Spatially distributed studies may follow two broad approaches: points that are randomly located and points that are clustered or clumped together. They are discussed within a discipline described as geo-statistics. Geo-statistics is applied in order to interpolate geoid undulation (N) and hence generate through a geoid model orthometric heights from scattered/random or closely/clustered located controls. Control points are coordinated points within a primary/secondary geodetic survey network. Kriging method was adopted to produce topographical maps of the both scattered and closely scenarios. Accuracy computed revealed that standard deviation (σ) of multiquadratic and bicubic models in the study area are respectively 11cm and 14cm in lopsided control study area while over clustered distributed located controls are respectively 12cm and 15cm. Standard deviation with the lowest values among the determined geometric geoid models is at all times preferable scattered/random than closely/clumped scenario. This implies that the multiquadratic models can be applied across the entire study area with high accuracy/reliability irrespective of spatial distribution of the points. Hence, the accuracy of the models are better when the total number of points distributed within the entire study area was used than when a limited number of points within a particular part of the study area was used.