Perils of correlating CUSUM-transformed variables to infer ecological relationships (Breton et al. 2006; Glibert 2010)

James E Cloern, Alan Jassby, Jacob Carstensen, William Bennett, Wim Kimmerer, Ralph Charles Mac Nally, David H Schoellhamer, Monika Winder

Research output: Contribution to journalLetterOther

10 Citations (Scopus)


We comment on a nonstandard statistical treatment of time-series data first published by Breton et al. (2006) in Limnology and Oceanography and, more recently, used by Glibert (2010) in Reviews in Fisheries Science. In both papers, the authors make strong inferences about the underlying causes of population variability based on correlations between cumulative sum (CUSUM) transformations of organism abundances and environmental variables. Breton et al. (2006) reported correlations between CUSUM-transformed values of diatom biomass in Belgian coastal waters and the North Atlantic Oscillation, and between meteorological and hydrological variables. Each correlation of CUSUM-transformed variables was judged to be statistically significant. On the basis of these correlations, Breton et al. (2006) developed the first evidence of synergy between climate and human-induced river-based nitrate inputs with respect to their effects on the magnitude of spring Phaeocystis colony blooms and their dominance over diatoms.
Original languageEnglish
Pages (from-to)665 - 668
Number of pages4
JournalLimnology and Oceanography
Issue number2
Publication statusPublished - 2012

Cite this