Atmospheric Sciences & Global Change Division
Continuous Dataset of Water Vapor Measurements Throws Water on Assumptions of Cirrus Cloud Formation
To study the link between water vapor, cirrus cloud formation (homogenous and heterogenous) mechanisms, and their potential climatic impacts, researchers at PNNL analyzed a one-year dataset of water vapor measurements obtained by a Raman lidar for the Department of Energy's Atmospheric Radiation Measurement (ARM) Program. Signals from the ARM Raman lidar can be used to distinguish ice and water in optically thin clouds (such as cirrus clouds). In addition, this lidar provides continuous, high resolution water vapor profiles with an accuracy of 5%, while simultaneously detecting the presence of clouds. As reported in Geophysical Research Letters in June 2004, the long-term continuous dataset provided by the Raman lidar allowed the researchers to provide the first analysis of reliable upper tropospheric water vapor profiles measured from a single location. Their findings confirmed one aspect of cirrus formation, while raising questions about another.
As the single largest percentage of greenhouse gases in the Earth's atmosphere, water vapor (or "humidity") is a critical component of climate change research. Because of its ability to absorb energy, water vapor plays an essential role in radiative feedback mechanisms and cloud formation. This is especially true in the upper troposphere—the highest point of weather conditions in Earth's atmosphere, at about 10 km high. At this altitude, small changes in moisture can significantly impact the amount of outgoing radiation, as well as influence the formation of ice-crystals in cirrus clouds. However, accurately measuring water vapor amounts in the upper troposphere (where cirrus clouds occur) is very difficult, particularly on a continuous basis. This is due to the affect of extremely cold temperatures on traditional measurement instrumentation and the irregular schedule of their deployment, as well as the lack of resolution in satellite measurements for studying cirrus nucleation processes.
Using the ARM Raman lidar data, PNNL researchers were able to analyze a significant climate sample (about 300,000 data points, or 9,500 profiles) of continuous upper tropospheric water vapor concentrations from the ARM Program's Southern Great Plains research site in Oklahoma. As quantified by aircraft in situ measurements, typical ice-generating cirrus clouds are structured with an ice-generating region near the top, an ice crystal growth or deposition region in the middle, and a sedimentation or sublimation region near the cloud base. Using this structure in their analysis, the researchers examined the relative humidity with respect to ice (RHI) in the three regions. They found that ice supersaturation (RHI > 100%) occured most frequently in the ice crystal formation region where cloud updraft velocities are typically the strongest, but also occurred frequently in the growth region.
Specifically, their study showed that ice supersaturation occurred about 31% of the time in cirrus clouds, confirming existing assumptions regarding the frequency of homogenous (non-aerosol related) cirrus formation. However, they also found that ice supersaturation often occurred at temperatures warmer than -40C, when heterogeneous (aerosol-related) cirrus formation typically occurs. This type of ice formation results in smaller ice particles, thereby increasing the resulting reflectivity of the cloud. This implies that heterogenous formation may play a larger role in the impact of cirrus clouds on the Earth's radiative energy budget than previously thought. These findings, and the dataset used to reach them, represent an important link between the measurement and modeling communities as they continue to improve scientific understanding of the effect of cirrus clouds on the Earth's global climate.
Reference: Comstock JM, TP Ackerman, and DD Turner. 2004. Evidence of High Ice Supersaturation in Cirrus Clouds using ARM Raman Lidar Measurements. Geophysics Research Letters 31:L11106. DOI:10.1029/2004GL019705.