Twenty-one years of crown area change in the Jasper Ridge Biological Preserve based on georeferenced multitemporal aerial photographs

A time series of large-scale aerial photographs of a 25 ha section of the J asper Ridge Biological Preserve (JRBP) in California was analysed for the q uantification of long-term changes in the orthogonally projected crown area s of a sample of mature canopy trees growing in closed woodland and open gr assland habitats. Crown area changes were computed on the basis of pairwise image comparisons for the years 1974, 1980, 1988, 1989, 1990, 1991, 1993 a nd 1995. Single frame image correction methods were assessed by comparing a high-resolution physical-system based method of photogrammetric orthorecti fication with a low-resolution method of georeferencing that involved the u se of ground control points (GCPs) for the derivation of best-fit statistic al transformations. The same crown digitization procedure was used for both methods, and a statistically significant difference was not found between the high-resolution and low-resolution measured crown areas. Crown area cha nges were not correlated with differences in sun position at the time of ae rial photo acquisition nor with differences in the off-nadir distances of e ach crown in the images being compared. The upper limit of digitizing error was 3.5% based on the mean coefficient of variation plus one standard devi ation computed from repeat digitization measures of each crown. Crown area changes exceeding 3.5% in the pairwise image comparisons were attributed to actual crown growth or crown reduction. Expressed as a percentage of the original crown area, the mean annual rate of change for each pairwise sample period was 4.9% yr(-1). In absolute term s, the mean annual rate of crown area change over the 21-year period 1974-1 995 was 1.4m(2) yr(-1). The long-term mean annual crown growth rate of the clustered woodland trees (1.3m(2) yr(-1)) was not significantly different f rom the isolated grassland trees (1.5m(2) yr(-1)). The mean annual rates of change for the periods 1974-1980, 1980-1989 and 1991-1993 were similar to the long-term average. Unusually low mean growth rates characterized the dr ought years of 1989-1990 and 1990-1991, while the relatively wet period of 1993-1995 exhibited the most dramatic increase in crown area at a mean rate of 3.0m(2) yr(-1).