Empirical fragility curves reinforced concrete and timber houses, using different intensity measures

Chile is a country affected by large scale seismic events, among which are those that occurred in the years 1960 (Valdivia, Mw=9.5, the biggest magnitude in history), 2010 (Maule, Mw=8.8), 2014 (Iquique, Mw=8.2) and 2015(Illapel, Mw=8.3) earthquakes. The 2010 Maule earthquake and tsunami caused damages in 370.051 houses. From these houses, 81.444 collapsed, 108.914 suffered severe damage, and 179.693 suffered slight damage. The 2014 earthquake caused damage in 9680 houses. Among these, 4.582 suffered severe damage and 38.100 inhabitants were estimated to be affected. The latest large earthquake in Chile occurred in 2015, affecting the Coquimbo Region. The estimated number of damaged houses in this earthquake is 6.763, where 1.420 houses collapsed. The damaged houses in this region represent 68% of the housing inventory. Chile has empirical data that can be used to estimated future damage. The estimation of the impact of earthquakes on the building inventory can support the development of risk reduction strategies. Past events are useful to calibrate risk models and they contribute to the understanding of the consequences of earthquakes. The objective of this study is to estimate empirical fragility curves of reinforced concrete and timber houses using damage information of the 2014 and 2015 earthquakes. The fragility curves are estimated using four Intensity Measures (IM) to identify which IM is better correlated with the observed damage. The methodology used has three-steps: (1) organize the damage observed for 2014 and 2015 earthquakes according to the Hazus damage scale (slight, moderate, extensive and collapse), (2) Using
the following IMs: Modified Mercalli Intensity (MMI), Peak Ground Acceleration (PGA), Peak Ground Velocity (PGV) and Spectral Acceleration at a period of 0.3 seconds (SA(0.3)), and (3)the empirical fragility curves are estimated using Maximum Likelihood Estimation and the methodology according to FEMA P-58. The obtained results indicate that, for the houses analyzed, SA(0.3) is the IM that better correlates with observed damage.