Abstract: The influence of landscape pattern on ecological processes is scale-dependent. Reasonable determination of the scale is the key for studying landscape pattern and ecological process. Taking Anshun, a typical karst mountainous city in the central Guizhou, as the research object, we analyzed the spatial grain effect of landscape pattern at the levels of type and landscape with GIS and RS technology, to reveal the variation of different landscape pattern indices with the increases of grain size. Furthermore, we explored the optimal spatial grain size for the landscape ecological study in karst mountainous city. Spatial grain size significantly affected the landscape pattern indices except for Simpson’s diversity index (SIDI), Simpson’s evenness index (SIEI) and total area (TA). The responses of each index with increasing grain size could be divided into types of monotone decreasing, monotone increasing, fluctuating decline, and basic stability. Most of the landscape pattern indices had obvious inflection point at 20 or 100 m, with a “critical threshold”. For Anshun central planning area, the optimal scale range of landscape pattern analysis was 5 to 20 m, with the appropriate grain size being 20 m. Mountain green space occupied the dominant position in the karst mountainous urban green space system, and class area (CA), total edge (TE), edge density (ED) and the other four landscape pattern metrics had similar grain size response curves. In conclusion, the suitable granularity of urban landscape ecological research in karst mountain areas based on high-resolution images was 20 m.

Key words: phosphorus application, degraded grassland, nitrogen application, community composition, nematode abundance