Abstract
Western boundary currents (WBC) balance the heat budget of global oceans by transporting warm waters poleward and vary in timescales from decades to days. WBC can interact with the bottom and provide deeper and nutrient-rich water to the continental shelf. Variability in WBC occurs on timescales of decades, seasons, or days due to external forcing (e.g. winds) or WBC intrinsic nonlinearity (sub- and mesoscale eddies).
The East Auckland Current (EAuC) is a WBC that connects the Tasman Sea to northeast New Zealand. Intra-annual variability in the EAuC was studied using a year-long timeseries of in situ and remotely-sensed velocity, temperature and salinity observations. In this study, we find that mesoscale eddies impact the long-term (> 30 days) EAuC variability. Four mesoscale eddies were observed between May 2015 and May 2016 (over a period of 260 days), generating distinct flows between the continental slope and rise. The EAuC dominated the circulation in the continental upper- and mid-slope and rise for 110 days and generated the most energetic events associated with wind forcing at periods between 4 and 32 days. Bottom Ekman transport, generated by the EAuC, caused the largest temperature anomaly (−1.5 °C) at the continental upper slope (500 m depth).
Ocean reanalyses using the Regional Modelling System in conjuction with 4-dimentional variational data assimilation method (ROMS 4D-Var) are developed to better understand cross-shelf exchange in the EAuC system. We assimilate sea surface height (SSH) and temperature (SST), subsurface temperature, salinity, and velocity from three moorings located at the continental upper- and mid-slope, and rise using a 7-day assimilation window. Sensitivity tests are conducted to elucidate the importance of subsurface observations in the quality of the reanalysis. Assimilation of velocity subsurface data is important for improving the mesoscale field up- and downstream of the moorings' locations in comparison to assimilating surface fields (sea surface height (SSH) and temperature (SST)) only. By improving the representation of the mesoscale eddy field, data assimilative runs increased complex correlation between modelled and observed water column velocity vectors in all experiments. In situ subsurface temperature is found to be of utmost importance to correctly simulate the top of the thermocline - one of the most difficult regions to simulate in ocean models. Assimilation of moored CTDs data have little impact in correcting model salinity, however, increasing the tracer decorrelation length scales and using a 2-day assimilation window (instead of a 7-day window) improves model salinity in comparison to independent Argo data. The 2-day window simulation overall better matches the observations but the 7-day window simulation is more dynamically consistent and better suited to study physical processes in the EAuC system.
The 7-day window ocean reanalysis and a freely evolving simulation are used to study drivers of cross-shelf exchange in the region. These one-year simulations are used to study the impact of the EAuC and its associated eddying variability on exchange between the continental shelf and slope. The EAuC is found to drive bottom Ekman transport with variability from 4 to 60 days in the first and last thirds of the New Zealand’s northeast shelf. In the mid northeast shelf region, other processes such as submesoscale and frontal eddies dominate cross-shelf velocities. We use an eddy tracking algorithm to identify, classify and analyse eddy-driven impact on cross-shelf exchange. Mesoscale eddies does not show an impact on cross-shelf exchange, however, smaller eddies (radius < 30 km) are ubiquitous on the shelfbreak (200-m isobath) and force on- and offshore velocities while travelling south. We find that cyclones represent most of the eddies that reach the shelfbreak (200-m isobath) and live longer (up to 70 days) compared to anticyclones (< 20 days). Coastal cyclones were smaller eddies (radius < 24 km), formed on the continental shelf and have submesoscale characteristics (Rossby and Richardson numbers O(1)) and can travel offshore representing export of shelf waters. On the other hand, slope cyclones formed in deep waters (>200 m), had transitional characteristics between submesoscale and large mesoscale eddies (radius > 50 km) and can hardly perform short onshore incursions. Coastal (slope) eddies tended to travel south along the shelf break and generate uplift of upper-slope water onto the shelf with vertical velocity of ~30 m/day (~15 m/day) causing negative temperature anomalies of -1.5°C (-1°C).
These results represent a substantial step change in knowledge of the understanding of the EAuC and its eddying variability impact on cross-shelf exchange. The EAuC is driven by local winds rather than basin-scale wind stress curl and has its flow largely modified by locally or externally formed mesoscale eddies. The EAuC can generate colder temperature anomalies compared to winds via Ekman dynamics. Data assimilation is vital for accurate simulation of the mesoscale field and subsurface temperature assimilation is of utmost importance for correcting the modelled top of the thermocline. Coastal and slope (frontal) eddies tend to travel south along the shelfbreak forcing colder waters onto the shelf which might impact primary productivity in the region.