GIScience 2016 Short Paper Proceedings Semi-parametric Geographically Weighted Regression (S- GWR): a Case Study on Invasive Plant Species Distribution in Subtropical Nepal Qunshan Zhao 1 , Elizabeth A. Wentz 1 , Stewart Fotheringham 1 , Scott T. Yabiku 2 , Sharon J. Hall 3 , Jennifer E. Glick 2 , Jie Dai 4 , Michele Clark 3 , Hannah Heavenrich 3 Center for Geographical Information Science, School of Geographical Sciences and Urban Planning, Arizona State University, Tempe, AZ 85287, USA Email: {qszhao; wentz; sfotheri}@asu.edu Department of Sociology and Criminology, Pennsylvania State University, University Park, PA 16802, USA Email: {sty105; jeg115}@psu.edu School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA Email: {sharonjhall; mdclar10; hheavenr}@asu.edu Department of Geography, San Diego State University, San Diego, CA 92182, USA Email: jdai@rohan.sdsu.edu Abstract Geographically weighted regression (GWR) is a spatial statistical methodology to explore the impact of non-stationarity on the interaction between spatially measured dependent and inde- pendent variables. In this paper we use a semi-parametric geographically weighted regression (S- GWR) and demonstrate the effectiveness of the method on a case study on socio-ecological fac- tors on forest vulnerability. The case study is based on community forests in and around the buffer zone of Chitwan National Park, Nepal, a biodiversity hotspot that is being rapidly degrad- ed by exotic invasive plant species. This research integrated heterogeneous data sources such as observational ecological surveys, household interviews, and remotely sensed imagery. These data were utilized to extract and represent invasive plant species coverage, human activity inten- sity, topographical parameters and vegetation greenness indices. Research findings both demon- strate the S-GWR method and offer possible interventions that could slow the catastrophic spread of invasive plant species in Chitwan, Nepal. 1. Introduction Geographically weighted regression (GWR) is a spatial analysis method that uses the spatial dis- tribution of dependent and independent variables to specify non-stationarity to quantify the driv- ers of spatially dependent processes. GWR has been widely applied in application domains such as species distribution modeling in ecology and crime analysis in sociology (Foody 2004; Zhang and Song 2013). This paper presents a semi-parametric geographically weighted regression (S- GWR) method to model the factors that influence the spatial distribution of invasive plants in Nepal. We chose to use a semi-parametric model to include both parametric and non-parametric variables in the model specification. The S-GWR is implemented in a case study on the relation- ship between socio-ecological factors and invasive plant species in Chitwan, Nepal. Invasive plant species are considered as a serious global environmental threat to ecosystem structure and function by creating disturbances in ecosystems, reducing native species diversity and abundance, limiting human usage of ecosystems and triggering environmental changes. It is an emergent research topic to prevent invasive plant species spread and alleviate their influences in forest ecosystems. The case study aims to quantify the relationships between invasive plant species coverage and socio-ecological factors in community forests (CFs) in Chitwan, Nepal. We