Title :
Predicting the potential distribution of bamboo with species distribution models
Author :
Jin, Jiaxin ; Jiang, Hong ; Xu, Jianhui ; Peng, Wei ; Zhang, Linjing ; Zhang, Xiuying ; Wang, Ying
Author_Institution :
Int. Inst. for Earth Syst. Sci., Nanjing Univ., Nanjing, China
Abstract :
Using history climate data and two representative climate change scenarios, we predicted the potential distribution of bamboo in China from 1961 to 2099 based on specie distribution models. Through evaluating the impact of presence-only, true-absence and pseudo-absence data on SVM models accuracy on the potential distribution of bamboo during 1981-2000, we found that the two-class SVM using presence and pseudo-absence data showed the finest performance in forecasting potential distribution of bamboo. The prediction results of spatial pattern and inter-annual variation of potential distribution of bamboo under history and future climate showed that, the potential distribution of bamboo increased by 91500 km2 from 1961 to 2000. In climate change scenario B1, the potential distribution area increased 2433 km2 per year. In A2 scenario, the annual increment of potential distribution area was 13825 km2. Furthermore, the potential distribution of bamboo showed a northward migration obviously in both history climate and future scenarios.
Keywords :
climatology; support vector machines; vegetation; AD 1961 to 2099; China; SVM model; bamboo potential distribution; history climate data; interannual variation; presence-only data; pseudo-absence data; spatial pattern; species distribution model; true-absence data; Biological system modeling; Meteorology; Predictive models; Smoothing methods; Support vector machines; Bamboo; Global change; Pseudo-absence data; Spatial prediction; Species distribution models;
Conference_Titel :
Geoinformatics (GEOINFORMATICS), 2012 20th International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
978-1-4673-1103-8
DOI :
10.1109/Geoinformatics.2012.6270307