Located in Nigeria’s southwestern zone, Lagos State is regarded as a hotspot in terms of urban expansion. Among all the 36 states in Nigeria, it is the smallest in area, comprising 19 local government areas, including the City of Lagos, the nation’s largest metropolitan area (Figure 1
). The United Nations predicted there would be a megacity in Africa by 2015 [1
], the City of Lagos made that mark by 2020 [2
]. The city currently ranks number seven in the fastest growing cities and urban areas globally, with an average annual growth of 4.4% in population from 2006 to 2020. It continues to grow in population density and urbanization, with a current population of over 21 million and over 6000 residents per square kilometer. A rise in population in urban areas is known to directly affect the demand for housing, which in turn leads to an increase in developed land. Urban development in Lagos State is taking place by land reclamation [3
] through dredging from lagoons and converting coastal wetlands into urbanized communities, which is a direct result of urban sprawl. Urban sprawl means the loss of wetland, forest, and agricultural land to houses, roads, and industries, leading to environmental challenges and changing demographics [4
]. Environmental challenges, such as flooding, are the major natural disaster that plagues Lagos State, which is assumed to be stimulated by urban sprawl. The expansion of developed land is taking place in many flood-prone areas and has also led to a change in the coastline geometry (Figure 1
). Monitoring urban sprawl and identifying the pace of the spread and spatial pattern is a primary concern for urban planners and policymakers, primarily due to insufficient data availability and lack of technical knowledge and tools. Additionally, there are no data from “standard” hydrology and climatology for most developing counties, such as Nigeria. It should be noticed that the accuracy of the presented study cannot be compared with modern techniques based on well-sampled water bodies. Still, in this case, those data are practically inaccessible.
Changes in Land Use and Land Cover (LULC) associated with urban sprawl can be measured and analyzed [5
] using remote sensing and Geographic Information Systems (GIS) [6
] but have not been done yet for the entire Lagos State due to the issues mentioned above. Commonly, LULC analysis aims to identify changes in the same geographical area between the two timeframes considered [14
]. Remotely sensed satellite imageries enable us to look at LULC retrospectively and could be used to monitor urban sprawl over time [15
]. Images from different timeframes can be compared after using satellite image-based land cover classification techniques to illuminate the developed and undeveloped land changes.
Several studies have linked urban expansion in Lagos, Nigeria, to flooding [17
]. The area comprises islands of different water bodies, ranging from lagoons and beaches to creeks, making it naturally susceptible to flooding. Urban flooding in Lagos State is expected, and the socioeconomic impact keeps increasing over the years. Urbanization is a common phenomenon in different parts of the world [20
]. However, intensive unplanned urban growth has negative consequences on many environmental aspects [21
]. For instance, urbanization resulting in a decrease in wetland, vegetation [22
], and soil cover leading to an increase in impermeable surfaces can reduce rainfall infiltration and increase runoff to streams and rivers, and eventually cause flooding [24
]. Vegetation can significantly affect hydrological fluxes due to variations in the physical characteristics of the land surface, soil, and vegetation [22
]. The adverse effects of flooding could be reduced with proper urban planning, starting with the identification of flood-prone areas through flood hazard mapping and assessment of LULC changes causing flooding in those areas [26
Flood hazard mapping requires specific hydrologic measurements [27
] from flood monitoring systems such as river and stream gauges, which are valuable but can be time-consuming, and expensive [29
]. More so, in most developing countries, such hydrologic records are insufficient or absent, and the cost of installing these systems could be limiting [30
]. Consequently, flood hazard studies based on direct measurements may be impossible in most developing countries because there are no historical data available to determine certain flood levels and recurrence intervals for a particular flood event [30
]. Flood hazard evaluation based on satellite data and damage reports could substitute for unavailable quantitative data [31
]. Here, we created a remote sensing and GIS-based flood hazard map using seven criteria combined in a weighted overlay analysis. Each criterion’s weight was estimated using the Multiple-Criteria Decision-Making (MCDM) methods which suggest the different influence each criterion has on the hazard delineation process.
Multiple-Criteria Decision-Making (MCDM) and Geographic Information System (GIS) methods have been considered by several researchers [34
] to be very versatile in terms of providing the techniques and strategies for analyzing complex decision-making problems comprising incomparable criteria. The MCDM methods are broadly categorized into objective and subjective methods [42
], with each category based on the role of the decision-maker in the context of determining the importance of a criterion. Of all the different MCDM methods for determining the weights of each criterion in GIS applications, the Analytic Hierarchy Process (AHP) [42
] is commonly used, while the Shannon Entropy weighting method is less widely used [55
]. A combined AHP and GIS approach was used in Kenya and Greece for urban flood vulnerability and risk mapping [59
]. This study applied a combined (hybrid) AHP and entropy MCDM method for flood hazard mapping in Lagos State, an approach suggested to be efficient for determining criteria weights for GIS-based applications [56
Therefore, this work aims to detect the link between LULC dynamics and flooding in Lagos State over 35 years through a multi-year (1986, 2000, 2016, and 2020) study. A map-matrix-based post-classification method was used to investigate LULC changes to identify the losses or increase in the specified land cover types. Additionally, a flood hazard map was created to evaluate LULC morphology and impacts to the continuous rise of flooding in the State using a hybrid weighting MCDM approach. This study could ultimately bring awareness to the general public, urban planners, and land-use managers on increasing flood hazard areas due to the loss of wetland and the expansion of the developed area, hence, promoting a better practice of land use in Lagos State.
We classified the study area into four major land cover types, including water, wetland, vegetation, and the developed areas, based on satellite images acquired in 1986, 2000, 2016, and 2020. Rapid growth in population results in urban sprawl into wetland and lagoons areas. The creation of impervious surfaces and ultimately more runoff could be the major reason for flooding in Lagos State.
The post-classification change detection analysis resulted in a map showing the transformation of one land cover type to another. The change detection process was estimated for the paired timeframes of 1986–2000, 2000–2016, 2016–2020, and 1986–2020, respectively. Figure 6
visually displays the LULC changes from 1986 to 2020. Our study revealed that of all the land cover types identified in Lagos State, decrement in wetland and increment in developed areas consist of the most land cover changes from 1986 to 2020. Our study further revealed that urban development encroachment into wetland areas is supported by estimating the wetland areas and developed land areas in 1986 and 2020, respectively. With 33% of the study area’s total area, the wetland was the primary land cover in 1986, and it became the least with 10.3% of the total area by 2020. The developed area also increased from 26% to 50% of the total area, almost doubling in the study area over 35 years. The areas of vegetation and water bodies remained approximately the same over the same period.
Seven causative criteria or contributing factors were used for flood hazard mapping. We used the weighted overlay MCDC method in combining these criteria and creating the flood hazard map. The weights of the seven criteria were estimated by a combined (hybrid) AHP and Shannon Entropy methods. We reclassified the resulting flood hazard map using the Jenks natural breaks classification into five categories: very high, high, moderate, low, and very low. Our results have also shown that over the years, urbanization has been happening in areas susceptible to very high to moderately high flood hazard zones, which is evident by the recent increase in flood occurrence in these areas.
We could not perform an accuracy assessment based on ground referencing or ground-truthing due to the lack of field data. Instead, we compared our classified maps to higher resolution images from Google Earth. An assessment of its accuracy was made to evaluate the reliability of the classified images. The outcome of the accuracy assessment of the LULC classification against the Google Earth imagery showed that 93 of 100 points were correctly classified. Evaluation of the accuracy of the post-classification change detection was also carried out by comparing the changed areas to higher resolution Google Earth imagery due to the unavailability of in situ data. Three areas within known LULC changes in Lagos State were used for the comparisons. These comparisons show good agreement. For validation of flooding analysis, the flood hazard map produced from this study was compared to flood inventory in Lagos State. The comparison showed that 92% of the flooding events fell within the very high to moderate flood hazard zones, while 8% was within the low flood hazard zone. In general, the flood inventory points show good agreement with our flood mapping and flood hazard classification. In addition, we also compared the flood hazard map to the DFO reported flood in Lagos State. The large flood events archived by the DFO were within the flood hazard areas of the map produced in this study. The highlights of our study include (1) introduced a map-matrix-based, post-classification LULC change detection method to estimate multi-year land cover changes over a few decades; (2) used a combined (hybrid) Analytical Hierarchy Process (AHP) and Shannon Entropy weighting method to carry out a weighted overlay MCDM flooding mapping; (3) put the LULC changes in the context of flood hazards and identify the possible causes of continuous rise in flooding in Lagos State, Nigeria.
Lagos State is bounded on the south by the Atlantic Ocean, which makes it flood-prone geographically., Our study reveals the extent to which different regions of the State are vulnerable to flooding and provides insights for urban planning, urban regulation, hazard awareness, insurance of property, and mitigation actions to alleviate flood hazards. Lagos is currently a megacity with a population of over 21 million people, and at the present rate of population rise, 2.5 billion people will live in urban areas by 2050. More so, our study revealed that there is a significant loss of wetland, and this has doubled in developed areas during the past few decades. Most developed areas are already in flood hazard zones.
Consequently, more attention should be given to flood mitigation measures for existing and proposed developed areas. Existing developed land cover, such as buildings already situated in flood hazard zones, could be flood-proof to minimize flood damages to properties. New developed areas could make provisions for flooding during the initial design and construction of structures by avoiding floodplains and/or incorporating necessary drainage measures to construct flood-proof facilities.
The assessment of the LULC changes in the context of flooding hazards in Lagos State brings insights into how these changes are worsening the already flood-prone areas. The situation can be mitigated by practicing adequate urban planning and regulation. Inadequate urban planning and lacking regulations have direct consequences of incessant flooding in the State. Additionally, our study emphasized the importance of remote sensing and GIS for LULC assessment and flood studies in developing countries, where there are no data from “standard” hydrology and climatology, and there is an increased rate of urban expansion in high flood hazard zones. Here, we present a low to no cost method to estimate LULC changes and identify flood hazard levels where ground-based data unavailability or security prohibits installing such systems. It should be emphasized that the accuracy of the presented method cannot be compared with modern techniques based on well-sampled water bodies. Still, in this case, those data are practically inaccessible. Our study tackled a complex problem with the help of publicly available data and their detailed analysis.