Authors' statement of overarching rationale
We believe that universities (private and public) should serve citizens' interests and bridge the disparity in knowledge available to community groups that have less resources to access tools/software/data to build effective arguments against both the systemic forces of history (struggles for affordable housing, better living conditions and the myriads of events and practices of life) and the - no longer - extraordinary climate-induced events like coastal and stormwater flooding.
Methods & Analysis
In order to map the H&A properties in current and future floodplain, we engaged in the following:
- Retrieved property data from Hope (not geo-located and with not practical BBL numbers) and from Ascendant (already geo-located but with projection issues)
- Re-coded BBL numbers for both properties from scratch through padding numbers and concatenation
- Conversion Tool - Convert Exl into Table
- Table join between properties BBL and Mapluto BBLs for Manhattan but 6/100 properties did not join. I figured out that some addresses that I had initially split into two (e.g. 242-244 East 106 Street) should be kept as one because splitting them generates a new BBL code that of course does not exist. After I joined the address back I solved 5 of 6 missing BBL which were not joining.
- Selected all null (not joint attributes) and selected inverse to obtain only BBLs in CD 10 and 11, where properties are located. Exported data selection as shapefile.
- Created 1 field per floodplain (2015, 2020, 2050) set with a short integer and all values at 0 (NOT IN FLOODLAIN). From the FEMA 2015 I selected by attribute only the 100 year floodplain
- Selected by location - from the BBL with condition of intersecting with floodplain 2015, 2020, 2050, 2080. Each time I selected with field calculator BBLs that intersected with each floodplain and I assigned the intersection at 1 (IN FLOODPLAIN)
- Created another column where I aggregated all numbers to understand the BBLs that fall in the floodplain at none, one, two or all projections.
- I decided to represent the different floodplains in three colors and code the properties (1 Ascendant; 0 Hope) and assign different colors to both. Another way I could have done this was to color code the chronology of flooding from green to red depending on 3 (floods in 2020), 2 (floods in 2050), 3 (floods in 2080) but I realized I also needed to distinguish the properties. I tried do choose "Many values" under Symbology and assign both coding for floodplain and coding for type of property (distinguishing them through a specific pattern applied to a solid color). The result was not satisfying because, especially in the overall map the zoom levels are not appropriate to show the details of the patterning.
- Exported Attribute table to Excel and performed a yearly rate of change analysis on the total number of buildings exposed to different floodplains as well as yearly rate of change for each property type. I used the following formula: =(new_value-original_value)/(original_value) where original_value represents the value that the percentage of change is based on, and new_value represents the value that has changed. I then divided what I obtained by the number of years (2080-2015) that is 65
- Created Map 1 showing all floodplains and the percentage of properties affected by flooding under each scenario
- Created 4 small multiple maps showing a zoomed in version of Map 1. The zooms were chosen based on where it most flooded in each projection.
In order to compare the percentages of H&A buildings flooding to those in Central and East Harlem I engaged in the following steps:
- Repeated steps 6 to 10 to calculate amount of BBLs in both Central and East Harlem that are flooded under each projection
- Created Map 6 where I show the percentage and yearly rate of flooding of all buildings in Central and East Manhattan
Finally in order to calculate the average H&A properties elevation in relation to floodplain hazard, I:
- Used Zonal Statistics as Table, where under "Input raster or feature zone data" I selected the H&A BBLs properties, in Zone field I selected 'BBL' under "Input value raster" the DEM clipped and mask to Central and East Harlem. Under statistics type I choose MEAN and I repeated the operation to obtain the MEDIAN elevation of the properties.
- Made a Table join on "BBL" between the table produced by zonal statistics and the table with H&A properties.
- Opened the table and on MEAN I chose ' Summarize SUM Floodplain' to the average MEAN elevation and saved it outside the geo-database as text file so I could work it on Excel.
- The result is Map 7 where I color coded different elevations (from 10 to 70 feet) of all properties and relate this to all floodplains.
In order to demonstrate socioeconomic vulnerability and exposure to urban change, we:
- Retrieved property data of remaining units from the NMC (46 in total) within East Harlem with expiring rent subsidizations from city, state or federal programs. We table joined the expiring subsidy data to Mapluto using the BBL number.
- Retrieved shapefiles for census tracts that have a median household income of $30000 or less for a family of four from the ACS 2017 5 year Estimates. We selected $30000 because that is the income limit for the federal poverty level.
- Retrieved shapefile for the 2017 East Harlem Rezoning from NYC Planning; residential and commercial districts within the rezoning were up-zoned for new higher density development and investments. Expansion of the Second Ave Subway also fall within this rezoning.
- We made a Vulnerability Ranking to visualize property lots that fall within various floodplains and exposed to certain socioeconomic conditions related to urban change on a temporal scale (EH rezoning and low income census tracts). Each indicator is given a score, described below, and property lots (BBLs) are assigned a score of 0-10, with the most vulnerable BBLs assigned at 10. BBLs that fall within FEMA 2015, NPCC 2020, 2050 and 2080 floodplains are given a score of 1 for each floodplain, with a max cumulative score of 4 if they fall within all four. BBLs with units that have affordable housing subsidies expiring between 2010-2020 are scored at 3, 2020-2030 at 2, and 2030 and 2040 at 1. BBLs that fall within the EH rezoning are scored at 2, given that rapid urban change and investment will likely to occur in 5-10 years. BBLs that fall within a low income census tract are scored at 1.
- Using the Vulnerability Scores, we then performed a cluster analysis using the Hot Spot Analysis (Getis-Ord Gi*) tool. The BBLs were clustered according to their vulnerability scores, which produced maps with cold and hot spots using polygon contiguity conceptualizations.
Results
In this project we asked four questions: 1) How many Hope and Ascendant properties are in the current and future floodplain? How do Hope and Ascendant buildings compare with the rest of the buildings in Central and East Harlem? 3) What is the average elevation of flooded and not flooded BBL under different floodplains? and 4) Where are high or low socio-economic vulnerability clusters located within East Harlem?
To answer question 1, the results in Figure 3 show that under the 2015 FEMA Floodplain 20% of Hope's buildings flood, while 12% of Ascendant's flood. In the NYPCC 2020 projected floodplain 37% of Hope's and 31% of Ascendant's properties may flood. Under the NYPCC 2050 projected floodplain 57% of Hope's and 35% of Ascendant's may flood. Finally under the NYPCC 2080 projected floodplain 84% of Hope's and 38% of Ascendant's properties may flood. Fig. 4 represents the same data in GIS.