Abstract
Oxygen-deficient sensitivity index （OSI）represents the hypoxia environment, which is calculated from the continuous values of dissolved oxygen saturation and water temperature. The macrobenthic community which dies from the development of hypoxia can be forecasted from the change of OSI by defining the initial biomass （Bmax）that has been observed before hypoxia. However, OSI is not a generalpurpose index, because it is necessary to set the four parameters estimated from field observations. In addition, the recovery amount of macrobenthos forecasted from the OSI method tends to be overestimated compared with the observed biomass. Therefore, the influence of hypoxia on each species of sampled macrobenthos is expressed as a new index that improved OSI （iOSI）, which is calculated from the two parameters and the continuous values of dissolved oxygen concentration and water temperature. Additionally, the iOSI method adopts the recruitment estimated from dissolved oxygen
concentration and subsequent growth rate. As a result, fluctuations in species biomass forecasted using the iOSI method is strongly correlated with field observations in June-July when hypoxia develops, as well as the OSI method. The recovery amount of macrobenthos in August-November when hypoxic water dissipates is able to be forecasted by the iOSI method more accurately than by the OSI method.
