Sensitivity of the SLEUTH cellular automata model to time intervals between land use inputs

Authors: Damilola Eyelade*, University of California, Ighodalo Ijagbone, University of Ibadan
Topics: Land Use and Land Cover Change, Urban Geography, Geographic Information Science and Systems
Keywords: Urban Geography, Land use, Cellular automata, Development, Nigeria
Session Type: Paper
Day: 4/4/2019
Start / End Time: 9:55 AM / 11:35 AM
Room: Congressional A, Omni, West
Presentation File: No File Uploaded

The SLEUTH genetic algorithm (GA) is implemented in Ibadan Nigeria, a heterogenous metropolitan area in the developing world. Similar regions have proved very challenging in prior implementations of SLEUTH. This is particularly so where calibration was done using the brute force implementation of SLEUTH which required hundreds of hours of CPU time and resulted in modest model fits in several developing region cities. The new GA implementation of SLEUTH, provides CPU resource and time savings over previous implementations, borne out by relatively fast calibration times. Prior studies examining the calibration of the SLEUTH cellular automata (CA) model have tended to focus on the urban growth modeling function with scant regard to the land use function controlled by the deltatron model. An understanding of how divisions of time within the calibration period influences model coefficients independent of the initiation and end time of the calibration period will help infer the sensitivity of the SLEUTH model to land use input layers. In this ongoing study the calibration time scale is held constant while varying the number of years within this period used for the land use matrix calculation. Results from this will help establish the sensitivity of the model to the temporal variation in input land use datasets and the impact on the accuracy of modeling and forecasting particularly for developing world cities.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login