Accurate and long term information on the physical properties of clouds is required to increase our understanding on the role of clouds in the current climate system, and to better predict the behavior of clouds in a changing climate. This thesis investigates if retrievals of cloud physical properties from satellite imagers can be used to prepare time series of these properties for monitoring climate change, and to evaluate parameterizations of cloud processes in weather and climate prediction models.
An algorithm for retrieval of Cloud Physical Properties (CPP) from visible and near-infrared
reflectances of the AVHRR instrument onboard NOAA and the SEVIRI instrument onboard METEOSAT is presented. This algorithm retrieves cloud optical thickness, effective radius, and liquid water path, whereas a cloud model is used to simulate cloud geometrical thickness and droplet number concentration. Due to the large differences found between the reflectances from the different instruments (up to 25%), a recalibration procedure is developed that successfully reduces the retrieval differences to less than 5%. The uniqueness of the SEVIRI cloud property retrievals is in its unprecedented sampling frequency (15 minutes) that ensures the statistical significance of the dataset.
One year of cloud liquid water path retrievals is validated against simultaneous Cloudnet microwave
radiometer observations over Europe. The results show that during summer the agreement is very good while during winter an overestimation of about 20% is observed. Possible reason for this overestimation is the plane-parallel assumption in the CPP algorithm used to simulate real clouds. For single-layer stratocumulus days, a sub-adiabatic cloud model is used to obtain cloud geometrical thickness and cloud droplet number concentration. During these days good agreement is found between geometrical thickness simulations and Cloudnet lidar and radar observations, and cloud liquid water path retrievals and Cloudnet microwave radiometer observations. The simulated droplet concentration is found to vary independently from liquid water path and the geometrical thickness, which suggests
possible interactions between aerosols and clouds. This shows potential in our dataset for studies of the indirect aerosol effect.
The SEVIRI dataset of cloud property retrievals is used to evaluate the Regional Climate Model (RACMO) over Europe during a six-month period. The results show that RACMO represents the spatial variations of cloud amount and cloud liquid water path realistically, but underpredicts cloud amount by 20% and overpredicts liquid water path by 30%. Examination of the diurnal cycle shows that the RACMO maximum liquid water path occurs two hours earlier than that observed by SEVIRI, while the RACMO maximum cloud amount agrees reasonably well with SEVIRI’s amount. The largest differences in the diurnal cycle between RACMO and SEVIRI are found in regions of alternating stratiform and convective regimes where RACMO has difficulty representing the transition between these regimes. The SEVIRI dataset of cloud physical properties proves to be a powerful tool for evaluating parameterizations of cloud and precipitation processes in weather and climate prediction models, and thus helps increase the confidence in these models.
RA Roebeling. Cloud physical properties retrieval for climate studies using SEVIRI and AVHRR data
published, Wageningen Universiteit, 2008