论文部分内容阅读
White sturgeon are an ecologically and commercially important species.Development of growth rate,body lipid,moisture,and energy models for white sturgeon fed at a given feeding rate(i.e.,50%to 110%of optimum feeding rate determined by a previously developed optimum feeding rate model provided in Lee et al.2014a)will allow researchers and farmers to predict feeding rate-dependent growth performance and will also enhance our predictability on a pattern of changes in nutrient and energy content in relation to feeding rate.Objectives of the current study were to develop and evaluate 1)growth rate models and 2)body lipid,moisture,and energy models at various feeding rates for white sturgeon.Nineteen datasets from 6 studies,containing initial body weights(g;weight class),corresponding specific growth rates(SGR;%body weight increase per day;response variable),and respective feeding rates(%body weight per day;predictor variable),were used for objective 1.Twelve datasets from 6 studies,containing initial and final body weights(g;weight class),corresponding body lipid(%),moisture(%),and energy(kJ g-1)(response variables),and respective feeding rates(%;predictor variable),were used for objective 2.In objective 1,proportions(i.e.,0.5 to 1.1)of optimum feeding rate(OFR;%),defined as the rate at which growth is maximal,calculated from a previously developed OFR prediction model were tested using a quadratic broken-line model,selected as the best-fit model,to estimate SGR at each proportion of OFR for 19 datasets.Using Box-Cox technique,a linear regression model with square root of estimated SGR as the response and the natural logarithm of the square root of the initial body weight as the predictor was selected as the best-fit model.Final SGR prediction models at each proportion of OFR were developed and are provided in the text.In objective 2,various regression models were fit to 12 datasets to investigate the relationship between response variables(body lipid,moisture,energy)and predictor variables(various feeding rates).Based on model selection criteria tests,a 2nd degree polynomial model was selected as the best model for all the response variables for 12 datasets.Proportions(0.5 to 1.1)of OFR calculated from a previously developed OFR prediction model were tested using the best model to estimate body lipid,moisture,and energy at each proportion of OFR.Then,a variety of regression models were fit to a given set of 12 estimates of each body lipid,moisture,and energy at 100%of OFR to find a best-fit model.Based on the model selection criteria,final prediction models for body lipid,moisture,and energy at each proportion of OFR were selected and are provided in the text.To our knowledge,these are the first models developed for growth rate,body lipid,moisture,and energy predictions at various feeding rates for white sturgeon.They can be applied for white sturgeon ranging from 0.05 to 800 g.