Evaluating Simulation Models - Measuring Landscapes

• A more robust suite of statistics needs to be assembled to measure landform morphology. Some possible types of parameters are:

  • Statistical moments of slope gradient and divergence (mean, variance, skewness, kurtosis)
  • Plots of gradient versus divergence (e.g. Roering et al., 1999)
  • Measures of channel network geometry: Basin shape, flowpath indirectness, etc. (Howard, 1971, 1990a,b) -- But NOT the usual Horton-Strahler statistics
  • Spatial autocorrelation and width functions
  • Drainage density as a function of critical divergence value

Previous slide

Next slide

Back to first slide

View graphic version

The use of a large suite of morphometric statistics to characterize landform shape was utilized by Howard to characterize river meanders (Howard, A. D., and Hemberger, A.T., 1991, Multivariate Characterization of Meandering: Geomorphology, 4, 161-186) and escarpment planform shape (Howard, A.D. 1995, Simulation modeling and statistical classification of escarpment planforms, Geomorphology, 12, 187-214).

The usual Horton-Strahler drainage basin statistics based upon stream order have proven ineffective in distinguishing stream networks with different shape or process suites. More effective parameters characterize the SHAPE of drainage basins (e.g., Howard, A.D. Simulation model of stream capture. Geological Society of America Bulletin, 82, l355-l376; Howard, A. D., 1990, Role of hypsometry and planform in basin hydrologic response: Hydrological Processes, 4, 373-385; Howard, A. D., 1990, Theoretical model of optimal drainage networks: Water Resources Research, 26, 2107-2117).