Woody Vegetation Cover, Attrition, and Patch Metrics over Eight Decades in Central Texas, United States

Abstract

Woody plant encroachment into rangeland ecosystems is a widespread and often unwelcomed circumstance affecting rangeland management decisions worldwide. In the rangeland management profession, varying philosophies have been employed in the management of woody plant encroachment. Following World War II, total eradication of woody plant cover was commonly practiced, eventually giving way to a mosaic approach that benefits livestock, wildlife, and recreational objectives, with cover increasing or even stabilizing in many areas. Cultural practices such as land fragmentation, lifestyles not dependent on agricultural income, and shifts in herbivory from predominately browsers to grazers may also be contributing factors. Modern image analysis technologies, such as object-oriented feature extraction and patch metric analyses, can shed light on past paradigm shifts through spectral and textural assessment of modern and historical aerial photography. In this study, woody plant cover and patch metrics were analyzed for a period spanning from 1938 to 1940 through 2018 in the Bennett and Sulphur Creek watersheds of the Lampasas Cut Plain of Central Texas. Object-based feature extraction was used to calculate woody plant cover, and Fragstats was used for landscape patch metrics. Total woody cover was compared with past stewardship paradigms. There was a net decrease of total woody plant cover from 1938 to 1940 through 2018, with variation in between as management paradigms shifted. A pattern of decline, regrowth, and stabilization, like that observed in other research, was noticed for the Bennett Creek watersheds but was not apparent in Sulfur Creek. Patch size/shape varied as well, but fractal patch complexity was relatively stable through time. Raster algebra analysis showed that < 10% of the initial woody cover from 1938 to 1940 remained in 2018, although total cover went through various expansion/reduction phases. This research underscores the importance of long-term datasets and locally based knowledge in the application and interpretation of historical management paradigms.