Developing Ice Nucleation Parameterizations for Large-Scale Models
Principal Investigator: Xiaohong Liu
Aerosol effect on clouds remains one of the largest uncertainties in understanding future climate change. Cirrus clouds composed of ice crystals have an annual global average frequency of occurrence of about 30 percent. Cirrus clouds scatter shortwave radiation and absorb and emit longwave terrestrial radiation, thereby modifying the global radiative balance. Precipitation formation in large part is initiated from ice phase processes. The limited state of knowledge of ice formation processes (including both primary and secondary ice particles) limits our ability to model cloud properties and precipitation production and to predict climate changes.
The goal of this project is to develop a new ice nucleation parameterization from the first principle using the classical nucleation theory (CNT). The CNT will be constrained with data obtained from PNNL's Atmospheric Measurement Laboratory, and also data available in the literature. The probability distribution function (PDF) for the ice nucleation capability of aerosols will be introduced in the CNT. The parameterization will be applied in both mixed-phase and ice clouds to represent mineral dust acting as ice nuclei through deposition nucleation mode. This new parameterization will be implemented in a cloud-resolving model to simulate the impacts of deposition ice nucleation on a cirrus cloud observed in the ARM SGP site for a cirrus case. We will report our progress in scientific meetings and publish our results in peer-reviewed journals.
Constrain classical nucleation theory with laboratory data
The newly developed ice chamber at AML was connected to particle generation set up to measure the ice nuclei (IN) active fraction as a function temperature (T) and relative humidity with respect to ice (RHi). Different size, 100 nm, 300 nm and 500 nm diameter, particles were selected and forwarded to the ice chamber. Active fraction was calculated as the ratio of IN number measured to total particles measured.
To incorporate the ice nucleating properties, determined in terms of ice nucleation fraction (Fice) as a function of T and RHi, into the CNT, we modified the original CNT to include the probability density function (PDF) of contact angles (Θ) distributed among the IN. The PDF parameters mean (µ) and standard deviation (σ) are iterated to fit computed Fice to measured Fice values.
Application to cloud-resolving models
We employed a cloud-resolving model, the System for Atmospheric Modeling (SAM), coupled with a spectral-bin microphysical scheme (SBM) to examine the sensitivity of cloud simulations to the PDF parameters of contact angle and IN concentration. CNT with the PDF of contact angle for deposition nucleation is implemented in the SAM-SBM. A cirrus cloud case observed from the ARM SGP site in March 09, 2000 was chosen to test deposition freezing since the cloud temperatures are lower than -22 °C. We conducted a group of sensitivity tests by varying the PDF parameters (mean µ, standard deviation s) and dust IN concentration. Cloud depth, ice number (Ni) and ice water content (IWC) are more sensitive to µ than s. As µ is decreased from 0.58 to 0.41, Ni increases by more than 2-3 times and about 25 percent for IWC, which are comparable to the 6-10 time increase of IN number. Our results suggest that accurately quantifying the PDF parameters of the contact angle is extremely important in determining ice nucleation and cloud properties.