...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>ICES Journal of Marine Science >Presence-only habitat suitability models for vulnerable marine ecosystem indicator taxa in the South Pacific have reached their predictive limit
【24h】

Presence-only habitat suitability models for vulnerable marine ecosystem indicator taxa in the South Pacific have reached their predictive limit

机译：南太平洋弱势海洋生态系统指标分类群的仅限栖息地适用性模型已达到预测极限

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Vulnerable marine ecosystems (VMEs) are typically fragile and slow to recover, making them likely to be substantially altered by disturbance. In the High Seas, regional fishery management organizations (RFMOs) are required to implement measures to prevent significant adverse impacts on VMEs. The objectives of the present study were to: update distribution models of VME indicator taxa in the South Pacific RFMO Convention Area; evaluate these against newly-collated independent field data to test the reliability of the presence-only habitat suitability models; and assess how well the updated models were able to predict into unsampled space. Ensemble habitat suitability models of 10 VME indicator taxa performed well using the newly collated data (AUC 0.95, TSS 0.76, and RMSE 0.34). There were no obvious patterns of decreasing model performance with decreasing environmental coverage; areas with few samples underpinning model predictions still had AUC 0.93, TSS 0.71, and RMSE 0.43. Despite these encouraging results, we also identify some important inherent issues with presence-only models that have profound implications for their use in management of VMEs. Future modelling efforts for VME management purposes should be based ideally on the use of survey presence-absence data and, preferably, abundance data.
机译：弱势海洋生态系统（VME）通常是脆弱的并且恢复缓慢，使得它们可能会因扰动而大大改变。在公海中，区域渔业管理组织（RFMOS）需要实施措施，以防止对VMES的重大不利影响。本研究的目标是：更新南太平洋RFMO会议区域的VME指标分类群的分布模型;评估这些针对新整理的独立现场数据，以测试仅限居住栖息地适用性模型的可靠性;并评估更新的模型能够预测到未夹持空间的程度。 10 VME指标分类群的集合栖息地适用性使用新整理数据（AUC＆GT; 0.95，TSS＆GT; 0.76和RMSE＆LT; 0.34）。随着环境覆盖率降低，没有明显的模型性能模式;有很少的样本的地区缺乏模型预测仍有AUC＆GT; 0.93，TSS＆ 0.71和Rmse＆ 0.43。尽管有这些令人鼓舞的结果，但我们还确定了一些重要的内在问题，这些内容仅适用于对VME的管理中使用的深远影响。 VME管理目的的未来建模努力应理想地基于使用调查存在性数据，而优选地，丰富的数据。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号