-PODs have a detection radius of 1.8 km, but can they accurately differentiate the various types of clicks within that entire radius? I also was curious how far apart were they placed, and how much their results were affected by what specific groups of foraging individuals they were able to detect (i.e., the sample size of individual dolphins’ behaviors within the larger population). As Crystal mentioned, I was also surprised at how little data they were able to obtain that was classified as foraging. Nevertheless, their results were interesting and informative for future monitoring, although the study left me wondering: does rate of encounter directly translate to successful predation, or just interaction with prey?
Although Schoener's work was rather dense, I did find his equations to be a bit clearer, providing a fairly straightforward, if sometimes brief, definition to the variables he used in constructing them. But considering how dense and simply how packed with information this paper was, I really would have wished there were some sort of discussion to synthesize what had been presented in the paper into a clear wrap-up. This would help both stylistically (a paper rather abruptly ending just reads a little strange) as well as aiding in the reader's takeaway from the paper by summarizing the main points, stressing the significance of those main points & of the work in general, and outlining future directions for the field of ecology. A paragraph's worth of conclusion would have helped a lot.
ReplyDeleteBailey et al gave an interesting applied study of area-restricted search behavior in a marine system, which fits nicely as a modern use of the ideas of feeding behavior presented by Schoener and other ecologists in the past. One small part I am uncertain of, as I am glancing back on the paper just now, is the significance of the 3 different colors used to notate different transitions in their Markov chain figure (Fig 2) - I may've missed a small detail, but this is not clearly explained in the figure description. While the experiment was interesting, I do agree with others that there were some disappointing bits, since the authors did not really expand out toward discussing broader significance of their results. However, it still was interesting to see foraging theory applied to a traditionally under-studied ecosystem.
- Elizabeth
Schoener’s paper presented models capturing animals’ feeding strategies under different circumstances. The author discussed efficient diets, foraging space, foraging period, and whether/not foraging by groups. I do like that the author listed some natural examples of the assumptions he made, such as hunger increased the range of foods, or predators preferred larger food if the distance increased. However, I think some of the situations he described, such as the time minimizer vs. energy maximizer (fig 1) is too nuance for me to come out of some examples in my mind and tell the difference (I understand they may be different mathematically, but I think it is rather 2 different perspective of the same spectrum).
ReplyDeleteThe area-restricted searching theory is interesting. Like others, I expected a very fascinating story after I read the introduction. Especially about one theory pointed there: the prior knowledge of potentially suitable area has stronger impact on ARS than prey capture. However, it does not seem to be addressed in this study. In fact, the authors spent an appreciable proportion of the discussion talking about some technical issues of detection. Also, I am not familiar with this area and I understand that this kind of experiment is really expensive, but I was still a little surprised that the authors tested the third hypothesis – the foraging pattern is consistent across species – based on the results from only 2 species.
I enjoyed these two papers, especially the newer companion paper by Bailey et al. In my own research, one of the things I'm interested in is how prey density shapes foraging behavior. The study by Bailey et al was a unique look at this for a predator-prey system that I don't know much about. However, like Annie, I was hoping for more coverage of the role of spatial memory in the behavior displayed by the dolphins, especially since dolphins are often touted as one of the smartest non-human animals. How long can dolphins remember where they successfully foraged in the past? This would be important information for determining whether ARS is happening or whether dolphins frequent the same prey-dense areas for other reasons (such as water temperature or predator avoidance).
ReplyDeleteI found Schoener's paper to be especially thought-provoking. I am curious as to how his predictions would change if he considered a complete food web, rather than just a predator with one or more prey species. For instance, if you consider that the predator itself may be at risk of getting eaten, you could include a probability of dying into the optimal foraging space components of his models. Perhaps an intermediate predator avoids a certain area - despite an abundance of energy-rich prey - because of the presence of higher level predators. Instead, the intermediate predator may be better off foraging on sparse, energy-poor prey in a safer setting.
Like others, I found the Schoener paper to be thought-provoking if a bit dense. I found it particularly interesting that he highlights independently the different dimensions of costs and benefits (time, energy, etc.) among which foragers must make trade-offs. He then demonstrates the need for considering foraging in an optimality context. I really enjoyed this train of reasoning because it obviously and immediately predates Charnov 1976, optimal foraging and the marginal value theorem. Reading the Schoener paper feels very much like watching history happen (I might be biased because I love both optimality theory and foraging).
ReplyDeleteI didn't enjoy the Bailey paper as much. I thought they had the potential to have a really interesting basic science question, but didn't really set it up to relate well to the larger body of theory or to any sort of bigger question. The modeling methods they used to discern foraging bouts were really interesting, but I didn't think that the actual connections to the behavior and ecology of the organisms were very robust. In particular, foraging effort does not necessarily correspond to foraging success in a linear way, so I did not find the argument that the dolphins were experiencing a depleting prey source, a vital component of ARS, to be convincing.
While I still don’t really understand optimal foraging theory, it’s clear that Schoener is a major synthesis of the OFT and it’s cool to see how much the field had developed since MacArthur and Pianka’s paper. I particularly appreciated how Schoener carefully linked the use of optimization models for energy or time budgets to fitness. The knowledge on optimal feeding period was clearly lacking development at the time, but I think that it has been more fleshed out since then by researchers using stochastic dynamic programming approaches. I don’t know if this work has been done, but it seems that it would be useful to tie the more recent and realistic physiological models (dynamic energy budget models) to link optimal foraging behavior to fitness.
ReplyDeleteI’m impressed with the Bailey et al. paper because of the methods they used to get past the challenges of studying dolphins, but because of these challenges I don’t think dolphins are a good study system for testing predictions of OFT. I also struggled to interpret the results of their null-hypothesis testing for whether dolphin behavior was Markovian and didn’t understand what the figures on encounter duration were supposed to show.
- David