MEPS 530:47-62 (2015)  -  DOI: https://doi.org/10.3354/meps11300

ABSTRACT: Ocean circulation models are useful for determining population connectivity, but are only available for a limited number of years. In contrast, meteorological reanalyses are available over decades. Since planktonic larvae are typically found in surface waters which are highly influenced by winds, the relationship between connectivity as estimated using an ocean circulation model and wind was used to develop a long-term hindcast of larval dispersal. The University of California Santa Cruz (UCSC) 31 yr Regional Ocean Modeling System (ROMS) hindcast of the California Current System was used to model inter-estuarine transport of larvae with a 6 d larval duration from 1981 to 2010, and between 3 and 8 connectivity patterns were identified using the self-organizing map (SOM) clustering algorithm. Regression models were developed for those connectivity patterns using meteorological reanalyses of winds. Training periods of 5, 10, and 30 yr were used for model development; in all cases there were strong associations between SOM connectivity estimates and winds. Regression models were validated using connectivity estimates from the ocean model. Validated regression models were used with winds from 1950 to 1980 to hindcast connectivity beyond the time range of the original ocean model. Connectivity as estimated from winds was correlated with the Pacific Decadal Oscillation and with upwelling from 1950 to 2010. Multi-decadal hindcasts of population connectivity can be carried out using meteorological reanalysis winds and statistical clustering of connectivity patterns derived from ocean hindcasts of 5 to 10 yr duration.

KEY WORDS: Connectivity · Dispersal · Self-organizing maps · Metapopulations

Full text in pdf format Supplementary material

Cite this article as: Oliver H, Rognstad RL, Wethey DS (2015) Using meteorological reanalysis data for multi-decadal hindcasts of larval connectivity in the coastal ocean. Mar Ecol Prog Ser 530:47-62. https://doi.org/10.3354/meps11300 Export citation Share:    Facebook - - linkedIn