Climate change is continuing to influence spatial shifts of many marine species by causing changes to their respective habitats. Habitat suitability as a function of changing environmental parameters is a common method of mapping these changes in habitat over time. The types of models used for this process (e.g. bioclimate models) can be used for projecting habitat if appropriate forecasted environmental data are used. However, the input data for this process must be carefully selected as less reliable results can incite mis-management. Thus, a knowledge of the organism and its environment must be known a priori. This paper demonstrates that these assumptions about a species’ life history and the environment are critical when applying certain types of bioclimate models that utilize habitat suitability indices. Inappropriate assumptions can lead to model results that are not representative of environmental and biological realities. Using American lobster (Homarus americanus) of the Gulf of Maine as a case study, it is shown that the choice of extrapolation data, spatial scale, environmental parameters, and appropriate subsetting of the population based on life history are all key factors in determining appropriate biological realism necessary for robust bioclimate model results.

No abstract

Stock assessments for a majority of the world’s fisheries often do not explicitly consider the effects of environmental conditions on target species, which can raise model uncertainty and potentially reduce forecasting quality. Model-based abundance indices were developed using a delta generalized linear mixed model that incorporates environmental variability for use in stock assessment to understand how the incorporation of environmental variability impacts our understanding of population dynamics. For this study, multiple model-based abundance indices were developed to test the incorporation of environmental covariates in a length-structured assessment of the American lobster stock in the Gulf of Maine/Georges Bank on the possible improvement of stock assessment quality. Comparisons reveal that modelled indices with environmental covariates appear to be more precise than traditional indices, but model performance metrics and hindcasted fishery statuses revealed that these improvements to indices may not necessarily mean an improved assessment. Model-based abundance indices are not intrinsically better than design-based indices and should be tested for each species individually.

The Yangtze River estuary (YRE) is an important migration channel and foraging habitat for Coilia nasus. Due to its ecological significance and a prioritization of this species' protection, the need to investigate and analyze environmental relationships of the abundance of Coilia nasus in the YRE as well as develop an understanding of their temporal and spatial distributions is becoming exceedingly important. Using fishery data and environmental survey data from 2009 to 2016, three models including generalized additive mixed models (GAMM), generalized additive models with zeroinflated Poisson distribution (ZIP-GAM) and two-step GAM were used to analyze relationships between environmental factors and the distribution of Coilia nasus in the YRE. The results showed that model fitting of GAMM was more consistent with observations and revealed influences of water temperature, salinity, chlorophyll, and pH on distribution. GAMM demonstrated that higher Coilia nasus abundances were located in waters with water temperature values at 15°C and 30°C, and lower Coilia nasus abundances were located in areas with water temperature values at 10°C and 20°C. All models indicated that the effect of salinity on abundance of Coilia nasus present a multimodal pattern including three peaks at 5, 15, and 25 ppt respectively.Additionally, abundance of Coilia nasus increased with the increase of chlorophyll A in its range of 0-4 mg/L. In a range of 8.0-9.5, higher PH value was more suitable for the aggregation of Coilia nasus. Cross validation was used to evaluate the predictive performance of models and GAMM was found to be the best. The predicted abundance distribution of Coilia nasus in the summer and autumn of 2016 was relatively higher overall than that in winter and spring. The predicted zero abundance distribution pattern was consistent with the sampling presence distribution which was obtained using fishery independent survey data of the year 2009-2015. Facing the urgency protection of Coilia nasus in YRE, results of this study could be used for Coilia nasus conservation and reserve planning. K E Y W O R D S Coilia nasus, GAM, spatio-temporal distribution, the Yangtze River estuary, zero-inflated model How to cite this article: Ma J, Li B, Zhao J, Wang X, Hodgdon CT, Tian S. Environmental influences on the spatio-temporal distribution of Coilia nasus in the Yangtze River estuary. J Appl

No abstract