River Water Level Research Articles

Pengaruh Cangkang Pomacea canaliculata L Terhadap Sifat Fisika dan Kimia Air Sungai Cisadane

The Cisadane River is a very important ecosystem for the people of the city of Tangerang. Its utilization has the potential to produce waste that can damage aquatic life so that it becomes one of the sources of the problem of decreasing water quality both physically and chemically. One of the efforts made to overcome the problem of Cisadane river water is to use water purifiers derived from natural materials so that they do not endanger health and can bind heavy metals. The purpose of this study was to determine the effect of adding snail shells to the water quality of the Cisadane river in industrial areas, urban areas and slums in terms pH value, TDS value and Pb heavy metal. The experimental procedures in this study were snail shell preparation, Cisadane river water sampling, sample analysis before application, snail shell powder application, and sample analysis after application. The results of analys show that the clarity level of Cisadane river water ranges from 19 NTU to 59 NTU, the pH value ranges from 7.2 to 7.9, the TDS value ranges from 60 mg/l to 170 mg/l and the heavy metal content is 0.0047 mg /l. The conclusion of this research is that there is an increase in the level of clarity with the addition of snail shells by 45%, there is not visible effect of the snail shells addition’s on the pH value and content of Pb heavy metal. And there is an increase in the TDS value after adding snail shells.

Satellite-based water surface slope over a small mountain river in northern China

Satellite altimetry has become a crucial alternative for measuring inland water surface elevations apart from ground observations. Water surface slope (WSS), a fundamental parameter in river discharge simulation, also allowing for assimilating satellite and ground-observed water level data with different observation positions. There are various satellite data products offering precise measurements of river water levels and estimates of WSS on a global scale. However, satellites observation of water surface over small rivers, particularly those in mountainous areas with narrow river channels, continues to be difficult. This study focuses on evaluating the precision of the ICESat-2 ATL03 photon height data in estimating WSS throughout the mountainous river section of Yongding River, which flows across the Hebei Province and Beijing City in northern China. The ICESat-2 ATL03 data with minimum along-track sampling interval of 0.7 m makes it possible to accurately estimate the WSS for small river sections that ranged from 50 to 100 m in width. A histogram-based statistical approach was used for identifying the satellite lidar photon height located right over the water surface. The eleven pairs of satellite virtual stations selected for extracting river water surface elevation and estimating river WSS demonstrated an overall validation correlation coefficient of 0.98. The range of relative error in WSS estimation is between 0.13 % and 9.02 %. The results of this work have important implications for using satellite-based measurements to estimate critical hydrological parameters in small and mountainous river basins in the absence of ground observations.

Flood Control Analysis with Polder Ponds in Tegal Made Village, Sukoharjo

A good drainage system should be able to drain back stagnant water into natural channels such as rivers and sea. Puddles and floods that occurred in Nawud Village in Sukoharjo Regency due to water from existing drainage channels cannot be directly discharged into the river because when the rain occure make the water level in river was higer than the water level from existing drainage. It causes water cant be discharged by gravity process and for this reason, a temporary water reservoir is needed. Design and capacity of the planned polder depends on the volume of water runoff. In determining the amount of runoff discharge to design a polder conduct a hydrological analysis based on rain that occurs in the area of ​​the study. Water runoff discharge analysis conducted using the HSS Nakayasasu method obtained runoff peak discharge of 33,3662 m3/sec and using the HSS SCS method obtained runoff peak discharge of 25,0379 m3/sec. Dimension plan of polder obtained that 15593 m2 is area of polder with 3,58 m pool hight which is planned to accommodate 55868,58 m3 a maximum puddle volume. the planned outflow volume is carried out using flow tracking analysis and using 3 pumps with each capacity of 2 m3 / sec.

Analysis of Sea and River Water Quality Standards Due to Operations and Domestic Activities in the Sanur Port Area, Bali

The development of an area has impacts, both positive and negative impacts. The positive impact is in the form of regional economic growth, employment and development of the surrounding regions. The negative impact is the disruption of the environmental ecosystem and a decrease in environmental quality due to waste generated from development operations, in the form of solid waste (garbage) and liquid waste. The increasing pollution load is also caused by public awareness and business activities in upstream areas that do not manage waste properly, which ultimately leads to all of the waste accumulating in water bodies and causing changes in water quality standards. Sanur Port, which is one of the development centers in its operations, is very vulnerable to environmental pollution impacts. The pollution mechanism originates from activities around Sanur Port and domestic activities which are discharged into rivers and seas in the Sanur area. This research aims to analyze the quality of river water and sea water caused by activities around Sanur Port and community activities based on Decree of the Minister of Environment Number 51 of 2004 and determine the pollution index. Water quality standards use Bali Governor Regulation no. 16 of 2016 concerning Environmental Quality Standards and Quality Standard Criteria for Environmental Damage. This research uses a quantitative approach by collecting primary data from direct observations in the field and secondary data based on previous research reports and government reports. The research results show that TDS in the rainy season tends to be lower than in the dry season, both in the results of testing samples of Abianbase river water, the results of testing samples of sea water in front of the Mak Beng Restaurant and sea water samples in the Sanur Port Pool. In front of Mak Beng, ammonia levels during the rainy season are higher than during the dry season. Ammonia levels in the Abianbase River and sea water in front of Sanur port are lower during the rainy season compared to the dry season. The presence of Coliform bacteria and E. coli shows a tendency to decrease during the rainy season. This decrease is thought to be caused by leaching by rainwater entering marine waters and retailing and changes in ocean currents during the rainy season.

Rainfall-Runoff Modeling Using Artificial Neural Network for Batu Pahat River Basin

This research delves into the effectiveness of Artificial Neural Networks with Multilayer Perceptron (ANN-MLP) and Nonlinear AutoRegressive with eXogenous inputs (NARX) models in predicting short-term rainfall-runoff patterns in the Batu Pahat River Basin. This study aims to predict river water levels using historical rainfall and river level data for future intervals of 1, 3, and 6 hours. Data preprocessing techniques, including the management of missing values, identification of outliers, and reduction of noise, were applied to enhance the accuracy and dependability of the models. This study assessed the performance of the models for ANN-MLP and NARX by comparing their effectiveness across various forecast timeframes and evaluating their performance in different scenarios. The findings of the study revealed that the ANN-MLP model showed robust performance in short-term prediction. On the contrary, the NARX model exhibited higher accuracy, particularly in capturing intricate temporal relationships and external impacts on river behavior. The ANN-MLP produces 99% accuracy for 1-hour prediction, and NARX yields 98% accuracy with 0.3245 Root Mean Squared Error and 0.1967 Mean Absolute Error. This study makes a valuable contribution to hydrological forecasting by presenting a rigorous and precise modeling methodology.

Inuit Traditional Ecological Knowledge of Anadromous Arctic Char, iqalukpik (Salvelinus alpinus) Under Changing Climatic Conditions in the Amundsen Gulf, Western Canadian Arctic

Inuit in the western Canadian Arctic have observed climate change impacts in marine and freshwater environments that have resulted in changes in the abundance, movement ecology, and health of Arctic char, iqalukpik (Salvelinus alpinus), with implications for their fisheries. This research was co-designed with Inuit to investigate reported changes in anadromous Arctic char in waters near Ulukhaktok, Northwest Territories, Canada. An analysis of semi-structured interviews with 26 Inuit fishers not only documented changes in Arctic char population abundance, spatial movement, appearance, and taste but also changes in access to the fishery. Over the past several decades and becoming pronounced in recent years, fewer fish and specifically fewer medium-sized fish that are preferred by fishers, have been caught using nets near the settlement, with some showing poor body condition, and others appearing to originate from lakes outside of the expected spatial range, earlier in an extended summer coastal fishing season. Inuit have observed changes in individual fish with broader environmental changes that are also disrupting fishing activities: the increasing prevalence of Pacific salmon, warmer air and marine sea surface temperatures, inconsistent sea and lake ice conditions, stronger and more frequent wind and wave activity, fluctuating water levels in rivers, and a seasonal tunicate bloom. Inuit have responded to these changes by altering personal subsistence fishing practices and temporarily halting a small-scale community-based commercial fishery because of observations in stock declines and in order to prioritize the subsistence fishery.

Hybrid deep learning models for multi-ahead river water level forecasting

Hybrid deep learning models for multi-ahead river water level forecasting

Seasonal changes of Chlorophyll-A concentration in Jiujiang city based on remote sensing

Abstract. The Chlorophyll-a (Chla) concentration is an important parameter characterizing the water quality of rivers and lakes. Satellite remote sensing provides new opportunities for the quantitative monitoring of Chla concentration in large-scale water bodies. In this study, Sentinel-2 satellite remote sensing data integrated on the Google Earth Engine (GEE) remote sensing big data platform are employed, along with hourly measured water quality site data, to establish a quantitative inversion model for Chla concentration water quality parameters in Jiujiang City within Jiangxi Province, China. The Chla concentration is estimated for each quarter from 2020 to 2022, and the spatial distribution is analyzed, revealing the changing trend of Chla concentration over the past three years. The key findings are as follows: (1) The quantitative inversion model for Chla concentration has been validated with measured data, achieving a model accuracy of 0.53; (2) The spatial inversion results of Chla concentration exhibit an increasing trend that is consistent with actual measurement site results; (3) Owing to the influence of human activities and the peak-low water level of rivers and lakes, Chla concentration shows a discernible seasonal variation pattern. This methodology offers a new perspective for analyzing the seasonal variation characteristics of Chla concentration in rivers and lakes, providing valuable insights for the sustainable management of river and lake water quality.

Impacts of High-Concentration Turbid Water on the Groundwater Environment of the Tedori River Alluvial Fan in Japan

The occurrence of high-concentration turbid water due to a large landslide in the upper reaches of the Tedori River Basin in Japan in May 2015 led to a rapid decline in the groundwater levels within the alluvial fan. However, factors other than turbid water, such as changes in precipitation patterns, can have a significant impact on groundwater levels but have not been thoroughly investigated. By analyzing the relationship between river water and groundwater levels, we found that by 2018, conditions had returned to those observed prior to the turbidity events. Regarding seepage, we found that approximately 24% of the Tedori River’s discharge contributed to seepage before the turbidity event. In contrast, during the post-turbidity years, seepage decreased between 2015 and 2017 and returned to the pre-turbidity levels by 2018. Furthermore, by constructing a hydrological model and examining the contributions of turbidity and precipitation, we found that in 2015, turbidity contributed to 76% of the groundwater level changes, whereas precipitation accounted for 24%. In contrast, in 2016, turbidity contributed to 67%, while precipitation contributed to 33%. In essence, the first year was characterized by a significant contribution from turbidity, while precipitation also played a significant role in groundwater level fluctuations in the second year.

Time-frequency connectedness between electricity prices in Romania and its determinants in the competitive markets

PurposeIn this paper, we aim to provide an extensive analysis to understand how various factors influence electricity prices in competitive markets, focusing on the day-ahead electricity market in Romania.Design/methodology/approachOur study period began in January 2019, before the COVID-19 pandemic, and continued for several months after the onset of the war in Ukraine. During this time, we also consider other challenges like reduced market competitiveness, droughts and water scarcity. Our initial dataset comprises diverse variables: prices of essential energy sources (like gas and oil), Danube River water levels (indicating hydrological conditions), economic indicators (such as inflation and interest rates), total energy consumption and production in Romania and a breakdown of energy generation by source (coal, gas, hydro, oil, nuclear and renewable energy sources) from various data sources. Additionally, we included carbon certificate prices and data on electricity import, export and other related variables. This dataset was collected via application programming interface (API) and web scraping, and then synchronized by date and hour.FindingsWe discover that the competitiveness significantly affected electricity prices in Romania. Furthermore, our study of electricity price trends and their determinants revealed indicators of economic health in 2019 and 2020. However, from 2021 onwards, signs of a potential economic crisis began to emerge, characterized by changes in the normal relationships between prices and quantities, among other factors. Thus, our analysis suggests that electricity prices could serve as a predictive index for economic crises. Overall, the Granger causality findings from 2019 to 2022 offer valuable insights into the factors driving energy market dynamics in Romania, highlighting the importance of economic policies, fuel costs and environmental regulations in shaping these dynamics.Originality/valueWe combine principal component analysis (PCA) to reduce the dataset’s dimensionality. Following this, we use continuous wavelet transform (CWT) to explore frequency-domain relationships between electricity price and quantity in the day-ahead market (DAM) and the components derived from PCA. Our research also delves into the competitiveness level in the DAM from January 2019 to August 2022, analyzing the Herfindahl-Hirschman index (HHI).

Double decomposition with enhanced least-squares support vector machine to predict water level

ABSTRACT As global climates undergo changes, the frequency of water-related disasters rises, leading to significant economic losses and safety hazards. During flood events, river water levels exhibit unpredictable fluctuations, introducing considerable noise that poses challenges for accurate prediction. A prediction of water level by using existing water level data makes a major contribution to forecasting flood. Enhanced least-squares support vector machine (ELSSVM) is utilized by integrating an additional extra bias error control term. In this study, least-squares support vector machine (LSSVM) and ELSSVM optimized by the genetic algorithm (GA) were chosen to be compared with the help of data decomposition methods to improve daily water level prediction accuracy. Double empirical mode decomposition (DEMD) will be integrated with LSSVM and ELSSVM. Thus, the models are named LSSVM-GA, ELSSVM-GA, empirical mode decomposition (EMD)-LSSVM-GA, EMD-ELSSVM-GA, DEMD-LSSVM-GA, and DEMD-ELSSVM-GA. The proposed models are used in forecasting the water level of Klang River in Sri Muda, Malaysia. The behavior proposed models are investigated and compared based on several performance metrics. The results demonstrated that the DEMD-ELSSVM-GA model outperformed the other models based on the performance analysis in forecasting the water level with RMSE = 0.2536 m and R2 = 0.8596 for testing data that indicate the forecasting accuracy.

Characterizing the impact of reservoir storage and discharge on nitrogen dynamics in an upstream wetland using a δ15N and δ18O dual-isotope approach

The identification of nitrate sources in reservoir water is important for watershed-scale surface pollution management. Significant fluctuations in river water levels arising from reservoir storage and discharge influence nitrate sources and transport processes. The Sanmenxia Reservoir, in the middle reaches of the Yellow River in China, undergoes significant water level changes (290–316 m), altering the composition of the nitrogen sources. This study employed a δ15N and δ18O dual-isotope method and MixSIAR modeling to quantify the contributions of nitrate sources. This reveals the impact of reservoir water impoundment and discharge on nitrogen dynamics in the upstream region of the wetland and the model sensitivity for each nitrate source. The results showed that the average concentrations of nitrate‑nitrogen (NO- 3-N) were elevated during the impoundment period compared to the discharge period. Nitrogen sources exhibited varying proportions in surface water, groundwater, and soil water during both the impoundment and discharge periods. The predominant sources include manure and sewage (MS), with a maximum proportion of 57.4 % in surface water. Soil nitrogen (SN) accounted for 25.8 % of groundwater nitrogen and 32.1 % of soil water nitrogen during the impoundment period, whereas, during the discharge period, soil nitrogen made up 41.4 % of surface water nitrogen, manure and sewage contributed 44.8 % of groundwater nitrogen, and manure and sewage dominated with 56.7 % of soil water nitrogen. Sensitivity analysis of the MixSIAR model revealed that the isotopic composition of the manure and sewage primary source most significantly influenced the apportionment results of the riverine nitrate source. Reservoir discharge facilitates the dissimilatory nitrate reduction to ammonium (DNRA). The migration of NO- 3 from surface water to soil water and groundwater occurred from the impoundment period to the discharge period.

Deep learning and LiDAR integration for surveillance camera-based river water level monitoring in flood applications

Deep learning and LiDAR integration for surveillance camera-based river water level monitoring in flood applications

Threshold-based flood early warning in an urbanizing catchment through multi-source data integration: Satellite and citizen science contribution

An effective flood early warning system is vital to take action to save lives and protect properties in urban areas which are increasingly prone to flooding. Despite substantial progress in flood early warning systems, limited available and accessible data often impede their advancement and reliability. Engaging communities affected by flooding can help address data and information gaps in flood early warning systems, facilitated by appropriate methods. This study developed and evaluated a flood threshold combination method to support a community-based flood early warning system in the Akaki catchment, home to Addis Ababa, the capital city of Ethiopia. Various flood threshold combinations were formulated, calibrated and validated by integrating multiple sources of data: rainfall, antecedent precipitation index estimates, Sentinel-1 Synthetic Aperture Radar satellite time series of flood extent, long-term simulated streamflow, citizen science data, river water level and three days lead-time numerical weather prediction rainfall forecast. During validation, the rainfall and river water level threshold combination outperformed other threshold combinations with probability of detection, false alarm ratio, and critical success index estimates of 0.74, 0.18 and 0.63, respectively. The flood threshold combination showed high detection performance for most flooding conditions. Flood forecasts with a 1-day lead-time exhibited a high likelihood in detecting historical severe flood events. The study provides a tested methodology for selecting suitable flood threshold-combinations, enhance the engagement of citizen scientists in a community–based flood early warning system in urban communities.

Analisis Penerapan Teknologi Sensor Berbasis Iot dalam Pemantauan Ketinggian Air Sungai

This research aims to compare conventional methods and Internet of Things (IoT) based sensors in measuring river water level. Conventional methods involve the use of manual measurement equipment and fixed measurement stations, while IoT-based sensors use sensor technology that is connected directly to an online monitoring system. This research was carried out by conducting a literature study and obtaining data on the water level of the Ciliwung River which was measured using the peilschaal method by BBWS Ciliwung Cisadane and the IoT-based electromagnetic sensor method by PANTIR (Pemantau Tinggi Air). Comparison of the two measurement methods is carried out from the aspect of accuracy of measurement data and performance based on the time required to obtain the data. The results obtained are in the form of a qualitative analysis which shows the advantages and disadvantages of each method.

PEMANFAATAN LIMBAH SERBUK GERGAJIAN KAYU ULIN (Eusideroxylon zwageri T & B) UNTUK MENURUNKAN KADAR FECAL COLIFORM AIR SUNGAI

Martapura river water is included in categories that have a fairly high bacterial content, which is 21,333 colonies/100 ml. The optimal value of water quality standards to be met with water treatment businesses is the use of ironwood sawdust waste. Bacteria in water attach most of their lives to suspension particles, so microscopic organisms are also separated or filtered. The purpose of this study was to analyze fecal coliform in river water and analyze the volume of ironwood sawdust to reduce fecal coliform levels in river water. Martapura river water contains the highest fecal 1,800MPN/100ml and the lowest 1,600MPN/100 ml, according to the Minister of the Republic of Indonesia No. 22 of 2021 concerning the implementation of environmental protection and management related to standards, including category two which is used as water tourism facilities/infrastructure, freshwater fish farming, animal husbandry, plant irrigation and other activities requiring the same water quality and use. Ironwood sawdust as much as 25% or 90 grams proved most effective to be used to reduce fecal levels with an average value on the first day of 1,650 MPN/100ml and on the third day with treatment carried out up to indigo 350 MPN/100ml, so that the difference reached a value of 1,300 MPN/100ml.

A Transfer Learning Approach Based on Radar Rainfall for River Water-Level Prediction

River water-level prediction is crucial for mitigating flood damage caused by torrential rainfall. In this paper, we attempt to predict river water levels using a deep learning model based on radar rainfall data instead of data from upstream hydrological stations. A prediction model incorporating a two-dimensional convolutional neural network (2D-CNN) and long short-term memory (LSTM) is constructed to exploit geographical and temporal features of radar rainfall data, and a transfer learning method using a newly defined flow–distance matrix is presented. The results of our evaluation of the Oyodo River basin in Japan show that the presented transfer learning model using radar rainfall instead of upstream measurements has a good prediction accuracy in the case of torrential rain, with a Nash–Sutcliffe efficiency (NSE) value of 0.86 and a Kling–Gupta efficiency (KGE) of 0.83 for 6-h-ahead forecast for the top-four peak water-level height cases, which is comparable to the conventional model using upstream measurements (NSE = 0.84 and KGE = 0.83). It is also confirmed that the transfer learning model maintains its performance even when the amount of training data for the prediction site is reduced; values of NSE = 0.82 and KGE = 0.82 were achieved when reducing the training torrential-rain-period data from 12 to 3 periods (with 105 periods of data from other rivers for transfer learning). The results demonstrate that radar rainfall data and a few torrential rain measurements at the prediction location potentially enable us to predict river water levels even if hydrological stations have not been installed at the prediction location.

From District to City Scale: The Potential of Water-Sensitive Urban Design (WSUD)

The summer of 2022 was one of the hottest and driest summers that Germany experienced in the 21st century. Water levels in rivers sank dramatically with many dams and reservoirs running dry; as a result, fields could not be irrigated sufficiently, and even power generation and supply were affected. The impact of abnormally high temperatures for extended periods (heatwaves) is not restricted to nature and the economy but is also a considerable public health burden. Experts worldwide agree that these extreme weather events are being driven by climate change and will increase in intensity and frequency in the future. The adverse impact of these extreme weather events multiplies among dense urban environments, e.g., through heat islands. This calls for cities to take action to heat-proof and water-secure their urban developments. Water-Sensitive Urban Design (WSUD) is one such approach to mitigate the aforementioned challenges by leveraging the urban water ecosystem with special attention to the subject of water reclamation, retention, treatment and distribution. This paper introduces and builds upon a prototype of WSUD that centers around an artificial lake as an integrated water resource management system (IWRMS) fed by treated grey water and storm water obtained from two housing blocks flanking the water reservoir. Based on the specifications of this prototype, indicators of site suitability are derived and applied to identify potential locations for replicable projects in the city of Darmstadt. The results confirm the impact WSUD can have: a total of 22 sites with 2527 apartments are found suitable for prototype implementation in Darmstadt. Savings in town water consumption from these 22 sites would add up to 147 million liters. Further benefits include the provision of 24 million liters of irrigation water, storm water retention, adiabatic cooling during heatwave, increased biodiversity and the improvement in livability of the sites and the city.

Floods and Land Use Land Cover Changes in the Kalu River Basin: A Call to Action


 
 
 Floods are one of the most common natural disasters worldwide. Apart from rainfall, Land Use Land Cover (LULC) changes too are a main contributory factor for floods. This study attempted to understand the link between floods and LULC changes in Kalu river basin, which is the second largest river basin and an area that experiences recurrent floods in Sri Lanka. We studied peak water levels, number of flood events, changes in land use types and impacts in rapidly urbanizing two districts, Rathnapura (upper basin) and Kalutara (lower basin) during 2001-2020. The satellite images (LANDSAT) were obtained for 2001, 2009, 2015 and 2020 and land use classification was done using ArcGIS and Remote Sensing Tools. Main land use types and their transformations were investigated and ground-truthing was carried out. Accordingly, the main types of land uses identified were Natural Vegetation and forests (NV), Settlements (ST- housing and industrial lands), Cultivated Lands (CL), Water Bodies (WB) and Bare Lands (BL). The results indicated that the most drastic change was found in the natural areas (NV) and they have diminished while the lands with anthropogenic impacts (ST, CL and BL) have increased across years. The NV had occupied the highest land area in 2001 (42.4%) and has reduced by 14.2% by 2020. The ST and CL have increased by 8.6 % and 5.2% respectively. The monthly rainfall of Rathnapura and Kalutara (Source: Department of Meteorology, Sri Lanka) has increased with time, which is a main reason for the increasing peak water levels of these areas (Source: Department of Irrigation, Sri Lanka). However, a significant correlation also exists between the change of the settlement area with the peak river water levels in the lower basin (p=0.03, R2=99%; regression analysis). Rathnapura has experienced 3 major floods (floods above the high water alert level) from 2001-2020, while 16 major floods have occurred in Kalutara. During the major flood in 2017, the number of child deaths in Rathnapura was 14 while in Kalutara it was 24. Accordingly, the LULC changes of the whole basin along with rainfall seem to influence on the severity of floods in Kalutara more, as it is located in the lowest elevation level. When natural lands are transformed to anthropogenic- impacted areas with disturbances to the water cycle, increased impervious surfaces, reduced water storage capacities and loss of natural drainage, the flood risk tends to increase. Proactive approaches including proper land use planning and rainwater storage are urgently needed as the climate change too would trigger more floods. Thus, the flood mitigatory actions, especially, in the lower river basin should be a priority to ensure resilience and sustainability.
 Keywords: Kalu river basin, Land Use Land Cover (LULC) changes, Floods
 
 

Flood prediction based on feature selection and a hybrid deep learning network

Abstract As climates change globally, water-related disasters increase, causing substantial economic losses and safety risks. During floods, river water levels show unpredictable fluctuations, introducing substantial noise that complicates accurate prediction. A hybrid model that uses eight-dimensional input data from hydrological and meteorological stations is proposed to address these challenges. Initially, the Variational Mode Decomposition preprocesses and denoises water level data, resulting in decomposed Intrinsic Mode Functions (IMFs). Then, the Pearson correlation coefficient between each IMF and input characteristics is computed, and the fluctuation factor for each IMF is defined. IMFs are categorized based on a threshold, leading to a hybrid prediction model. This model integrates convolutional neural networks (CNNs) for spatial information and bidirectional long short-term memory (BiLSTM) networks with an attention mechanism for learning from past and future data points. Comparative evaluations of mean absolute percentage error, root mean square error, mean absolute error, and goodness of fit (R2) show that the proposed model outperforms existing LSTM and CNN–BiLSTM frameworks, reducing RMSE by at least 20% and increasing R2 by approximately 10% on average. The model's practical significance lies in improving the accuracy and efficiency of meteorological forecasting and flood warning systems, contributing substantially to global disaster preparedness and response strategies.