Uploading your talk
Link to upload talks: https://www.dropbox.com/request/tpb5jBTRXaLXTlWKViMr Name your file as: TalkNumber_Firstname_lastname
Schedule
Day 1
Session 1 – Chair: Irene Mendoza
Keynote talk: Network science: Applications for sustainable agroecosystems and food security
Presented by: Prof. Darren Evans
The global challenge of feeding two billion more people by 2050, using more sustainable agricultural practices whilst dealing with uncertainties associated with environmental change, requires a transformation of food systems. As part of a Royal Society Challenge-led Grant, which established a consortium of researchers in the UK and South America, I present our perspective for how advances in network science can provide novel ways to better understand, harness, and restore multiple ecological processes in agricultural environments. I describe: (i) a network-focused framework for managing agro-ecosystems that accounts for the multiple interactions between biodiversity and associated ecosystem services; (ii) guidance for incorporating socio-economic factors into ecological networks; and (iii) the potential to upscale network methods to inform efforts to build resilience, including global food-supply chains. In so doing, I aim to facilitate the application of network science as a systems-based way to tackle the challenges of securing an equitable distribution of food, suggesting novel ways in which network scientists can get involved.
Talk 1: Merging molecular data into ecological networks to assess ecosystem services across agricultural habitats
Presented by: Jordan Cuff
Our food systems depend on the provision of ecosystem services such as pollination and pest management. These services in turn depend on the management of habitats adjacent to crops which act as a reservoir for pest and beneficial arthropods. To balance ecosystem services against the disservices performed by pests, we must first understand the impact of different semi-natural habitat margin structures and types on their provision. Many of these interactions are difficult to observe or identify, so we must extend beyond traditional approaches to investigate the complex networks that exist within and between agricultural habitats.
Using a national-scale survey of oilseed rape fields in the UK, we are constructing networks using a combination of traditional and molecular approaches to assess ecosystem service provision within and between semi-natural habitats and adjacent crops. We surveyed crops with grassland, hedgerow and woodland margins across a land-use intensity gradient to identify interactions between pollinators, plants, herbivores and natural enemies. We recorded interactions observed between plants and arthropods in the field, and individual arthropods and community samples were collected for molecular analysis. Using multi-marker DNA metabarcoding and high-throughput sequencing, we have identified the ecosystem service and disservice providers present across crop and semi-natural habitats. I will present our findings and the pitfalls and promises of merging traditional and DNA-based data for constructing highly-resolved networks to assess ecosystem service provision.
Talk 2: The effect of coevolution on the robustness of communities to co-extinctions
Presented by: Fernando Pedraza Perez
Species are going extinct at an alarming rate. The loss of biodiversity may be further amplified by species interactions. This puts communities at risk of co-extinction cascades. Yet, we still ignore all the factors that shape the robustness of communities to the loss of species. Here we analyse how coevolution influences the robustness of mutualistic and antagonistic communities. We find coevolution increases robustness in mutualism but reduces it in antagonism. Communities become more robust because coevolution increases their density of interactions. The largest changes to robustness occur when coevolutionary selection is strong. Yet, the effect size of coevolution on robustness depends on the size of the community. Our theoretical findings suggest that coevolution is a potential factor that affects the robustness of communities. These insights may inform efforts to reduce the risk of species loss in communities.
Session 2 – Chair: Maria Martignoni, Jordan Cuff
Talk 3: Using accessible network approaches to quantify ecosystem service flows in a smallholder agricultural – tropical forest landscape in Papua New Guinea
Presented by: Stanworth Anna
Ecosystems services are under increasing stress worldwide. Understanding how ecosystem services are provided throughout landscapes allows for better informed assessment and management. Network approaches provide detailed insights into ecosystem service provisioning by elucidating how ecological interactions drive ecosystem functioning and, in turn, ecosystem service flow through landscapes. However, challenges exist in sampling ecological interactions, limiting effective use of network ecology in theoretical and applied ecosystem service contexts. Collating suitable interaction datasets for analysis and application is time and labour intensive, relying on human observation over long periods. Other methods including camera trapping or DNA analyses are also financially demanding. One way to combat this inaccessibility of network methods could be local and indigenous ecological knowledge, using interviews to collate information about local ecological interactions. Here, we will present data on ecosystem services in a subsistence farming community in Papua New Guinea, collected using interviews with local farmers. Information about local ecological interactions and ecosystem service values will be used to build ecological networks representing the flow of ecosystem services (e.g. pollination, seed dispersal) which contribute to food and livelihood provision. Using local knowledge to value and assess ecosystem services through ecological networks is novel but has the potential to increase the use of network approaches in ecosystem services contexts. Analysis will determine network robustness (and subsequent resilience of ecosystem service provision), and the effectiveness of this sampling technique in building informative ecological networks.
Talk 4: Variation in host-microbiome network structure across land use
Presented by: Matan Markfeld
Anthropogenic land use change degrades natural habitats and alters wild microbial communities
including the host microbiome. The host microbiome plays a major role in the host’s health and disease states. Hence, the compositional and structural changes of individual hosts’ microbiomes induced by land use change can affect disease dynamics at the host population level. Here, we ask what is the effect of land use on the host-microbes structure. We focused on the gut microbiome, sampled from rats in rural Madagascar across habitat degradation gradient, encompassing primary and secondary forests and different agricultural crops. We found no change in alpha- and beta-diversity across land use at the host individual level.
However, when exploring the host-microbes network at the meso-scale, using community detection methods, we found a weak signal of land use in the network structure. This pattern is potentially driven by rare microbes that occur in specific land uses. These results shed more light on the relative effect of land use on the host microbiome and enable a better understanding of the factors that shape the host microbial community.
Talk 5: Functional traits of birds and propagules are important in waterbird-mediated dispersal networks
Presented by: Ádám Lovas-Kiss
The destruction and degradation of wetlands in recent years have made it harder for water-bound organisms to disperse. To build up the first detailed waterbird-mediated dispersal network, our research investigates the dispersal interactions of plants and invertebrates by waterbirds in a given area. Our aim was to detect external (epizoochory) and internal (endozoochory) propagule dispersal by waterbird species in the study area and to map relationships among species. Since spring 2020 we have collected 924 faecal samples from 21 bird species. We found that endozoochory plays a much bigger role in propagule dispersal than epizoochory. 16.67% of faecal samples contained at least 1 intact propagule, including a total of 735 plant seeds and 769 invertebrate eggs. The plant dispersal network was modular and nested, while the invertebrate dispersal network was only nested. Several traits of dispersing birds influence their role as dispersers. Body mass, beak size, tarsus length and feeding method of dispersing birds influence the identity of the propagules dispersed, their abundance, and different network metrics. Our networks improve our understanding of the dispersal of propagules by waterbirds, including weeds and invasive species, and allow us to make recommendations for site management.
Talk 6: Maintaining ecological stability for the sustainable economic yield of multispecies fisheries in complex food webs
Presented by: Alexandra Werner
More and more fish stocks are being (over-)exploited in an effort to feed the growing world population. As these species are embedded in complex food webs, single-species management plans must be replaced with models integrating multispecies fisheries, economic market feedbacks, and fisher behaviour into complex ecological interaction networks to promote sustainable resource use. Here, we develop such a dynamic model containing three open-access fisheries in a complex food web. We find that selectively choosing similar species is more beneficial than the balanced harvesting of species at different network positions. Targeting low or high trophic levels risks reducing basal biomass or unchaining trophic cascades, respectively, which undermines first ecological stability (food-web biomass and persistence) and then the sustainability of economic yield (sustained total catch and revenue). Win-win situations with high sustainable economic gain and low negative ecological impact arise when similar mid-trophic level species are caught in the multispecies fishery. Complex system analyses are thus an important tool for combining ecological stability and multispecies fisheries to achieve a sustainable food supply for the world.
Session 3 – Chair: Alexandra Werner, Johannes Nauta
Keynote talk
Presented by: Prof. Laura Dee
Talk 7: Detection of rapid shifts between top-down and bottom-up regulated phases through temporal dynamics in predator-prey interactions and intraguild predation
Presented by: Pedro Leote
Species communities are constantly shifting in response to resource abundance and environmental changes. For consumers, and predators in particular, these resources are prey that can either limit the population’s density or be limited by it. Such shifts from bottom-up to top-down and vice-versa are likely to occur often, providing stability and resilience to the community. Despite this, competition among predators and intraguild predation can modify the interaction strength of predators with herbivore prey, as competitors constrain their foraging behaviour to avoid competition and/or predation, but this has yet to be demonstrated in a field setting. These two factors can affect the timing of regime shifts, destabilizing the system and compromising its function. To assess this, we have studied predator-prey interactions and intraguild predation in agricultural invertebrate food webs, with generalist predators, pest and alternative prey. We experimentally manipulated prey availability to determine how alleviating competition affects regime shifts and ecosystem functioning. Moreover, to find how such alleviation affects shift timing, we conducted our sampling at a high temporal resolution throughout entire seasons for two years. This will allow us to determine the shifting frequency and pinpoint the transition from bottom-up to a top-down regulation, with direct implications for the system’s stability, and the provisioning of ecosystem services such as biological control.
Talk 8: Invasive species modulate the structure and stability of a multilayer mutualistic network
Presented by: Agustin Vitali
Species interactions are the backbone of the structure and dynamics of communities. The extensive research into the link between structure and stability has been primarily theoretical and focused on monotrophic networks. Therefore, how the disruption of multitrophic interactions alters communities` response to perturbations in nature remains an open question. Here, we explored how non-native ungulates affect pollination-seed dispersal multilayer networks in Patagonia, Argentina. Ungulates disrupt a hummingbird-mistletoe-marsupial keystone interaction, which alters community composition. We calculated interlayer connectivity, modularity, and species’ roles in connecting modules for intact vs. invaded networks. To link structural changes to stability, we quantified network tolerance to a single random species removal (disturbance propagation) and sequential species removal (robustness) using a stochastic coextinction model. Non-native ungulates reduced the connectivity between pollination and seed dispersal and produced fewer modules with a skewed size distribution. Moreover, species shifted their structural role, primarily from connectors to peripherals, thereby fragmenting the network by reducing the “bridges“ among modules. These structural changes altered the dynamics of cascading effects in the community, increasing disturbance propagation and reducing network robustness. Our results highlight the importance of understanding the mechanisms that alter the structure and subsequent stability of multitrophic communities in nature.
Talk 9: Community and network ecology of biting Diptera-host interactions
Presented by: Talya Hackett
Biting flies (Diptera) transmit a variety of human and animal pathogens, and while most species are generalist in host choice, some display a consistent preference at the class or species level, with important implications disease transmission. Biting Diptera-host interactions can now be reliably identified from blood meals; however, most research has targeted individual species and specific hosts. Here we investigate the community ecology of biting Diptera–host interactions determined through DNA metabarcoding of blood meals. Initially, using interactions derived from the literature, we demonstrate how DNA metabarcoding and ecological network analysis can answer questions on biting Diptera community composition, structure, and associated disease transmission risks, and how these might change in response to perturbations. We also examined latitudinal and local anthropogenic habitat modification effects on network structure. Finally, using data from collected biting flies in Ghana, we looked at the effect of anthropogenic landscape modification and proximity to human habitation on biting Diptera community composition and network structure. We identified interactions between competent disease vectors and susceptible hosts, and thus potential disease transmission events. Our findings highlight how ecological networks combined with metabarcoding of blood meals can provide insights into anthropogenic landscape modification impacts and the risk of established and emergent disease transmission.
Day 2
Session 4 – Chair: Isabel Donoso, Noa Kan
Keynote talk: The evolution of virulence: insights from digital coevolution
Presented by: Dr. Miguel Fortuna
The reduction in host fitness caused by a parasite is called virulence. Trade-offs between different component of parasite fitness provides the dominant conceptual framework to understand why some parasites cause high levels of damage to their hosts whereas others are relatively benign. The trade-off theory assumes that if a parasite is very virulent and prolongs infection, it will increase the likelihood of host death before its own replication and then it will not spread through the population. Surprisingly, empirical evidences of trade-offs are uncommon because of the difficulty of conducting the appropriate experiments. Digital evolution is a complementary approach to experiments in the lab. In this computational framework, self-replicating computer programs, called digital organisms, compete to each other for limited resources (computer processor time and memory space), mutate and evolve within a user-defined computational environment. Some of them (a kind of computer viruses) operate inside others by matching their genetically encoded phenotypes, stealing CPU cycles from them to execute their own genome’s instructions and, hence, reducing their host fitness. By performing in silico coevolutionary processes that are analogous to the coevolution that takes place between bacteria and their phages we test experimentally the existence of trade-offs in the evolution of virulence.
Talk 10: Experimental testing of the network structure-stability relationship
Presented by: Grégoire Proudhom
How diversity is maintained in ecological networks and what are the factors influencing it is one of the key questions in ecology. One of these factors is network complexity, and regroups the size and the arrangement of the species interactions in the network. Many network indices have been created to better understand which exact component of network complexity is the most important. Among these indices, we chose to consider nestedness and modularity as previous studies show both could be important drivers of stability. Here we used a host-parasitoid system to build different communities in mesocosms, varying in modularity, nestedness and species composition and exposed them to heat waves and invasions to try to get a better understanding of this structure stability relationship. We kept communities for multiple generations and sampled them regularly to get snapshots of the community composition and interactions. We show that our communities respond differently depending on the treatment and their structure, with little effect of the species composition.
Talk 11: Using network centrality measures to predict the outcome of gene drive deployment
Presented by: Keith D. Harris
Gene drives are genetic constructs with super-Mendelian inheritance that can spread deleterious alleles in wild populations. Gene drives could potentially be used to suppress or eradicate disease vectors, crop pests and invasive species, with deployment programs expected to commence within a decade. However, prior to deployment, it is crucial to understand how gene drives are expected to spread in order to design safe deployment programs. Previous gene drive models that investigated gene drive spread focused on simple population structures, either in homogenous continuous space or in two-population models. In order to study the behavior of gene drives in more complex population structures, we developed a discrete-space network-based model of gene drive spread with an arbitrary number of populations that are connected by arbitrary gene flow patterns. Under this framework, we studied the relationship between the properties of the network and the release site, and the properties of gene drive spread. By using different generative network models, we studied the effect of network topology on gene drive spread, and identified centrality measures of release sites that can predict the outcome of deployment. Our results demonstrate that population structure can crucially alter gene drive spread dynamics. In addition, we highlight centrality measures that are correlated with specific deployment outcomes. Our results identify the key aspects of population structure that should be measured in wild populations that are currently being considered as candidates for gene drive deployment programs, and also demonstrate the ability to develop predictors of deployment outcomes based on ecological factors.
Talk 12: An Agent-Based Model to understand the processes that structure mutualistic networks in Doñana
Presented by: Javier Galeano
Pollinator-plant interactions are of enormous importance in ecology, being one of the most important types of biotic interactions on our planet. Once the community has been characterized, the next step in the study would be to have a mathematical model to be able to predict the global image of the interactions based on some simple working hypotheses. For this, the simulations of agent-based models can be a first step to understanding what are the basic elements that can give us this complex ecological community. These types of models have the advantage that with simple rules in the interactions of the agents, it is possible to obtain emergent behaviour.
In this work, data on pollinator-plant interactions have been collected, obtained in a study that was carried out in an area of the Doñana National Park, on the Atlantic coast of southwestern Spain. Agent-based models have two different classes of agents: plant individuals and pollinators. Different scenarios were designed where the spatial distributions of the plants were varied. For pollinators, the distributions were made randomly within each of the plots studied. The initial abundances of the agents that represent the plants were obtained from the real data, while the abundances of the pollinators were made by defining two types of roles: generalists, more abundant and with a greater search field, and specialists, with less abundance and small range. The movement of the pollinators was simulated with two types of movements.
Session 5 – Chair: Shai Pilosof
Keynote talk: Feasibility and stability of ecological communities
Presented by: Prof. Stefano Allesina
The structural stability of an ecological dynamical system is measured as the range of parameters leading to coexistence. In the past, researchers have studied the case in which the environmental conditions change, resulting in independent, identically distributed perturbations to the parameters. Naturally, parameters such as growth rates and interactions strength are likely to change in concert whenever environmental conditions are altered. We introduce a toolbox to study correlated changes to the parameters dictating population dynamics, and use it to study how different network structures impact the structural stability of ecological networks.
Session 6 – Chair: Sam Ross
Keynote talk: The architecture of multifunctional ecological networks
Presented by: Prof. Anna Traveset
Understanding how biotic interactions affect ecosystem functioning and ecological resilience has been a long research goal in the natural sciences. Yet, traditional assessments of ecological complexity generally focus on species-species interactions that mediate a particular function (e.g. pollination or seed dispersal), incompletely integrating the multiple other functions that underpin ecosystem multifunctionality. While ecological network theory holds a high potential to explore ecosystem multifunctionality, this potential remains poorly materialized, mainly due to challenges in measuring different interactions and establishing their relevance across multiple functions. Such lack of quantitative studies therefore limits our ability to determine which species and interactions are more important to maintain the multiple functions of ecosystems. Here we develop a framework, derived from a resource-consumer-function tensor analysis, that bridges these gaps by framing biodiversity-ecosystem multifunctionality in terms of multilayer ecological network theory. Its application to recently collected ecological data reporting weighted interactions between plants, animals and fungi across multiple function types allows to (i) unveil and quantify the existence of both (multi-functional) keystone species and a dual function keystoneness pattern, and (ii) project plants and functions into a similarity space where clear clusters emerge and the importance of weak links is manifested. This dual insight from species and functional perspectives will better guide conservation efforts to halt biodiversity loss.
Talk 13: Lessons using network ecology as an empirical predictor of ecosystem health in the search for solutions to real world problems
Presented by: Oskar Rennstam Rubbmark
Understanding how animals interact with their environment and how humans impact these animals, is key to understanding how we can apply our knowledge of ecosystems to real world problems. Which, is something that is especially true if we want prepare the diversity of animals that provide the functions that we rely on for human food production for future climate change.
However, even though we over the years have developed a pretty good general understanding of how to use the diversity available in agricultural landscapes to provide these functions, we have a relatively poor understanding of how these ecosystems are regulated over shorter time scales. This short term regulation is however key to develop a good understanding of which mechanisms it actually is that we should manage, when we try to support the resilience or functional contribution of service providers like the pollinators, generalist or specialist predators that farmers need in these landscapes.
Here we provide an overview of the research that we have been doing on how to use trophic interactions and network ecology to detect the signatures of well functioning ecosystems with a special focus on agricultural systems. Arguing that how well these systems work perhaps only indirectly depend on the diversity within these field or landscapes that can contribute (e.g. functional redundancy), but more directly on “how this diversity behaves” (e.g. how quickly and completely service providers adapt and include pest or flowering crops in their diets).
Talk 14: Structural dynamics of plant-pollinator mutualistic networks
Presented by: Albert Solé Ribalta
The discourse surrounding the structured interaction of species within mutualistic communities predominantly centers on modularity and nestedness. The former is known to enhance the stability of communities while the latter is related to their feasibility, albeit compromising the stability. But, their joint structural emergence poses challenges and consequently limits the inheritance of their respective dynamical properties. We hypothesize that compound structures, combining modules with internal nested organization, may offer valuable insights in this debate.
Analyzing the temporal structural dynamics of 20 plant-pollinator interaction networks and find that compound structures are highly prevalent during the peak of the season with approximately 50% of the communities modifying their predominant structural pattern throughout the year. Motivated by these empirical findings emphasizing the temporal plasticity of plant-pollinator networks, we synthetically investigate the dynamics of the structural patterns observed in the data across two control parameters—community size and connectance levels—mimicking the progression of the pollination season.
Our analysis reveals contrasting impacts on the stability and feasibility of species’ interaction networks. We characterize the consistent relationship between network structure and stability, which follows a monotonic pattern. However, in terms of feasibility, we observe non-linear relationships. Compound structures exhibit a favorable balance between stability and feasibility, particularly in middle-sized ecological communities, suggesting they may effectively navigate the simultaneous requirements of stability and feasibility. These findings allow us to resolve that the assembly process of mutualistic communities may be driven by the delicate balance between multiple properties, rather than the dominance of a single property.
Day 3
Session 7 – Chair: Lyle Poley
Keynote talk: Exploring the role of groups for functioning and ecosystem management in ecological communities
Presented by: Prof. Anna Eklöf
Processes and functions in ecosystems depend on how species interact with each other. Therefore, by identifying broad-scaled interaction patterns, important information on species’ ecological roles may be revealed. The group model, in network science more generally named stochastic block model, is a useful tool for identifying and understand how groups of species share interactions across several trophic steps. Here I will give a brief overview of how we, and others, have used the group model to explore ecological relevant groups of species in complex ecological networks. I will then particularly focus on analyzes of i) how the group structure differs between different subregions in an ecosystem, ii) how the solution landscapes differ between different food webs, and iii) how sensitive the group structure is to uncertainties in interaction data. I will highlight both possibilities and limitations and point out important future directions for how the group structure of networks can guide us to a better understanding of ecosystem’s functionality.