Search

The objectives of this study were to analyse spatiotemporal patterns of reported typhoid cases between 2005 and 2009 and to model the spatial relationship of demographic and socioeconomic factors associated with the occurrence of typhoid in Dhaka. The lowest level census units were used as the scale of analysis. Data in relation to typhoid was collected from 11 major hospitals by scanning individual medical reports, while demographic and socioeconomic variables were encoded in GIS from the population censuses of 2001 and 2011. Global (Moran’s I) and local models (G i *) were used to test how census districts were dispersed or clustered over space. The spatial relationships were modelled through ordinary least square (OLS) and geographically weighted regression (GWR) techniques. Spatial pattern analysis as measured by Moran’s I demonstrated that the distribution of the affected communities with typhoid was spatially autocorrelated across the study period, 2005–2009. Hotspot analysis using local G i * indicated large variation in the locations and sizes of clusters. The demographic model outperformed the socioeconomic and demographic + socioeconomic models in predicting the occurrence of typhoid in the study area. The results of this study are of great aid to identify spatial risk factors, essential to develop the control and prevention measures to specific areas.

Population studies worldwide have suggested that urban population densities generally follow an exponential decay pattern as one travels outwards from the central business district (CBD). Dhaka has experienced phenomenal population growth over the past two decades. This chapter uses econometric and GIS techniques to map and model recent population dynamics using census data for three successive census years (1991, 2001 and 2011) aggregated at the lowest level of census geography. Linear and non-linear regression models were tested to examine urban density form. The study found that a negative exponential function was best suited for the study area since it produced the highest coefficient of determination (R 2). Additionally, temporal trends of the population density gradient for the study area revealed gradual flattening. Further, it was found that the y-axis intercept (an indicator of CBD density) did not drop over time as general theories for cities experiencing economic growth would suggest. The visualisation of population change was conducted through standard deviational ellipses and simple spatial analysis. The results revealed that, with the exception of a few census tracts, the magnitude of population change is (are) still high in the area, and that a suburbanisation trend has set in over the period since the penultimate census.

The objectives of this chapter are to quantify the impact of land use and land cover (LULC) changes on land surface temperature (LST) and to map the changes in the spatial and temporal distribution of LST. A transition matrix was used to determine LULC changes in a multi-temporal manner. LST was retrieved using Landsat TIR data from 1990 to 2011, and statistical analyses were carried out to determine the relationship between LSTs and biophysical parameters. The results showed that the expansion of urban built-up surface over natural land cover such as floodplain and agricultural land has become conspicuous, significantly affecting the spatial and temporal distribution of surface temperature. Annual mean land surface temperature estimation revealed that urban built-up surface consistently has the highest ambient radiant temperature during the study period. A decrease in vegetation cover and subsequent increase in urban land cover were found to be associated with increased LST, suggesting an amplification of the UHI effect with time.

This chapter demonstrates the use of remote sensing and spatially referenced population data to estimate and model urban sprawl, growth and urban structures. Using spatial analytical tools within a GIS, the typology of urban growth and Dhaka’s spatial structure from 2000 to 2011was quantified. The results revealed a 33 % expansion of urban areas during the study period. Analysis of urban growth types showed that the extension growth type being the dominant followed by leapfrogging development. The amount of low-density development is increasing with time, indicating sprawling development. Investigation of changes in the population per unit area of built-up surface indicated that overcrowding and lack of space in the urban core are compelling people to settle in peripheral areas, thereby exerting tremendous pressure on a limited resource base.

The numbers of patients with cholera and Aeromonas-positive diarrhoea show similar bimodal seasonality in Dhaka, Bangladesh. Considering the association between these two pathogens, our previous finding that the number of cholera cases increased following the period of high rainfall led us to investigate the potential role of rainfall on the transmission of Aeromonas-positive diarrhoea. This study quantifies the impact of rainfall on the number of cases of Aeromonas-positive diarrhoea, to gain a deeper insight into the mechanisms of the seasonality of the disease. We examined a time series of the number of hospital visits due to Aeromonas-positive diarrhoea per week in relation to weekly rainfall from 1996 to 2000, using Poisson regression models and adjusting for seasonal and between-year variation, public holidays and temperature. The weekly number of cholera cases increased by 20.7 % (95 % confidence interval, 10.6–31.6) for a 10-mm increase in average rainfall over lags of 0–16 weeks. There was no clear relationship between the number of cases and river level or temperature.

The objective of this chapter is to analyse the spatiotemporal patterns of crime in the Dhaka Metropolitan Area (DMA). Crime data for the period of August 2011 to July 2012 were acquired from Dhaka Metropolitan Police (DMP). Apart from descriptive statistics, spatial pattern analysis was carried out through local indicators of spatial autocorrelation (LISA). Linear regression was conducted using the crime incidence rate as a dependent variable and ten socio-demographic factors as potential explanatory variables. Descriptive analysis showed distinct temporal variation with the pre-monsoon season having the highest crime occurrences. Furthermore, there are three single-centred criminal activity hotspots of crime. Multivariate regression analysis showed that the size of male population and poverty rate were the best predictors of crime incidence in DMA; however, simple regression suggested that total population, number of males and size of unemployment population in the police districts were also suitable predictors. To our knowledge this is the first study that employed spatial techniques to analyse crime in DMA, and the results of this work should be valuable for informed decision-making for the law enforcing agency as well as relevant authorities whose task is to prevent increasing criminal activities in Dhaka.

This chapter looks at the use of a Markov chain–cellular automata method to model and then predict land-use change in Dhaka. Initially land-use/land-cover maps for three separate time periods were derived from satellite images and evaluated against ground truth. The Markov chain method was then used to establish transition probability matrices between land-cover categories for the time periods represented. The use of cellular automata in this work enables neighbourhood interactions to be accounted for. After an initial calibration run, the combined method is then used to predict land use and land cover in 2022 and 2033.

Using census and spatial databases, this chapter demonstrates the assessment of flood vulnerability and risk zoning in the megacity of Dhaka. A place-based approach was adopted to evaluate flood vulnerability at the community level. A method comprising the AHP and the WLC was used to operationalize the conceptual model within a GIS framework. A number of biophysical variables were used to derive the physical vulnerability index (PVI). In addition, nine variables were extracted from diverse sources to derive the social vulnerability index (SVI). To determine the coping capacity of a community, five variables were employed and a coping capacity index (CCI) was developed. Using the PVI, SVI, and CCI, a composite vulnerability index (CVI) was prepared, which revealed that 28 and 14% of the population in the study area were located between high and very high vulnerable zones, respectively. Thirty-six percent were in katcha housing units, which are largely home to marginal people and were in the very high vulnerable zone, revealing the susceptibility of human settlements to floods. Flood hazard data were multiplied with the CVI to delineate the flood risk zone. A total of 18.5% of people lived in high and very high flood risk zones, which was further compounded by the number of females, young, elderly, and illiterate persons. At least 25.8% of housing units were located in these zones. The flood risk map developed in this study could be used as baseline information for the development of pertinent mitigation measures to ameliorate flood loss in the days ahead.

Natural hazard terminologies with existing models of vulnerability, hazard, and risk are presented in this chapter. A conceptual framework has been developed based on hazard literature. The basic premise of the framework is based on Cutter’s place of hazard theory with inputs from recent literature. In the framework, hazard is viewed as a threat that has the potential to overwhelm people, property, and the environment. It is a pre-existing condition that can turn into a catastrophe depending on the influence of exogenous and endogenous factors. Exposure to hazard is treated as given and is an implicit element. The vulnerability element is perceived as the interactive effects of the social and physical aspects of a system (e.g., urban) regarding the causal process of hazards. Contrary to some conceptualizations, the framework views that the total vulnerability of a community depends on physical, social, and existing coping capacity attributes, and therefore, the calculation of the total vulnerability should consider these elements simultaneously. Risk is conceptualized as the product of hazard and vulnerability. To minimize the effects of disasters, it is imperative to take appropriate measures to reduce vulnerability rather than risk.

This chapter illustrates the physical environment together with the built and socioeconomic characteristics of the megacity of Dhaka. An account of historical floods with damage statistics is discussed, revealing that flood loss in the city has increased over time. This is clearly attributed to unplanned and ill-structured development planning. A number of factors are accountable for the increase in flood vulnerability, including extreme population density, intense inequality in resource distribution, and dilapidated drainage systems. Although various adaptation strategies have been suggested for flood management, their success depends on a number of issues such as restricting urban expansion in flood-prone areas.

Flood mapping and monitoring was performed using both Landsat and Radarsat data from 1988 to 2009. Multi-temporal satellite data were digitally classified using a threshold algorithm to determine the spatial and temporal distribution of floods. In addition, flood depths were calculated using the highest water-level data with a DEM. Synthetic depth–damage curves were prepared for three housing categories and for major roads. Flood maps were evaluated using classified satellite images with ground truth data. It was found that 25% of the study area is flooded every year which could rise to more than 40% during abnormal events such as those that occurred in 1988 and 1998. Flood damage estimations revealed that the greatest damage occurred to katcha houses rather than semi-pucca and pucca houses. Variable accuracies were found for satellite-derived flood maps; however, the overall accuracy was highest for radar-based classifications.

Satellite-derived flood maps from 1988 to 2009 were used to construct a flood-affected frequency map of the study area. Similarly, a series of flood depth maps were used to generate a flood damage map for flood depth data. A flood hazard map was then developed by considering both the flood-affected frequency and flood depth data, simultaneously using a 2-D ranking matrix. The final flood hazard map was then combined with population and housing data to calculate the flood exposure for these two elements. The analysis revealed that 23% of the population is located in low- to high hazard zones, while 25% of human settlements are located in different flood hazard zones.