A self-calibrating effective drought index (scEDI): Evaluation against social drought impact records over the Korean Peninsula (1777–2020)

Abstract

The Effective Drought Index (EDI) is designed to monitor and characterize the drought condition at a daily scale, using the last 30 years of daily precipitation records as a climatic yardstick. A critique of the EDI is that the behavior of the index depends on the reference period, making comparisons of EDI values difficult over a long record period. Here, a self-calibrating EDI (scEDI) is calculated using the 244-year daily precipitation records (1777–2020) in Seoul, the Republic of Korea. The scEDI is evaluated by comparing with drought damage reports from the Annals of the Joseon Dynasty (1807–1907) and relevant online search activity volumes from Google Trends and NAVER DataLab data (2013–20). The scEDI automatically calibrates the behavior of the index over time by calculating the normal condition of antecedent precipitation based on a rolling 30-year period. As a result, the scEDI is more temporally comparable than the EDI, that is, droughts with the same intensity have the same frequency throughout the entire record period, regardless of wet and dry decades. Results show that a majority of drought damage records and spikes of public interest in droughts are found when droughts are moderate (-1.4 of scEDI) and severe (-2.0), respectively, implying that social drought impacts/response might occur during different intensities of an emerging drought. This study highlights the importance of temporally self-calibration when it comes to detecting and characterizing “social droughts” with social impact/response data over multi-centuries.