Stat In Fisheries

STATISTICAL TOOLS IN FISHERIES RESEARCH T.M.Sankaran The Science, Statistics is essentially a branch of Applied Mathematics. To be exact it may be considered as Mathematics applied to observational numerical data. The aim of application of Statistical techniques is mainly for studying about the population, studying about the variations present and studying about the methods of reduction of data. All the Statistical tools are developed to suit the above aims. Just as in any other field, in fisheries also observational data provide very large extent of valuable information. Starting from designing of collection of such data, the application of such tools play a very useful role in any scientific enquiry in fisheries. Collection of data become an important function when any study is planned scientifically, needless to say about a study aimed at understanding the characteristics of a population living organisms such as fish. Fish population which is self-regulating and renewable, require a continuous monitoring system in order to provide with the latest status,which is needed for a judicious exploitation of the resource. This calls for a wellplanned statistical design for collection of necessary information. Different sampling designs are used for this purpose depending on the situation. The Statistical sampling design developed by the Central Marine Fisheries Research Institute (CMFRI), India for collecting the data on marine fish landings on the Indian coast is a typical example of a statistical tool applied for collection of data in fisheries. Fishery Biologists wanted to know more about the growth pattern of different species of fish, when they started studying the biology of the same. Quantitative approach using morphometric characters came to their help. With the help Mathematical and Statistical tools different models were developed which helped the interpretations of the relationships between various body parts, the legth and the weight, etc.. Among these models developed the most useful and widely applied model is the von Bertalanffy's Growth model, which help estimating the length of fish at a given age. The model is based on three parameters, viz. Theoretical maximum length of the fish, growth parameter and the time at which the fish has zero length. This model has been widely applied in biological studies such as stock assessment of different species, growth study, etc.. Certain authors have improved the model to suit the changes in climatic conditions, temperature variations, etc. Biometricians found applications of length-weight relationship model in the biological studies of fish. The condition factor based on this model helped them interpret the changes in conditions of fish due to factors like non-availability of food, egg carrying, spawning, and such other ecological and biological situations. The models always become handy when interpolations or extrapolations are to be done for estimating some missing value or for projection purposes. The length frequency data help a lot in the stock assessment problems. A periodical plot of length of length frequency data provides the biologist with an age length key from the modal

progression interpretations. Estimates of natural mortality and fishing mortality are arrived at using statistical techniques. Details regarding recruitment, selection of gears, yield, effort, catch per unit effort, etc., which are needed for an understanding of any fishery in order to guarantee a sustainable catch, are obtained through statistical enquiries and analyses. When it comes to the question of fisheries research one may have to conduct specific experiments in order to obtain observations to find an answer to the question posed. Here the statistical tools starts playing its role from the very stage of designing the experiment. This is very important, since any experiment conducted without proper planning is unlikely to yield useful results. Actuallym this branch of Statistics, viz., Design of Experiments, has been developed as a separate topic mainly because of the demand from the field studies in biology and agriculture. The main problems that attracted the attention of design developers were, how many times an experiment is to be repeated for the reliability of the result, how to avoid a biased approach in the distribution of treatments into the experimental area and how to take into account the locally existing heterogeneity. Since the analysis of the results are also to be done using statistical tools, the samples or observations are to have a minimum size. The analysis of variance technique is a very important tool, which is used in the analysis of the results of experiments based on statistical designs. The simple principle of this technique is that the total variation present in the observed data can be split into different components and compared. This will be possible only if the experiments are conducted using proper designs. In fisheries the application of designing techniques get more relevance with the advent of aquaculture and related experiments. Different topics coming under statistical inference are widely used in fisheries and related studies. Almost all of the statistical tests have got wide applications in fisheries. The parametric and non-parametric tests and the small and large sample tests do find place in fisheries research, like in any other subject. Similar is the case with estimating different population parameters from the available sample observations. Both the value (point estimation) and the interval (interval estimation) are estimated depending on the requirements. This become more relevant since fish population is something, which is not accessible directly, and has to depend on samples alone. Statistical quality control has found an application in fish processing field. The famous '3 sigma' property of normal distribution help deciding the lower and upper quality level in fish food items in processing industry. This help deciding whether the defects or defective items are due to assignable causes or chance causes. Perhaps the latest '6-sigma' interpretation is yet to be introduced in fisheries. Multivariate analysis techniques also have applications in fisheries field of research. Very often observations gathered from more variables, related among themselves, can provide with better information than from those based on single variables. Multiple regressions and correlations are widely used for interpreting the inherent relationships of different morphometric characters of fish. Similarly the influence of different physical or chemical parameters, present in the water body, on the growth of fish can be studied with the multivariate analysis. If grouping of ecosystems based on such physico-chemical parameters

are to be done, advanced statistical tools, such as Distance function, T² statistic, etc., can be used. Factor analysis, Time Series analysis, Bioassay methods and other statistical tools can also be made use of in fisheries in appropriate contexts. ----------------