Date of Award


Degree Name

Doctor of Philosophy



First Advisor

Eichholz, Michael


The main proponent of management of any animal species is habitat management. The ability of habitats to maintain species communities will depend on the variation in both habitat structure and composition. While spatial variation in habitat resources plays a critical role in determining the distribution of species, an equally important consideration that must be accounted for is temporal variation in the needs of the target species. Nutritional requirements, and thus the habitats used to fulfil those nutritional needs, will be different depending on if the individual is breeding, migrating, molting, enduring winter, or establishing a new range. For waterfowl, we currently assume that winter and migration are nutritionally stressful and are consequently the periods most limiting to populations. The theory of ideal free distribution assumes that animals distribute themselves according to the factor most limiting to their fitness. In the case of non-breeding waterfowl, this factor is believed to be food. We assume if habitats with abundant food resources are provided, waterfowl will make use of those habitats. Deviations from an ideal free distribution based on food become problematic for managers since these deviations will keep some areas from being exploited to their potential, while other areas may become over-exploited. Recent observations have made it clear that this assumption may need to be reconsidered for effective waterfowl management. In this dissertation I quantitatively examine the degree to which spring migrating waterfowl conform to, or deviate from, an ideal free distribution based on food. Since food availability was not expected to account for 100% of waterfowl distribution, I further investigated what other potential habitat components influence the distribution of spring migrating waterfowl. In the first chapter of this dissertation, I explicitly tested the influence of food availability on waterfowl distribution. A series of paired 0.42 ha (1 acre) plots were established in various habitat types. One plot in each pair was treated with corn to a density of 2000 kg/ha, while the remaining plot was used as a control. Background food availability was controlled for by taking core samples from each plot, and estimating the natural seed and invertebrate biomass. The abundance each species of waterfowl using the plots was recorded during morning and afternoon observation periods. Linear mixed models were used to assess how variations in food availability influenced distribution of waterfowl. Although the waterfowl community showed a significant preference for treatment plots, our ability to influence abundance was low. Food availability accounted for minimal variation in abundance of the waterfowl community as a whole or for each focal species. Since the results of the first chapter showed food availability to be a poor predictor of waterfowl distribution, in the second chapter I set out to determine other potential habitat variables could be responsible for driving waterfowl distribution during spring migration. After each observation period, a series of habitat structural measurements were made within each paired plot. Habitat measurements included water characteristics, vegetation structure, vegetation type, habitat type, and weather conditions. Linear mixed models and model selection were employed to determine which of the habitat characteristics showed the greatest ability to predict waterfowl abundance on study plots. Models containing precipitation and Wabash River flood stage predictor variables were the best performing, and were the best predictors of waterfowl abundance on study plots. The results from this chapter encouraged investigation into how environmental factors shape the formation of local duck communities are structured from regional pools. In the third chapter of this dissertation I investigate the relationship between local and regional waterfowl community structure and how this relationship is mediated through environmental filters which dictate what proportion of the regional species pool exists at local scales. To address this relationship, I tested three hypotheses: 1) resource availability drives species diversity at local scales; 2) similarity between local and regional habitats will result in a similar species community occurring at both scales, and; 3) increased heterogeneity of local habitat structure will result in more diverse waterfowl communities at local scales. I used Mahalanobis distance and cumulative standard deviation of habitat variables in conjunction with mixed models and model selection to compare hypotheses and determine which had the greatest potential for mediating local community structure from regional pools. Increasing resource abundance appeared to have the greatest influence over local duck diversity, but the model indicated that although species diversity could be increased by increasing food abundance, diversity at local scales would become saturated before becoming representative of the regional community.




This dissertation is Open Access and may be downloaded by anyone.