USGS Home PageGigalopolis Home Page
Home | About | Implement | Download | Discussion | Publications | Site Map
Home
About
Implement
functions & modes
data base setup
download and verify
calibrate
selecting ranges
      coarse
      fine
      final
      forecasting
predict
build products
Download
Discussion
Publications
Site Map
UCSB Home Page
 

Selecting coefficient values from a "derive coefficient values" run

This is an example of how to select a coefficient set from a derive coefficient values run of Demo_city using the demo200 input images.

Using file avg.log:

  1. For the STOP_DATE year (the last row in avg.log) find the diffus, breed, spread, slp_res, and rd_grav values
  2. Round the floating values to integers
  3. These values are used to initialize a prediction run of SLEUTH

Partial contents (some columns have been romoved) of demo200 derive coefficients calibration run:

run year index sng sdg og rt pop edges clusters cl_size diffus spread breed slp_res rd_grav %urban grw_rate leesalee grw_pix
0 1950 1 1.15 0.1 22.27 2.28 993.09 490.46 64.65 15.02 1.2 13.13 3.58 80.95 1.61 4.72 2.59 0.58 25.8
0 1970 2 1.55 0.1 38.09 4.17 1664.3 723.58 102.21 15.88 1.47 16.03 4.37 70.02 2.71 7.49 2.63 0.48 43.91
0 1990 3 1.65 0.14 63.07 3.73 2777.29 1033.16 141.46 19.21 1.79 19.55 5.33 52.38 4.47 11.19 2.46 0.41 68.59

(While we're here, notice that this file contains data showing how the metric values change over time. Several of these metrics are created for the control data and written to the LOG_# file if the LOG_BASE_STATISTICS flag in the scenario file is set to YES.)

Coefficient values used to predict growth:

coefficient type PREDICTION_*_BEST_FIT
dispersion 2
breed 5
spread 20
slope 52
road gravity 4