Indicators
This project is focused on developing baseline indicators and related maps for several UrbanShift cities. Selected indicators are in four themes:
- Biodiversity (BIO)
- Land protection and restoration (LND)
- Greenspace access (GRE)
- Climate mitigation (GHG)
| Theme | Indicator core code | Indicator name | Related indicators | Status | Notebook |
|---|---|---|---|---|---|
| Biodiversity | BIO-1 | Natural Areas | SICB-1 | link to notebook | |
| Biodiversity | BIO-2 | Connectivity of ecological networks | SICB-2 | link to notebook | |
| Biodiversity | BIO-3 | Biodiversity in built-up areas (birds) | SICB-3 | link to notebook | |
| Biodiversity | BIO-4 | Change in number of vascular plants | SICB-4 | link to notebook | |
| Biodiversity | BIO-5 | Change in number of birds | SICB-5 | link to notebook | |
| Biodiversity | BIO-6 | Change in number of arthropods | SICB-6 | link to notebook | |
| Land protection and restoration | LND-1 | Permeable areas | SICB-10 | link to notebook | |
| Land protection and restoration | LND-2 | Tree cover | SICB-11; SDG 15.1.1 | link to notebook | |
| Land protection and restoration | LND-3 | Change in vegetation and water cover | SDG 6.6.1 | link to notebook | |
| Land protection and restoration | LND-4 | Proportion of natural areas restored | SICB-7A | link to notebook | |
| Land protection and restoration | LND-5 | Number of habitat types restored | SICB-7B | link to notebook | |
| Land protection and restoration | LND-6 | Protected areas | SICB-8 | link to notebook | |
| Land protection and restoration | LND-7 | Protection of Key Biodiversity Areas | link to notebook | ||
| Land protection and restoration | LND-8 | Built-up Key Biodiversity Areas | link to notebook | ||
| Greenspace access | GRE-1 | Recreational Services: Area of recreational space per capita | SICB-12 | link to notebook | |
| Greenspace access | GRE-2 | Built-up area that is open space for public use | SDG 11.7.1; C4F-3.1 | link to notebook | |
| Greenspace access | GRE-3 | Proximity to Parks: population with access to public open space within walking distance (400m) | SICB-13; C4F-3.2 | link to notebook | |
| Greenspace access | GRE-4 | Population with threshold level (10%+) of tree cover within walking distance (400m) | C4F-3.3 | link to notebook | |
| Climate mitigation | GHG-1 | Level and share of GHG emissions by pollutant and sector | SDG 13.2.2; C4F-5.1 | [link to notebook] | |
| Climate mitigation | GHG-2 | Average annual carbon flux from trees per hectare of city area | C4F-5.2 | link to notebook |
Most of the indicators in this category are described in the Singapore Index on Cities’ Biodiversity and can be used for assessment of current efforts and to identify areas of action to prioritize in future efforts. The complete methodology for computing Singapore Biodiversity Index is provided in this publication. The data are drawn from public, global datasets published by reputable organizations. In many cases, cities will have access to local data that are of higher quality or are more specifically suited to local contexts and needs than what we can provide from global data.
Percent of land that is within natural area land classes
Natural areas support biodiversity by providing habitat. They also provide human beings with ecosystem services. The portion of the total city area that is close to a natural state thus provides information both about a city’s biodiversity and about the benefits provided by biodiversity.
Natural ecosystems are defined as all areas that are natural and not highly disturbed or completely human-altered landscapes. Examples of natural ecosystems include forests, mangroves, freshwater swamps, natural grasslands, streams, lakes, etc. Parks, golf courses, cropland, and roadside plantings are not considered natural. This indicator is calculated as the percent of natural area within the city boundary: (Total area of natural, restored and naturalized areas) ÷ (Area of city) × 100%. We calculated this indicator using the ESA WorldCover 10 m 2020 V100 land-classification map. We included as natural area all land classified as trees, shrubland, grassland, herbaceous wetland, mangrove, or moss and lichen.
In general, contiguous habitat benefits biodiversity better than habitat that is subdivided by roads, buildings, and other built infrastructure. Connectivity of habitat patches mitigates the effects of habitat fragmentation by allowing animals to access more habitat without having to cross inhospitable terrain. The fragmentation of natural areas affects different species differently. For example, a road might not be a barrier for birds but it can seriously fragment a population of arboreal primates. While consideration of dispersal ability and habitat requirements are important in the management of particular species, the Singapore Index adopts a generalist approach to quantifying connectivity based purely on patch geometry.
The method involves a two-step process: calculating the effective mesh size, followed by coherence that will normalize for the size of the city. Step 1: calculate the effective mesh size (EMS): EMS = 1 / Atotal (AG12+AG22+...+AGn2) Where: Atotal is the total area of all natural areas AG1 to AGn are the sizes of each group of connected patches of natural area that are distinct from each other n is the total number of groups of connected patches of natural area. AG1 to AGn may consist of areas that are the sum of two or more smaller patches which are connected. In general, patches are considered as connected if they are less than 100m apart. This equation was derived from Deslauriers et al. (2018).
Step 2: calculate the coherence: Coherence = EMS / Atotal We calculated this indicator using the UMD GLAD Landsat Analysis Ready land cover classification map. We included as natural area all land classified as trees, shrubland, grassland, herbaceous wetland, mangrove, or moss and lichen.
Percent of birds observed in all areas of the city that were observed in built-up areas
Cities often include large amounts of built-up land and brownfield sites. These areas are not typically thought of as high-quality habitat, but they can support some species. This indicator reflects the ability of built-up areas to support biodiversity, using birds as an indicator group. Built-up areas include impermeable surfaces like buildings, roads, drainage channels, etc., and anthropogenic green spaces like roof gardens, roadside planting, golf courses, private gardens, cemeteries, lawns, urban parks, etc.
The indicator is calculated as: (Number of bird species observed in built-up areas)/(number of bird species observed in all areas of city). For built-up areas we used the ESA WorldCover land-use classification for built-up land. For bird species number, we estimated the saturation levels of species-area curves as described for indicators BIO-4, BIO-5, and BIO-6, using research-grade observations of birds in the years 2016-2021 in the iNaturalist database (accessed through the Global Biodiversity Information Facility) that occurred on built-up land. Note that because we lack reliable data on which species are native, our estimates probably include introduced species.
Number of observations of vascular plants, birds and arthropods (2016-2021)
Direct measurement of biodiversity is important, but it would be impossible to assess directly the diversity and health of all wild populations in a city. We therefore provide diversity data on the three indicator taxa chosen by the Singapore Index as proxies for overall biodiversity: birds, arthropods, and vascular plants.
- Birds (the Aves) are one of the most visible species groups in urban areas, and they are often considered as a key indicator of environmental quality. They are well studied by academics and amateur naturalists worldwide, they are sensitive to environmental and habitat changes, and they are comparatively easy to observe and count. The Singapore Index includes the change in the number of bird species over time.
- Arthropods (the Arthropoda) include insects, spiders, crustaceans, and other animals with exoskeletons and jointed limbs. They are important actors in terrestrial ecosystems, driving critical processes such as pollination, food-web relations, and nutrient cycling.
- Vascular plants (the Tracheophyta) are plants with vascular tissues, which conduct water, minerals, and the products of photosynthesis throughout the plant. In contrast to nonvascular plants (for example, the mosses), vascular plants can grow tall and in locations with little moisture on the surface of the soil. Vascular plants include more than 90% of the earth’s vegetation.
The Singapore Index recommends reporting the number of native species in each of these groups in regular time intervals. A local data-collection process would typically involve partnering with a local research institution to carry out a species survey. This survey would be repeated every few years; the indicator is the change in time from a baseline level to the current year. Here we provide baseline species counts estimated from crowdsourced species-occurrence datasets, but this method has several important limitations.
Biologists often interpret species-survey data by examining the number of species found as a function of effort, plotted as a species-accumulation curves (SAC). The SAC plots the cumulative number of species recorded as a function of sampling effort (i.e. number of individuals collected or cumulative number of samples). At the start of a species survey, the total number of species found typically grows quickly with every unit of effort. After some time, however, effort expended yields more and more species that have already been found earlier in the survey—and the total number of species grows more slowly per unit of effort. A plot of species number vs effort will come to a plateau, and this saturation level is a common estimator for the number of species in an area, as shown in the figure below (source).
We estimated species richness by estimating the saturation level of species-saturation curves using individual crowdsourced observations as our effort unit. The estimated counts are the mean saturation levels of 100 species-saturation curves, each generated by randomizing the order of all research-grade iNaturalist observations of the Aves, Tracheophyta and Arthropoda taxa in the years 2016-2021. Only species’ observations with enough data to generate species-saturation curves are considered for estimating species’ richness.
We use the SAC curve-fitting approach because it is impossible to conduct an exhaustive survey of all species present even in the indicator taxa. However, our reliance on crowdsourced species observations introduces serious limitations, mostly stemming from an inherent bias toward observations in heavily populated or frequently visited locations.
Note also that the Singapore Index Indicators 4-6 require calculation of changes in species number. Our species-number estimates are based on the most recent five full years of observation data, and data from more than five years ago are sparser. We therefore can only provide baseline species numbers. Calculation of these indicator scores will require revisiting species-number estimation in several years’ time.
Percent of land area that has permeable surface
Impervious areas are areas in which pavement or other barriers to water movement prevent water from percolating into the soil. As climate change in many places will change precipitation regimes, many cities with large areas of impervious surface will experience high peaks in water flow and flooding damage to infrastructure and natural areas. This type of flooding also lowers the quality of receiving waters, impacting the biodiversity of aquatic and marine systems.
This indicator is calculated by measuring the proportion of all permeable areas to total terrestrial area of city: (Total permeable area) ÷ (Total terrestrial area of the city) × 100%
Percent of land area that has tree cover
Trees provide numerous services to cities: they provide cooling, improve air quality, store carbon, reduce noise pollution, and regulate the water cycle. Trees also provide habitat for birds, insects and mammals, and generally improve local ecosystem health.
This indicator calculated is based on tree canopy cover data from Trees in Mosaic Landscapes. The trees included may be planted or naturally occurring. The formula is: (Land area under tree canopy) ÷ (Total terrestrial area of the city). This indicator differs from Cities4Forests indicator GRE 1.4 in three ways: it is calculating the mean tree cover (rather than the percent of pixels with any tree cover), it is calculated over the entire area of interest (not just builtup areas within it), and it counts the areas with tree cover (rather than those without).
Net increase/decrease in area of vegetation and water cover between 2016 and 2021 as percent of 2016 area with vegetation and water cover
Vegetation and water cover provide many ecosystem services to cities, including groundwater accumulation, flood management and temperature moderation. They are also critical prerequisites to habitat.
This indicator uses a trendline of the measurements of spectral indices from 2016-2021 to estimate the change in the presence of vegetation and water at points within each city. We use 10-meter Sentinel-2 data, as accessed on Earth Engine and cloudmasked, to calculate spectral indices associated with vegetation (Normalized Difference Vegetation Index, NDVI) and water (Normalized Difference Water Index, NDWI) and then create annual greenest/bluest pixel mosaics for each year. We calculate trendlines of greenness and blueness for each pixel over the six year period and flag pixels with an average annual change (slope) of at least (-)0.03 in its index value. We then mask these trend layers by layers with all pixels that meet the threshold for vegetation (NDVI of at least 0.4) or water (NDWI of at least 0.3) in at least one year so as to exclude pixels that are not likely vegetation or water in any year. Next we separate these change layers into gain and loss layers, combine the separate vegetation and water layers to produce one gain and one loss layer, and apply 30-meter resolution reductions to count the number of pixels of loss and gain for either vegetation or water and in each city and sub-city area. We subtract the count of loss pixels from the count of gain pixels to produce a count of net change in vegetation and water pixels. To normalize this value so as to represent a percent increase or decrease in vegetation and water area, we divide it by the total number of 30-meter pixels that met the index value threshold for vegetation or water in 2016 for the same area of interest.
Note that this indicator measures both change in the area of vegetation/water cover as well as change in the strength of the vegetation or water signal provided by the index. Most importantly, pixels classified as vegetation change include both areas of gained (e.g., conversion from bare ground to green field) or lost vegetation (e.g., a forest converted to buildings) and changes to existing vegetation areas that have become observably more green (e.g., a meadow that is growing tree cover due to succession) or less green (e.g., an existing field that is now used to grow a different crop). Additionally, while these values do represent measured changes and an observed trendline, we do not apply a filter of statistical significance to assess how likely it is that a pixel’s trendline represents a robust pattern of change or if it may represent some other inter-year variation.
Area of habitat land restored between 2000 and 2020 as % of 2000 habitat land
A direct way for a city to improve biodiversity is to convert poor habitat or non-habitat land to good habitat. Good habitat generally includes multispecies vegetation and enough structural complexity to provide shade and cover. The Singapore Index’s Indicator 7 calls on cities either to estimate the area of land restored to “good ecological functioning” or to enumerate the types of habitat restored. We provide indicators to support the habitat-type enumeration method.
We are unable to discern habitat quality using global data, but we are able to provide data on changes in land classification from nonhabitat to habitat. We used the Landsat Analysis Ready Data from Global Land Analysis and Discovery research group at the University of Maryland. We defined habitat as aquatic or non-cropland vegetated classes, and we compared classifications from the years 2000 and 2020. We defined new habitat as pixels that were classified as non-habitat (urban, cropland or bare) in 2000 but as habitat in 2020. For this indicator we calculated (Area of new habitat) ÷ (area of non-habitat in 2000).
Note that this method differs somewhat from what is described in the Singapore Index indicator 7. Our denominator is non-habitat rather than degraded habitat. Our estimate probably underestimates restoration as measured by this indicator.
Number of habitat types restored between 2000 and 2020 as % of all habitat types in the city
This indicators calculated by the formula (Number of habitat types restored) ÷ (number of habitat types in the city).
The Landsat Analysis Ready Data allows discernment of six land-cover classes which we include as types of habitat: short vegetation, forest, tall forest (taller than 20m), wetland with short vegetation, wetland forest, and open water. (We treated the classes for bare ground, snow or ice, cropland, and built-up area as non-habitat.) We calculate our indicator by (1) identifying areas of new habitat, where habitat existed in 2020 but not in 2000; (2) counting the number of habitat types (i.e., aquatic and vegetated land-cover classes) in these areas of new habitat; and (3) comparing this number to the total number of habitat types in the city in 2020.
Note that our method differs from that in the Singapore Index: we use new habitat instead of habitat that has been improved from degraded to good ecological function. Cities that have access to local data specifically detailing the extent and status of habitat-restoration projects would probably benefit from calculating the Singapore Index Indicator 7 using that data.
Percent of land area that is designated as protected area
Protected or secured natural areas indicate local and other governments’ legally formalized commitment to biodiversity conservation. Protected areas are lands (or waters) with legal restrictions on development or use, and sometimes physical barriers to entry.
This indicator uses the following formula: Area of protected or secured natural areas) ÷ (Total area of the city) × 100%
Percent of Key Biodiversity Area land area that is designated as protected area
Key Biodiversity Areas (KBAs) are areas that have been identified by the KBA Partnership as being important to global biodiversity, generally because they include either a globally important type of habitat or because they include habitat for a globally important species. Cities can contribute to the global persistence of biodiversity by limiting development within KBAs, or possibly formally protecting land within KBAs.
In addition to the map of local KBAs, we provide two indicators related to KBAs. Neither indicator corresponds directly to any indicator in the Singapore Index.
KBAs can often benefit from formal protection (but see this guidance on protection of KBAs on the relationships between KBAs and protected areas). In this indicator we provide information on how much of the KBA area in the city is currently under formal protection. The formula is: (Area of protected KBA within city) ÷ (Total area of KBA within city).
Percent of Key Biodiversity Areas that are within built-up areas
Habitat quality in built-up areas tends to be lower than in natural areas. Habitat quality in KBAs might therefore be improved by habitat-restoration efforts that convert built-up areas to natural ecosystem types.
We provide an indicator that describes how much of the KBA area within the city is built-up. The formula is: (Area of built-up KBA within city) ÷ (Total area of KBA within city). The data on built-up areas are from the ESA WorldCover. Note that while we are not allowed to provide spatial data on KBAs for download, all KBA data can be requested from BirdLife International at http://keybiodiversityareas.org/kba-data/request.
Hectares of recreational space (open space for public use) per 1000 people
Parks, natural areas and other green spaces provide city residents with invaluable recreational, spiritual, cultural, and educational services. They are essential for human physical and psychological health.
The recreational services indicator is calculated as (Total area of recreational space) ÷ (1000 population ÷ 1000). Data on recreational areas were taken from the crowdsourced data initiative OpenStreetMap.
Percent of built-up area that is open space for public use.
Same as Cities4Forests indicator GRE-3.1.
Percent of population with access to public open space within walking distance (400m).
Same as Cities4Forests indicator GRE-3.2.
Percent of population with 10% or greater tree cover within walking distance (400m).
Same as Cities4Forests indicator GRE-3.3.
GHG emissions (CO2e) from city disaggregated by pollutant and sector.
Same as Cities4Forests indicator GRE-5.1.
Average annual carbon flux from trees (2001-2021) per hectare of city area (Mg CO2e/ha).
Same as Cities4Forests indicator GRE-5.2.