Browsing by Subject "Vegetation mapping"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Senescent vegetation ground-cover estimates from corrected scene brightness measurements(Texas Tech University, 1993-05) Sanden, Eric MRemote sensing techniques have been successfully used to estimate live, green vegetation parameters. The estimations of senescent vegetation with remote sensors, however, have been less successful due to the reflectance characteristics of dead and dormant vegetation. The goal of this study was to develop a remote sensing vegetation index to estimate senescent vegetative ground cover. Spectral reflectance measurements were collected on a rangeland site with a hand-held radiometer, under a variety of vegetation and soil conditions. The spectral measurements were used to develop senescent vegetative ground-cover prediction models, based on scene brightness. These models gave relatively accurate estimates during certain sample periods, with r^ values ranging from 0.52 to 0.65. Under heterogeneous plant canopy conditions, with high total ground coverage and when a large senescent vegetation component was present, the model accuracies declined substantially, with r^2 values ranging from 0.003 to 0.29. An additional nuisance variable affecting the scene brightness of rangeland scenes is the variation in soil moisture content of the sample plots. A correction factor was designed to standardize scene brightness factors with respect to soil moisture content. Alternative senescent ground-cover prediction models were developed for all sample periods, incorporating the corrected scene brightness factors. The correction factor, however, generally had a negative influence on the accuracy of the senescent groundcover prediction model. A multivariate approach was conducted to determine if additional independent variables would increase the effectiveness of the prediction model. The best 3-variable model included uncorrected scene brightness factors, actual average green ground-cover estimates, and perpendicular vegetation index values as the independent variables The multivariate model yielded only slightly higher accuracies in all but 1 of the sample period data sets. Aerial video data, similar to the data collected with the hand-held radiometer, were obtained during the last sample period to investigate the methodology of using this cover estimation approach with aerial remote sensing data. Although the aerial video data set provided a model with a higher r^ value than the corresponding ground-level model, interpretation of the results was difficult due to the overall low model accuracies and the non-significance of the ground-level model. The procedures for estimating senescent ground cover with scene brightness factors, developed in this study, appear relatively accurate under certain vegetation conditions. The best model results were obtained under conditions of low total vegetative ground coverage, with a small senescent vegetation component.