Straws in the Wind Blog Articles
As a spatial practitioner familiar with many aspects of spatial data capture and processing, I have been watching the responses to the opportunity that the National Broadband Network (NBN) offers the people of Australia and its federal, state, local and private spatial practitioners.
One hears very little from NBN Co about what it is doing with regards to using spatial data and processing in its rollout of its fibre and wireless services. In fact, one hears very little at all unless one regularly reads The Australian newspaper’s IT and Telecommunications section!
I have a particular interest in how well the NBN does its job.
OK, let’s put all this together.
Topography and Forests
Tasmania is of interest to me due to its highly topographic and forested nature. Its varied topography and vast areas of Private and Crown forests, and National Parks, makes the provision of Television, Radio, Wireless and Mobile Phone technologies difficult in all but the major residential areas.
Having an accurate model of this topography with its forest cover is therefore important.
The current digital elevation dataset for Tasmania are very coarse and also do not take into account the vegetation/forest that covers our hills.
If we do not have a good model of our topography and vegetation cover we will not be able to determine the optimal television, radio, mobile and wireless tower locations that would ensure maximal coverage in terms of signal and quality.
I have done some simple tests with the current coarse elevation dataset, vegetation and mobile towers in my area and they confirm that the coverage I experience at my house is, as expected, predictable (I live in the bottom of a cleared valley 25 minutes drive from Hobart).
How can the models and the coverage area be improved?
By having better models and a clearer understanding of the location of all customers.
Modelling Topography and Forests through LiDAR
Forestry Tasmania (FT) has led the state of Tasmania for the past 10 years or so in the application of LiDAR to its forest management. With LiDAR, and clever algorithms, it is now able to measure the standing volume of timber in native – and plantation – forest doing away with the unbiased statistical sampling methods.
FT, as a good corporate citizen, has asked other agencies to join with it in funding the rollout. But its vision was ignored until very recent. Even so, there is still, to my knowledge, no moves by any of the Government departments and agencies in Tasmania to fund the capture and use of LiDAR even though its applications, particularly for asset capture for Local Government, is phenomenal.
Besides mapping its own forests, FT is also starting to build quite a nice income stream by providing its LiDAR interpolation skills to external agencies and companies.
The cost to LiDAR the state is not high – of the order of $5M.
This amount is rough and depends on clear specification of:
But actual cost could be born by many interested parties such as: Google, NBN Co, Telstra, Optus, Vodaphone, State and Local Government, Forestry Tasmania, private agricultural companies etc.
Determining Service for Occupied Buildings.
If one is going to provide good service one needs to know where one’s customers are. The current method of providing mobile coverage is to site the towers where one can in the hope that whoever created the tower there in the first place knew it to be a good site to provide coverage to the majority of customers.
This approach is a top-down approach (a 180 degree approach) which is known to be insufficient to providing coverage.
How do we know the locations are optimal?
We can only know this by knowing the location of occupied buildings within the area being covered.
For example, even using simple line-of-sight algorithms it is relatively easy to create a view-shed from one or more mobile towers that would allow one to identify which occupied buildings within the view-shed would get service and a prediction of the quality of the service.
But, critically, such a dataset would allow one to do a reverse view-shed to identify optimal tower locations.
Building Dataset Availability
However, we don’t have a good, publically available dataset holding these locations.
The current Geographic Names Address File (G-NAF) from the PSMA is great but it doesn’t define a point that locates an occupied dwelling accurately via a point at the centre of its footprint. Also, the dataset only holds rateable property locations and not the location of all possible residential buildings.
Now, the Australian Bureau of Statistics (ABS) each census visits all residential dwellings and locates them. However, the locational aspect of each dwelling is thrown away after census to ensure, as with its other data, that individuals cannot be identified.
This is a pity because individual confidentiality would not be compromised by having a dataset that contained nothing more than:
Note that one does not need any other information: address, name of occupier, age, sex etc.
But such a dataset cannot be sourced from the ABS. So, as they say “get over it!”.
Where could we get one from? Possibly by the following method.
(The roads dataset would be great for determining quality of wireless/radar coverage along major or minor roads.)
Coupled with a decent occupied building point location dataset the elevation and vegetation height datasets would provide a critical base set of data that is targeted to the needs of a modern digital economy.
I have not seen any moves towards the capture and provision of such datasets.
Are we failing to grasp this due to competition and privacy issues? Or because Governments and their agencies rarely ever provide the sort of leadership that is required?