Uncertainty in Global Drought Severity

Abstract:

Global depictions of drought are often used in historical climate analysis, future climate projections, and hazard risk assessments. Uncertainties surrounding the classification of drought severity are rarely accounted for in descriptions of drought at the global scale. One potential source of uncertainty is that methods for classifying drought are not standardized to represent local conditions. Drought is a complex and multidimensional natural hazard and the wide variety of global climate types makes it difficult to standardize drought severity classifications. Despite this, operational drought definitions are frequently used in global drought research. This study provides a critical evaluation of potential biases and uncertainties that arise when using the same fixed severity thresholds across diverse climate classifications. The results of this study demonstrate that fixed thresholds are inadequate for depicting drought at a global scale because distributions of drought indices vary with space and time. Objective drought thresholds are formulated globally by matching drought index distributions to percentile classifications designated by the US Drought Monitor (USDM). In comparison to fixed index values, objectively defined thresholds vary significantly and better represent local drought conditions in a climatological context. Uncertainty is greatest in arid and saturated climate classifications where precipitation distributions are difficult to standardize. The different methods used for drought classification are a significant source of uncertainty when monitoring global drought events.