期刊
DEEP-SEA RESEARCH PART II-TOPICAL STUDIES IN OCEANOGRAPHY
卷 110, 期 -, 页码 84-91出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.dsr2.2014.10.017
关键词
Repeated measures; Regression analysis; Environmental monitoring; Oil and gas fields; Statistical analysis
类别
资金
- Suncor
This paper provides a general overview of repeated-measures (RM) regression designs and analysis for marine monitoring programs, in support of sediment chemistry, particle size and benthic macroinvertebrate community analyses provided as part of this series. In RM regression designs, the same n replicates (usually stations in monitoring programs) are re-sampled (i.e., repeatedly measured) at t > 1 Times (usually years). The stations provide variation in the predictor, or X variables. In the Terra Nova environmental effects monitoring (EEM) program, n = 48 stations were sampled in each of t = 7 years from 2000 to 2010. Two distance measures from five drill centres (sources of drilling wastes) were fixed predictor variables. RM regression designs are rarely used in environmental monitoring programs, but are often suitable and would be appropriate if applied to data from many monitoring programs. For the Terra Nova EEM program, carry-over effects, or persistent and usually small-scale variations among stations unrelated to distance, were strong for most sediment quality variables. Whenever natural carry-over effects are strong, RM designs and analysis will usually be more powerful and suitable than alternative approaches to the analysis. (C) 2014 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据