Files
Abstract
This dissertation consists of a study that examines the importance of indirect energy flows in food webs and a multi-taxon gradient analysis study informed by a network perspective.Food Webs and Network Analysis Compartment models are widely used to represent ecological networks of stocks and flows of conserved substances. Network environ analysis (NEA) has revealed several interesting properties of flowstorage networks but can only be applied to systems at a constant steady state. I developed a computational analog of NEA called dynamic environ approximation (DEA), which can be used away from steady state. I used DEA to examine the effects of system size and connectance on the importance of indirect energy flows in a commonly studied theoretical food web model. Over the full range of parameter values examined, the mean fraction of energy traveling over indirect paths was 9.2%, but could be as high as 30%. This quantity increased with system size but peaked at intermediate connectance levels, a pattern explained by the availability of more pathways at intermediate connectance levels.Multi-Taxon Gradient AnalysisThe extent to which ecological communities are coherent entities as opposed to mere intersections of species distributions is one of the fundamental questions of ecology. Gradient analysis is commonly used to address this question; however, all such studies have used organisms from a single guild. This risks missing connections due to non-competitive interactions, which should be most common among functionally different organisms. I used two different methods of analyzing species abundance data, elements of metacommunity structure (EMS) and causal discovery, to examine the importance of species interactions in structuring communities. The EMS analysis found that the distributions of study taxa commonly exhibited high coherence, turnover and boundary clumping, the pattern termed Clementsian in EMS. Also, pairs and triplets of directly interacting guilds had higher-than-expected boundary conjunction values, while those that did not directly interact generally did not. I also produced a causal interaction network for my study taxa and found that inter- and intra-guild interactions were equally common. These results highlight the importance of inter-guild interactions in structuring patterns of cooccurrence.