Incorporating demographic diversity into food web models: Effects on community structure and dynamics

Adaptive foragers and community ecology: linking individuals to communities and ecosystems

Natural ecosystems comprise an innumerable amount of different organisms. These organisms are not separated, they interact and depend on each other. Today’s ecosystems are facing an enormous decline in biodiversity due to human impacts with thus far unknown consequences. One key objective of ecological research is to understand the mechanisms generating and maintaining this incredible amount of diversity. However, comprehensive analyses of natural ecosystems are impeded by their complexity and diversity. Food webs, therefore, provide an excellent tool to analyze the complexity of ecosystems. They depict the system‘s diversity and species interactions in a condensed form. Furthermore, food-web structure can help to predict the interaction strengths between species and the energy pathways through the system. In my thesis, I use food web structure to analyze structural properties which separate food webs from other network types and furthermore I investigate generalities and differences of food-web structure across different ecosystems. 
\nOne of the most important ecosystems is the soil ecosystem, as it provides the base for aboveground productivity. However, detailed soil food webs are scarce. In chapter 2, I assembled the complex food webs of 48 forest soil communities and analyzed if soil food webs differ in their topological parameters from those of other ecosystems. I found that soil food webs are characterized by a higher number of omnivorous and cannibalistic species. Moreover, they comprise more trophic chains and intraguild-predation motifs than food webs from other ecosystems. Finally, soil food webs showed high average and maximum trophic levels. These differences in network structure to other ecosystem types may be a result of ecosystem-specific constraints on hunting and feeding characteristics of the species that emerge as network parameters at the food-web level. Despite these differences, soil food webs showed the same scaling of their properties with connectance and size. In a second analysis of land-use effects, I found significant but only small differences of soil food web structure between different beech and coniferous forest types, which may be explained by generally strong selection effects of the soil that are independent of human land use. This study has unravelled systematic structures of soil food-webs, extending our mechanistic understanding how their environmental characteristics determine patterns at the community level. Additionally, I have shown that the general scaling laws also apply for soil food webs.
\nIn addition to purely topological properties, I analyzed another important aspect of food webs. The distributions of body masses and degrees across species are key determinants of food-web structure and dynamics. In chapter 3, I analyzed body masses of species and their systematic distributions across food-web structure. In particular, allometric degree distributions combine both aspects in the relationship between degrees and body masses. They are of critical importance for the stability of complex ecological networks. I used an entirely novel global body-mass database including food-web structures of four different ecosystem types to analyze body-mass distributions, cumulative degree distributions, and allometric degree distributions regarding differences among ecosystem types. My results demonstrate some general patterns across ecosystems: the body masses are either roughly log-normally (terrestrial and stream ecosystems) or multimodally (lake and marine ecosystems) distributed, and most networks exhibit exponential cumulative degree distributions except stream networks that most often possess uniform degree distributions. Additionally, with increasing species body masses we found significant decreases in vulnerability in 70% of the food webs and significant increases in generality in 80% of the food webs. Overall, these analyses document striking generalities in the body-mass and degree structure across ecosystem types as well as surprising exceptions (uniform degree distributions in stream ecosystems). This suggests general constraints of body masses on the link structure of natural food webs irrespective of ecosystem characteristics. 
\nWhile I revealed general patterns of food-web topology in chapter 2 and 3, I investigated the drivers of these general patterns in chapter 4. Therefore, I analyzed the influence of different external factors on community (beta diversity) and food-web structure. Two main theoretical bodies explain β-diversity, the niche theory and neutral theory. However, neutral theory predicts only distributions for trophically identical species, whereas influences of local niches or neutral effects on food-web structure as a crucial part of the multitrophic structure of ecosystems are not taken into account. In chapter 4, I therefore analyzed the effects of spatial distance and environmental dissimilarity on the species dissimilarity (beta diversity) and food web dissimilarity (structural dissimilarity) of multitrophic forest communities. I showed that the mechanisms proposed by neutral theory can adequately predict the beta diversity of multitrophic species communities. Furthermore, food-web structure was robust and affected neither by spatial distance (random dispersal, neutral theory) nor by environmental filtering (niche theory). I additionally analyzed model food webs (random and niche topology) and compared their dissimilarities to empirical food webs. The highest dissimilarity was reached by random food webs whereas niche model food webs were in between and the lowest distances were expressed by empirical food webs. Further, random food webs displayed the highest mean trophic level (115), while niche model food webs showed lower (5) and empirical food webs the lowest (4) mean trophic level values. Hence, food-web structure appears to be energetically optimized with local species adapted to energetic niches within the food web while species identity within these niches remains random. This suggests that different species could be adapted to the same energetic niches and, while following random drift, still assemble into similar food web structures.
\nAltogether, the results of this thesis demonstrate the practicality of food-web structure in unravelling generalities across different ecosystems. Furthermore, food-web structure explains species distributions across the environment and provides additional important information on the ecosystem. 
\nThe observed generalities indicate constraints on food-web structure. The allometric degree distributions demonsrate such constraints on food-web structure by distributing the links in dependence of the species body masses. Finally, my results from chapter 4 indicate that, additionally to global topological constraints, local communities have to meet certain energetic constraints to explain the similarity found across food webs.

Human induced global change has greatly altered the structure and composition of food webs through the invasion of non‐native species and the extinction of native species. Much attention has been paid to the effects of species deletions on food web structure and stability. However, recent empirical evidence suggests that for most taxa local species richness has increased as successful invasions outpace extinctions at this scale. This pattern suggests that food webs, which represent feeding interactions at the local scale, may be increasing in species richness. Knowledge of how food web structure relates to invasive species establishment and the effect of successful invaders on subsequent food web structure remains an unknown but potentially important aspect of global change. Here we explore the effect of food web topology on invasion success in model food webs to develop hypotheses about how the distribution of biodiversity across trophic levels affects the success of invasion at each trophic level. Our results suggest a connectance (C) based framework for predicting invasion success in food webs due to the way that C constrains the number of species at each trophic level and thus the number of potential predators and prey for an invader at a given trophic level. We use the relationship between C and the proportion of species at each trophic level in 14 well studied food webs to make the following predictions; 1) the success of basal invaders will increase as C increases due to the decrease in herbivores in high C webs, 2) herbivore invasion success will decrease as C increases due to the decrease in the proportion of basal species and increase in intermediate species and omnivores in high C webs. 3) Top predator invasion success will increase as C increases due to the increase in intermediate prey species. However, it is not clear how the relative influence of trophic structure compares to empirically known predictors of invasion success such as invader traits, propagule pressure, and resource availability.

Food webs

Food web structure and interaction strength pave the way for vulnerability to extinction

Carbon flows in Baltic Sea food webs — a re-evaluation using a mass balance approach

Nonpoint source pollution entering rivers will pollute water quality, degrading the health of aquatic ecosystems. However, owing to the lack of quantitative research on the effects of nonpoint source pollution on the structure of aquatic food webs, there is a lack of quantitative basis for river management. Nonpoint source pollution is not only difficult to control effectively, but also the success rate of water ecological restoration projects is low. With the increasing proportion of nonpoint source pollution in water environmental problems, it is urgent to quantitatively assess and predict the impact of nonpoint source pollution on the structure of food webs. Therefore, this thesis presents a method for quantitatively assessing and predicting the impact of nonpoint source pollution on the structure of food webs through using fuzzy clustering to screen the typical points of the impact of nonpoint source pollution, then using canonical correspondence analysis (CCA) and partial least squares regression analysis to comprehensively filtrate the driving factors affect food web that results in nonpoint source pollution, and then determining the impact of each driving factor on the structure of food webs. Finally, the change trend of food web structure is predicted. The results show that (1) the driving factors that the nonpoint source pollution that affects the food web structure is NH3‐N and chemical oxygen demand (COD). The increase in NH3‐N and COD promotes the growth of phytoplankton, causing the change of the primary productivity of the ecosystem, and ultimately changes the entire food web structure; (2) NH3‐N and COD affect the stability, maturity, connectivity and complexity of the aquatic food web structure. The increase of NH3‐N increases the connectivity and maturity of the food web structure but reduces complexity and stability; the increase of COD increases the connection of the food web structure, while reducing the other three indicators; (3) in some areas with good water quality, aquatic species diversity is high, the relationship of interspecies dietary is complex, food web structure level index is high and the structure of food web is stable. The food web structure in the rainy season will be better than that in the dry season. In some areas with severe water pollution and poor food web structure, the ability of the food web to resist external interference is weak. The food web structure in the rainy season will be worse than that in the dry season owing to rainfall into the river. The methods and conclusions in this treatise can provide a reliable and quantitative scientific basis for river ecosystem management and ecosystem restoration and can improve the success rate of ecological restoration projects.

Climate change negates positive CO2 effects on marine species biomass and productivity by altering the strength and direction of trophic interactions

Summary Food webs and trophic dynamics of coastal systems have been the focus of intense research throughout the world, as they prove to be critical in understanding ecosystem processes and functions. However, very few studies have undertaken a quantitative comparison of entire food webs from a key consumer perspective across a broad geographical area, limiting relevant comparisons among systems with distinct biotic and abiotic components. We investigate the structure and functioning of food webs in four tidal ecosystems of international importance for migratory shorebirds along the East Atlantic Flyway: Tejo estuary in Portugal, Sidi Moussa in Morocco, Banc d'Arguin in Mauritania and Bijagós archipelago in Guinea‐Bissau. Basal food sources, shorebirds and their prey (benthic invertebrates) were sampled in all areas, and Bayesian stable isotope mixing models and community‐wide metrics were used in a comparative analysis among areas. Significant differences among study areas were found in the structure of food webs, as well as in the relative importance of basal resource pools supporting each food web. Overall, the food web of Banc d'Arguin was characterized by lower trophic diversity and higher functional redundancy than the other sites. This result might be explained by the low number of trophic pathways of organic matter transfer in this seagrass‐dominated system which, as a fossil estuary, lacks inputs from both freshwater and nutrient‐rich offshore oceanic waters. Structure of shorebird communities was consistent with the main organizational patterns found for each food web, highlighting the less diverse character of the community of Banc d'Arguin. At Banc d'Arguin and Bijagós archipelago, which displayed the smallest and largest isotopic niche widths in bird assemblage, respectively, mean niche overlap among species was low, suggesting high interspecific partitioning in resource use. Tropical systems typically offer comparatively lower harvestable prey biomass for shorebirds and might thus strengthen interspecific competition, leading to low niche overlap among species. Our study reveals relevant differences in the structure of food webs and shorebird communities in coastal areas along an avian flyway. While differences in trophic redundancy of food webs point to distinct levels of ecosystem resilience, contrasts in the organization of shorebird communities highlight the plasticity in the foraging behaviour of species inhabiting areas with distinct environmental conditions.

Understanding how marine microbial food webs and their ecosystem functions are changing is crucial for projections of the future ocean. Often, simplified food web models are employed and their solutions are only evaluated against available observations of plankton biomass. With such an approach, it remains unclear how different underlying trophic interactions affect interpretations of plankton dynamics and functioning. Here, we quantitatively compare four hypothetical food webs to data from an existing mesocosm experiment using a refined version of the Minimum Microbial Food Web model. Food web representations range from separated food chains to complex food webs featuring additional trophic links including intraguild predation (IGP). Optimization against observations and taking into account model complexity ensures a fair comparison of the different food webs. Although the different optimized model food webs capture the observations similarly well, projected ecosystem functions differ depending on the underlying food web structure and the presence or absence of IGP. Mesh‐like food webs dominated by the microbial loop yield higher recycling and net primary production (NPP) than models dominated by the classical diatom‐copepod food chain. A high degree of microzooplankton IGP increases NPP and organic matter recycling, but decreases trophic transfer efficiency (TTE) to copepods. Copepod production, the trophic role of copepods, and TTE are more sensitive to initial biomass changes in chain‐like than in complex food webs. Measurements resolving trophic interactions, in particular those quantifying IGP, remain essential to reduce model uncertainty and allow sound conclusions for ecosystem functioning in plankton ecosystems.

Lake Ellasjøen, located in the Norwegian high arctic, contains the highest concentrations of polychlorinated biphenyls (PCBs) ever recorded in fish and sediment from high arctic lakes, and concentrations are more than 10 times greater than in nearby Lake Øyangen. These elevated concentrations in Ellasjøen have been previously attributed, in part, to contaminant loadings from seabirds that use Ellasjøen, but not Øyangen, as a resting area. However, other factors, such as food web structure, organism growth rate, weight, lipid content, lake morphology, and nutrient inputs from the seabird guano, also differ between the two systems. The aim of this study is to evaluate the relative influence of these factors as explanatory variables for the higher PCB fish concentrations in Ellasjøen compared with Øyangen, using both a food web model and empirical data. The model is based on previously developed models but parameterized for Lakes Ellasjøen and Øyangen using measured data wherever possible. The model was applied to five representative PCB congeners (PCB 105, 118, 138, 153, and 180) using measured sediment and water concentrations as input data and evaluated with previously collected food web data. Modeled concentrations are within a factor of two of measured concentrations in 60% and 40% of the cases in Lakes Ellasjøen and Øyangen, respectively, and within a factor of 10 in 100% of the cases in both lakes. In many cases, this is comparable to the variability associated with the data as well as the efficacy of the predictions of other food web model applications. We next used the model to quantify the relative importance of five major differences between Ellasjøen and Øyangen by replacing variables representing each of these factors in the Ellasjøen model with those from Øyangen, in separate simulations. The model predicts that the elevated PCB concentrations in Ellasjøen water and sediment account for 49%-58% of differences in modeled fish PCB concentrations between lakes. These elevated sediment and, to a lesser extent, water concentrations in Ellasjøen are due to PCB loadings from seabird guano. However, sediment-water fugacity ratios of PCBs are consistently greater in Ellasjøen compared with Øyangen, which suggests that internal lake processes also contribute to differences in sediment and water concentrations. We hypothesize that the nutrients associated with guano influence sediment-water fugacity ratios of PCBs by increasing the stock of pelagic algae. As both these algae and the guano settle, their organic carbon content is degraded faster than PCBs, which causes an extra magnification step in Ellasjøen before these detrital particles are consumed by benthic organisms, which are in turn consumed by fish. The model predicts that the remaining approximately 50% of the differences in PCB concentrations observed between the fish of these lakes are due to other subtle differences in their food web structures. In conclusion, based on the results of a food web model, we found that the most dominant factors influencing the higher PCB fish concentrations in Lake Ellasjøen compared with Øyangen are the higher sediment and water concentrations in Ellasjøen, caused by seabird guano. Together, sediment and water are predicted to account for 49%-58% of differences in fish concentrations between lakes. Although seabird guano provides a source of nutrients to the lake, in addition to contaminants, empirical data and indirect model results suggest that nutrients are not leading to decreased bioaccumulation, in contrast to what has been observed in temperate, pelagic food webs. The results of this study emphasize the importance of considering even small differences in food web structure when comparing bioaccumulation in two lakes; although the food web structures of Ellasjøen and Øyangen differ only slightly, the model predicts that these differences account for most of the remaining approximately 50% of the differences in PCB fish concentrations between the two lakes. This study further demonstrates the utility of food web models as we were able to predict and tease apart the influence of various factors responsible for the elevated concentrations in the fish from Lake Ellasjøen, which would have been difficult using the field data alone.

To test the effects of temperature on food web structure and productivity, Mary O'Connor (above, checking temperatures) and colleagues placed five microcosms of food webs (shielded from full sunlight and UV) in eight independent water tables, each filled with a temperature-conditioned water bath.

Trophic interactions are central to ecosystem functioning, but the link between food web structure and ecosystem functioning remains obscure. Regularities (i.e. consistent patterns) in food web structure suggest the possibility of regularities in ecosystem functioning, which might be used to relate structure to function. We introduce a novel, genetic algorithm approach to simulate food webs with maximized throughput (a proxy for ecosystem functioning) and compare the structure of these simulated food webs to real empirical food webs using common metrics of food web structure. We repeat this analysis using robustness to secondary extinctions (a proxy for ecosystem resilience) instead of throughput to determine the relative contributions of ecosystem functioning and ecosystem resilience to food web structure. Simulated food webs that maximized robustness were similar to real food webs when connectance (i.e. levels of interaction across the food web) was high, but this result did not extend to food webs with low connectance. Simulated food webs that maximized throughput or a combination of throughput and robustness were not similar to any real food webs. Simulated maximum-throughput food webs differed markedly from maximum-robustness food webs, which suggests that maximizing different ecological functions can generate distinct food web structures. Based on our results, food web structure would appear to have a stronger relationship with ecosystem resilience than with ecosystem throughput. Our genetic algorithm approach is general and is well suited to large, realistically complex food webs. Genetic algorithms can incorporate constraints on structure and can generate outputs that can be compared directly to empirical data. Our method can be used to explore a range of maximization or minimization hypotheses, providing new perspectives on the links between structure and function in ecological systems.

Population's response to environmental noise: the influence of food web structure

The structure of food webs is frequently described using phenomenological stochastic models. A prominent example, the niche model, was found to produce artificial food webs resembling real food webs according to a range of summary statistics. However, the size structure of food webs generated by the niche model and real food webs has not yet been rigorously compared. To fill this void, I use a body mass based version of the niche model and compare prey-predator body mass allometry and predator-prey body mass ratios predicted by the model to empirical data. The results show that the model predicts weaker size structure than observed in many real food webs. I introduce a modified version of the niche model which allows to control the strength of size-dependence of predator-prey links. In this model, optimal prey body mass depends allometrically on predator body mass and on a second trait, such as foraging mode. These empirically motivated extensions of the model allow to represent size structure of real food webs realistically and can be used to generate artificial food webs varying in several aspects of size structure in a controlled way. Hence, by explicitly including the role of species traits, this model provides new opportunities for simulating the consequences of size structure for food web dynamics and stability.