Please help: Very small urban growth in SLEUTH

Dear professors and colleagues,

I am Jie Song, a Ph.D. student from the University of Florida. I am using SLEUTH for urban growth in a case study region in Florida. But I found that after coarse calibration the urban growth (the number of urban pixels) for the control years was almost the same. The same was true for test and prediction mode. I have attached relevant statistics as follows.

The following excerpt is from ave file:

run year index  sng sdg sdc og rt             pop               area
0 1995 1 0.40 0.00 0.00 12.00 0.00 163006.20 163006.20
0 2004 2 0.40 0.00 0.00 11.40 0.00 163113.60 163113.60
0 2007 3 0.40 0.00 0.00 10.60 1.20 163145.40 163145.40
0 2013 4 0.20 0.00 0.00 9.80 0.40 163212.80 163212.80

The following is from log0 file:

***********************LOG OF BASE STATISTICS FOR URBAN INPUT DATA********************
Year   Area
1974 162676.00 
1995 262011.00 
2004 299143.00
2007 311818.00 
2013 324697.00

It is evident that simulated area was very different from actual one. But the control_stats file gave a very high pop metric:

Run Product Compare Pop
0 0.00207 0.50266 0.98176

I have tried several possible methods that I could by my own  try best to came up. For example, I tried to change the resolution, change raster files to .tiff and to .gif in Photoshop, or change the value of slope cells into integers before exporting them as .gif files. But none of these work. I am currently totally being stopped at this point. Any comments and/or help would be extremely appreciated. I could be also reached at the following email address at: Many thanks again.

Best regards,

Jie Song

One comment on “Please help: Very small urban growth in SLEUTH

  1. Update:

    I would appreciate very much for Dr. Clarke’s help on this matter. I finally get the urban growth. I made two mistakes. First, I stretched the slope raster, so the average slope value is too high. Second, the default parameter values were not suitable for my case study region. So the growth is not obvious.



Leave a Reply