Research ArticleGEOPHYSICS

Global climate change and local land subsidence exacerbate inundation risk to the San Francisco Bay Area

See allHide authors and affiliations

Science Advances  07 Mar 2018:
Vol. 4, no. 3, eaap9234
DOI: 10.1126/sciadv.aap9234
  • Fig. 1 SFBA and data sets.

    (A) Footprint of SAR images (Envisat C-band and ALOS L-band) and GNSS stations used for estimating 3D displacement field and validation. We randomly assign about half of the sites to be tie points relying on the others as independent checkpoints. E, east; N, north, U, up; GPS, Global Positioning System; BARD, Bay Area Regional Deformation. (B) Time series of monthly (dashed line) and annually averaged (solid line) sea level measured at three tide gauges within the Bay Area with locations marked in (A). Time series have been offset for clarity. Asc, ascending; Des, descending.

  • Fig. 2 Average 3D velocity field across the SFBA from 13 July 2007 to 17 October 2010.

    (A) East, (B) north, and (C) up. In (A) to (C), displacement rates are relative to central Bay Area GNSS station LUTZ (37.28685°, −121.86523°, 96 m). Circles in (A) and (B) show the location of GNSS checkpoints (see Fig. 1) color-coded with their respective horizontal GNSS velocities. In (A) to (C), STD represents the SD of the difference between 3D displacement obtained from InSAR and GNSS measurements at the location of checkpoints. (D) Land subsidence transferred into the North American reference frame (NA12) using measurements of four continuous GNSS stations provided by Blewitt et al. (24). Triangles show location of four continuous GNSS stations color-coded with their respective vertical velocities. Circles show location of continuous GNSS station used for validating 3D displacement field time series shown in fig. S3. Note that for sake of visualization, the scattered data points are interpolated on a regular grid using an inverse distance interpolation algorithm, which causes a somewhat patchy appearance. The background in (D) is a shaded relief generated from Shuttle Radar Topography Mission (SRTM) 1 arc sec digital elevation model (DEM) (www2.jpl.nasa.gov/srtm/).

  • Fig. 3 Inundation maps and areas for different emission scenarios.

    The lower and upper bounds of the likely ranges (67% probability) for projected SLR under various RCPs are considered (see table S1). In each scenario, we consider only SLR (yellow) and the combined effect of SLR and LLS (red). (A) Area that will be inundated in 2100 considering the upper bound of SLR projection under RCP 2.6 scenario, which represents the goals of United Nations Framework Convention on Climate Changes 2015 Paris agreement. (B) Results using the upper bound of RCP 8.5 projection, under which there will be no significant effort to mitigate or remove emissions. Examples of key areas with significant inundation are marked, and close-ups are shown in Fig. 4. Background in (A) and (B) is the 1-m resolution black and white aerial imagery provided by National Agriculture Imagery Program (NAIP) (to obtain the data, see https://lta.cr.usgs.gov/NAIP). (C) Total estimated areas of potential inundation considering the ranges of projected SLR alone (yellow) and the combined effect of SLR and LLS (red) in 2030, 2050, and 2100. Letter labels show the estimates of inundated areas shown in (A) and (B).

  • Fig. 4 Examples of areas characterized by significant inundation by 2100.

    (A) Foster City, (B) Union City, and (C) San Francisco International Airport (SFO). Locations are marked in Fig. 3B. Top row shows the area that will be inundated considering only LLS, middle row indicates the inundated areas considering the upper bound of SLR projection under the RCP 8.5 scenario, and bottom row illustrates the combined effects of LLS and SLR. Background is the 1-m resolution black and white aerial imagery provided by the NAIP (to obtain the data, see https://lta.cr.usgs.gov/NAIP).

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/3/eaap9234/DC1

    table S1. Projected SLR (in meters) for the Golden Gate tide gauge in San Francisco.

    table S2. SAR data set used in this study, including Envisat data acquired in descending orbit track 70 and ascending orbit track 478, as well as ALOS SAR data obtained in ascending orbit, frame 740 and track 222.

    fig. S1. Validation test using synthetic data sets.

    fig. S2. LOS velocity associated with each data set.

    fig. S3. Validation test of InSAR time series with GNSS measurements.

    fig. S4. The distribution and LIDAR DEMs used in this study, as well as the associated sources.

    fig. S5. One hundred–year projected uncertainty of the land subsidence from the InSAR- and GNSS-derived subsidence rates.

    fig. S6. Present-day compaction rates.

    fig. S7. Inundation map at 2030 given the lower bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S8. Inundation map at 2030 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S9. Inundation map at 2050 given the lower bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S10. Inundation map at 2050 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S11. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 2.6 scenario (table S1).

    fig. S12. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S13. Inundation map at 2100 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).

    fig. S14. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 8.5 scenario (table S1).

    fig. S15. Inundation map at 2100 given the SLR projection under H++ scenario (table S1).

    Reference (44)

  • Supplementary Materials

    This PDF file includes:

    • table S1. Projected SLR (in meters) for the Golden Gate tide gauge in San Francisco.
    • table S2. SAR data set used in this study, including Envisat data acquired in descending orbit track 70 and ascending orbit track 478, as well as ALOS SAR data obtained in ascending orbit, frame 740 and track 222.
    • fig. S1. Validation test using synthetic data sets.
    • fig. S2. LOS velocity associated with each data set.
    • fig. S3. Validation test of InSAR time series with GNSS measurements.
    • fig. S4. The distribution and LIDAR DEMs used in this study, as well as the associated sources.
    • fig. S5. One hundred–year projected uncertainty of the land subsidence from the InSAR- and GNSS-derived subsidence rates.
    • fig. S6. Present-day compaction rates.
      fig. S7. Inundation map at 2030 given the lower bound of the likely range of SLRprojection under RCP 4.5 scenario (table S1).
    • fig. S8. Inundation map at 2030 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).
    • fig. S9. Inundation map at 2050 given the lower bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).
    • fig. S10. Inundation map at 2050 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).
    • fig. S11. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 2.6 scenario (table S1).
    • fig. S12. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).
    • fig. S13. Inundation map at 2100 given the upper bound of the likely range of SLR projection under RCP 4.5 scenario (table S1).
    • fig. S14. Inundation map at 2100 given the lower bound of the likely range of SLR projection under RCP 8.5 scenario (table S1).
    • fig. S15. Inundation map at 2100 given the SLR projection under H++ scenario (table S1).
    • Reference (44)

    Download PDF

    Files in this Data Supplement: