Grassland Productivity Simulation: Integrating Remote Sensing and an Ecosystem Process Model

Abstract

The heterogeneous nature of semi-arid grasslands in Canada creates significant challenges in monitoring grassland conditions, especially in light of increasing human activities and rapid environmental changes. It is thus imperative to develop a spatially-explicit tool to monitor and predict grassland productivity and to examine its responses to land-use and environmental change processes. In response to this need, we use a spatial BIOME-BGC model to estimate spatially distributed net primary productivity (NPP) for a mixed semi-arid grassland in Canada. Given the importance of the foliar C:N ratio in modelling terrestrial biochemical cycles and the ability of remote sensing in deriving spatially distributed data, a C:N ratio map is first produced from MODIS data which is then used to drive the spatial BIOME-BGC model. The simulated NPP driven by the fixed foliar C: N (i.e., C:N = 24.0) has an average of 112.53 g C m⁻² years⁻¹, while simulated NPP driven by MODIS-derived spatial foliar C:N has an average of 107.36 g C m⁻² years⁻¹. The latter better reflects the actual NPP on the ground which is 98.29 g C m⁻² years⁻¹. The results demonstrate that spatial foliar C:N can produce a more accurate simulation of grassland biogeochemical cycles thus improving NPP simulation accuracy.