**6. Discussion**

The engineering design process by Nigel Cross [91], points to the importance of fully understanding the "problem definition" and to develop a "conceptual solution". Following this, clear functions of the object to be designed have to be defined. For almost a decade, Seattle and Portland worked hard on these two tasks; the keystone of green street design. The SEA street project in Seattle served as the initial credible evidence of the effectiveness of green streets for stormwater management. Experimentation determined the features of different GI technologies (ponds, bio swales, trenches, etc.), and performance rates. Philadelphia had a much shorter problem definition period and conceptual process, given that the technologies had already been defined before.

The large-scale analysis serves to determine the overall demand of GI (e.g., in the watershed) and some relevant citywide characteristics for the design, such as precipitation, imperviousness rates, infiltration rates, or sewer performance in different city zones. Conversely, demands for traditional street designs (e.g., traffic demand), are not accountable for the entire city, but rather estimated more locally (e.g., corridor). Recall that traditional design practices usually consider traffic volumes and infrastructure capacity (for different transportation modes), road safety, aesthetics, or other variables, which can be resolved at local/corridor scale.

The meso-scale stage is the least developed of the three stages. Worthy of note are the contrasting points of view of interviewees regarding the most suitable type of road (locations) for constructing green streets. Should green streets be located on low-traffic roads, busy streets, or both? Streets with low volumes of traffic seem to be the right location for GSI, since there is more space available for reallocation. In general, such streets account for the vast majority of city road networks. This type of street in the US is characterized by generous rights of way [92], by oversupply of parking [93], and by a desire for traffic calming to enhance livable areas. In addition, the literature suggests that green infrastructure should be located close to people, as benefits can be attained by mere proximity to green infrastructures [31]. On the other hand, interviews with the Deputy Water Commissioner of the Philadelphia Water Department and the director of environmental services (DES) on Gresham, Oregon [94] provided reasons to locate green infrastructure in busy corridors. The DES director pointed out that locating the GSI on busy roads would do more to reduce the discharge of pollution into rivers than locating it on quiet streets since most pollutants come from vehicle operation. In addition, people tend to walk along busy roads, highlighting them as places where people would benefit most from green streets [94].

A concern in the lower-scale analysis deals with developing a standardized manner for finding spaces within existing consolidated streets for GI. Different strategies become feasible depending on the hierarchy of roads selected in Stage II. These include the narrowing of vehicle lanes, reclaiming underutilized pavement, conversion of borders, furnishing zones, and traditional landscape areas, and eventually substituting parking boxes to function as GI. Overall, the interviewees agreed that it is important to maintain (or affect at the minimum) the street performance and services provided by the street before the addition of GI. Tackett [95] reinforces this idea when she suggests that an implicit design rule should be to ensure a similar service for motorists as in original designs while accommodating infiltration swales. One of the most sensitive topics is parking. Any change in the amount of space devoted to parking could generate opposition to green streets. The principle this implies is that GI should be built "at the edges of the right-of-way."

Street design has been experiencing ground transformations in recent years. One of these changes is due to the evolution of street functions; a process that began in the transportation field with a broader understanding of the street, beyond movement for cars. Under a new mobility paradigm presented by Banister [96], he supports the idea that a much broader notion of the street is needed, where streets are no longer considered mere roads but also as spaces for people, active modes, and public transport. In the environmental field, little has been reported in the context of street design with regards environmental functions. This paper provides empirical evidence of the effect in street design of the environmental function of the street introduced in the past by Hui et al. [97] and other authors [98].

It is important to remark that neither multi-modal street designs nor green streets come from the self-evolution of traditional street design. Both movements, multi-modal streets and green streets, succeeded quite independently in breaking the rigid code-based, transportation-focused street design. Both movements proved that a well-defined, understandable, and sharply focused workhorse is a strength that increases power and that allows goals to be reached more easily. This workhorse is safety for complete streets and, in this case, stormwater managemen<sup>t</sup> for green streets. Transportation officials and policymakers, who are more concerned with the right-of-way, are more likely to be motivated to change if the problem is framed as one that the city or society cares about: safety and water quality.

GI provision in several US cities has been rooted in stormwater management, especially in those with combined sewer overflows. For this initial group of green street implementations in US cities, the workhorse was stormwater management. Since new GI in cities can produce other benefits, the workhorse can potentially involve many other aspects, such as public health (improved air quality), property damage prevention (flood control), or the reduction of heat wave mortality (for urban cooling). In terms of design, the citywide analysis and meso-scale step are required to define the overall amount of GI required and to efficiently allocate that infrastructure within the city.

Paradoxically, the insertion of green infrastructure within the ROW, instead of harming or diminishing the movement function, might strongly benefit pedestrians and cyclists. These modes, also known as active modes, are important within the urban transportation strategy for sustainability. Evidence suggests that there is a causal effect between tree planting and cyclist satisfaction [99,100]. Another transportation-related benefit provided form bio-swales installed next to intersections, is that the crosswalk distances for pedestrians can be shortened, reducing risk exposure, while slowing traffic down.

GI in streets provides many other benefits beyond stormwater management, making green street programs an attractive option in working towards sustainability. In practice, cities in the US might still have a partial or incomplete grasp of the concept, beyond stormwater management. In other countries, the lack of incentives and dissemination to policymakers can be reasons for the limited implementation of green streets in many other countries. The case of Philadelphia shows that a mono-target green streets program (for stormwater management) is not actually accurate when evaluating a massive green streets program, as long as many other important benefits of GI are not considered. This paper contributes in the practical scope by summarizing these benefits (addressing other environmental functions of green streets) and expanding on the practical design process.
