Remote sensing-based characterization of land management and biophysical factors in grassland
Abstract
Land use and management are important factors influencing ecosystem functions, including the cycling of carbon (C) in plant/soil systems. Information about land use and management, needed to prioritize conservation efforts in managed grasslands of the Central Great Plains, can be obtained using remote sensing techniques, but this process is complex in grasslands because of the subtle class differences, large within-class variability, and complex seasonal changes in canopy spectral characteristics. In this study, time-series of remotely sensed data were used to derive vegetation index (VI) and image texture measures. The utility of these measures for classification of five managed grassland types was assessed using ANOVA and stepwise discriminant analysis methods. Image texture was found to improve the accuracy of classification by ∼13% over the use of VI alone. The optimal timing of data acquisition for classification with VI was found to be in April/May and in October; optimal timing for acquisition of texture was in June. Remotely sensed VI have been commonly used to model photosynthetic capacity and net primary production in ecosystems. Since VI theoretically assume canopy conditions of uniform geometry and greenness, seasonally variable management-induced changes in the grassland canopy can potentially influence the VI response and therefore the strength and stability of the model. This study examined the seasonal and inter-annual stability of the relationship between VI and photosynthetic capacity under both idealized and realized conditions. With regression analysis, the relationship between VI and field-measured estimates of photosynthetic capacity was established and evaluated. This work identified two types of management activity strongly influencing the stability of this relationship: (1) Conservation management, in which the vegetation is neither hayed nor grazed, results in accumulation of senescent canopy material and leads to lower than expected VI response; (2) Heavy grazing management leads to elevated levels of forb (non-grass species) cover in the canopy coupled with low photosynthetic capacity and high levels of bare ground, resulting in higher than expected VI response. When sites exhibiting these characteristics were removed, the relationship between VI and photosynthetic capacity was found to be stable seasonally and between years.
- Publication:
-
Ph.D. Thesis
- Pub Date:
- 2006
- Bibcode:
- 2006PhDT........42R