A Multi-scale Approach for Characterizing Habitat Selection of Tidal Creek Fish in Charlotte Harbor, Florida
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science, Department of Environmental Science, Policy and Geography, College of Arts and Sciences, University of South Florida St. Petersburg.
This study supplemented existing fish catch data by assessing habitat characteristics in four tidal creeks in Charlotte Harbor. Habitat and fish data were analyzed to determine spatial patterns at three different spatial scales; at the smallest mesohabitat scale, between segments within each creek, and at the largest scale between creeks. Fish abundance and diversity were calculated from fish catch data collected by Mote Marine Laboratory from 2005 through 2009. Twenty-three habitat variables were derived from data collected during field surveys and via GIS analyses. Field sampling methodology was adapted from the MesoHABSIM approach to delineate eleven habitat types at the meso-scale based channel geometry and other hydro-morphologic characteristics. Field survey variables included intertidal habitat type, substrate, depth, and channel profile characteristics. Habitat variables also included sampling depth, dissolved oxygen, temperature, and salinity that were acquired at the time of capture by Mote Marine Laboratory. Spatial differences in habitat and fish variables were compared between three spatial scales using one-way ANOVA and multi-factor, nested ANOVA tests. Field sampling and analysis methods were designed to address the following research questions; • Question 1: How do habitat characteristics vary between groups at three spatial scales; creeks, segments within creeks, and mesohabitats within segments? • Question 2: How does fish abundance and species richness vary between groups at the creek, segment, and mesohabitat scales? • Question 3: At which spatial scale is variation in habitat characteristics and fish abundance and diversity the greatest? Results of ANOVA tests addressing Questions 1 and 2 revealed significant differences in several habitat and fish catch variables at each of the three spatial scales; between creeks, segments, and mesohabitat patches. Species richness and CPUE were generally higher in North and South Silcox Creeks and in the middle and lower segments among creeks. Mesohabitats within these creeks and segments were generally shallower, larger, with wide canopy zones and intertidal areas dominated by mangroves and higher salinity and temperature. Several habitat variables and fish CPUE and richness also varied between mesohabitat types, however, trends among variables and between HMU types were less consistent than differences at the larger segment and creek scales. It was difficult to determine if differences between HMU types were driven by habitat conditions unique to that HMU type or if between-habitat differences were largely a result of location within a particular segment or creek, especially, because the distribution of HMU types was disproportionate between creeks and segments. Despite differences between creeks, nested ANOVA tests suggest that variation in habitat conditions and fish species richness was most important at smaller scales between segments and mesohabitat patches. The only exception to this was for depth variables, where most variation was explained by larger scale differences between creeks.