The impact of remote sensing observations on cross-shelf transport estimates from 4D-Var analyses of the Mid-Atlantic Bight

This paper explores the impact of the individual components of a coastal ocean observing system on estimates of the circulation derived from a state-of-the-art analysis and forecast system for the Mid-Atlantic Bight and Gulf of Maine. The foundation of these activities is the Regional Ocean Modeling System 4-dimensional variational (4D-Var) data assimilation platform, which is run in support of the Mid-Atlantic Regional Association Coastal Ocean Observing System as part of the U.S. Integrated Ocean Observing System. The specific focus of this study is on the impact of remote sensing observations from both space-and land-based platforms on estimates of cross-shelf transport in the vicinity of the National Science Foundation Ocean Observatories Initiative Pioneer array. Sea surface temperature (SST) and sea surface height (SSH) were found to have, on average, a similar impact on the transport estimates. However, during a typical 3-day 4D-Var assimilation cycle, approximately two orders of magnitude more observations of SST than SSH are used in the model, and closer analysis shows that each altimeter measurement has approximately 50 times more impact on the transport estimates than an individual SST observation. This highlights the value of altimetry data for ocean state estimation, and the significance of expanding the altimeter constellation. The observations that are most impactful of all are in situ measurements of temperature and salinity, which have typically 3-4 times more impact than an individual SSH datum. A robust geographical distribution of the observation impacts emerges across a range of transport metrics which results from the combined influence of space-time dynamical interpolation and error covariance information within the 4D-Var system. The observation impact calculations suggest that High Frequency (HF) radar estimates of surface currents have relatively little direct influence on cross-shelf transport estimates. However, quantification of the sensitivity of these same estimates to changes in the observing system indicate that HF radar observations indirectly provide important information. This is understood in the current system by appealing to the idea of borrowing strength from the field of statistics in which some observations (satellite remote sensing in the case considered here) can borrow strength from other, seemingly less important observations. (C) 2019 COSPAR. Published by Elsevier Ltd. All rights reserved.

Automated summary

This study investigates the impact of individual components of a coastal ocean observing system on circulation estimates, highlighting the differing influences of various observations such as remote sensing, temperature, and salinity measurements. By utilizing the 4D-Var system in combination with parameters like spatial interpolation and error covariance information, a robust geographical distribution of observation impacts is established.