کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
6411512 1629924 2015 11 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Probabilistic drought classification using gamma mixture models
ترجمه فارسی عنوان
طبقه بندی خشکسالی احتمالی با استفاده از مدل های مخلوط گاما
موضوعات مرتبط
مهندسی و علوم پایه علوم زمین و سیارات فرآیندهای سطح زمین
چکیده انگلیسی


- A method is proposed for probabilistic drought classification.
- The method modifies SPI analysis using gamma mixture model.
- The method uses Bayesian framework to avoid overfitting.
- The method propagates model uncertainties to drought classification.

SummaryDrought severity is commonly reported using drought classes obtained by assigning pre-defined thresholds on drought indices. Current drought classification methods ignore modeling uncertainties and provide discrete drought classification. However, the users of drought classification are often interested in knowing inherent uncertainties in classification so that they can make informed decisions. Recent studies have used hidden Markov models (HMM) for quantifying uncertainties in drought classification. The HMM method conceptualizes drought classes as distinct hydrological states that are not observed (hidden) but affect observed hydrological variables. The number of drought classes or hidden states in the model is pre-specified, which can sometimes result in model over-specification problem. This study proposes an alternate method for probabilistic drought classification where the number of states in the model is determined by the data. The proposed method adapts Standard Precipitation Index (SPI) methodology of drought classification by employing gamma mixture model (Gamma-MM) in a Bayesian framework. The method alleviates the problem of choosing a suitable distribution for fitting data in SPI analysis, quantifies modeling uncertainties, and propagates them for probabilistic drought classification. The method is tested on rainfall data over India. Comparison of the results with standard SPI show important differences particularly when SPI assumptions on data distribution are violated. Further, the new method is simpler and more parsimonious than HMM based drought classification method and can be a viable alternative for probabilistic drought classification.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Journal of Hydrology - Volume 526, July 2015, Pages 116-126
نویسندگان
, , ,