Knowledge Center
  1. Home
  2. Knowledge Center
  3. Applied Research

Toward a Standardized Land-Use Coding Standard

March 30, 1994

Introduction

The Research Department of the American Planning Association assisted the Federal Highway Administration in determining the interest of federal agencies in updating the 1965 Standard Land Use Coding Manual (SLCUM). This scoping project results are summarized below.

Working Paper

In 1965, the Federal Highway Administration and the Department of Housing (then the Bureau of Public Roads and the Urban Renewal Administration, respectively) published the SLUCM. The manual provided a detailed listing of land-use categories with numeric codes assigned to them. The categories were based on the Standard Industrial Classification (SIC) system. This coding procedure became the typical method for land-use coding in urban areas throughout the country. The manual was reprinted in 1972. Beginning in the late 1970s, the manual was used less frequently because land-use planning emphasized short-term, small-scale projects and the long-term horizon for planning was de-emphasized.

The 1965 SLUCM provided a general numeric coding scheme that used two, three, four, or more digits to identify land-use activities and an additional two to eight digits to identify ownership, type of structure the activity is housed in, auxiliary use codes for secondary land uses, etc. The primary emphasis of the SLUCM coding was to provide an exhaustive set of land uses derived from the SIC codes and a limited set of attribute data to further define some of the land-use categories. The manual provided illustrations of three attributes: ownership types; type of structures for residential uses; and crop types for farm uses.

The coding system as developed in 1965 was a recommended system; participating agencies or programs were not required to use it.

In brief discussions with several federal agencies, the following primary purposes surfaced. The contents of a SLUCM update could potentially include all of them. However, the manual will be limited to addressing, either through the main update of the SLUCM or an addendum, only those aspects in which a specific agency expresses interest. The purposes of revision are:

  • To develop an up-to-date and comprehensive list of land uses, and a flexible approach to categorizing new land uses in urban, suburban, and rural areas.
  • To provide a system of coding land uses which is consistent with and supporting of recent legislation. New Federal initiatives — such as the Clean Air Act Amendments of 1990 (CAAA), the Environmental Justice Order and the Intermodal Surface Transportation Efficiency Act (ISTEA) — and virtually all environmental data used to do geographically based policy analyses require land-use/land-cover data that is specific to program needs. To address such varying levels of detail, the update may have several parts, and each part could address a specific area of concern (e.g., transportation, environmental, economic, coastal management, etc.).
  • To facilitate the sharing of computer-based land-use/land-cover data by standardizing the coding system. Any sharing of such data will have to recognize the differences of studies between regulatory jurisdictions, as well as the issue of spatial resolution and levels of detail. The update, if necessary, will include computer-based conversion packages to assist in such conversion/sharing of land-use data.
  • To develop a coding system that will recognize the source of data or nature of data acquisition. For instance, land-use/land cover data could be acquired through digital images derived from remote sensing; digital or manual data derived from the interpretation of aerial photography, or site information gathered from surveys that include characteristics such as ownership, parcel boundaries, zoning, etc. Coding schemes should be easily adaptable to address such differences in scale.
  • To make possible periodic updates so that each part of the manual could be updated, revised, or modified independent of other parts.
  • To enable regional agencies (e.g., Councils of Government (COGs), Metropolitan Planning Organizations (MPOs), Regional Planning Agencies and Regional Special Purpose Districts) to acquire or disseminate land-use/land-cover data to local city or county governments by using a well-defined coding scheme to satisfy the needs of municipal planning and management, small area forecasts, monitoring regional trends, etc.
  • To create a coding system that is easily adaptable for use in Geographic Information Systems (GISs). The widespread use of GISs and other computer-based systems for spatial data analyses has put a premium on the use of existing available data and the portability of such data between various computer systems, as well as the aggregation of land-use categories and some uniformity and consistency in coding types.

Case Studies

Included here is a summary of 21 case studies of successful coding schemes currently in use for land-use and land-cover information. A wide range of coding schemes representing a broad spectrum of applications and sources of land-use/land-cover data was collected and reviewed during the research. The schemes included in this report are currently in use in both the public and private sectors for land-use/land-cover applications.

For the purposes of this study and to better understand the differences between systems, the coding schemes are classified based on their current use and suitability for use as a standard. This classification scheme and a table that summarize all of the major coding schemes are presented below. Additionally, each of the major coding schemes has been more fully explained in a narrative. Those narratives are followed by a listing of each of the coding schemes. All the examples are cited, and, where available, the source or contact name, address, and other details are also provided. The coding schemes used in the survey came from the following agencies:

Fairfax County, Virginia

County of Los Angeles, California

Department of Natural Resources, North Carolina

City of St. Louis, Missouri

Orange County, California

Washoe County, Nevada

Clark County, Nevada

U.S. Geological Survey

Institute of Transportation Engineers

Maryland Water Resources Administration, Department of Natural Resources

Atlanta Regional Commission, Atlanta, Georgia

Southern California Association of Governments

Army Corps of Engineers

Executive Office of Environmental Affairs, Massachusetts

Ohio Remote Sensing Program

Department of Natural Resources, Michigan

Note: All of the above examples are available in a downloadable Excel file.

Classification of Coding Schemes

Land-use/land-cover coding schemes are generally used to categorize land uses to study specific phenomena, such as commercial activity, transportation impacts, or environmental impacts. In the study of such phenomena, land-use/land-cover data is ascribed to spatial or geographic entities, such as parcels, census tracts, or one-acre grids. The choice of a specific coding scheme is based on the types of phenomena and the geographic entity that is being analyzed.

While it is relatively easy to determine the geographic extent of a coding scheme, and thereby its adaptability as a standard, the purposes for which land-use/land-cover information is to be used are not always clear. Some coding schemes can become quite complex when they seek to describe a large variety of attributes. For example, a coding scheme may further differentiate between public and private ownership (public and private parks, office buildings housing government offices and private offices, public and private schools). The complexity of a coding scheme also generally tends to increase as resolution levels increase. For instance, land-cover data acquired through remote-sensing has a significantly lower resolution than parcel-based land-use data. Some coding schemes counter this complexity by using a hierarchical coding scheme that makes it easier to aggregate various data.

In sum, the type of coding scheme that an agency chooses is typically based on two factors — the scale (geographic extent) at which the land-use/land-cover data is used and the source (data acquisition methods) of land-use/land-cover information. The other factor in making a choice, and one that is becoming increasingly critical due to advances in use of computer-based technologies, is the ease of manipulation of the various land-use/land-cover classes. This manipulation makes it possible to share digital information. Some classification methods, such as hierarchal methods, are easier to use when aggregating data than are non-hierarchal methods.

Scale

The four common scales for coding schemes are:

  • Neighborhood or small area scale: Coding schemes employed at this geographic level are generally very specific to the location and therefore not easy to standardize. Most coding schemes at this scale do not provide land-use classes uniformly for all major land-use types. Typically, neighborhood-scale urban planning efforts, such as urban renewal, special tax districts, etc., employ such coding schemes.
  • Citywide or countywide scale: Land-use coding schemes at this geographic extent are by far the most common. Almost every urban and suburban municipality in the country has some form of coding system. Land-use codes are generally assigned to a parcel-based database that often contains other social and cultural information, such as land values, housing types, taxes, ownership, etc. Coding schemes at this scale are often adaptable, especially by other cities and urbanized counties. Many of them are derivatives of coding schemes in the 1965 SLUCM and the Standard Industrial Classification Manual (SIC), with modifications for local conditions and application needs.
  • Regional scale: At this scale, several political jurisdictions may be covered to describe a set of common phenomena. Typical examples are coding schemes used by regional agencies, such as Councils of Governments (COGs) and coastal area management agencies. Generally, these schemes are tailored to specific program requirements. Common examples include environmental issues, such as coastal zone, floodplain and wetlands mapping. Other examples in the near future include transportation, air quality, land-use planning, and regional fair-share housing. In light of an increasing emphasis on regional coordination and planning, coding schemes at this level are currently growing fastest. The need for standardizing land-use/land-cover coding also appears to be most acute at this scale. This is mainly due to pressures to integrate data collection and subsequent dissemination in a uniformly acceptable manner among member jurisdictions. Lack of a coherent standard is generally attributed to the differences in the focus of land-use/land-cover classes among regional agencies.
  • Statewide scale: The coding schemes at this and higher (national, continental or global) are used by state agencies to aggregate information from individual counties and cities and/or disseminate information to such constituents. Such coding schemes are typically used by state agencies, such as the Departments of Natural Resources in Maryland, Ohio, Michigan, North Carolina, etc., and Federal agencies that are mandated to monitor phenomena at this scale, such as EPA, USGS, etc.

Source of Data

The primary source of land-use/land-cover data (or the data acquisition method) also affects the type of coding scheme used. The resolution (or the minimum mapping/coding unit) and accuracy of the data are dependent on the type of source. The three common methods are: site survey, aerial photography, and satellite-based remote sensing.

  • Site survey/land records: Site surveys and municipal land records are the primary means of data acquisition for parcel-based, land-use information. The 1965 SLUCM at the most detailed level was designed for such purposes. Land records data are the primary source of all land-use information at the neighborhood and citywide/countywide scales. Resolution for this type of data acquisition is accurate at a parcel level. Accuracy standards (both in terms of resolution level and timeliness of data) are generally the highest for this type of source. Coding schemes with 200 to 800 land-use classes are typical at this resolution level.
  • Aerial photography: Communities are increasingly using this method of data acquisition to verify or update site-surveyed, parcel-based, land-use information. This method is also used to automate the scanning and mapping of land uses for applications requiring thematic maps. The resolution of such sources varies from the census-block level to four feet for land-cover information. When combined with parcel-based data, the resolution and accuracy could be similar to that of the site-surveyed information. Coding schemes for land-use/land-cover data acquired only by this method generally contain 12 to 100 land-use classes.
  • Satellite-based remote sensing: This method of land-cover data acquisition is undergoing rapid change as improvements and accessibility in satellite, imaging, and scanning technologies increase. Typically resolutions (sometimes called pixel sizes) range from one- to 100-acre grids. Global imaging systems use significantly larger grids. The coding schemes generally contain no more than 10 to 15 types of land cover. Some statewide coding schemes use as many as 30 land-cover classes. Land-cover maps generally start at 1:24,000 scale.

Classification Methods

Coding schemes generally group together similar or related land-uses and sometimes land-cover types. The two common methods of classification are a simple exhaustive listing and a hierarchy of land uses.

  • Exhaustive listing: This method uses a list of land uses with sufficient codes to meet the application needs. This is the simpler of the two methods. This method is easier to customize for local conditions but is harder to share or aggregate. Typically, geographic extents at the neighborhood or citywide scale adopt this method. Certain types of land uses tend to have too many or too few classes. For example, typical urban and suburban jurisdictions may have up to 50 different codes for classifying residential uses and no more than one or two codes for all water-related uses. Applications that require an inventory of existing land uses generally use an exhaustive listing.
  • Hierarchy: In a hierarchal coding system, all land-use/land-cover types are grouped into classes, and each land-use class may contain additional subclasses or levels. In a hierarchal scheme there are 10 or 12 base classes for broad categories, such as urban land, agriculture, forest, water, wetlands, and barren lands. Under each of the first-level categories, there would be further classes. For example, urban land would contain residential, commercial, and industrial classes at the second level. At the third level, residential use may contain single-family, duplex, town houses, and multifamily classes. The 1965 SLUCM was loosely based on a hierarchal system. Hierarchal coding systems impose a structure by using 1-, 2-, 3-, or 4-digit codes. The preferred method for computer-based manipulation and the one in use in most new systems is the hierarchal coding system. Most of the coding schemes use numerical values and prefix the base-class numeric code to its subclasses. For instance, base-class numerical codes may range from 1 through 9, level-two classes range from 11 through 99, level-three classes range from 111 through 999, etc.

Summary of Coding Schemes

The Summary of Coding Schemes table compares the schemes. The schemes listed in the table are grouped by scale (or geographic extent) namely, Neighborhood/Area Plan Scale, Citywide or Countywide Scale, Regional Scale, and Statewide Scale.

Each of the coding schemes is identified by its location of use, the name of agency or jurisdiction, and the following:

  • Coding Scheme: Describes the literal classification method used in this example.
  • Levels: The number of levels in the coding scheme (in nonhierarchal systems, there is only one level).
  • Base Classes: The number of major land-use types or base classes in the first level.
  • Total Codes: The total number of land-use/land-cover codes in the coding scheme.
  • Primary Data Source: The source of data or the data acquisition method.
  • Smallest Unit: The smallest unit of measurement to indicate the level of detail.
  • Primary Appl.: The primary application(s) of the coding scheme.
  • Data Sharing: Other applications that share land-use/land-cover information collected for the primary application.

Descriptions of Coding Schemes Surveyed

Fairfax County, Virginia

Fairfax County uses four different coding schemes for land-use applications. Each coding scheme evolved over several years and was customized as needs changed. The primary source of land-use data for the entire 400-square-mile area of the county is the Land Master File, which is a parcel/land-records database. The Land Master File contains land-use data for more than 70 million square feet of built commercial space and 400,000 dwelling units spread over several urban cores, towns, and rural areas of the county. For assessments, zoning, and other routine county parcel-based land-use needs, the Existing Land Uses Coding Scheme is used. This is a three-level coding scheme that consists of about 150 three-digit codes. Every record in the Land Master File is a tax parcel and contains a list of all existing land uses for each parcel. Applications that require existing land-use information primarily use this coding scheme.

For projections and land-use-based impact analyses, the Planned Land Uses Coding Scheme is used. In this scheme, planned land-use designations, obtained from the Comprehensive Plan, are grouped differently. While the Existing Land Uses Coding Scheme is an exhaustive list of all different types of land uses loosely based on the Standard Industrial Classification, the Planned Land Uses Coding Scheme groups similar types of uses together. Applications that study impacts of growth and project demand for public services, such as schools, safety, transportation, etc., use this coding scheme. Because data requirements for such studies typically emphasize the magnitude of impact rather than precise land uses, each of the major land-use types is classified according to the intensity or density of development. For example, residential uses are grouped by the number of dwelling units per acre. In the Existing Land Uses Coding Scheme, they are grouped by the dwelling unit type (single-family, town houses, apartments, condominiums, etc.). The Planned Land Uses Coding Scheme also uses a three-digit code and contains more than 200 distinct codes. Depending on specific application needs, this coding scheme is frequently modified.

For thematic land-use mapping, such as the Comprehensive Plan Maps, simplified versions of the above two coding schemes are used. A coding scheme that contains 25 character-based codes is used for aggregating existing land uses. For planned uses, the Conceptual and Area Plan Land Uses coding scheme is used. This coding scheme contains 12 three-digit land-use codes and is generally aggregated from the Planned Land Uses Coding Scheme.

County of Los Angeles, California

The County of Los Angeles, like many large jurisdictions in the country, uses several land-use coding schemes to meet a diverse range of applications. Since the county consists of distinct geographic and development patterns, coding schemes for each planning area were customized to address specific planning needs. Three coding schemes included for comparisons from the survey are from: Malibu/Santa Monica Mountains Area Plan; Santa Clarita Valley Area Plan; and Hacienda Heights and Rowland Heights Area Plans.

Each of the coding schemes is very specific to its area and contains no more than 26 land-use codes. All of them use unique mixed numeric and character codes. As part of the county's General Plan/Zoning Consistency Program begun in 1986, the land-use information was incorporated in a land-use layer of a countywide GIS system. Each area's unique land-use codes were maintained in the system. The land-use layer of the GIS system, which also includes 31 other similar areas throughout the county, was used to identify areas where the zoning was not consistent with the General Plan land-use designation. Because of the complexity of the system and lack of funds to maintain and update the land-use portion of the database with information from each of the small area plans, the program is no longer used actively.

Department of Natural Resources, North Carolina

Superconductor/Supercollider Project (pdf)

This coding scheme was adopted for use in the feasibility analysis of a superconductor/supercollider facility in an area roughly equal to 16 quad (7.5' USGS) maps. The seven land-use categories represent the rural nature of the study area. Some categories, such as Urban Land, contain further delineation of land uses in an hierarchal system. The primary source of land-use data was aerial photography with a mapping unit of three acres. Although the source data was grid based, the 28 land-use codes were used to map typical overlays through a GIS system.

The coding scheme as used is being modified for other statewide applications in cooperation with counties developing such an information base.

City of St. Louis, Missouri

Map

St. Louis Land Records Management System (pdf)

The coding scheme used by the City of St. Louis Land Records Management System is the closest to a full implementation of the 1965 SLUCM and is typical of systems used in several cities across the country. The City of St. Louis is approximately 61 square miles in area with a population of about 400,000 and 195,000 dwelling units. The city is losing population. Most commercial uses are concentrated in the downtown area of the city. Although several new land uses were added or existing definitions modified, the coding scheme retains the four-digit numeric coding scheme of the 1965 SLUCM. Currently there are over 800 codes the city uses to designate all land uses in a parcel-based database. To accommodate secondary land uses, a new database field was added to the Land Record Management System. The secondary land uses information provides the city with more accurate land-use data.

The primary source of data collection is the Assessor's Office. Because of the sheer number of codes, there is a move by the Assessor's Office to use a simplified one- or two-digit coding system at the expense of losing land-use detail. In commercial areas, the License Collector's office uses a modified and simplified four-digit coding system. The primary challenge is to maintain a consistent land-use coding scheme across all city agencies.

Orange County, California

The 1990 Land Use Inventory program adopted a modified Level II Anderson coding scheme with about 26 land-use codes. The Anderson coding system was originally developed at the USGS and is 4-level hierarchial system. Because the primary source of data was aerial photography, the entire 800 square miles (22 USGS quads) were divided into 15,000 cells, and each cell was given one of the land-use codes. The cells were aggregated to census tracts, community planning areas, and cities to meet specific program needs. This coding scheme is a combination of a three-digit system mixed with a character code.

Principal applications of this system are to obtain a countywide inventory of land uses that could identify undeveloped land in relation to other developed land uses.

Washoe County, Nevada

The county's integrated terrain unit mapping project used a simplified Level II Anderson coding scheme to map the entire 6,600-square-mile county area for a vegetation layer of the GIS system. This layer is to be used in conjunction with a land-use layer that consists of both existing and planned land uses. The primary source of land-cover data is aerial photography with a mapping unit of two acres. The primary application of this information is the monitoring of water resources and the relationship of irrigated to nonirrigated land uses. Other applications were developed to assist in forest fire control and water-quality studies.

The coding scheme consists of two levels with a total of 29 codes. The county is currently trying to update the coding scheme to identify dedicated and nondedicated open space and forest lands.

Digital Line Graph — Enhanced, U.S. Geological Survey

Digital Line Graph — Enhanced (pdf)

DLG-E Model Overview from USGS (pdf)

Complete DLG Standards

The U.S. Geological Survey enhanced the Digital Line Graph model to a feature-based model. This model, DLG-E, uses a comprehensive approach to represent all types of features. The model uses five basic groups or views for all features — Cover, Division, Ecosystem, Geoposition, and Morphology. The views are based on some common defining characteristics. Cover reflects physical or material features; Division reflects cultural/political boundaries; Ecosystem reflects environmental features; Geoposition reflects locational characteristics relative to know or established points; and Morphology reflects the form (beach, basin, dune, island, valley, etc.) of land. Within each view there are subviews that clarify the distinctions between the various features. There are a total of 202 features in all five views. Most land-use/land-cover features are in the view Cover. Cover consists of 122 features grouped in five subviews — Barren Land, Built-up Land, Cultivated Cropland, Vegetation and Water.

The DLG-E model is used to generate the digital cartographic productions based on the National Digital Cartographic Data Base. This effort, which is part of the National Mapping Program, will be periodically reviewed by the USGS. A significant effort is being made to ensure that information based on the DLG-E data model can be exchanged with other federal, state and local agencies.

Institute of Transportation Engineers

ITE Manual, 5th Edition (pdf)

The ITE Manual provides trip generation statistics for various land uses based on data obtained by the ITE. The ITE Manual is a standard guide for estimating the number of trips that may be generated by a specific land use. Land uses are identified by a one-level coding scheme. Some of the uses have multiple codes to account for diversity in traffic impacts based on their size or density of development. There are 10 major land-use types with a total of about 127 land-use codes. With each new edition of the ITE Manual, the land-use codes are altered or expanded.

This coding scheme is entirely based on a single characteristic of land uses — trip generation. Typically, such coding schemes based on a single characteristic or attribute of a land-use/land-cover type are very application specific and not conducive to land-use data sharing.

Passaic, New Jersey

Passaic River Basin (pdf)

The Passaic River Basin is a 132 square-mile area of ten counties, 132 municipalities and a population of 2.4 million. The U.S. Army Corps of Engineers maintains a GIS containing hydrologic and land-use information for environmental analyses of the basin. The primary source of data is aerial photography although some land-use information for urbanized counties is provided or cross-checked by local planning agencies. The minimum mapping unit is a 10.33 acre grid unit. The coding scheme uses 13 basic land-cover classes for the GIS layer. The land-use information for urbanized areas is aggregated to these classes. The Corps provides hydrologic information to local agencies in addition to using this for its own programs.

MassGIS

The Massachusetts GIS Program (pdf)

The MassGIS program uses a statewide land-cover database to support various state and local program needs. The land-use data layer of the GIS was developed using 1:25,000 scale color infrared aerial photos with a minimum mapping unit of one acre. The coding scheme used for the land-use data layer has 21 land-use classifications and several additional classes were added for parts of Massachusetts. The coding scheme evolved from the MacConnel Coding Scheme, an earlier coding scheme developed by Prof. MacConnel and was widely used throughout the state of Massachusetts. The MacConnel coding system was a simple list of land uses that contained 104 classes. In the MassGIS adaptation, there are an estimated 70,000 polygons in the land-use data layer using the 21 classes (and additional classes for some parts of the state). The land-cover classes in use closely meet the environmental program needs around the state. Typical applications of this scheme are regional and statewide scale such as watershed protection buffer zones, air quality monitoring applications, pesticide regulation, solid waste sites, and monitoring coastal ecosystems. The MassGIS program also distributes its GIS data through computer files or published maps to other local and state agencies.

Note: Complete references for contacts and sources appear in the printed report.

Disclaimer: This material is based upon work supported by the Federal Highway Administration under AGREEMENT No. DTFH61-92-P-01590. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the Federal Highway Administration.

Prepared by the Research Department, American Planning Association, for the Federal Highway Administration, U.S. Department of Transportation. © 1994 American Planning Association