MODELING OF SEASONAL VARIATION OF GROUNDWATER LEVELS USING ARTIFICIAL NEURAL NETWORK – A CASE STUDY (APO/GUDU) METROPOLIS, ABUJA

ABSTRACT

Artificial Neural Networks (ANNs) have been successfully used for predicting and modeling groundwater levels for a good period. In this paper, I use Artificial Neural Network (ANN) and MultiLinear Regression (MLR) to model groundwater levels in Abuja, Nigeria using Apo-Gudu as a case study. After setting the models with the various parameters that affect groundwater level in Abuja, the developed ANN and MLR should be able to produce near accurate predictions of the groundwater level variations using just the input variables that have been identified as maximum and minimum temperature, rainfall, relative humidity, and wind speed. To accomplish this, the models are first aligned on a preparation dataset to perform as they should as neural networks to forecasts future groundwater levels utilizing past noticed groundwater levels and outside inputs. Reproductions are then delivered on another informational index by iteratively taking care of back the anticipated ground-water levels, alongside genuine outer information. The outcomes show that the created ANN and MLR can precisely imitate groundwater levels and accurately model them. With the results obtained a comparison will be made of which neural network is more accurate. The examination proposes that such tools can be utilized as a suitable option in contrast to physical-based models to recreate the reactions of the seasons under conceivable future situations or to remake extensive stretches of missing perceptions gave past information to the impacting factors is accessible.

CHAPTER ONE

INTRODUCTION

1.1     Background of the Study

 

Water is a universal solvent and natural resource tapped by man, animals, and plants to meet their need on the earth, either in vapor, liquid, or solid form. Water is one of the essential compounds for all forms of plants and animals, thus its pollution is generally considered more important than soil. Studies show that about 80% of communicable diseases are either water-borne or water-related. Water is an indispensable resource for the existence of man, animals, and plants. Demand for groundwater has been on the increase due to rapid growth in population as well as the accelerated pace of industrialization and urbanization in the last few decades especially in developing countries like Nigeria (Abimbola, A. P. and Odukoya, Abiodun M. and Olatunji, 2012). The inadequate supply of pipe-borne water and the paucity of surface water has led to an increase in demand for groundwater in Abuja. People around the world have used groundwater as a source of drinking water and even today more than half the world’s population depends on groundwater for survival.

Groundwater has long been considered as one of the purest forms of water available in nature and meets the overall demand for rural and semi-rural people (Amadi et al., 2012). The increase in groundwater demand for various human activities has placed great importance on water science and management practice worldwide. (UNESCO, 2003) estimates that globally, groundwater provides about 50% of current potable water supplies, 40% of the demand of the self-supplied industry, and 20% of water use in irrigated agriculture. Over much of Africa, groundwater is the most realistic water supply option for meeting water demand.

However, increasing demand and withdrawal, significant changes in land-use patterns, vast industrial and agricultural effluents entering the hydrological cycle as well as seasonal variation, affect the quality and quantity of groundwater (O. M. Idoko, 2010). The determination of groundwater quality for human consumption is important for the wellbeing of the ever-increasing population. Groundwater quality depends, to some extent, on its chemical composition (Al-Ariqi & Ghaleb, 2010; M. Idoko & Oklo, 2012) which may be affected by natural and anthropogenic factors. Changes in groundwater recharge, due to seasonal variation, also affect the concentration of the water parameters. Rapid urbanization, especially in developing countries like Nigeria, has affected the availability and quality of groundwater due to waste and effluent disposal practices, especially in urban areas. Once groundwater is contaminated, its quality cannot be restored by just stopping the pollutants from the source, this is because groundwater contamination may continue years after the waste source is in place (Ramakrishnaiah et al., 2009; Water & For, 2012).

As groundwater has a huge potential to ensure the supply of future demand for water, it is important that human activities on the surface do not negatively affect the precious resource. Agricultural activities, especially abattoir operations, produce a characteristic highly organic waste with relatively high levels of suspended solid, liquid, and fat. The improper disposal of these wastes onto lands and into water bodies leads to the contamination of the environment, one of which is the impairment of water quality (Abdul-Gafar, 2016). There is a high possibility that the effluents from the abattoir will percolate into the ground and pollute the groundwater. This study, therefore, seeks to determine the extent of pollution of the groundwater from the abattoir effluents through the qualitative analysis of groundwater samples taken from different existing wells at various distances from the abattoir. It also evaluates the influence of seasonal variation on the concentrations of the parameters.

The groundwater level is a key parameter for evaluating spatial and temporal changes in groundwater environments (Iwasaki et al., 2013). The groundwater level is governed by various factors. Climate change, as reflected in precipitation and evaporation rates, influences the groundwater level fluctuation. (Z. Chen et al., 2002). (Z. Chen et al., 2004) also found that climate trends have high correlations with groundwater level variations in southern Manitoba (Z. Chen et al., 2004). In plain areas, precipitation infiltration and evapotranspiration in the vertical direction are the major recharge and discharge processes of the water cycle. In the study area, most of the rainfall falls between July and October. Seasonal variation in climate is obvious. So, the focus is on the influence of short-term seasonal variation in climate on groundwater level in this place. The impact of climate variability on groundwater levels can be investigated by analyzing the relationship between climate records and groundwater level fluctuations. Hence, the study aims to investigate the modeling of seasonal variation of groundwater levels using artificial neural networks across Abuja using Apo and Gudu District of Abuja metropolis as a case study.

1.2     Statement of the Problem

 

Abuja is underlain by crystalline basement rocks with rocks that include different textures of granites, coarse to fine, consisting essentially of biotite, feldspars, and quartz. In most cases, the rock has weathered into reddish micaceous sandy clay to clay materials capped by laterites. Generally, only a small amount of water can be obtained in the freshly un-weathered bedrock below the weathered layers. Groundwater is found mainly in the variable weathered/transition zone and in fractures, joints, and cracks of the crystalline basement. Fissure systems in Nigeria rarely extend beyond 50m, as evidenced by the available drilling data. The local water table depth is controlled by textural and compositional changes within the regolith vertical profile and the bedrock topography.

However, the poor management of waste arising from industrialization and urbanization has led to contamination of groundwater hence the need for the present study. It, therefore, becomes imperative to evaluate the quality of groundwater from shallow aquifers in Abuja, to prevent the occurrence of water-borne diseases such as typhoid, cholera, diarrhea, and dysentery as well as cancer-related diseases due to contamination by heavy metals.

 

Figure 1.1: Aerial map of the specified location.

 

1.3     Objectives of the Study

 

The main objective of this study is to examine the high-resolution spatial modeling of seasonal variation of groundwater levels across Abuja using the Apo-Gudu metropolis as a case study.  

The specific objectives include;

1. To perform the sensitivity analysis to determine the most dominant GW parameters.
2. To determine the performance of ANN and MLR model in modeling GWL

• To develop an independent model for the prediction of GWL at APO/GUDU

1. To detect the dynamic of GWL in both wet and dry season
2. To compare the performance linear model (MLR) and Nonlinear model (ANN) for the simulation of GWL

1.4     Research Questions

 

1. What are the dynamics of groundwater level and salinity in the wet and dry seasons?
2. Is there any relationship between the groundwater level, maximum and minimum temperature, rainfall, relative humidity, wind speed in the Apo and Gudu metropolis in Abuja?

• What is the relationship between the geophysical survey on Apo-Gudu metropolis and its results with drilling data for better resistivity interpretations for productive and effective borehole construction?

1.5     Significance of the Study

 

This study will benefit the management of groundwater level companies in Nigeria, especially in the Abuja metropolis to understand the importance of using artificial intelligence such as artificial neural networks in modeling seasonal variation of groundwater levels in Nigeria.

This study will be of immense benefit to other researchers who intend to know more about this study and can also be used by non-researchers to build more on their research work. This study contributes to knowledge and could serve as a guide for another study.

1.6     Scope of the Study

This study is on the modeling of seasonal variation of groundwater levels using an artificial neural network across Abuja using Apo and Gudu metropolis as a case study.

Gudu is an established district in phase 2 of Abuja city. It is also sometimes known as Apo-Gudu. The district occupies a strategic location just outside Abuja’s city center. Gudu is residential but has quite an extensive commercial part to it. Although well-populated it is not as dense or as busy as neighboring areas like Garki. Gudu sits close to Garki in the north, Guzape in the northeast, Apo Dutse in the southeast, Gaduwa in the southwest, and Durumi to the west.

Being a developed neighborhood, the infrastructure is good. The road network is extensive, giving easy access to most parts of the Abuja metropolis. The main roads are Oladipo Diya Street, Ahmadu Bello Way, and the Nnamdi Azikiwe Expressway. An excellent location (away from the hustle and bustle of the inner city) and a high level of development have attracted many people to the neighborhood. Residents include civil servants, politicians, traders, and other middle-class citizens.

The research study will cover the Apo/Gudu metropolises in Abuja.

1.7     Limitations of the study

The demanding schedule of respondents at work made it very difficult to get the respondents to participate in the survey. As a result, retrieving copies of a questionnaire in a timely fashion was very challenging. Also, the researcher is a student and therefore has limited time as well as resources in covering extensive literature available in conducting this research.

Financial constraint: Insufficient fund tends to impede the efficiency of the researcher in sourcing for the relevant materials, literature, or information and in the process of data collection (internet, questionnaire, and interview).

Time constraint: The researcher will simultaneously engage in this study with other academic work. This consequently will cut down on the time devoted to the research work.

Information provided by the researcher may not hold true for all businesses or organizations but is restricted to the selected organization used as a study in this research especially in the locality where this study is being conducted. Finally, the researcher is restricted only to the evidence provided by the participants in the research and therefore cannot determine the reliability and accuracy of the information provided.

1.8     Definition of Terms

 

Artificial Neural Network: An artificial neural network (ANN) is the piece of a figuring framework

intended to mimic the manner in which the human mind investigates and measures data. It is the establishment of man-made reasoning (AI) and tackles issues that would demonstrate outlandish or troublesome by human or factual guidelines.

Modeling:                   Modeling involves making a representation of something.

Seasonal Variation: Seasonal variation is variation in a time series within one year that is repeated

more or less regularly. The seasonal variation may be caused by the temperature, rainfall, public holidays, cycles of seasons, or holidays.

Groundwater:	Groundwater is water that exists underground in saturated zones beneath the land surface. The upper surface of the saturated zone is called the water table. It fills the pores and fractures in underground materials such as sand, gravel, and other rock, much the same way that water fills a sponge.
Metropolis:	An urban area that has a name, defined boundaries, and local government, and

that is larger than a village and generally smaller than a city.

MODELING OF SEASONAL VARIATION OF GROUNDWATER LEVELS USING ARTIFICIAL NEURAL NETWORK – A CASE STUDY (APO/GUDU) METROPOLIS, ABUJA