Center for Environmental Science Applications (CESA)

Finding solutions to sustainability challenges by linking ideas, people and resources

 The importance of humans as major ecological, geological, hydrological, and climatic agents of change is gaining recognition across a wide range of scientific disciplines. A logical starting point to gain understanding of anthropogenic forcing factors and feedbacks is to investigate human-dominated ecosystems, i.e. cities. We define three research objectives to achieve this goal:

Objective 1: Timely and continuous monitoring of rapidly urbanizing regions using EOS (and other) remotely sensed datasets. Such an approach is necessary to characterize the rates of urban processes, and to provide policy-makers with information that allows them to seize opportunities for direction of urban development trajectories.
 
Objective 2: Mapping and modeling of urban/peri-urban biogeophysical and climatic properties (both current and historic) that are impractical to obtain using field-based studies.
 
Objective 3: Use these mapping and modeling results to advance the understanding of urban development trajectories and urban "futures".
The primary application of remote sensing data in this study is to provide a means for extrapolating detailed measurements at local sites to a regional context for comparison. Multispectral image classification is used with vegetation density, spatial texture, and ancillary data within an integrative expert system to identify urban and peri-urban land cover types. These data can then be used to create regional land use thematic maps that depict different processes. For example, urban versus native materials, permeable versus impermeable surfaces, and transportation systems (asphalt and concrete materials) can be mapped.
 
If you have questions or would like to participate in the 100 Cities Project please contact us.

Selected 100 Cities case studies

 

Urban Land Cover Classification and Textural Classification of Cities
Abstract
Recent studies indicate that human activities are a significant component of ecosystem processes that need to be incorporated into existing ecological models. The Central Arizona-Phoenix Long Term Ecological Research (CAP LTER) site was recently established (along with a sister LTER site in Baltimore, Maryland) to study biogeochemical, geophysical, and social processes operating in a human-dominated ecosystem. Remote sensing is an integral component of CAP LTER research, and drives the development of innovative approaches addressing a diverse array of urban ecological questions. A prime example of this is the development and application of a knowledge-based system incorporating Landsat Thematic Mapper data, vegetation indices, image variance texture, and ancillary datasets to classify land cover in the Phoenix metropolitan region for a 1985-1998 time series. An associated research program that provides the opportunity to apply CAP LTER research results on a global scale is the Advanced Spaceborne Thermal Emission and Reflection Radiometer Urban Environmental Monitoring (ASTER UEM) program. The ASTER instrument onboard the Terra satellite acquires data in the visible to near infrared (15 m/pixel), short-wave infrared (30 m/pixel), and thermal infrared (90 m/pixel). The primary goal of the UEM program is to monitor land cover and land use change over a six-year period for 100 global urban centers. Land cover classification techniques developed for the CAP LTER project are being applied to ASTER UEM data. Initial results using image variance texture analysis suggest that urban centers can be classified into decentralized (ex. Phoenix), centralized (ex. Baltimore), and intermediate (ex. Madrid) textural types.
 
Spatial Structure Variation and Landscape Fragmentation of Urban Centers
Abstract   
The primary application of remote sensing data in this study is to provide a means for extrapolating detailed measurements at local sites to a regional context. Specifically, multi-spectral image classification and texture analysis are used to identify land cover types, such as different densities of vegetation, soils, man-made materials and water. This information is combined within a classification matrix, using an expert system framework, to obtain final pixel classifications. A methodological approach using and analyzing landscape metrics is presented, initially applied to the land cover classifications of a set of pilot study cities. We demonstrate that monitoring and evaluation of landscape diversity in urban - suburban landscapes is feasible on the basis of medium to high resolution satellite data. Thus a 10*10 km grid is used to calculate the metrics on a regular and comparable basis. We place the detailed landscape metric analyses in the global context using a fragmentation metric for comparison between 55 urban centres with varying geographic and climatic settings including North America, South America, Europe, central and eastern Asia, and Australia. Temporal variations in land cover and landscape fragmentation are assessed for a smaller subset of urban centres. These data provide a useful baseline for comparison of human-dominated ecosystem land cover and associated regional landscape fragmentation. Continued collection of ASTER data throughout the duration of the Terra mission will enable further investigation of urban ecosystem trends.
 
Vegetation Density Measurements in the Paris, France Urban Area Using Astronaut Photography and ASTER Data
Abstract  
Astronaut photography of cities collected during Apollo, Skylab, Shuttle, Mir, and International Space Station missions provides a useful dataset for urban analysis that complements the satellite data archive. Recent astronaut photography acquired with digital cameras is now approaching the ground resolutions of commercial satellites such as IKONOS (i.e. less than 6 m/pixel). Astronaut photographs are a relevant source of data for urban analyses, particularly for studies that do not have the resources to purchase commercial-quality data. The CCD image sensors in the cameras currently used for astronaut photography are sensitive to the infrared portion of the electromagnetic spectrum, but infrared signal is filtered out above 700 µm. As such, the digital camera data contain less information on actively synthesizing vegetation than they would with an infrared signal included. We present an analysis of aboveground biomass (i.e. actively photosynthesizing vegetation) derived from astronaut photography of the Paris, France metropolitan area acquired on April 24, 2002 using a Kodak DCS 760C electronic still camera aboard the International Space Station. The accuracy of biomass estimation obtained from the digital camera data is demonstrated by comparison with Advanced Spaceborne Thermal Emission and Reflection Radiometer visible to near infrared data for Paris acquired on April 8, 2002. Correlations of bands between the two instruments allow interpretation of the identified vegetation and soil classes. Collection of astronaut photography over global urban centers is ongoing and planned for future orbital missions, and will be a useful addition to ongoing studies of urban ecosystem change, sustainability, and resilience.
 
Investigation of Human Modifications of Landscape and Climate in the Phoenix Arizona Metropolitan Area Using Master Data
Abstract   
Humans directly alter surficial processes and climate at the local or “neighborhood” scale (typically on the order of hundreds of hectares) where process - response is not well understood. Investigation of surficial processes at this scale requires very high resolution (both spatial and spectral) data over a wide wavelength range. Commercial data from satellite-based sensors such as IKONOS and Quickbird now provide excellent spatial resolution in the visible through near-infrared wavelengths; however data with high spectral and spatial resolution at longer wavelengths, particularly the mid-infrared, are still the province of multispectral to hyperspectral airborne sensors. Superspectral data acquired by the NASA MASTER airborne sensor is being used to investigate social-biogeophysical microclimate interactions in Phoenix, Arizona neighborhoods. This sensor acquires data in 50 bands in the visible through mid-infrared wavelengths, placed to match the bandpasses of the satellite-based MODIS and ASTER instruments. Ground resolution of data acquired over the Phoenix metropolitan region varies from 5 – 12 m/pixel depending on aircraft height. Surface temperature and vegetation density spatial variations between neighborhoods spaced along an income gradient in Phoenix have been mapped using 12 m/pixel data. These data correlate with ethnicity and income level, and demonstrate inequity in the microclimates experienced by Phoenix residents.