Maintaining and Improving Biodiversity in Urban Centers: Landscape Spatial Design in Mexico City and New York City


 ContextLarge cities contain different sizes and distributions of green spaces in a sea of buildings and roads. This urban landscape establishes the habitat for different species that persist in cities.ObjectivesHow does this “archipelago” of habitat space function? How does the arrangement of green spaces affect plant and animal species' biodiversity and movement through this urban pattern?MethodsBy using Patch Analyst Metrics, we propose a novel method to analyze and improve the current spatial arrangement of green spaces using Mexico City and New York City, long-established urban areas.ResultsThe two cities differ in the number, size, and spatial distribution of green spaces. Frequency analysis suggests that Mexico City has a high number of large green spaces for native species conservation; but most of them are in one vast cluster of green areas at the south. In New York City large spaces are distributed along the whole territory, comprising most potential habitats, but it has much more small areas. This spatial analysis shows particular areas in which both cities have the potential to add connectivity among existing green spaces for dispersal of many taxa of plants and animals. ConclusionsMuch data is available on the potential dispersion through cities, but a better framework for understanding the existing distribution is needed for future landscape decisions. Results suggest ways that new urban areas can better increase plant and animal movement patterns.


Introduction
The presence of green spaces in large cities provides many ecological services that improve the quality of urban life and the health of the human population (Elmqvist et al., 2015;TEEB, 2010). Ecological sustainability of urban green spaces is strengthened by high biodiversity and complex food webs so that changes in physical and biotic conditions do not diminish ecological functioning. Urban biodiversity can be high with many species present, even in large, historical cities (Aronson, Patel, O'Neill, & Ehrenfeld, 2017).
The distribution of green spaces directly impacts the ecological community of organisms that live there (Gilpin & Hanski, 1991). For example, species move at different rates and have different generation times. Genetic studies (e.g., Johnson and Munshi-South 2017) show population isolation among even closely placed urban patches. Consequently, biodiversity in a new urban green space will not replicate other spaces immediately but will build as surrounding species immigrate. This may happen very slowly or not at all, depending on the rate of disturbance and the presence of barriers in the landscape matrix, in a classic metapopulation process. hardscape barriers and heights make functional distances in cities longer than in rural areas, even though the spatial scale is the same. This is similar to the "mountain passes are higher in the tropics" concept (Ghalambor, Huey, Martin, & T, 2006). Some urban barriers affect the height that animals or seeds must overcome and create shade, arti cial light, wind tunnel, and heat island stressors to species' movements. Together, these spatial forces make the ability of species to navigate and colonize in cities quite different from in rural areas.
The ndings of high urban biodiversity have been surprising as urban green spaces are often small and separated by this hardscape/infrastructure, which impedes the movement of many species and their propagules. The small area of many urban green spaces results in small population sizes that can lead to local extirpation by biotic or stochastic pressures (Collinge, 2009). These smaller areas have a relatively large edge/center ratio, creating an adverse effect on their quality, favoring species that can persist in those changed spatial conditions. These problems of sustaining urban green spaces have led to the consideration of generating corridors or additional "stepping stone" green parcels to increase biodiversity (Bennett, 2003;Collinge, 2009). These types of connections may be more easily designed in new cities, where biodiversity planning can be arrayed from the start into the spatial pattern of the landscape. In contrast, old cities are constrained to support biodiversity by the pattern of urban infrastructure around existing, often isolated, green spaces.
In existing cities, ecological constraints are framed by several forces that in uence the spatial distribution of habitats: the type of ecosystem in which the city was established; the culture that de nes uses of urban space from time of rst settlement to today; -history (growth rate, episodic disruptions) of the urbanization; the socio-ecological zoning of the urban area; the economic forces that mold urban development over the decades.
These multiple determinants of urban structure constrain the possibility of maintaining or improving biodiversity (Clifton, Ewing, Knaap, & Song, 2008). Ecological improvements, however, can occur if planning concepts are better married to ecological principles (Harris, 1984;Hobbs & Saunders, 1993;Parris et al., 2018;Zipperer et al., 2000). Our study frames a method to map existing green spaces to improve the design and management of urban green areas for ecological structure and function within these constraints. Biotic communities in cities are affected by a foundation of at least four spatial characteristics: 1. The number of green spaces ( gure 1a). Species diversity usually increases as the number and heterogeneity of areas increase, creating a mosaic of potential urban biota spaces. More areas buffer against stochastic extirpation in any one area. Also, different taxa require diverse niche spatial axes.
Consequently, more taxa can be found if more individual spaces are present.
2. Area of green spaces ( gure 1b). Larger areas may increase survivorship as each population usually can be more numerous, avoiding stochastic local extinctions. Larger areas also increase the probability that the parcels will contain soil, nutrient, water, and refuge conditions that can safely harbor species during unfavorable and changing climatic conditions. 4. Quality of green spaces ( gure 1d). The ability to support species' niches must be present. "Quality" is a metric that varies with each space's ability to provide the unique niche axes of each species. Urban areas so often have had past land uses that modi ed original conditions. The space now may be inadequate, even for regionally common species. Restoration science may not be able to remediate the space adequately to return to those conditions.
Many older cities have numerous green spaces, but this varies enormously among cities with different developmental histories (Fuller & Gaston, 2009). Similarly, many new urban greening projects have been designed over the past decades, but often without attention to the spatial interplay among new and old green spaces (Kemp, 2006). Here, we explore new methods to map and analyze the spatial conditions of urban green spaces using data from two very large, established cities, Mexico City and New York City, whose current forms and infrastructure began 400-500 years ago. We ask for each city, what is the pattern of existing green spaces and what practical actions can be made to secure existing biodiversity and improve biodiversity? This is of importance in a time when climate and sea levels are rapidly changing, and current species diversity and abundance are expected to change signi cantly (Grimm et al., 2008). New land management actions (addition of green spaces, or new management of existing green spaces, for example) may be needed to maintain or improve the current ecological functions. The analyses aim to establish a method that can be used elsewhere to reach these goals.

Methods
Mapping protocols of the two cities were different because the data sources were created at different times In both cities, those spaces smaller than 0.5 hectares were selected and eliminated since a 0.5 ha area could be the minimum size for many species' survival capacities and home range (Rudd, Vala, & Schaefer, 2002a).
The spaces featuring the Brooklyn Green-wood Cemetery, Floyd Bennett Field, and the green space between Hendrix and Betts Creeks were added by tracing and editing. Although these are not parks, they have landscape vegetation that is consistent with mapped parks. The natural protected area in the south of Mexico City also was included as a large bulk of green space. Then by exporting the Attribute Table of each shape le, the number of green spaces and their respective areas were obtained.
Each variable was divided into ve categories, which helped to analyze frequency patterns of green space size and the whole city's distribution. Geographically classes helped to visualize the green spaces' characteristics within each city. We emphasized differentiating small values of the variable by splitting more categories than Finally, the Total Core Area (TCA) variable, in this case we used a percentage of green space covering the total of the hexagon area. The rst three categories embrace up to 20% of the hexagon covered by green space, while the last category covers more than 50% of the hexagon.

Results
The two cities have contrasting patterns of green space that have been developed over the centuries, based on the geologic and topographic patterns of the landscape. In both cities, the existing green spaces are surrounded by the urban matrix, roads, residential districts, and other infrastructure, constrained from growing by urban needs and history.
In MXC ( gure 2a), substantial green spaces are in the southern section of the city. This is an area of mountainous topography that had small settlements during the pre-Colombian era. During the colonial period, settlement continued to be focused on the northern half, the lowlands. Consequently, much of the land in the south is a continuous forest and grassland mosaic occupying more than 90% of the green space areas.
However, in the past 70 years, new settlements have arisen throughout this southern sector (Graizbord & González Granillo, 2019). In the northern half, the vast majority of the 2,132 small scales (0.5-5ha) green areas are present, surrounded by the urban matrix. Even though there are thousands of small areas, these represent only 4% of the city's green space area (Table 1).
In NYC ( gure 2b), there are 22 extra-large (>100ha) green spaces scattered over the landscape, not concentrated in one area; these represent 62% of the total green area. The majority (472) of green spaces are small in NYC, but they represent 7% of the total area ( Table 1). The NYC extra-large green areas represent planning decisions based on geology (the land on the terminal moraine and the glacial outwash plain was low quality for agriculture and was subsequently used for parks such as Prospect, Green-wood, Forest, and Marine) and by political and social actions (e.g., Central and Van Cortland Parks) (Kieran, 1982;Schuberth, 1968). Political decisions reserved the smaller parks for local neighborhoods as the city grew from its original location in the southern tip of Manhattan.
The analyses based on our 500ha hexagon data sets catalog each cities' green space spatial characteristics help to evaluate the capacities of each city to keep habitat for native species and their dispersal capabilities.
Frequency gures of each variable helped to compare green spaces presence, number, and distance in both cities' whole area ( gure 3). The number (NumP) of green spaces appear to have a log-normal distribution.
The number of smaller green spaces per hexagon is more common than large ones in both cities. On the contrary, the mean patch size (MPS) frequency gures are contrasting in both cities. While NYC has a large number of hexagons with small areas, and many fewer hexagons with larger areas, in MXC, the number of hexagons with different sized green spaces is almost constant.
Consequently, these ndings relate to the distance between green space neighbors. When the number of spaces in a hexagon is greater, the distance must decrease. In NYC, the number of hexagons with different The nearest neighbor between any two green spaces helps to understand the potential for organisms' movement between the spaces in a city's layout. In MXC ( gure 6a), most green spaces are close, <200m, from the next one across the entire landscape. This is caused by a large number of small green spaces in the northern section and the vast contiguous green space that predominates in the city's southern part. In contrast, in NYC ( gure 6b), the lack of green spaces across Brooklyn and Queens, particularly ( gure 2b), yields a pattern of relatively vast distances between adjacent green spaces. This would correlate with a more di cult movement of organisms between adjacent green spaces. Also, there is a band of more widely spaced green areas across the east-west center of the Brooklyn-Queens geography. The data from NYC present several medium-sized "mainlands," capable of hosting many species (Gargiullo, 2007;Kieran, 1982). In NYC, green spaces are closely positioned to waterways such as Hudson and East Rivers, and Newark, New York, and Jamaica Bays. These have high biodiversity; Jamaica Bay is a vast short, 15-20m, for most bird-dispersed species (Hoppes, 1988;Howe & Smallwood, 1982). A 2km distance would move a diaspore among many urban parcels in our study, but 20m would typically land on the pavement, not a recruitment site. There is wide variation for wind-dispersal plants, but most seeds fall close, less than 100m to the mother plant (e.g., Vittoz  The invertebrate biodiversity is essential for urban food web structure. In NYC, bee species diversity in small urban gardens is signi cant, 54 species, although this is smaller than surveys in the larger NYC urban parks To continue to rebuild our historical cities' design using these two goals, we need to understand which information, data sets, and institutional changes are needed. The spatial analyses we report illustrates base maps upon which biodiversity improvements can be molded. Urban design decisions must include small-scale green spaces that accommodate to improve urban biodiversity, such as bioswales and infrastructure that comprise "green streets" (Beatley, 2017;Kemp, 2006).
The known biodiversity in cities can be mirrored by a diversity of design decisions to break down the "greygreen dichotomy" that is the caricature of recent urban planning (Parker, 2015).
Historical cities' constraints challenge improvement of urban biodiversity. It is better to have an ecological presence in decision making early in planning and design (Dunnett & Hitchmough, 2004). New city designs may be a more accessible template upon which to create biodiverse urban centers. Patterns that will allow sustainability of urban biodiversity have been explored theoretically and by site analyses (e.g., ( There are short-term actions to advance urban biodiversity, but our perspective must have a longer view, even years. This hexagon-based analysis supports decisions and policies for the addition of green areas with particular sizes, distances, and locations. New design decisions must demonstrate short term progress to secure public and political approval, but the ecological time scale of change is long and must outweigh the typical emphasis on annual or short-term goals. Declarations