[talk-au] Vic State Forest boundary files

Little Maps mapslittle at gmail.com
Thu Oct 21 06:29:34 UTC 2021

Hi all, I started a new thread so these links can be easily found in the future. This builds on a recent thread on permissible data sources  to answer 3 questions: (1) Which data layers from the Vic gov contain State Forest (SF) boundary data? (2) Which of these layers do we have permission (with waiver) to use? (3) Do these permissions enable editors to use of State Forests boundaries on the Vic gov MapShare site? Happy to hear back from everyone and to fix up any stuff-ups.

As stated on the Aus data catalogue and confirmed on the forum by Andrew Harvey, we have a waiver to use all of the CC by 4.0 datasets that are listed on DELWP's Vic Map catalogue. (I'll call these the 'core' VicMap datasets for brevity). DELWP has many other CC4.0 datasets that we don't have explicit permission to use according to the Aus data catalogue (I'll call these 'non-core' DELWP/VicMap data).

Aus data catalog: https://wiki.openstreetmap.org/wiki/Australian_data_catalogue
DEWLP VicMap Catalogue: https://www.land.vic.gov.au/maps-and-spatial/spatial-data/vicmap-catalogue

Core VicMap datasets

I checked through the metadata files and/or downloaded the GIS datasets for all of the core VicMap datasets to try to find a full set of State Forests boundaries. The core VicMap files contain 3 versions of the State Forests boundaries.

The VicMap Lite dataset [VMLite_PUBLIC_LAND_SU5] has a simple, low-res dataset that shows the boundaries of all SFs and other types of public land. It doesn't show their names. A single polygon might contain many contiguous SFs (each of which has a different name in reality). The dataset is old, and was last updated in 2008. The Lite datasets were intended for use  at broad zoom levels on web devices.  The dataset isn't much use for our purposes (although some of the Lite boundaries are of comparable 'accuracy' to some of the coarser SF boundaries that currently exist in osm.)


The VicMap Crown Land Tenure dataset [CL_TENURE_VIEW] is where you'd expect to find SF boundaries. It does contain some, but it's a mess. Most SFs aren't included in the dataset, and when the dataset is compared against more recent (non-core) DELWP data, there doesn't seem to be any rhyme or reason to why most SFs were omitted. [I'm guessing that this resulted from inter-agency rivalries between DELWP and the privatised VicForests, but that's speculation on my part. Public land tenures other than SFs seem to be well represented on the Crown Land Tenure map and are also shown on a few other core datasets. SFs are conspicuously absent.]


The VicMap Property dataset (VMPROP.PARCEL_CROWN_APPROVED) contains detailed, accurate boundaries of all land parcels on crown lands. But it doesn't show which parcels are in SFs and which are other crown lands. Nevertheless, if local editors know where a SF is on the ground, this dataset does provide accurate boundary data which could be used I guess. This dataset can also be accessed and data subsets can be downloaded in MapShare.


More recent, non-core DELWP datasets

The best sources of boundaries for SFS and other public lands are two, recent datasets: (1) Public Land Management (PLM25) and (2) Public Land Management (PLM25) Generalised.


PLM25 is described as "... a derived layer, using VMPROP.PARCEL_CROWN_APPROVED as the base layer. It relies on tabular and spatial sources to get the land description. PORTAL, PRIMS are the tabular sources of information. PARKRES, VEACRECS25 and CL_TENURE are the spatial sources. <i>Additional layers were also created to describe areas of state forests</i>, plantations, water frontages and commonwealth land." [my italics]

The "generalised" dataset is said to have fewer tags and some dissolved boundaries. However, the line work for the two seems to be identical, and equally precise along complex, curved boundaries, with one key difference. Many SFs contain a number of parcels of land, some contiguous (i.e. adjoining with shared boundaries), and some adjacent, but separated by easements. In both datasets, adjacent parcels that are separated by easements are shown as separate polygons. However, the base PLM25 dataset also contains separate polygons for contiguous, adjoining parcels that have shared boundaries, whereas the Generalised dataset dissolves all the internal boundaries to form one single polygon for the contiguous parcels. 

(Hope this makes sense. If not, imagine two rows of 5 residential house blocks on opposite sides of a road, giving 10 houses in total. Now imagine that these house blocks (=land parcels) are in a SF and the ‘road' is a gov easement. In the PLM25 dataset, every house block (land parcel) would be shown as a separate polygon, giving 10 polygons in total, each tagged as part of the SF. In the Generalized dataset, all of the house blocks on one side of the street would be combined into one polygon, and those on the other side would be in another polygon, giving 2 polygons in total, both tagged as being part of the SF.). Apart from this , and fewer fields/tags, the two datasets seem identical.

The two PLM datasets can be downloaded as GIS datafiles or viewed as WMS files, which could be used as a base layer image in iD or JOSM to trace boundaries from, for those who don't use a GIS. The dept provides wms files for the SFs boundaries (i.e. excluding other types of public land) at:


PLM25 vs VicMap Crown Land Tenure

As noted earlier, the VM Crown Land Tenure dataset contains only a fraction of the SFs in Vic. This image - https://imgur.com/E6FtbB1 - compares the two datasets. The Blue areas show all SFs in Vic, as provided in the new PLM25 dataset. The Red polygons show all polygons in the older Crown Land dataset that overlie SF boundaries that are shown in PLM25. (Thus all red patches have a blue patch beneath them). As can be seen, most SFs in the new PLM25 dataset are not shown in the older, core VicMap Crown Land Tenure dataset that we have permission to use.

MapShare boundaries

MapShare contains a copy of the SF boundaries. On MapShare, users can click on a polygon (or point or line), add it to a 'drawing layer', add more polygons to the drawing layer, and then download a shape file that contains all of the features on the drawing layer. Hence, a user could download the complete boundary files for one or more SFs. This raises the question, which DELWP dataset is used in MapShare? As best I can tell, MapShare uses the detailed PLM25 dataset.  When you click inside a SF, each parcel of land inside a SF forms a separate polygon, as in the PLM25 dataset and not the Generalised dataset. The source data is derived from a "State Forest Holding Layer", which is the same field name used in the PLM datasets, but not in the older VicMap Crown Land Tenure dataset. The boundaries are extremely precise and definitely aren't derived from the older, generalised 'Lite" dataset.

Implications for OSM

The PLM25 and PLM25 Generalised datasets contain the most accurate SF boundaries. I'll leave it to the group to decide whether these are covered by the existing waiver. If not, it would be great to get permission to use them. These would be the best datasets to use if a formal import of SF boundaries was ever considered. If the PLM25 dataset is considered to be not covered by the existing waiver, then I imagine it would also be considered to be inappropriate to get SF boundaries from MapShare.

We do have a waiver for the VicMap Crown Land Tenure dataset but I suggest that editors avoid using this data for SF boundaries. Most SFs aren't on it anyway, as the image above shows.

We also have a waiver for the Vic Property and VicMap Lite datasets (although VicMap Lite isn't listed on the Aus data catalogue atm). If editors are keen to improve the accuracy of boundaries of a local forest, the boundaries in the VicMap Lite dataset, while coarse when compared to the PLM25 data, may still be much better than some of the coarsest approximations to SF boundaries that currently exist in osm. If matched to features on satellite images, the VicMap Lite data could still provide a useful way to refine existing boundaries. Similarly, if you're really keen, boundaries could be extracted from the Vic Property dataset (which is on MapShare), although it would take some work to match these to on-ground boundaries.

Hope this helps. Thanks to everyone who provided suggestions on the earlier topic thread. I'm happy to chase up more info if it's useful. Cheers Ian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-au/attachments/20211021/b6982745/attachment-0001.htm>

More information about the Talk-au mailing list