Capability of meteorological drought indices for detecting soil moisture droughts

Study region: Eastern Australia Study focus: Long-term monitoring of soil moisture is a time- and cost-intensive challenge. Therefore, meteorological drought indices are commonly used proxies of periods of significant soil moisture deficit. However, the question remains whether soil moisture droughts can be adequately characterised using meteorological variables such as rainfall and potential evaporation, or whether a more physically based approach is required. We applied two commonly used drought indices the Standardized Precipitation Index and the Reconnaissance Drought Index to evaluate their performance against soil moisture droughts simulated with the numerical soil water model Hydrus-1D. The performance of the two indices was measured in terms of their correlation with the standardised simulated monthly minimum soil water pressures, and their capability to detect soil moisture droughts that are potentially critical for plant water stress. New hydrological insights for the region: For three typical soil types and climate zones in Eastern Australia, and for two soil profiles, we have found a significant correlation between the indices and soil moisture droughts detected by Hydrus-1D. The failure rates and false alarm rates for detecting the simulated soil moisture droughts were generally below 50% for both indices and both soil profiles (the Reconnaissance Drought Index at Melbourne was the only exception). However, the complexity of Hydrus-1D and the uncertainty associated with the available, regionalised soil water retention curves encourage using the indices over Hydrus-1D in absence of appropriate soil moisture monitoring data.