Kiarie, F.2016-05-162016-05-162016-04-30Vol-3, Issue-4 PP. 303-3082394-5788http://ir-library.ku.ac.ke/handle/123456789/14713Various parametric models have been designed to analyze volatility in river flow time series data. For maximum likelihood estimation these parametric methods assumes a known conditional distribution. This paper considers the problem of nonparametric estimation of critical streamflow discharge levels of a river regime based on quantile regression methodology of Koenker and Basset (1978).In particular, the paper demonstrates the use of kernel estimators for conditional quantiles resulting from a kernel estimation of conditional distribution function. It is finally proved that the estimate of the nonparametric quantile function is consistent and asymptotically normally distributed and under suitable conditions, the estimator converges uniformly with an appropriate rate.enConditional quantileKernel estimateQuantileautoregressionConsistencyAsymptotic normalityCritical discharge level.Article