PROOF COPY [HE/2001/022449] 002305QHE PROOF COPY [HE/2001/022449] 002305QHE Nonparametric Approach for Estimating Return Periods of Droughts in Arid Regions Tae-Woong Kim 1 ; Juan B. Valde ´ s 2 ; and Chulsang Yoo 3 Abstract: Droughts cause severe damage in terms of both natural environments and human lives, and hydrologists and water resources managers are concerned with estimating the relative frequencies of these events. Univariate parametric methods for frequency analysis may not reveal significant relationships among drought characteristics. Alternatively, nonparametric methods provide local estimates of the univariate and multivariate density function by using weighted moving averages of the data in a small neighborhood around the point of estimation and opposed to parametric methods. A methodology for estimating the return period of droughts using a nonparametric kernel estimator is presented in order to examine the univariate as well as the bivariate behavior of droughts. After evaluating and validating a nonparametric kernel estimator, a drought frequency analysis is conducted to estimate the return periods of droughts for the Conchos River Basin in Mexico. The results show that, for the univariate analysis, the return periods of the severe drought occurring in the 1990s are 100 years or higher. For the bivariate analysis, the return periods are approximately 50 years for joint distributions and more than 120 years for the conditional distributions of severity and duration. DOI: 10.1061/ASCE1084-069920038:51 CE Database subject headings: Droughts; Arid lands; Frequency; Estimation. Introduction Hydrologic systems are impacted by extreme events such as floods and droughts, which cause severe damage to natural envi- ronments and human lives. For example, the 1987–1989 drought, which lasted 3 years and covered 36% of the United States, caused approximately $39 billion in losses of energy, water, the ecosystem, and agriculture National 2000. These extreme events are usually expressed by return periods in hydrology and water resources engineering. When the concept of the return period is applied to drought-related variables, the return period will be the average time between events with a certain magnitude or less Haan 1977. In general, a drought is defined as a sustained period of significantly lower soil moisture levels and water supply rela- tive to normal levels. During droughts, water supplies are inad- equate to meet the water demand of management systems, and lack of rainfall adversely affects the environment and human so- ciety Dracup et al. 1980. A significant number of studies on droughts deal with the defi- nition of droughts, low-flow frequency analysis, and climatic im- pacts of droughts Smakhtin 2001. Lee et al. 1986developed a practical approach for the frequency analysis of multiyear drought durations of annual streamflow series. A technique that smooths the frequency-curve irregularity of drought durations was used to reduce the statistical uncertainties associated with sample-size limitations. Nathan and McMahon 1990evaluated the applica- tion of the Weibull distribution to low-flow frequency analysis. They investigated the differences between low-flow frequency es- timates based on calendar and hydrologic years, and they made recommendations for the selection of an appropriate subset of the data. In recent years, several papers were published addressing the methodology for estimating drought return periods Ferna ´ ndez and Salas 1999a,b; Chung and Salas 2000. Ferna ´ ndez and Salas 1999a,bsummarized the definitions of the return period and estimated the risk of failure of hydraulic structures, especially applied to drought events. They estimated the return periods and the associated risks of failure of hydrologic events related to me- teorological droughts, low flows, annual maximum floods, and hydrological droughts, which are either dependent or indepen- dent. Chung and Salas 2000dealt with drought occurrence prob- abilities, return periods, and risks of drought events for dependent hydrologic processes. Rather than using Markovian models, which traditionally have been used for modeling hydrologic pro- cesses, low-order discrete autoregressive moving average ARMAmodels were used for modeling wet and dry years, since they are adequate for processes exhibiting longer time depen- dence. They concluded that low order ARMA models are useful for representing the occurrence of wet and dry periods and, con- sequently, are capable of modeling and simulating drought condi- tions that are observed historically. A general approach used in drought-related frequency analyses is to derive the distributions of drought durations and severities separately Sen 1976; Mathier et al. 1992; Stedinger et al. 1992. A drought event, however, is a multivariate event characterized by its duration, magnitude, and intensity Salas 1992, which are 1 Graduate Research Assistant, Dept. of Civil Engineering and Engi- neering Mechanics, and Center for Sustainability of Semi-Arid Hydrol- ogy and Riparian Areas SAHRA, The Univ. of Arizona, Tucson, AZ 85721-0072. E-mail: taek@email.arizona.edu 2 Professor and Head, Dept. of Civil Engineering and Engineering Me- chanics, and Center for Sustainability of Semi-Arid Hydrology and Ri- parian Areas SAHRA, The Univ. of Arizona, Tucson, AZ 85721-0072. E-mail: jvaldes@u.arizona.edu 3 Associate Professor, Dept. of Civil and Environmental Engineering, Korea Univ., Seoul, Korea 136-701. E-mail: cyoo@korea.ac.kr Note. Discussion open until February 1, 2004. Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and pos- sible publication on October 18, 2001; approved on January 21, 2003. This paper is part of the Journal of Hydrologic Engineering, Vol. 8, No. 5, September 1, 2003. ©ASCE, ISSN 1084-0699/2003/5-1–10/$18.00. JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / SEPTEMER/OCTOBER 2003 / 1 PROOF COPY [HE/2001/022449] 002305QHE