A multivariate analysis of biophysical parameters of tallgrass prairie among land management practices and years.

Six treatments of eastern Kansas tallgrass prairie--native prairie, hayed, mowed, grazed, burned and untreated--were studied to examine the biophysical effects of land management practices on grasslands. On each treatment, measurements of plant biomass, leaf area index, plant cover, leaf moisture and soil moisture were collected. In addition, measurements were taken of the Normalized Difference Vegetation Index (NDVI), which is derived from spectral reflectance measurements. Measurements were taken in mid-June, mid-July and late summer of 1990 and 1991. Multivariate analysis of variance was used to determine whether there were differences in the set of variables among treatments and years. Follow-up tests included univariate t-tests to determine which variables were contributing to any significant difference. Results showed a significant difference (p < 0.0005) among treatments in the composite of parameters during each of the months sampled. In most treatment types, there was a significant difference between years within each month. The univariate tests showed, however, that only some variables, primarily soil moisture, were contributing to this difference. We conclude that biomass and % plant cover show the best potential to serve as longterm indicators of grassland condition as they generally were sensitive to effects of different land management practices but not to yearly change in weather conditions. NDVI was insensitive to precipitation differences between years in July for most treatments, but was not in the native prairie. Choice of sampling time is important for these parameters to serve effectively as indicators.