Title

Land Use History data (1837-2005) 

Alternate title

LANDUSE_HISTORY_1837_2005_25m 

Purpose

Sinclair, S.J., White, M.D., Medley, J., Smith, E. and Newell,G.R. (2012) Mapping the past: constructing a digital land-use history map for Victoria, Australia. Proceedings of the Royal Society of Victoria 124(3): 193-206 

Credit

Sinclair, S.J., White, M.D., Medley, J., Smith, E. and Newell,G.R. (2012) Mapping the past: constructing a digital land-use history map for Victoria, Australia. Proceedings of the Royal Society of Victoria 124(3): 193-206 

Supplemental Information

Related Documents: None 

Status

Completed  

Code

4283 

Maintenance and update frequency

Unknown  

Title

DIGITAL 2 

Topic category

• Society

Use limitation

Public access 

Classification

Unclassified  

Statement

Dataset Source: Individual `time slices' representing land use at the following dates have been sourced from primary sources (paper copies in archives), and brought into a GIS environment: (1837, 1843, 1866, 1888, 1901, 1911, 1923, 1944, 1988). The 2005 time slice was remote-sensed. Dataset Originality: Primary & Derived 

Description

Individual `time slices' representing land use at the following dates have been sourced from primary sources (paper copies in archives), and brought into a GIS environment: (1837, 1843, 1866, 1888, 1901, 1911, 1923, 1944, 1988). Most map sheets were scanned, brought into a GIS environment and geo-referenced (Arcview 3.2, ESRI & ImageWarp 2.0). These were overlayed with the current cadastre in the GIS, and every polygon was assigned to a class by hand, based on the geo-referenced image. Some sheets were physically unable to be scanned, and were photographed in segments. Those sheets were not geo-referenced, and the parcels were assigned to their class by visual comparison. Some land use boundaries cut across the current cadastral parcels. To assign historic information to the current cadastre, we used a simple `majority rule': whatever category applied to most of the parcel by area was assigned to the whole parcel. We also used the `majority rule' to deal with parcels that have changed their size and shape over time (due to subdivision of freehold land, resumption of redundant parcels by the crown, etc.). The 2005 time slice was remote-sensed. A chrono-sequence of cloud free, Landsat Thematic Mapper (LTM) summer images of the state of Victoria were constructed by mosaicing contemporaneous LTM scenes. Ten separate images for the years 1989, 1991, 1992, 1995, 1998, 2000, 2002, 2004 and 2005 inclusive were processed to obtain the median values for each 25 m x 25 m pixel across the 16 years for each of the 5 visual and thermal bands. From these data we created a median Normalised Difference Vegetation Index (NDVI) surface for Victorian for the years 1989 to 2005 using the near infrared (NIR) and visible red (VIS) LTM bands. The NDVI image was then subjected to unsupervised classification employing a K-means clustering procedure, set-up to circumscribe 150 classes. These classes were further aggregated to 5 land-uses (built-up, irrigated agriculture, native tree cover, dry-land agriculture and water-bodies) based on a supervised classification informed by the examination of aerial photographs and field visits. Following the identification of these five primary land-uses and/or land cover types, 4 additional mapped land uses were interposed on the grid surface; 1.plantation forestry (based on recent aerial photographic interpretation), 2.conservation reserves - based on public land cadastral mapping, 3.dryland cropping, and 4.unimproved or native pastures based on supervised and thresholded models of these often sub-parcel features. 

Classification

Unclassified  

Attribute Quality

Comments

The land-use categories are generalised, and may not capture regional variation. 

Positional Accuracy

Comments

- The dataset is intended for use at the state-wide level (where it is assumed to show real patterns), and is not considered accurate for assessing individual parcels. - The dataset does not show within-parcel patterns. - 25m grid 

Conceptual Consistency

Comments

The dataset relies on historical mapped sources, of varying scale and accuracy, and has not been checked in the field. 

Missing Data

Comments

VIctoria 

Excess Data

Comments

The field VALUE contains all data. Land use is encoded as follows: 110 - Poorly-known indigenous use 120 - None, or resource or nature conservation, Crown 130 - Dryland Agriculture, type unknown 131 - Grazing, native vegetation 140 - Miscellaneous Reserve 151 - Goldfields, mines and quarries 152 - Timber production (mostly hardwood) 230 - Dryland Agriculture, type unknown 231 - Grazing, native vegetation 232 - Grazing, improved pasture 233 - Cereal cropping and dryland horticulture 253 - Timber production (mostly plantation) 260 - Built (private) 270 - Irrigated, type unknown 271 - Irrigated pasture 272 - Irrigated horticulture