Record:   Prev Next
Author Anderson, Eric S
Title Spatial prediction of forest soil carbon: Spatial modeling and geostatistical approaches
book jacket
Descript 158 p
Note Source: Dissertation Abstracts International, Volume: 65-06, Section: B, page: 2720
Director: James A. Thompson
Thesis (Ph.D.)--North Carolina State University, 2004
Understanding the carbon cycle is one of the most difficult challenges facing scientists studying the global environment. Efforts to balance the global C budget have focused attention on terrestrial carbon storage in temperate ecosystems. Historically, most estimates of soil organic C (SOC) are based on means extrapolated from broad categories of soils and vegetation on a regional scale. Forest ecosystems of North America are of particular interest because of their ability to provide long-term C storage in both the forest vegetation and soils. Understanding spatial patterns in forest SOC may result in future development of techniques for conserving and enhancing terrestrial C pools. A series of studies were undertaken to explore a number of current issues that contribute to our inability to model SOC on a regional or landscape scale. Investigation into the spatial distribution of SOC occurred on a 32,500 ha forest ecosystem located entirely within the bounds of Hofmann Forest. Hofmann Forest is located in Jones and Onslow Counties of eastern North Carolina, USA. The objectives of the research were (i) to compile and compare a remotely-sensed high-resolution digital elevation model (DEM) to other commonly available DEM sources; (ii) to utilize landscape attributes and selected soil properties to develop and validate an explicit, quantitative, and spatially realistic model of SOC for a 32,500 ha forest ecosystem; (iii) to determine if the spatial scale of environmental variables affects model predictions; and (iv) to quantify SOC on an areal basis using the newly parameterized spatial models. The first issue at hand was to derive a highly precise, highly accurate DEM. A newly emerging technology, light detecting and ranging (LIDAR), was selected as a source of highly precise and accurate source of elevation data. Attempts to produce landscape scaled DEM lead to a series of issues that inhibited their production. To resolve these issues a series of geostatistical approaches were developed to reduce the LIDAR data sets while maintaining their precision and accuracy. A study was conducted to evaluate the effects of inverse distance weighted (IDW) and ordinary kriging (OK) linear interpolators on datasets of various levels of data reduction. (Abstract shortened by UMI.)
School code: 0155
Host Item Dissertation Abstracts International 65-06B
Subject Biogeochemistry
Agriculture, Soil Science
Alt Author North Carolina State University
Record:   Prev Next