February 10, 2019 at 8:02 pm #6631
A question I recently received via e-mail:
I have a question regarding Adjacency Multipliers. This might be quite basic but I’m happy to put it in the forum if you think it suitable (although I’ve added lots of details of my specific case).
In my example, I’m trying to model urban growth of Bogota. I have 3 state classes: 1) Non-urban: rural, 2) Urban: Low density (LDC) and 3) Urban: High density (HDC). Two transition pathways: High-urbanisation which is either LDC->HDC or Rural -> HDC, and Low-urbanisation which is Rural -> LDC (see screenshots below). I want to use transition targets rather than probabilities. I have separate targets for growth of LDC (27km2/yr) and HDC (15km2/yr) which is why I have created two transition pathways. Before adding planning constraints, I want the model to ‘spread’ the urbanisation from existing LDC and HDC areas. So I used the transition-spatial- adjacency multipliers copying the values from the Daniel2016 files. However, as you can see from the resultant maps, it didn’t work.
Is this perhaps due to the Attribute Value? So far in the model, the only map input is a raster of Rural, LDC and HDC with values of 3, 2 and 1 respectively. Are the attribute values referring to those values and so as there are no zeros, it’s not working? The details about the adjacency-multipliers in the supplementary info was specific to the zoning map for which additional maps were inputted. I’ll need to do this down the line but for right now, a bit confused about how it works at the initial stage.
I hope I have explained that adequately but please let me know if you need further clarification.February 10, 2019 at 8:15 pm #6632
I suspect that you are correct that you have not setup attribute values correctly. the steps to use adjacency multipliers are the following:
1. create the state attribute types (Project Properties | Advanced | Attributes | State Attributes) that you will need to apply your adjacency multipliers. For example if your transition depends on a derrived function of the state classes in the neighborhood of the cell that might transition such as population density, then you would create a state attribute type called “population density”.
2. define the “State Attribute Values” associated with each state class (Scenario Properties | Advanced | Attributes | State Attribute Values). This is the value of for example “population density” associated with each state class in your model. You need only include state classes that have a population density.
3. Define the transition adjacency settings (Advanced | Transitions Spatial | Transition Adjacency Multipliers | Transition Adjacency Settings). These include, the radius for your moving window analysis. The attribute type that applies to each transition group and the frequency at which you want to apply your moving window analysis.
4. Define the transition adjacency multipliers (Advanced | Transitions Spatial | Transition Adjacency Multipliers | Transition Adjacency Multipliers). This is the relationship between the average attribute values (i.e., population density in this example) and the multiplier you want to apply to the transition probability. Note that the model interpolates between values. In your example, your multiplier will likely be zero or very low if the attribute value is zero and increase as a function of an increase in population density within the moving window analysis.
Things could go wrong if: your radius is too small, ie. it does not include neighboring cells; if you did not define state attribute values or there may be other use cases where things are not setup properly. If you are still not able to figure out the problem, you could send me your library to take a quick look.February 10, 2019 at 8:18 pm #6633
Also here is a link to the documentation page for adjacency multipliers:
You must be logged in to reply to this topic.