Spatial Modeling of Forest Foliar Nitrogen Concentration across the Southern US Region

spatial modeling;GIS;linear regression;stepwise regression;foliar N concentration

Mao, Zhiyan ; Hugh Devine, Committee Member,Montserrat Fuentes, Committee Member,Steven McNulty, Committee Member,Ge Sun, Committee Co-Chair,Siamak Khorram, Committee Co-Chair,Mao, Zhiyan ; Hugh Devine ; Committee Member ; Montserrat Fuentes ; Committee Member ; Steven McNulty ; Committee Member ; Ge Sun ; Committee Co-Chair ; Siamak Khorram ; Committee Co-Chair

University:North Carolina State University

关键词: spatial modeling;    GIS;    linear regression;    stepwise regression;    foliar N concentration;   

Others  :  https://repository.lib.ncsu.edu/bitstream/handle/1840.16/2377/etd.pdf?sequence=1&isAllowed=y

美国|英语

来源: null

【 摘 要 】

Foliar N concentration represents a primary link between C and N dynamics in forest ecosystems, and it is widely used as a critical parameter for ecosystem modeling. This study was designed to investigate the relationships between foliar N concentrations and a set of environmental factors, and to detect the spatial pattern of forest foliar N concentrations in the southern US region. A field observation foliar N concentration database was generated by combining literature search data with field study data. A spatial GIS database was developed from diverse data sources, which contained: climatic variables such as mean January temperature, mean July temperature and precipitation; soil variables such as soil organic matter and soil available water capacity; and total N deposition. Totally 104 data points were obtained including 61 for deciduous forests and 43 for coniferous forests after processing both foliar N concentration and GIS data. Various model fit approaches were employed for exploring the regression relations between foliar N concentrations and the selected environmental variables, including simple linear regression, variable-transformed linear regression and stepwise-based multiple regression. It was found that in this region, foliar N concentrations of deciduous forests were correlated primarily with mean July temperature, secondarily with mean January temperature, while foliar N concentrations of coniferous forests were correlated primarily latitude, secondarily with mean January temperature. A set of predictive equations were developed based on the regression analysis results and validated against previously reserved data points using split-sample approach. These equations could be applied across the region for spatially explicit estimation of forest foliar N concentrations.