【摘要】 海洋牧场是实现海洋环境保护与渔业资源养护的重要举措。本研究以2014年春(5月)、夏(8月)、秋(10月)对海州湾海洋牧场示范区水环境数据为基础,选取高锰酸盐指数(CODMn)、溶解无机氮(DIN)、溶解无机磷(DIP)、生化需氧量(BOD)作为评价指标,利用BP神经网络模型对该海域水环境承载力进行评价。结果显示,2014年海州湾海洋牧场水环境承载力指数平均值高于0.6,承载状态较理想。水环境承载力存在明显的季节变化,呈现出夏季>春季>秋季;海洋牧场区域水环境承载力优于对照区海域;CODMn、DIN浓度过高是导致部分站点轻度超载的主要原因,这可能与陆源污染有关。研究表明,航运对水环境承载状态有负面影响;BP神经网络模型构建方便快捷,评价结果客观合理,可应用于海洋牧场等海域水环境承载力的研究。更多还原

【Abstract】 Marine ranching is an important measure to achieve marine environmental protection and efficient production of fishery resources. Based on the water environment data of Haizhou Bay marine ranching demonstration area in spring(May), summer(August) and autumn(October) of 2014, selecting permanganate index(CODMn), dissolved inorganic nitrogen(DIN), dissolved inorganic phosphorus(DIP), biochemical oxygen demand(BOD) as evaluation indexes, neural network models were developed using BP(back propagation) to evaluate the water environmental carrying capacity of the area. The results showed that the average water environmental carrying capacity index of Haizhou Bay marine ranching in 2014 was higher than 0.6, indicating an ideal level of the carrying status. Water environmental carrying capacity varied seasonally as follows: summer >spring > autumn. The water environmental carrying capacity of marine ranching area was obviously better than that of the control area. The high concentrations of CODMn and DIN were the main cause for the slight overload of some stations, which might be related to land-based pollution. In addition, shipping also had a certain negative impact on the water environment. BP neural network model has the advantages of convenient construction,objective and reasonable evaluation results, which could be applied to the study of water environmental carrying capacity in marine pastures and other sea areas.更多还原