Cloud fraction: can it be defined and measured?

 

Author

Stephen E. Schwartz — Brookhaven National Laboratory

Category

Cloud Properties

Description

<em>a)</em> Hourly mean CF at SGP determined by several techniques over a two-day period in May 2009. ARSCL (Clothiaux, <em>Journal of Applied Meteorology and Climatology</em> 2000) is time-average based on vertically pointing lidars and millimeter cloud radars; SIRS (Long, <em>Journal of Geophysical Research</em> 2006) is time-average based on downwelling SW irradiance within nominal 160° field of view. TSI is based on fraction of cloudy pixels within hemispheric field of view. GOES is based on average of all pixels (4-km size; satellite) within 20 km of the surface measurement site (Genkova, <em>Proceedings of the 14th ARM Science Team Meeting</em> 2004). Gray denotes nighttime; TSI and SIRS not available. Modified from Stevens and Schwartz (<em>Surveys in Geophysics</em> 2012). <em>b)</em> Scatterplot matrix of the several data sets for May 2009. Data provided by W Wu (Brookhaven National Laboratory, 2011).
Clouds greatly affect radiation transfer in the atmosphere and consequently climate. Globally, clouds enhance reflected shortwave (SW) flux by 47.5 ± 3 W m-2 and reduce outgoing longwave (LW) flux by 26.4 ± 4 W m-2 for a net cooling influence of 21.1 ± 5 W m-2 (Harrison, Journal of Geophysical Research 1990). The amount and properties of clouds are expected to change with increasing global temperature, but the amount and even the sign of resultant flux changes are not known, giving rise to much uncertainty in estimates of climate sensitivity and projections of climate change. Consequently it is essential that representation of clouds and their radiative influences in climate models be accurately assessed.

The conventional measure of the amount of clouds is "cloud fraction" (CF) the fraction of the atmosphere volume or column occupied by clouds. This raises the question of whether CF can be defined and how well it can be measured. If average CF is 0.5, then in round numbers, 1% error in CF corresponds to 1 W m-2 in SW and 0.5 W m-2 in LW globally. This sets the scene for how well CF must be known and provides context for differences in measurements by different techniques (Figure 1). Observationally, cloud fraction depends on resolution and threshold and on whether CF is evaluated as area average or as temporal average in a narrow-field-of-view vertical column.

This poster examines issues associated with definition and measurement of CF, compares different measures of CF at the ARM Southern Great Plains (SGP) site, and examines some of the literature. Given the importance of CF in models and the need for measurements, it might be useful to establish a Cloud Fraction Focus Group in ASR. The author invites discussion of this and suggestions for measurements or alternatives to CF, at the poster and/or by email.