Date of this Version
Chavez-Ramirez, F. Patterns of relationships among whooping crane fitness, blue crab abundance, and freshwater inflows: an exploratory assessment of available data sets. In Chavez-Ramirez, F, ed. 2005. Proceedings of the Ninth North American Crane Workshop, Jan 17-20, 2003. Sacramento, California: North American Crane Working Group. Pp. 254.
This paper evaluates the possible relationships of whooping crane (Grus americana) fitness parameters to blue crab abundance and freshwater inflows based on independently gathered data sets from several agencies between 1978 and 1999. I ordered all possible analysis models into two separate types based primarily on implied relationships. The 1st order exploratory relationship models were those where evaluated parameters were potentially affected by freshwater inflows, such as salinity and blue crab abundance. The 2nd order exploratory relationship models evaluated parameters that could, more directly, have potential affects on whooping crane fitness parameters for which there was data available. Freshwater inflow showed a significant negative relationship with salinity (P = 0.0001, R2 = 0.59) and a significant positive relationship with San Antonio Bay (SAB) trawl captures (P = 0.03, R2 = 0.29). whooping crane Detectability Index (defined as proportion of cranes observed during a complete day aerial survey in relation to total number of cranes wintering in the area) was strongly and significantly associated (P = 0.01,) with a combination of three independent variables, SAB trawls, salinity, and commercial landings which explained approximately 74% of variation (R2 = 0.74). The model that best explained the number of nests found the following spring consisted of; SAB commercial landings, whooping crane Detectability Index, and peak population size (P < 0.00001) which explained almost 90% of the variation (R2 = 0.89). The number of chicks hatched the next spring was only significant with a model involving SAB (lbs) and WC detectability (P=0.03), however the variation explained by this model was only around 31% (R2 = 0.31). No models were found that significantly related any variables to over-winter mortality.