USGS Western Ecological Research Center

Home Who We Are Where We Are What We Do Products Search Director's Message Outreach Jobs Contacts

With climate change and sea level rise, how will the coastal habitats of the San Francisco Bay Estuary change over the next 100 years? Mapping and modeling studies by WERC scientists have produced scenarios for this important coastal ecosystem.

Click on the links below to go to the project pages.


Conceptual model of sea-level rise modeling.

Coastal ecosystems have been identified by the International Panel on Climate Change (2007) as areas that will be disproportionally impacted by climate change. Recent sea-level rise projections range from 0.57 to 1.1 m (Jevrejeva et al. 2012) or 0.75 to 1.9 m by Vermeer and Rahmstorf (2009) by 2100, which are contingent upon the ambient temperature conditions and CO2 emissions.  Sea-level rise projections for San Francisco Bay are 1.24 m by 2100 (Cayan et al. 2008).  The expected accelerated rate of sea-level rise through the 21st century will put many coastal ecosystems at risk, especially those in topographically low-gradient areas.


Sea-level rise response modeling was conducted at 12 tidal salt marshes around San Francisco Bay Estuary where marsh accretion and plant community state changes were assessed to 2100.  Detailed ground elevation, vegetation, accretion, and water level data was collected at all sites between 2008 and 2011 and used as model inputs.  A modification of the Callaway et al. (1996) model, the Wetland Accretion Rate Model for Ecosystem Resilience (WARMER), was developed to run sea-level rise response models for all sites.


San Francisco Bay estuary contains the largest remaining expanse of tidal salt marsh on the Pacific coast of the United States. We collected baseline elevation, tidal inundation and vegetation data at 12 marsh sites to examine potential effects of climate change. We used Real-Time Kinematic Global Positioning System (RTK GPS; ± 2 cm vertical accuracy) to survey elevation. Vegetation was surveyed within 0.25 m2 quadrats for species composition, percent cover, and height. We established water level monitoring stations at all sites to capture local annual variation in tidal inundation and to determine elevation relative to site-specific tidal datum.

Study sites included San Pablo Bay National Wildlife Refuge, China Camp State Park, Corte Madera Ecological Reserve, Fagan Ecological Reserve, Cogswell Marsh, Arrowhead Marsh, Colma Creek Marsh, Laumeister Marsh, Coon Island Marsh, Black John Marsh, Petaluma Marsh, and Gambinini Marsh. Across all sites, 19.94 km2 of marsh was surveyed or roughly 12% of the remaining tidal marshes in the San Francisco Bay estuary.

Ground elevation data was interpolated in ArcGIS 9.3 (Kriging method) into continuous 3 x 3 m gridcell rasters. Mean root-mean-square error for all interpolations was < 0.10 m.

Sediment cores collected along elevation transects at four representative sites (China Camp, Coon Island, Petaluma, and Laumeister) were used to determine historic accretion rates with horizon marker dating techniques (210Pb, 137Cs) (Callaway et al. 2012). Percent organic matter, porosity, compaction and decomposition were incorporated with accretion rates, elevation, and tidal range data input into a Wetland Accretion Response Model for Ecosystem Resilience (WARMER). Results at the representative sites were extrapolated to the remaining eight sites.

Across all sites, 7,437 elevation points were surveyed with 88% of the data between 1.5 and 2.1 m (NAVD88). Sites were located at different elevations within the San Francisco Bay tidal range. Sites ordered by decreasing mean elevation were: Fagan (highest mean elevation), Laumeister, Gambinini, Coon Island, Petaluma, San Pablo, Black John, China Camp, Cogswell, Arrowhead, and Colma (lowest mean elevation).

A total of 3,302 vegetation plots were surveyed across all sites. Low plant species richness was recorded with 21 plant species identified within plots across all sites. The plant community composition also reflected variation in tidal inundation and water salinity. For example, the highest species diversity was found in marshes along the Napa River presumably due to lower water salinity levels.

Pickleweed (Sarcocornia pacifica) was the most common species recorded in 91% of the plots, followed by Schoenoplectus spp. (12%), Distichlis spicata (9%), Spartina spp. (9%), Jaumea carnosa (7%), Grindelia stricta (7%), and Frankenia salina (4%).

Plant communities were categorized based on observed elevation relative to local tidal datum to develop marsh zones for sea-level rise response model interpretation. Upland transition was defined by Baccaris pilurais (> 1.0 m relative to MSL). High marsh was defined by Jaumea carnosa, Distichlis spicata, and Frankenia salina (0.1 – 1.0 m MSL). Mid marsh was defined by a transition of high marsh species and the upper edge of Spartina spp (0.45 – 0.7 m MSL). Sarcocornia pacifica was found over a large elevation range and therefore was present in both high and mid marsh categories. Low marsh was defined by Spartina spp. (0.2 – 0.34 m MSL). Mudflat was defined by anything less than 0.2 m MSL.

Results from the WARMER modeling suggested that 95% (1,942 ha) or nine of the sites would become mudflats by 2100 with a 1.24 m sea-level rise. Three sites with the remaining 4% (85 ha) is projected to be low marsh habitat dominated by Spartina spp. by 2100. All upland transition, high and mid marsh habitats were projected to be lost by 2100.

Accretion rates used in the WARMER model were relatively high in south San Francisco Bay (Callaway et al. 2012), and those marshes withstood sea-level rise longer with areas transitioning from high to low marsh vegetation by 2100 (e.g. Cogswell, Laumeister, and Colma).

Napa River sites were parameterized with higher sediment accretion rates and higher starting elevations with WARMER and showed the maintenance of high marsh until 2030 (+0.24 m SLR) for Coon Island and 2040 (+0.32 m SLR) for Fagan marsh. Between 2040 (+0.32 m SLR) and 2060 (+0.57 m SLR) mid marsh vegetation was maintained. Low marsh was dominate to 2090 (+1.05 m SLR), at which time both marshes transitioned to mudflats.

All other marsh sites had relatively low accretion rates relative to sea-level rise. WARMER projected all of these marshes to transition to mudflat by 2100. Corte Madera, China Camp, and San Pablo Bay NWR marshes lost all high marsh by 2030 (+ 0.24 m SLR), briefly transitioned to mid and low marsh plant communities, but ultimately transitioned to areas dominated by mudflat by 2080 (+ 0.85 m SLR). The three marshes located on the Petaluma River (Gambinini, Petaluma, and Black John) lost most high marsh habitat by 2030 (+ 0.24 m SLR) and transitioned to mostly mudflat by 2080 (+0.85 m SLR).

Projected loss of pickleweed (Sarcocornia pacifica) habitats by 2100 could affect many tidal marsh wildlife species such as the federally endangered salt marsh harvest mouse (Reithrodontomys raviventris) and state threatened California black rail (Laterallus jamaicensis coturniculus). Only 4% of the marsh area was projected to support cordgrass (Spartina spp.) habitats by 2100 that also would affect distribution of the federally endangered California clapper rail (Rallus longirostris obsoletus).

The modification of the San Francisco Bay estuary makes it especially susceptible to sea-level rise, since there are few areas available for upslope marsh transgression. Seven of the marsh sites we surveyed had adjacent open space, but five marshes were surrounded by urban infrastructure prohibiting upslope movement.

Updated 2012.09.04

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information:

References to non-U.S. Department of Interior (DOI) products do not constitute an endorsement by the DOI.

* DOI and USGS link policies apply.