Diana C. Roman
H.O. Wood Chair of Seismology

Diana C. Roman

Research Interests

Source processes of volcanic earthquakes; volcano-fault interaction; structure and dynamics of magma transport and storage systems; volcano monitoring and eruption forecasting


B.S., Applied Economics, Cornell University, 1997 M.S., Geological Sciences, University of Oregon, 2001 Ph.D., Geological Sciences, University of Oregon, 2004

Contact & Links

  • (202)478-8834 | fax: (202) 478-8821
  • droman at carnegiescience.edu
  • Earth and Planets Laboratory
    Carnegie Institution for Science
    5241 Broad Branch Road, NW
    Washington, DC 20015-1305
  • Personal Website


Diana Roman's research straddles the boundary between volcanology and seismology, with a dual focus on understanding the nature of magma ascent and eruption and of volcanic microearthquake swarms. Specifically, Roman works to understand, from a mechanical perspective, the formation, evolution, and dynamics of crustal magmatic systems and the source mechanisms and causes of microearthquake swarms occurring in the vicinity of active volcanoes. These two lines of research are tied together through development of conceptual and numerical models of the interaction of tectonic and volcanic processes.

Roman's primary research efforts have been aimed at identifying, documenting, and understanding systematic patterns in precursory volcanotectonic (VT) earthquake swarms and crustal stress field orientations during episodes of magma ascent, and at exploring the potential of volcanic stress field analysis as an eruption forecasting technique. A major result has been the recognition, though detailed analysis of precursory VT swarms and split S-wave polarizations, that changes in the orientation of local maximum compression precede many eruptions by weeks to months (e.g., Roman et al. 2004, BSSA; Roman et al. 2006, EPSL; Roman et al. 2008, JVGR; Lehto et al., 2010, JVGR; Roman and Gardine 2013, EPSL) and that these changes are caused by the dilation of the volcanic conduit system during magma influx (Roman 2005, GRL; Roman et al., 2011, JGR). Roman has also demonstrated that identification of local stress field reorientations may be used to assess the intermediate-term likelihood of impending eruptive activity (Roman et al. 2006, EPSL). Additional study has shown that one of two general patterns of VT seismicity may be observed prior to an eruption, depending on the mechanism of earthquake generation (earthquakes generated at the tip of a propagating dike vs. earthquakes generated on faults in the rock surrounding an inflating dike) (Roman and Cashman 2006, Geology). Detailed numerical modeling has shown that the observed pattern of precursory seismicity is determined or strongly influenced by the relative strength of regional (tectonic) stresses (Roman and Heron 2007; GRL).

A second major avenue of research has been the analysis of non-eruptive volcano-seismic unrest, which occur beneath potentially active volcanoes but do not correspond to episodes of eruption. This line of research considers both discrete episodes of unrest ('swarms' - Roman et al. 2004, BSSA; Roman et al. 2004, JVGR; Roman et al. 2011, Bull. Volcanol.; Kilgore et al. 2011, GJI; O'Brien et al. 2012; JVGR), and persistent restlessness (Rodgers et al. 2013, JVGR; Geirsson et al. 2014; JVGR; Rodgers et al., 2015; JVGR). Through detailed observational study, it has been shown that, while some non-eruptive swarms are caused by stalled intrusions of magma in the seismogenic crust, others appear to be driven by tectonic processes in combination with increased fluid and/or gas circulation above deeper magma bodies. More recently, detailed study of persistent seismic restlessness at the open-vent Telica Volcano, Nicaragua has led to development of a model for decreases in seismicity rates preceding explosions involving shallow sealing of gas pathways.