Green infrastructure (GI), a network of nature, semi-natural areas and green space, delivers essential ecosystem services which underpin human well-being and quality of life. Maintaining ecosystem services through the development of GI is therefore increasingly recognized by policies as a strategy to cope with potentially changing conditions in the future. This paper assessed how current trends of land-use change have an impact on the aggregated provision of eight ecosystem services at the regional scale of the European Union, measured by the Total Ecosystem Services Index (TESI8). Moreover, the paper reports how further implementation of GI across Europe can help maintain ecosystem services at baseline levels. Current demographic, economic and agricultural trends, which affect land use, were derived from the so called Reference Scenario. This scenario is established by the European Commission to assess the impact of energy and climate policy up to 2050. Under the Reference Scenario, economic growth, coupled with the total population, stimulates increasing urban and industrial expansion. TESI8 is expected to decrease across Europe between 0 and 5 % by 2020 and between 10 and 15 % by 2050 relative to the base year 2010. Based on regression analysis, we estimated that every additional percent increase of the proportion of artificial land needs to be compensated with an increase of 2.2 % of land that qualifies as green infrastructure in order to maintain ecosystem services at 2010 levels.

Dense and mixed land-use configurations are assumed to encourage high and prolonged activity levels, which in turn are considered to be important for the condition of urban neighbourhoods. We used mobile phone usage data recorded in Amsterdam, the Netherlands, as a proxy for urban activity to test whether the density in different forms of urban land use increases the level of activity in urban areas, and whether mixed land uses can prolong high levels of activity in an area. Our results indicate that higher densities correspond with higher activity levels, mixed land uses do indeed diversify urban activity dynamics and colocating particular land uses prolongs high activity levels in the evening hours. We proceed to demonstrate that mixed activity provisions and high urban activity levels coincide with urban neighbourhoods that are considered attractive places in which to live and work, while lower activity levels and markedly low activity mixes coincide with neighbourhoods that are considered disadvantaged.

In the majority of EU Member States, agricultural land is expected to decrease not only due to land-use changes in favour of urban expansion and afforestation but also to land abandonment processes. The knowledge on location and extent of agricultural land abandonment is relevant for estimating local external effects and adapting policy interventions. Currently, multi-level land-use models are able to capture determined processes of demand-driven redevelopment. However, land abandonment is much more difficult to capture because of its more ambiguous definition and the lack of data on its spatial distribution. This paper presents a method to explicitly model agricultural abandonment as a choice of disinvestment, which in turn is embedded in a utility-based land-use modelling framework that projects land-use changes for the EU and the UK. Validation exercises using observed spatial distribution of abandoned farmland show that the proposed method allows to model abandonment with acceptable accuracy.

Analyses of spatial interaction are to some degree plagued by uncertainty regarding the impact of spatially dispersed interaction masses within zones on travel times. In this paper, interaction-weighted travel times are computed from a matrix between regularly distributed points at fine resolution, and used together with secondary data to improve estimates of interaction weighted travel time based on commonly applied methods. The paper proposes a method for computing intra-zonal, interaction weighted travel times that is considerably less sensitive to spatial aggregation than existing approaches, and demonstrates that population-weighted centroids are to be preferred over geographically-weighted centroids.