Data-centric science can be treated as purely informatics problems, but our focus is on treating it as multidisciplinary problems to solve problems in the real world. In this approach, a general rule is that domain experts identify and specify important problems and informatics experts solve and propose better processes for data collection, analysis, workflow, and presentation using our knowledge on algorithms, abstractions, and digital tools. This presentation introduces our past experiences on the application of data-centric science to earth environment-related domains, namely agriculture, biodiversity and disasters, and how we worked with domain experts. These three domains are among nine domains identified as “societal benefit areas” (agriculture, biodiversity, climate, disasters, ecosystems, energy, health, water and weather) by GEOSS (Global Earth Observation System of Systems), in which earth observation data and information can support better decision making in the society. We introduce case studies on agricultural field monitoring and optimized harvesting, biodiversity monitoring on invasive species, and large-scale heterogeneous disaster databases on typhoons and earthquakes. We also discuss data-related issues such as data mining, quality control, and participatory science.