The Sixth Generation (6G) network will rely on the integration and cooperation of a plethora of communication and networking systems. Space-air-ground integrated networks (SA-GIN) which integrate satellites, aerial networks, and terrestrial communication, can provide seamless and continuous delivery of services and applications with high levels of performance. While SAGIN provides promising solutions for data collection and communication, it poses security concerns given that it may allow untrusted devices to be part of the cooperative system. In this paper we introduce a blockchain-enabled SAGIN framework that authenticates both devices and the data collected by SAGIN participants. Federated Learning (FL) is supported by the framework to enable network distribution and accurate real-time data analysis. As a proof-of-concept, the solution considers a use-case for earthquake event prediction and monitoring. Satellite sensors are used to collect data and identify the presence of thermal anomalies used for earthquake forecasting. Furthermore, a post-earthquake monitoring technique is used which relies on Unmanned Aerial Vehicles (UAVs) to provide access to remote places in order to collect high quality images. The collected images are then processed locally on UAV swarms or through edge devices for real-time decision making. An experimental study is conducted on models and approaches needed to increase prediction accuracy for earthquake events.