气候变化领域集成服务门户

Growing concern over climate and management induced changes to soil nutrient status has prompted interest in understanding the spatial distribution of forest soil properties. Recent advancements in remotely sensed geospatial technologies are providing an increasing array of data sources (e.g., LiDAR, hyper spectral imagery) relating to forest biophysical properties. While these data sources have the potential to improve spatial predictions of forest soil properties, considerable uncertainty exists regarding which remotely sensed (RS) indices are correlated to soil variability and what underlying pedogenic processes connect them. The main objective of this study was to identify and interpret RS indices that account for soil variability within a 2300 ha forested watershed. Redundancy analysis (RDA) and variation partitioning methods were used to uncouple the complexity of soil-environmental relationships. Thirty-two soil pedons were described, sampled, characterized and analyzed for 22 soil properties within the 0-50 cm soil depth interval. A suite of environmental covariates, comprised of LiDAR derived canopy metrics, land-surface and hydrologic terrain indices, broad-band remotely sensed indices (GeoEye-1), and narrow-band hyper-spectral indices (HyMap), were used as covariates in our RDA models. Principal coordinates of neighbor matrices (PCNM) was used to disentangle the contribution of spatial autocorrelation among sampling locations to the total variance explained by our RDA models. Two groups of soil properties were identified using discriminate analysis of principal components, with each soil property group (SPG) relating to different pedogenic processes occurring with the watershed (SPG1: organic matter metal cycling; SPG2: base-cation cycling). Our results show there was a relatively strong correspondence between soil properties and terrain/spectral indices; with 61% and 81% of the total variance explained by the first four RDA axes for SPG1 and SPG2, respectively. Variation partitioning analysis revealed that both SPG1 and SPG2 were most strongly related to terrain and canopy indices; although spectral indices were also important, especially for SGP2. Variation in the types of RS indices correlated to each SPG results from variation in the degree to which each environmental covariate relates to the pedogenic process (es) driving soil property development. The approach used in this study can help improve our understanding of soil spatial variability through identifying the most significant environmental covariates related to soil variation. Given the growing demands placed upon forest ecosystems (e.g., timber, recreation, carbon sequestration), improved knowledge of soil variability and the factors that affect the soil resource is essential to facilitate more effective forest management. Published by Elsevier B.V.