Browsing by Author "Ojasalo, Amanda"

Climate change and urbanization are among the largest environmental challenges facing the world today. The role of vegetation in urban environments is substantial from the perspectives of climate, ecology, and human-wellbeing. Plant phenology plays a key role in the functionality and feedbacks related to these ecosystem services and the characteristics of urban phenology can considerably differ from rural areas due to Urban Heat Island (UHI), vegetation composition, hydrological changes, light pollution, and air pollutants. Several previous studies using coarse resolution remote sensing data have reported longer Growing Season Length (GSL) in urban areas compared to their rural counterpart and the UHI effect is generally considered as the main driver for these differences. However, urban phenology studies have not been implemented on a regional European scale and high-resolution remote sensing data is needed to understand the characteristics of heterogenous and sparse urban vegetation. Therefore, the objectives of this study were (1) to analyse GSL along the urban-rural gradients in 38 European capital cities using Copernicus HR-VPP phenology dataset on a 10-meter spatial resolution, (2) to analyse GSL along the urban-rural gradient between the 38 European capital cities, and (3) to find out how Land Surface Temperature (LST), land cover and dominant leaf type influence on the GSL variation. The GSL pattern along the urban-rural gradients were classified into six categories based on linear and quadratic fits. The results showed that the GSL along the gradient in European capital cities is highly variate. It shortens along the gradient in 8 cities and the urban GSL is longer in 26 cities when compared to the overall surrounding zone, contradictory to the general outcome of previous studies. The influence of LST, land cover and dominant leaf type was examined with a Geographically Weighted Regression (GWR) model which considers the spatial nonstationary of variables. The final GWR model variables included LST, the proportion of urban land cover above 30 % of sealed surface and the proportion of broadleaved trees which all had spatially varying and nonlinear influences on GSL. These results and the variate gradient patterns suggest that despite the significance of LST, GSL variation along the urban-rural gradient is more driven by changes in land cover and vegetation characteristics. Spatial modelling techniques are needed to understand these locally varying relationships. There are several potential methodological and site-related explanations for the divergent findings of this and previous studies. The key methodological difference is the better spatial resolution which improves the accuracy of GSL detection, agricultural land cover masking and urban area definition. Site-related explanations include the different background climate and vegetation types, urban vegetation composition and species, and urbanization intensity. In addition, several heatwaves took place during the study period potentially contributing to the early senescence of urban vegetation. This study highlights the need for a high-resolution remote sensing data when analysing urban vegetation phenology and provides new information about the complex dynamics of urban phenology in general level and in European capital cities. These results can be beneficial for developing sustainable cities where urban vegetation plays a key role in adapting and mitigating climatic, ecological, and societal challenges.