Wavelet Analysis of Seasonal Rainfall Variability of the Upper Blue Nile Basin, Its Teleconnection to Global Sea Surface Temperature, and Its Forecasting by an Artificial Neural Network (original) (raw)
Related papers
Monthly Rainfall Prediction Using Wavelet Neural Network Analysis
Water Resources Management, 2013
Rainfall is one of the most significant parameters in a hydrological model. Several models have been developed to analyze and predict the rainfall forecast. In recent years, wavelet techniques have been widely applied to various water resources research because of their timefrequency representation. In this paper an attempt has been made to find an alternative method for rainfall prediction by combining the wavelet technique with Artificial Neural Network (ANN). The wavelet and ANN models have been applied to monthly rainfall data of Darjeeling rain gauge station. The calibration and validation performance of the models is evaluated with appropriate statistical methods. The results of monthly rainfall series modeling indicate that the performances of wavelet neural network models are more effective than the ANN models.
Journal of Hydrologic Engineering, 2007
Wavelet-based empirical orthogonal function ͑WEOF͒ analysis was used to analyze the nonstationary spatial, temporal, and frequency regimes of the regional variability in southern African summer ͑October-March͒ rainfall. The leading modes of rainfall variability were then used to establish associations with gridded scale-averaged wavelet power of the sea surface temperature ͑SST͒ for the Indian and Atlantic Oceans. The WEOF revealed that southern African rainfall is out of phase between areas north and south of 25°S and that areas north of 25°S and northern South Africa experienced decreased rainfall between 1970 and 1997. The decrease in rainfall was modulated by periods of between 2 and 8 years. Using judiciously selected windows of April-May-June SST data for the Atlantic and Indian Oceans as predictors in the artificial neural network-genetic algorithm ͑ANN-GA͒, high prediction skill of standardized summer rainfall of southern Africa was achieved. For the validation period 1988-97, Pearson correlation between 0.83 and 0.98 ͑i.e., 69-96% of observed rainfall variability͒, Hanssen Kuipers skill scores of between 0.2 and 1.0, and root-mean-square errors of between 0.25 and 0.72 mm of standardized rainfall were found between observed and predicted summer rainfall at a 3-month lead time.
2014
Since 2000, western drought caused sharp drop by about 100 feet in the largest reservoir of North America, Lake Mead. About 97% of inflow into Lake Mead is supplied by Colorado River Basin which is extremely sensitive to changes in precipitation and temperature. Oceans play an important role on earth's climate via oceanic-atmospheric interactions known as climate teleconnections, which deeply affect the terrestrial precipitation patterns. This issue signifies the necessity of developing a modern hydroinformatics tool-precipitation forecasting model-to account for teleconnection signals from climate change and mitigate drought hazards impact on lake water, quantitatively and qualitatively, which cannot be achieved by using traditional Global Circulation Model. Therefore, understanding the relationship between precipitation and teleconnection patterns could be the first step for precipitation forecasting. However, highly non-linear and non-stationary nature of teleconnection patterns result in large uncertainties in estimates, since simple linear analyses failed to capture underlying trends at sub-continental scales. For this purpose, high-resolution remote sensing imagery, spectral analysis techniques, and wavelet analysis were integrated to explore the nonstationary and nonlinear behavior of teleconnection signals between the Pacific and Atlantic sea surface temperature (SST) on a short-term basis (10 years) from which the precipitation pattern shift in the Upper Colorado River Basin can be elucidated. These processes lead to the creation of correlation maps which specify index regions within the Atlantic and Pacific Oceans where SST anomaly can be statistically significant in correlation with terrestrial precipitation. These indexed regions delivering some kind of memory effects of SST were extracted to be inputs into an Artificial Neural Network (ANN). Advances in Wavelet-based Empirical Orthogonal Function and Artificial Neural Network (WEOF-ANN) model for rainfall prediction assists the local water management agencies to mitigate the drought impacts and obtain sustainable development strategies a month ahead of the time in urban drinking water infrastructure assessment plan around Lake Mead area.
This study investigates the accuracy of three data-driven models for short-term forecasting of the Total Dissolved Solid (TDS) loads at the outlets of three of the main drainage systems in the Nile Delta in Egypt. The three data-driven models considered in this study are multiple linear regression (MLR), artificial neural network (ANN) and a hybrid model that utilizes the wavelet transform (WT) and ANN (WT-ANN). The three models were applied to forecast TDS loads at one and three months lead times using historical records at the outlets of the Nashart, Bahr El-Bakar and Umom agricultural drains in the Nile Delta. For the MLR and ANN models, autocorrelation and partial autocorrelation analyses were used to identify the potential influencing input variables corresponding to different time lags. In the hybrid WT-ANN models, the Haar à trous wavelet transform was used. The TDS approximation and detail components (derived via the WT) were used as input variables in the ANN model. The performance of the three models was compared based on the mean absolute error, mean absolute relative error, root mean square error, and relative root mean square error. The results showed that the hybrid model produced significantly better results than the traditional MLR and ANN models for the short-term forecasting of the TDS loads.
Polish Journal of Environmental Studies
Climate change, population growth, industrialization, and environmental impacts cause spatiotemporal changes in the availability of regional water resources [1, 2]. In particular, climate change will affect the streamflow, temperature, amount of precipitation, and variability, which are the main components of the hydrological cycle [3-5]. For example, Jiao and Wang [6] state in their study that the streamflow and rainfall are in a decreasing trend while the temperature is in an increasing trend in the last decades. Modeling and outlining streamflow is a crucial process in water management and planning, and accurate streamflow prediction is a vital tool for optimal water quantity and quality management [7]. Studies on accurate projections of temporal streamflow patterns can aid in understanding the properties of hydrological processes in basins and improving basin modeling [8]. Many studies have been conducted that examined the relationship of streamflow with precipitation and temperature, and evaluated its changes and forecasts [9, 10]. Xu et al. [11] found that periodic changes in streamflow were closely correlated with temperature
A wavelet-SARIMA-ANN hybrid model for precipitation forecasting
Journal of Water and Land Development, 2016
Given its importance in water resources management, particularly in terms of minimizing flood or drought hazards, precipitation forecasting has seen a wide variety of approaches tested. As monthly precipitation time series have nonlinear features and multiple time scales, wavelet, seasonal auto regressive integrated moving average (SARIMA) and hybrid artificial neural network (ANN) methods were tested for their ability to accurately predict monthly precipitation. A 40-year (1970–2009) precipitation time series from Iran’s Nahavand meteorological station (34°12’N lat., 48°22’E long.) was decomposed into one low frequency subseries and several high frequency sub-series by wavelet transform. The low frequency sub-series were predicted with a SARIMA model, while high frequency subseries were predicted with an ANN. Finally, the predicted subseries were reconstructed to predict the precipitation of future single months. Comparing model-generated values with observed data, the wavelet-SARI...
Journal of Hydroinformatics
Streamflow forecasting is highly crucial in the domain of water resources. For this study, we coupled the Wavelet Transform (WT) and Artificial Neural Network (ANN) to forecast Gilgit streamflow at short-term (T0.33 and T0.66), intermediate-term (T1), and long-term (T2, T4, and T8) monthly intervals. Streamflow forecasts are uncertain due to stochastic disturbances caused by variations in snow-melting routines and local orography. To remedy this situation, decomposition by WT was undertaken to enhance the associative relation between the input and target sets for ANN to process. For ANN modeling, cross-correlation was used to guide input selection. Corresponding to six intervals, nine configurations were developed. Short-term intervals performed best, especially for T0.33; intermediate intervals showed decreasing performance. However, interestingly, performance regains back to a decent level for long-term forecasting. Almost all the models underestimate high flows and slightly overe...