Quantifying horizontal length scales for surface wind variability in the tropical Pacific based on reanalyses

The characteristic spatial sampling length scales (for example, in the longitudinal direction) to adequately resolve the variability associated with a quantity to be observed, depends on the typical length scales of its variability. For example, if temporal variations in the longitudinal direction are correlated over thousands of Kilometers then an observing network with a similar spatial density may be enough to sample the variability with adequate fidelity. Given the importance of surface wind stress in the equatorial tropical Pacific in determining sub-surface oceanic variability, particularly in relation to the evolution of El Nino—Southern Oscillation (ENSO), in this analysis the horizontal length scales of surface wind stress variability based on two reanalysis products—the Climate Forecast System Reanalysis and the ERA-Interim—are documented. The analysis is of relevance to the future evolution of the Tropical Pacific Observing System. As the current design of the Tropical Atmosphere Ocean [that has a wider (narrower) separation in the longitudinal (latitudinal) direction] was originally envisioned following estimates of the length scales of wind variability based on the observational record from tropical Pacific Islands (Harrison and Luther in J Clim 3:251–271, 1990), we find a similar characteristics for length scales for the surface wind variability based on reanalyses datasets. Further, the inferences based on two reanalysis products have a large degree of similarity, and thereby, give us some confidence in the reanalysis products for the purposes of analyzing the length scales of surface wind stress variability, and further, providing some basic information about the requirements for a sustained observing system in the tropical Pacific to monitor and to predict ENSO.