Biogeosciences Discuss., 10, 4225-4270, 2013
www.biogeosciences-discuss.net/10/4225/2013/
doi:10.5194/bgd-10-4225-2013
© Author(s) 2013. This work is distributed
under the Creative Commons Attribution 3.0 License.
Review Status
This discussion paper has been under review for the journal Biogeosciences (BG). Please refer to the corresponding final paper in BG.
Effects of vegetation heterogeneity and surface topography on spatial scaling of net primary productivity
J. M. Chen1,2, X. Chen1,3, and W. Ju1
1International Institute of Earth System Science, Nanjing University, Nanjing, Jiangsu 210093, China
2Department of Geography, University of Toronto, Ontario, M5S 3G3, Canada
3State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China

Abstract. Due to the heterogeneous nature of the land surface, spatial scaling is an inevitable issue in the development of land models coupled with low-resolution Earth system models (ESMs) for predicting land-atmosphere interactions and carbon-climate feedbacks. In this study, a simple spatial scaling algorithm is developed to correct errors in net primary productivity (NPP) estimates made at a coarse spatial resolution based on sub-pixel information of vegetation heterogeneity and surface topography. An eco-hydrological model BEPS-TerrainLab, which considers both vegetation and topographical effects on the vertical and lateral water flows and the carbon cycle, is used to simulate NPP at 30 m and 1 km resolutions for a 5700 km2 watershed with an elevation range from 518 m to 3767 m in the Qinling Mountain, Shaanxi Province, China. Assuming that the NPP simulated at 30 m resolution represents the reality and that at 1 km resolution is subject to errors due to sub-pixel heterogeneity, a spatial scaling index (SSI) is developed to correct the coarse resolution NPP values pixel by pixel. The agreement between the NPP values at these two resolutions is improved considerably from R2 = 0.782 to R2 = 0.884 after the correction. The mean bias error (MBE) in NPP modeled at the 1 km resolution is reduced from 14.8 g C m−2 yr−1 to 4.8 g C m−2 yr−1 in comparison with NPP modeled at 30 m resolution, where the mean NPP is 668 g C m−2 yr−1. The range of spatial variations of NPP at 30 m resolution is larger than that at 1 km resolution. Land cover fraction is the most important vegetation factor to be considered in NPP spatial scaling, and slope is the most important topographical factor for NPP spatial scaling especially in mountainous areas, because of its influence on the lateral water redistribution, affecting water table, soil moisture and plant growth. Other factors including leaf area index (LAI), elevation and aspect have small and additive effects on improving the spatial scaling between these two resolutions.

Citation: Chen, J. M., Chen, X., and Ju, W.: Effects of vegetation heterogeneity and surface topography on spatial scaling of net primary productivity, Biogeosciences Discuss., 10, 4225-4270, doi:10.5194/bgd-10-4225-2013, 2013.
 
Search BGD
Discussion Paper
    XML
    Citation
    Final Revised Paper
    Share