Choice of Hydrologic Variables for Probabilistic Drought Classification

Abstract

Management of irrigation practices takes on a more-prominent role under drought conditions. Watershed-scale drought characterization is performed using cumulative density function (CDF)-based probabilistic drought indices in this case study. To investigate the role of hydrologic variables, either singly or in combination, copulas are used for multivariate joint cumulative density functions (CDFs) combined with graphical models for probabilistic drought classification. Adopting a multivariable, multiscalar approach in the proposed framework yields a drought index that allows for examining the roles of hydrologic variables on integrated drought assessment. The methodology is demonstrated using streamflow, precipitation, and soil-moisture anomalies to develop univariate and multivariate CDF-based indices at one-month, three-month, and six-month time scales to analyze the drought events over an Indiana watershed. Drought characterization varied across the univariate, bivariate, and trivariate drought models in the case study. The multivariate models were able to capture the early onset of drought events and persistence of the drought states, features that are contributed by different components of the hydrologic cycle. While short-term drought monitoring is facilitated by one-month models, threats to long-term water storage in the watershed can be assessed better with longer-time-scale models.