Applying Remote Sensing/GIS to Arab Fertility

This research aims to add to our understanding of the Arab fertility transition by investigating the existence of spatial patterns of fertility differentials and change over time in urban and rural settings in two Arab nations: Egypt and Jordan. The research is guided by a conceptual framework that explains the fertility transition as a combination of human capital changes in local contexts that affect the opportunity structure for households and the members of those households (consistent with the supply-demand framework) and the spatial diffusion of ideas and behavior regarding family size (the “horizontal” component of the cultural diffusion perspective). The major thrust of the research is oriented toward the exploration of the spatio-temporal component of fertility change in rural and urban areas in Arab countries, predicated on the more general hypothesis that reproductive behavior is a function of both who you are and where you are. The project aims to extend the work that the researchers have already begun incorporating an explicitly spatial component to the analysis of the Arab fertility transition. This spatial component has two important aspects: (1) measuring the extent to which where you are influences reproductive behavior, net of who you are; and (2) quantifying the environmental context in which reproductive decisions are being made. These objectives are undertaken by applying techniques of remote-sensing, geographic information systems, and local indicators of spatial association and combining them with census data for study sites in Egypt and Jordan.

The satellite imagery offers the ability to generate otherwise unavailable information about the ecological/environmental context of the local areas in which reproductive behavior is occurring. The incorporation of these variables into a GIS with the census data offers a way to statistically analyze the information using emerging techniques of local indicators of spatial association (LISA). These techniques permit the quantitative assessment of spatial clustering of low and high fertility, and they also permit the calculation of spatially filtered regression models which are able to distinguish between variability in the dependent variable (fertility) that is due to the spatial component (where you are) and that which is due to the non-spatial component (who you are). These new data and analytical approaches can then be utilized to answer research questions about spatial and temporal changes in fertility that can point the way toward further, more detailed, research into the underlying explanations for the Arab fertility transition. Although the project is not individually pioneering the use of any of these techniques, no demographers have yet put all of these pieces together in this way.

The project therefore seeks to (1) explore the importance of spatio-temporal patterns of fertility in Arab nations, specifically Egypt and Jordan; (2) examine the relative importance of “who you are” and “where you are” in the spatial distribution of fertility levels, and in the change in fertility over time; (3) continue the development of and demonstrate the demographic applications of remote sensing, geographic information systems and local spatial statistical techniques; (4) apply these techniques to rural villages in Egypt and Jordan and urban neighborhoods in Greater Cairo, Egypt, and Amman, Jordan. Among the research questions to be answered are the following: (1) Are fertility levels within both urban and rural settings in Arab nations distributed in a spatially-dependent pattern? (2) Do changes in fertility levels over time in both rural and urban settings exhibit a spatially dependent pattern? (3) Will areas of high fertility tend to be spatially clustered, and will areas of low fertility tend to be spatially clustered, in urban as well as in rural regions? (4) Can fertility decline over time be modeled as a process of geographic (“horizontal”) diffusion from identifiable nodes to specific surrounding areas? (5) Can remotely sensed images be classified in both urban and rural areas to generate variables that represent potentially relevant aspects of the local context in which human reproductive decisions are made? (6) Is the spatial diffusion of fertility change positively associated with indices of community well-being (local contextual factors as measured through the classification of remotely sensed images) and aggregated indices of human capital (including education and labor force characteristics, as derived from census data)?

This work extends the investigators' long-term interest in the demography of the middle east, as well as being part of a broader program of research linking GIS, remote-sensing, and LISA to demographic analysis. These techniques are still quite new in demographic research and the conduct of and products from this research project will provide models for advancing the role of spatial perspectives and methods in demographic education and research. These models will be available to all potential users through the website that will created as one of the several means by which work will be disseminated. The project aims to demonstrate that an understanding of regional fertility transitions requires an understanding of the way in which fertility levels and their change over time exhibit spatial clustering, and of the way in which spatial clustering is related to specific social ecological environments. The methods employed offer a model of how an increase in local prediction could increase the effectiveness of locally-applied policies that may influence reproductive decisions. At the broader societal level, the substantive results from our work will help to guide intelligent use of always scarce resources in improving reproductive health in developing countries.