From process study to parameterization (Ice Cloud effective radius)
Ice clouds play an important role in the energy balance of the atmosphere. They can either cause cooling or warming depending on their altitude, ice water content (IWC) and microphysical properties like the particle effective radius ( ). The effective radius (here defined in terms of the mass and cross-sectional area of the particles) is intended to describe the effective size at which radiation interacts with individual particles.
Figure 1. Example of lidar-radar results for a Cirrus cloud over the ARM SGP site
Describing properly is important, as it, combined with the IWC, determines the optical thickness and emissivity of ice clouds. In atmospheric models, Ice cloud effective radius commonly parameterized in terms of temperature and/or IWC.
The combined Lidar and Radar method described by Donovan et al. (2001) was applied to data from Cabauw and compared to an earlier analysis performed using data from the ARM programs SGP site (Donovan and van Lammeren 2003). The method is based on estimating the so-called lidar radar effective radius (R’eff) which is directly related to the ratio between radar reflectivity and optical extinction. R’eff may then be used to estimate if certain assumptions related to the average crystal morphology are made. Example results for a 3 month time span are shown in Figure 2. As shown in the figure it was found that the distribution of lidar-radar effective radius with temperature was similar at Cabauw and at the ARM site. However, the mean dependence was found to be sufficiently different to warrant further investigation.
Further investigation revealed that representing cloud effective radius as a function of normalized depth from cloud top for different cloud thickness regimes (see Figure 3) yielded a parameterization that seemed equally valid for both the Cabauw data and the ARM data set. (van Zadelhoff, 2004) The radiative effect of implementing this parameterization was then investigated by using radar derived IWC profile time-series data together with surface-based SW transmissivity measurements. It was found that the new parameterization yields the least bias compared to the default temperature dependent parameterization as well as a parameterization based on both IWC and T.
|
Figure 2. Cumulative probability of occurrence of lidar-radar effective radius as a function of temperature for optically thin clouds at Cabauw. The greyscales, from dark to light, show the 10, 30, 60, 90 and 99 % probability of occurrence. The error-bars indicating the 1-sigma level of the distribution. The dashed line shows the mean particle size for the ARM data
|
|
Figure 3. Mean effective radius versus normalized cloud depth (H [km]) regimes (H>4.5, 3.0<H<4.5, 1.5<H<3.0, H<1.5 from light-grey to black points) observed at the ARM site. Note that Cloudtop is at 0 and Cloud bottom at -1. The error bars show the error in the mean and the solid lines a second order polynomial fit to the data
|
|
|
|
|
Having assessed the direct influence of the new effective radius, parameterization on the SW radiative transfer using observations of shortwave radiation fluxes at the surface, the impact of the different parameterizations on the performance of the KNMI regional climate model (RACMO2) was investigated (van Zadelhoff, 2007). The parameterization was implemented within RACMO2 and forecast runs for an entire year (1995) in the domain in between 62W, 62E, 27N and 75N were performed. Differences in planetary albedo for , and , compared to the , are found to be in the order of 2.4 % and 1.3 % respectively. In should be noted that effects found in this study are relatively small compared to what the effects would look like for a climate run. In a climate run, differences in Reff, on the radiation flux profiles would impose a long-term feedback on the model dynamics. For the forecast runs presented here this effect is absent as the model is reset once a day.
In conclusion, the new parameterization has a definite effect on the energy balance relative to the temperature based formulation, which is currently used in RACMO2. The parameterization used in the latest release of the ECMWF-IFS (cy30r1) shows a smaller but still significant difference compared to the parameterization. To quantify this effect on the shortwave flux in such a way that one can decide which parameterization is more suitable for use in a regional climate model is problematic due to the lack of knowledge about the ‘truth’ and all the interactions and feedback mechanisms present in a (regional) climate model. A more general issue would be that the present day climate models are balanced in such a way that the best results or skill score is achieved with the standard setup. Changing one parameterization requires a re-balancing of other parameters like IWC before a complete assessment could be carried out.
References
Donovan, D.P., Ice-cloud effective particle size parameterization based on combined lidar, radar reflectivity, and mean Doppler velocity measurements, J. Geophys. Res., 108(D18)}, 4573, 10.1029/2003JD003469, 2003.
van Zadelhoff, G.-J., D.P. Donovan, H. Klein Baltink, and R.Boers, Comparing ice cloud microphysical properties using CloudNET and Atmospheric Radiation Measurements Program data., J. Geophys. Res. 109, 24,214--24,229, 10.1029/2004JD004967, 2004.
van Zadelhoff, G-J, E. van Meijgaard, R. Boers, D. P. Donovan, W. Knap , Sensitivity of the shortwave radiative budget to the parameterization of ice crystal effective radius, 2007, J. Geophys. Res., 112, D08213, doi:10.1029/2006JD007791
)
)
)
