The world currently experiences rapid urbanization and an increase in the number of megacities, particularly in developing countries. The (often uncontrolled) process of urbanization always results in changes in land use and cover and causes serious problems including environmental pollution, destruction of ecosystems, and waste disposal. Moreover, urbanization and related changes in lifestyle increase the per capita demand for energy, goods and services. Thus, there is a critical need to map urban land cover composition for urban growth modelling and estimating the climatic and hydrological consequences of urban activities.
EnMAP will open up new opportunities to describe and monitor land cover composition in urban areas and along urban-suburban gradients, assisting in the understanding of the dynamics of global urbanization. However, fundamental research is needed to determine the most suitable classification scheme for EnMAP. Given the 30 m ground sampling distance of EnMAP data, urban mapping will often require quantification of the sub-pixel land cover composition. The given spatial resolution and high spectral information content requires new concepts for describing land cover composition. While occurrences of spectrally pure surface materials are rare, urban spectral mixtures contain compositional information that might be characteristic of certain urban structures, such as inner city centres or commercial areas.
Another possible field of methodical development is the combined analysis of hyperspectral data and high-resolution remote sensing data, which enables detailed analysis of urban structures combined with surface material information. The challenging task in this context is to preserve spectral information content while spatially upscaling the tasks. Novel fusion techniques combining future EnMAP data with thermal sensors (e.g., ASTER, HyspIRI) can open up new opportunities for large-scale urban climate studies. Results may reveal the relationship between thermal patterns, urban surface materials, and urban structure, and thus help to understand their influence on the urban climate. Overall, the synergistic use of hyperspectral, thermal, and optical data with advanced data analysis techniques may result in enhanced socioeconomic and environmental indicators to model urban dynamics and their social and environmental consequences.
Accordingly, the main scientific tasks related to urban areas include:
- Map and monitor urbanization and its dynamics on a global scale;
- Implement a comprehensive spectral library to analyse urban land cover based on EnMAP data;
- Develop and improve classification algorithms to map urban land cover (including classes that are spectrally ambiguous in multispectral data) at the spatial resolution of EnMAP;
- Develop algorithms for quantitative analysis of urban land cover composition at the spatial resolution of EnMAP with regard to mixed pixel problems and rare pure endmember availability;
- Investigate new concepts for information extraction based on the compositional information of spectral mixtures;
- Apply the V-I-S concept to monitor impervious surface fractions, e.g. in the context of urban climate studies;
- Support urban climate studies with hyperspectral data; and
- Map the abundance of hazardous materials such as asbestos, e.g. in the context of risk analysis.