High-resolution images reveal rate and pattern of shrub encroachment over six decades in New Mexico, USA

TitleHigh-resolution images reveal rate and pattern of shrub encroachment over six decades in New Mexico, USA
Publication TypeJournal Article
Year of Publication2003
AuthorsGoslee S.C., Havstad K, Peters DC, Rango A., Schlesinger W.H.
JournalJournal of Arid Environments
Date PublishedAugust 1, 2003
ARIS Log Number139463

Encroachment of the shrub Prosopis glandulosa (honey mesquite) into semiarid grasslands is a serious concern in the southwestern United States, yet little is known about the long-term dynamics of the invasion process. We used 10 high-resolution aerial and satellite images taken from 1936 to 1996 to track population dynamics and spatial pattern of all P. glandulosa greater than 2m in diameter on a 75-ha area in southern New Mexico. Shrub cover and patch numbers increased from 1936 to the 1970s, then stabilized at 43% cover and 83 patches/ha. Individual patches were extremely persistent: 95% of the area occupied by shrub patches in 1936 was still occupied in 1996. Recruitment into the 2m size class was more variable: 0.6 to 5.2%/yr (mean 0.8% yr). Patch shape complexity increased from 1936 to 1983 as adjacent shrubs merged and then declined as those clusters filled in and became rounder. Spatial pattern of shrubs showed a distinct trend over time: strongly clustered in 1936 at lag distances up to 250m, then random arrangement at all scales; and, by 1983 the pattern was regular at lag distances greater than 100m. There was no clear relationship with precipitation. The use of remote sensing imagery allowed us to examine one site over time and revealed patterns in population dynamics and spatial pattern that would not have been visible otherwise. Comparison of field estimates collected in 2001 with 1996 image data suggest the canopy cover estimates were accurate, but shrub densities were seriously underestimated in the satellite photographs which do not show shrubs smaller than 2m diameter. As long as limitations of the imagery are understood, these methods can be applied over a larger and more heterogeneous area to examine environmental correlates of invasion success.