Abstract
In this paper, a stochastic model is presented to generate monthly inflows into the Hirakud Dam by using the principles of entropy and copulas. The marginal distributions obtained using the principle of maximum entropy (POME) helps the model to retain the higher-order moment statistics of the historical data, whereas the copula-based joint distributions ensure the preservation of the dependence between the adjacent months. On analysing the moments and dependence structures of the generated data samples and comparing with the corresponding observed data properties, it is found that the developed stochastic models gives good performance by preserving flow properties. Hence the entropy-copula-based model can be used as an effective approach for stochastic generation of hydrological variables.
Acknowledgement
This article is among the selected papers presented at the “Hydro-2012” conference held at IIT Bombay on December 7-8, 2012, and was short-listed by the Editor for publication in this Journal after re-review and revisions where necessary.