Quoted from: https://doi.org/10.1371/journal.pone.0224998  Yunqiang Liu,Long Li ,Longqian Chen,Liang Cheng,Xisheng Zhou,Yifan Cui,Han Li,Weiqiang Liu. 2019. Urban growth simulation in different scenarios using the SLEUTH model: A case study of Hefei, East China. Plos One.

The SLEUTH model consists of two sub-models: the urban growth model (UGM), which can be run independently, and land cover deltatron model (LCD) [15]. The SLEUTH model requires a minimum of four periods of city-wide layers, two periods of transportation layers, slope layers, hill- shade layers, and exclusion layers [29]. Based on grid cells, the SLEUTH model assigns the attributes of the city or non-city to each cell and simulates urban growth by four conversion rules—spontaneous growth, new spreading center growth, edge growth, and road-influenced growth [24]. Spontaneous growth defines the occurrence of random urbanization of land, i.e., randomly selected non-urbanized cells may be transformed into urbanized cells when slope conditions are appropriate. An urban spreading center is defined as a location with three or more adjacent urbanized cells. New spreading center growth determines whether any of the new, spontaneously urbanized cells would become new urban spreading centers. Edge growth defines the growth of an existing spreading center, which simulates the urban's fill-in growth and the expansion of boundaries [24,28]. Road-influenced growth encourages urbanized cells to develop along road network, which simulates the impact of existing transportation infrastructure on urban growth. The mentioned four urban growth rules are performed sequentially in each growth cycle and are controlled by five growth coefficients [44], namely dispersion coefficient, breed coefficient, spread coefficient, road gravity coefficient, and slope coefficient. The relationship between the four growth rules and the five coefficients is shown in Table 1.

The execution of the SLEUTH model is divided into four major steps (Fig 2): input data preparation, model calibration, model prediction, and model output [45]. Three urban growth scenarios were planned by changing the exclusion layer, and both the calibration and prediction of the three scenarios were performed independently. We used the SLEUTH 3.0 beta_p01 module (http://www.ncgia.ucsb.edu/projects/gig/) and completed the model compilation and operation with the help of the software of Cygwin.

 

Fig 2

Technical flowchart of the study.