@article {KNZ00558, title = {Forest spread and phase transitions at forest-prairie ecotones in Kansas, U.S.A}, journal = {Landscape Ecology}, volume = {11}, year = {1996}, pages = {225 -235}, abstract = {The spread of gallery forest habitat into upland areas is of substantial interest to resource managers because such spread has many implications for the management of grassland and forest habitats. This study used a dynamic percolation model to examine the potential rates of spread or invasion of forest in eastern Kansas. Aerial photos taken 16 years apart at the Fort Riley training base were used to calibrate a spatially explicit contagion model of forest spread to interpolate and extrapolate the forest spread processes. Results fit the actual pattern of spread well, as measured by both visual inspection and a multiscale fractal measure of pattern. Comparisons to a long-term fire-exclusion experiment in Geary County, Kansas, and to the Konza Prairie also provided validation. Both the simulation and the 100-year Geary County series showed an interesting pattern of forest spread. Spread was slow and steady until about 20\% forest cover was reached, at which point the rate increased. We conclude that this self-accelerating response is due to spatial patterns created by the spreading forest that tend to accelerate the growth process after a critical point is reached. On the basis of theoretical study and experimental simulation of the percolation phase transitions, we suggest that fractal dimensions in a transient ecotone of binary mixtures (e.g., trees and grasses) should range between 1.56 and 1.8958, and the critical fractal dimension during ecotonal phase transitions should be 1.7951. This critical point of about 18.5\% forest cover that we predicted was close to the observed result and might represent a phase transition at the forest-prairie ecotone}, keywords = {LTER-KNZ, gallery forest}, doi = {10.1007/BF02071813}, author = {Loehle, C. and Li, B.L. and Sundell, R.C.} }