University of Twente Student Theses

Login

Investigating spatial bias in citizen science phenological data

Qiu, Xinyi (2020) Investigating spatial bias in citizen science phenological data.

Link to full-text:https://ezproxy2.utwente.nl/login?url=https://library.itc.utwente.nl/login/2020/msc/gfm/qiu.pdf
(only accessible for UT students and staff)
Abstract:Volunteered phenological observations collected by citizen scientists are an important source of information for phenological studies. These observations are a type of volunteered geographic information (VGI). VGI has considerable value to various scientific research but contains inherent limitations. In phenology, the spatial bias in the distribution of observations may bring uncertainty to the representativeness of the observed geographic phenomenon. Spatial bias with respect to the uneven spatial distribution of observations is the focus of this study. The objective of this study is to identify and quantify the causes of spatial bias in sampling design and the influence on the spatial pattern of collected data. The variation in volunteer behavior and social-economic background which influence data collection is assessed through a statistical model. A point process model is adopted to model the relationship between observation intensity and a series of spatial covariates. The model primarily focuses on the first-order property of a point process – intensity, which only influenced by spatial covariates. The further analysis of residuals enables the visual interpretation of an unusual spatial pattern. The data is extracted from a national phenological network in the USA from 2003 to 2019. Seventeen geographic variables (human population density, road density, 15 different land cover classes) are included in the model. In general, road density and human population density are demonstrated significantly affect the collection of observations. Also, some particular land cover types show a higher probability for volunteers to make observations, such as the mixed forest of deciduous and evergreen trees, open space, and high-medium intensity developed areas. Overall, point process modelling shows a useful framework to analyse the spatial bias in VGI. The results demonstrate the effects of such bias, which can provide guidance for future volunteered data collection.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:https://purl.utwente.nl/essays/85208
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page