Date of Award
Doctor of Philosophy
The National Resources Inventory (NRI) is a longitudinal survey of non federal lands in the US. The objectives of the survey are to produce estimates for variables related to land use, land cover and soil erosion at the national and sub-national level. Three extensions of existing small area models are proposed for estimating soil erosion for counties;A transformed Fay-Herriot model is developed to estimate wind erosion for the counties in Iowa. A soil erodibility index is available from administrative records for each county and is used as the predictor. The response variable is the soil loss as recorded in the 2002 NRI. An iterative approach is proposed to obtain a calibrated small area estimator. The small area estimates and the standard errors are reported;A class of estimators based on local polynomial regression is proposed. The assumptions on the area level regression are considerably weaker than those of standard small area models. Both the small area mean function and the between area variance function are modeled as smooth functions of the area level covariates. A composite estimator that is a convex combination of the direct mean and the predicted mean is used as the small area estimator. The estimator is shown to be asymptotically consistent under mild regularity conditions. An approximation for the mean squared error based on Taylor linearization is developed;An estimation model is developed for the cover and crop management factor (C factor) that can be used for small area estimation for counties. The NRI data set contains a significant proportion of imputed values, where the unobserved values are determined by the sampling design. The variance due to the current imputation procedure is estimated using an explicit imputation model. An existing small area procedure is adjusted for the C factor to reflect the effect of imputed values and is applied to the NRI.
Digital Repository @ Iowa State University, http://lib.dr.iastate.edu
Mukhopadhyay, Pushpal, "Extensions of small area models with applications to the National Resources Inventory " (2006). Retrospective Theses and Dissertations. 1548.