Kabul River Flow Prediction Using Automated ARIMA Forecasting: A Machine Learning Approach

Musarat, Muhammad Ali; Alaloul, Wesam Salah; Rabbani, Muhammad Babar Ali; Ali, Mujahid; Altaf, Muhammad; Fediuk, Roman; Vatin, Nikolai; Klyuev, Sergey; Bukhari, Hamna; Sadiq, Alishba; Rafiq, Waqas; Farooq, Waqas

doi:10.3390/su131910720

Cited by 22 publications

(11 citation statements)

References 61 publications

(58 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These techniques include probability characteristics, time series methods, synthetic data generation, multiple regression, pattern detection, and neural network methods [16]. It is commonly acknowledged that time series modelling is a better choice in the areas of prediction problems [17], which describe the pattern or stochastic behavior of a non-linear problem [18,19]. According to the time series modelling, reasonable results have been reported for most areas of the contiguous United States (US) and China.…”

Section: Introductionmentioning

confidence: 99%

Water Level Prediction through Hybrid SARIMA and ANN Models Based on Time Series Analysis: Red Hills Reservoir Case Study

et al. 2022

Self Cite

View full text Add to dashboard Cite

Reservoir water level (RWL) prediction has become a challenging task due to spatio-temporal changes in climatic conditions and complicated physical process. The Red Hills Reservoir (RHR) is an important source of drinking and irrigation water supply in Thiruvallur district, Tamil Nadu, India, also expected to be converted into the other productive services in the future. However, climate change in the region is expected to have consequences over the RHR’s future prospects. As a result, accurate and reliable prediction of the RWL is crucial to develop an appropriate water release mechanism of RHR to satisfy the population’s water demand. In the current study, time series modelling technique was adopted for the RWL prediction in RHR using Box–Jenkins autoregressive seasonal autoregressive integrated moving average (SARIMA) and artificial neural network (ANN) hybrid models. In this research, the SARIMA model was obtained as SARIMA (0, 0, 1) (0, 3, 2)12 but the residual of the SARIMA model could not meet the autocorrelation requirement of the modelling approach. In order to overcome this weakness of the SARIMA model, a new SARIMA–ANN hybrid time series model was developed and demonstrated in this study. The average monthly RWL data from January 2004 to November 2020 was used for developing and testing the models. Several model assessment criteria were used to evaluate the performance of each model. The findings showed that the SARIMA–ANN hybrid model outperformed the remaining models considering all performance criteria for reservoir RWL prediction. Thus, this study conclusively proves that the SARIMA–ANN hybrid model could be a viable option for the accurate prediction of reservoir water level.

show abstract

Section: Introductionmentioning

confidence: 99%

Water Level Prediction through Hybrid SARIMA and ANN Models Based on Time Series Analysis: Red Hills Reservoir Case Study

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…There is a crucial necessity to perform an accurate forecast of drought occurrence especially for a longer timescale [81]. Using past data to predict future drought behavior, hidden information can be disclosed that is of vital significance for mitigating the effects of droughts on water resources [75,82]. This paper adopts the SPEI at a 12-month timescale as a drought index to describe the annual dry-wet conditions in China.…”

Section: Climate Change Characterized By Global Warming Has Become An...mentioning

confidence: 99%

“…To achieve this, the time series was examined for stationarity using the augmented Dickey-Fuller (ADF) test. The formulation of the ADF test is given by [82]:…”

Section: Autoregressive Integrated Moving Average Modelmentioning

confidence: 99%

“…When the parameters are well estimated, the tentative model accuracy is validated by examining the autocorrelations of the residuals in order to simulate the white noise process [32]. If the validation fails to satisfy the parameters of the best-fit model, the analysis is repeated for a different model selection using appropriate lags [82]. Finally, the SPEI-based drought forecasting was performed for the year 2022 with a forecast step of 12 months.…”

Section: Diagnostic Checkingmentioning

confidence: 99%

See 1 more Smart Citation

Evidence for Intensification in Meteorological Drought since the 1950s and Recent Dryness–Wetness Forecasting in China

Yang

Bing

2022

Atmosphere

View full text Add to dashboard Cite

Drought is one of the major environmental stressors; drought is increasingly threatening the living environment of mankind. The standardized precipitation evapotranspiration index (SPEI) with a 12-month timescale was adopted to monitor dry–wet status over China from 1951 to 2021. The modified Mann–Kendall (MMK) and Pettitt tests were used to assess the temporal trend and nonlinear behavior of annual drought variability. The analysis focuses on the spatio-temporal structure of the dry–wet transition and its general connections with climate change processes. In addition, the seasonal autoregressive integrated moving average (SARIMA) model was applied to forecast the dry–wet behavior in the next year (2022) at 160 stations, and the hotspot areas for extreme dryness–wetness in China were identified in the near term. The results indicate that the dry–wet climate in China overall exhibits interannual variability characterized by intensified drought. The climate in the Northeast China (NEC), North China (NC), Northwest China (NWC), and Southwest China (SWC) has experienced a significant (p < 0.05) drying trend; however, the dry–wet changes in the East China (EC) and South Central China (SCC) are highly spatially heterogeneous. The significant uptrend in precipitation is mainly concentrated to the west of 100° E; the rising magnitude of precipitation is higher in Eastern China near 30° N, with a changing rate of 20–40 mm/decade. Each of the sub-regions has experienced significant (p < 0.01) warming over the past 71 years. Geographically, the increase in temperature north of 30° N is noticeably higher than that south of 30° N, with trend magnitudes of 0.30–0.50 °C/decade and 0.15–0.30 °C/decade, respectively. The response of the northern part of Eastern China to the warming trend had already emerged as early as the 1980s; these responses were earlier and more intense than those south of 40° N latitude (1990s). The drying trends are statistically significant in the northern and southern regions, bounded by 30° N, with trend magnitudes of −0.30–−0.20/decade and −0.20–−0.10/decade, respectively. The northern and southwestern parts of China have experienced a significant (p < 0.05) increase in the drought level since the 1950s, which is closely related to significant warming in recent decades. This study reveals the consistency of the spatial distribution of variations in precipitation and the SPEI along 30° N latitude. A weak uptrend in the SPEI, i.e., an increase in wetness, is shown in Eastern China surrounding 30° N, with a changing rate of 0.003–0.10/decade; this is closely associated with increasing precipitation in the area. Drought forecasting indicates that recent drying areas are located in NWC, the western part of NC, the western part of SWC, and the southern part of SCC. The climate is expected to show wetting characteristics in NEC, the southeastern part of NC, and the eastern part of EC. The dry–wet conditions spanning the area between 30–40° N and 100–110° E exhibit a greater spatial variability. The region between 20–50° N and 80–105° E will continue to face intense challenges from drought in the near future. This study provides compelling evidence for the temporal variability of meteorological drought in different sub-regions of China. The findings may contribute to understanding the spatio-temporal effect of historical climate change on dry–wet variation in the region since the 1950s, particularly in the context of global warming.

show abstract

“…Studies on river ow prediction using deep learning [39][40][41][42][43][44] show that the algorithms used to predict river ow are limited in predicting outliers using data-based algorithms. Therefore, we propose the PNP algorithm, which can predict the ow of rivers using climate change-sensitive precipitation.…”

Section: Related Workmentioning

confidence: 99%

Deep-learning algorithm development for river flow prediction: PNP algorithm

Bak

Bae

2022

Preprint

View full text Add to dashboard Cite

Deep-learning algorithms developed in recent decades have performed well in prediction and classification using accumulated big data. However, as climate change has recently become a more serious global problem, natural disasters are occurring frequently. When analyzing natural disasters from the perspective of a data analyst, they are considered outliers, and the ability to predict outliers (natural disasters) using deep-learning algorithms based on big data acquired by computers is limited. To predict natural disasters, deep-learning algorithms must be enhanced to be able to predict outliers based on information such as the correlation between the input and output. Thus, algorithms that specialize in one field must be developed, and specialized algorithms for abnormal values must be developed to predict natural disasters. Therefore, considering the correlation between the input and output, we propose a positive and negative perceptron (PNP) algorithm to predict the flow rate of rivers using climate change-sensitive precipitation. The PNP algorithm consists of a hidden deep-learning layer composed of positive and negative neurons. We built deep-learning models using the PNP algorithm to predict the flow of three rivers. We also built comparative deep-learning models using long short-term memory (LSTM) to validate the performance of the PNP algorithm. We compared the predictive performance of each model using the root mean square error and symmetric mean absolute percentage error and demonstrated that it performed better than the LSTM algorithms.

show abstract

Kabul River Flow Prediction Using Automated ARIMA Forecasting: A Machine Learning Approach

Cited by 22 publications

References 61 publications

Water Level Prediction through Hybrid SARIMA and ANN Models Based on Time Series Analysis: Red Hills Reservoir Case Study

Water Level Prediction through Hybrid SARIMA and ANN Models Based on Time Series Analysis: Red Hills Reservoir Case Study

Evidence for Intensification in Meteorological Drought since the 1950s and Recent Dryness–Wetness Forecasting in China

Deep-learning algorithm development for river flow prediction: PNP algorithm

Contact Info

Product

Resources

About