Covariance between subsample mean and variance as related to storage variables Article

Bayazit, M, Obeysekera, JTB, Yevjevich, V. (1985). Covariance between subsample mean and variance as related to storage variables . JOURNAL OF HYDROLOGY, 78(1-2), 137-150. 10.1016/0022-1694(85)90158-1

cited authors

  • Bayazit, M; Obeysekera, JTB; Yevjevich, V

authors

abstract

  • The covariance, cov( x ̄, s2x), between mean and variance of subsamples is conceived as a single parameter which integrates the effects of asymmetrical extremes and dependence structure on storage capacity required for seasonal flow regulation. An exact expression for this covariance for the first-order autoregressive processes is derived. The use of cov( x ̄,s2x) for estimation of skewness is investigated by simulation for both independent and dependent stationary processes. This approach is shown to result in a lesser bias in estimated skewness than the moment estimator. The relationship between cov( x ̄,s2x) and the storage related variables, such as range and deficit, is also studied by simulation. It is found that cov( x ̄,s2x) is positively correlated with mean range and deficit for skewness smaller than 3, and negatively correlated with adjusted range. It is concluded that these storage related variables are influenced in different ways by the sampling fluctuations in the coefficients of skewness and serial correlation. © 1985.

publication date

  • May 30, 1985

published in

Digital Object Identifier (DOI)

start page

  • 137

end page

  • 150

volume

  • 78

issue

  • 1-2