Development and applications of a comprehensive land use classification and map for the US.

Adina Maya Merenlender,David M. Theobald

doi:10.1371/journal.pone.0094628

Adina Maya Merenlender, David M. Theobald

Open Access

https://doi.org/10.1371/journal.pone.0094628

Copy DOI

Journal: PloS one	Publication Date: Apr 11, 2014
Citations: 77	License type: CC BY 4.0

Affiliation: Colorado State University, Conservation Science Partners

Abstract

Land cover maps reasonably depict areas that are strongly converted by human activities, but typically are unable to resolve low-density but widespread development patterns. Data products specifically designed to resolve land uses complement land cover datasets and likely improve our ability to understand the extent and complexity of human modification. Methods for developing a comprehensive land use classification system are described, and a map of land use for the conterminous United States is presented to reveal what we are doing on the land. The comprehensive, detailed and high-resolution dataset was developed through spatial analysis of nearly two-dozen publicly-available, national spatial datasets – predominately based on census housing, employment, and infrastructure, as well as land cover from satellite imagery. This effort resulted in 79 land use classes that fit within five main land use groups: built-up, production, recreation, conservation, and water. Key findings from this study are that built-up areas occupy 13.6% of mainland US, but that the majority of this occurs as low-density exurban/rural residential (9.1% of the US), while more intensive built-up land uses occupy 4.5%. For every acre of urban and suburban residential land, there are 0.13 commercial, 0.07 industrial, 0.48 institutional, and 0.29 acres of interstates/highways. This database can be used to address a variety of natural resource applications, and I provide three examples here: an entropy index of the diversity of land uses for smart-growth planning, a power-law scaling of metropolitan area population to developed footprint, and identifying potential conflict areas by delineating the urban interface.

Highlights

The majority of the global biosphere is occupied by humanmodified landscapes of agricultural and urban land uses, with less than 20% semi-natural and about 25% ‘‘wild’’ [1]
Classification system Ideally, a classification system should be applicable over broad extents, contain hierarchical classes that would allow aggregation and a comprehensive range of classes so that all areas fit within a class; result in repeatable and consistent results through time and by interpreter; and have reasonably-balanced levels of accuracy amongst classes [14]
Using the four-group structure, I developed sub-group and classes building on Anderson level 2 codes as well as the typology defined by the North American Industrial Classification System (NAICS)

Summary

Introduction

The majority of the global biosphere is occupied by humanmodified landscapes of agricultural and urban land uses, with less than 20% semi-natural and about 25% ‘‘wild’’ [1]. Some land cover maps depict ‘‘developed’’ or built-up land cover types, but these typically miss low-density development that is either diffuse but has direct and permanent human activities or obscured by canopy cover [6]. This underrepresentation has been estimated to be roughly twice the extent of built-up land cover in the US [7]. Data products designed to resolve different land use types and patterns would complement land cover datasets and likely improve our ability to understand the extent and complexity of human modification

Methods

Results

Conclusion