FAO Home>Fisheries & Aquaculture
EspañolFrançaisРусский
 Food and Agriculture Organization of the United Nationsfor a world without hunger Fisheries and Aquaculture Division
 Home About us Activities Statistics GeoInfo Meetings and News Publications Fact Sheets
 Statistics VONBIT 2005 - Von Bertalanffy Iterative ApproachVONBIT stands for von Bertalanffy Iterative Approach and it is a computer software designed to run under Windows. The presented linear regression method for fitting the von Bertalanffy growth function to data on size at age is also directly applicable to tag and recapture data, thus offering a more integrated approach than other similar methods. Several comparison tests have revealed that the present approach shows lower error and bias and more flexibility of action than other commonly used models and systems.The methodological notes included in the tutorial component of the system were taken from the original work that was published in 1989 by ICES with reference J. Cons. int. Explor. Mer. 45: 200-208.Methodology In the method proposed here the driving parameter is k, trial values of which are used in simple linear regressions with age as the independent variable and observed size as the dependent variable. The advantage over general-purpose non-linear procedures is that the optimization process concentrates on values of k only and deals with stationary points on a curve, rather than on a surface. This makes it possible to select a single population value for k as a global optimum. The method involves very few computational steps, and although it remains an iterative approach (in that trial values of k are examined), the number of iteration steps as well as the accuracy of parameter estimation are under the control of the user.Fitting data from tagging experiments Extensive comparisons, some reported here, were made between the proposed method and other methods in current use. These comparisons reveal that the latter procedures may sometimes underestimate the true value of the overall population k. For instance, some methods make the basic assumption that all individuals have the same maximum asymptotic length and only their individual values of k are different. Others use the length at capture as independent variable, thus assuming that it is free of measurement errors. Such assumptions tend, in general, to underestimate k .Data handling Creating new data sets for size-at-age data or tagging data Frequencies or weighting factors may be included in size-at-age data Editing/modifying of datasets Ages/time intervals may refer to years, months or days   Computations Flexible set-up of ranges and accuracy requirements for the driving growth parameter k; Users verify the outcome of the iterative approach by means of plots and system messages; Displays of the resulting estimated parameters and associated coefficient of determination ; Displays of input and estimated values for lengths (size-at-age data); Displays of input and estimated values for lengths and absolute ages (tagging data); VONB plots for both size-at-age and tagging data; Consistency checks for size-at-age-data; Optional non-asymptotic approach for both size-at-age and tag-and-recapture data.  Statistics and Information Branch (FIAS)Fisheries and Aquaculture DepartmentFood and Agriculture Organization of the United NationsViale delle Terme di Caracalla00153RomeEmail: Source | Citation | XML | Print | PDF