Department of Surveying and Geoinformatics
Institute of Technology and Environmental Studies
Nigerian Army School of Military Engineering
Email: [email protected]oo.com
Department of Surveying and Geoinformatics
Faculty of Environmental Sciences
Nnamdi Azikiwe University
Email: [email protected]
The heterogeneity and complexity of census datasets requires appropriate tools such as remote sensing (RS), global positioning system (GPS) and geographic information system (GIS) to handle it. This is in view of challenges posed by global population explosion and its attendant stress on scarce available resources. Since population census is about people and their welfare, any development that seeks to improve the living conditions of the people must therefore take into consideration the rare insights provided by population census. A vital component of census work is the delineation of statistical areas sometimes referred to as enumeration areas (EAs) for the field enumeration, which is the spatial foundation for census datasets. The results of this vital census work are sometimes contested (rejected) on account of massive inaccuracies, distortion and lack of spatial credibility. This is due to poor handling of census operation, resulting from are lack of adequate data, dearth of knowledge of the new technological tools, and crude data processing, among others.
This paper highlights this aspect with particular reference to 2005/06 EAs demarcation (EAD) exercise in Enugu, a state situated on the southeast of Nigeria. This pertinently illustrates the ideal problems and processes over full implementation and use of these (RS, GPS and GIS) tools for planning, understudy, delineation and management of census EAs. It uses satellite image data from IKONOS (1-Meter resolution) and modeling hardware and software and other spatial characteristics (data) to demonstrate how management could use these technologies as decision-support tools. Delineation of EAs in the field involved azimuths (bearings) and distance measurements using conventional surveying principles. GPS technique was used for point positioning, to enable appropriate geo-relation of EAs and the locality maps. Salient aspects of the data planning and processing methods were explored in terms of image (data) compatibility and modelling through appropriate hardware and software.
The study reveals that due to these problems of poor handling of census operation, population estimate results are distorted and not a true reflection of the spatial reality. This also has some effect on policy decisions, resource allocation, investments, development of a viable work force and sustainable economic development. Therefore, this work is recommended to serve as information medium and standard for delineation and management of EAs and planning of future population censuses by relevant stakeholders. This is to add value to decision making and planning processes in Nigeria.
Accurate knowledge of population and its spatial distribution provides enabling platform on which man’s natural and acquired needs could be optimally met and also provide appropriate data in quality and quantity on which an account on how and extent of man’s utilisation of resources within his living domain could be made. Census operations in Nigeria are faced with the problems of massive inaccuracy, lack of spatial credibility, completeness, and undoubtful integrity among others. This has led to the contest of population figures for validation and possible rejection in some quarts. This has direct bearing on the implications of poor handling of census operations. Inaccurate census poses serious threats to economic, social and political stability as well as endangers global environmental health, resource planning, allocation and distribution. In tackling these problems, there is a need for availability and appropriate use of geospatial modelling tools for real world scenarios (census data). Unfortunately, many census stakeholders in Nigeria lack suitable capacity for census datasets acquisition and manipulation. In overcoming these problems, census planning and execution are now exploring various new geo-spatial modelling tools available for quick production of much needed interactive census maps.
A vital component of census work is the delineation of statistical areas sometimes referred to as enumeration areas (EAs) for field enumeration, which is the spatial foundation for census datasets. The results of this vital census work are sometimes contested (rejected) on account of massive inaccuracies, distortion and lack of spatial credibility. The heterogeneity and complexity of census datasets requires appropriate geo-spatial tools such as remote sensing (RS), GPS and GIS to handle it. This is in view of challenges posed by global population explosion and its attendant stress on scarce available resources. Although in all these the methods of census enumeration are accordingly applied, it has been found that this geospatial modelling tool on which population census is assessed is affected by some constraints. These constraints are lack of adequate data, dearth of knowledge of the new technological tools and crude data processing, among others.
This paper discusses these aspects with particular reference to 2005/2006 EAs demarcation (EAD) exercise in Enugu, a state situated in the southeast of Nigeria. This pertinently illustrates the ideal problems and processes over the implementation and use of these tools for planning, understudy, delineation and management of census EAs. This is to add value to decision making and planning processes in Nigeria. The study reveals that due to these problems of poor handling of census operation, population estimate results are distorted and not a true reflection of the spatial reality. This also has some effect on policy decisions, resource allocation, investments, development of a viable work force and sustainable economic development.
Enugu State is one of the constituent thirty-six states of the Federal Republic of Nigeria, located within latitudes 5º 53’ to 7º 05’N of the equator and longitude 6º 46’ to 8º 00’E of the Greenwich meridian. It occupies an area of 7,627.20sq km, and is bordered in the north, east, south and west by Benue and Kogi, Ebonyi, Abia and Imo, and Anambra states respectively (Figure 1). Enugu State formed the hub of the former Eastern Region of Nigeria. It came into existence with the break-up of old Eastern State of Nigeria during the states creation exercise of 1967 when the South-Eastern State (made up of Cross-Rivers and Akwalbom) and Rivers States were created from the Eastern State (Region) and what remained of the region was renamed East-Central State. Again in 1976, Owerri Division was carved out of East-Central State to form old Imo State and what remained of the state was renamed Anambra State. With further creation of further states in 1991, Awka and Onitisha Divisions were carved out and named what is today known as Anambra State and the remainder as Enugu State. In 1996, Abakiliki Division was merged with Afikpo Division of Abia State to form Ebonyi State and the remainder (Enugu and Nsukka Divisions) formed the present day Enugu state, with 17 local government areas (LGAs). The state is characterised by natural scenery, sun-dappled landscape and contrast of hills, escarpments, valleys and waterfalls with a moderate tropical continental climate type. The ethnicity is Igbo with a population of about 3.3 million (2006 estimate). Enugu state has been undergoing rapid development in terms of expanding road networks, construction of new estates and complexes. Thus, creating environmental and planning consequences that implies critically on census figures with spatial credibility. Given the scenic landscape and warm climate, other sectorial growth is inevitable and should be guided through well informed research.
Fig. 1: Map of Nigeria Showing Enugu State
Reasons for rejection of census results
The study unraveled certain criterion for rejection of census results to include:
- Lack of spatial credibility: Sometimes NPC are not able to relate the census figures with what is on the ground (no spatial credibility) and as a result, could not substantiate their results (figures) in a competent law court in Nigeria.
- Crude data processing technique: Non-inclusion of state-of–the-art technology in mapping and demographic analysis, due to lack of manpower and or corruption. This tends to aid the cooking-up of census figures in some cases.
- Socio-political tendencies: The ethnic, political, religious ranglings and acrimonious tendencies aimed at scuffling figures which may have revenue sharing undertones.
- Disproportional houses: The existing threshold across the EA in a locality is a function of area concerned (Enumerators Manual, 2005). Therefore, a population census could be rejected or questioned based on disproportional houses to population declared.
- Correlation: Building information as declared in any EA must be correlated with “unbiased satellite information.”
- Attribute verification: Other attribute information could still be verified with the Federal Office of Statistics – an office charged with collection of social statistics for the nation.
These criteria were preceded by the resulting events of population census which began in Nigeria in 1866 when enumeration was restricted to Lagos area (Umoh, 2001). The censuses of 1866, 1871, and 1896 were restricted to Lagos Island and parts of Lagos mainland. The censuses of 1901, 1911 and 1921 covered, in addition to Lagos, a few more urban towns in the colony. Earlier censuses were not complete counts of population, but estimates based on annual returns of tax-payers (Oboli et al, 1971). The 1952/53 census was elaborate enough to cover Enugu and its environs, but lacked simultaneity, which had multiple counting. The first post independence census in 1962 was outrightly cancelled, in spite of huge resources invested and another conducted in 1963. The most officially accepted are the 1963 and 1991 exercises, with population figures of 55.7 and 88.9 millions respectively (Enumerator’s Manual, 2006). There was census in 1973 which was equally unacceptable on account of massive inaccuracies and lack of spatial credibility. This was attributed to attempts by each section (Northern and Southern states) to out-do the other in terms of comparison of relative sizes (Umoh, 2001). The result of the 2006 population estimate (120 million) was recently gazetted by the government of Nigeria (FRNOG, 2007) despite the contentious issues raised in some states.
An accurate knowledge of the population and its attributes (age, occupation, education, etc) including distribution and spatial content augurs well for the standard of development planning in any particular city. But how reliable are population figures? For the last 40 years or more, national census operation has come under severe criticism for failure of the exercise to produce the expected results. At the height of these criticisms, post enumeration surveys (PES) often suffer setback and are therefore rejected as well (EAD Manual, 2005). Against this background, it does not mean that Nigeria as a nation with her abundant human and material resources cannot organise and execute successfully a census that will produce acceptable results that is beyond any reasonable doubt. What is at stake is for the country to produce a locally and internationally accepted result through the use of geospatial modelling tools and inputs. Lack of acceptable, reliable, current and accurate results in various national census exercises is partly due to non-inclusion of substantial levels of the state-of-the-art techniques in mapping and spatial modelling.
The conventional (traditional) method of handling map and map-data (pure analogue) which has been the case in the past has not yielded the desired output, because it is in most cases not integratable in terms of linking mechanisms. Analogue maps which should be the bases of EA and SA demarcation are not easily updated (revised). This was one of the biggest obstacles the 2005/06 EAD encountered. Recently, population studies have taken advantage of new technological tools (RS, GPS and GIS) in planning and execution of census data processes for an acceptable and reliable census results needed for physical planning and resource sharing. These technological capacities have not been maximised in various national census exercises in Nigeria.
Materials and methods
The materials used for this study were mainly the high-resolution image data (2005 IKONOS-1 meter in multi-spectral band) covering the demonstration area, analogue topo maps of 1:50,000-1:25,000, cadastral maps of scales 1:2000-1:3000 and EAD forms and questionnaires were used to generate the desired products (tables and maps) in an interactive GIS environment. These questionnaires were sometimes complemented by oral questions and explanations were necessary, and the relevant data appropriately analysed and interpreted using relevant data hardware and software such as a handheld (Garmin GPSMAP 76c) GPS receiver, A420 Canon power shot digital camera, HP Pavilion dv6000 (HDD 250GB), AutoCAD MAP 2000i, ArcGIS 9.0 environment operating in Window XP.
Data acquisition and processing
This study is based on field-based surveys conducted between February 2005 and August 2007. The planning and execution of the study was achieved using relevant data and software and hardware outlined above. Maps are produced by the combined efforts of many professionals using a variety of technologies (Lo and Yeung, 2002). Typically the mapping process involves the following phases of work: planning, data acquisition, production and product delivery. Therefore, the field data were captured employing basic survey principles and techniques while using a handheld (Garmin GPSMAP 76c) GPS receiver as a tool for point positioning. The GPS points are necessary for georeferencing the delineated polygon in the AutoCAD Map 2000i window. The inbuilt coordinate conversion enabling module of the handheld (Garmin GPSMAP 76c) GPS receiver resolved in-suite the challenges of coordinate conversion in the field. An A420 Canon power shot digital camera was used to picture the area, for higher visual appreciation and spatial analysis.
Secondary data such as the satellite image (Ikonos 1-meter resolution in multispectral band) covering the delineate area and lines (graphic) maps of the demonstration site in required digital format were employed. A newly digitised map typically requires editing and geometric transformation. The editing removes digitising errors which may relate to the location of spatial data such as missing polygons and distorted lines, or the topology such as dangling areas and unclosed polygons (Chang, 2007). The preprocessed raster (Ikonos) image on WGS 84 projection datum was converted to local (Mina) datum. The study area was framed and extracted using the input criteria of centre coordinates (latitude and longitude). For the purpose of achieving desirable (higher) image data quality to be used for geo-registration and feature digitising, the framed image data was inserted into ArcGIS 9.0 environment for reprocessing, using the image reprocessing module of this enabling software. Consequently, a three band (RGB) recombination was done in conjunction with other image reprocessing techniques such as contrast stretching, histogram equalisation, edge enhancement, etc. The EAD forms and questionnaires were merged with some modifications as desired to create the attribute database (tables) of the delineated area. House listing and enumerations were carried out using appropriate EAD forms and questionnaires, which were filled in-suite at the site. The data (raster, vector and attribute) were processed and analysed by creating objects and linking the resulting shape-sheet file (DXF) to DBMS. These were implemented through ODBC or Active-X technology (Loney et al, 2004).
Database development and modelling
Based on relational data model, an EA attribute database was modelled and developed for the purpose of interactiveness. A relational database is efficient and flexible for data search, data retrieval and creation of tabular reports (Chang, 2007). Each table in the database was prepared, maintained and edited separately from other tables. The tables remained separated until linked up by query or analysis. The respective forms (EAD 1A through to locality identification and classification questionnaire) were used for the basis of table creation and database modelling. In addition, some forms were created digitally using spreadsheets, the database management system capability in which platform the databases were modelled internally and linked for data manipulation and analysis. In the process, five tables were created and linked to ArcView GIS shape file using geo-relational technology. The modelled tables were created in consideration to linking mechanisms since the tables were properly normalised (Ndukwe, 2001). Unique fields that serve as keys (primary and foreign) were inserted accordingly, although there are few entries in which tables that make up the database and which ordinarily do not necessitate indexing. The entries in the table were indexed. This was because the system, if fully deployed, is supposed to handle census data, which has a complex and large information. Before the databases were created, the geographic features and their classes (shapes), their entities and their attributes and field data type listing (data declarations) were specified accordingly.
Results and data analysis
AutoCAD 2000i mapping tool was used for drafting the polygons (delineated area) and the map product in vector form was imported into the ArcView environment for cartographic annotations and editing. Both manual and automatic text annotation methods were used to process graphic components. This generated two major data (map) products, namely the EAs and SAs map of the demon sites (figure 2(a -b)). The map is useful to businesses looking for prospective markets; developers in search of home or office, to planners concerned with population trends and anyone else with interest in the numbers, distribution, income, life styles, etc of the delineated area.
Figure 2(a): Generated EAs Map of the Delineated Area Figure 2(b): Another Generated EAs Map of the study area
However, until the generated map layout (figure 2) above, the map product of the delineated area was overlaid on the raster data (Ikonos) using the georelational mapping tools in ArcGIS 9 (figure3). This ensures accurate geo-location and positioning of points for more accurate estimation of population and housing census with spatial credibility. The composite map of the EA (demo site) with the street network, buildings, attribute table, EAs and SAs, in single software, makes spatial relationship of points in both raster and vector overlays a convenient task (figure 4).
Figure3: Overlaid Map product on the raster data (Ikonos) using the geo-relational mapping tools in ArcGIS 9
Figure 4: EA Map with the attribute table in ArcGIS 9.0
The required analyses were performed on the dataset to determine the age, type or purpose of buildings and other facilities within the EAs. Spatial queries were carried out based on user specified needs. The basic queries were either single or multi criterion query, performed using the query-selection command in the ArcGIS environment. These resulted into other new themes (in form of maps, structured (attribute) database, figures and tables). Samples of such themes or results are shown below (figures 5-7). Figure 5 shows a single criterion query on the estimated population in an EA with population more than 3003. This is indicative of the result of digital map interrogation through the dbase with the help of the prompts (query selection). The theme was queried to know the number of buildings that were of mixed use. This created a new theme which showed that more buildings (39) were found to be of mixed purposes with over 20 household in each of the buildings (figure 6). The land areas of the EAs were equally shown (figure 7). These visual and tabular (themes) outputs will be used in decision process for shared spatial perception of the area.
Figure 5: Enumeration Area with population estimate of over 3003
Figure 6: Mixed Purpose Buildings with household estimate more than 20
Figure 7: Enumeration Area with Land area less than 5289.636
Conclusion and recommendations
The main key contributions derived from this paper are that:
- The integrated use of high-resolution remote sensing data, GPS, and GIS tools to delineate enumeration area (EA) and supervisor area (SA) boundaries for the organisation of a population and housing census were practicable;
- The capabilities and advantages of using GIS software to store and manage geographic data relevant to census operations (RS data, EA and SA boundaries, streets, building boundaries) together with attribute data on buildings and households collected during field activities;
- The fact that the use of RS, GPS and GIS improve the overall quality of a census operation and the quality of census results;
- The assertion that results of a population and housing census conducted in Nigeria, have to be rejected or contested for many reasons, including that regarding the inaccurate results at regional and small area level which could have been limited if RS and GIS would have been used for the delineation of EAs.
Therefore, the technique of RS, GPS and GIS for EAD and management was found effective in detection, monitoring, storing, retrieving, querying and analysing census spatial patterns. This work uses a query language which combines the power of both SQL and graphic data in object relational model. It provides a visual environment which allows users to display, query and browse their data. It provides a high level data abstraction and combines rules and procedures into data model that makes census data more manageable. It’s thus recommended for implementation, to serve as information medium and standard for creation and management of EAs and planning of future population censuses by relevant stakeholders.
- Chang. K. (2007): Introduction to Geographic Information Systems, Fouth Edition. McGraw-Hill Publications, a business unit of the McGraw-Hill Co. NewYork, USA, ISBN 0-07252811-7. Pp 1-18.
- Enumerator’s Manual (2006): Federal Republic of Nigeria, Enumeration Manual for 2006 Population and Housing Census 2006. National Population Commission, Abuja, Nigeria. 122pages.
- Federal Republic of Nigeria Official Gazette (2007): “Federal Republic of Nigeria Official Gazette”. No. 24, Vol. 91, Supplementary Copy, Printed and Published – the Federal Government Printer, Lagos, Nigeria. 15th May 2007
- Igbokwe, J.I; Akinyede, J; Dang, B; Ono, M; Anika, O.L; Nnodu, V. (2007): “Mapping the Spatial Distribution of Gully Erosion in Southeastern Nigeria with Satellite Imageries”. (https://www.fig.net/pub/costerica/paper/ts20_03_igbokwe_2398.pdf).
- Jensen, J.R. (2007): Remote Sensing of the Environment: An Earth Resource Perspective, 2nd Edition, Prentice Hall. htt://en.wikipedia.org/wiki/Remote_Sensing
- Kroenke, D.M. (1983): Database Processing: Fundamentals, Design, Implementation. Second Edition, Sceince Research Associates Inc. a Subsidiary of IBM. USA, ISBN 0-574-21320-1 pp 436-486.
- Loney, K. (2004): Oracle Database 10g: The Complete Reference. Tata McGraw-Hill Publishing Company Ltd, New Delhi. ISBN 0-07059425-2.
- Lo, C.P. and Yeung Albert, K.W. (2002): “Concepts and Techniques of Geographic Information Systems”, Original US ed, Published by Pretice-Hall, Inc., NewJersey, USA, ISBN-81-203-2230-4. Pp 19-62.
- Ndukwe, K.N (2001): Digital Technology in Surveying and Mapping: Principles, Applications and Legislative Issues. Rhyce Kerx, Publications, Enugu, Nigeria.ISBN 978-35346-9-0, pp250.
- Oboli, H.O.N. and Harrison-Church, R,J. (1971): An Outline Geography of West Africa. Seventh Edition Revised. Published by African University Press, in association with Harrap London. Page, 148.
- Umoh, B.D. (2001): Population Studies For Nigeria: A New Perspective. Published 2001 by Institute for Developmental Studies, University of Nigeria, Enugu Campus, Enugu, Nigeria