Zoning Before Zoning: Land Use and Density in Mid-Nineteenth-Century New York City

Abstract

Not until the beginning of the twentieth century did U.S. city governments turn to comprehensive zoning to gain control of their land use and built environment. Nineteenth-century cities had comparatively unregulated land-use systems, where proprietors and builders found minimal restrictions to their choices to develop urban land. This article exploits newly digitized geographic information systems (GIS) data, at the level of building footprints, made available by the New York Public Library, to study the land-use geography of mid-nineteenth-century Manhattan, the Western world's then third largest city. We ask: What was the spatial order of the nineteenth-century city? Beyond the case, what can we learn about land use in a political economy where market forces operated with much greater freedom? Addressing these issues, we introduce a variety of advanced GIS methods to the original data set. Specifically, we examine the separation and mixing of the three basic land-use types of commerce, industry, and residence, by the spatial units of both blocks and streets. In addition, we measure at a new level of precision the enormous variations in residential density and crowding that defined the growing sociospatial inequalities of nineteenth-century cities. Documenting systematically and in detail these spatial patterns, including their socially undesirable outcomes, helps understand how nineteenth-century cities developed and the conditions they produced to warrant increasing land-use controls, from building codes to government-mandated zoning.

美国的市政府, 直到二十世纪初才开始以综合性的分区制度来管制土地使用及建成环境。十九世纪的城市, 相对而言则具有不受规范的土地使用系统, 其中产权所有者和建商在选择发展城市土地方面拥有极小的限制。本文利用纽约公共图书馆所提供的建蔽率层级的最新数码化地理信息系统 (GIS) 数据, 研究曼哈顿这个在十九世纪中叶身为西方世界第三大城市的土地使用地理。我们质问: 十九世纪的城市空间秩序为何? 在这个案例之外, 我们能从市场力量以更大的自由度运作的政治经济中, 获悉什麽样的土地使用? 为了应对这些议题, 我们将各种先进的 GIS 方法引进原本的数据集。我们特别检视商业、工业与居住三种基本的土地使用类别在街廓与街道上的隔离及混合。此外, 我们在新的精确层级上, 测量居住密度及拥挤的巨大差异, 而这些差异则定义了十九世纪城市加剧的社会空间不均。以系统化和细緻的方式记录这些空间模式, 包含它们对社会造成的不良后果, 有助于理解十九世纪的城市如何发展及其造成的境况, 导致批准了从建筑法规到政府强制的分区制度等逐渐增加的土地使用管制。

Solo hasta el comienzo del siglo XX optaron los gobiernos de las ciudades americanas por la zonificación completa para asumir el control de los usos del suelo y del entorno construido. Las ciudades del siglo XIX tuvieron sistemas de uso del suelo comparativamente no regulados, donde los propietarios y constructores enfrentaban mínimas restricciones a sus selecciones para desarrollar el espacio urbano. Este artículo aprovecha datos de los recientemente digitalizados sistemas de información geográfica (SIG), al nivel de las huellas de los edificios, suministrados por la Biblioteca Pública de Nueva York, para estudiar la geografía del uso del suelo del Manhattan de mediados del siglo XIX, tercera ciudad más grande del Mundo Occidental. Nos preguntamos: ¿cuál era el orden espacial de la ciudad del siglo XIX? Más allá del caso, ¿qué podemos aprender nosotros acerca del uso del suelo en una economía política donde las fuerzas del mercado operaron con mucha mayor libertad? Abocando estos temas, introducimos una variedad de métodos avanzados de SIG para el conjunto de datos original. Específicamente, examinamos la separación y mezcla de los tres tipos básicos de uso del suelo, comercial, industrial y residencial, por las unidades espaciales de manzanas y calles. Fuera de esto, medimos con un nuevo nivel de precisión la enorme variación de densidad residencial y congestión que definían las crecientes desigualdades socioespaciales de las ciudades decimonónicas. La documentación sistemática y detallada de estos patrones espaciales, incluso sus resultados socialmente indeseables, ayuda a entender cómo se desarrollaron las ciudades del siglo XIX y qué condiciones generaron para justificar los crecientes controles al uso del suelo, desde los códigos de construcción hasta la zonificación por mandato gubernamental.

Key Words:

• historical GIS
• land use
• New York City
• urban history
• zoning

关键词：:

• 历史 GIS
• 土地使用
• 纽约市
• 城市历史
• 分区制度。

Palabras clave:

• SIG histórico
• uso del suelo
• ciudad de Nueva York
• historia urbana
• zonificación

Acknowledgments

We are thankful to the four anonymous reviewers and the editor for their insightful comments and suggestions that helped to significantly improve this article.

Notes

1. The three commissioners were Gouverneur Morris, John Rutherfurd, and Simeon De Witt, and the chief engineer and surveyor was John Randel, Jr.

2. Undoubtedly, crowdsourcing introduces concerns of data quality that are widely documented in the literature (e.g., Girres and Touya 2010; Heipke 2010; Rice et al. 2012). To the extent possible, we attended to issues of geometric or positional accuracy using automatic batch processes. Where necessary, we employed manual corrections for problems in both geometric and attribute accuracy. A more unusual challenge was to address fading colors on the historical Perris Atlas, which made land-use coding in a few cases ambiguous even after consulting various copies of the original maps. Importantly, the total number of these instances was negligible, and none of these problems amounted to a serious concern of data quality once individual data points were aggregated. It is important to note that several data quality concerns with crowdsourced spatial data (temporal accuracy, completeness, and lineage) and even with historical GIS are mitigated by the completeness of the data set and the constraints of its digitizing (Plewe 2002; Rice et al. 2012). Digitizers trace high-resolution source material against which researchers can verify results before potentially propagating errors through analysis, while attribute definitions are predetermined by the source material rather than by the digitizers.

3. On the orientation map and on the subsequent maps the shoreline has been generalized, leaving out the countless slips on both the Hudson and East River frontages. This decision is justified because the slips do not form part of the article's analysis.

4. Street-level analysis was also performed by Olson and Thornton (2011), using annual median rents by street as one of their variables to study sociospatial patterns in nineteenth-century Montreal.

5. Given the frequency of irregularly shaped buildings digitized as multiple polygons, this method attends to the possibility that not all errors were corrected in the original cleaning process, as it does not rely on the accuracy of the count of buildings but rather on their location, shape, and area relative to other buildings and the street system.

6. It is also true that exploiting the finer detail of a city directory, Davey and Doucet (1975) found considerable internal spatial differentiation within the CBD of Hamilton, Ontario, in 1853, a much smaller city than New York or Boston, between retailing, wholesaling, manufactories and crafts, professional services, financial institutions, administrative establishments, hotels, and boarding houses.

7. The single-use clusters were defined as groups of contiguous (including corners) blocks with significantly high percentages of each of the three land uses considered, using the Getis–Ord G_i* statistic for local clustering (Getis and Ord 1992). The maps in Figure 2 outline the blocks that cluster at a confidence level of 90 percent or higher (z score greater than or equal to 1.645).

8. To accommodate for differences in building sizes, their presence and prominence along the street frontage, and variation in street length, the maps of commerce on the street (Figures 3B–3C) represent the ratio of the footprint area of commercial buildings facing each street segment and the length of the corresponding segment. The analysis opts for this approach over considering the number or percentage of street-facing buildings with commercial uses for the quality reasons noted earlier (cf. note 2).

9. The authors are currently working on an article analyzing how the spatial algorithm of the Manhattan grid determined land-use patterns.

10. The best known example is Gramercy Park, where the land owner Samuel B. Ruggles imposed precisely such deed restrictions to boost property values and create an elite residential enclave.

11. Alongside this article's land-use analysis, the effects of the modifiable areal unit problem for the historical account of citywide population density are evident (Openshaw 1984). Relative to the census ward, the block as a unit of analysis is less prone to inaccurate generalizations both because of its smaller size and, importantly, because it more accurately reflects the descriptive unit of population density.

12. Specifically, the estimation method used here is designed as a hybrid filtered areal interpolation and dasymetric map to take advantage of the building-level scale of the data set and its land-use classification as ancillary data (Maantay, Maroko, and Herrmann 2007). Lacking desired information (e.g., the number of dwelling units per residential building), a dasymetric approach based on the size, density, or location of individual residential buildings alone was deemed insufficient as a basis for estimation. This was particularly true given the inability to validate our findings against other sources. Thus, the estimated population values were calculated at the unit of the block, reaggregating the data and weighting each block by its proportion of the corresponding ward's residential building area.

13. This underestimation can be attributed to the fact that with an areal interpolation of residential building area, each on-the-ground square meter is of equal weight regardless of the size of the building to which it belongs. Given that the narrative account describes several larger residences in less densely built neighborhoods housing single families, whereas many smaller footprint buildings in denser neighborhoods might be home to multiple families, this equivalency of area likely creates underestimated values within the most densely populated blocks.

14. More precisely, according to the 2010 Census, densities in the Lower East Side were 395 people per hectare in census tract 22.01, 158 in 22.02, and 144 in 26.01 (Social Explorer, http://www.socialexplorer.com/ [last accessed 1 February 2015]).

15. In principle, the method serves to identify tenement crowding across Manhattan. It is also true that lower ratios in urbanizing northern areas can be misleading. These blocks were often not yet developed into their interior, and although they might have contained several tenements, these were typically not located at the rear lots. As such, existing tenement buildings in the northern portion of the city might have been overcrowded, but they did not suffer the same lack of light and ventilation of their downtown, rear-lot counterparts. Additionally, the findings here could serve as baseline calculations for future research on the development of these blocks as more historical data become digitized.

16. We would like to express our gratitude to one of the anonymous reviewers, who in an exceptionally careful reading of our original submission, proposed this dynamic interpretation of five residential developmental patterns in the production of Manhattan spatial markets.