An Optimal Population Modeling Approach Using Geographically Weighted Regression Based on High-Resolution Remote Sensing Data: A Case Study in Dhaka City, Bangladesh

Abstract

Traditional choropleth maps, created on the basis of administrative units, often fail to accurately represent population distribution due to the high spatial heterogeneity and the temporal dynamics of the population within the units. Furthermore, updating the data of spatial population statistics is time-consuming and costly, which underlies the relative lack of high-resolution and high-quality population data for implementing or validating population modeling work, in particular in low-and middle-income countries (LMIC). Dasymetric modeling has become an important technique to produce high-resolution gridded population surfaces. In this study, carried out in Dhaka City, Bangladesh, dasymetric mapping was implemented with the assistance of a combination of an object-based image analysis method (for generating ancillary data) and Geographically Weighted Regression (for improving the accuracy of the dasymetric modeling on the basis of building use). Buildings were extracted from WorldView 2 imagery as ancillary data, and a building-based GWR model was selected as the final model to disaggregate population counts from administrative units onto 5 m raster cells. The overall accuracy of the image classification was 77.75%, but the root mean square error (RMSE) of the building-based GWR model for the population disaggregation was significantly less compared to the RMSE values of GWR based land use, Ordinary Least Square based land use and building modeling. Our model has potential to be adapted to other LMIC countries, where high-quality ground-truth population data are lacking. With increasingly available satellite data, the approach developed in this study can facilitate high-resolution population modeling in a complex urban setting, and hence improve the demographic, social, environmental and health research in LMICs.producing high-resolution gridded population surfaces (HGPS), which have large benefits for many [5]. Traditionally, the census has been the main data source for such mapping. Yet, the census is normally conducted every 10 years, which is insufficient to fully capture the fast dynamics of the population changes over time. Using traditional methods to update population data frequently is time-consuming, costly, and even infeasible at a large scale [6].Over the past decade, researchers have developed a wide range of methods for mapping the spatial distribution of the population with the aid of Geographic Information Systems (GIS) technology, which is used to create choropleth maps. As the population is not uniformly distributed over space within the administrative units [7], GIS-based discrete and aggregated choropleth maps cannot represent the spatial distribution of the population accurately [8]. Recently, several projects incorporated remote sensing (RS) derived datasets to model population distribution, such as Gridded Population of the World (GPW) [9], Global Rural-Urban Mapping Project (GRUMP) [10], LandScan Global [11], and WorldPop (Asia, Africa, and South America) [12]. Other studies that have be...

Reconstruction of Residential Land Cover and Spatial Analysis of Population in Bursa Region (Turkey) in the Mid-Nineteenth Century

“…From TAE statistics, we found that GWR models perform better than OLS models. This confirms the findings in the literature [22,[92][93][94][95], as GWR considers spatial non-stationarity from the reconstructed residential land cover/use map. The gridded population distribution maps obtained using the GWR approach can be used to analyse spatio-temporal patterns of population density in the Bursa region and can be used as an input in future studies focusing on exploring population dynamics in Turkey.…”

“…Dasymetric modelling methods use ancillary data along with geographic information systems (GIS) and remote-sensed data to refine the geographical representation of the census variable reported as coarse spatial aggregations. Land cover/use, night lights, geophysical factors, urban/rural areas, building data and roads have been used as ancillary data to disaggregate population to fine-scale maps [21,22]. Although land cover/use data have been recognised as the best option to reflect population density [23], the data that are required from remote sensing images do not exist to model the distribution of historical population.…”

Classification of High-Resolution Remote Sensing Images in the Feilaixia Reservoir Based on a Fully Convolutional Network

“…The WorldView-2 satellite was launched into space on October 6, 2009, and has accumulated sufficient data so far. Its primary advantages include flexible operation, a large capacity, fast return visits, accurate shooting, high-definition image capturing, and the availability of multiple color bands [46]. These advantages provide us with sufficient and reliable high-resolution remote sensing data sources.…”

Delineation of the Urban-Rural Boundary through Data Fusion: Applications to Improve Urban and Rural Environments and Promote Intensive and Healthy Urban Development

“…Therefore, the accurate delineation of urban–rural boundaries helps with distinguishing urban from rural areas for the formulation of reasonable urban and rural development planning methods, to then further alleviate the environmental problems in the process of urban and rural development, and achieve the healthy development of urban and rural areas [ 15 ]. However, due to the significant differences between urban and rural areas in terms of the economic development level and infrastructure, urban–rural boundaries were previously delineated based on economic, population, and other statistical data [ 16 ]. In previous studies, the statistical data of the whole urban and rural areas were generally divided into different levels, with the higher level being the urban area and the lower level being the rural area, then the urban and rural areas were delineated.…”

“…As data that can reflect the differences in urban spatial surface information, urban remote sensing has gradually replaced the use of traditional statistical data, such as administrative indicators, economic, and population statistical data, due to their advantages, including a larger spatial scale, wider collection range, and more connected time scale [ 16 , 18 , 19 ]. As one of the most widely used remote sensing data, NTL data mainly reflect the difference between urban and rural development levels by representing the light brightness of urban buildings and infrastructure at night [ 20 , 21 ].…”