Abstract Prior to 1970 there were fewer than 10 seismographs sited in the southeastern United States—a study region defined herein as all of the states of West Virginia, Virginia, North Carolina, South Carolina, Georgia, and Alabama and the eastern portions of Kentucky and Tennessee (see Sibol and others, 1986, for a detailed description). Multistation seismic networks began to be installed in 1974, and by the end of 1977 there were 53 operational seismograph stations. As of January 1986, the Southeastern U.S. Seismic Network (SEUSSN) Seismicity Bulletin (Sibol and others, 1986) listed 147 stations in the region (Fig. 1). Thus, several years of instrumentally monitored seismicity data for the region have been acquired to supplement the historical data base that extends back to as early as 1698. The purpose of this chapter is to review the seismicity of the southeastern United States as documented by the studies to date of the historical and network data bases. Data acquisition and analyses are derived in large measure from the collective and individual efforts of the members of the SEUSSN coalition—a group of some 10 member institutions that operate seismographs in the region (Sibol and others, 1986). We divide consideration of the data base into a Historical Seismicity period (pre-network: 1698 to June 30, 1977), and a recent period of Instrumental Seismicity (post-network: July 1, 1977 to January 1, 1986). The reasons for that division are the differences in the levels of completeness and accuracy with respect to earthquake size and location. Prior to the

Abstract Problems with and approaches to seismic-hazard estimation in the midcontinent of the United States are evaluated by using recent data on stress regime, crustal age and structure, and seismicity of other stable continental regions. Evaluating earthquake hazard in the central U.S. is difficult because of the lack of identifiable seismogenic faults and because of the low rate of seismic activity. Furthermore, the recurrence intervals of large earthquakes are poorly known, in part because of the short historical record that spans only a fraction of the repeat times of these quakes. The seismotectonic regime of the central U.S. is dominated by the Reelfoot rift complex and the associated New Madrid, Missouri, seismic zone. However, there are other major tectonic structures in the region such as the Nemaha ridge, the Midcontinent rift system, and the Wichita-Ouachita orogenic belt; earthquakes generating damaging ground motion (approximately magnitude 5.0 or greater) have occurred in the states of Ohio, Illinois, Oklahoma, Texas, Kansas, Nebraska, Kentucky, Alabama, and Arkansas, as well as Missouri. Opinions vary widely about the best way to delineate seismic source zones in such a diffuse and varied seismotectonic environment. Moreover, detailed paleoseismic or neo-tectonic data that could improve hazard assessments are extremely sparse in the central United States. The Meers fault scarp in southwestern Oklahoma, with its evidence for Holocene displacement and its lack of background seismicity, highlights a new set of assessment problems. Development of site-specific probabilistic hazard curves are further hampered by the lack of strong ground-motion data and high-resolution attenuation data. We address aspects of the overall seismic-hazard assessment problem for which neotectonic information provides constraints. These include a seismic source zonation for the central U.S. and estimates of maximum possible earthquakes for these zones, especially for the New Madrid region.