Last updated: 05.07.04 Nadav_Carmel@Brown.edu

LIMITATIONS

Before any discussion of the results of the project can commence, some limitations of the methodology must be mentioned. First it should be said that the methodology is purely qualitative and suggestive, by no means quantitative. However, since quantification of specific environmental hazards would require comprehensive testing of every parcel and a quantitative map of contamination in Providence has yet to be assembled, the cumulative contamination map produced by this project could still possibly be used with demographic, epidemiological or other geographic information systems overlays in order to provide some guidance in policy-making and cleanup decisions.

Next, every step of the methodology embodied certain assumptions, especially those pertaining to the Sanborn Fire Insurance maps. As noted by the Boston Industrial Archaeology Mapping Project, the Sanborn maps may be incomplete, as they only document areas where there would be a demand for fire insurance, that is, in economically active areas. But while the industrial corridor along the Woonasquatucket River is economically depressed now, the area was highly active through the most of the period under study. Still, as detailed as the maps are, the possibility of incompleteness or error is still present. As well, the maps might not correspond exactly to the 1999 parcel map of Providence, and some changes have certainly occurred in the street layout of the city since 1889.

Further, ascertaining an Occupational Safety and Health Administration Standard Industrial Classification (SIC) code from the information on the Sanborn maps was not always a simple process. In some cases, the type of products manufactured at and the industrial processes used on a site were listed in full detail; in other cases only the name of the property owner was listed, with no further information. Most sites could be located on an informational continuum somewhere between those two poles. In cases closer to the latter, as much research as possible was done in order to determine the correct (or the closest-fitting) SIC code to use, but ultimately, in some cases, a best estimate was made. Also, some SIC codes included a very narrow range of industries or industrial processes, while others were more broad. (The newer, six-digit North American Industry Classification System corrects for this in some ways.)

The most tenuous step in the methodology involved attributing pollutant releases to certain SIC codes. While several different databases exist that catalog contaminant discharges and cleanups, obviously none can claim to be completely comprehensive. The methodology does not posit that all releases definitely occurred at each and every site, or even that certain pollutants are likely to be found at a site. It merely tries to predict the relative likelihood of finding soil contamination at a site now. The shading on the resulting maps indicates the potential presence or absence of contamination, rather than specific contaminants in specific concentrations.

Finally, no data was available later than 1956, which severely limited the analysis. Also, since the project depended on the availability of the Sanborn maps, the lengths of time elapsed from one map to the next varied, potentially making some changes difficult for the methodology to capture (e.g., the slumps in industry preceding both Worlds Wars, the revitalization of industry during the Wars, and the sharp decline again afterwards).