An ensemble genetic programming model for seasonal precipitation forecasting
- PDF / 2,382,350 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 32 Downloads / 250 Views
An ensemble genetic programming model for seasonal precipitation forecasting Ali Danandeh Mehr1 Received: 25 May 2020 / Accepted: 5 October 2020 © Springer Nature Switzerland AG 2020
Abstract Seasonal precipitation forecasting is one of the most challenging tasks in stochastic hydrology. This article proposes a new ensemble model, called EGP, to a season-ahead forecast of total seasonal precipitation. The EGP integrates evolutionary genetic programming (GP) and gene expression programming (GEP) techniques with multiple linear regression to increase forecasting accuracy of standalone GP and GEP models, while it secures their explicit structure. The EGP model was trained and validated using 88 years (1930–2017) of measured precipitation data from Muratpasa Station, Antalya, Turkey. The model performance was evaluated in terms of different statistical error measures and cross-validated with two other ensemble models as well as the state-of-the-art random forest developed in this study as the benchmark. The results showed that the proposed model can increase the forecasting accuracy of the best standalone GP and GEP models up to 30%. The EGP was also found to be superior to random forest, particularly in predicting low and high seasonal precipitation amount. This model is explicit, easy to evolve, and therefore, motivating to be used in practice. Keywords Precipitation · Genetic programming · Seasonal forecast · Gene expression programming · Multiple regression · Random forest
1 Introduction It is expected that the recent increase in the concentration of atmospheric carbon dioxide and global temperature will have a significant impact on precipitation patterns in both global and regional scales [1, 7, 40]. Like the other water cycle components, the precipitation process is highly complex owing to the stochastic attributes of its triggering factors, such as temperature, wind speed, and humidity. The level of uncertainty in precipitation forecasting is significantly higher than the other water cycle components such as temperature and streamflow [19, 38]. Consequently, the problem has been addressed in a variety of recent studies (e.g., [8, 22, 33, 39]). Many efforts have been made over the past decades to model precipitation patterns at specific locations. Using the theory of Markov chains and classic statistical models, observed precipitation series have
been used to model and predict precipitating in different lead times [12, 43]. However, they are not well enough for long-lead-time forecasting owing to the highly nonlinear structure and presence of non-stationarity in the monthly or seasonal rainfall series [10, 15]. Nowadays, emerging artificial intelligent (AI) techniques such as artificial neural networks (ANNs), support vector machine (SVM), extreme learning machines, and genetic programming (GP) are from the AI techniques that were used for precipitation forecasting (e.g., [2, 11, 24, 34]). However, the techniques were criticized as a black box and the results were not precise enough, particularly in the pred
Data Loading...
