Periodic Copula Autoregressive Model Designed to Multivariate Streamflow Time Series Modelling
- 51 Downloads
It is a challenge to develop models that can represent the stochastic behaviour of rivers and basins. Currently used streamflow models were constructed under rigid hypotheses. Hence, these models are limited in their ability to represent nonlinear dependencies and/or unusual distributions. Copulas help overcome these limitations and are being employed widely for modelling hydrological data. For instance, pure copula-based models have been proposed to simulate univariate hydrological series. However, there have been few studies on the use of copulas to model multivariate inflow series. Thus, the aim of this study is to develop a pure copula-based model for simulating periodic multivariate streamflow scenarios, wherein temporal and spatial dependencies are considered. The model was employed in a set of 11 affluent natural energy series from Brazil. We used the model to simulate many scenarios and analyze them through statistical tests such as Levene’s test, the Kolmogorov-Smirnov test, Kupiec test, and t-test. In addition, we investigated the spatial and temporal dependence of the scenarios. Finally, the critical periods of the simulated scenarios were investigated. The results indicated that the proposed model is capable of simulating scenarios that preserve historical features observed in the original data.
KeywordsNon-linear models Stochastic streamflow simulation Copula models for multivariate streamflow time series Periodic multivariate copula model
The authors would like to thank the National Council for Scientific and Technological Development (CNPq) of Brazil for support through the CNPq-PDJ program (grant number 150866/2017-8).
Compliance with Ethical Standards
Conflict of interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
- Hipel KW, Mcleod AI (1994) Time Series Modelling of Water Resources and Environmental Systems. Elsevier, AmsterdamGoogle Scholar
- Joe H (1996) Families of m-variate distributions with given margins and m(m-1)/2 bivariate dependence parameters, Distributions with fixed marginals and related topics, 120–141, Institute of Mathematical Statistics,L. Ruschendorf and B. Schweizer and M. D. Taylor, 28, Lecture Notes-Monograph SeriesGoogle Scholar
- Kupiec P (1995) Techniques for verifying the accuracy of risk measurement models. J Deriv 73–84Google Scholar
- Liu Z, Zhou P, Chen X, Guan Y (2015) A multivariate conditional model for streamflow prediction and spatial precipitation refinement. Journal of Geophysical Research:, Atmospheres 120Google Scholar
- Penna DDJ (2009) Definition of the streamflow scenario tree to long-term operation planning, D.Sc Thesis, Department of Electrical Engineering Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, BrazilGoogle Scholar
- Pereira GAA, Souza RC (2014) Long Memory Models to Generate Synthetic Hydrological Series, Mathematical Problems in Engineering, https://doi.org/10.1155/2014/823046, Article ID 823046, 1-8
- Salas JD, Delleur JW, Yevjevich JW, Lane WL (1980) Applied Modeling of Hydrologic Time Series. Water Resources Publications, First, MichiganGoogle Scholar
- Schepsmeier U, Stoeber J, Brechmann EC, Graeler B, Nagler T, Erhardt T (2017) VineCopula: Statistical Inference of Vine Copulas,R package version 2.1.1, http://CRAN.R-project.org/package=VineCopula
- Sklar M (1959) Fonctions de répartition à n dimensions et leurs marges. Publications de l’Institut de Statistique de L’Universit? de Paris 8:229–231Google Scholar
- Souto M, Moreira A, Veiga A, Street A, Garcia J, Epprecht C (2014) A high-dimensional VARX model to simulate monthly renewable energy supply. In:IEEE Power Systems Computation Conference, Wroclaw, Poland, pp 1–7Google Scholar