Publication: Two-variate drought frequency analysis based on modified twodimensional drought index
Loading...
Date
2024-09-01
Authors
Ihsan Faseeh, Hasan Al-Mostafa
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Drought events are predicted to occur more frequently due to climate change,
that may adversely affect surface water resources and rainfed agriculture. Therefore,
understanding drought and reliable assessment of its risk is crucial for an early alert
system in arid and semiarid areas. In the current study, the Standardized Effective
Precipitation Index (SEPI) and Standardized Runoff Index (SRI) were utilized to
investigate the two-variate frequency of agricultural and hydrological drought
characteristics adopting the copula methodology. Utilizing Kendall's tau, Spearman's
rho, and Pearson's correlations, the dependence between severity and duration was
estimated. The marginal distributions that optimally fit the drought's severity and
duration have been identified according to the Chi-Squared, Kolmogorov-Smirnov,
and Anderson-Darling statistics. To choose the most appropriate copula model, three
Archimedean family and Gaussian copulas were contrasted applying Nash–Sutcliffe
Efficiency (NSE), Bayesian information criteria (BIC), and Root Mean Square Error
(RMSE). Uni-variate and two-variate return periods are estimated and compared.
Generally, the results revealed that Archimedean copulas outperformed Gaussian
copulas, with Archimedean Gumbel RMSE of 0.08 and BIC of -97, versus Gaussian
RMSE of 0.09 and BIC of -92, in the case of SRI-12. The results also revealed an
increase in the frequency of drought events, with most drought events having short
joint return periods of less than 10 and 5 years for agricultural and hydrological
droughts, respectively. The challenges of inadequate or non-existent soil moisture
data in the study area can be overcome by the SEPI, which only needs effective precipitation data. Depending just on one index or variable is inadequate to perform
an accurate assessment of drought risk. Multivariate indices can reliably analyze
droughts. Therefore, the Multivariate Standardized Effective Precipitation Runoff
Index (SEPRI) was modified and constructed through a combination of effective
precipitation and runoff variables based on the copula methodology. The significant
drought characteristics of the SEPRI were identified utilizing the theory of run. The
drought-events' two-variate frequency is computed in terms of the joint probability
and joint return period. The results indicate that SEPRI is applicable and can
comprehensively and effectively characterize drought events as contrasted to the
outputs of other drought indices. The correlations between SEPRI and both SEPI and
SRI were 91% and 90%, 89% and 88%, and 96% and 94% at Tusan, Eski Kalak, and
Dokan Stations, respectively. The novelty of present study emerges from the
drought-events' two-variate frequency analysis adopting the modified two-
dimensional SEPRI. As well as the novelty of adopting the copula model in the two-
variate frequency analysis of characteristics of agricultural drought utilizing effective
precipitation. The current study can give helpful information for reliable assessment
of drought-risk and water-resources management under climate change.