UiT The Arctic University of Norway has established the UiT Aurora Centre for Nonlinear Dynamics and Complex Systems Modelling (DYNAMO). This is an interdisciplinary centre with thematic focus on the most profound challenges for our society – climate change, its impact on ecosystems, and its mitigation by development of fusion power as an inexhaustible source of sustainable energy.
The enormous success of natural sciences in the industrial age has been built on the paradigm of reductionism, which attempts to explain systems in terms of their constituent parts and the individual interactions between them. Empirical laws have been explained by fundamental principles that have been experimentally verified with remarkable precision. However, largely due to the revolution in scientific computing, we are now discovering new scientific laws that emerge from non-linear interactions in systems with many degrees of freedom which cannot be deduced from the first-principles-based equations that define the systems. Such complex systems have many components which may interact with each other resulting in a behaviour which is intrinsically difficult to model and predict due to the dependencies, competitions, and relationships between their parts or with their environment. Distinctive properties of complex systems are chaos, nonlinearity, emergence, spontaneous order, adaptation, and feedback loops, among others.
A major part of complex systems science is devoted to universality and commonalities between systems with strong non-linear interactions and collective behaviour. However, a general theory of such systems has not been developed. The most important advances in the field has come through a combination of domain-specific modelling as development and application of conceptual mathematical and statistical models, and general mathematical concepts such as symmetry, invariance and stability. Theories of turbulence, intermittency, fractal geometry, multifractal measures and self-organized criticality are all developed as tools to describe emergent phenomena in complex systems.
The UiT Aurora Centre DYNAMO is organized in three research themes: Arctic ecosystems, the climate system, and fusion plasmas. These are areas of research where complex systems modelling is particularly fruitful. The implementation is structured into four work packages corresponding to the main methodological frameworks: long range dependence and temporal scale invariance, bifurcations and transients, instabilities and chaotic dynamics, and ensemble studies. The main objective is to develop and combine dynamical, phenomenological and stochastic models in order to describe non-linear dynamics and emergent structures common to complex systems.
The centre project period is January 1, 2020 to December 31, 2024. Centre leader is professor Odd Erik Garcia and deputy leader is professor Martin Rypdal. The centre is based on the research group Complex Systems Modelling at the Department of Mathematics and Statistics and the Department of Physics and Technology.
April 21, 2021 DYNAMO enter research collaboration with MIT Plasma Science and Fusion Center with funding from Equinor. Three new doctoral and postdoctoral fellows will be hired in the center. The agreement gives UiT employees access to measurement data from MIT, and DYNAMO doctoral and postdoctoral fellows will have possibilities for a research stay at MIT. At UiT, we will use our expertise to predict how MIT-PSFC will maximize scientific result from the SPARC experiment. UiT published a news article about the collaboration agreement (in Norwegian).
December 11, 2020 News article: Newspaper, Dag og Tid, publishes a news article about fusion energy, including interview with centre leader Odd Erik Garcia (in Norwegian).
November 14, 2020 News article: NRK publishes a news article about fusion energy commitment in Norway, including interview with centre leader Odd Erik Garcia (in Norwegian).
August 10, 2020 DYNAMO welcomes Magdalena Anna Korzeniowska! Magda has just started her PhD in complex systems modelling at DYNAMO. She will be working on modelling self-organized critical (SOC) dynamics and long-range dependence in physical systems, supervised by centre leader, prof. Odd Erik Garcia. Magda completed her Master´s degree in applied mathematics at the Department of Mathematics and Statistics at UiT The Arctic University of Norway.
June 15, 2020 News article: Local newspaper, iTromsø, publishes news article about centre leader Odd Erik Garcia's role in Equinor's recent invenstments in fusion energy (in Norwegian).
June 03, 2020 News article: Local newspaper, Nordlys, publishes interview with centre leader Odd Erik Garcia about fusion energy and Equinor´s investments in fusion energy research (in Norwegian)
May 26, 2020 Debate: Tu publishes a debate post by centre leader Odd Erik Garcia, discussing Norwegian contribution to develpoment of fusion energy (in Norwegian).
February 14, 2020 News article about DYNAMO: Local newspaper, iTromsoe, pusblishes a news article about the Dynamo centre, including interviews with centre leader Odd Erik Garcia and deputy leader Martin Rypdal (in Norwegian).
February 6, 2020 Prediction of Corona virus development: UiT publishes a news article with DYNAMO-scientist Martin Rypdal, who has used a non-linear mathematical model in order to predict the spread of the corona virus in China (in Norwegian).
February 4, 2020 News article about DYNAMO: The regional newspaper Nordlys has published a news article about the DYNAMO-centre, including interviews with centre leader Odd Erik Garcia and deputy leader Martin Rypdal.
January 30, 2020 Tenure-track positions announced: We are looking for 1-2 Associate Professors in applied mathematics, statistics, climate science, scientific computing, or complex systems modeling. For each associate professor we offer an excellent start-up package with reduced teaching, funding for PhD students, and mentoring. See more under Open Positions.
January 27, 2020 UiT publishes a news article about DYNAMO. Here you can read interviews with the UiT Aurora Centre research theme leaders explaining the motivation behind the research activities, the link between the research themes and the goal of the investigations (in Norwegian).
January 24, 2020 A new RCN project strengthens research in the DYNAMO center: The Research Council of Norway announced that they will fund the project "NORTHERN FOREST: A multi-driver framework for near- term iterative forecasting of ecosystem states" with approximately 12 MNOK. DYNAMO research theme leader Nigel Gilles Yoccoz, who is a work-package leader in the new project, says that the proposed research will provide more and better data for model building and thereby benefit the DYNAMO-centre. Congratulations to everyone involved!
January 15, 2020 First PhD position announced: The DYNAMO center has announced its first open position – a joint PhD Fellowship between UiT The Arctic University of Norway and the Niels Bohr Institute at the University of Copenhagen. Application deadline is February 16. See more under Open Positions.
January 1, 2020 UiT Aurora Centre DYNAMO is established: UiT The Arctic University of Norway has allocated funds for research groups that demonstrate excellence and that need to strengthen their capacity in order to be successful in the competition for larger external funding. Applications for two new UiT Aurora Centres were subject to an external peer review process. The DYNAMO project description received the highest score on all evaluation points and its funding was announced in a UiT news post on December 18.
Small mammal, insect and fish dynamics have provided some of the best-known examples of complex dynamics in ecosystems due to interactions within and between populations, including harvesting by humans. One common characteristic to all Arctic ecosystems is the strong seasonality (long winters and short summers), which plays an important role in the transition from stable or cyclic dynamics to strongly chaotic or more complex dynamics. Climate change will obviously impact the seasonality of Arctic ecosystems, and therefore the dynamics of key populations at different trophic levels. There is, however, only a limited understanding of how seasonality affects ecological models, such as for predator-prey and host-parasite interactions, and this project will aim at understanding under which conditions season length and contrasts affect transitions between dynamical regimes. We will in particular investigate how the complexity of the modelled trophic interactions (e.g., including or not age or size structure, how different trophic level respond to changes in seasonality) and the way stochastic components are incorporated and affect the resulting dynamics, and how short-term and long-term predictions depend on such model complexity.
Ecosystems can also respond abruptly to slowly changing conditions, with for example temperature change leading to changes in system stability. Lake eutrophication, fishery collapse, insect outbreaks, algal growth on corals or desertification are well-known examples. Most ecological time series are, however, either too short or of too poor resolution to immediately understand such transitions, or to evaluate model predictions using for example historical data. In particular, the concept of early-warning signs of abrupt transitions, in the form of critical slowing down, has been argued to be non-robust for abrupt ecological transitions. Early-warning signs are not only important with respect to mitigation and adaptation, but also to determine the dynamical mechanisms that drive these abrupt transitions. They also provide insight into the structure of the underlying processes. Understanding of these mechanisms is crucial for building conceptual models that describe the effects of drivers on species, and they can be used to predict the evolution of systems in a warming climate. It is of critical importance to determine the type and quality of data that are needed to predict and detect tipping points in ecosystems that cannot be manipulated at a relevant scale. In the context of fisheries, an increased variability of fish abundance can result from harvesting, but this may depend on complex interactions with environmental stochasticity and demography, as well as the specific model used. Change in variability does not necessarily lead to irreversible regime shifts. To address how the ecological context can influence the critical slowing down we will focus on the following two topics:
Fish are important resources, and reliable detection of critical slowing down is essential as a management tool. A range of different ecological models exist of various degree of complexity and input needed, depending on the aim. Ongoing changes in the Barents Sea ecosystem include all trophic levels, and management-based monitoring combined with process and food web studies as part of projects like ClimeFish (EU funded) and the Nansen Legacy (RCN funded) provide data on shorter and longer time scales.
Research theme leader: professor Nigel Yoccoz
Multi-scale behaviour is a defining property of complex systems and an inherent property of the climate system, both spatially and temporally. A complete description of the Earth’s climate would span spatially from the size of micro-particles in the atmosphere to the diameter of the planet, and temporally there are non-trivial fluctuations on scales ranging from seconds to millions of years. In the time period since the last glaciation, the Holocene, the climate has been stable, but complex, nevertheless. Perturbations of the Earth’s radiative balance, caused by e.g. volcanic eruptions or internal variability, cause the temperatures in the atmosphere and mixed ocean layer to respond quickly to restore energy balance. But the full equilibration takes thousands of years due to the large thermal inertia of the deep oceans. The resulting climate fluctuations exhibit variability on all time scales, and characteristic scales are difficult to reveal. The power spectral density of global temperature records is an approximate power-law (scale invariant) function of the frequency over several orders of magnitude. Similar results are found for past temperature reconstructions, with power-laws being accurate approximations across a vast range of temporal scales. Accordingly, long-range dependent, scale-invariant, stochastic processes are characterizing the global temperature fluctuations.
Most research on long-range dependence in climate science has focused on its implications for detection and estimation of trends, whereas some work has dealt with the more controversial notion of scaling regimes. Here the conventional wisdom has been that long-range dependence in temperature fluctuations is produced by non-linear energy transfer, analogous to turbulence. In recent years however, the research group at UiT has demonstrated that long-range dependence can be fully explained as a linear multiscale response to radiative forcing and chaotic atmospheric dynamics. The idealized versions of such models provide linear responses where perturbations decay algebraically in time. These fractional energy-balance models do not conserve energy, and the equilibrium climate sensitivity is not well-defined. Paradoxically, these models may still hold the key to obtain better constraints on the equilibrium climate state, which despite years of extensive research has proven difficult to estimate accurately. A solution proposed here is to combine the fractional approach with the notion of scale-dependent sensitivity. First results in this direction show that scale-dependent sensitivity estimated from historical data correlates strongly with equilibrium climate sensitivity in complex climate models when it is evaluated at centennial and millennial time scales.
In addition to temporal scale invariance, there are non-linear feedback mechanisms that alter the system characteristics as the planet warms. These feedbacks may act differently in different climate states, potentially giving rise to tipping cascades that will invalidate the linear notion of climate sensitivity. The effect of non-linear feedbacks is especially evident in the Arctic region which exhibits high variability and is warming rapidly in response to current increase of greenhouse gas concentrations in the atmosphere. The warming can trigger abrupt and potentially irreversible climate transitions. The Greenland ice sheet is melting at an accelerating rate, and may already have crossed a critical threshold where an almost complete meltdown is inevitable. Retreating Arctic sea-ice cover can further accelerate Greenland runoff, and fresh meltwater from the Greenland ice sheet is already slowing down the upper cell of the Atlantic meridional overturning circulation, a non-linear component of the climate system. Abrupt changes in the Atlantic meridional overturning circulation will have immediate global effects, likely inducing accelerated melting of the East Antarctic ice sheet. Such sequences of abrupt transitions in coupled climate components are observed in paleoclimatic records, and recent evidence suggests that polar tipping cascades drove the early part of the last Interglacial warming. Thus, if anthropogenic forcing pushes the Arctic climate beyond a certain warming threshold, the global implications may be dramatic. Understanding the role of the Arctic and determining the thresholds for abrupt climate change are among the greatest challenges in climate science of today.
Complex climate models contribute substantially to our understanding of the climate system and the feedback mechanisms in the Arctic, but advances in modelling capabilities are unlikely to resolve key questions. There is insufficient understanding of the plethora of physical subsystems and their interactions, and it has also been argued that complex climate models are inherently built for stability. Subgrid-scale processes are not resolved and have to be parameterized to obtain a closed description of the system. The overwhelming complexity of the Earth system is commonly handled by a modular approach where the system is partitioned into suitable subcomponents such as the atmosphere, the oceans, the cryosphere, vegetation, and other land surface processes. Hence, in addition to subgrid-scale parameters, a second class of free parameters is introduced in the course of coupling different model components. These parameterizations are mainly performed against the instrumental record of the last 150 years, and in some cases against paleo-reconstructions for the last few thousand years. Indeed, the calibrated models perform well in reproducing the comparably smooth climate variability of the last few thousand years. However, state-of-the-art climate models driven by orbital forcing have problems in simulating abrupt transitions that are evidenced in paleoclimatic records. Prominent examples are the abrupt Dansgaard-Oeschger events in Greenland during the last glaciation. Recently, such transitions have been reproduced in complex models, but always by specifically conditioning the models for this purpose. For instance, stadial to interstadial transitions have been simulated in Earth System Models by cutting off an imposed fresh water forcing applied to a Marine Isotope Stage 3 control simulation.
In contrast to complex climate models, simplified models can exaggerate non-linear effects. A well-known example is the sea-ice albedo feedback, which in simple energy-balance models gives rise to instability of a small ice cap. This instability is not seen in more complex models. There is a need for better understanding of the relation between model complexity and stability of the modelled Arctic climate. In order to achieve this, we need a new methodology for quantifying stability and state-dependent sensitivity in complex climate models. In simple low- dimensional models, dynamical systems theory provides a range of tools for this purpose, but these are generally not applicable to the complex models. Loss of stability can only occur in non-linear systems, and in climate models it is generally associated with an increase of the state-dependent climate sensitivity. In complex models, probing of stability requires model experiments starting from a wide range of background states. So far model studies have mainly focused on background states close to the present climate. Long model runs subject to instantaneous quadrupling of CO2 concentration show that “fast” feedbacks, primarily the cloud response to changing sea-surface temperature, exhibit increased sensitivity in a warmer state. But there is little knowledge about the state-dependence of the Arctic feedbacks on different time scales. These questions call for studies of the ocean-, ice- and atmosphere interactions, in particular of the resilience of ocean circulation patterns induced by rapidly melting ice sheets.
Research theme leader: professor Martin Rypdal
For more than half a century, there has been an internationally coordinated research program to develop controlled thermonuclear fusion as an unlimited and sustainable source of electrical power without emission of greenhouse gases or long-lived radioactive waste. The leading methods are based on magnetic confinement of the hot plasma in toroidal geometry. However, research has shown that there is substantial transport of particles and heat across the magnetic field lines, leading to detrimental plasma interactions with the reactor main chamber walls.
The traditional approach to describe the cross-field transport of particles and heat is based on standard mixing length estimates, assuming small relative fluctuation levels and that the fluxes are proportional to local gradients in the particle density and temperature profiles. This is the basis of so-called transport modelling of the boundary plasma, using simplified models to estimate the average particle densities and temperatures. These models are sophisticated in their treatment of realistic geometry and interaction between electrically charged and neutral particles, and are presently used for predictions and design of the next generation fusion experiments. The transport across magnetic field lines is treated as a simple diffusion process. This approach is used despite the ample experimental evidence that the boundary region is always in an inherently turbulent state with relative fluctuation levels of the particle densities and temperatures of order unity.
Recent experimental measurements and theoretical modelling have revealed that the cross-field transport of particles and heat is dominated by radial motion of filament structures through the boundary region. The average radial particle and heat fluxes evidently depend on the filaments size, velocity and amplitude distributions as well as their frequency of occurrence. Revealing the statistical properties of plasma fluctuations in the boundary layer is thus crucial for predictions of transport and plasma–wall interactions. It follows that both first-principles-based modelling and statistical treatments are required in order to fully describe these phenomena.
The fluctuation-induced transport results in high particle densities and temperatures in the boundary and enhanced interactions between the plasma and the main chamber walls. This leads to erosion of wall material and release of impurities into the confined plasma. The former will ultimately define the plasma facing components life-time, while the latter may cause significant radiation in the core plasma and limit the power performance of fusion reactors. The radial transport is an emergent property of fusion plasmas produced by instabilities and turbulent motions. The complex plasmas exhibit chaotic dynamics, intermittent fluctuations and in some cases long-range temporal dependence and scale-invariance.
In order to reveal the statistical properties of the fluctuations, dedicated experiments will be performed to record exceptionally long fluctuation data time series at fixed spatial positions in the relatively cold boundary layer under stationary plasma conditions. International collaborative partners in the project have committed to perform dedicated measurements on all the leading fusion experiments in order to obtain fluctuation data time series of unprecedented duration and with high sampling rate. A major goal of this project is to obtain similar data sets across multiple experimental devices and to compare the fluctuation statistics for conventional tokamaks, spherical tokamaks and stellarator designs. The resulting data sets will allow unambiguous identification of the fluctuation statistics, including the fluctuation probability density functions, frequency power spectra, amplitude distribution, waiting time distribution, rate of level crossings and excess times.
Output data from long-run turbulence simulations will be analysed in the same manner as the experimental measurement data and this part of the project will be an essential contribution to code validation, eventually leading to first-principles-based predictive capabilities of fluctuation-induced material erosion for the reactor chamber walls. Both the experimental measurement and the numerical simulations data will be analysed and interpreted in the context of a stochastic modelling framework developed at UiT. Successful code validation requires that all fluctuation statistics comply with those derived from the experimental measurements. Such stringent tests of turbulence simulations have never before been undertaken.
Research theme leader: professor Odd Erik Garcia
Manuscripts in preparation:
2 Postdoctoral Research Fellows within climate science and complex systems modelling will be announced.
5-6 PhD positions in climate science, fusion plasma physics and complex systems. We will announce a number of PhD Fellow positions within these research themes in the UiT Auroura Centre.
All positions, when announced, will be made available at Jobbnorge. Use "UiT Aurora Centre DYNAMO" as search string to view them all.