Flood Vulnerability Mapping of Lokoja Metropolis Using Geographical Information System Techniques

Adequate geographic information on flood vulnerability is required to be able to prepare for flood disasters. This study applies Geographical information System Techniques to produce flood vulnerability map of Lokoja metropolis due to its confluence nature and its potential to cause devastating effect to the surrounding communities. This study is aimed at mapping flood vulnerable areas within Lokoja metropolis, for an effective flood disaster management and proper planning. Satellites imageries MODIS of 2011and 2012, SPOT 5 of 2011, location map of Lokoja Metropolis, SRTM DEM, rainfall data, water discharge/gauge data, and GPS coordinates; acquired during field survey were integrated to map areas vulnerable to flooding. In this study Rank Sum method alongside Principal Component Analysis (PCA) is used to calculate the weight of factors that contributed to flooding within Lokoja metropolis. The study is limited to environmental factors such as hydrology, slope, soil type, drainage density, landform and landuse/landcover. Different maps were generated; composite map of the study area, flood extent map, flood plain map, slope map, flow direction map, flow accumulation map, Triangular irregular network, flood vulnerability map and also pie chart showing percentage area impacted, histogram showing the pattern of rainfall within Lokoja metropolis was generated. The approach resulted in four classes of flood vulnerability ranging from not vulnerable, less vulnerable, more vulnerable and most vulnerable areas. The area not vulnerable accounted for 20.25%, less vulnerable area accounted for 34.57%, more vulnerable area accounted for 28.57%, and the most vulnerable area accounted for 16.61%. The study concludes by proffering a number of recommendations aimed at addressing the issue of flooding within Lokoja metropolis. The recommendations includes; construction of levee along areas that are vulnerable to flooding, widening and construction of standard drainages around Lokoja metropolis, dredging of surrounding water bodies to deepen their depth, among others.


Introduction
Remote sensing data from satellite image offers an opportunity for improved observation and added systematic analysis of various factors affecting groundwater. Again GIS is one of the most important tools for integrating and analyzing spatial information and multidisciplinary database of any resource for planning of resource development environmental protection, scientific researchers and investigations [2].
Lokoja, Kogi State is both a highland/mountainous and lowland state. The river Niger in most cases and the flow of seasonal streams used to have undisturbed flow under cool vegetation from the Mount Patti and environ has since not only been exposed but its courses have always been blocked by construction and refuse from different land uses. The drainage systems originate from the centrally situated highlands and make their way down to the peripheral or outlying lowlands, especially during the rainy season (April-October). During rains, the water becomes so much that it leads to excess water occupying and eventually submerging the streets. The state metropolis experiences majorly riverine flooding and sometimes flash floods depending on geology and the drainage pattern of the area [1]. On the other hand, much of the flood disasters in Lokoja metropolises are attributed to rivers that overflow or burst their banks and inundate downstream plain lands. The 2012 flood that assaulted Sarkin-Noma, Gadumo, Kporoka, Jingbe, Galilee, Kpata, Ganaja village and Saudana areas is a typical manifestation of riverine floods. Therefore, owing to its topographic and altitudinal characteristics, flooding, as a natural phenomenon, is not new to Lokoja [1]. They have been occurring at different places and times with varying magnitude. Scientific studies suggest that climate change is likely to cause shifts in the global pattern and intensity of flood events in some regions, thereby increasing the exposure of populations to severe flooding and that the impacts of future changes in climate extremes are expected to fall disproportionately on the poor [3].
Certain sections of society are particularly vulnerable, especially the poor who have to resort to living in areas that are at high risk from natural disasters, such as flooding, landslides, etc. [4]. Flood hazard comprises many aspects which include structural and erosion damage, contamination of food and water, disruption of socio economic activity including transport and communication, as well as loss of life and property [5]. According to IPCC [6], flood hazards are natural phenomena, but damage and losses from floods are the consequence of human action. Flash flooding/urban flooding destroys the produce e.g. crop, rice paddy, fruit tree and vegetables thereby posing the risk of hunger to those engaged in subsistence farming and great loss to those engaged at a commercial scale [7]. Some of the causal factors of flood disasters in Nigeria have been attributed to land inundation from heavy rainfall, Dam failure/release of dams' water, climate change, and blockage of drainages with refuse, construction of buildings across drainages, inadequate drainage networks, and population increase in urban areas [8].
It is generally understood that there is a relationship between urbanization and hydrological characteristics; decreased infiltration, increase in run-off, increase in frequency and flood height. All disasters have spatial component; adequate geographic information on hazards and areas vulnerable to hazards is required to be able to prepare for disasters [28]. Hazard risk vulnerability mapping is considered as one of the most important steps in disaster risk reduction because it identifies areas vulnerable to disaster so as to plan for disaster risk management [28]. Remote sensing offers a synoptic view of the spatial distribution and dynamic of hydrological phenomena such as flood and erosion. They are used to measure and monitor the extent of flooded areas, provide a quantifiable estimate of the land area and infrastructure affected by flooding and erosion [9]. Nigeria has been slow to realize the potential possessed by remote sensing and GIS in flood disaster management [28]

Methodology
In order to carry out this research methodogically, data acquisition is the primary operation as far as digital mapping or GIS is an operation is concerned. Figure 2.1. Shows the step by step procedures carried out to arrive at the desired result. GIS and remote sensing techniques were used in mapping flood vulnerability areas within Lokoja metropolis. Table 2.1. Shows the primary and secondary data used to execute this work.

Data Acquisition
All data used for this research work, comprises of both secondary and primary data sets as shown in Table 2.1. The type of data needs depends upon the user requirements. Geometric data are usually collected in digital form from a variety sources. For this project, the following data were collected and used with the application of GIS and Remote Sensing techniques in the mapping of the flood vulnerable areas within Lokoja metropolis.

Acquisition of Satellite Imagery
The satellite imagery used are SPOT-5 of 2011 and MODIS of 2011 and 2012. SPOT-5 satellite is a high resolution satellite operated by space imaging. The satellite imagery was chosen because of its spectral characteristics. The Moderate Resolution Imagining Spectroradiometer (MODIS) in surface reflectance with 250m resolution of Lokoja of 2011 and 2012 downloaded from NASA website; http//modis.gsfc.nasa.gov. The MODIS images comprises of visible and infrared channels which makes it possible to distinguish between water and land. The images used were for two years; the first was acquired in October, 2011 and the actual river width and the second was acquired in October, 2012 during the peak of the flood menace in Lokoja.

Acquisition of Elevation Data, Meteorological Data, and Hydrological Data
The shuttle Radar Topographic Mission (SRTM) of the study area used, is the digital elevation model (DEM). In addition to the satellite data was the topographic map of Lokoja (sheet 247NE, NW, SE, and SW). The rain fall data covering the period of twenty years (1995-2014) and the hydrological dataset were also obtained. All these data set were gotten from each of the Lokoja branch office. Table 2.2 and Table 2.3 shows the data used respectively.
Also, the soil raster map data of the Nigerian soil map was use to extract soil information of the study area. It is generally accepted that soil types in Nigeria are grossly influenced by the climate and vegetation zone of the country. This is expected because of the degree of available moisture in the soil, is an important factor in soil reactions.
Immediately the data were acquired and prepared they were taken to the GIS environment for further data processing, where the following were done; image pre-processing, filtering, colour composite, georefferencing and digitization, extracting the spatial extent of the flooded areas, terrain processing using DEM, probability rating of flooding in flood vulnerable area and classification of 2011 spot-5 image, training sample, classification and processing of soil, land form processing and classification, processing of drainage density and classification, slope processing and classification, GIS weighted operation and multi-criteria analysis for flood vulnerability assessment and processing of hydrological data for flood frequency.

Journal of Geosciences and Geomatics
The following parameters were considered during data processing to achieve maximum accuracy as shown in Table 2 Drainage density L / A = Where; L is the total length of drainage channel in watershed in km A is the Total area of watershed in km 2 The calculation was done in Arc-map. The data used for this calculation are the length of the main stream and the area of the river respectively. Further insight to flood frequency is provided by the return period analysis using the Weibull equation; Where m is the rank from lowest to the highest (ascending order), P is an estimate of the probability of the value being equal to or less than the ranked value.
For the probability expressed in percentage; P m / n 1x100 = +

Result
The results are presented in form of maps, charts and statistical tables as obtained from the step wise processing of the DEM, MODIS, SPOT images and non-spatial data of the study area. The statistical tables include the Landcover of Lokoja metropolis, wet and dry area in Lokoja metropolis between 2011 and 2012 as obtained from the MODIS image, and flood extent of Lokoja metropolis as obtained from the 2012 MODIS image. Presented below are the various maps, tables, and charts showing the output of the techniques in order in which they were carried out.   From the map it can be seen that areas ranging from 26 to 58 and 58 to 123 are relatively low and fairly flat; its suffice to say that such areas are vulnerable to flooding as run-off from the higher elevations tend to accumulate at the area of lower concentrations and when rivers over flow their banks. Areas ranging from 155 and above are not vulnerable to flooding. Considering the height variation, Lokoja metropolis can be considered to be low except for the surroundings as shown in Figure 3.4. Its then suffice to say that the flood menace in Lokoja metropolis can partly be attributed to poor drainage system. Figure 3.5. the slope map of Lokoja metropolis. The slope map identifies the overall rate of downwards movement of water, in other words, the higher the elevations, the faster the downward flow of water. It can be seen from Figure 3.5 that the low slope areas range from 0 to 4.44 to degrees, the moderate slope areas range from 4.44 to 13.85 degrees and high slope areas ranging from 13.85 to 34.28 degrees. One of the important factors to be considered in determining areas vulnerable to flooding is the steepness of the slope. The calculations of the slope angle of the DEM reveals that areas within and without the stream buffer have varying slope angles. Therefore, it's possible that some areas within the stream buffer zone are well above water level than the other areas and such areas are said to be less vulnerable to flooding than the later. The Figure 3.5 clearly shows the slope steepness of the area reclassified into three categories. The lower the slope angles of a particular area, the closer to water level. Therefore, areas within the slope angle of 0 -4.44 and within the stream buffer zone were classified to be the most vulnerable to flooding which can arise from the event of sudden release of Dams or heavy rainfall.    Figure 3.9, gives a first order initial idea of the areas vulnerable to flooding within Lokoja metropolis. Areas like Crutcher village, Zango, part of Lokongoma, Otokiti, and other parts very close to the above places listed stand as being above flood level. Others are ranked as either "Less Vulnerable", "More Vulnerable", or "Most Vulnerable to flooding respectively.      The Triangular Irregular Network (TIN) of the study area shown in Figure 3.8 further buttressed on the fact that some areas are more vulnerable than the other. From the map, the areas ranging from 35-83.889m are considered to be the most vulnerable areas to flooding, areas with height ranging from 83.889-132.778m are considered to be more vulnerable to flooding, areas ranging from 132.778-181.667m are considered as areas less vulnerable to flooding, while areas with heights ranging from 181.667 and above are considered as areas not vulnerable to flooding. More also, the presence of excess water may be as a result of reduction in capacity of natural water course, this obstruct the natural water flow and deregulate topography thereby leading to local depression.

Discussion of Results
The Pre and during flood mapping as well as the analysis of the study area using the MODIS Satellite imageries (2011 and 2012) as shown in Figure 3.9 serve as reality check. From Figure 3.9, almost all the areas considered to be dry in 2011 were converted to wet areas in 2012. It can be seen in Table 3.2 that the wet area accounted for 4.91% in 2011 which is the size of part of River Niger and Benue increased to 40.47% in 2012. These clearly show that the flood took over the area and therefore reduced the dry area from 95.09% in 2011 to 59.53% in 2012. This result attest to the fact that there was a very serious flood disaster in 2012 which took over areas considered to be dry by the communities living in such zones.
The rainfall data analysis carried out shows that over the years, the rainfall amount in the study area is inconsistence. There has been a variation in the rainfall pattern over the years. The mean monthly rainfall is 104.2cm while the mean annual rainfall is 1224.11cm. (See Table 3.4). Finally, in order to create the flood risk map of Lokoja Metropolis as shown in Figure 3.10 rank sum method was used and each factor is weighted according to the estimated significance for causing flooding. From the GIS weighted analysis which result is shown in Table 2.6, factors that influences the flood were ranged from one to four and was categorized from Not vulnerable to Most vulnerable. A composite flood risk index (FRI) is computed using the overlay analysis and raster calculator in the Arc GIS 9.3 software. Areas considered not to be vulnerable accounted for 20.25% of the total area, area considered to be less vulnerable accounted for 34.57%, area said to be more vulnerable accounted for 28.57% while area considered to be the most vulnerable accounted for 16% of the total area.

Conclusion
Geographical Information System and Remote Sensing techniques were adopted in this work and they are in phases in order to map out areas vulnerable to flood. The research identified the factors that contributed much to the flooding within Lokoja metropolis. Multi criteria analysis for flood vulnerability was carried out in a GIS environment. In the multi criteria analysis, Weighted Linear Combination which involves the ranking of each contributing factor was used. The datasets for this work are those whose spatial references can be determined. These datasets are satellite images (Both high and medium resolution), topographic maps, soil map, hydrological and meteorological data. The study has raised awareness on areas that are vulnerable to flood and their magnitudes. The study fully appraised the role of GIS and Remote Sensing Techniques in decision making process. The study concludes by proffering a number of recommendations aimed at addressing the issue of flooding within Lokoja metropolis.

Summary of Findings
The various phases of analysis carried out above have given an in-depth understanding of the areas that are vulnerable to flood within Lokoja metropolis and the factor that has played the major impact. This insight will pre-empt the recommendations of the study on how flood could be controlled and properly managed. The key findings are itemized below.
i. Lokoja metropolis is linked with Lagdo dam in Cameroon, Jebba Dam, Kainji Dam and Shiroro Dam in Niger State, Nigeria. ii. The water level as revealed by the research is on the high side at the time of flood (See Table 3.4). iii. The research also showed that the flood extent in the study area accounted for the large land cover of about 37.19% and about 230062500m 2 of the total area while the River only accounted for 5.06% and about 31312500m 2 . The land on the other hand including the built up area, farm land, vegetation cover and others accounted for a total of 57.75% and about 357250000m 2 of the total area (See Table 3.1).
iv. The study also revealed that the low slope areas range from 0 to 4.44 degrees, the moderate slope areas range from 4.44 to 13.85 degrees and high slope areas ranging from 13.85 to 34.28 degrees (See Figure 3.5). v. Areas ranging from 35-83.889m are considered to be the most vulnerable areas to flooding, areas with height ranging from 83.889-132.778m are considered to be more vulnerable to flooding, areas ranging from 132.778-181.667m are considered as areas less vulnerable to flooding, while areas with heights ranging from 181.667 and above are considered as areas not vulnerable to flooding (See Figure 3.8). vi. Five contributing factors were examined (soil, landuse/landcover, topography, slope and mean annual rainfall) (MAR)). The impact of each of the above contributing factors as revealed by the Principal Component Analysis (PCA) shows a very close representation of the reality. This is because the flood vulnerability class progression in the PCA corresponded with the areas of observable impact as earlier visited on the field when compared with the PCA values. The contributing percentage for the factors going by the overall result in the PCA are soil (38.65%), landuse/landcover (32.59%), topography (16.02%), slope (9.19%), and MAR (3.55%) (Appedix A). vii. The research also revealed that areas considered not to be vulnerable accounted for 20.25% of the total area, area considered to be less vulnerable accounted for 34.57%, area said to be more vulnerable accounted for 28.57% while area considered to be the most vulnerable accounted for 16% of the total area (See Figure 3.10 and Table 3.3).

Recommendations
Based on the findings of this study, it is important to strictly adhere to the following recommendations: i