Physics of Climate

Romps Group at the University of California, Berkeley

The Fixed Anvil Temperature (FAT) hypothesis is widely accepted, but a paper in GRL shows that FAT is supported neither theoretically nor empirically. Instead, a Fixed Tropopause Temperature (FiTT) is found: for a 50-K change in sea-surface temperature, the temperature of the tropopause changes by less than 2 K.
In the tropics, the highest frequency of cloudiness is in the upper troposphere. But, why? The conventional wisdom says that the peak in cloudiness occurs where convective detrainment is highest. A paper in GRL argues this is wrong. Instead, cloudiness is greatest in the cold upper troposphere because the sublimation of ice is so slow there.
CAPE x P has been used to predict changes in lightning with global warming, but how good is this proxy? As it turns out, quite good! As shown in a paper in GRL, the CAPE x P proxy correctly predicts the seasonal maps of lightning and its spatially varying diurnal cycle. Shown here is the hour of maximum (top) CAPE x P and (bottom) lightning for JJA.
Why does precipitation increase at 2% / K? A paper in PNAS explains why. The net upwelling radiative flux F is a fixed function of temperature, F = Ft - k (T - Tt)n, where k and n are positive constants and subscript t denotes the tropopause. Since n = 2 and the depth of the troposphere is 100 K, precipitation increases at 2% / K.
Shallow cumulus clouds provide a large and uncertain feedback on global warming, but the 4D observations needed to characterize their lifecycles are nearly impossible to obtain with current instruments. A paper in BAMS describes a new ring of stereo cameras that is now providing the needed 4D data on a 50-meter grid with a 20-second time step.
Convective aggregation is a bizarre phenomenon where, in numerical simulations, rain clouds clump together into a single blob. How relevant is this numerical oddity to the real world? In a paper in PNAS, the feedbacks responsible for convective aggregation are shown to dramatically accelerate the formation of hurricanes, pointing to a real-world impact.
For nearly 200 years, atmospheric scientists have relied on approximations to the height of the lifting condensation level (LCL). In a paper in JAS, the exact, explicit, and analytic expression for the LCL is derived, as is an analagous expression for the lifting deposition level (LDL). These expressions depend only on basic physical parameters.
What happens to convective heating after it is deposited in the troposphere? This question is fundamental to atmospheric dynamics, but the only known solutions have assumed an artifical rigid lid at the tropopause. A paper in JAS presents the first Green's function for buoyancy in a troposphere overlain by a realistic stratosphere.
A cloud's size plays an important role in its dynamics, its ability to form rain, and its radiative forcing. But, the distribution of cloud sizes can be difficult to measure. A paper in JAS explores ways to calculate the size distribution from linear sampling, e.g., from vertically pointing lidar or from aircraft.
Severe weather is predicted to become more frequent with global warming, but those predictions are based on warming-induced increases in CAPE, for which we have had no theory. A paper in JAS presents the first analytical expression for CAPE. It explains the Clausius-Clapeyron (CC) scaling of CAPE, and its divergence from CC scaling at very warm temperatures.
On average, the largest convective buoyancies and vertical velocities occur in the upper troposphere. Common wisdom states that this results from the extra buoyancy provided by the latent heat of fusion. A paper in GRL shows that this is a fallacy: the profiles of convective buoyancy and vertical velocity are the same in a world with and without ice.
In global climate models, the poor treatment of "entrainment" (the mixing of clear air into clouds) is the largest source of uncertainty in how much the Earth will warm for a given increase in atmospheric CO2. A paper in JAMES describes the Stochastic Parcel Model, a convective parameterization that treats entrainment the way it operates in nature and that is also deterministic and fast enough for use in global climate models.
Many aspects of Earth's climate (e.g., cloud cover, precipitation, and radiation fluxes) depend on the sizes and lifetimes of cold pools, which can trigger storm clouds when they collide. A paper in QJRMS presents analytical expressions for the evolution of cold pools, including their maximum sizes and lifetimes. These results are successfully validated with large-eddy simulations.
The energy available to storms, which is measured by CAPE, is predicted to increase dramatically with warming, leading to more severe storms and more frequent lightning. But, why does this increase in CAPE occur? A paper in GRL shows that the increases in CAPE are driven by a ballooning of upper-tropospheric buoyancy, which is correctly predicted by the zero-buoyancy plume model.
Stationary buoyant parcels of fluid do not accelerate at a rate equal to their buoyancy. Instead, they accelerate at a smaller rate, which is found by solving a Poisson equation. In a paper in QJRMS, it is shown that proximity of buoyant parcels to the surface greatly impacts their acceleration. This finding has implications for the triggering of new thermals, which originate at the surface.
Rings of high-humidity air at the edges of cold pools serve as important precursors for deep convection. But, how do those rings form? The conventional wisdom is that evaporation of precipitation is responsible. A paper in GRL overturns this conventional wisdom and identifies enhanced surface fluxes as the true source of the rings of high humidity.
Stereo photogrammetry allows us to make quantitative three-dimensional measurements of cloud sizes and speeds. In a paper in GRL, data on dozens of cloud thermals are used to show that drag -- recently argued to be negligible -- is, to the contrary, a dominant term in their momentum balance, and wave drag is the likely culprit.
What physical process gives birth to convective clouds? There are two possibilities: warm and humid air launches off the surface due to its buoyancy, or that air gets pushed up by colder air that collides with it. In a paper in the Journal of Atmospheric Science, we cleanly decompose the forces into those due to buoyancy and inertia to find the answer.
Vertical velocities in clouds are important for many phenomena, including hail, lightning, and stratospheric moistening. But, there is no consensus on the dominant balance of forces giving rise to these velocities. In a paper in the Journal of Atmospheric Science, we identify, track, and analyze thousands of individual cloud thermals to identify whether clouds are slippery (acceleration equals buoyancy) or sticky (buoyancy equals drag).
GCMs disagree on the answer to the following question: will global warming cause an increase or decrease in summertime severe weather over the United States? In a paper in the Journal of Climate, we test a hypothesis that the models that do a good job of simulating today's severe weather will be in better agreement about the future of summertime severe weather.
For hundreds of millions of years, lightning has shaped the evolution of terrestrial species through its triggering of wildfires, and lightning has altered atmospheric chemistry through its production of NOx. Now, with rapid global warming projected for the 21st century, the frequency of lightning is poised to increase dramatically. In a report featured on the cover of Science, a new proxy for lightning is validated against observations and then used in global climate models to predict a 50% increase in U.S. lightning strikes over the 21st century.
The relative humidity of the atmosphere depends on the precipitation efficiency (PE) of convective clouds. The paper "Lagrangian investigation of the PE of convective clouds" measures precipitation efficiency in unprecedented detail using water-following Lagrangian particles. The results show how the PE of water vapor varies depending on where it enters the cloud.
Water vapor is the dominant greenhouse gas in the atmosphere, but what processes set its concentration? The paper "An analytical model for tropical relative humidity" provides a simple set of equations for relative humidity and its changes with global warming. In particular, it predicts that relative humidity will remain a fixed function of temperature as the atmosphere warms.
Using methods from the field of computer vision, the paper "Stereo photogrammetry of oceanic clouds" details how to reconstruct oceanic clouds in 3D using two digital cameras.
The paper "An improved weak-pressure-gradient scheme for single-column modeling" eliminates a spurious gravity-wave resonance in a method used to test convective parameterizations for GCMs. The resonance appears as the dashed-blue spike in the figure, while the solid-red curve shows the new scheme, which matches the solid-black benchmark.
The paper "Rayleigh damping in the free troposphere" shows that wind profiles with large wavelengths damp slower and descend slower than those with short wavelengths.
The paper "Convective self-aggregation, cold pools, and domain size" finds that cold pools are responsible for the abrupt transition to convective aggregation as domain size increases. In their absence, aggregation occurs at all scales.
The paper "Measurement of convective entrainment using Lagrangian particles" finds that clouds rapidly entrain both environmental air and their own detritus. Click on the image to watch Lagrangian particles in a cumulus congestus.
Particles in large-eddy simulations (animated here with and without precipitation) confirm large entrainment rates of ~100% per kilometer. Animated GIFs here and here.
The paper "On the equivalence of two schemes for convective momentum transport" finds that a very simple scheme, which neglects the horizontal pressure force, performs better than the schemes currently used in GCMs.
Cloud-resolving simulations, presented in "Numerical tests of the weak pressure gradient approximation", validate the underlying theory and predictions of WPG, including stacked layers of negative buoyancy in regions of steady-state ascent.
The paper entitled "Weak pressure gradient approximation and its analytical solutions", in press at JAS, uses pencil-and-paper derivations to argue that WPG is more accurate than WTG.
A new study, "CAPE in tropical cyclones" provides an explanation for the asymmetrical distribution of lightning in hurricanes. Using data from thousands of dropsondes, condensate loading and entrainment are found to be key mechanisms in convective ascent.
In the paper published in JAS, "A transilient matrix for moist convection", it is found that boundary-layer parcels detrained by clouds in the free troposphere come primarily from within 100 meters of the surface. This has important implications for how convective instability is calculated in the atmosphere.
The first study of precipitation extremes in a high-resolution cloud-resolving model, "Response of tropical precipitation to global warming," is now in JAS. Contrary to results from GCMs, convective updrafts become more vigorous when CO2 is doubled, contributing to higher rain rates.
The first investigation of stratospheric HDO using a steady-state cloud-resolving simulation with isotope microphysics, "Isotopic composition of water in the tropical tropopause layer," is published in JGR. Simulating a Walker cell over an 8000-km-wide domain, the convective injection of ice and the generation of cirrus by gravity waves are found to be dominant controls on HDO in the stratosphere.
The first direct measurement of convective entrainment. From the paper published in JAS, "A direct measure of entrainment," this figure shows the azimuthally averaged entrainment rate at three intervals in the life cycle of a deep cumulonimbus.
An animation of 2005 brightness temperatures and best-track data from the GRL paper "Overshooting convection in tropical cyclones."
An animation from a paper in the May 2010 issue of JAS, "Nature versus nurture in shallow convection."
Animations with and without precipitation of deep convection from work in progress.