Dieser Text ist zur Zeit nur auf Englisch verfügbar...
How do clouds influence the climate system? How do they form intense precipitation? Such questions have challenged atmospheric science for decades. High-resolution simulations and observations bring us closer to answering them, yet, fundamental processes are not understood.
This is because clouds organize on a range of scales, scales interact and precipitation is a result of abrupt changes of phases, leading to abruptness, e.g. intermittancy, in the moisture dynamics. Furthermore, convective-type cloud has been shown to produce unexpectedly strong precipitation intensity, not explained by equilibrium thermodynamics.
The team uses methods from theoretical physics to describe and model atmospheric processes. The aim is to capture emergent aspects, e.g. self-organization, that originate from small scales but can impact on larger scales. We use high-resolution simulations and observational data and make simplified, conceptual, models to capture key aspects of atmospheric complexity.
What is convection?
Convection is buoyancy driven
Convection is a consequence of strong heating of moist surface air. Warmer air expands and is therefore subject to buoyancy relative to the surrounding, more dense, air. This buoyancy causes rapid lifting of surface air. In contrast, stratiform cloud is often caused by cyclonic activity, where the collision of fronts causes large-scale lifting of moist air. Such lifting processes are generally slow compared to convective updrafts.
Convective precipitation is local
In general terms, the convective precipitation of a single convective updraft (see the image to the right) falls over relatively short times and covers areas of only few kilometers across. For extreme events, it is not uncommon to measure 50 mm of rain in one hour. Such strong rain can cause flash floods, local floods that arise within hours and can be devastating to human lives and infrastructure.
Convection is hard to simulate in climate models
Because of its local nature and complex dynamics, convection continues to pose basic challenges in coarse-scale models. Standard climate models have horizontal resolutions of roughly 50 km, while typical scales of convective events are far less than that, often around 1-10 km. To simulate single events, or their interaction, properly, models must have resolutions better than 1 km.
Stating the problem
Rayleigh-Bénard convection is a classical problem in complex systems science: When a fluid is placed in between two horizontal plates and the lower one is heated relative to the upper, so-called convective rolls will eventually appear. These rolls constitute a form of symmetry breaking, by which some areas see locally rising, others locally sinking fluid parcels.
The dynamics of the atmosphere can be modeled as a fluid for many practical purposes.
However, in contrast to classical convection, that of the atmosphere arises due to a fluid containing varying phases: The water vapor contained in the air can condense to produce cloud droplets, thereby releasing latent heat. Further, when rain falls to the ground by the action of gravity, latent heat is redistributed vertically. These are processes that can act to break the initially rising motion of moist air parcels, thereby - in a sense - destroying convective rolls.
Tackling the challenge
State-of-the-art global climate models are still far from capable of resolving scales of convective clouds througout the extended periods needed for climate simulations (at least 100 years). Yet, for short periods and smaller areas, so-called large-eddy simulations (see left column) can resolve turbulent processes down to tens or hundreds of meters. We use such simulations, which also represent simplifications of the real atmosphere, but do allow us to extract processes responsible for the organization of the atmosphere through convection and intensification of precipitation extremes.
While simulations are convenient in accessing processes, the most direct way to study convection is through meteorological observations of the atmosphere. The atmospheric state consists of a large array of variables, e.g. temperature, moisture, wind speed and pressure, just to name a few. Precipitation, however, is the quantity that most directly affects the living environment, e.g. humans. As it plays a crucial role in re-distributing energy and moisture within the climate system, and occurs as an intermittant process, it is particularly important to gather detailed, high-resolution observations on it. Some observations we currently use are those from ground-based stations, radar reflectivity and satellite observations.
Simulations and observations can tell us a whole lot about convective transport of moisture and energy. However, there is always a risk of taking a realistic simulation as meaning that the system is "understood". The more complex the simulation output, and perhaps the more it is visually compatible with observations or intuition, the less we actually grasp the abstract processes behind the formation and organization of clouds in the atmosphere.
We use methods from physics of complex systems to describe the self-organization of convective clouds. Requiring simplified models we can re-enact, at least qualitatively, some of the emergent phenomena seen in observations or simulations. We leave out some of the complexity — leading to better understanding and sometimes inspiring new predictions for future climate change.
Beyond this, simplified models can be more universal, describing similarities of disparate fields, such as atmospheric science and biology or even social science, fields that are also explored at Niels Bohr Institute.