Research ArticleENVIRONMENTAL STUDIES

High-tide flooding disrupts local economic activity

See allHide authors and affiliations

Science Advances  15 Feb 2019:
Vol. 5, no. 2, eaau2736
DOI: 10.1126/sciadv.aau2736
  • Fig. 1 The study site in Annapolis, Maryland.

    (A) Annapolis, located adjacent to the Chesapeake Bay. The red box marks the historic downtown and City Dock area. (B) Aerial image of the City Dock parking lot, with the locations of two storm drains marked in yellow and two entrances marked with red arrows. The blue rectangle marks the Market Space parking lot.

  • Fig. 2 Flood documentation as a function of maximum daily water level and total daily rainfall.

    For each day in 2016 and 2017, the maximum daily water level and total daily rainfall are plotted. A day is considered documented if there is photographic documentation of flooding on that day. Days with a photograph of flooding are plotted as blue dots, while days without a photograph of flooding are plotted as gray dots. The concentration of flood days with zero rainfall (in the upper left corner) demonstrates that high water levels alone can cause flooding.

  • Fig. 3 High-tide flood extent at water levels of 1.73, 2.03, 2.33, and 2.83 feet (the maximum observed water level in our data) based on LiDAR (light detection and ranging) data.

    See Materials and Methods for additional documentation. Corresponding flood categories are specified in the first three images, and the fourth image depicts the maximum water level observed in our data, 2.83 feet. Points are aggregated to 5 feet by 5 feet cells, and cells are shaded red if the minimum elevation within the cell is lower than the specified water level. Storm drains are marked in yellow.

  • Fig. 4 Estimated changes in visits due to high-tide flooding for minor, moderate, and major flood hours; during rain hours; and during postflood time periods (1 to 6 hours postflood and the day after a flood).

    Errors bars indicate 95% confidence intervals using standard errors for the negative binomial models and robust standard errors for the Poisson models. Effect of rainfall is represented as the effect of 0.085 inches/hour, which is the average rainfall in an hour during hours with rain.

  • Fig. 5 Cumulative impacts of high-tide flooding on visits to City Dock, as observed to date and projected under additional sea level rise.

    (A) Annual mean sea level (feet NAVD88) over time in Annapolis (NOAA tide gauge #8575512). (B) Hours above the high-tide flood threshold each year. Only hours in which Annapolis City Dock parking lot is operational are included. (C) Cumulative impacts of high-tide flooding on visits to City Dock based on mean annual sea level. Estimates are based on a uniform increase or decrease compared with 2017 water levels, holding all else constant. Error bars mark 95% confidence intervals. For observed mean sea levels, each point represents 1 year, with color reflecting the decade in which the sea level was observed. Higher sea levels and associated visits are projected in blue. Note that a maximum of one additional foot of sea level rise is included in this plot; 1 foot is the lower bound of projected global mean sea level rise for the 21st century. The high and extreme sea level rise scenarios are 6.6 and 8.2 feet, respectively (1).

  • Table 1 Summary statistics of key variables.

    Total observations (hours) in data = 4584.

    VariableMeanMinMax
    Water level (feet NAVD88)0.406−2.132.59
    Visits per hour52.600120
    Rainfall (inches)0.00501.03
    Flood*0.02001
    Minor flood*0.01101
    Moderate flood*0.00701
    Major flood*0.00201
    First hour postflood*0.00401
    Second hour postflood*0.00501
    Third hour postflood*0.00601
    Fourth hour postflood*0.00601
    Fifth hour postflood*0.00601
    Sixth hour postflood*0.00701
    Day after flood*0.08901

    *Binary variables.

    Supplementary Materials

    • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/2/eaau2736/DC1

      Supplementary Material

      Table S1. Summary of interviews.

      Table S2. Regression results for model specifications.

      Table S3. Regression results for alternative specifications.

      Fig. S1. Documentation of flooding by time of day and by water level.

      Fig. S2. Documentation of flooding on City Dock.

      Fig. S3. Visits as a continuous function of water level.

      Fig. S4. Model results showing estimated changes in visits to the Market Space parking lot during minor, moderate, or major flood hours; during rain hours; and during postflood time periods (1 to 6 hours postflood and the day after a flood).

      Fig. S5. Estimated visits in past years using hourly water level measurements.

      References (3236)

    • Supplementary Materials

      This PDF file includes:

      • Supplementary Material
      • Table S1. Summary of interviews.
      • Table S2. Regression results for model specifications.
      • Table S3. Regression results for alternative specifications.
      • Fig. S1. Documentation of flooding by time of day and by water level.
      • Fig. S2. Documentation of flooding on City Dock.
      • Fig. S3. Visits as a continuous function of water level.
      • Fig. S4. Model results showing estimated changes in visits to the Market Space parking lot during minor, moderate, or major flood hours; during rain hours; and during postflood time periods (1 to 6 hours postflood and the day after a flood).
      • Fig. S5. Estimated visits in past years using hourly water level measurements.
      • References (3236)

      Download PDF

      Files in this Data Supplement:

    Stay Connected to Science Advances

    Navigate This Article