Modeling Rainfall with Respect to Land Cover and Population in the Niger Delta Area Nigeria for the Period 1990-2040

ABSTRACT


INTRODUCTION
Water flooding in the Niger Delta region is not new. However, the fact that the population in the region increasing, could imply that the number of human lives and property under the risk of flooding could be on the rise. As people settle in the region, some land covered by vegetation is converted to developed land to carter for human settlement and industrial expansion. This leads to soil compaction and an increase in the quantity of surface runoff into streams and rivers during rainfall. Communities that live in the Niger Delta region are highly vulnerable to impacts of floods for about 50% of each year. Land cover under vegetation, especially forest contributes to reducing pollutants in water, protecting the water quality, and controlling the quantity of silt in surface water bodies, etc. (Twumasi and Merem, 2006) The increase of silt in surface water bodies reduces their depth, leading to an increase in flooded areas. The variation between human population and land cover under vegetation, and human population and surface area covered by water temporally in the Niger Delta has not been published, which contribute to the gap in this study. Vegetation land cover, especially forest, plays a major role in protecting surface water from pollution. Hence, its destruction of could result in deterioration of surface water quality. As human population increases, the demand for land to develop residential and business structures increases, leading to a decrease in the vegetation land cover through the conversion of some of it to developed land. The main objectives of this study were to model rainfall with respect to the percentage land cover type and forecast the percentage land cover type in 2040 for river Niger-Delta. This study assesses and models the temporal relationship between land cover and increase human population in the Niger Delta region of Nigeria. Also, similar assessments and modeling can be carried out for other places of the world where human settlements interact with rivers, streams, and lakes, etc. This study proposed policies for the mitigation of the regions' environmental degradation by anthropogenic activities based on the modeling on the results of the study.

LITERATURE REVIEW Lower Niger River Drainage Basin
River Niger is the third longest river in Africa (4 200 km 2 ). Its basin has a total surface of 2.2 million km 2 , and hence, the ninth largest in the world. It is also an important linkage between West and Central Africa and among the nine ABN countries, some of which are among the poorest in the world (Diallo, 2003). The river flows through four climatic zones (humid tropical, dry tropical, semi-arid and arid zones), regions of rainfall spanning from 4 000 mm in the Guinea Gulf to 200 mm in the Sahel, respectively. Along the regions that the river flows through, rainfall is very variable in time and space (Adeaga et al., 2012). Its watershed is affected by a widespread environmental degradation process and a deterioration of the natural resource base. Some of the main environmental threats are unsustainable agricultural and ranching practices, bush fires, deforestation, pollution from different sources; aeolian and water erosion of rangelands; silting of water courses; and the proliferation of aquatic plants (water hyacinth, water lettuce, etc. (Diallo, 2003). The lower Niger Basin system has an area of area is about 629,545 square kilometers, with a discharge contribution of about 117 cubic kilometers per year, which is about 64.3% of the river's total flow ( Andersen, and Golitzen, 2005). The system begins at the en¬try point of the River Niger into Nigeria at approximately 162 km north of Lake Kainji and then extends to the outlet into the Gulf of Guinea through the Niger Delta region at about 75 km downstream of the Ni¬gerian border, the Niger river is joined by the Sokoto River. It then extends upstream with a broad floodplain for approximately 387 km (Laë et al., 2004). The rivers Rima, Kaduna, Gbako, Gurara and Anambra are among the other tributaries of the River Niger in Nige¬ria. The Niger forms a confluence with the Benue River at Lokoja in its lower course, af¬ter which the Benue River remains its major tribu¬tary, as well as significant local precipitation, which strongly increases the flow (Laë et al., 2004). The rainfall received annually by the lower Niger basin has been estimated to be between 1,000 and 4,000 mm with inter-an¬nual rainfall variability ranging from 10% to 20% (Laë et al., 2004). The Niger River has been subjected to several natural and anthropogenic perturbations since the 19th century resulting from the Sahelian drought of the 1970s (Laë et al., 2004). Extensive environmental pollution caused by increased anthropogenic activi¬ties including, untreated industrial and human effluent and waste in the lower Niger basin, especially downstream of the confluence at Lokoja is evidently clear (Adeaga et al., 2017).

Water Pollution
Oil spills in the Niger Delta have occurred frequently and the resultant contamination of the surrounding environment has caused significant tension between the residents of the delta and the international companies operating there. Oil pollution has a negative impact on water bodies and agricultural land when it occurs leading to the destruction of the aquatic ecosystems (Twumasi and Merem, 2006;Merem et al., 2017). Crops on affected lands rarely survive.

Impact of Climate Change
A report by (Bernstein et al., 2008) explained that because of global changes, the change of temperature on a global level is expected to increase between 0.2 to 0.5°C per 10 years. Expansions of seas due to thermal energy and decrease of polar ice through melting is expected. The changes will result in a rise of the sea level by approximately 3 to10 cm per decade during the next century (Bernstein et al., 2008), leading to increases in soil erosion by water at alarming rates. The effects of climate change are threatening the way of life for millions of people along the Niger River. According to a head of the Niger Basin Observatory, decades of drop in rainfall in the 1970s and 80s left water levels at up to 30 percent below normal, but in recent times, the trend has shifted and increasingly heavy rainfall during the wet season is making it difficult for people to live along the river. According to (Amosu et al., 2012), apart from climate change, coastal environment is subject to various anthropogenic impacts, which have been frequently associated with high human population, industrial and agricultural activities.

Impact of Forests to Rivers
Forests play a major role in protecting surface water bodies. Loss of forest cover can affect surface water bodies in multiple ways. According to (Neary et al., 2009), forests are the major source for the highest quality and most sustainable water resources. The forested land cover, which involved multiuse forests including timber production, has been found to be absolutely/positively correlated to good water quality. In Nigeria, after mass deforestation during the settlement and post-settlement periods, forest cover has generally been decreasing progressively (Neary et al., 2009). Forest lands provide economic and environmental resources. Forests provide for, wood products (fuel wood lumber building poles, etc.), plus food for animal and humans (Saleem, 2003). Environmental resource of forests includes, biodiversity, wildlife habitat, climate benefits, water quantity and quality, and aesthetics' component. The ecological importance of forests can be recognized in their beneficial contribution on water catchment areas, where they have a regulatory effect. Forests also protect soils from erosion, hence, protecting canals and dams from siltation. Although it may by argued that biodiversity and other ecological features are equally important, natural forests are the most stable systems that offer hydrologically services on earth. Forestry and related development projects should attempt to capture the hydrologic and erosion control benefits of natural forest systems (Saleem, 2003). Hence, with proper management, forests can maintain or improve Stream, dam, or river water quality.

Sand Mining in Nigeria
Sand was among the top four highest mined solid minerals in Nigeria in 2017 (Abdulazeez, 2021). The figure released of 1.3 million metric tons is likely to have been very far less than the actual quantity of sand mined in the country, since this practice is mainly informal, illegal, unrecorded, and unchecked in many states and local governments (Abdulazeez, 2021). According to) (Lawal, 2011), sand mining is rapidly becoming an ecological problem as demand increases in many states of Nigeria's industry and construction sectors. The practice is carried out both legally and illegally leading to environmental devaluation It occurs both on small and large-scale in major parts of the country. According to studies by (Ezekiel, 2010) and (Isah, 2011), it was estimated that the country's housing deficit was 16 million. Hence, the demand for sand for construction in developing areas such as Ado-Ekiti the capital of Ekiti state in southwest Nigeria was expected to rise (Omole, 1998) A study to investigate the degree of land degradation due to sand and laterite excavations in eleven selected locations in Ondo State Nigeria was carried out by (Ofunim-Omoruyi et al., 2017). This study revealed that there was significant loss of forest in the excavated areas compared to unexcavated (control) locations. The study also showed that the level of reclamation of the excavated areas by the operators was zero. (Aromolaran, 2012), many people who lived on agricultural land in Ogun State, Nigeria supported the good uses of sand however, the negative impacts on their land were more than the benefits.

MATERIAL AND METHODS
This study aimed at analyzing the land coverage by water and built-in, and built-in area versus land cover area under water respectively for the study area in River Niger Delta, in the lower region of the river. The methods that were used to collect and analyze the data for the study are presented here. The study area was bounded by the following coordinates, coordinates, 5026'58.12" N, 50 17' 33 .69" E, 50 26' 58.12 "N,60 43' 45.1" E, C 40 35' 38. 69" N, 60 43'48 .4 "E, 40 35' 38. 69 "N, and 40' 45.60 "E, respectively.

Collection of Land Cover Data
A region of Nigeria bordering the Atlantic Ocean provided the area for this study. All the data for the area was acquired from Google Earth Engine (GEE, https:// earthengine.google.org)). The images for this study were downloaded from Google Earth Engine (GEE, https:// earthengine.google.org) which is a cloud platform that stores satellite imagery and supports geospatial analysis on a global scale. The area of interest was imported into the coding platform (Gorelick et al., 2017) and the remote sensing data acquired by Landsat and Sentinel-2 (MSI) was used for 1992, 2000, 2013 and 2020. Specifically, images from the Landsat Enhanced Thematic Mapper (ETM+) were selected for this analysis. The Landsat images were atmospherically corrected surface reflectance images with a resolution of 30m. Scanlines in Landsat was corrected using the fix Landsat 7 scanline tool in ArcMap. The Sentinel-2 image consisted of 13 UINT16 spectral bands top of atmosphere (TOA) images scaled by 10000. These images were exported into Google drive and imported in the GIS environment for further analysis.

Image Processing
The different scenes obtained from Sentinel 2 were mosaiced in ArcMap to get a single image for the year 2020. The image clipping tool in ArcMap was used to subset the images to the shapefile of the study area for the Southern Niger Delta region in Nigeria. The Supervised classification tool was used in ArcMap to classify the images into four different Land use/cover types, Vegetation, mangroves, developed area and water bodies were the main Land use / cover types classified as these are the predominant land uses in the area.    Figure 5).
Since the study area was in Nigeria, it was selected by clicking on the country icon and then clicking on map labeled Nigeria. Once the map of Nigeria has been clicked the data download process is initiated by clicking on the download button.

Analysis Land Cover Change Analysis
Land cover imagery data was as processed to get extract data for analysis. Land cover data analysis was carried out by regression analysis using Microsoft Excel tool kit. The process involves modeling of a land cover type with respect to time (years). Initially linear models will be developed. These will be accepted if their R square is at least 60%. For linear models with R square less than 60%. Second order polynomial models will be developed.

Analysis of Climate Data
Climate data was analyzed by aggregation (Twumasi et al., 2020). Before modeling, rainfall and temperature trends, tables for blocks of several years, for example 10-year, 4-year, etc. blocks were be made. The formation of the tables involved the following steps. To determine rainfall trend for 1915-2020 the data was divided in blacks of 10 years. The average each 10 year was computed and recorded in a table for rainfall versus time in years. For the modelling process, regression analysis was used. First linear regression was used to fit the data using statistical software (SPSS, Microsoft Excel data analysis toolkit or any other) to lines or curves and the equations and corresponding R squared (coefficient of correlations) determined. R squared is a measure of the strength of association between the variables being modeled. The table for the pixel counts for each land cover type for the study are presented in Table 2.

RESULTS AND DISCUSSION Land Cover
The area for each land cover was calculated using the following formula. A (area of each land cover) = (number of pixels x pixel size) (30m 2 ) (Radwan et al., 2021). The areas for the land cover types on the Niger Delta were computed and presented in Tables 2, To simplify comparisons, the calculated land cover areas in square meters for the different classes of land cover arranged in a single table (Table 3). Modelling for the given land cover categories versus years was carried out and their corresponding predictions for the year 2040 presented, using Microsoft statistical tool kit.

Built-In
Land cover area under built-in was modeled using Microsoft excel spreadsheet by linear curve fitting regression analysis for data displayed in Table 3. The model is presented by Figure 10. The correlational coefficient was found to be about 72.4%. The built-in area is expected to increase to 132000000 square meters in 2040.
To determine whether the relationship between the change in area covered by built-i versus years was statistically significant, the approach by (Namwamba, 2021) was used. Data extracted from the model was included to the original data set of 1990-2020 and the model was used to run a linear regression analysis using Microsoft Excel statistical tool kit. The summary for the analysis is presented in Table 4.  The increase in built-in area is due to the increase of land being developed to expand businesses and residential housing, etc., yearly as the population in the region rises. Vegetation area versus time in years was modeled from the year 2000 since it adopts a linear trend in 2000. Its model and prediction for 2040 are shown on the Figure 11.
In 2040 the area under is expected to be 420000000 square meters. To test whether the relationship between the change in area covered by vegetation and years is statistically significant the field data and predicted data for the years 2010, 2030 and 2040 were used as input data for linear regression using Microsoft Excel and presented in Table 5.   The land cover area under mangrove for  was modeled using Microsoft excel spreadsheet by linear curve fitting. The model was extrapolated to predict the coverage by 2040. The predicted land cover area is 360000000 square meters (Figure 12). The correlational coefficient was found to be about 78% (Figure 12). According to the collected data, between 1990 and 2010 there was a slight increase in land coverage by mangrove forests. However, from 2010 onward, there was a remarkable increase in its land coverage. This could have been a result of the increase of land under commercial crops such as oil palm, raffia palm, etc.
To determine whether the relationship between the changes in area covered with mangrove and the years were statistically significant, linear modeling of the data for 2020 was carried out, and then extrapolated to predict 2050 using Microsoft Excel statistical tool kit. The table for the regression analysis is shown in Table 6. Based on predictions of the models, the land coverage and corresponding percentage by 2040 is as given in Table 7. Each given land cover type was expressed as a percentage of total land using the following formula. The percentage of each land cover type (each year) = (Area of each type)/(total land cover size) x 100 The formula can be rewritten as follows.
is being carried out and C, the land cover type. The percentages are presented in Table 8. There was an increase in percentage coverage by builtin, water while a decrease was realized in land coverage by mangrove and vegetation categories. The increase in percentage coverage by built-in category could be a result development of land to expand businesses and residential area as a response to increase in population. The increase in percentage coverage by water could be a result of increase of impermeable surface caused by soil compaction as the land is developed and changes in rainfall patterns that could be global. (Percentage land cover area by Land cover C) Y = ((Area under land cover C) Y / Total land cover area) x100%.
Where, Y represents the year for which the computation

Modeling Rainfall Versus Land Cover Types
Rainfall data for the year range 1993 to 2021 was used to generate corresponding data for 1990-2020. A correlation analysis for percentage land covers by categories, years and rainfall are shown in Table 9.
Rainfall was found to be strongly correlated with % of land coverage by vegetation, and % land coverage by water, respectively. The corresponding models are presented in Figures 13 and Figure 14, respectively. The model explained only 35.5% of the data. A regression analysis on the data was carried out to determine whether the model was statistically significant. The summary for the analysis is presented in Table 10.    Hence, the surface model for area cover under water versus rainfall is statistically significant.

Modeling Rainfall Versus Land Cover Types
Rainfall data for the year range 1993 to 2021 was used to generate corresponding data for 1990-2020. A correlation analysis for percentage land covers by categories, years, and rainfall for the data is presented in Table 12. Rainfall was found to be strongly correlated with percentage of land coverage by vegetation, and percentage land coverage by water, respectively. Their corresponding models and regression summary tables are presented by Figure 15 and Figure 16, Table   13, and Table 14, respectively. The model explained only 35.5% of the data. Since p<0.05 for the model of the percentage rainfall versus the percentage area under water (Table 14), the model is statistically significant.   Hence, the surface model for area cover under water versus rainfall is statistically significant.

DISCUSSION
The analysis indicated different variations in the images over the years mainly due to oil spill, development, other anthropogenic activities, and water retention in the area. Increased water stagnation over the years has also led to an increase in coverage of land by water, as shown in the images for 2010 and 2020, and Table 3, respectively. While the land cover by water has increased, the Niger River itself and other surface water bodies have witnessed decreases in water depths (as indicated by the thinning blue lines from between 1990-2020). This has mainly been caused by surface water bodies' siltation and debris due to anthropogenic activities. According to the land cover versus years' models, the percentage coverage by built-in and water increased, while a decrease was realized in land area coverage by mangroves and other vegetation categories. The rise in the built-in category percentage coverage may have resulted from the expansion of commercial and residential spaces due to carter for population and businesses As a result of the changes, the soil is compacted and its ability to hold water and allow water to be transmitted through it decreases. Hence, there is an increase in quantity of water on the ground or surface. Also, due to worldwide changes in rainfall patterns, there may be an increase in the proportion of land covered by water since the water drainage properties of soil have deteriorated. A correlation analysis between rainfall and percentage land coverage for the defined categories yielded results suggesting strong correlations between rainfall and only vegetation and water, respectively. This can be attributed to the decrease in the soil's ability to take in water has over years of its structural manipulation through anthropogenic activities on land. These activities lead to an increase of land area covered by impervious material, and compaction of soil. As the surface of water cover increases because of rainfall and decreased ability of the soil to take in water, the area covered by vegetation decreases. The water covers some of the surfaces that had short vegetation and hence, the vegetation coverage decreases.

CONCLUSIONS
In this study, the land cover changes of specified categories were modeled with respect to time in years for the study area in River Niger Delta. Rainfall over the River Niger Delta region was modeled with respect to the land cover categories and presented. The study also modeled the human population in the River Niger Delta region, and precipitation data of the region versus time in years and forecasted the region's precipitation over the period 1990-2040, respectively. The results of this study found that the areas of land covered for, built-in, water and mangrove increased, respectively, whereas the area under vegetation decreased. The increase in Built-in was driven by economic growth together with rural-urban migration, mainly. According to the results produced by the analysis, the area under mangroves increased significantly between, 2000-2020, from about 20% to 32%, respectively, which reflects possible conservation efforts by either the Nigerian government, oil companies, area residents or both.