Effects of spatial aggregation on forest landscape model simulation in northeastern China

Issues of scale and aggregation become important when large range of space and time scales is considered in landscape models. However, identifying appropriate levels of aggregation to accurately represent the processes and components of ecological systems is challenging. A raster-based spatially explicit forest landscape model, LANDIS, was used to study the effects of spatial aggregation on simulated spatial pattern and ecological process in Youhao Forest Bureau of the Small Khingan Mountain in Northeastern China. The model was tested over 500 simulation years with systematically increased levels of spatial aggregation. The results show that spatial aggregation significantly influences the simulation of fire disturbance, species abundance, and spatial pattern. Simulated fire regime was relatively insensitive to grain size between 30 m and 270 m in the region. Spatial aggregation from 300 m to 480 m dramatically decreased fire return interval (FRI) and increased mean fire size. Generally, species abundance and its aggregation index (AI) remained higher level over simulation years at the fine-grained level of spatial aggregation than at coarser grains. In addition, the simulated forest dynamics was more realistic at finer grains. These results suggest that appropriate levels of spatial aggregation for the model should not be larger than 270 m.