Revisiting land use classification and spatial aggregation for modelling integrated urban water systems

Peter Marcus Bach, Severin Severin Staalesen, David Thomas McCarthy, Ana Deletic

Research output: Contribution to journalArticleResearchpeer-review

18 Citations (Scopus)

Abstract

Land use and different scales of spatial data aggregation are important for integrated urban water modelling. Inadequacy of land use classification in such models, however, prompted development of a new water-centric classification. A thirteen-category system (considering land cover, utilisation, and urban zoning aspects) was devised and applied to Melbourne, Australia. Sensitivity of the classification to spatial aggregation was tested by analysing randomly sampled subsets of Melbourne’s urban fabric for variation in landscape patterns, land use dominance and fragmentation. Significant differences were found when comparing inner city (highly fragmented, dominated by many land use categories) to fringes, (only few dominant categories e.g. residential, reserves), which are challenging to accurately represent in urban water models if oversimplified. The new classification encompasses key integrated urban water
management aspects, but does not capture local-scale features of the urban environment. Significant shortcomings for using spatial aggregation methods to simplify data for urban water models were found. Solely focussing on spatial coverage of land use instead of land use mix neglects important water interactions between categories (e.g. open space and residential). Trade-offs between spatial resolution and computational efficiency, for example, can degrade accuracy of representing urban water characteristics as less-dominant land uses are systematically removed. Finally, spatial aggregation methods likely result in the loss of minor (yet important) land uses (e.g. Civic). More effective methods for representing the variability of urban environments, which depart from simple aggregated rasters and sole reliance on land use, may be required to better define urban complexity in our water models.
Original languageEnglish
Pages (from-to)43 - 55
Number of pages13
JournalLandscape and Urban Planning
Volume143
DOIs
Publication statusPublished - Nov 2015

Keywords

  • Integrated urban water management (IUWM)
  • Land cover
  • Landscape patterns
  • Land use
  • Spatial numerical modelling

Cite this

@article{7f4c3e2529e0404c986750dca65a7d27,
title = "Revisiting land use classification and spatial aggregation for modelling integrated urban water systems",
abstract = "Land use and different scales of spatial data aggregation are important for integrated urban water modelling. Inadequacy of land use classification in such models, however, prompted development of a new water-centric classification. A thirteen-category system (considering land cover, utilisation, and urban zoning aspects) was devised and applied to Melbourne, Australia. Sensitivity of the classification to spatial aggregation was tested by analysing randomly sampled subsets of Melbourne’s urban fabric for variation in landscape patterns, land use dominance and fragmentation. Significant differences were found when comparing inner city (highly fragmented, dominated by many land use categories) to fringes, (only few dominant categories e.g. residential, reserves), which are challenging to accurately represent in urban water models if oversimplified. The new classification encompasses key integrated urban watermanagement aspects, but does not capture local-scale features of the urban environment. Significant shortcomings for using spatial aggregation methods to simplify data for urban water models were found. Solely focussing on spatial coverage of land use instead of land use mix neglects important water interactions between categories (e.g. open space and residential). Trade-offs between spatial resolution and computational efficiency, for example, can degrade accuracy of representing urban water characteristics as less-dominant land uses are systematically removed. Finally, spatial aggregation methods likely result in the loss of minor (yet important) land uses (e.g. Civic). More effective methods for representing the variability of urban environments, which depart from simple aggregated rasters and sole reliance on land use, may be required to better define urban complexity in our water models.",
keywords = "Integrated urban water management (IUWM), Land cover, Landscape patterns, Land use, Spatial numerical modelling",
author = "Bach, {Peter Marcus} and Staalesen, {Severin Severin} and McCarthy, {David Thomas} and Ana Deletic",
year = "2015",
month = "11",
doi = "10.1016/j.landurbplan.2015.05.012",
language = "English",
volume = "143",
pages = "43 -- 55",
journal = "Landscape and Urban Planning",
issn = "0169-2046",
publisher = "Elsevier",

}

Revisiting land use classification and spatial aggregation for modelling integrated urban water systems. / Bach, Peter Marcus; Staalesen, Severin Severin; McCarthy, David Thomas; Deletic, Ana.

In: Landscape and Urban Planning, Vol. 143, 11.2015, p. 43 - 55.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Revisiting land use classification and spatial aggregation for modelling integrated urban water systems

AU - Bach, Peter Marcus

AU - Staalesen, Severin Severin

AU - McCarthy, David Thomas

AU - Deletic, Ana

PY - 2015/11

Y1 - 2015/11

N2 - Land use and different scales of spatial data aggregation are important for integrated urban water modelling. Inadequacy of land use classification in such models, however, prompted development of a new water-centric classification. A thirteen-category system (considering land cover, utilisation, and urban zoning aspects) was devised and applied to Melbourne, Australia. Sensitivity of the classification to spatial aggregation was tested by analysing randomly sampled subsets of Melbourne’s urban fabric for variation in landscape patterns, land use dominance and fragmentation. Significant differences were found when comparing inner city (highly fragmented, dominated by many land use categories) to fringes, (only few dominant categories e.g. residential, reserves), which are challenging to accurately represent in urban water models if oversimplified. The new classification encompasses key integrated urban watermanagement aspects, but does not capture local-scale features of the urban environment. Significant shortcomings for using spatial aggregation methods to simplify data for urban water models were found. Solely focussing on spatial coverage of land use instead of land use mix neglects important water interactions between categories (e.g. open space and residential). Trade-offs between spatial resolution and computational efficiency, for example, can degrade accuracy of representing urban water characteristics as less-dominant land uses are systematically removed. Finally, spatial aggregation methods likely result in the loss of minor (yet important) land uses (e.g. Civic). More effective methods for representing the variability of urban environments, which depart from simple aggregated rasters and sole reliance on land use, may be required to better define urban complexity in our water models.

AB - Land use and different scales of spatial data aggregation are important for integrated urban water modelling. Inadequacy of land use classification in such models, however, prompted development of a new water-centric classification. A thirteen-category system (considering land cover, utilisation, and urban zoning aspects) was devised and applied to Melbourne, Australia. Sensitivity of the classification to spatial aggregation was tested by analysing randomly sampled subsets of Melbourne’s urban fabric for variation in landscape patterns, land use dominance and fragmentation. Significant differences were found when comparing inner city (highly fragmented, dominated by many land use categories) to fringes, (only few dominant categories e.g. residential, reserves), which are challenging to accurately represent in urban water models if oversimplified. The new classification encompasses key integrated urban watermanagement aspects, but does not capture local-scale features of the urban environment. Significant shortcomings for using spatial aggregation methods to simplify data for urban water models were found. Solely focussing on spatial coverage of land use instead of land use mix neglects important water interactions between categories (e.g. open space and residential). Trade-offs between spatial resolution and computational efficiency, for example, can degrade accuracy of representing urban water characteristics as less-dominant land uses are systematically removed. Finally, spatial aggregation methods likely result in the loss of minor (yet important) land uses (e.g. Civic). More effective methods for representing the variability of urban environments, which depart from simple aggregated rasters and sole reliance on land use, may be required to better define urban complexity in our water models.

KW - Integrated urban water management (IUWM)

KW - Land cover

KW - Landscape patterns

KW - Land use

KW - Spatial numerical modelling

UR - http://goo.gl/vZZccG

U2 - 10.1016/j.landurbplan.2015.05.012

DO - 10.1016/j.landurbplan.2015.05.012

M3 - Article

VL - 143

SP - 43

EP - 55

JO - Landscape and Urban Planning

JF - Landscape and Urban Planning

SN - 0169-2046

ER -