The greatest uncertainty in future projections of climate arises from cloud and their interactions with radiation. Traditional approaches for evaluating and improving cloud-radiative feedbacks in climate models have yielded little advances over the past decade. A new method based on statistical information obtained from large ensemble of cloud objects observed by NASA Earth Observing System satellite combined with matched 4-dimensional atmospheric state data have recently been developed to overcome many of the problems associated with the traditional approaches (Xu et al., 2004; submitted to J. Climate). Statistical comparisons between cloud model results with CERES satellite observed cloud objects from this new approach have yielded significant insight into the deficiencies of current cloud models. In addition, statistical comparisons based on large ensemble of CERES observed tropical deep convective cloud objects during the peak of 1998 El Nino in March 1998 and those from the climatological near normal conditions in March 2000 have yielded important scientific information relating to the observed sensitivity of these cloud systems as climate state changed from El Nino to near normal conditions.

As part of our on-going effort to distribute these Earth Science Information to the broader community, a web-based cloud object data retrieval system has been developed to allow users to pick and choose the cloud object data and download them directly from the Internet. The front-end of this data system is the NASA Cloud Object Data Website, which is the interactive human interface to this system and consists of a set of user-definable parameters. This is linked to the back-end of the data system through a set of JavaScript procedures. The back-end of this system consists of database system powered by MySQL, which retrieves the user-defined cloud object data. These cloud objects are then download to the user system through the Internet using FTP. Users are encouraged to use these data to verify and improve their current cloud models and to study the observed sensitivity of cloud objects to different climate conditions.