Deliver geospatial models direct to decision makers with SyncroSim › Forums › ST-Sim & State-and-Transition Simulation Models › Tracking disturbances for a pixel to trigger transition?
- This topic has 1 reply, 2 voices, and was last updated 2 years, 5 months ago by sarah-chisholm.
-
AuthorPosts
-
June 14, 2022 at 5:45 pm #21046sgreenlerParticipant
Hello! I am working with ST-SIM for a project and would like to add a transition that only occurs if there have been multiple disturbances while a pixel is within a specific state. More specifically, I would like to be able to model a transition from forest to woodland that is triggered by three or more low severity fires occurring while a pixel is within a specific state. I would like the first and second low severity fires to not trigger a transition (this is easy to do), but the third to trigger a transition. Is this something that is possible? It’s not clear to me if disturbances are tracked in a way that would allow this.
If it is possible, I would also be interested in modeling the inverse as well (if there have not been x number of low severity fires while a pixel is within a state it deterministically ages to a different age class than it does if there has been x number of fires). Ecologically, I’m thinking of a young woodland state being maintained by multiple low severity fires. With fire it would deterministically age to a mid-woodland state but with few or no fires it would age to mid-open state.
Thanks for your help with this!
June 17, 2022 at 6:58 pm #21047sarah-chisholmModeratorHi there, Here’s an example library of worked solutions to your questions, followed by a description of the library below:
Example Library: https://s3.us-west-2.amazonaws.com/apexrms.com.public/stsim-forum/Forum-Post-2022-06-14.ssim.backup.2022-06-16-at-14-23-45.zip
Part 1: Repeated fires required to transition to woodland
There are two approaches to conditionally trigger a transition based on the number of disturbances that occur in a pixel. The simpler approach requires that you create more state classes (see the ‘Definitions – States’ Project in the example library). For example, to trigger a transition from Forest to Woodland only after three low severity burns, you will need three Subclasses of the Forest Class representing 1) unburned Forest cells, 2) Forest cells burned once, and 3) Forest cells burned twice (see the ‘States’ tab in the ‘Definitions – States’ Project). In the Project Definitions, you’ll also need to define a ‘Fire – Low Severity’ transition. The last step is to add the appropriate Transition Pathways to a Scenario (see the ‘[22] Forest to Woodland – Using States’ Scenario). In this non-spatial single cell model, the State Class is initialized as unburned Forest. In the Transition Pathways, an unburned Forest cell has a 0.1 probability of transitioning to a Forest burned once after a low severity fire (note that all transition probabilities in this example are arbitrary). Once in the ‘Forest:Burned Once’ Subclass, low severity fires that occur within 20 years will transition this cell to the ‘Forest:Burned Twice’ and finally to the ‘Woodland:All’ Subclass. Burned forest cells can transition back to unburned forest if a low severity fire does not occur within 20 years following the previous fire (See Result Scenario [23]).
The second solution requires fewer State Classes, but more Transition Types and a Transition Group (see the ‘Definitions – TST’ Project). In this Project, only two State Classes are defined (‘Forest:All’ and ‘Woodland:All’). Under the ‘Transitions’ tab in the ‘Transition Type’ node, you’ll find that three types of low severity fires indicating the number of previous burns are defined. These Transition Types also belong to a single Transition Group, ‘Low Severity Fire’. In the child Scenario ([1] Forest to Woodland – Using TST), three Transition Pathways are defined, where the different low severity fire types sequentially transition Forest cells to Woodland after a cell has experienced a third burn. Under the ‘Initial Conditions’ tab, the ‘Initial TST Randomize’ node, TST for all Fire types are set to a minimum of 1000 years to set the initial time since a fire to a very large number; in other words, this landscape has not experienced recent fire. Lastly, under the ‘Advanced’ tab, the ‘Transition – Multipliers’ node is used to define at what time since the last fire each Transition Type and Group may occur.
Part 2: Transition from woodland to forest in the absence of fire, maintenance of woodland with fire
To model the succession of Woodland to Forest, two State Classes (Forest:All, Woodland:All) and two Transition Types (Fire, Succession to Forest) are needed (see the ‘Definitions – Woodland Succession’ Project). In the ‘[25] Woodland Succession’ baseline Scenario, the landscape is initialized in the Woodland State Class under the ‘Initial Conditions’ tab. Two Transition Pathways are defined, where 1) Woodland cells transition to Forest via succession after 10 years, and 2) Woodland is maintained if there is a fire. Under the ‘Initial Conditions’ tab, the ‘Initial TST Randomize’ node, time since transition is randomized for Fire when cells are initialized so that all cells have had a Fire between 0 and 10 years ago. Lastly, under the ‘Advanced’ tab in the ‘Time Since Transition Group’ node, the Succession to Forest Transition Type is associated with the Fire Transition Type. This sets the probability for Succession to be contingent on the time since a fire. In the ‘[28] Woodland Succession – Double Fire’ Scenario, a Transition Multiplier (under ‘Advanced’) is used to double the probability of fire, resulting in a slower progression out of Woodland to Forest (see Result Scenarios [27] and [29] to compare).
You can find a similar question posted to the forum here: https://syncrosim.com/forums/topic/using-tst-to-conditionally-set-transition-probabilities/
For more information on the Time Since Transition Group Datasheet, visit the online ST-Sim Reference pages: https://docs.stsim.net/reference/prop_time_since_transition_groups.html -
AuthorPosts
- You must be logged in to reply to this topic.