Urban growth modeling

The Clarke urban growth model (UGM) is used to simulate urban growth for a study area and is the main component of SLEUTH. UGM functions independently of the land transition, deltatron module of SLEUTH. If modeling only urban/non-urban transitions, simply remove or comment out any reference to land use transition files in the input images list of the scenario file. The fmatch value in the control_stats.log file, which is a measure of land class best fit, will be 0.0.


Land use/cover modeling

The deltatron model is used to simulate non-urban land class transitions. Deltatron is driven by, and is loosely coupled with UGM. Whether land use or land cover is being modeled is dependent upon input type. The developers of SLEUTH are of the opinion that the class transition dynamics simulated by the deltatron approach can handle either classification system equally well.




Test mode is intended to give the user an easy way to execute a single run on a data set to confirm that the model is performing correctly, or produce output files for a specific set of coefficients.

Test mode will perform a single run through the historical data using the CALIBRATION_*_START values to initialize growth. It will complete the the set number of MONTE_CARLO_ITERATIONS, and then conclude execution.

All output files that can be created in calibrate mode can also be produced in test mode. It depends only upon the scenario file flag settings.

Unlike calibrate mode, in test mode annual GIF images of simulated land cover change are generated and written to the location indicated by the OUTPUT_DIR flag. These images may be used to visually assess how the coefficient values are affecting growth, as well as generate an animation of modeled historic urban growth.
(See also: Test Process Flow)



Calibrate mode is intended to narrow the large number of possible coefficient sets to a reasonable estimate of best fit values using brute force calibration methods.

Calibrate will perform monte carlo runs through the historical data using every combination of the coefficient values indicated. The CALIBRATION_*_START coefficient values will initialize the first run. A coefficient will then be increased by its _STEP value, and another run performed. A coefficient will no longer be incremanted by it's _STEP value when it's _STOP value has been exceded. This will be repeated for all possible permutations of given ranges and increments. (See also: Calibration Process Flow)



Predict mode will perform a single run, over the set number of MONTE_CARLO_ITERATIONS, using the PREDICTION_*_BEST_FIT values for initializaion. (See also: Prediction Process Flow)



back to Implement