基于改进ConvLSTM网络太平洋长鳍金枪鱼时空分布趋势预测

Spatial and temporal distribution trends of albacore tuna in the Pacific Ocean based on complex deep learning

中文关键词:  长鳍金枪鱼  时空分布  ConvLSTM  太平洋

英文关键词:albacore tuna  spatial and temporal distribution  ConvLSTM  The Pacific Ocean

中文摘要:

      渔场资源与位置的变动由空间与环境因子共同驱动，远洋渔场时空演变信息的精准预测是远洋捕捞的关键支撑。该研究考虑渔业生产统计数据，并兼顾同期海洋环境数据包括海表面温度（Sea surface temperature, SST）、海表面盐度（Sea surface salinity, SSS）、初级生产力（primary productivity, PP）和溶解氧浓度（dissolved oxygen concentration, O2），提出了一种融合卷积长短期记忆网络（ConvLSTM）和卷积神经网络（CNN）的渔场时空分布预测模型。首先对时空因子进行编码，提取高层时空特征；其次采用CNN提取海洋环境变量的抽象特征，并基于ConvLSTM提取渔业数据的时空特征，最后融合高层时空关联信息对渔场时空演变趋势进行预测。以1995－2018年太平洋海域的延绳钓生产数据对模型进行验证，模型的根均方误差为0.1036，实验对比发现较传统渔场预报模型的预测误差降低15%~40%，预测的高产渔区与实际作业的高渔获量区匹配度高。该研究构建的渔场时空预测模型能够准确地预测出太平洋长鳍金枪鱼的时空分布，为太平洋长鳍金枪鱼的延绳钓渔业提供科学参考依据。

英文摘要:

      The changes of fishery resources and location are driven by both spatial and environmental factors. Accurate prediction of spatio-temporal evolution information of pelagic fishery is the key support of pelagic fishery. The study considered fishery production statistics, And factors the Marine environmental data including Sea surface temperature (SST), Sea surface salinity (SSS), primary productivity, PP) and dissolved oxygen concentration (O2), and a spatial-temporal distribution prediction model based on ConvLSTM and CNN was proposed. Firstly, the spatial and temporal factors are encoded to extract the spatial and temporal characteristics of the high-rise. Secondly, abstract features of Marine environmental variables were extracted by CNN, and spatiotemporal features of fishery data were extracted based on ConvLSTM. Finally, high-level spatiotemporal correlation information was fused to predict the spatiotemporal evolution trend of fisheries. The model was verified with the longline production data of the Pacific Ocean from 1995 to 2018, and the root mean square error of the model was 0.1036. The experimental comparison showed that the prediction error of the model was 15% ~ 40% lower than that of the traditional fishery prediction model, and the predicted high-yield fishery area had a high matching degree with the actual high-yield fishery area. The spatio-temporal prediction model constructed in this study can accurately predict the spatio-temporal distribution of Pacific albacore tuna and provide scientific reference for the longline fishery of Pacific albacore tuna.