Comparison of MesoHABSIM with two microhabitat models (PHABSIM and HARPHA)

Abstract

AbstractThis study demonstrates how using different habitat models can influence the results of instream habitat assessment and conclusions for river management. We used three models for a portion of the Quinebaug River (Connecticut and Massachusetts, USA): a simplistic microhabitat model with univariate habitat‐use criteria and substrate‐based channel index (the Physical Habitat Simulation Model (PHABSIM)); a microhabitat model using multivariate criteria including a wide range of cover attributes (HARPHA); and a mesohabitat model with multivariate habitat‐suitability criteria (MesoHABSIM). The flow‐habitat rating curves produced by each model were compared at two scales: site and study segment. To investigate the impact of model choice on answering questions such as which location or flow provides more habitat, we applied Spearman's correlation of ranks. The relationship between habitat‐suitability predictions and fish presence at the same location was investigated with dedicated fish observations. The study showed that: (1) of the tested models, only MesoHABSIM predictions correlated with fish observations; (2) the variation within microscale models (PHABSIM and HARPHA) was greater than between micro‐and mesoscale models (HARPHA and MesoHABSIM); and (3) simple univariate habitat‐use criteria provided the largest source of discrepancies among the models. We suggest that these differences may lead to erroneous conclusions, especially if flow‐habitat rating curve analysis is considered an endpoint of instream flow study. Copyright © 2007 John Wiley & Sons, Ltd.