Synopsis of Research in Shanghai (May 26 - July 25):
During his summer GRI fellowship, Professor Mingzhen Lu will collaborate with Prof. Kangning Huang to investigate the spatial mismatch between China’s rapidly growing urban populations and their built environments—a phenomenon marked by both “ghost towns” and overcrowded urban villages. The project involves three core objectives:
1. Mapping Urban Infrastructure and Population:
They will map building volumes in 100 Chinese cities by processing remote sensing data from sources like Sentinel-1, Sentinel-2, PALSAR, VIIRS, and SRTM. A machine learning model (Random Forest regression) will estimate building heights, validated against real-world data. In parallel, they will use mobile phone signal data to construct probabilistic maps of population distribution via a Voronoi tessellation approach. Overlaying these layers will reveal mismatches between infrastructure and population density.
2. Analyzing Temporal Trends:
By extending the mapping techniques over multiple years, they will create time series analyses that capture changes in building age-structure and track population migration patterns. These dynamic maps will help model trends in material flows, highlighting how urban development imbalances have evolved—exemplified by rising Gini coefficients in building volumes.
3. Developing Predictive Ecological Models:
They will adapt ecological models to urban building demographics, establishing age cohorts and determining survival and maintenance rates. This framework will simulate future urban trajectories, predicting construction demands, demolition waste, and related emissions. These predictions aim to guide policy on waste management and sustainable urban planning.
Ultimately, their work will produce openly accessible maps, novel analytical tools, and multiple academic publications, while enhancing our understanding of urbanization dynamics with broad applications for cities undergoing rapid growth.