渔情预报技术及模型研究进展

Review of fishery forecasting technology and its models

渔情预报是渔场学的重要研究内容,对渔业科学生产和渔业资源管理具有重要的意义。近年来,随着现代统计理论、数值计算方法、数据挖掘和人工智能等理论和技术的发展,使渔情预报技术和模型的发展焕发出了新的活力,为此本文对渔情预报技术及模型研究进展进行了回顾,并对今后发展提出了展望。本文简要概述了渔情预报建模的理论和方法,包括渔情预报相关的渔场学基础、数据模型和预报模型,重点介绍了基于统计和机器学习、人工智能方法的渔情预报模型的应用和研究现状,并对各种模型在渔情预报应用中的优势与缺陷进行了综合分析,针对存在的问题提出了建议。主要建议包括:建立专为渔业服务的海洋环境预报系统;进行长期系统的渔业资源调查,针对不同鱼种和海区对数据获取和处理方法进行标准化和规范化;借助随机模拟方法降低模型不确定性,提高预报精度。本总结与分析将为国内的渔情预报模型研究工作提供参考。

Fishery forecast is an important research content of oceanography,which has the vital significance to the scientific production and management of fishery resources in fishery science. In recent years, with the development of modern statistics theory, the numerical calculation method, and data mining and artificial intelligence theory and technology, the development of fishery forecasting technology and model has displayed a new vitality. Therefore,the studies on the fishery forecasting technology and model development are reviewed, and the future development of fishery forecasting was put forward. In this paper, the theory and methods of fishery forecasting are summarized, including fishery oceanography, data models and prediction models related to this subject. Prediction models based on statistics methods and machine learning and artificial intelligence methods are emphasized, as well as the advantages and drawbacks of each kind of the forecasting model. Some research perspectives of fishery forecasting models are also proposed, i. e. developing ocean environments forecasting system, conducting systematic fishery resources survey of long standing and the standardization and normalization of fishery data acquisition and processing, reducing the uncertainty of prediction models with stochastic simulation methods and improving the prediction accuracy.