Creation of Hydrodynamic Models for Flood Disaster Mitigation Along a River Basin

1.0 Introduction

Due to increase in climatic change, the use of flood predictive models is increasingly gaining importance for flood impact mitigation. In the present study, an attempt has been made to understand flow dynamics along the lower Niger basin in Nigeria by drawing inferences from statistical analysis of flood transmission using past flood hydrological data and geometries of the lower Niger River flow route. The major objective was to investigate flow dynamics along the lower Niger basin in Nigeria and develop reliable statistical models for predicting water level and flood transmission time between different locations along the basin during extreme weather. The frequency of flood occurrences and its devastating impacts on global communities, eco-environment and bio-diversity have been of great concern worldwide [1]. Flood is defined as a large amount of water covering an area that is usually dry [2]. It occurs when inflow exceeds the carrying capacity of a channel or Dam [3; 4].  Flooding has been recognized as one of the most dangerous disasters detrimental to the environment and human [5; 6]. Global annual average loss caused by floods is estimated to be 104 billion United State Dollars [7]. In recent times the rate and impacts of floods is increasing due to several factors including rising sea level, increased physical developments on flood plains etc. There is consensus within the scientific community that climatic change caused by anthropogenic activities has increased the brutality of precipitation events [8; 9; 10].  Floods by nature are complex events caused by improper development planning, climatic change and anthropogenic activities which changes river morphology [11]. Climate dynamic can change the hydrologic regime of river basins including low, high, and medium stream flows [12].

The International Panel on Climate Change (IPCC) reported that global temperatures will rise by 1.4–5.8 °C by 2100 and this will result to changes in climatic variables (rainfall, temperature, etc.) at regional and sub-catchment levels. Although flood hazard is a natural phenomenon, human modification and alteration of natural flood plains accentuates the severity of its consequences. Of all geo-hazards, flooding is the most common and costly [13]. Flood can destroy the natural eco-environment. The depletion of plants with its attendant soil erosion and loss of habitats for animals, may result to a decline in biodiversity and other environmental degradation problems [14]. Floods can result to destruction of landscape, natural habitats, bio-diversity as well as the survival strategies of human communities due to soil erosion and rapid water runoff. [15; 16]. Floods do not only damage properties and endanger the lives of human and animals but also produce other secondary effects like displacement of residents, and disruption of livelihoods as well [17]. According to [18] floods are one of the most frequent natural hazards and have led to human losses, large economic losses, the destruction of fertile land, and damages to properties and infrastructure.

Flooding has become a global phenomenon averaging four billion dollars annually in property damage alone [19]. In the period from 1963 to 1992 the world suffered from a total of 202 floods (most of which were flash floods, monsoon floods, headwater floods or landslides) with more than one million victims [20; 21]. In the year 1970 the coastal area of Bangladesh recorded one of the worst flood disasters of the 20th century with a cyclone which killed 500,000 people [22]. Within the period 1996 and 2005 several African countries were affected by a total of 290 flood disasters which left 8,183 people dead, displaced 23 million people and caused economic losses of $1.9 billion [23]. A report by the UN-HABITAT in 2010 predicted that within ten-year period (2010-2020) more than 25% of Africa’s population living within 100km from the coasts will be at risk from sea level rise and coastal flooding [19; 24]. In line with this prediction, the Natal Province of South Africa experienced heavy flood caused by a record-breaking five-day rainstorm in 2009 which killed at least 400 people and left 55,000 people homeless.

In most cities of Nigeria, extreme rainfall, high urbanization rate with the attendant obstruction of streams and channel due to poor waste disposal behaviors and other human activities at flood plains have been found to constitute the major causes of floods [25]. The 2012 flood event in Nigeria was attributed to the release of water from Lagdo Dam in Cameroun and extreme weather. The resultant effect was devastation of low-lying areas along the lower Niger-Benue basin. Despite the benefits of Dams sudden release of water or its failure can result to a high magnitude of flooding in downstream areas of a basin [26]. 

The 2012 flood was very disastrous, affecting a total of 14 States with 1.3 million people displaced and about 431 people lost their lives. It affected most communities living within 5kilometers from the River Niger and River Benue trough. The National disaster impact assessment rating (IAR) instituted by National Emergency Management Agency (NEMA) categorized the affected areas into group A, B, C and D in descending level of damage and part of the study area falls under group A and others in group B. In 2018 and 2022 Nigeria experienced similar levels of flood disaster. These events were described as the worst in the nation’s history with over 30% of the entire country completely submerged,   resulting to the death of over 600 persons and displaced more than 2 million people from their homes in each episode. [27]. The 2018 and 2022 flood in Nigeria was also attributed to extreme rainfall which coincided with the period of excess water release from Lagdo Dam and Jebba Dam. When disaster of a particular magnitude occurs in a country, efforts should be made to prevent future re-occurrence or at least minimize the impact through various preventive programs and mechanisms. For the fact that the magnitude of flood of 40-year return interval re-occurred within a space of 6 and 4 years with similar impact on the same communities implies that adequate mitigation measures have not been put in place.

Sustainability can only be achieved if there is efficient management of natural environment. In Africa, floods are among the most serious natural disasters, due to inadequate forecasting technologies [28]. As per  [29], lack of good predictive capabilities is among the problems faced by disaster management agencies in this region in putting in place early warning sign and response mechanisms to tackle flooding. For example, communities along the lower Niger basin explained that the floods caught them by surprise and gave them no time to prepare, while others reported that they woke up in the morning to discover that everywhere has been toppled by flood [27].

As per [30], accurate predictions of water stages during high flow periods are fundamental to water resources management and flood control operations. Predictive models are invaluable in disaster analysis. It is never an easy task, because to develop a hydrodynamic model, the behavior of the river processes must be known from historical records. A hydrological model is a mathematical simulation of the hydrological process in a catchment. It can be used to better understand and explain hydrological dynamics [31]. In an ungauged catchment where stream flow data are lacking, the use of simulation models to generate runoff volumes and extreme flows on the intended stream from rainfall data is a necessity [32; 33]. Statistical models have proved very useful for flood impact mitigation. For example, [34] conducted flood frequency analysis of Ikpoba River catchment in Benin using Log Pearson type III distribution model. [35], also carried out flood frequency analysis of Osse River using Gumbel’s distribution. Studies have shown that there is relationship between flow speed and channel geometrics. For example, [36] noted that geometric method is an alternative mode of estimating flood. Flow route geometrics are crucial parameters for alternative techniques of discharge estimation in an un-gauged river. As per [37], hydrological response from each catchment assists in flood routing vis-a-vis flood modeling and flood forecasting. [38], estimated the peak discharge of un-gauged river Kunur in Bendal river Basin during heavy flood using indirect methods with channel geometrics. Flood impacts can be mitigated or minimized by a combination of both the structural and nonstructural control measures. The structural mitigations are designed to divert flood water away from people; for example, building of retention ponds, levees, channel improvements, dams etc. [39]. Non-structural measures place people away from flood. They are designed to reduce the impact of flooding to society and economy, and include flood insurance, flood forecasting and warning schemes, building codes, land use planning and zoning, use of hydrological models etc. [39].

In the run up to any disaster, two things are of key importance to improve safety: these are the spatial component and time component. Two questions that should be of concern in this situation are: where is the disaster expected to occur and when? Spatial components determine where the disaster is likely to occur while the time component predicts the time of occurrence of disaster in space. Knowledge of time of occurrence of flood events improves safety and is of great importance for a variety of planning purposes. It allows for precautions to be taken and for people to be warned so that they can be prepared in advance to respond to flood events. For example, farmers can remove animals from low-lying areas, the elderly and children can be relocated in advance, and emergency services can also make provisions for enough resources ahead of time to respond to disaster as it occurs. Communities, faced with flooding challenges, develop diversified resilience strategies.  Although, many studies have been conducted on flood risk across the world but most of them focused on the spatial component. There is dearth of research involving the use of hydro-dynamic models for flow transmission time and river stage prediction to support flood impact reduction particularly in Nigeria therefore, this underscores the need for this study. In this work, flow transmission time and river stage hydrodynamic models have been developed using River channel’s geometrics and flow historical records. These derivatives are useful for flood impact mitigation and management of future flood disasters within the study area

2.0 Material and Methods

2.1 Description of study area

The basin under consideration in this study covers a total land mass of approximately 70,959 square kilometers. It cuts across four major catchment areas along the lower Niger basin. These include Kogi, Edo, Anambra, and Delta River Niger basins. The most important hydro-geological features are River Niger and River Benue. The Niger River originates from Guinea in West Africa, runs through Mali, Niger, enters Nigeria through Niger State and extends to about 4000 km. The Benue River originates from the Adamawa Plateau of Northern Cameroon, flows west through the towns of Garoua and Lagdo reservoir, through Jimeta, Ibi and Makurdi in Nigeria before meeting the Niger River. These two Rivers formed a confluence at Lokoja, Kogi State in Nigeria before flowing southwards, dividing the basin into two wings and finally discharge into Atlantic Ocean in the southern coasts of Delta state. Geographically the study area is located between latitudes 5°.00`N and 8°.45`N, then longitudes 5°.00`E and 7°.45`E and lies within elevation range of 70m to 520m above mean sea level. Figure (1) is the location map of the study area.

2.2 Data Description

The Shuttle Radar Topography Mission (SRTM) imagery of one arc seconds resolution covering the study area was acquired through the U.S. Geological Survey Earth Resources Observation and Science Center (http://earthexplorer.usgs.gov/). Other datasets include hydrological data which comprised of bathymetric (sounding) data, time series flood discharge and River stage data collected from National Inland Water

Ways (NIWA) Lokoja. These River stage data are information recorded at permanent gauging stations along the basin. Additional River stage records were collected from temporary gauging stations installed along the basin by our research group. The permanent stations are located at Lokoja, Ajeokuta, Idah, Illushi and Onisha while the temporary stations were sited at Ogbaru, Ndokwa east, Patani and Bomadi at lower part of the basin in. Other datasets include LANSAT and IKONOS Satellite imagery of the study area collected before and during the flood episodes. These datasets have been analyzed in GIS window and the spatial extent of the river before and during the flood episodes were investigated to gain understanding of flood extent. Secondary data were collected from Ground Truthing. These include communities residing within 5kilometers from the river which were affected by the floods. This study has been carried out as per the methodology shown in Figure 2

2.3 Flood and riverbed topography analysis

In hydrological studies the rule of thumb is that speed and direction of water flow are governed by the channel’s (geometrics) configuration and shape of the surface in which it flows. Proper understanding of the terrain morphology is invaluable for prediction of the direction and speed of flowing water because water flows naturally by gravity. Satellite imageries that captured the floods were used to map the flood extents. The pre-flood imageries were used as references to determine the spatial extent of the floods through change detection analysis and later validated by ground truthing.  The river layer and flood layer from each flood episode were superimposed on digital elevation model (DEM) of the study area to assess the  elevation of the inundated areas and its influence on spatial spread of the floods (figure 4). Assessment of the riverbed morphology along the flow route was also carried out using bathymetric data. The reduced levels (Z) of the riverbed were processed from the chart depths (H) and reduced level (X) of the chart datum (see   figure 3). The reduced hydro-terrain datasets were used for graphical assessment of the riverbed’s topography (figure5a), longitudinal profile (figure 5b-c) and cross profiles (figure 5d). The area of the cross sections, the wetted perimeters and the longitudinal slopes of the Riverbed in the direction of flow at different locations were digitally measured and calculated. Progressively, the hydraulic radius at the various cross sections has been obtained as a function of the cross-sectional area and the wetted perimeter.

Predictive analysis was made in two categories. The first category was based on Gauge Relation. This is an empirical hydrological modeling approach that estimates flows at downstream based on those of upstream. The idea here was to develop statistical models which are capable of estimating flood levels at downstream locations based on water level at upstream locations along the basin. Accurate river stage prediction is a crucial component in the early flood warning system and plays a key role in flood disaster mitigation. The flood warning system is a vital mitigation technique during natural disasters that can be used by river managers to make decisions before the peak flow. Two main approaches are used to establish flood prediction models. These are physics-based model and data-driven model. The physics-based model requires cross-sectional bed elevation data or digital elevation model data for establishing the simulation domain. Therefore, simulation results obtained using a physics-based model are highly dependent on the quality of topographic data. Data-driven model is based on the collection and analysis of data and does not require riverbed elevation data. It is the approach applied in river stage forecasting in this present study. The second category applied physical based approach and is based on the notion that flow speed between different locations along a River course is a function of the geometrics of the channel between these locations. Therefore, the idea here was to statistically model the relationship between flow transmission times along different river segments with the metrics of the flow route along these segments. This is done with the view of using the developed models as predictive tools during future storms.

2.4 Statistical background.

Statistical background of this study is based on least square regression principle which states that of all statistical models defining relationship between variables, the one having the property that R₁² + R₂² + R₃² +—R_n² is minimum is said to relate the data in least square sense and is called best fitting model [40]. Where R is the difference between the observed and the estimates. As per [40], any relationship between events (variables) in space, can be statistically modelled. The model can be used to predict any of the missing variable (unknown) provided others are known

2.5 Predictive modelling through Gauge Relations

A total of 1080 daily water level data observed at different gauge stations along the basin collected during 2012, 2018 and 2022 flood episodes were used for this study. 540 hydro data points of these datasets captured at the gauging stations during 2012 and 2018 flood episodes were statistically combined to develop the models by applying least square regression technique while the same number of hydro-datasets collected at the same locations during 2022 flood event were used for validation. Estimation of flood levels at point B in response to flood stage at point A is the problem the model is expected to address. Both linear and nonlinear models were derived and tested using least square criterion however, this study reported and presented the models that met the least square requirement. Estimates of the observed were computed from the models developed. These estimates were compared with their observed equivalents. The residuals were fundamental to the assessment of model’s result using different evaluation metrics. (For want of space, only part of the data is tabulated in appendix A and B).The goodness of fit of the models developed were assessed by applying least square criterion. The statistical models that met least square requirement were selected.  Validations were carried out in relative and absolute senses to investigate the reliability of the models developed. Relative validations were made using hydrographs (figure 6) and absolute validation was made by comparison between the model’s estimates and flood level at the same location captured during 2022 flood episode and accuracies of the models were calculated at 95% confidence limit.

2.6 Modeling based on channel’s geometrics and time lag of transmission

This model involves statistical estimation of flow transmission time as a function of geometrics of the flow route. It is well known that speed of flowing water is a function of the character of surface in which it is flowing. During the flood episodes, communities along the basin reported that the flood caught up with them by surprise and gave them no time to prepare therefore this call for attention. Flow transmission time between different locations along the basin observed during 2012 and 2018 flood disaster was tabulated. Its variation in response to change in metrics of river flow route was assessed and statistically modelled. The channel’s metrics used include average hydraulic radius and slopes of these channel segments. Hydraulic radius is the ratio of the cross-sectional area to the wetted perimeter. Hydraulic radius and slope are fundamental factors that influence flow speed along a river course. For example, Manning developed an empirical model that estimates flow speed as a function of slope and hydraulic radius (see equation 3). The observes quantities (flow transmission time (Tr) along the river segments) were made dependent (predicted) variables while average slope (S) and hydraulic radius (R) along these segments were the independent variables. Value of the constants (k’s) was derived by solving the normalized equation (see equation 1). Accuracy assessment was made, and the model was further evaluated by comparing the estimated variables (time) with their observed equivalents recorded between the same locations during 2022 flood episode.

2.7 Flow velocity analysis

Additional validation was made using estimated flow velocities along the basin  carried out by applying Manning’s flow model (see equation 2).

2.8 Accuracy Assessment of Model results

Accuracy is the most important in determination of the fitness of geographic data. In statistical analysis, it is extremely helpful to use a single number to judge a model’s performance. Five evaluation criteria were used for result assessment these are Nash–Sutcliffe efficiency (NSE), Coefficient of correlation (R²), Mean absolute error (MAE), root-mean-square error (RMSE), Peak water-level error (PWE). Graphical assessment was also utilized as additional evaluation criterion (see figure 6 a, b, c, and d). NSE was initially proposed by Nash–Sutcliffe and is widely used to evaluate hydrological models. It is one of the best performance metrics for assessing the overall fit of a hydrographic record [42]. NSE is the ratio of the mean square error to the variance of observed data subtracted from unity. Its value ranges from – ∞ to 1. The closer the value of NSE to 1, the better the prediction during modeling. A negative NSE value indicates poor predictive performance. The coefficient of correlation more commonly known as R² quantifies the strength of the relationship between variables in a model. R² value ranges from 0 to 1. Zero (0) value indicates no relation, while the value of 1 indicates that the predicted values are equal to the observations. MAE indicates the number of errors obtained by a model, and the optimal value is zero. RMSE measures the average difference between values predicted by a model and the actual values. It provides an estimation of how well the model can predict the target value (accuracy). It aggregates the residuals into a single measure of predictive value. PWE is the difference between highest observed water level and highest predicted water level during the period under investigation.

where, n is the total number of data; O^p is the predicted water stage; O^m is the observed water stage; is the predicted peak water stage and is the observed peak water stage.

3.0 Results and Discussion

3.0.1 Result

Figure 4 a and b are the real-world views of 2012 and 2022 flood episodes respectively in the study area showing the spatial extents of floods and elevation of the affected areas. Water accumulates in areas of low-elevation making these regions more exposed   to flooding. Figure 5a, b, c and d are respectively, the topography of the riverbed, 3-D riverbed’s longitudinal profile, 2-D, longitudinal profile of the riverbed and part of the cross profiles of the flow route. These terrain derivatives were fundamental to generation of the channel’s geo-metrics which served as input to the various statistical analysis.

Figure 6a, b, c, and d  are hydrographs showing observed (Solid black line) and predicted (red broken lines) daily water level at four permanent gauging sites during 2012 flood period. The solid blue lines are the observed water level at the  reference station. Considering the time series plots, it is evidenced that the predicted water levels approach the observed significantly.

Table 4a-11a, 4b-11b (Appendix A) and       4c -11c (Appendix B) are part of the observed and predicted river stage during 2012, 2022 and 2018 flood episode respectively showing the discrepancies between the observed and estimated water levels at stations where physical observations were made. The residuals (discrepancies) are fundamental to assessment of model’s result. 

3.2 Discussion

3.2.1 Topographical assessment

The Topographical models (figure 4a and 4b) revealed undulating hydro-terrain of the basin. It showed that areas of elevation ranging from 20m to 120m above mean sea level are exposed to high level of flooding. These areas are low-lying areas located within 5 kilometers from the river course. Some of the communities located within this regions are listed in the last column of table2. Below. The riverbed topography (figure5) is made of slopes of various degrees with sharp decrease in elevation from the upstream towards the downstream region of the basin. The varied slopes influences flow movement along the basin.  The cross profiles (figure 5d) revealed a varying channel shape with depths varying between 50m to 100m at various locations. These attributes play significant roles in flow dynamics along the basin.

3.2.2 Analysis of river stage simulation models

The river stage predictive models and their performance indicators are shown in table 1 below. We utilized five performance indicators to determine the models’ performances. This includes the coefficient of correlation (R²), Nash–Sutcliffe efficiency, Mean absolute error (MAE), root-mean-square error (RMSE), and Peak water-level error (PWE). The coefficient of Correlation quantifies the strength of the relationship between the model’s variables. A value of 1.0 indicates a perfect relationship, however this cannot be achieved because physical measurements are never perfect. The recorded R² values in this study are within the range of 0.9291 to 0.9905. This signifies excellent relationships between the model variables hence reliability of the estimates.  [43] and [44] have defined performance ratings for NSE. Based on their evaluation scheme, 0.75<NSE≤1 is rated very good, 0.65<NSE≤0.75 is rated good, 0.5<NSE≤0.65 is rated satisfactory, and NSE ≤0.5 is rated unsatisfactory. This study recorded NSE range of 0.8215 to 0.9959. We considered this performance to be very good based on the above rating scheme.

MAE and RMSE are indicators showing the number of errors obtained by a model. Perfect fit for these variables is zero, however, this can never be obtained since every geophysical measurement is exposed to certain level of insecurity.  Based on the rule of thumb, RMSE values between 0.2 and 0.5 shows that the model can predict the data accurately. RMSE range of 0.0657 to 0.4159 resulting from this study and the values of MAE as shown in table1 indicate that the models’ estimates are reliable.

 Peak water level is of significant concern to water resources planners and decision-makers. This is because pre- knowledge of peak flow is very crucial for preparing for emergency response and other impact reduction measures. Peak water error (PWE) is the difference between the observed and predicted peak water level.  PWE determined at the various stations are between the range of 0.01m to 0.75m. This shows that the models were able to accurately predict the highest water levels.

3.2.3 Flow Time Prediction

Result of flow transmission time analysis is presented in table2 below. It captured the observed and predicted time of flow transmission during the flood episodes. Accurate timing of the incoming flow is required to prepare for mitigation measures. We evaluated flow transmission time predictive performance using four evaluation metrics (see table 3). The predictive model (equation 8) was trained using data collected during 2012 and 2018 flood event and validated using data collected during 2022 flood episode. The performance evaluators, which include R2, NSE, MAE and RMSE, showed values ranging from 0.9996 to 0.9999, 0.9996 to 0.9998, 0.0985 to 0.1238 and 0.1935 to 0.2168 respectively. These statistics indicate excellent performance as per the standard we utilized in section 3.2.2 above. The observed total time of flood transmission (T_obTTr) along the basin during 2012, 2018 and 2022 are 244.2 hours, 244 hours, and 244.3 hours respectively and the predicted equivalent (T_pTTr) is 243.2 hours.  More details can be seen in table 2 below

Figure7 below shows the flow velocity map generated by applying Manning’s flow model which was used for further validation. The study revealed that flow velocity along the

channel falls within the range of 9.48*10-²ms^-1 to 1.129ms. High flow velocities were observed in areas of low transmission time and vice versa and these show correlation with the slopes and river depth

4.0 Conclusion

Flood emergency management agencies (both public and private) in this region are often reactive rather than proactive in response to crises. They are usually not concerned with the fundamental issues of mitigations or impact reduction measures. One of the primary reasons for this is lack of effective prediction of events. Flood cannot be prevented but its impacts can be reduced if adequate preparations are made. Although flood risk analysis has been conducted in many cities around the globe, most of them focused on spatial components. This has to do with the determination of areas or locations under different flood risk potential levels. However, none of these studies (especially in Nigeria) incorporated predictive models. As a result, many communities along the Lower Niger Basin in Nigeria are adversely affected during extreme climatic condition. Accurate and reliable flow prediction of river dynamics during storms is essential for flood control, mitigation of flood hazard, evacuation of people from flood hazard areas etc. Considering the high records of flooding events along Lower Niger Basin, determination of the hydrological character of the river was essential. Thus, identification of such hydrological behaviors benefits modelers with an efficient tool to predict the water level and flow transmission time and so as to reduce potential damages of flood events therefore this is the rationale for this study. In this study Morphology of the basin and riverbed was investigated and displayed using different visual format and reliable hydrodynamic models for river stage and flow transmission time prediction have been developed. Least square regression analysis was used to model the relationship between hydrological variables along the basin and most suitable models were selected by applying least square criterion. The models developed were validated, graphically and statistically to verify their predictive capabilities. Evaluation metrics which include R2, NSE, MAE , RMSE and PWE were applied to compare the predicted results with the observed data and significant level of agreement was observed. These models are invaluable for reliable water level and flow transmission time forecasting along the basin without relying totally on observations from gauging stations especially in areas lacking river stage measuring instruments. The results of the analysis have direct application in the formation of policies on disaster management, flood risk assessment and reduction along the Basin in future

References

1. Destaw, K.  and Assayew, N. (2025.) Flood monitoring and damage assessment in agricultural field using Sentinel-1 SAR data in Gumara river catchment, Blue Nile basin, Environmental Systems Research, 14(32) pp1-19   https://doi.org/10.1186/s40068-025-00427-1
2. Olajuyigbe, E. A. (2012). Mapping and analysis of 2012 flood disaster in Edo State using geospatial technique. Journal of environmental sciences. 6 (5), pp 32-44.
3. Emmanuel, U. A.,  Aniekan, E. (2017). River inundation and flood hazard zonation in Edo State using geospatial technique. The International Journal of Engineering and Science. 6, (8) PP48-59
4. Aldardasawi AM, Eren B (2021) Floods and their impact on the environment. Acad Perspective Procedia 4(2):42–49
5. Masoud, M., Amini, M. H., Maghsoudi, S. F., & Alipour, S. (2020). Spatial multicriteria decision analysis for flood susceptibility mapping in Talesh Coastal Plain, Iran. Iran. International Journal of Environmental Research, 14(2), 193–208
6. Kurapati, P.V., Babu, A., Pyla, K.R, Nsr, P., Mandla, V.R.B (2023) Flood mapping and damage assessment using Sentinel – 1 & 2 in Google Earth engine of Port Berge & Mampikony Districts, Sophia Region, Madagascar. Adv Scalable Intell Geospatial Analytics: Challenges Appl 1:288–289 https://doi.org/10.1201/9781003270928-20
7. Sehrish, K, Q. A, Muhammad Z. A, Syed, A. M. A, Ahsan, U/ B., Muhammad, S. B (2021). Socioeconomic determinants of climate change adaptations in the flood-prone rural community of Indus Basin, Pakistan. Environmental Development.V37. Available online at https://doi.org/10.1016/j.envdev.2020.100603
8. Lyu, H., Wang, G., Shen, J., and Lu, L. (2016). Analysis and GIS mapping of flood hazard in Guangzhou china. Water journal  (8)447 pp1-17
9. Al-Quraishi, A. M., & Babel, M. S. (2021). Flood hazard mapping in Iraq using multicriteria decision analysis and machine learning techniques. The Egyptian Journal of Remote Sensing and Space  Science, 24(2), 339–353.
10. Christo, G., Loannis, T., Anthi-Irini, V., and George, P. K.  (2022). Flood risk assessment and flow modeling of the Stalos Stream area. Journal of hydro informatics. 24 (3) pp 677-696
11. Aramburú-Paucar, J. M., Martínez-Capel, F., Puig-Mengual, C. A., Muñoz-Mas, R., Bertagnoli, A., & Tonina, D. (2024). A large flood resets riverine morphology improve connectivity and enhances the habitats of a regulated river. Science of The Total Environment, 919, 170717. https://doi.org/10.1016/j.scitotenv.2024.170717
12. Alodah, A., Seidou O (2019) Assessment of climate change impacts on extreme high and low flows: an improved bottom-up approach. Water 11:1236. https:// doi. org/ 10. 3390/ w1106 1236
13. Emmanuel, U. A., Ojinnaka, O.C., Baywood, C.N .(2015). Flood Hazard analysis and damage assessment of 2012 Flood in Anambra State using GIS and remote sensing approach. American Journal of Geographic Information System, 4(1): pp38-51
14. Benefit O (2024) The Impacts of Flooding on the Depletion and Dispersal of Plants. Greener Journal of Agronomy, Forestry and Horticulture Vol. 8(1), pp. 1-5.  https://gjournals.org/GJAFH
15. Mahé, G., Liénou, G., Olivry, J. C., Orange, D., Ardoin-Bardin, S., & Servat, E. (2013). The impact of land use change on soil water holding capacity and river flow modeling in a tropical humid and semi-arid region. Hydrology and Earth System Sciences, 17(6), 2253-2275
16. Alexis D, Ezechiel K. (2025). Impact of Floods on Biodiversity and Resilience of Affected Local Communities in the Far North Region of Cameroon. HAL Open Science. https://hal.science/hal-05004422v1
17. Nji, T. M., Balgah, R. A., & Vubo, E. Y. (2019). Coping with flood hazards in Cameroon: The role of community-based strategies.Sociology International Journal,3(5)pp372382.  DOI:10.15406/sij.2 019.03.00202
18. Masoud, M., Amini, M. H., Maghsoudi, S. F., & Alipour, S. (2020). Spatial multicriteria decision analysis for flood susceptibility mapping in Talesh Coastal Plain, Iran. Iran. International Journal of Environmental Research, 14(2), 193–208
19. Nwilo, P.C. (2012). Survey practice in the Nigerian contemporary society. Paper presented on the occasion of the opening and dedication of ‘Surveyors House’ of Nigerian Institution of Surveyors (NIS), Enugu State Branch.
20. Kwak, Y. and Kondoh, A. (2008). A Study on the Extraction of Multi-Factor Influencing Floods from RS Image and GIS Data: A Case Study in
21. Ahuchaogu U. E., Ojinnaka O. C., Njoku R. N. and Baywood C. N. (2021) Earth Observation System-Based Impact Assessment of 2012 Flood in Delta State Nigeria. Civil and Environmental Research.12, (9)79-86
22. Ikusemoran, M., Anthony, D. and Maryah, U. M.(2013). GIS based assessment of flood risk and vulnerability of communities in the Benue floodplains, Adamawa State. Nigeria. Journal of Geography and Geology. 5(4) pp148-160
23. Satterthwaite, D., Huq, S., Pelling, M., Reid, H. and Lankao, P. (2007).  Adapting to climate change in urban areas. International Institute for Environment and Development, London.: http://pubs.iied.org/pdfs/10549IIED.pdf.
24. Ejikeme, J.O., Igbokwe, J.I., Ojiako, J.C.,Emengini, E.J., and Aweh, D.S. (2015). Modelling the impact of flooding using geographic information system and remote sensing. International Journal of Technical Research and Applications. 3(4) , PP. 67-72
25. Knuth, R. (2025) GIS-Based Flood Risk Mapping and Disaster Preparedness. J Remote Sens GIS (14) pp 1-2 DOI: 10.35248/2469-4134.24.14.390,
26. Anu, A., Utsav, B., Vishnu, P. P. and Pawan, K. B (2024) Downstream impacts of dam breach using HEC-RAS: a case of Budhigandaki concrete arch dam in central Nepal. Environmental Systems Research Environmental Systems Research 13 (37) 1-16 https://doi.org/10.1186/s40068-024-00358-3
27. Ominabo, W. D. (2022). 2022 Flood: A tragedy foretold.https://www.thisdaylive.com/index.php/2022-flood-a-tragedy-foretold/Accessed 23 November 2022
28. Ejigu, N. F. (2016) . Flood Risk Analysis With Regards To Crop Yield in Upper Awash River Basin, Ethiopia (Case Study of Teji River)
29. Ogbonna, C. G. Eleazu, E. I., Obinka, A. N., and Ukpabi, J. I. (2016). Features of Urban Drainage Systems in Aba, Nigeria. Asian Journal of Science and technology. 07(11), pp.3922 – 3931
30. Wen-Cheng, L and Chuan-En, C (2014) Enhancing the Predicting Accuracy of Water Stage Using physical based Model and Artificial neural network and artificial Neural Network Genetic Algorithm in a River System, Water 6(6) pp1642-1661
31. Haan, C. T., Johnson, H. P., and Brakensiek, D. L. (1982). Hydrologic modeling of small watershed. St Joseph American society of Agricultural Engineers (ASAE). 13(2), pp123-124
32. Salau1, O. B., Salaudeen, A., Gana, B. Zubairu, I.,  Musa, S. I (2021).  Assessment of a New Dam Site for Water Supply Potential in Bauchi Metropolis, Nigeria. Nigerian Journal of Technological Development 18 (4) pp312-321
33. Salaudeen, A.; S. Shahid, T. Ismail, E.-S. Chung, M. Mohsenipour and X.-J. Wang (2016). Prediction of Flow Duration Curve in Ungauged Catchments using Gene Expression Programming. Procedia Eng. 154: 1431-1438
34. Izinyon, O.C., and Igbinoba G. E. (2011). Flood frequency analysis of Ikpoba River catchment at Benin city using log Pearson Type III distribution. Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 2 (1), pp 50-55
35. Okonofua . (2013). Flood frequency analysis of Osse River using Gumbel’s distribution. Journal Civil and Environmental Research. 3(10), pp 55-59
36. Bhatt, V.K. and Tiwari, A.K (2008). Estimation of peak stream flows through channel geometry. Hydrological sciences Journal. 53(2) PP 401-403
37. Anupam, K. S., Sudhakar, S., Utkarsh, V., and Arun, K. S. (2011). Estimating Hydrological Parameters in Ungauged Basins Using Field Measurements and GIS.  National Symposium Empowering Rural India through Space Technology November 9-11, 2011. Available at Bhopalhttps://www.researchgate.net/publication/257866377
38. Suvendu, R. and Biswaranjan, M. (2013). Estimation of peak flood discharge for an Ungauged River: A case study of the Kunur River,
39. West Bengal. Geography Journal. 2013 pp 1-11.  DOI:10.1155/2013/214140
40. ESCAP/UN. (1991). Manual and Guide Line for Comprehensive Flood Loss Prevention and Management. https://repository.unescap.org
41. Murray, R. P. and Larry, J. S. (2008). Theory and Problems of Statistics, 4th Edition, Schum’s Outline Series, McGraw-Hill, USA, pp316 – 357
42. Hakan, T. and Martin J. B (2018), Simulation and Forecasting of stream Flow using machine learning Models Couple with Base Flow Seperation. Journal of Hydrology 564 (2018) pp266-282
43. Pan, T., Wu, S., Dai, E and Liu, Y. (2013) Estimating the daily global daily radiation spatial distribution from diurnal temperature ranges over the Tibetan plateau in China, Applied Energy, 107, 384-393
44. Cho, J., Bosch, D., Vellidis, G. Lawrance, r. and Strickland T (2012). Multi-site evaluation of hydrology component of SWAT on coastal plain of south west Geogia. Hydrological Process, 27(12), 1691-1700