Investigating Rainfall Persistence in the Seybouse Watershed of Northeastern Algeria using the ARFIMA model

Balah Belkacem; Zeroual Abdelatif

doi:10.56093/aaz.v65i1.170211

Authors

Balah Belkacem Department of Hydraulic, Faculty of Sciences and Applied Sciences, University of Oum El Bouaghi, Oum El Bouaghi https://orcid.org/0000-0002-6400-1737
Zeroual Abdelatif Department of Hydraulic, Faculty of Sciences and Applied Sciences, University of Oum El Bouaghi, Oum El Bouaghi https://orcid.org/0000-0001-7290-5157

https://doi.org/10.56093/aaz.v65i1.170211

Keywords:

ARFIMA, Seybouse watershed, rainfall, hurst exponent, persistence

Abstract

This research paper examines the persistence of precipitation forecast memory by estimating the fractional differentiation parameter (d) of the AutoRegressive Fractionally Integrated Moving Average (AutoRegressive Fractionally Integrated Moving Average (ARFIMA)) model and the corresponding Hurst exponent (H) using the maximum likelihood method (LM) and nonlinear least squares estimation method (NLS). The optimal method is selected based on criteria such as minimal AIC, portmanteau, and residue tests. The study focuses on monthly rainfall data from four representative stations in the Seybouse watershed area in the Algerian Northeast, covering the period from January 1, 1998, to December 31, 2007. To ensure data normality, the Cox-Box transformation is applied to model the studied process. The constructed Autoregressive Fractionally Integrated Moving Average (ARFIMA) models exhibit the following characteristics: Annaba station: (5; 0.13; 14) with a Hurst exponent (H) of 0.63, indicating long forecast memory; Oum El Bouaghi station: (0; 0.11; 0) with (H) of 0.61, indicating long forecast memory; Guelma station: (9; -0.16; 4) with (H) of 0.34, indicating long forecast memory; and Souk Ahras station: (2; -0.37; 2) with (H) of 0.13, indicating anti-persistence.

Downloads

Download data is not yet available.

References

Anctil, F. and Coulibaly, P. 2004. Wavelet analysis of the interannual variability in southern Québec streamflow. Journal of Climate 17:163-173.

Aouad, H.S., Taouli, M.K. and Benbouziane, M. 2012. Modélisation du comportement du taux de change du dinar algérien: une investigation empirique par la méthode ARFIMA. International Research Journal of Finance and Economics 87: 117-133. https://mpra.ub.uni-muenchen. de/38605/1/MPRA_paper_38605.pdf

Balah, M. 2016. Analyse statistique de la persistance des séries pluviométriques dans le bassin versant de la Seybouse. Ph.D. Thesis, University of Annaba, Algeria.

Beran, J. 1995. Statistics for long-memory processes. Journal of the Royal Statistical Society: Series B 57: 659-672.

Beveridge, S. and Oickle, C. 1993. A new approach for testing the random walk hypothesis. Journal of Money, Credit and Banking 25(1): 1-14.

Bouanani, A. 2004. Hydrologie et modélisation du bassin versant de la Seybouse. Ph.D. Thesis, University of Annaba, Algeria.

Box, G.E.P. and Cox, D.R. 1964. An analysis of transformations. Journal of the Royal Statistical Society: Series B 26(2): 211-252.

Cao, J., Cai, X. and Guo, S. 2018. Long memory characteristics of hydrological time series. Hydrological Sciences Journal 63(6): 865-878.

Cekim, H.O., Gunes, H. and Karaca, M. 2021. Long memory processes in precipitation records. Theoretical and Applied Climatology 143: 1245-1258.

Chung, C.F. and Baillie, R.T. 1993. Small sample bias in conditional sum-of-squares estimators of fractionally integrated ARMA models. Empirical Economics 18: 791-806.

Corduas, M. 2000. ARFIMA models in hydrological time series analysis. Statistical Methods and Applications 9: 139-156.

D’Elia, A. and Piccolo, D. 2004. Statistical methods for long-memory processes. Computational Statistics & Data Analysis 45(3): 549-564.

Doornik, J.A. and Ooms, M. 2004. Inference and forecasting for ARFIMA models with an application to US and UK inflation. Studies in Nonlinear Dynamics and Econometrics 8(2):1-34.

Ghachi, A. 1986. Étude hydrologique du bassin versant de la Seybouse. Engineering Report, Algeria.

Granger, C.W.J. and Joyeux, R. 1980. An introduction to long-memory time series models and fractional differencing. Journal of Time Series Analysis 1(1): 15-29.

Guo, J., Zhou, J. and Wang, Y. 2019. Long-term persistence in hydrological time series. Ecological Indicators 96: 669-683.

Hosking, J.R.M. 1981. Fractional differencing. Biometrika 68(1): 165-176.

Kadi, A. 1997. Étude hydrologique et climatique de la région de la Seybouse. Technical Report, Algeria.

Kharbbouch, I. and Daoui, A. 2018. Long memory modelling of rainfall data. Arabian Journal of Geosciences 11: 1-12.

Lai, D. 2004. Estimating the Hurst effect and its application in monitoring clinical trials. Computational Statistics & Data Analysis 45(3): 549-564. https://doi.org/10.1016/S0167- 9473(03)00085-9

Ludwig, W., Serrat, P., Cesmat, L. and Garcia- Esteves, J. 2004. Mediterranean river systems and climate variability. Hydrological Processes 18: 3449-3466.

Machiwal, D., Santra, P., Meena, H.M. and Singh, A. 2025. Long-term rainfall variability and persistence analysis in arid regions. Journal of Hydrology 640: 131112.

McCoy, E.J. and Walden, A.T. 1996. Wavelet analysis and synthesis of stationary long-memory processes. Journal of Computational and Graphical Statistics 5: 26-56.

Meddi, M., Hubert, P. and Meddi, H. 2010. Hydrological variability in North Africa. IAHS Publication 340: 95-102.

Melard, G. 1979. Algorithm AS197: Maximum likelihood estimation of ARMA processes. Applied Statistics 28: 324-329.

Na, L., Na, R., Zhang, J., Tong, S., Shan, Y., Ying, H., ..., Bao, Y. 2018. Vegetation dynamics and diverse responses to extreme climate events in different vegetation types of Inner Mongolia. Atmos 9(394).

Olbermann, B., Sibbertsen, P. and Reimers, H.E. 2006. Forecasting long-memory time series models. Computational Statistics 21: 445-462.

Rana, R., Staron, M., Berger, C., Hansson, J., Nilsson, M. and Törner, F. 2013. Comparing between maximum likelihood estimator and non-linear regression estimation procedures for NHPP software reliability growth modelling. In: 2013 Joint Conference of the 23rd International Workshop on Software Measurement and the 8th International Conference on Software Process and Product Measurement, pp. 213-218, IEEE.

Retia, M. and Gaidi, K. 2017. L’estimation des modeles ARFIMA avec erreurs Garch du cours du Dow Jones. Cahiers du CRÉAD 33: 27.

Robinson, P.M. 1995. Gaussian semiparametric estimation of long range dependence. Annals of Statistics 23: 1630-1661.

Sadaei, H.J., Enayatifar, R. and Lee, M. H. 2016. Forecasting long memory time series with neural networks. Neurocomputing 175: 782-796.

Serinaldi, F. and Kilsby, C.G. 2015. Long-term persistence in rainfall records: implications for flood risk estimation. Advances in Water Resources 77: 17-36.

Tang, S., Dufour, J. and Gauthier, G. 1991. Simulation of long-memory processes. Simulation 57(5): 303- 310.

Valerie, M. 1998. Hurst exponent estimation methods and their applications in hydrology. Economy and Forecasting 1-2: 193.