Absence of scale dependence in dolphin–habitat models for the eastern tropical Pacific Ocean

Marine Ecology Progress Series

L. T. Ballance¹  E. A. Becker¹  J. V. Redfern¹  J. Barlow¹  T. Gerrodette¹ 

关键词: Dolphin density;    Habitat modeling;    Striped dolphin;    Stenella coeruleoalba;    Eastern spinner dolphin;    Stenella longirostris orientalis;    Short-beaked common dolphin;    Delphinus delphis;    Risso’s dolphin;    Grampus griseus;   

DOI  :  10.3354/meps07495

学科分类：海洋学与技术

来源: Inter-Research

【 摘 要 】

ABSTRACT: Research into the effects of scale on cetacean–habitat relationships is limited and has produced ambiguous results. We explored the effects of spatial resolution (a component of scale) on dolphin–habitat models using 4 yr of data collected in the eastern tropical Pacific Ocean (ETP). We developed generalized additive models of dolphin–habitat relationships for 4 species at 6 spatial resolutions using oceanographic and geographic habitat variables. For all species, the ecological patterns in the models built at the different resolutions were similar: the same variables frequently occurred at multiple resolutions and had similar functional forms, and maps of predicted distributions identified similar high and low density regions. Additionally, interannual habitat variability, which is most likely related to the El Niño–Southern Oscillation, had a greater impact on the predictive power of dolphin–habitat models than spatial resolution. Although it is common to find scale dependence in species–habitat relationships, domains of scale exist in which ecological patterns do not change. The absence of scale dependence in the models for the 4 dolphin species suggests that resolutions from 2 to 120 km occur within a single domain of scale in the ETP. This domain of scale may be determined by the physical oceanography of the ETP, which is generally defined by large-scale processes. Although resolutions from 2 to 120 km appear to occur within a domain of scale, building models at the larger resolutions we investigated may reduce the noise in the data due to false absences.

【 授权许可】

Unknown