Mapping benthic ecological diversity and interactions with bottom-contact fishing on the Flemish Cap (norhtwest Atlantic)

The ecological diversity of benthic invertebrates from bottom trawl surveys was mapped for the Flemish Cap, a plateau of ~200 km radius in the northwest Atlantic. Species density (SpD), the exponential Shannon diversity index (eH′) and Heip’s index of evenness (E') were measured at different spatial scales. Continuous surfaces of each were created to 2000 m depth using predictive distribution models based on random forest (RF) algorithms. When fishing effort was included as an independent variable in the RF models, it was the most important predictor of sample SpD but unimportant in predicting eH′ and only a minor predictor of E'. In the absence of a historical baseline, we used a novel approach to evaluate spatial impacts of fishing on diversity by simulating and comparing spatial SpD prediction surfaces using response data associated with different levels of fishing effort. Although it is not possible to fully evaluate the precise nature of the impact of fishing on the ecological diversity, our models have identified Sackville Spur, Flemish Pass and south of Flemish Cap as the areas of greatest impact. Combining minimum bottom salinity, annual primary production range, fishing effort and biomass of sponges and small gorgonian corals, resulted in the best performing generalized additive model, explaining 73% of the total variance in SpD. Although current closures to protect vulnerable marine ecosystems from the adverse impacts of bottom fishing activities protect an important part of the ecological diversity associated with the deeper communities, unique and representative habitats on top of the Cap remain unprotected.

本研究针对西北大西洋半径约200公里的弗拉芒海台（Flemish Cap），基于底拖网调查（bottom trawl surveys）数据绘制了底栖无脊椎动物（benthic invertebrates）的生态多样性分布图谱。研究在不同空间尺度下测算得到物种密度（Species Density, SpD）、指数化香农多样性指数（exponential Shannon diversity index, eH'）以及希普均匀度指数（Heip’s index of evenness, E'）。基于随机森林（Random Forest, RF）算法构建预测分布模型，生成了该海域水深2000米以浅各指标的连续空间分布面。当将捕捞努力量作为独立变量纳入RF模型时，其是样本物种密度的最重要预测因子，但对指数化香农多样性指数的预测无显著影响，仅对希普均匀度指数的预测具有微弱作用。由于缺乏历史基线数据，本研究采用创新方法：通过模拟并对比不同捕捞努力量水平对应的响应数据所生成的物种密度空间预测面，评估捕捞活动对多样性的空间影响。尽管无法完全精准解析捕捞对生态多样性影响的具体机制，但本研究模型已识别出萨克维尔海脊（Sackville Spur）、弗拉芒海峡（Flemish Pass）以及弗拉芒海台南部为受捕捞影响最严重的区域。结合底层最低盐度、年初级生产力波动范围、捕捞努力量以及海绵与小型柳珊瑚生物量构建的广义可加模型（Generalized Additive Model, GAM）表现最优，可解释物种密度总变异的73%。尽管当前为保护脆弱海洋生态系统免受底拖网活动不利影响而设立的禁捕区，已保护了与深水区群落相关的重要生态多样性，但弗拉芒海台顶部的独特典型生境仍未得到有效保护。