Marsh Vulnerability in the San Francisco Bay-Delta Estuary

Improving our understanding of tidal marshes and their vulnerability to sea-level rise

Tidal marshes support important coastal food webs, improve water quality, and offer a buffer against storm and wave damage. Rising sea levels are negatively impacting the structure and function of these habitats.

This project is enhancing a locally relevant marsh model to better understand the impacts of sea-level rise (SLR). Results will allow coastal managers to evaluate vulnerability and inform restoration of San Francisco Bay-Delta tidal marshes. For example, we are helping  SF Bay National Research Reserve  and  East Bay Regional Park district  to better understand impacts of SLR and other changing environmental conditions (e.g., salinity, sediment, and vegetation) on their wetlands.

Objectives

This collaborative project, funded by NOAA, aims to better understand ecosystem processes in Bay-Delta wetlands and model their vulnerability to sea level rise. Our research had three main objectives:

  1. Identify how salinity and inundation drive wetland processes
  2. Improve the way elevation data are corrected
  3. Refine predictions of wetland vulnerability to sea-level rise with improved elevation and salinity information

Study Sites

Study sites, shown in the map at right, were distributed across the San Francisco Bay-Delta Estuary to capture the strong salinity gradient present from the highly saline Bay to the nearly freshwater Delta. Along this estuarine gradient wetland vegetation composition, tide range, and sediment supply also vary considerably. 

A researcher collecting data in tall, dense vegetation at Brown's Island.

Wetland Plant Communities

Wetlands & Elevation

Tidal marshes in the SF Bay Estuary occur across a range of elevations, experiencing anywhere from near-constant to very infrequent flooding. Salinity and inundation influence wetland plant distribution as well as species composition. The results at right highlight differences in species distribution and composition with inundation in saline and fresh water marshes.

The tide range from mean higher high water (MHHW) to mean lower low water (MLLW) is about 6-7.5 ft (1.8-2.3 m) in San Francisco Bay, 4.2-5.5 ft (1.3-1.7 m) in Suisun Bay, and smaller in the Delta.

We conducted marsh elevation and vegetation surveys at our study sites to assess how species cover changes across the estuary and elevation gradients within sites.

Find out more, read our recent publication:

Plant Distributions & Inundation

To understand how plant growth may change with sea-level rise, we examined plant productivity-elevation relationships for four common species in the Bay-Delta estuary. Our results show that higher inundation levels – as would occur with high relative sea-level rise – led to reduced productivity in all the common species we tested. However, species sensitivity to inundation varied. Notably, pickleweed (Salicornia pacifica) which is common throughout SF Bay and other saline marshes in California, did not grow well when inundation time exceeded about 20%.

Understanding these productivity-elevation relationships is important for assessing ecosystem change with sea-level rise since plants help build wetland elevations by producing organic matter and trapping mineral sediments. Find out more, read our recent publication:

For several common plant species, we used field mesocosms called 'marsh organs' (above) to test responses to inundation gradients.

Measuring Elevation

Improving Remotely Sensed Elevation Data

To measure and model sea-level impacts to tidal wetlands it is necessary to have accurate maps of elevation because large differences in tidal inundation occur with relatively small differences in elevation. Surveying by leveling or high precision GPS technology is one method to obtain elevation maps, but it can be costly and time-consuming. Lidar is another technique that allows mapping of large areas, but vegetation can interfere with lidar accuracy, as shown in the graphic at right.  

Where vegetation is dense or tall, lidar vertical bias can be large enough to affect the accuracy of ground elevations.

Raw lidar data needs to be corrected for these errors to ensure accurate species modeling. We use the LEAN method, described at right, to correct these lidar errors and produce a precise lidar-derived digital elevation model (DEM).

Get LEAN-corrected DEMs for the following areas in California:

DEMs are also available from  Chesapeake Bay  and  Collier County, FL. 

Comparing Inundation

Shown at right are inundation estimates from DEMs using raw lidar and LEAN-corrected lidar. Applying the LEAN-correction addresses the erroneously high elevations caused by dense plant cover. In some cases the LEAN-correction results in shifts from never inundated to partially inundated, representing a significant change in the area's vulnerability to sea-level rise. Use the legend below for more information on inundation time differences.

Color scale used in the maps at right to highlight differences in inundation time at study sites.

Try This:

Zoom in/out of a map to see how inundation time differs without adjusting the lidar digital elevation model (DEM) (left) and the LEAN-corrected lidar DEM (right).

Select the bookmark icon in the left map to see results for our other study sites. At Rush Ranch, the National Estuarine Research Reserve (NERR) boundary is included for reference (blue line). Regions without elevation data or beyond the study domain are transparent, with the underlying map visible.

Comparing Model Outputs

When modeling with elevation surfaces, using corrected or uncorrected elevation data can result in dramatically different habitat predictions and thus perceived vulnerability to sea-level rise.

Shown in the maps to the right are habitat classifications from our sea-level rise model (WARMER) using uncorrected lidar-derived digital elevation model (DEM) (left) and LEAN-corrected DEM (right) at our study sites.

These are results for 2020 and 2080 under a sea level rise scenario of 104 cm (3.4 feet) by 2100. Use the bookmark icon (top left map) to see our other study sites. The red line identifies the extent of the modeled area at each site. At Rush Ranch, the National Estuarine Research Reserve (NERR) boundary is included for reference (blue line).

Habitat classification output from the WARMER model, that predicts wetland elevation and habitat changes with sea-level rise.

Modeling Sea-level Rise

WARMER (Wetland Accretion Rate Model of Ecosystem Resilience) is a model that incorporates both biological and physical processes to assess marsh accretion. It can be used to evaluate changes in elevation, the impact of elevation changes on marsh habitats, and carbon accumulation over time.

WARMER considers the dominant processes that control elevation (shown at right). The model is run across a range of initial elevations under future sea-level rise scenarios and the elevation and carbon results are interpolated across a digital elevation model. 

WARMER Results

Shown in the maps to the right are WARMER results under a sea-level rise scenario of 104 cm (3.4 feet) by 2100. Four different time steps are shown (2020, 2050, 2070, and 2100). Red outlines identify the modeled area and at Rush Ranch, the National Estuarine Research Reserve (NERR) boundary is included (blue line). These data are available in 10 year increments from 2000 - 2100. See the link above to download the data.

WARMER habitat classifications shown in the maps at right.

Try This:

  • Zoom in/out of a map to see how WARMER results differ over time.
  • Change the background map using the top right icon (four small squares).
  • See results from other sites by selecting from the bookmark list (top left map)

Additional Outputs

Sediment, Marsh Elevation, & Carbon Sequestration

Given the uncertainty of future conditions, there are additional results for different sediment input configurations available for download (link coming soon).

Shown at right are data summaries of mean carbon accumulation (g/m2/yr) for 104 cm SLR. Additional summaries of carbon sequestration and average marsh elevation (cm above or below Mean Sea Level) are available for three sea-level rise scenarios (48cm (1.6 ft), 104 cm (3.4 ft), and 174 cm (5.4 ft)) and three sediment configurations.

These sediment configurations were derived from the work of  Cloern et al. (2011). 

Team & Funding

Core Project Team

  • Christopher Janousek (OSU)
  • Kevin Buffington (USGS)
  • Karen Thorne (USGS)
  • Bruce Dugger (OSU)

Funding

  • NOAA EESLR grant (NA15NOS4780171)
  • USGS Western Ecological Research Center

Collaborators

  • NOAA Sentinel Site Cooperative - San Francisco
  • National Estuarine Research Reserve
  • East Bay Regional Parks
  • CA Department of Fish & Wildlife

Website and data visualizations developed by NOAA's National Centers for Coastal Ocean Science (NCCOS)

A researcher collecting data in tall, dense vegetation at Brown's Island.

For several common plant species, we used field mesocosms called 'marsh organs' (above) to test responses to inundation gradients.

Color scale used in the maps at right to highlight differences in inundation time at study sites.

Habitat classification output from the WARMER model, that predicts wetland elevation and habitat changes with sea-level rise.

WARMER habitat classifications shown in the maps at right.