• Salzano, E.; Iervolino, I.; Fabbrocino, G., "Seismic risk of atmospheric storage tanks in the framework of quantitative risk analysis", 2003
• A. Di Carluccio, I. Iervolino, G. Manfredi, G. Fabbrocino, E. Salzano, "Quantitative probabilistic seismic risk analysis of storage facilities", 2006
• Eidinger J. M.; Avila, E. A.; Ballantyne, D.; Cheng, L.; der Kiureghian, A.; Maison, B. F.; O'Rourke, T. D.; Power, M., "Seismic fragility formulation for water systems", 2001

See Part 1, Table 5.7 on page 69 in Eidinger et al. (2001). Source data is given in Table 5.6 on page 68.

Low standard deviation (beta) values for the curves ≥DS2 (μ = 0.15, β = 0.12) and =DS5 (μ = 1.13, β = 0.10) were indicated to reflect the sample set. However, a beta value of 0.80 is recommended for use for all damage states for regional loss estimates unless otherwise justified.

See Di Carluccio et al. (2006) Table 2 (page 219) for probit constants (PGA in g), which supercedes Salzano et al. (2003) Table 5 (page 408) (PGA in g · 10²). Use of probit constants for ≥DS2 and =DS5 should be avoided, because they are for the low beta values.