The growth of urban populations and a changing climate are two defining environmental challenges of the 21st century, especially for the developing world and for human health. An increase in urbanization poses new challenges for the spread and control of communicable diseases. Several major infectious diseases, including malaria, exhibit significant variation in the size of seasonal outbreaks across years in urban environments. To date, questions regarding inter-annual variation have focused on the temporal variability of environmental factors (such as climate ones) and on the nonlinearity of disease dynamics (laneri et al. 2010, Roy et al. 2013). However, the intersection of environmental forcing and socioeconomic heterogeneity in the context of relevant spatial scales for transmission remains unexplored in the field of infectious disease dynamics, especially for urban landscapes (Reiner et al. 2014).,Heterogeneity in socio-economic and associated demographic conditions can interact with environmental variation in space and time. These factors create differential conditions across cities, which can modulate the effect of climate factors at local spatial scales in large urban environments of the developing world (Santos-Vega et al. 2016). It is my general working hypothesis that the understanding, control, and prediction of the population dynamics of climate-sensitive infectious diseases within cities require consideration of the pronounced spatial heterogeneity of urban environments. In concert, the environmental and socio-economic dimensions could define the relevant spatial scales at which to address malaria transmission dynamics. Thus, understanding and controlling transmission dynamics within cities will necessitate consideration of their pronounced spatial heterogeneity and of the relevant spatial scales at which to address temporal variation in incidence.,Despite an increasing interest in the role of spatial heterogeneity, the population dynamics of vector-transmitted diseases has typically been addressed with temporal surveillance records aggregated at the level of whole towns, cities, or regions, including the development of recent statistical inference methods for confronting process-based models to time series data. Such coarse resolutions have been the norm in part because climate variability is thought to operate at relatively large, regional scales, synchronizing dynamics in space (the Moran effect). These (‘mean-field’) models assume for the most part that host populations are well mixed, so that each individual is equally susceptible. In particular, they do not take into account how spatial variation in environmental, climatic and socio-economic conditions affect vector habitat, biting rates, vector control, contact rates and host susceptibility.,Here, we used urban malaria to understand the respective roles of host density, economic level, and environmental variation in urban malaria prevalence, and to determine the critical spatial scales of these drivers. In chapter 2, I studied the determinants of within-city spatiotemporal heterogeneity in the incidence patterns of malaria by combining statistical analyses and a phenomenological transmission model applied to an extensive spatio-temporal dataset on cases of malaria in the city of Ahmedabad, India. These results showed that climate forcing and socio-economic heterogeneity act synergistically at local scales on the population dynamics of urban malaria in this city. With these models, in chapter 3 I described and explained patterns of spatial variation at the resolution of units (wards), as well as both higher or lower resolutions, and identified the spatial scales of aggregation at which the spatiotemporal variation in reported cases is best captured. Finally, the fourth chapter addresses the role of humidity in influencing the seasonality and inter-annual variability of urban malaria. Our analysis shows a strong and significant effect of humidity in the inter-annual variability of the disease. Simulations of the transmission model for twenty years capture remarkably well the observed epidemic patterns when humidity is incorporated, and also show that forecasting malaria risk based on a dynamical transmission model driven by this climate factor is possible.