Implicit within the metadata strategy of most archiving or preservation institutions is the use of format registries to contain important information, usually technical in nature, that is common among like formats or data types.At present, there are several approaches to data models, policies, and implementation models for format registries that are in various stages of implementation and/or conception including The National Archives' (UK) PRONOM Technical Registry, the Global Digital Format Registry (GDFR), funded by the Mellon Foundation and led by Harvard University Library, and the Library of Congress (USA).The National Geospatial Digital Archive (NGDA) project funded by the Library of Congress' National Digital Information Infrastructure and Preservation Program (NDIIPP) has been investigating whether existing or planned format registry efforts do or would support the often quite complex geospatial data formats which NGDA and other institutions are collecting for long term preservation.This paper discusses results of a comparative study of the data models of pertinent format registries in which instances of over 20 proprietary and open geospatial formats were examined to assess whether the elements within the data models could adequately describe a given format and its relatives, and if not, what other kinds of information would be important to include.The paper places the findings and recommendations into the context of previous work done by the NGDA team and others about what preservation metadata would be appropriate for geospatial resources. In addition, the paper discusses differences in definitions for key format registry concepts describing relationships among formats.Finally, we identify related research questions yet to be answered regarding the usability of existing and potential format registry efforts for geospatial resources, and the broader questions about the practicability of gathering this and other pertinent preservation metadata for geospatial resources.