Large-scale forcing dominates interannual variability of Agulhas leakage

Sun, 30 Jul 2017 00:00:00 +0000

Motivation

In our early work, we have shown that Agulhas leakage is highly variable at daily to seasonal time scales, presumably dominated by the passage of Agulhas Rings. Decadal trends of Agulhas leakage are oftend attributed to the changes of Southern Hemisphere westerlies by previous studies. We hypothesize that, at interannual time scales, leakage variability is constrained by large-scale wind forcing, minifesting itself as changing large-scale velocity fields.

Method

Our work is based on a high-resolution configured Community Climate System Model (CCSM3.5) output. The spatial resolution of this simulation for the atmospheric and ocean component is 0.5 degree and 0.1 degree respectively. All the velocity fields used in this project is from an ongoing twentieth century climate change simulation with observed CO2 forcing. Agulhas leakage is quantified using a Lagrangian particle tracking approach, more details can be found in Cheng et al. [2016].

Beside the control case with actual vecolity fields, we designed three testing cases with slighly modified velocity fields. The velocity fields were first decomposed into large-scale fileds using a 500x repeating 9-point spatial smoother (Fig.1). The difference between the original and the large-scale fields are then taken as the eddy-fields. We shuffled or shifted the eddy fields’ temporal order, before plugged them back to the sequential large-scale fields. Forty years of data from 1961 to 2000 are used. By comparing leakage time series using different integration fields, we can show that, at interannual time scales, leakage time series among cases are highly coherent, which is attributed to the commonly-varying large-scale fields.

Diagram for shuffling and shifting experiments.

Time series comparison

Applying a 24-months low-possed fileds allow us to compare the interannual leakage variability between the testing cases and the reference case. As shown in Fig. 2, the timeseries correlate well above 0.6 (p<0.01). A Coherence analysis further support the hypothesis: leakage time series between cases are highly coherent over time scales longer than 1000 days (Fig. 3).

24-months low-passed leakage timeseries

coherence square of Agulhas leakage timeseries

Regional climate imprints

Leakage variability at this time scale is highly associated with the poleward shift of Southern Hemisphere westerlies, as shown in the regression of such to the low-passed Agulhas leakage time series (Fig. 4).

Westerlies response to interannual leakage variability

By regressing various climate variables from both atmosphere and ocean, we are able to show some regional climate patterns associated with interannual leakage variability (Fig. 5)

Regional imprints of low-passed Agulhas leakage

Future work

As the high-resolution climate change simulation still ongoing, we are eager to find out the long-term trends of Agulhas leakage and its effects on the regional climate, South Atlantic Ocean, and Atlantic Meridional Overturning Circulation. We would like to investigate the heat content and northward heat transport variability in the Southern Ocean in our coupled simulation, and relate such to Agulhas leakage.

Quantifying Agulhas leakage in a high-resolution coupled climate model

Wed, 30 Jul 2014 00:00:00 +0000

Motivation

The Agulhas current and its leakage of warm and salty indian-ocean water, play crucial role in the climate system by influencing Atlantic Meridional Overturning Circulation stability. While the ocean circulation models are not capable due to lacking of couplings, and the fully-coupled climate models fail because of inability to resolve mesoscale features, the high-resolution coupled model is the only viable tool to access this topic.

Dataset

Due to limitation of computing resources, and the interests of longterm time-scale variabilities, current generation coupled-climate models are conventionally configured to horizontal resolution of 1.0 degree, and outputs are usuall stored monthly. This configuration leads to poor performances in simulating mesoscale features. In Agulhas current context, these models fail to reasonably simulate features such as Agulhas return current, Auglhas Ring, and therefore Agulhas leakage. Here, my work is based on a high-resolution configured Community Climate System Model (CCSM3.5) output. The spatial resolution of this simulation for the atmospheric and ocean component is 0.5 degree and 0.1 degree respectively.

Quantifying Agulhas leakage

The first difficulty we must overcome is quantifying Agulhas leakage robustly.In numerical models, where we have gridded data, a widely accepted approach to quantify Agulhas leakage is Lagrangian particle tracking.

With the aid of Connectivity Modeling System (CMS), we release particles along the ACT array. Each particle is tagged with a volume flux equivalent to the local velocity times the corresponding grid cell size. They are advected by local velocity fields from CCSM3.5 forward in time for two years. Tracking every particle trajectories, we can determine whether one particle crossing the GoodHope line, which basin it ends up in, and, above all, the timing of last-crossing. Summing up the number of last-crossing particles of every time-step, we can establish time-series of Agulhas leakage.

Temporal Resolution

The approach mentioned above is fairly simple; however, there are many technical diffculties we must solve. First, lagrangian calculations in previous studies are mostly based on toolbox, which is not applicable to CCSM3.5 due to grid setup. Therefore, we use CMS instead. In addition, most of previos studies use Ocean General Circulation Model outputs with high temopral resolution (typically 5-daily output.), while most of Coupled Model outputs are archieved monthly, which is not ideal for lagrangian particle tracking.

To Adress these problems, we have tested different strategies, including releasing frequency, offshore boundary of releasing, particles with same the volume and so on. We also conducted several experiments using different temopral-resolution velocity fields to adress the degree of degradation of Agulhas leakage estimates of temporal-averaging.

p2d vs m2d Agulhas leakage timeseries

Future work

We’ve established a robust way to quantify Agulhas leakage in high-resolution CCSM3.5 output, with which we obtain much more resonable Agulhas Leakage estimates (~20Sv) than previous estimates based on coarse-resolution CCSM4 simulation. Our next plan is to apply such method to a complete 54year high-resolution pre-industrial CCSM3.5 run, and to a currently-running 20-Centuray climate change run to evaluate the variability of Agulhas leakage and possibliy to link such variability to other climate modes, such as El Niño, AMO and Indian Ocean Dipole. In particular, we are very interested in seeing the variability of Agulhas Leakage in the climate-change condition, because the inter-basin exchange is believed to counteract the global-warming effect on the AMOC.

Climate Variability |

Large-scale forcing dominates interannual variability of Agulhas leakage

Motivation

Method

Diagram for shuffling and shifting experiments.

Time series comparison

24-months low-passed leakage timeseries

coherence square of Agulhas leakage timeseries

Regional climate imprints

Westerlies response to interannual leakage variability

Regional imprints of low-passed Agulhas leakage

Future work

Quantifying Agulhas leakage in a high-resolution coupled climate model

Motivation

Dataset

Quantifying Agulhas leakage

Temporal Resolution

p2d vs m2d Agulhas leakage timeseries

Future work