Overview: Michael Baker International has initiated a project with PennDOT to determine flooding vulnerability for state owned bridges and roadways in three pilot Pennsylvania Counties: Allegheny, Delaware, and Lycoming. Flooding vulnerability for the 1% annual chance flood was determined for current, existing conditions and for future conditions in the years 2050 and 2100. Bridges were analyzed under existing and future conditions to determine flooding vulnerability, including the estimated water depth on the bridge. Roadways were also analyzed under existing and future conditions to determine flooding vulnerability, including the estimated water depth along the roadway.
Data Sources:
Allegheny County | Delaware County | Lycoming County | |
---|---|---|---|
3m Digital Elevation Model (DEM) | PASDA, 2006-2008 LiDAR | PASDA, 2006-2008 LiDAR | PASDA, 2006-2008 LiDAR |
Flood Hazard Areas (including BFEs and Cross Sections) | National Flood Hazard Layer, 1/29/2015 (Cut-off date for LOMRs and prelims: 8/1/2016) | National Flood Hazard Layer, 7/13/2016 (Cut-off date for LOMRs and prelims: 8/1/2016) | National Flood Hazard Layer, 6/2/2016 (Cut-off date for LOMRs and prelims: 8/1/2016) |
Bridge Data | PennDOT, 2016 | PennDOT, 2016 | PennDOT, 2016 |
Roadway Data | PennDOT, 2016 | PennDOT, 2016 | PennDOT, 2016 |
WSEL and Depth Grid Creation
1) Water Surface Elevation (WSEL) Grids and Depth Grids were created for each county using the latest available FEMA Flood Insurance Rate Map GIS data. For Allegheny and Lycoming Counties the National Flood Hazard Layer (Effective) was used. For Delaware County, the National Flood Hazard Layer (Effective) was used in combination with Preliminary riverine analysis data for the Brandywine-Christina watershed, Preliminary riverine analysis data for the Chester Creek Levee de-accreditation, and Preliminary coastal analysis for the Delaware River. A custom ArcGIS tool was developed to produce the WSEL and Depth grids. For the riverine areas, the WSEL grids were developed using Inverse Distance Weighting (IDW) interpolation of the latest water surface elevation data in the cross section (S_XS) and base flood elevation (S_BFE) layers for detailed flood zone areas (i.e Zone AE), and IDW interpolation of the floodplain boundary elevations for the approximate flood zone areas (i.e. Zone A). For the coastal areas, the static Base Flood Elevation (BFE) was translated directly into the WSEL. The Depth grids were developed from the WSEL grids and the ground surface digital elevation model (DEM) grids. Only 1% annual chance outputs were produced.
Bridge Vulnerability Analysis
2) A custom ArcGIS tool was developed to produce bridge and floodplain elevation data for all bridges in the County that were within or near the exiting conditions 1% Special Flood Hazard Area (SFHA) data. Bridge point data for state bridges were supplied by PennDOT, and Bridge Centerlines were developed for all applicable bridge points that were within or near the 1% SFHA. Using the Bridge Centerlines, WSEL grids, Depth grids, ground surface DEM, and the latest SFHA data, the custom ArcGIS tool calculates the average bridge elevation using the ground surface DEM elevations at the end points of the Bridge Centerlines. This approach accounted for the variability in the existence of the bridge surface being removed or not removed in the ground surface DEM. The tool also calculates the average WSEL for the bridge centerline using Zonal Statistics on the WSEL grid, extracts the maximum depth along the bridge centerline using the Depth grid (In cases where the bridge is removed from the DEM, this value will actually be the water underneath the bridge, instead of on top), and calculates average water depth on the bridge by subtracting the average elevation from the average WSEL.
3) The tool produces a bridge point shapefile as an output. The following fields and attributes are included. All numeric fields are rounded to the nearest tenth of a foot (existing SFHA depths may be less than 0.05 feet in some areas and the result in the bridge point shapefile field may show 0.)
Roadway Vulnerability Analysis
4) A custom ArcGIS tool was developed to produce floodplain elevation data for all roadways in the County that were within the exiting conditions 1% Special Flood Hazard Area (SFHA) data. Roadway data for state roadways were supplied by PennDOT. The tool first removes sections of roadway that correspond to bridge centerlines using a 40 foot buffer of the Bridge Centerlines (these sections were already analyzed in bridge vulnerability analysis), then extracts the minimum, maximum, and average water depths along the remaining roadway segments using zonal statistics on the WSEL and Depth grids. Individual roadway segments are defined using a combination of the PennDOT roadway feature attributes ([cty_code]+[st_rt_no]+[seg_no]). This segment ID (ROAD_ID) can be used to link the output layer back to the original PennDOT dataset.
5) The tool produces a roadway segment shapefile as an output. The features in this shapefile are roadway segments within the SFHA that have flooding on them. Roadway segments within the SFHA that do not have flooding on them are not included. The following fields and attributes are included. Note fields d-g are rounded to the nearest tenth.
Three future conditions hydrology models were run for the detailed riverine studies in each county. The average WSEL increase from the three models was applied to the existing conditions cross section layers to develop a future conditions cross sections layer that included water surface elevations for year 2050 and year 2100. The WSEL increase values in the cross sections were then applied to the existing condition BFE layer using a spatial join operation in ArcGIS based upon the "closest" cross section. This output was then modified to account for instances where the "closest" cross section was pulled from adjacent reaches rather than the closest cross section within its reach. This occurred mainly where cross sections were sparsely available along a reach and at confluences and backwater areas. The Future BFE layer is also modified to capture static AE areas (such as lakes and backwater ponding areas) and their future elevations. A custom ArcGIS tool was developed to produce future floodplain boundary polygons for years 2050 and 2100 using the future cross section layer and future base flood elevation layer, and temporary WSEL and Depth grids. The temporary WSEL grids were produced through IDW interpolation of the future conditions WSEL data. The temporary Depth grids were produced from the WSEL grids and ground surface DEM. The future 2050 and 2100 floodplain boundary layers were produced from the output temporary Depth grids. The floodplain boundary layers were then modified manually to account for adjustments at confluences, floodplain limits, and hydraulically disconnected areas. Other than these edits, the floodplain boundary layers are raw output. No smoothing operators or filling techniques were used.
Allegheny County and Lycoming County contain riverine flooding. Delaware County contains coastal flooding as well as riverine flooding. Just as with Lycoming and Allegheny Counties, the Delaware County riverine results included vulnerability analysis for Existing Conditions, and analysis for two Future Conditions scenarios, year 2050 and year 2100. The Future Conditions coastal analysis for Delaware however, included two sea level rise scenarios for each future scenario. Therefore, Delaware County has four analysis results for Future Conditions for the coastal areas, two for 2050 and two for 2100.
To create the coastal future conditions WSELs, the sea level rise values for each scenario were added directly to the present day 1% annual chance Stillwater Elevation (SWEL) surface, from the 9/2/2015 Effective FEMA coastal study, using raster math. Since the SWEL surface was used for the coastal future conditions analysis, the effects of waves on the WSELs were not included in the future conditions mapping. See the Limitations section below for details. This created temporary WSELs for each scenario that were used to create a coastal floodplain boundary by subtracting the topographic data from each WSEL using the Surface Difference tool in ArcGIS. Outputs from this tool are vector polygons that show areas above and below the flood elevation. The polygon below the flooding elevation that exhibits hydraulic connectivity with the Delaware River was isolated and smoothed to be used as the floodplain boundary. This was repeated for each of the four future year coastal scenarios.
In areas where riverine flooding meets coastal flooding, the coastal floodplain boundary extents were manually clipped at the inland location in which the coastal BFE matched the riverine BFE. This clipping location varied depending on the coastal scenario. Once the floodplain polygons were manually edited, they were used to clip the temporary WSELs. The clipped coastal WSELs were then merged in with the riverine WSELs to create combined riverine and coastal WSELs that included interpolations in the elevations between the riverine and coastal areas as needed to provide a seamless transition. The Depth Grids were created using raster math to subtract the topographic data from the WSELs.
To create the final WSELs and Depth Grids, the floodplain boundaries for both the coastal and riverine scenarios were then used to clip the WSELs and Depth Grids and produce final grids. The result was four WSELs and four Depth Grids, two for 2050 and two for 2100, in which the coastal differs for all four (due to the different sea level rise scenarios used) but the riverine is the same for both 2050 scenarios and both 2100 scenarios.
These grids were then used as input to the Bridge Analysis and Roadway Analysis Tools (discussed in the Existing Conditions Methodology) to create future conditions vulnerability outputs for the bridges and roadways. (Note that only detailed SFHA areas are included in the future conditions output).
LiDAR Data
LiDAR data that has been processed and converted to a "bare earth" DEM was readily available and used for this project. Part of what makes a DEM "bare earth" is removal of trees, buildings, bridges and other infrastructure not part of the ground. This was the case with the DEMs used for Allegheny, Delaware, and Lycoming. Large bridges were removed from the DEM preventing extraction of an exact elevation value for the bridge and leading to the development of an automated estimated elevation calculation which is detailed in the methodology section above. When LiDAR is collected the "Last Returns" of the dataset can be also be processed into its own raster grid. The Last Returns are tops of buildings, bridges, and trees. Last Returns were not readily available for this project, but for a more accurate analysis, they can be added as an input dataset to the analysis.
Accuracy of Data
In some instances the bridge points provided by PennDOT were not spatially located on the bridge. While digitizing the bridge centerlines, in areas of multiple bridges, it may have been unclear what point corresponded to what bridge. Best assumptions were made in preparing this data for the analysis, however, it is possible inaccuracies may exist within the Bridge IDs.
Anomalies were also found within the roadway dataset. Small, isolated, roadway segments that corresponded to bridge locations were found, however, some of these locations did not have a bridge point in the bridge dataset. This was observed in Delaware County and may be present in Allegheny and Lycoming as well. At this time these segments were left as is. As a result they may be present in the output roadway dataset and appear as inundated even though they may not be.
Local inaccuracies may exist in the WSEL and depth grids. Their creation is automated using best available FEMA FIRM GIS data inputs.
This is a higher level analysis so local inaccuracies in the overall results may be present.
The future conditions WSEL grid datasets are developed from the detailed future conditions cross section and base flood elevation data. The existing conditions WSEL grid datasets are developed from the detailed cross section and base flood elevation data AS WELL AS approximate floodplain elevation data extracted from the ground surface DEM along the boundaries of the approximate floodplains. This additional elevation data being used as water surface elevation data in the development of the WSEL grids results in a differences between the Existing Conditions WSEL grids and the Future Conditions WSEL grids around confluences where the Zone As meet the detailed reaches. These differences will also result in differences between Existing Conditions and Future Conditions vulnerability results in the bridge and roadway data. Anomalies in the vulnerability results in these areas should be expected.
Stream Profiles
Bridges can be included on detailed stream profiles located in the Flood Insurance Study (FIS) report. In instances where bridges are shown on the profile, one can ascertain whether each flood event shown on the profile overtops the bridge or not. Due to the manual nature of such a review, bridge profile determinations are not included in the analysis and output for this project. In most cases, the automated results will match the profile determination, however in certain instances where the WSEL is very close to the bridge deck, the automated results may not match the profile due to the way the automated tool averages values. For a more refined analysis, this can be a recommended task.
Zone A
The results for bridges and roadways in Zone A areas should be interpreted cautiously. Zone A areas are approximate 1% annual chance flood zones without any published elevation information. Newer Zone A areas are required to have model-backed cross section elevations available, however many of the Zone As in the pilot counties did not have this information. These Zone As are older, sometimes delineated decades ago on topography that was much coarser than the LiDAR-based topography available today. The GIS tools developed for this project were able to estimate water surface elevations and water depths for Zone A areas, but these are only estimates. In order to easily identify bridges and roadways in Zone A areas, a "Notes" field has been added to the GIS deliverables and "Zone A" included in that field where applicable.
Future Conditions WSEL and Depth Grids in Zone A Areas
Future conditions hydrology models were only run for the detailed riverine and coastal studies in each county. Zone As are not included due to lack of available model information.
Future Conditions WSEL and Depth Grids in Accredited Levee Areas
Lycoming County has several levee systems that are currently shown as accredited in the NFHL. Levee freeboard analysis was not performed using the future conditions estimates as part of this project. As such, bridges and roadways within the current levee "protected areas" continue to be shown as free of flooding in the 2050 and 2100 results, even though they may actually be at risk due to future conditions flooding.
Future Conditions Roadway Vulnerability Results
Vulnerable roadway segments produced in the 2050 and 2100 analyses may not match extents of the existing conditions roadway results and as such should not be compared in a 1:1 fashion. The roadway analysis is based upon a clip of the roadway segment within the floodplain. The 2050 and 2100 floodplains will be larger in most areas. There will also be small areas within the floodplains that may have been smoothed or filled in for the existing FEMA 1% floodplain, but have not for the 2050 and 2100 floodplains. As a result vulnerable roadway segments may be a larger or smaller geographical extent for 2050 and 2100 compared to the existing conditions results. The minimum, maximum, and average flood depth along these segments of roadway may be different and in some cases lower than the existing conditions results.
Coastal Wave Action
The present day coastal analysis for Delaware County used FEMA 9/2/2015 Effective mapping and associated BFEs for the bridge and roadway vulnerability analysis. The BFEs in the mapping include the impacts of waves on the WSEL. The future conditions coastal approach however, did not address wave impacts.
The future conditions mapping incorporated sea level rise on top of the present day 1% SWEL surface and did not include additional wave modeling to incorporate the effects of future conditions waves. Wave action along the Delaware River in Delaware County is minimal and predominantly concentrated as runup along the developed shoreline. Because of this, waves generally do not impact bridges or roadways in the County as much as elevated waters due to storm surge, which is included in the 1% SWEL.
Since waves are not included in the future conditions analysis, WSELs in some areas may be higher in the present day mapping than in the future conditions mapping, even though the future conditions mapping includes sea level rise.