How much vegetation density in green infrastructure do people prefer?

Abstract

Green Infrastructure (GI) refers to the natural spaces in a city that improve urban ecology and bring ecosystem services to citizens (Kondo, Low, Henning, & Branas, 2015). These spaces include rain gardens, bioswales, and green roofs. A few studies have explored people’s preferences for and reactions to green infrastructure elements (Nassauer, Wang, & Dayrell, 2009; Sullivan & Lovell, 2006), but they do not explore the factors that influence this preference. One factor that may influence GI preference is vegetation density. Studies show that people prefer natural to urban environments (Velarde, Fry, & Tveit, 2007), but these studies mostly focus on coarse categorizations, such as nature vs. no nature, rather than varying amounts of nature (Sullivan, Frumkin, Jackson, & Chang, 2014). We do not know how much vegetation density is needed to achieve a moderate or high preference rating. Without this knowledge, we might create landscapes that people do not prefer, and people are unlikely to gain health benefits from places they do not like (Korpela & Hartig, 1996). This study aims to determine the nature of the relationship between vegetation density in GI photographs and preference. For example, would preference increase steadily as the density increases, or would it decrease when the density gets too high? To answer that, we acquired data from 3 different preference questionnaires where participants rated images in 1-5 scale. The participants are recruited through emails and crowdsourcing website. Using 152 photographs with varying densities of GI including street trees, bioretention, green roods, etc. , we ran curve estimation plots and regression analyses between the vegetation density index and the photographs’ average preference scores. The photographs contain different types of GI including street trees, rain gardens, bioswales, green roofs, retention ponds, etc. More than 1,000 participants rated the images. Our findings suggest that when vegetation density is low (0-25%), an additional increase in vegetation yields a dramatic increase in preference. Preference rates level off but do not decline as density increases to moderate and high levels (25-80%). We found that 45% vegetation density is needed to achieve a moderate preference value of 3. The results of this study are similar to those describing the dose-response relationship between tree canopy density and preference (Jiang, Larsen, Deal, & Sullivan, 2015). These results can be useful to designers seeking to create green infrastructure that people prefer. Limitations such as possible recruitment biases and GI types are discussed.