About – University of Copenhagen


Please find here some information about Atmospheric Complexity as well as the team leader.

Atmospheric complexity

Climate research currently faces a dilemma: in a quest to make simulations more realistic, models are becoming increasingly sophisticated by incorporating more and more processes. Yet, despite increased apparent realism of the model output, basic understanding of the underlying physics becomes more challenging — the analysis of the model output itself sometimes resembles that of observational data. 

LES Simulation of moisture convergence in the boundary layer: Red areas indicate strong convergence and are associated with precipitation formation.

LES Simulation of moisture convergence in the boundary layer: Red areas indicate strong convergence and are associated with precipitation formation.

The Atmospheric Complexity team combines methods from non-equilibrium statistical physics, high-resolution simulations, and the best available observational climate data to improve process understanding of clouds and precipitation.

The team is led by Jan O. Haerter and is based at the Niels Bohr Institute, Copenhagen, Denmark. We collaborate locally with the Center for Ice and Climate, the Climate and Geophysics Group, the Center for Models of Life, and externally with the Max Planck Institute for Meteorology, Hamburg, Germany, the Swedish Meteorological and Hydrological Institute, Norrköping, Sweden, KNMI and TU Delft, Netherlands, as well at the ETH Zürich  

At Niels Bohr Institute there is a strong tradition for the study of self-organization in complex systems. Examples include self-organized criticality, evolutionary models, and conceptual models in biology

research funding: We are currently aiming at understanding the role of convective precipitation in a changing climate, an initiative funded by a generous Villum Young Investigator grant. More generally, we study self-organization of convection from small scales to large, within an ERC Consolidator Grant.

Team leaderTil toppen

Jan O. Haerter holds a PhD in theoretical condensed matter physics from the University of California at Santa Cruz, USA (2007). His thesis, advised by B. Sriram Shastry, dealt with the quantum mechanics of strongly interacting electron systems in two-dimensional frustrated systems, e.g. the Sodium-Cobalt oxide compound. 

Since his PhD, Jan has ventured into a range of complex systems topics. Predominantly, he has been engaged in the physics of the atmosphere, in particular the global hydrological cycle and precipitation extremes on varying spatial and temporal scales. He followed these topics during his postdoc at Max Planck Institute for Meteorology (MPI-M) during 2007-2010. During this time, Jan was involved in the European Water and Global Change Project (WATCH).

In 2011, Jan joined the Center for Models of Life (CMoL) at Niels Bohr Institute. At CMoL, Jan has studied a range of additional complex systems, taken from biology, ecology, and social network science. He followed this thread of theoretical complex systems science during a research visit at the University of Barcelona (2015-2016).

Since 2016 Jan leads the Atmospheric Complexity group at Niels Bohr Institute.

Publication list (google scholar)

Jan's further research interests:


Matrix showing possible interactions between species (blue squares), allowed interactions between adjacent trophic levels (gray) and forbidded interactions (white areas).

Food web assembly rules

In a sequence of papers, we have studied the classical generalized Lotka-Volterra equations for a food web consisting of set of many interacting species, from primary producers (plants, bacteria) to their consumers, higher-level consumers and top predators. Given a simplified food web, where each species only consumes species at the trophic level below, we were able to construct a set of rules for the numbers of species that can exist on any level. One outcome of this analysis is that the observed larger richness at intermediate levels, but relatively low richness at the bottom and top, come out naturally from our conditions.

Press release >>

relevant papers:

  • J. O. Haerter, N. Mitarai, K. Sneppen, Phys. Rev. E, 97, 022404 (2018)
  • J. O. Haerter, N. Mitarai, K. Sneppen, Food web assembly rules for Generalized Lotka-Volterra Equations.  PLoS Computational Biology, 12(2), e1004727 (2016) 
  • J. O. Haerter, N. Mitarai, K. Sneppen, Phage and bacteria support mutual diversity in a narrowing staircase of coexistence. The Nature ISME journal (2014).
  • J. O. Haerter and K. Sneppen: Spatial structure and Lamarckian adaptation explain extreme genetic diversity at CRISPR locusmBio3(4), e00126-12 (2012)
Social Networks

Three stages during the experiment, showing the presence of strong communication links between human subjects. Each line represents the connection between two individuals, stronger (weaker) links are shown in red (yellow). No color between nodes means that the link is not stronger than expected at random.

Trust and cooperation in social networks

We have studied, both using large email datasets as well as lab experiments with human subjects, how electronic communication between professionals is maintained and can evolve. A special focus lies on the effects of reciprocity, how this can lead to trust and social capital, and how time constraints force humans to prioritize some communication partners while neglecting others.

Press release >>

relevant papers:

Bistability, Noise, Cooperativity

Genomal regions surrounding a CpG island (center). Methylation level shown on vertical axis, note the decline in the vicinity of the CpG island.

Bistability in Epigenetics

Epigenetics is short for phenotypic modifications of organisms, e.g. humans or other mammals, that go beyond the information encoded in the genome (the DNA sequence). CpG methylation is one prominent epigenetic mark. CpG sites are sequences of Cytosine and Guanine base pairs along the DNA. They are capable of being modified by chemically bound methy groups, CH3. Generally, there are far fewer CpG sites along the DNA than expected by chance, but some regions, often in the proximity of genes, show unusually high CpG concentrations (about one every ten base pairs). Such so-called CpG islands tend to have generally low methylation levels (few methyl groups attached to the CpGs), while most other genomal regions have many methyl groups. Experimental results contain evidence for a process we term "collaboration", where the presence of one methylated CpG site may influence its surroundings to also attract methylation - and vice versa for the absence of methylation.

More generally, the process we describe can be seen as bistability in the presence of noisy transitions. Mathematically, bistability is the capability of a complex system of many constituents to maintain two meta-stable system states, i.e. an infinite-system analog would be able to have two fixed points. However, matters become much more complicated when noting that the system is neither infinite nor coarse grained, mean field, descriptions to apply.

relevant papers:

  • J. O. Haerter, A. Díaz-Guilera, M. Serrano, Noise-Induced Polarization Switch in Single and Multiplex Complex Networks, arXiv:1608.00428
  • C. Lövkvist, I. B. Dodd, K. Sneppen, J. O. Haerter, DNA methylation in human epigenomes depends on local topology of CpG sites, Nucleic Acids Research, gkw124 (2016)
  • J. O. Haerter, C. Lövkvist, I. Dodd, K. Sneppen, Collaboration between CpG sites is needed for stable somatic inheritance of DNA methylation states. Nucleic acids research, gkt1235 (2013)