Observations and Processes
Activity Leader: Dr Bernadette Sloyan
Staff and associates
Ensemble forecasts are produced by integrating a climate model forward in time from a set of observation-based initial conditions and fundamental understanding of the dynamical processes in the ocean, ice and atmosphere that result in long-lived coherent variability in the climate system. Thus the observations and process understanding are part of the same activity, although there a strong links/crossover to all activities. The observation and process theme will provide observational data for model initialization, and use observations and model s to identify key ocean-atmosphere processes that have the ability to drive multi-year coherent variability in the climate system
The technologies comprising ocean observing systems include a wide array of instruments and platforms (Figure 1). The heterogeneous nature of the ocean observing system
requires a sophisticated data integration and interpretation activity for all available in situ and satellite observations. We will provide integrated satellite and in situ observational
data sets for initializing the forecasting system, providing withheld data sets and data products for model validation and, providing process resolving observations to inform model
physics assessments and improve the realism of our models.
Figure 1. Main components of the Global Ocean Observing System (GOOS) that will provide in situ data to CAFE
for ocean initialization, validation, ocean process understanding.
As the forecast timescale increases (seasonal to decadal), processes in the ocean become increasingly more important. However, the generation of coherent ocean anomalies does not
guarantee a skilful forecast. Knowledge of the dynamical processes that generate coherent long-lived ocean is key to understanding the potential predictability for the climate system.
Here we will identify key ocean processes that drive multi-year coherent variability in the ocean and assess what is their impact on climate variability and climate forecasting.
Ocean Storm Tracks
Connecting the slowly evolving ocean signals to the tropics and mid-latitudes and how these anomalous ocean signals re-emerge at the ocean-atmosphere boundary is a key challenge
for a seasonal to decadal climate forecasting. Discovering how the ocean drives the low frequency variability of the mid-latitude atmospheric circulation is still very much an open
question, but of utmost importance is understanding droughts and climate extremes and in determining the predictability of the climate system.
There remain many unanswered questions on the fundamental nature and drivers of the production of ocean variability with the potential for it to remain coherent on the multi-year
to decadal time scale and how these anomalous ocean signals re-emerge at the ocean-atmosphere boundary. Our work will investigate the dynamical ocean-atmosphere-ice-coupling to
better understand the connection and feedback on the climate system.
The Decadal Climate Forecasting Project (DCFP) has embarked on building a “big data” climate science platform to address the opportunities and challenges in the large ensemble datasets generated by our models.
Given the explosive growth of these datasets it’s becoming increasingly impossible to interactively explore our climate model results and this is hindering scientific inquiry and discovery. The resulting fragmentation of workflow solutions renders most geoscience research effectively unreproducible and prone to error. Ultimately this “big data crisis” is not limited to the field of climate science. The solutions we are developing will almost certainly be directly applicable to CSIRO-wide data science challenges.
The vision, borrowed from Pangeo – a community platform for big data geoscience, is an interactive data exploration workflow that scales to meet our current and future challenges. The workflow is more open and reproducible, containerised and portable, and ultimately, collaborative. We have adopted a Pangeo-style approach and related python-based software stack that we’ve deployed on CSIRO HPC resources. The performance we have seen in our trials has been encouraging and also justifies the employment of CSIRO’s new parallel cluster file system, which is ideally suited for a scalable workflow.
|Bax, H.J; Ward Appletons, W; Brainard, Duffy; R. J. E; Dunstan, P; Hanich, Q; Harden Davies, H; Hills, J; Miloslavich, P; Muller-Karger, F; Simmons, S; Aburto-Oropeza, O; Batten, S; Benedetti-Cecchi, L; Checkley, D; Chiba, S; Fisher, A; Andersen Garcia, M; Gunn, J; Klein, E; Kudela, R; Marsac, F; Obura, D; Shin, Y-J; Sloyan, B; Tanhua, T; Wilkin, J; (2018) Linking capacity development to monitoring networks to achieve sustained ocean observation. Frontiers in Marine Science|
|Alford, M; Sloyan, B; Simmons, H; (2017) Internal waves in the East Australian Current. 44 12,280 – 12,288 Geophysical Research Letters https://doi.org/10.1002/2017GL075246|
|Sloyan, B; Craw, P; King, E; Neill, C; Kloser, R; Bodrossy, L; (2017) Future Technologies in B. Mapstone (ed) Oceans: Science and solutions for Australia 179-186 http://www.publish.csiro.au/book/7724/|
|Feng, M; Zhang, X; Sloyan, B; Chamberlain, M (2017) Contribution of the deep ocean to the centennial changes of the Indonesian Throughflow. 44 2859-2867 Geophysical Research Letters https://doi.org/10.1002/2017GL072577|