• Clouds and Climate Tipping Points

    Tapio Schneider’s Watson Lecture at Caltech (April 24, 2019). Low clouds over subtropical oceans cool Earth’s climate because they reflect much of the sunlight shining on them back to space. They are fundamentally important for regulating Earth’s climate, yet difficult to capture in global climate models. This lecture presents results from high-resolution simulations of subtropical stratocumulus clouds. The results points to a possible tipping point of the climate system: if greenhouse gas concentrations rise high enough, subtropical low clouds may break up, triggering dramatic global warming. This may explain how the Arctic once could have been warm enough to harbor crocodiles, as we know it did 50 million years ago, when CO2 concentrations were 3 or 4 times higher than they are today.

  • https://vimeo.com/120035135

    Challenges in Climate Dynamics

    Tapio Schneider’s introductory lecture at ETH Zurich, October 2013, on some recent results and remaining challenges in climate dynamics.

  • https://vimeo.com/136707147

    Where the Wind Comes From, on Earth and Other Planets

    Tapio Schneider’s Watson Lecture at Caltech (Nov. 11, 2009). It covers a history of ideas about how winds on Earth arise, leading to a modern perspective of what controls winds and their structure on Earth and other planets.

  • https://vimeo.com/115070686

    Distant Worlds – Strange Climates

    A workshop discussion of climates in our solar system, how we know about other planets and moons, and what we can learn from them about climate dynamics more broadly. (Produced by Oliver Stebler. Originally the interview was in German. This version is voiced over in English.)

Temperature changes can be decomposed into slow interdecadal components and faster intradecadal components. The slowest components of temperature changes have the largest ratio R of interdecadal variance to intradecadal variance; faster components of temperature changes have a smaller ratio R of interdecadal variance to intradecadal variance. By filtering out components with smaller variance ratios R, we can isolate the slow component of temperature changes. Any anthropogenic component of temperature changes is expected to be contained in this slow component.


  • https://vimeo.com/19457865

    Annual Mean Temperatures, 1850-2014

    The upper panel of the animation shows the slow component of temperature changes between 22.5°S and 67.5°N, an area with sufficient data coverage since 1890 to allow this multivariate analysis. The indicated temperature changes are changes relative to the period 1850-1900.

    The lower panel shows the time series of area-mean temperature changes (black) and the area-mean temperature change accounted for by the slow component in the main panel (red).

    Gray areas in the upper panel are areas with insufficient data coverage for this analysis.

    See also the following seasonal mean temperature animations:

    Several of the temperature changes are suggestive of human influences on climate. For example, the relatively uniform and steady warming of the ocean surfaces, the generally enhanced warming of continents, and the strong warming of high northern latitudes, particularly in the transition seasons, is consistent with expected effects of increases in greenhouse gas concentrations. The localized cooling between about 1950 and 1970 over industrial regions such as Europe and Southeast Asia, where anthropogenic sulfate aerosol loadings were high, is consistent with the expected cooling effect of sulfate aerosols. Also recognizable are numerous apparently natural climate variations, for example, strong temporary cooling in the North Atlantic from the 1950s through the 1970s, which contributed to the lack of global-mean temperature increase during that time.

    The animations are produced using the methods described in Schneider and Held (2001). As in the paper, the data are from the Climatic Research Unit at the University of East Anglia (dataset HadCRUT3v). The analysis in the paper has been extended to the annual mean and to all seasons, including data through 2014. The overlapping decadal data groups used to determine the slow temperature variance are defined similarly to the analysis in the paper but are centered on the years 1858, …, 2008, with 15 years between successive group centers. The temperature changes represented in the animations are those accounted for by all discriminants with a variance ratio R corresponding to a p-value less than 0.1 determined by a bootstrap procedure (i.e., with a less than 10% chance of occuring if there is no coherent decadal variability). There are 2 to 4 such discriminants in each animations (i.e., the temperature changes in the animations have 2 to 4 spatial degrees of freedom).

    Schneider, T., and I. M. Held, 2001: Discriminants of twentieth-century changes in Earth surface temperatures. Journal of Climate, 14, 249-254.
    [PDF] [Official version]

We use the large-eddy simulation code PyCLES to simulate the dynamics of clouds and boundary layers, to elucidate their response to climate changes, and to develop closure schemes for representing their smaller-scale dynamics in larger-scale climate and weather forecasting models.

  • https://player.vimeo.com/video/184179593

    Large-eddy simulation of precipitating cumulus clouds

    Simulation of cumulus clouds under conditions observed during the Rain in Cumulus over the Ocean (RICO) field campaign. Within the gray volume, there is cloud liquid water, simulated with the PyCLES code developed by Kyle Pressel et al. Blue colors indicate rain. The simulation uses 5th-order weighted essentially non-oscillatory (WENO) schemes for discretization of all fluxes (of momentum, entropy, and total water), at a resolution of 50 m in the horizontal and 40 m in the vertical.

  • https://player.vimeo.com/video/147025160

    Large-eddy simulation of stratocumulus clouds

    Simulation of stratocumulus clouds in the Dynamics and Chemistry of Marine Stratocumulus (DYCOMS) field study. Colors show the liquid water path (amount of liquid water in atmospheric columns) simulated with PyCLES. The simulation uses weighted essentially non-oscillatory (WENO) schemes for discretization of all fluxes (of momentum, entropy, and total water), which leads to improved fidelity of simulations already at relative low resolution.

The atmospheric circulation of a planet that has no zonal asymmetries in boundary conditions will remain axisymmetric if the initial condition is. So we can calculate the “ideal” Hadley circulation that would result in that case by choosing an axisymmetric initial condition in an idealized GCM without zonal inhomogeneities in boundary conditions. However, for an Earth-like planet, the resulting axisymmetric circulation is unstable with respect to non-axisymmetric perturbations.


  • https://vimeo.com/19513572

    Spinup of General Circulation from Axisymmetric State

    This animation (produced by Tim Merlis) shows the spinup of a macroturbulent circulation from an axisymmetric circulation. The upper panel shows zonal-mean zonal wind and the lower panel shows surface air temperature, with the zonal mean in the small white panel on the right.

    In the macroturbulent simulation, there is no subgrid-scale diffusion of heat or momentum above the planetary boundary layer. Vertical subgrid-scale diffusion of heat and momentum is necessary for stability in the axisymmetric simulations. Because the subgrid-scale diffusion is turned off when the three-dimensional perturbation is added to the axisymmetric circulation, the Hadley circulation first weakens before it strengthens when large-scale eddies form. (See Schneider (2006) for a description of the simulation.)

    Schneider, T., 2006: The general circulation of the atmosphere. Annual Reviews of Earth and Planetary Sciences, 34, 655-688.
    [PDF] [Official version]

(Junjun Liu and Tapio Schneider, 2010)

We have carried out the first 3D simulations of all four giant planets (Jupiter, Saturn, Uranus, Neptune) with closed energy and angular momentum balances that are consistent with observations (Liu and Schneider 2010). The simulations reproduce many large-scale features of the observed flows, such as equatorial superrotation on Jupiter and Saturn and equatorial subrotation on Uranus and Neptune.

The simulations resolve the flow in the upper atmospheres of the giant planets, with implicit links to the (convective) flow, mean meridional circulations and (likely magnetohydrodynamic) dissipation mechanisms at depth. Below are animations of the zonal wind and vorticity at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flows are inferred.

Coherent vortices form in these simulations and are recognizable in the vorticity fields. They are particularly large in the polar regions, where the largest vortices are cyclonic: see the polar projections of the vorticity in the Jupiter and Saturn simulations. Similar cyclonic vortices have recently been observed by the Juno mission.

Earlier we had carried out simulations of Jupiter with the same model (Schneider and Liu 2009), but with slightly different choices of dissipation parameters (which are poorly constrained by data). This leads to weaker and narrower jets than in the later simulations: [zonal wind] [vorticity]

Liu, J., and T. Schneider, 2010: Mechanisms of jet formation on the giant planets. Journal of the Atmospheric Sciences, 67, 3652-3672. [PDF] [Official version]

Schneider, T., and J. Liu, 2009: Formation of jets and equatorial superrotation on Jupiter. Journal of the Atmospheric Sciences, 66, 579-601. [PDF] [Official version]

  • Vorticity around north pole in Jupiter simulation

    The vorticity is shown in polar view at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flow is inferred (0.75 bar). Coherent cyclones are recognizable in the simulation, similar to those observed by the Juno mission in 2016. [See Liu and Schneider (2010) for a discussion of these vortices.]

  • https://vimeo.com/19397937

    Zonal wind in Jupiter simulation

    The zonal velocity is shown at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flow is inferred (0.75 bar). Clearly recognizable is the prograde (superrotating) equatorial jet and alternating off-equatorial jets. Their strengths and widths depend on observationally poorly constrained drag parameters (Liu and Schneider 2015). Also recognizable are Rossby wave packets especially on the equatorial jet, with retrograde propagation and phase lines that have the characteristic chevron shape indicating angular momentum transport into the jet, as is observed to occur.

  • https://vimeo.com/19457837

    Vorticity in Jupiter simulation

    The vorticity is shown at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flow is inferred (0.75 bar). As in the zonal wind animation, Rossby wave packets are clearly recognizable, as are a host of smaller coherent vortices.

  • https://vimeo.com/19397070

    Zonal wind in Saturn simulation

    The zonal velocity is shown at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flow is inferred (0.1 bar). As in the Jupiter simulation, a prograde (superrotating) equatorial jet and alternating off-equatorial jets are clearly recognizable, as are large Rossby wave packets. As observed, the jets are stronger and wider in the Saturn than in the Jupiter simulation, especially near the equator.

  • https://vimeo.com/19397135

    Vorticity in Saturn simulation

    As in the Jupiter simulation, the vorticity (at 0.1 bar) exhibits Rossby wave features on the jets.

  • https://vimeo.com/19397582

    North polar vorticity in Saturn simulation

    The vorticity is shown in polar view at the level in the model that corresponds roughly to the estimated level of the cloud tops from which the observed flow is inferred (0.1 bar). As in the Jupiter simulation, coherent vortices (the largest being cyclones) are visibly in the polar regions. The polar jet exhibits polygonal (wavenumber 8 and 9) undulations that are reminiscent of the wavenumber-6 “polar hexagon” observed on Saturn. Indeed, the simulations spontaneously produce similar hexagonal polar jets when the drag at the lower boundary is slightly lowered.

  • https://vimeo.com/19397260

    Zonal wind in Uranus simulation

  • https://vimeo.com/19397373

    Vorticity in Uranus simulation

  • https://vimeo.com/19397430

    Zonal Wind in Neptune simulation

  • https://vimeo.com/19397514

    Vorticity in Neptune simulation

(Timothy Merlis and Tapio Schneider, 2010)

To explore climate dynamics of tidally locked Earth-like exoplanets and to illustrate how the climate would adjust if Earth abruptly entered a tidally locked state, we conducted simulations with the GFDL CM2.1 coupled climate model (Delworth et al. 2006). Starting from a typical January 1 initial condition in the present-day climate, we set the planetary rotation rate to 1/365th of the present value (so that 1 day is equal to 1 current Earth year). We fixed the insolation so that there is a perpetual subsolar point on the equator at 88°W (near the coast of Ecuador). The simulation was run for 50 years. This simulation with a comprehensive coupled climate model illustrates and expands upon the dynamics discussed in the context of an aquaplanet atmosphere-only model in Merlis and Schneider (2010).

Delworth et al. 2006: GFDL’s CM2 global coupled climate models. Part I: Formulation and simulation characteristics. Journal of Climate, 19, 643-674. [Official version]

Merlis, T. M. and T. Schneider, 2010: Atmospheric dynamics of Earth-like tidally locked aquaplanets. Journal of Advances in Modeling Earth Systems, 2, Art. #13, 17 pp. [PDF] [Official version] [Correction]
Zonal wind animations: [Slow rotation] [Rapid rotation]

  • https://vimeo.com/19458103

    Surface Temperature

    The surface temperature changes rapidly over land masses to day-side values of about 290 K and to night-side values of about 240 K. The ocean surface temperature changes more slowly; the night-side ocean remains near the freezing point for the length of the simulation.


  • https://vimeo.com/19458068


    Precipitation patterns change rapidly (within a few months) from the zonally-elongated intertropical convergence zone that is typical of Earth today to a configuration in which there is substantial precipitation near the subsolar point and little precipitation on the night-side of the planet. The regions of large precipitation are determined by the atmospheric circulation: the near-surface atmospheric flow converges near the subsolar point, leading to strong ascending motion and condensation (see Merlis and Schneider (2010) for a detailed discussion). The topography (e.g., the Andes) modulates the precipitation patterns so that they are less concentric about the subsolar point than in the aquaplanet simulations in Merlis and Schneider (2010).


  • https://vimeo.com/19458029

    Evaporation Minus Precipitation

    The net evaporation field (evaporation minus precipitation) shows that atmospheric water vapor is transported from the night side to the day side. Regions on the day side of the planet away from the subsolar point, such as Canada, experience net drying. They become relatively warm because of the loss of evaporation as a cooling mechanism for the surface. Sufficiently far from the subsolar point, there is net evaporation, which eventually would lead to the formation of deserts and complicating habitability of those regions.


  • https://vimeo.com/19458085

    Sea Ice

    Because the ocean is actively convecting where it is near freezing, no sea ice forms during this simulation, except near the coasts. However, the ocean surface will eventually freeze (on the much longer timescales needed to cool the entire water column to freezing). It may seem surprising that no widespread sea ice forms on the night side of the planet within 50 years; after all, new sea ice forms every winter in Earth’s high latitudes. However, Earth’s polar regions currently experience net precipitation, and the fresh water effect on the ocean density allows the surface to freeze without the need for the entire column of ocean water to reach the freezing point. In contrast, the night side of the tidally locked Earth experiences net evaporation, so the ocean surface is becoming cooler and saltier, so convection penetrates deeper into the interior ocean as the simulation progresses and prevents the surface from freezing.