Widespread Expansion Of Salt Marsh Pools Observed On Maine Marshes Since 2009 (Data Files)

Authors

Katelyn DeWater, University of New EnglandFollow
William Kochtitzky, University of New EnglandFollow
Ruth Ellis, University of New EnglandFollow
Paige-Marie Merrill, University of New EnglandFollow

Document Type

Data

Publication Date

2024

Abstract

Salt marshes contain standing water, which we refer to here as pools. We manually digitized salt marsh pools on 12 salt marshes in Maine (given as Name and corresponding ID number (see ID in Attributes)) in 2009 and 2021: Scarborough (1), Webhannet (2), Spurwink (3), Marshall Point (4), York River (5), Ogunquit (6), Brave Boat Harbor (7), Cousins River (8), Biddeford Pool (9), Little River (10), Goosefare Brook (11), Gooch’s Beach (12). Marsh extents were delineated with the Maine Natural Areas Program Current Tidal Marshes layer. All pools were digitized at a scale of 1:400 to 1:1000. Features were digitized using National Agricultural Imaging Program (NAIP) aerial imagery from 2009 and 2021. The spatial resolution of the aerial photos is 1 m for 2009 and 0.6 m for 2021. For more NAIP imagery metadata such as imagery ID and tide information, see included Supplementary_table.xlsx. There were 10,566 pools in 2009 and 14,156 pools in 2021. These layers were created to document changes in salt marsh pools from 2009 to 2021:

• ‘2009_Pools_Not_Georectified’ (24,992 KB) is the original 2009 pool layer in which pool polygons align with the 2009 NAIP imagery.
• ‘2009_Pools_Georectified’ (27,936 KB) includes pool polygons moved to correctly overlap with their 2021 counterparts (affected less than 850 pools which were all less than 5 m wide). Note, this means that the pool location is inconsistent with imagery and should only be used to spatially join attributes.
• ‘2021_Pools’ (35,252 KB) contains all pool polygons in 2021.

Comments

Main file is a .zip file of geopackage pool data and supplementary tables; individual data group and table files can be accessed under Additional Files. Pool layers are geopackages (.gpkg) and can be viewed in any GIS software such as QGIS or ArcGIS. These files were created in QGIS version 3.32.0. Python scripts (.ipnyb) under additional data can be viewed in software such as Jupyter Notebook.

Attributes (in all geopackages)

• Area: area of each polygon (m²).
• PoolType
• MS: Mega-pool - Defined to disjointed borders, relatively large, typically in higher marsh, surrounded by smaller expanding pools as exterior vegetation dies off.
• IP: Individual pool - Defined borders, relatively small, typically circular, primarily naturally occurring.
• PP: Perimeter pool - Defined to disjointed borders, long, thin pools along embankments on a marsh’s border or roads.
• ID: Reported as MARSH NUMBER_YEAR_UNIQUE NUMBER.
• Example: 1_2009_20 indicates Scarborough (1); pool extent in 2009; 20 is the unique polygon number.
• Perimeter: perimeter of each polygon (m) used in error calculations.
• Digitizer_consistency (m²): Calculated for each time frame of imagery and for each pool type (mega, individual, and perimeter pool). Three mega, individual, and perimeter pools were digitized ten times for 2009 and 2021. The standard deviation of area was used to quantify digitizer accuracy by pool type and year.
• Pixel_error (m²): Uncertainty due to pixel (spatial) resolution of the aerial imagery and the perimeter of the pool.
• Uncertainty (m²): Uncertainty of each polygon. Calculated as the square root of the sum of digitizer_consistency and pixel_error squared.

Details for Python notebook scripts (Additional files)

17 June, 2024. This code is associated with a publication that will be updated when the manuscripts are published and available online. The code was developed in 2023 to calculate salt marsh pool expansion on 12 Maine salt marshes by Katelyn DeWater at the University of New England. This script requires many inputs that are combined to compute expansion, count, and elevations. Regretfully, you will not be able to run the script on your own without these specific inputs (see file paths in the jupyter notebook). If you are interested in working with this script on your own, please feel free to contact Katylyn DeWater (kdewater@une.edu) or Will Kochtitzky (wkochtitzky@une.edu), and we could help. We make this script available for complete transparency and bad code is better than no code. We are happy to answer any questions about these scripts.

• ‘Pool_Expansion.ipynb’ will compute the percent cover of pools and will generate a plot of expansion from 2009 to 2021.
• ‘Pool_Density.ipynb’ will compute the number of pools per square kilometer of marsh and will generate a plot of density from 2009 to 2021.
• ‘Pool_Changes_Intersection.ipynb’ will compute how the pools have changed over time based on how they overlap. For example, if one pool overlaps with more than one pool, it split apart. This will generate a graph of how the pools have changed over time and the percent change in pool cover grouped by type of change.
• 'Pool_elevations' will add an elevation attribute to each pool which finds the mean elevation within each pool using a lidar derived digital elevation model. It then will categorize each pool into > highest astronomical tide (HAT), HAT > mean higher high water (MHHW), MHHW > mean high water (MHW), MHW > mean low water (MLW), MLW > mean lower low water (MLLW), MLLW > lowest astronomical tide (LAT), and LAT >.
• 'Ditch_marsh_analysis' will find which pools intersect with the ditch vector layer from the USGS Marsh Ditches Layer and assign True of False as an attribute based on that intersection.

We acknowledge that this code is not commented or well documented, and you will not be able to run it without the file structures that exist on my computer. The script will not be updated as it is the script for a specific project and will need to be written differently for future projects. If you need to make these types of calculations, it will be much easier to write a script on your own rather than to use these scripts. This database contains 2009 and 2021 pool geopackage data files used int he scripts. See NOAA Data Access Viewer for 2020 lidar-derived DEM data for each marsh. We are happy to share any additional files not provided here that might be useful in your work.

Recommended Citation

DeWater, Katelyn; Kochtitzky, William; Ellis, Ruth; and Merrill, Paige-Marie, "Widespread Expansion Of Salt Marsh Pools Observed On Maine Marshes Since 2009 (Data Files)" (2024). Geographic Information Systems Data Projects. 1.
https://dune.une.edu/gis_data/1