Comparative “citiology”: ideas for looking at natural and urban patterns and processes

J Ramón Arrowsmith

With input from Jon Fink

Department of Geological Sciences

Arizona State University

April 23, 2002

Large spatial scale correlations between patterns of “urbanness” and geology

Urban growth in the US will be largest in the Rocky Mountain region (  Five states, NV, AZ, UT, CO and ID, plus Georgia, grew twice as fast as the national average over the 1990-2000 period ( This growth pressure will, at various levels, be influenced by the natural environment. The ecological and urbanization patterns and their connectivity are directly influenced by the geological template which is clearly due to the integrated geological history of a region.  In addition, it strongly influences the surface processes operating over the landscape, thus mediating the interaction between ecological elements with the substrate.  Note that this “ecology” includes humans and our patterns of interaction with the two regions. These relationships can be explored with large and diverse earth science datasets and those that are available for biota and urban/social patterns.  One of the powerful aspects of the application of geospatial information technologies is the bringing together of these high resolution and diverse datasets.  We can test this idea that high quality information about the geology and geophysics of a region can be a predictor for the ecological patterns, pathways, and processes there. 

Lela Prashad (a new Ph.D. student working with me) and I have been working on geospatial data for the Tucson-Phoenix-Flagstaff corridor as part of our small NSF-ITR with UTEP ( and  We have taken the higher quality and more detailed geologic map data for the greater Phoenix area and mapped it out here:  Look especially at:  In the state geologic map, only the crudest distinctions around Phoenix are evident (; whereas those on the local map are much more detailed.  In particular, they have much more detail in the Quaternary/basin units—on which most of the development occurs.

These geological data are good enough that if we now take them and do simple comparisons with measures of “urbanness” such as assessed value, demographics, growth rate/patterns, something about transportation, we might find important correlations.  We have had a few discussions with Laura Mussachio over in Planning and Landscape Architecture and I have concluded that the main areas of potential influence of the geology (and topography) on the urban parameters are these:  1)  basic engineering (construction costs, infrastructure emplacement, service/resource delivery), 2)  scenery (why things look the way they do), and 3)  geospatial data convergence. What I think I mean by the last is if these data are sufficient quality and detail and are easy to get at and are somewhat understandable for Planners and Landscape Architects, they will use them and that will influence decisions over time (this is an area in which Laura is quite interested).

I would like to gather similar high quality geologic and topographic data sets for a number of major southwestern cities (Las Vegas, Phoenix, Albuquerque, San Diego/Tijuana, El Paso/Juarez) and possibly Baltimore (just for the cross-site LTER interest) and put together these basic queries and see what is there.  When one brings together diverse high quality data sets, you can never really tell what you will find and it should be interesting.

A related area that has more of a hazards focus is to learn more about piedmont processes and flood hazards in these regions.  The rate of urban growth especially in the southwest is so high that it has outstripped the hazard (floods mostly) recurrence rate and so many developments have not seen the natural events that will test their planning.  Furthermore, given the nature of some development in which “pioneer” (or “leapfrog”) settlements jump out beyond the fringe and are followed by more steady encroachment onto the open lands, are there unanticipated changes in surface properties of the landscape that have significant downstream effects?  Mapping of surficial geology should be combined with the best available subsurface maps of hydrology and its geologic controls, as these most directly affect development.

Immediate tasks:

High resolution geological data and urban patterns

            Gather largest scale easily available digital geologic and topographic data for above cities and compile into common GIS framework

            Assess urban parameters and compile into GIS framework

            Explore correlations among urban and geologic/geomorphic parameters

            Seek synergies, especially with ecological datasets/models and remotely sensed representations of surficial materials

            Combine with initial ASTER data compilations (visible, near-IR, TIR) for 100 Cities project

Piedmont processes and flood hazards

Develop synergy with Maricopa County Flood Control District to see what their data needs and thoughts on theses subjects are and where technology development and training that we could do at ASU might be transportable to them.  For example, note Sarah Robinson’s recent collaboration with Fuller and Associates’ MCFD supported floodplain delineation project in which she evaluated remote sensing analyses to flood plain delineation and its utility for increases in efficiency and quality.

Gather historic information for southwestern US on major flood events and their effects on engineered and developed structures.  These data include the character of storms and their hydrologic responses as well as the actual damage assessments.

Gather information on strategies employed by local agencies in the various SW US study areas to manage flood hazards

Develop overview and mapping of other urban hazards and the risks they pose to the cities of the SW US.


Landcover map above. How does this measure of urbannes depend on other urban and natural parameter ssuch as the (Phoenix area) geology or topography as shown below.


High resolution depictions of natural processes and urban responses:  implications for real time monitoring of urban/natural systems

We can tap into data streams of information sampled at frequencies of daily or higher and attempt to provide an estimate of the energy expenditure and mass flux per square meter for various portions of the greater Phoenix area.  Given the geographic isolation of the greater Phoenix area, we can take total traffic (including trucking) in and out of the area on the major highways, and couple that with air and rail traffic, solid waste, sewage, water, recycling, shipping and receiving, construction, gravel mining, power and power demand, and other data to depict the urban system in an unprecendented ecological light.  See figure above.


If one takes these diverse data sets, could you track the effect for example of a dust storm or thunderstorm as it crossed the valley?  In other words, take data from EPA/NOAA/MCFD sampled at high spatial and temporal rates and watch the storm move, and combine that in the same geospatial framework with traffic data (see above).  Could we then go even further and look at things like cell phone traffic or even effects such as accidents or some measure of purchases (people were planning to go out to eat, but they decided to stay in because it was so ugly out, or something?). Also, spatial data combinations that include health effects of natural and anthropogenic processes: e.g., correlations of cumulative air quality rankings and epidemiological data for pulmonary or coronary problems.


Data sets

Objective versus derivative and synoptic versus localized data

Datasets in which we might be interested are either objective or derivative and are either sampled broadly and regularly (~synoptic) or locally and thus require interpolation at non sampled points.  All different data sets are interesting and useful in different ways.  Remote sensing data are definitely objective and synoptic, whereas the freeway speeds are localized and also somewhat derivative (likewise the air quality data that are shown above which have also been interpolated).  Combining the different datasets into a common spatial and temporal framework could permit us to look for correlations among various parameters.

Seismic sounds of the region (idea from Jon Fink)

Along with the remote sensing data (sensing electromagnetic properties of the surface), one might sample the vibrations of the city and its surrounding region.  A seismometer measures ground accelerations at a given point over different frequency ranges. Usually multiple seismometers are used to characterize earthquake or explosion sources, the earth structure through which the seismic waves travel, and responses at the receiver itself. Some seismologists have done other studies such as measure the sonic boom from the space shuttle or sub aerial explosions (Oklahoma City Bombing, World Trade Center attacks—“forensic seismology”).  From afar, with a seismometer, one might measure the total vibrations due to a city as a point source.  However, as one got closer and had more seismometers, it might be possible to directly measure and locate the major vibrational energy sources of a city and their temporal and spatial variations.  This is an example of a hitherto unexamined objective data set that could be used to characterize a given city, how that city evolves over time, cyclic behavior on either diurnal or annual time scales, similarities and differences among various cities.


“Local” remote sensing of urban emissions

New chromatographic techniques allow different constituent species in the atmosphere to be distinguished, after careful calibration. One potential application of this capability would be to generate maps (real-time or retrospective) showing the gas emissions from common commercial sources, like fast-food restaurants, gas stations, highways, airports. With suitable visualization algorithms and connectivity, one can imagine the creation of dynamic electronic maps showing the different discharges that affect one’s personal sub-region of the urban environment.


Visualization of comparative value systems

Recent studies of environmental decision-making (e.g., studies by Carl Steinitz of Harvard) have demonstrated how multiple points of view can be summarized spatially. A prime example comes from the large collaborative studies of the San Pedro River watershed in southern Arizona and northern Mexico. Steinitz prepared GIS datasets and maps of the San Pedro corridor showing contours of economic valuation (as determined by real estate developers) and by members of the environmental community. The result was a series of maps that include top priorities for each group, as well as areas that are of interest to only one group. The result is a set of tools that greatly facilitates reaching consensus.

Data mining

Obviously these data are voluminous.  How do we look at them?  We use the tools of data mining: 

An information extraction activity whose goal is to discover hidden facts contained in databases. Using a combination of machine learning, statistical analysis, modeling techniques and database technology, data mining finds patterns and subtle relationships in data and infers rules that allow the prediction of future results. Typical applications include market segmentation, customer profiling, fraud detection, evaluation of retail promotions, and credit risk analysis. 



We have colleagues in the CSE (Computer Science and Engineering Department) who specialize in these methods (e.g., Associate Professor Huan Liu; 


Web accessible real time (or nearly so) data

Just a quick search on “real time Arizona” in GOOGLE comes up with the following sites from where such data might come:

Water and weather Real-Time Data for Selected Stream-Gaging Sites in Arizona Other USGS H2O data Real-time Weather, Rainfall and Flood Information for Maricopa County, AZ (MCFD) NOAA Worldwide Integrated Sensor Hydrometeorology (WISH) NOAA National Severe Storms Laboratory


Traffic ADOT Freeway Management system Arizona Traveller Information


Air Quality and EPA and ADEQ air quality


Local images and other information Central Scanners and web cameras Cameras on Mill Ave Phoenix360 - Phoenix and Arizona News, Live Weather, Web cams and more


Interesting Real time links CEAC Real-Time Greenhouse Environment Monitoring System (Tomatoes Live!) Real-Time Digital Libraries based on Widely Distributed, High Performance Management of Large-Data-Objects (read abstract) Internet accessible machines Real time science data access page


Immediate tasks

            Search for archive data in a few of these and other data streams in the greater Phoenix area and work on a single event to capture it by placing the different datasets in a common temporal and spatial framework and visualize it and test for correlations of events and their potential triggers

            Develop these ideas for monitoring and explore how remote sensing can be incorporated (more applications of Phoenix Sat concept)



2 views of comparative citiology:

1)      large spatial scale correlations between patterns of urbanness and geology

2)      high resolution depictions of natural processes and urban responses:  implications for real time monitoring of urban/natural systems


            If these are fishing expeditions, will we catch anything?  In other words, are the research questions well enough articulated?  Where is the SCIENCE?  But, what is urban science anyway?

            Where could interesting results from these projects lead us?  Can we gather either geoscience or cities related funding from NSF (+/-ITR) or from NASA?

            Obviously the remote sensing work from Phil’s group in particular and the 100 cities project provides very valuable objective and synoptic data; is it better to just push the analysis of those data further and not worry about these ancillary and somewhat more interpreted datasets?  I happen to think not and that they will be complementary.

            How can we work from disciplinary strengths so that studies of these sorts can be done with confidence and we don’t attribute cause and effect in these complicated systems incorrectly?