TY - JOUR
T1 - Pan-European urban green space dynamics
T2 - A view from space between 1990 and 2015
AU - Xu, Fei
AU - Yan, Jingli
AU - Heremans, Stien
AU - Somers, Ben
N1 - Publisher Copyright:
© 2022 Elsevier B.V.
PY - 2022/10
Y1 - 2022/10
N2 - Creating and managing urban green infrastructure is an essential part of the sustainable urban development initiative promoted by the EU governments. To achieve a synergistic development of urban green spaces in various countries, acquiring spatially explicit information on urban green areas on a continental scale in Europe thus becomes necessary. Yet, the currently available European urban land use/cover products may fail to provide this comprehensive view as, due to their coarse spatial resolutions, they only pick up the large patches of green but fail to account for sparsely distributed urban green types. To address this issue, we applied a support vector regression model trained on samples collected from 50 European cities to urban green space mapping on a subpixel scale based on Landsat imagery. The model was validated in 10 other European cities, yielding a robust model performance (RMSE for the modeled green area fractions ranging from 0.09 to 0.16). Thus, we successfully mapped the spatial distribution of European urban green areas in 1990, 2000, and 2015. Given these results, we obtained three major conclusions: (i) previous studies might have underestimated the green space growth in Eastern European countries; (ii) most European countries have been building new settlements with more green spaces as the primary way to increase their urban greenness; (iii) the gap in urban green space per capita is widening among European countries, and Southern European countries are experiencing a decrease in access to urban green spaces.
AB - Creating and managing urban green infrastructure is an essential part of the sustainable urban development initiative promoted by the EU governments. To achieve a synergistic development of urban green spaces in various countries, acquiring spatially explicit information on urban green areas on a continental scale in Europe thus becomes necessary. Yet, the currently available European urban land use/cover products may fail to provide this comprehensive view as, due to their coarse spatial resolutions, they only pick up the large patches of green but fail to account for sparsely distributed urban green types. To address this issue, we applied a support vector regression model trained on samples collected from 50 European cities to urban green space mapping on a subpixel scale based on Landsat imagery. The model was validated in 10 other European cities, yielding a robust model performance (RMSE for the modeled green area fractions ranging from 0.09 to 0.16). Thus, we successfully mapped the spatial distribution of European urban green areas in 1990, 2000, and 2015. Given these results, we obtained three major conclusions: (i) previous studies might have underestimated the green space growth in Eastern European countries; (ii) most European countries have been building new settlements with more green spaces as the primary way to increase their urban greenness; (iii) the gap in urban green space per capita is widening among European countries, and Southern European countries are experiencing a decrease in access to urban green spaces.
KW - Continental-scale analysis
KW - Landsat
KW - Support vector regression
KW - Urban green space
UR - https://www.scopus.com/pages/publications/85136278214
U2 - 10.1016/j.landurbplan.2022.104477
DO - 10.1016/j.landurbplan.2022.104477
M3 - 文章
AN - SCOPUS:85136278214
SN - 0169-2046
VL - 226
JO - Landscape and Urban Planning
JF - Landscape and Urban Planning
M1 - 104477
ER -