The methodological framework for DRIP: Drought representation index for CMIP model performance

Abstract

This paper presents a methodological framework designed to evaluate the ability of CMIP climate models to simulate drought characteristics. The approach is based on the Drought Representation Index for CMIP Model Performance (DRIP), which assesses models using three key drought parameters-average duration, severity, and return period-by comparing simulated outputs with historical observations. The methodology encompasses four main stages: data acquisition and preparation, drought characterization, DRIP calculation, and model ensemble generation (E-DRIP). This approach provides a systematic method to identify models that best represent regional drought dynamics and reduce uncertainty in climate projections. By leveraging DRIP as a selection criterion, E-DRIP ensembles outperform traditional CMIP ensembles in both reliability and precision. The method's flexibility allows adaptation to various drought indices and temporal scales, making it applicable across diverse climatic contexts. Validation in a climatically uncertain area, the Paraíba do Sul River Basin in Southeast Brazil, demonstrates DRIP's effectiveness in enhancing model performance assessment and improving drought scenario projections. This study contributes a replicable tool for climate modelling, supporting water resources management strategies amid increasing climate variability.•DRIP index assesses CMIP models' performance in representing drought characteristics.•E-DRIP ensembles reduced drought projections uncertainties by up to 63 % in the validation study area.•DRIP enhances decision-making in climate model selection, improving its reliability for regional water planning.

Keywords: Climate ensemble; Drought Representation Index for CMIP Climate Model Performance (DRIP); Drought index; Global climate models.

Graphical abstract

Fig. 1

Flowchart summarizing the methodological framework.

Fig. 2

Detailed flowchart for Stage 2, showing: (2.1) Drought index calculation and (2.2) Drought pattern characterization.

Fig. 3

Detailed flowchart for Stages 3 and 4, showing (3) DRIP application and (4) E-DRIP ensemble generation, along with a comparative framework between E-CMIP and E-DRIP ensembles.

Fig. 4

Comparison of boxplots for drought duration and severity: observed data versus those identified by the models with the highest DRIP (best model) and lowest DRIP (worst model) by [28].

Fig. 5

Reduction in projection uncertainty for drought characteristics between E-CMIP and E-DRIP across all combinations of scenarios, time periods, and sub-basins.