On the sensitivity of Optimum Multiparameter Analysis: a California Current System case study

Abstract

Eastern boundary upwelling systems, like the California Current System (CCS), represent a confluence and mixing of water masses whose variability and composition play a key role in modulating their high biological productivity and ecosystem variability. In the southern CCS, the relative contribution (i.e. proportion) and variability of its source water masses was quantified previously using an extended Optimum MultiParameter (eOMP) analysis, which is an inverse modeling technique that solves a set of linear mixing equations using quasi-conservative and non-conservative properties. However, there are several oceanographic decisions implicit in eOMP analysis that can generate uncertainties in the representation of the mixing and proportions of the source waters in a region. Here we quantify the sensitivity of these previous eOMP results in the southern CCS to varying oceanographic assumptions based on the uncertainty of the water mass properties, modified Redfield ratios, and alternate locations of the eastern tropical Pacific source waters. We show that the mean relative contributions of the main CCS source waters are more sensitive to the location of their selected source region (∼20–25 %) and the Redfield ratio (∼15–20 %) than to the uncertainty in the source water properties (∼2–5 %). Understanding the uncertainties of the eOMP assumptions benefits similar studies in other regions, especially in other eastern boundary upwelling systems (EBUS) where water masses characteristics and composition strongly impact the ecosystem.