Abundance exchange models of fish assemblages along the Hudson River Estuary Gradient, New York

Abstract

The spatially explicit abundance exchange model (AEM) was built for four fish species: winter flounder (Pseudopleuronectes americanus), Atlantic silverside (Menidia menidia), eastern silvery minnow (Hybognathus regius), and striped bass (Morone saxatilis) along the Hudson River estuary gradient, New York. The fish and habitat data during 1974-1997 were used to develop and calibrate the AEM; and the fish data during 1998-2001 was used to validate the model. Preference indexes of fish species for dissolved oxygen, salinity, water temperature, and bottom substrates along the gradient were estimated; and these were used to compute habitat preference (HP) of the associated fish species. The species HP was a key variable in the AEM to quantify abundance and distribution patterns of the associated species along the gradient. The AEM could efficiently predict abundance and distribution patterns of all modeled species except striped bass. The model ability for predicting a local distribution range of a fish species with broad tolerance on changing environment like striped bass should be improved.

Predicting estuarine use patterns of juvenile fish with Generalized Linear Models

“…Distribution of juvenile fish in estuarine and coastal areas has been addressed with diverse methodologies including: habitat suitability indices (Brown et al, 2000;Vinagre et al, 2006); statistical habitat suitability models using Generalized Linear Models (GLM) (Riou et al, 2001;Le Pape et al, 2003b;Le Pape et al, 2007;Nicolas et al, 2007;Whaley et al, 2007;Rochette et al, 2010;Vasconcelos et al, 2010), Generalized Additive Models (GAM) (Stoner et al, 2001;Kupschus, 2003;Francis et al, 2005;Florin et al, 2009;Zucchetta et al, 2010), regression trees (Fodrie and Mendoza, 2006;Francis et al, 2011;Froeschke and Froeschke, 2011), mixed approaches (Norcross et al, 1999;Cabral et al, 2007;Lauria et al, 2011) for assessing preferential habitat; as well as regression quantiles for assessing potential habitat (Eastwood et al, 2003;Martin et al, 2009). In parallel, dynamic modelling approaches, such as spatially explicit Abundance Exchange Models based on habitat preferences (Singkran and Bain, 2008) are more seldom applied. Habitat-preference models have shown to be useful tools to model fish species presence/absence and/or abundance in estuaries, given knowledge of the local environmental conditions.…”