New models for storm track variability

Lead Research Organisation: University of Reading
Department Name: Meteorology


Storms tracks are the localized regions in the atmosphere, largely over the oceanic basins, where midlatitude weather systems grow and develop. They are the dominant dynamical feature of the midlatitude atmosphere. For example, the North-Atlantic storm track determines the weather in Western Europe and its mean properties are a central factor in determining Northern Hemisphere climate. Anomalous storm track structure is the origin of persistent weather regimes such as cold periods or wet periods.

Our understanding of storm tracks goes back to detailed conceptual models developed from the 1970s onward. These models describe how midlatitude weather systems freely develop into their mature phase and how the jet stream is modified by this development. More recently, the focus has been shifting to questions about the structure of storm tracks under changed climatic forcings. This is still very much an open question; in fact we are not even sure about the basics, such as: will the storm track move in a future climate and if so in what direction?

The reason for this uncertainty is that the temporal and spatial structure of the storm track is the result of strong interactions between climatic forcings, storms and the jet stream. To think of storm tracks as a set of storms growing on a fixed jet stream misses this key structural property of storm tracks. This highlights the need to understand the storm track as a non-linear forced-dissipative system.

This proposal builds on very recent developments published by the PI and a NERC funded PhD student, where it was shown that under forced-dissipative conditions, storm tracks in fact do not satisfy the traditional grow-and-mature model, but rather act as a non-linear oscillator where the dynamics is dominated by a periodic exchange of "energy" between weather systems and the underlying jet. This non-linear oscillator model has been shown to have many realistic properties observed in real data and for the first time has allowed an understanding of the variability of a mature storm track in the context of a relatively simple model.

The ultimate aim of this proposal is to understand and predict how the spatial and temporal structure of storm tracks is established under given climatic forcings. There are several gaps in our new model which need filling and which will allow us to start using it in a more predictive way:

1) We need to build a solid theoretical underpinning of the oscillatory model for the storm track in order to pin down parameter dependencies in the model. The present motivation of the model is based on physically plausible but ultimately heuristic arguments. We also need to extend our framework to capture some of the spatial variance in the storm track, specifically to get a better handle on possible storm track shifts under climate change scenarios.

2) The predictions of the oscillator model need to be tested in a hierarchy of model contexts. The oscillator model has been described in theoretical terms and demonstrated in real observations, and by extension in fully fledged weather forecast models, but it has not yet been examined systematically in a model context. To do this we will work with simplified general circulation models, building on modeling configurations that have been published in the literature.

3) One of the most promising features of our work will be that it provides a new framework to interrogate data from reanalysis or climate models. For example, our model predicts a different behaviour for the response of the potential for storm growth under climate change scenarios compared to our current understanding of the storm track. We will be examining several of these issues in the set of CMIP5 runs, which were used in the IPCC report on climate change.

The present proposal will form the basis of a new understanding of the observed storm track and will have fundamental consequences to our description of the midlatitude climate system.

Planned Impact

Although the proposed research is focussed on academic impact, we envisage a useful interaction with partners outside academia. The proposal is about the dynamics of the eddy driven jet which determines the surface winds on the downstream end of the storm track, that is, W. Europe in the case of the N. Atlantic storm track. Because of this focus our work will be of interest to energy suppliers and traders as well as insurance companies who all rely heavily on accurate predictions and understanding of statistical properties of surface winds.

Energy traders and suppliers, and insurance industry would benefit from the proposed research on immediate timescales. Both of these industries rely heavily on small competitive advantages in predicting and understanding low level winds. For example, we are currently running a joint MSc project about storm track variability precursors with EDF Trading, and they want to continue this collaboration in relation to the present project project. EDF Trading have seen our proposal and impact plan and are supporting it.

The aforementioned industries rely on small competitive advantages in using weather forecast products. Such competitive advantages follow from smart uses of products that all competitors have access to. Therefore there is a great interest in details of, for example, the statistics of low level wind variance or possible precursors thereof. Such information feeds directly in the weather forecast advice that meteorologists produce for their energy traders or their insurance brokers.


10 25 50
publication icon
Novak L (2017) Marginal stability and predator-prey behaviour within storm tracks in Quarterly Journal of the Royal Meteorological Society

publication icon
Tailleux R (2018) On the Local View of Atmospheric Available Potential Energy in Journal of the Atmospheric Sciences

publication icon
Harvey B (2018) Baroclinic Adjustment and Dissipative Control of Storm Tracks in Journal of the Atmospheric Sciences

Description The aim of our project was to identify to what extent a particular simplified model for the interactions between the jet stream and waves on the jet stream can be used to explain observed variability of storms hitting western Europe, and also what kind of changes we can expect in those storms in a fututre climate. We found that our simplified model works incredibly well. In fact, it works better than our theoretical understanding currently supports. We have found evidence of so-called "frictional control" of the mid latitude jet stream, giving the counterintuitive effect that an increase in surface friction actaully can lead to an increase in the jet speed. We also found evidence that a change in forcing of the mean climate (such as in a future climate) will actually show up in changed storm intensity and frequency, and not so much in a change in mean climate. We have also checked whether this view is supported in the present state of the art climate model predictions for our future climate. We indeed found that all climate models individually behave as we expect, but that climate model biases cannot be explained by the dynamical biases that our model would suggest: the important implication is that model biases in the jet stream (a key problem in current climate models) cannot be repaired by improving local dynamics of the jet, but must be improved by adjusting remote and global forcing mechanisms.
Exploitation Route Several research groups have picked up our analysis and are taking this further.
Sectors Energy,Environment

Description Roy. Met. Soc. Legacies fund
Amount £600 (GBP)
Organisation Royal Meteorological Society 
Sector Charity/Non Profit
Country United Kingdom
Start 03/2017 
End 04/2017
Description international exchange grant
Amount £11,000 (GBP)
Funding ID IES\R1\180099 
Organisation The Royal Society 
Sector Charity/Non Profit
Country United Kingdom
Start 07/2018 
End 08/2020
Title Phase space averaging for wave-mean flow interactions 
Description We have develped, tested, and applied a novel way of taking climatic averages, but not in time, or in space, as is routinely done, but in a two-dimensional phase space describing the mean jet stream structure and the perturbations in the et stream. This way these two key variables, which are linearly unrelated (approximately) using normal analysis tools, show physically realistic and important interactions. 
Type Of Material Data analysis technique 
Year Produced 2017 
Provided To Others? Yes  
Impact Besides the paper dscribing the tool and initial analysis (plus two papers in preparation), we have applied our analysis to ocean dynamics with a research group in Oxford (leading to a publication) and have influenced several other research groups to start using this analysis, including atmospheric sciemce at Caltech, Weizman, Oxford, Exeter, Bologna. 
Description l'Aquila/OpenIFS 
Organisation University of L'Aquila
Country Italy 
Sector Academic/University 
PI Contribution An intensive collaboration has been started with the group at l'Aquila/ICTP Trieste, to test some of our ideas in the idealized models as well as the OpenIFS they use. Novak has visited Trieste for one week for collaboration. We have hosted Paulo Ruggieri for extended visits.
Collaborator Contribution Paulo Ruggieri has made extended visits to our Department to work on this topic, and he has hosted Novak for a week for collaboration in Trieste. He co-organized a workshop in Trieste at which Ambaum is a keynote speaker.
Impact Collaborative visits
Start Year 2016