Analytical Approach to Examining Drivers of Residential Land Use Development in Lokoja, Nigeria

Aim: To analyse the statistical relationship between residential land-use development and socio-economic and bio-physical variables.
Study design: Case study.
Place and Duration of Study: Lokoja in Kogi State of Nigeria ,which lies within latitude 7O45’N and 7O51N and longitude 6O41’E and 6O45’E of Greenwich meridian, February 2010 and May 2011
Methodology: The two sets of data source used in the study were questionnaire responses to get information on the socio-economic background of the people, and the satellite images to access information on the physical characteristic of the area. 996 questionnaires were returned out of the 1000 distributed to 10 neighborhoods zones in the town. The Arc View 9.2 was utilized with the help of its geo-processing extension to analyse the land sat image of Lokoja recorded in September 2010. This landsat thematic mapper (ETM+) satellite image of 2010 was used to derive different landuse types, it was rectified terrain corrected and geo-referenced to local UTM zone. The land sat image was processed using ERDAS imaging 8.6 software. Logistic Regression Analysis (LRA) was used analyze the selected factors of landuse determinants and factors of land use change , in this case emphasis is on bio-physical and socio-economic factors . The logit model is used to relate the probabilities and the locational characteristics of a particular land use. This is done by finding the coefficient (ß) of the logit model. A logistic regression procedure is used with actual landuse as dependent variable. The SPSS statistical package is then used to regress land use upon the change factors.
Results: The result shows that proximity to infrastructure is significant at a probability level of 1 %, the significant Chi Square value (55.4 df 14, P < 100) and high correct classification percentage of 78%, which is indicative of the perfect fit of the model in explaining the relationship between independent and dependent variable. The R square of Nagelkerke also indicates 38.2% of the total variance in the dependent variable (residential development) was explained with independent variables. The variable elevation shows a significantly negative relationship which implies that residential land development will decrease as elevation increases i.e. one unit increase in elevation will cause the odds of residential development to decrease by a factor of 0.480. Also the Logit result shows that road condition has a significantly negative relationship with residential development , indicating that the higher the density of roads the less the probability of residential land development in the city, where the increase in one unit road presence causes the odds of residential land use development to decrease by a factor of 0.840, this is however contrary to the initial assumption .The result also shows a significant negative relationship between residential land development and population density, which implies the lower the population density the higher the probability of residential land development, the decrease of one unit population density causes the odds of residential area development to increase by a factor of 0.926. However the result did not indicate any significant relationship between drainage, education, land price, proximity to water and soil type or flood potential.
Conclusion: The findings indicated that elevation variables and nearness to infrastructure has a positive significance in residential land development which is supported by previous assumptions. The result did not indicate significant relationship with the soil type, and most physical variables. The inclusion of socioeconomic variables made a difference in the total variation which is indicative of its strength as a strong predictor of residential land development as compared to biophysical variables. The result could be utilized to generate spatially explicit explanation to study residential land use change.