Disaster Damage Records of EM-DAT and DesInventar: A Systematic Comparison

This study compares the disaster damage records of two of the most widely used disaster databases - EM-DAT and DesInventar for a selected sample of 70 countries over the period 1995–2013. We consider four types of natural disasters – droughts, floods, earthquakes and storms and use descriptive statistics to compare the records of the two databases for the selected datasets in terms of all recorded events, matched events and large-scale events along with few country-specific comparisons. We note significant differences in the damage estimates reported in the two selected datasets. The comparison of the damage estimates for all recorded events shows that the DesInventar dataset has greater number of recorded events than the EM-DAT. The descriptive statistics of the former exhibits larger mean and standard deviation across all disaster types compared to that for the latter. The same is true for the comparison of large-scale disaster events and country-specific comparisons with the DesInventar dataset having higher number of recorded events and larger values for descriptive statistics for the selected country-year. On the contrary, for the hand-matched events data, EM-DAT shows larger mean disaster damages along with a higher statistical range compared to the DesInventar dataset. The basic structure of the datasets and the data collection methods may influence the magnitude of the recorded damages along with possible human errors while data entering. The present study further highlights the need for a more systematic and standardized disaster damages database which is critical to achieve the first priority action ‘understanding disaster risk’ of the Sendai Framework for Disaster Risk Reduction.