A domestic earthquake impact alert protocol based on the combined USGS PAGER and FEMA Hazus loss estimation systems

Abstract

The United States Geological Survey’s Prompt Assessment of Global Earthquakes for Response (PAGER) alert system provides rapid (10–20 min) but general loss estimates of ranges of fatalities and economic impact for significant global earthquakes. FEMA’s Hazus software, in contrast, provides time-consuming (2–5 h) but more detailed loss information quantified in terms of structural, social, and economic consequences estimated at a much higher spatial resolution for large domestic earthquakes. We developed a rapid hybrid post-earthquake product that takes advantage of the best of both loss models. First, though, we conducted a systematic comparison of loss estimates from PAGER with Hazus for all significant, relatively recent, domestic earthquakes for which adequate loss data exist—augmented by a dozen ShakeMap scenarios. The systematic comparison of Hazus and PAGER losses provided the basis for selecting the specific loss metrics to present from each system. The signature product will serve as a supplement to the widely deployed PAGER alert product for significant domestic earthquakes.

Use of scenario earthquakes for seismic hazard assessment in low-seismicity, stable continental regions: A case study from Indiana, USA

“…Several studies have worked on validating and comparing losses from Hazus to real events (Kircher et al, 2006; National Institute of Building Sciences (NIBS), 2001; Remo and Pinter, 2012; Wald et al, 2020). The NIBS (2001) study examined only earthquakes in California and found that using site-specific soil types and actual building inventories could improve loss estimates by 10%–30% and up to 28%, respectively.…”

“…This study also showed that utilizing site-specific soil data resulted in a reduction of 8%–41% in estimated losses. Most recently, Wald et al (2020) compared Hazus and PAGER losses to each other for scenario events and to actual losses for real events. They compared 14 actual events and 10 scenario events, and of those 5 of each were located in the CEUS and concluded that Hazus estimates are consistently larger than reported economic losses (with loss ratios ranging from 1.3 to 16), especially for lower-magnitude events ( M < 6).…”

“…Nighttime Hazus-estimated fatalities are nearly equivalent for the Wabash scenario, and they are about a factor of 2 less for the Anna and Indianapolis scenarios. Wald et al (2020) concluded that the nighttime Hazus estimates are generally closer to the PAGER estimates than the daytime estimates due to PAGER using an ambient 24-h average population. We note, however, that the total number of fatalities for the Anna and Indianapolis scenarios is very low (only ;5-10 fatalities), whose differences result in large ratios.…”

“…PAGER uses intensity values to calculate losses and Hazus uses spectral acceleration values. Wald et al (2020) warn that the difference in how losses are calculated makes comparison between the two, particularly in the CEUS, difficult. Given all of these caveats and limitations, we nonetheless find that the independent estimates are readily comparable, with absolute values falling within the same order of magnitude for both parameters for all of our five main scenarios (Table A4 in supplementary material).…”

“…Given all of these caveats and limitations, we nonetheless find that the independent estimates are readily comparable, with absolute values falling within the same order of magnitude for both parameters for all of our five main scenarios (Table A4 in supplementary material). Based on their comparison of outputs of both PAGER and Hazus to real events, Wald et al (2020) concluded that Hazus better estimates economic losses while PAGER better estimates fatalities, with the caveat that for larger events, Hazus tends to overestimate economic losses and PAGER tends to underestimate fatalities. Figure 11a shows total economic losses calculated with Hazus versus those calculated with PAGER.…”