Application of the Sea-Level Affecting Marshes Model (SLAMM 6) to Galveston Bay

Prepared For: The Nature Conservancy Gulf of Mexico Initiative Corpus Christi, TX

June 30, 2011

Warren Pinnacle Consulting, Inc. PO Box 315, Waitsfield VT, 05673 (802)-496-3476

Application of the Sea-Level Affecting Marshes Model (SLAMM 6) to Galveston Bay Executive Summary ...... 1 Introduction ...... 2 Model Summary ...... 3 Sea Level Rise Scenarios ...... 6 Temporal Aspect ...... 7 Methods and Data Sources ...... 8 Results and Discussion ...... 26 Hindcast ...... 26 Forecast ...... 37 Erosion Map ...... 79 Sensitivity Analysis ...... 80 Elevation Uncertainty Analysis ...... 85 Conclusions ...... 89 References ...... 90

Funding for this model application was provided by The Nature Conservancy, who also provided extensive GIS processing in support of this analysis. Funding for TNC was provided through a grant from the Gulf of Mexico Foundation to support the Habitat Conservation & Restoration Team, a part of the Governor’s Gulf of Mexico Alliance. We would like to acknowledge the Natural Resource Use and Monitoring and Research Subcommittees of the Galveston Bay and Estuary Program for their support and help in organizing the stakeholder SLAMM workshop and gathering input to improve in developing this final report.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Executive Summary

In order to understand the potential future impacts of sea-level rise (SLR) on the ecology of Galveston Bay, the Sea-Level Affecting Marshes Model (SLAMM) was applied. A grant from the Gulf of Mexico Foundation to support the Habitat Conservation & Restoration Team, a part of the Governor’s Gulf of Mexico Alliance, was used to fund research to parameterize, calibrate, and run forecast simulations of multiple SLR scenarios until the year 2100. Historical wetland coverage information from a 1979 National Wetlands Inventory (NWI) survey was used as a starting point for model hindcasting. Hindcast simulations were run until 2004, the photo date of the newest NWI data layer. Despite significant changes in the wetland coverages in large areas of the bay due to anthropogenic changes and/or changes in NWI coding techniques, after multiple iterations of model calibration hindcast results indicated the model able to predict wetland loss in critical areas of the bay with reasonable accuracy (e.g. observed salt marsh loss in West Bay = 14%, SLAMM predicted 12%). The calibrated model was run for five individual SLR scenarios: 0.39, 0.69, 1, 1.5, and 2 m of SLR by 2100. At 1 m of SLR by 2100, what scientists consider to be the “most-likely” scenario, SLAMM predicts 67% losses of irregularly-flooded (or brackish) marsh throughout the bay. Under the same SLR scenario, 92% of the tidal swamp in the Trinity River delta is predicted to be lost. A map of erosion was produced to indicate the areas that are predicted to undergo the highest amount of marsh erosion, and showed the marshes protecting the Trinity Delta and Christmas Bay to be the most threatened by erosional forces. Simulations were also completed to assess the effect of elevation data layer uncertainty and the sensitivity of model results to individual parameters. Based on a LiDAR RMSE of 18 cm and a VDATUM correction RMSE of 10.1 cm, elevation uncertainty analysis indicated that model results are not very sensitive to error. Sensitivity analysis suggested tidal swamp, tidal flat, ocean beach, irregularly-flooded marsh, and estuarine beach were the most sensitive land cover types to changes in the historic trend parameter, which represents the rate of land subsidence. Overall, the application of SLAMM to Galveston Bay indicated marsh communities are susceptible to sea-level rise.

Prepared for Gulf of Mexico Alliance 1 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Introduction

Tidal marshes are among the most susceptible ecosystems to climate change, especially accelerated sea-level rise (SLR). The Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) suggested that global sea level will increase by approximately 30 cm to 100 cm by 2100 (IPCC 2001). Rahmstorf (2007) suggests that this range may be too conservative and that the feasible range by 2100 is 50 to 140 cm. Rising sea levels may result in tidal marsh submergence (Moorhead and Brinson 1995) and habitat “migration” as salt marshes transgress landward and replace tidal freshwater and irregularly-flooded marsh (R. A. Park et al. 1991).

In 2010 and 2011, The Nature Conservancy and Warren Pinnacle Consulting, Inc. worked together to produce an analysis of the potential effects of SLR on Galveston Bay, TX. Funding for TNC was provided through a grant from the Gulf of Mexico Foundation to support the Habitat Conservation & Restoration Team, a part of the Governor’s Gulf of Mexico Alliance.

This modeling analysis included calibrating the SLAMM model to historical data, and then using this model to provide forecasts of various potential SLR scenarios through the year 2100. A first draft of this report was created and distributed to project stakeholders who gave their feedback regarding parameter and data choices at a workshop held in the Galveston Bay Estuary Program offices on April 7, 2011. This is the final report for this project and includes an uncertainty analysis designed to ascertain the effects of elevation-data uncertainty on model predictions and also a sensitivity analysis in which the effects of various model parameters on final predictions are evaluated.

Prepared for Gulf of Mexico Alliance 2 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 1. Galveston Bay SLAMM modeling boundaries (in orange)

Model Summary

Changes in tidal marsh area and habitat type in response to SLR were modeled using the Sea Level Affecting Marshes Model (SLAMM 6) that accounts for the dominant processes involved in wetland conversion and shoreline modifications during long-term sea level rise (Park et al. 1989; www.warrenpinnacle.com/prof/SLAMM).

SLAMM predictions are generally obtained by two consecutive steps: (1) calibration of the model using available historical wetland and SLR data, referred to as the “hindcast;” (2) starting from the most recent available wetland and elevation data, the calibrated model is run to predict wetland changes in response to estimated future SLR.

Successive versions of the model have been used to estimate the impacts of SLR on the coasts of the U.S. (Titus et al. 1991; Lee et al. 1992; Park et al. 1993; Galbraith et al. 2002; National Wildlife Federation & Florida Wildlife Federation 2006; Glick et al. 2007; Craft et al. 2009).

Within SLAMM, there are five primary processes that affect wetland fate under different scenarios of sea-level rise:

• Inundation: The rise of water levels and the salt boundary are tracked by reducing elevations of each cell as sea levels rise, thus keeping mean tide level

Prepared for Gulf of Mexico Alliance 3 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay (MTL) constant at zero. The effects on each cell are calculated based on the minimum elevation and slope of that cell. • Erosion: Erosion is triggered based on a threshold of maximum fetch and the proximity of the marsh to estuarine water or open ocean. When these conditions are met, horizontal erosion occurs at a rate based on site- specific data. • Overwash: Barrier islands of under 500 m width are assumed to undergo overwash during each specified interval for large storms. Beach migration and transport of sediments are calculated. The overwash model was not used for the Galveston Bay analysis. • Saturation: Coastal swamps and fresh marshes can migrate onto adjacent uplands as a response of the fresh water table to rising sea level close to the coast. • Accretion: SLR is offset by sedimentation and vertical accretion using average or site-specific values for each wetland category. Accretion rates may be spatially variable within a given model domain or can be specified to respond to feedbacks such as frequency of flooding.

SLAMM Version 6.0 was developed in 2008/2009 and is based on SLAMM 5. SLAMM 6.0 provides backwards compatibility to SLAMM 5, that is, SLAMM 5 results can be replicated in SLAMM 6. However, SLAMM 6 also provides several optional capabilities.

• Accretion Feedback Component: Feedbacks based on wetland elevation, distance to channel, and salinity may be specified. This feedback is used where adequate data exist for parameterization. • Salinity Model: Multiple time-variable freshwater flows may be specified. Salinity is estimated and mapped at MLLW, MHHW, and MTL. Habitat switching may be specified as a function of salinity. This optional sub-model is not utilized in simulations of Galveston Bay. • Integrated Elevation Analysis: SLAMM will summarize site-specific categorized elevation ranges for wetlands as derived from LiDAR data or other high-resolution data sets. This functionality is used in the current simulations to test the SLAMM conceptual model. The causes of any discrepancies are then tracked down and reported on within the model application report. • Flexible Elevation Ranges for land categories: If site-specific data indicate that wetland elevation ranges are outside of SLAMM defaults, a different range may be specified within the interface. If such a change is made, the change and the reason for it are fully documented. • Many other graphic user interface and memory management improvements are also part of the new version including an updated Technical Documentation, and context sensitive help files.

For a thorough accounting of SLAMM model processes and the underlying assumptions and equations, please see the SLAMM 6.0 Technical Documentation (Clough et al. 2010). This document is available at http://warrenpinnacle.com/prof/SLAMM

Prepared for Gulf of Mexico Alliance 4 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay All model results are subject to uncertainty due to limitations in input data, incomplete knowledge about factors that control the behavior of the system being modeled, and simplifications of the system (Council for Regulatory Environmental Modeling 2008). Site-specific factors that increase or decrease model uncertainty may be covered in the Discussion section of this report.

Prepared for Gulf of Mexico Alliance 5 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Sea Level Rise Scenarios

SLAMM 6 was run using scenario A1B from the Special Report on Emissions Scenarios (SRES) – mean and maximum estimates (IPCC 2007). The A1 family of scenarios assumes that the future world includes rapid economic growth, global population that peaks in mid-century and declines thereafter, and the rapid introduction of new and more efficient technologies. In particular, the A1B scenario assumes that energy sources will be balanced across all sources. Under the A1B scenario, the IPCC WGI Fourth Assessment Report (IPCC 2007) suggests a likely range of 0.21 to 0.48 m of SLR by 2090-2099 “excluding future rapid dynamical changes in ice flow” (IPCC 2007). The A1B- mean scenario that was run as a part of this project falls near the middle of this estimated range, predicting 0.39 m of global SLR by 2100. A1B-maximum predicts 0.69 m of global SLR by 2100.

The latest literature (Chen et al. 2006; Monaghan et al. 2006) indicates that the eustatic SLR is progressing more rapidly than was previously assumed, perhaps due to the dynamic changes in ice flow omitted within the IPCC report’s calculations. A recent paper in the journal Science (Rahmstorf 2007) suggests that, taking into account possible model error, a feasible range by 2100 of 50 to 140 cm. This work was recently updated and the ranges were increased to 75 to 190 cm (Vermeer and Rahmstorf 2009). Pfeffer et al. (2008) suggests that 2 m by 2100 is at the upper end of plausible scenarios due to physical limitations on glaciological conditions. A recent US intergovernmental report states "Although no ice-sheet model is currently capable of capturing the glacier speedups in Antarctica or Greenland that have been observed over the last decade, including these processes in models will very likely show that IPCC AR4 projected sea level rises for the end of the 21st century are too low" (Clark 2009) A recent paper by Grinsted et al. (2009) states that “sea level 2090-2099 is projected to be 0.9 to 1.3 m for the A1B scenario…” Grinsted also states that there is a “low probability” that SLR will match the lower IPCC estimates.

To allow for flexibility when interpreting the results, SLAMM was also run assuming 1 m, 1.5 m and 2 m of eustatic SLR by the year 2100. The A1B- maximum scenario was scaled up to produce these bounding scenarios (Figure 2).

Prepared for Gulf of Mexico Alliance 6 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

200 180 A1B Mean 160 A1B max 140 1 meter 1.5 meter 120 2 meters 100 80 60 SeaLevel Rise(cm) 40 20 0 1990 2015 2040 2065 2090

Figure 2. Summary of SLR scenarios utilized

When the model was run to estimate wetland changes in the recent past (“hindcasting”), the global (i.e., eustatic) rate of SLR from 1990 to the present was estimated to be 3 mm/year (Grinsted et al. 2009)1. Prior to 1990, a global rate of 1.7 mm/year was combined with the local subsidence rate of 3.05 mm/yr., resulting in 4.75 mm/yr. of SLR.

Temporal Aspect

Each SLAMM simulation is run starting from the National Wetland Inventory (NWI) photo date as the initial condition. In this study the time-step was selected to be 25 years for forecasting and 5 years for hindcasting. This choice resulted in model output for years 2025, 2050, 2075, and 2100 for forecasting simulations. Model output was produced for 1979, 1984, 1989, 1994, 1999, and 2004 for the hindcast.

1 Due to the predicted increase of SLR over the next 90 years, this is achieved by entering a “custom” eustatic SLR of 0.54 by 2100 within the SLAMM interface.

Prepared for Gulf of Mexico Alliance 7 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Methods and Data Sources

The digital elevation map used in this simulation was derived from Sanborn 2007 and Tropical Storm Allison Recovery Project (TSARP) 2002 LiDAR (received from Harte Research Institute) and 2009 1/9 arc second NED (Figure 3) (Texas Water Development Board 2010).

Figure 3. Shade-relief elevation map of Galveston study area

Two wetlands datasets were used in this simulation, one historical map for model hindcasting and the most current data used both for model projection and evaluation of model hindcast results. The wetlands layer used for hindcasting was produced by the NWI in 1979 and was based on aerial photographs taken in November of 1979. Data from 1956 were available; however, as the 1956 maps were missing large sections of the study area the 1979 data were used instead. The wetland layer used for projection was also NWI-produced with date of 2009, but was based on aerial photos taken in August and October of 2004. Therefore, in this report the 2009 NWI layer will be referred to as the 2004 NWI layer. Figure 4 presents the 2004 wetlands data layer. Because this data layer is derived from several smaller images with various dates, it is difficult to state the tide level that corresponds to the wetland layer. Conversion of this survey into 10 m cells for preliminary model

Prepared for Gulf of Mexico Alliance 8 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay calibration and analysis indicated that the approximately two-million acre study area is composed of the categories shown in Table 1.

Table 1. Wetland Classes According to 2004 NWI Data Layer

Open O cea n Open Ocean 42.7%

Estuarin e O pen Wa ter Estuarine Open Water 19.8%

Undevelo ped Dry Lan d Undeveloped Dry Land 18.0%

Inland-fresh Marsh Inland-fresh Marsh 6.2%

Develo ped Dry La nd Developed Dry Land 5.4%

Irregularly-flooded Marsh Irregularly-flooded Marsh 3.1% Inland Open Water 1.3%

Regularly-flooded Marsh Regularly-flooded Marsh 1.1%

Swam p Swamp 1.1%

Tidal Swamp Tidal Swamp < 1%

Tidal Fresh Marsh Tidal Fresh Marsh < 1%

Prepared for Gulf of Mexico Alliance 9 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 4. 2004 NWI layer (released by NWI in 2009 based on 2004 aerial photography)

According to the NWI data layer and project stakeholders, there are various impounded and/or diked areas within Galveston Bay. SLAMM assumes that land behind dikes will be managed and that these regions will not be subject to inundation until local SLR exceeds two meters. The areas designated as diked in the SLAMM model of Galveston Bay are represented by the red areas in Figure 5. Dikes beyond those designated in the NWI data layer were added based on stakeholder feedback. Additional impounded areas included:

1. Area surrounding High Island, TX. The dike near High Island was added based on information obtained during a project applying the SLAMM model in modeling Jefferson County, TX (Clough and Larson 2009). 2. Wetlands adjacent to Lake Anahuac and the Trinity River Dam. The diked designation of the land to the west of Lake Anahuac was determined to be protected based on elevation data and satellite imagery which suggests the area is surrounded in part by a berm (Dick 2010). 3. The area surrounding Texas City, TX. 4. Dredge spoil placement areas in the Houston Ship Channel.

Prepared for Gulf of Mexico Alliance 10 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay 5. Portions of the Anahuac NWR. These areas in the eastern portion of the study area are diked and subject to considerable management (Walther 2011).

Figure 5. Land in Galveston Bay protected by dikes (red)

Prepared for Gulf of Mexico Alliance 11 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

In addition to adding diked areas, a “freshwater-flow polygon” was incorporated in the Trinity River delta to delineate the Wallisville Lake project. The extent of the freshwater-flow polygon is shown in Figure 6. Within this region, swamp is predicted to convert to tidal swamp and tidal swamp to tidal fresh marsh when they are predicted to fall below their lower elevation boundary (relative to water levels). While the freshwater-flow polygon affects the habitat switching functions, SLAMM does not directly integrate rates of freshwater inflow into the switching functions.

Figure 6. Diked areas (in yellow) near the mouth of the Trinity River. Wallisville Lake salt-barrier project area (freshwater-flow region) outlined in yellow

Historic SLR trends have been measured at two sites in the study area: Galveston Pier 21 (6.39 ± 0.28 mm/year) on the Bay side of Galveston Island and Galveston Pleasure Pier (6.84 ±0.81 mm/year) on the Ocean side of Galveston Island. The observed rate of SLR at these gauges has been significantly higher than the average for the last 100 years (approximately 1.7 mm/year, IPCC 2007).

The higher-than-average historic SLR observed in Galveston Bay can be attributed to land subsidence. However, because of decreased groundwater withdrawals, the pattern of subsidence in the Galveston area significantly changed after 1978 (Gabrysch and Coplin 1990). In addition, recent measurements in East Houston have shown that historic subsidence in this area has stopped completely (Buckley et al. 2003). Given that both the simulations started after 1978, a rate of 0.305 m/century (1 ft/century) was applied to both the hindcast and forecast modeling efforts. This parameter choice was based on information from the Harris-Galveston Subsidence District who advised that subsidence from anthropogenic sources is not anticipated in the future (Michel 2010).

Prepared for Gulf of Mexico Alliance 12 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay This “natural subsidence rate” of 3.05 mm/yr was applied within the model by modifying the “Historic Trend” parameter for model forecasts (Table 4)2. A rate of 3.05 mm/year is lower than subsidence that would be estimated using measured historic SLR trends from Galveston Island (5.1 mm/year at Galveston Pleasure pier and 4.7 mm/yr at Pier 21)3. This discrepancy may be caused by the averaging period for these gauges as they include years prior to 1978. Additionally, there may be a small region of higher subsidence in the Galveston Island portion of the study area. Lacking spatial data showing a local pocket of high subsidence in that part of the study area, it was determined that applying the “natural subsidence rate” across the study area was the best approach.

Applying the projected “natural subsidence rate” of 3.05 mm/year across the entire study area may be considered to be somewhat conservative with respect to subsidence in the region. If groundwater withdrawals resume or if there are local pockets of higher subsidence, SLR effects would be presumed to be more severe in those cases.

Several tide gauges and tide tables (2010 predictions) were used to define the tide ranges for Galveston Bay (Table 2, Figure 7). The great diurnal tide range at this site varied from 0.25 m to 0.65 m. Tide data were assigned over twelve input subsites (Figure 8).

Table 2. NOAA gauges used for determining Tide Range and Salt Elevation Great Description Diurnal Tide Data sources/NOAA station ID Range (m) Middle Bay 0.3444 Average of 8770933 and 8771013 East Bay 0.37 Tide Table Urban Tributaries 0.4998 8770777 Open Ocean 0.5973 Average of 8771416 and 8771341 Smith Point/Anahuac 0.249 8772132 Outer Bay Entrance 0.5151 8771341 West Bay & West Bay-2 0.32 Average of 8772132 and Tide Table values Inner Bay Entrance 0.4282 Average of 8771328 and 8771450 HSC (back bay) 0.4511 8770733 Trinity Delta 0.3 Tide Table Northern Bay 0.415 Average of 8770559 and 8770613

2 The “Historic Trend” parameter is used to input an estimate of historic local SLR. The difference between this historic local trend and the historic eustatic trend is then used to adjust global estimates of SLR utilized by SLAMM. In model forecasts the “Historic Trend” parameter was set to 4.75 mm/yr, which is equal to the 1.7 mm/year historic eustatic SLR trend plus the 3.05 mm/yr local subsidence rate. The model then interprets this parameter by applying a subsidence rate of 3.05 mm/year throughout the study area. 3 For example, at Galveston Pleasure Pier, 6.8 mm/year observed minus 1.7 mm/year of eustatic SLR observed would suggest a rate of 5.1 mm/year due to subsidence.

Prepared for Gulf of Mexico Alliance 13 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 7. Great Diurnal Tide Range (GT). Red squares (purple labels) represent values from NOAA tidal datums, Blue squares (pink labels) represent values from NOAA tide prediction tables

Prepared for Gulf of Mexico Alliance 14 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Trinity Delta

Urban Tributaries HSC

Northern Bay Smith Point /Anahuac

Middle Bay East Bay

Outer Bay Entrance

West Bay Inner Bay Entrance

Open Ocean

West Bay - 2

Figure 8. Input subsites

The “salt elevation” parameter within SLAMM designates the boundary between coastal wetlands and dry lands or fresh water wetlands. An estimate of this elevation may be derived by examining historical tide gauge data to determine how frequently different elevations are flooded with ocean water. Within SLAMM modeling simulations this elevation is usually defined as the elevation over which flooding is predicted less than once in every 30 days. Dry lands and fresh-water wetlands are assumed to be located above that elevation.

Prepared for Gulf of Mexico Alliance 15 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

In areas with microtidal regimes such as Galveston, the salt elevation depends on both the lunar effects on the tide as well as wind (Lester and Gonzalez 2002). Therefore a relationship that takes into account the effect of wind was determined to calculate the salt elevation in this study area. The equation:

(1)

where: Salt Elevation = calculated estimate of the “salt elevation” in m above MTL; X = unitless factor that accounts for the effect of local tide range on inundation height; and W = constant effect of wind or barometric pressure, not a function of tide range.

The Excel solver was used to identify where the parameters X and W provided the closest data fit between the observed salt elevations and model predictions. This resulted in a relationship where X = 105% and W = 0.17 m. Data for this analysis were determined from NOAA gauges at Morgan’s Point, Eagle Point, and the Galveston Pleasure Pier.

Equation 1 was used to estimate the salt elevations from the available tide range data, and results were added to the model subsites as shown in Table 3.

Table 3. Salt Elevations Input Subsite Salt Elev. (m above MTL) Middle Bay 0.351 East Bay 0.354 Urban Tributaries 0.432 Open Ocean 0.484 Smith Point/Anahuac 0.301 Outer Bay Entrance 0.440 West Bay & West Bay 2 0.401 Inner Bay Entrance 0.395 HSC (back bay) 0.407 Trinity Delta 0.384 Northern Bay 0.371

Accretion values for Galveston Bay were determined based on a review of eleven peer-reviewed studies, four of which were determined to be applicable for SLAMM modeling: Ravens et al. 2009; White et al. 2002; Williams 2003; Yeager et al. 2007. Figure 9 and Figure 10 present maps indicating where in the study area these accretion rates were measured. These studies/accretion values were selected because they were directly spatially relevant to the study area, were measured in well- described marsh communities and had been published in peer-reviewed journals.

Prepared for Gulf of Mexico Alliance 16 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Location Source Marsh Type Accretion Rate (mm/yr) A Ravens et al. 2008 Salt 3.5 ± 0.10 B Ravens et al. 2008 Salt 1.4 ± 0.04 C Williams 2003 Salt 10.2 D Williams 2003 Salt 5.3 E Williams 2003 Fresh 2.5 F White et al. 2002 Fresh - high 6.5 F White et al. 2002 Fresh - low 3.2 G Yeager et al. 2007 Fresh 1.3 Figure 9. Map of accretion values used

Prepared for Gulf of Mexico Alliance 17 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Location Source Marsh Type Accretion Rate (mm/yr) C Williams 2003 Salt 10.2 D Williams 2003 Salt 5.3 E Williams 2003 Fresh 2.5 F White et al. 2002 Fresh - high 6.5 F White et al. 2002 Fresh - low 3.2 G Yeager et al. 2007 Fresh 1.3 Figure 10. Detail of Accretion Map showing Trinity Bay. Orange line shows the boundary of the study area

Based on these studies, as well as comments from project stakeholders, the study area was divided into high and low sediment areas. The East and West Bays were designated as low sediment supply areas while the remainder of the study is considered a relatively high sediment area.

While running the model calibrations, feedbacks between cell-elevations and predicted accretion rates were incorporated. The basic idea is that cells at lower elevation are inundated more frequently and may also trap sediment more efficiently due to increases in above-ground biomass (for more

Prepared for Gulf of Mexico Alliance 18 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay information on the SLAMM accretion model, see Clough et al. 2010). Figure 11 and Figure 12 show the accretion vs. elevation curves applied to the regularly and irregularly flooded marsh and tidal fresh marsh, respectively. The curves shown in Figure 11 were applied for to both regularly and irregularly-flooded marsh categories. The feedback curve for the low sediment-supply areas was based on data reported by Ravens et al. 2009. The curve for the high sediment-supply areas were derived from data collected by Williams (2003).

Accretion feedback curves do not change the mean accretion rate as estimated from the data, but rather incorporate relationships between accretion rates and marsh elevations. The maximum and minimum accretion values were adjusted in order to calibrate the model. Adjusting the maximum and minimum accretion rates changes the shape of the curves shown in Figure 11 and Figure 12 while keeping the average accretion rate constant. For the high-sediment supply areas the maximum accretion rate applied was 10 mm/yr and minimum was 3.8 mm/yr resulting in an average rate of 7.7 mm/yr. For low sediment supply areas, the maximum accretion rate applied was 4 mm/yr and minimum was 1.6 mm/yr, resulting in an average rate of 3.1 mm/yr.

Figure 11. Accretion feedback curves applied for regularly and irregularly-flooded marshes

Tidal Fresh Marsh accretion feedbacks were also applied to all subsites based on data collected in the Wallisville Lake Project reported by White and coworkers (2002). Figure 12 presents the accretion feedback curve applied to Tidal Fresh Marsh. This curve resulted in an average accretion rate of 4.86 mm/yr. The accretion rate of 2.9 mm/yr applied to Inland Fresh Marsh was derived from the average accretion rate of all fresh marsh values reported by White and coworkers (2002; 4.9 mm/yr) averaged with the rate of 2.5 mm/yr observed by Williams et al. (2003) and the rate of 1.3 mm/yr observed by Yeager and coworkers (2007).

Prepared for Gulf of Mexico Alliance 19 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 12. Feedback curve applied to Tidal Fresh Marsh accretion. Maximum = 6.5mm/yr, Minimum = 3.2 mm/yr

Erosion rates observed from 1931-2000 were applied to the SLAMM model based on data from the Texas Hazard Mitigation Package (Texas Geographic Society, http://www.thmp.info/data_layers/coastal-erosion.html). Using GIS analysis, individual erosion rates were determined for each input subsite. In addition, erosion rates for the West Bay were taken from a more detailed erosion study conducted by Gibeaut and coworkers (2003). Their work suggested the western shore of West Bay and the northern shore of Galveston Island (represented by subsite West Bay 2) had a much higher erosion rate (approximately 2.3 m/yr).

The same observed rate of erosion was applied to all land-cover categories. However, marshes and swamps are only predicted to undergo erosion if they directly interface open water with over 9 km of fetch available for wave setup (see page 79).Table 4 presents a summary of the input parameters used for Galveston Bay, including the erosion rates applied to each input subsite.

Prepared for Gulf of Mexico Alliance 20 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 13. Map of shoreline retreat 1931-2000 (ft per year) from which erosion rates were derived (Texas Hazard Mitigation Package).

Prepared for Gulf of Mexico Alliance 21 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Table 4. Summary of SLAMM input parameters for Galveston Bay

Smith Urban HSC Northern Trinity Middle Subsite Description Point/ Tributaries (back bay) Bay Delta Bay Anahuac NWI Photo Date (YYYY) 2004 2004 2004 2004 2004 2004 DEM Date (YYYY) 2007 2007 2007 2007 2007 2007 Direction Offshore [n,s,e,w] East East South South East East Historic Trend (mm/yr) 4.75 4.75 4.75 4.75 4.75 4.75 GT Great Diurnal Tide Range (m) 0.50 0.45 0.42 0.30 0.34 0.25 Salt Elev. (m above MTL) 0.43 0.41 0.37 0.38 0.35 0.30 Marsh Erosion (horz. m /yr) 2.44 1 1 1 1 0.77 Swamp Erosion (horz. m /yr) 2.44 1 1 1 1 0.77 T.Flat Erosion (horz. m /yr) 2.44 1 1 1 1 0.77 Inland-Fresh Marsh Accr (mm/yr) 2.9 2.9 2.9 2.9 2.9 2.9 Tidal Swamp Accr (mm/yr) 1.1 1.1 1.1 1.1 1.1 1.1 Swamp Accretion (mm/yr) 0.3 0.3 0.3 0.3 0.3 0.3 Beach Sed. Rate (mm/yr) 3 1 1 1 1 1 Hindcast - Use Elev Pre-processor TRUE TRUE TRUE TRUE TRUE TRUE [True,False] Forecast - Use Elev Pre-processor FALSE FALSE FALSE FALSE FALSE FALSE [True,False] Reg Flood Max. Accr. (mm/year) 10 10 10 10 10 4 Reg Flood Min. Accr. (mm/year) 3.8 3.8 3.8 3.8 3.8 1.6 Reg Flood Elev a coeff. (cubic) -1 -1 -1 -1 -1 -1 Reg Flood Elev b coeff. (square) 0.8 0.8 0.8 0.8 0.8 0.8 Reg Flood Elev c coeff. (linear) 1 1 1 1 1 1 Irreg Flood Max. Accr. (mm/year) 10 10 10 10 10 4 Irreg Flood Min. Accr. (mm/year) 3.8 3.8 3.8 3.8 3.8 1.6 Irreg Flood Elev a coeff. (cubic) -1 -1 -1 -1 -1 -1 Irreg Flood Elev b coeff. (square) 0.8 0.8 0.8 0.8 0.8 0.8 Irreg Flood Elev c coeff. (linear) 1 1 1 1 1 1 Irreg Flood D.Effect Max (meters) 0 0 0 0 0 0 Irreg Flood D min. (unitless) 1 1 1 1 1 1 Tidal Fresh Max. Accr. (mm/year) 6.5 6.5 6.5 6.5 6.5 6.5 Tidal Fresh Min. Accr. (mm/year) 3.2 3.2 3.2 3.2 3.2 3.2 Tidal Fresh Elev a coeff. (cubic) 0 0 0 0 0 0 Tidal Fresh Elev b coeff. (square) 0 0 0 0 0 0 Tidal Fresh Elev c coeff. (linear) 1 1 1 1 1 1

Prepared for Gulf of Mexico Alliance 22 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Outer Bay Inner Bay Open Subsite Description East Bay West Bay West Bay 2 Entrance Entrance Ocean NWI Photo Date (YYYY) 2004 2004 2004 2004 2004 2004 DEM Date (YYYY) 2007 2007 2007 2007 2007 2007 Direction Offshore [n,s,e,w] West East East East East East Historic Trend (mm/yr) 4.75 4.75 4.75 4.75 4.75 4.75 GT Great Diurnal Tide Range (m) 0.37 0.32 0.32 0.52 0.43 0.60 Salt Elev. (m above MTL) 0.35 0.40 0.40 0.44 0.39 0.48 Marsh Erosion (horz. m /yr) 0.77 0.77 2.3 1 1 1.68 Swamp Erosion (horz. m /yr) 0.77 0.77 2.3 1 1 1.68 T.Flat Erosion (horz. m /yr) 0.77 0.77 2.3 1 1 1.68 Inland-Fresh Marsh Accr (mm/yr) 2.9 2.9 2.9 2.9 2.9 2.9 Tidal Swamp Accr (mm/yr) 1.1 1.1 1.1 1.1 1.1 1.1 Swamp Accretion (mm/yr) 0.3 0.3 0.3 0.3 0.3 0.3 Beach Sed. Rate (mm/yr) 1 1 1 3 1 1 Hindcast - Use Elev Pre-processor TRUE TRUE TRUE TRUE TRUE TRUE [True,False] Forecast - Use Elev Pre-processor FALSE FALSE FALSE FALSE FALSE FALSE [True,False] Reg Flood Max. Accr. (mm/year) 4 4 4 10 10 4 Reg Flood Min. Accr. (mm/year) 1.6 1.6 1.6 3.8 3.8 1.6 Reg Flood Elev a coeff. (cubic) -1 -1 -1 -1 -1 -1 Reg Flood Elev b coeff. (square) 0.8 0.8 0.8 0.8 0.8 0.8 Reg Flood Elev c coeff. (linear) 1 1 1 1 1 1 Irreg Flood Max. Accr. (mm/year) 4 4 4 10 10 4 Irreg Flood Min. Accr. (mm/year) 1.6 1.6 1.6 3.8 3.8 1.6 Irreg Flood Elev a coeff. (cubic) -1 -1 -1 -1 -1 -1 Irreg Flood Elev b coeff. (square) 0.8 0.8 0.8 0.8 0.8 0.8 Irreg Flood Elev c coeff. (linear) 1 1 1 1 1 1 Irreg Flood D.Effect Max (meters) 0 0 0 0 0 0 Irreg Flood D min. (unitless) 1 1 1 1 1 1 Tidal Fresh Max. Accr. (mm/year) 6.5 6.5 6.5 6.5 6.5 6.5 Tidal Fresh Min. Accr. (mm/year) 3.2 3.2 3.2 3.2 3.2 3.2 Tidal Fresh Elev a coeff. (cubic) 0 0 0 0 0 0 Tidal Fresh Elev b coeff. (square) 0 0 0 0 0 0 Tidal Fresh Elev c coeff. (linear) 1 1 1 1 1 1

Prepared for Gulf of Mexico Alliance 23 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Barrier island overwash was not simulated due to the lack of site specific data and the high uncertainty of this particular submodel. In addition, hindcast results for barrier islands compared reasonably well with observed data when the overwash model is excluded.

In April 2011 NOAA released VDATUM correction maps for the shoreline of Texas. For Galveston the VDATUM coverage is split into two correction maps: one covering the inner shoreline of the bay (called lagoons) and one covering the open-ocean facing shores of the Bolivar Peninsula and Galveston Island (shelf). In order to incorporate the best data available, MTL- NAVD88 correction values were determined for each area and the two files joined together to produce a map of corrections covering the entire study area, as shown in Figure 14.

Uncertainties in the correction between MTL and NAVD88 were also evaluated as part of the elevation uncertainty analysis (see the Elevation Uncertainty Analysis section on page 80).

Figure 14. Map of VDATUM correction applied to Galveston Bay (m)

For the hindcast, lack of high vertical-resolution historical elevation data required use of the SLAMM pre-processor to estimate all wetland elevations for the hindcasting and model calibration effort. This causes the elevation data for hindcasting to be subject to considerable uncertainty.

The cell-size used for the hindcast, elevation uncertainty, and parameter sensitivity analyses was 30 m by 30 m. For the model forecast a 10 m2 cell size was used. Note that the SLAMM model will track partial conversion of cells based on elevation and slope.

Prepared for Gulf of Mexico Alliance 24 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Four output subsites were used to assess areas of particular interest within Galveston Bay. The location of these subsites is shown in Figure 15.

Houston Ship Channel

Trinity Delta

East Bay

West Bay

Figure 15. Output sites used to analyze Galveston Bay

Prepared for Gulf of Mexico Alliance 25 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Results and Discussion

Hindcast

The analysis of Galveston Bay included a “hindcasting” analysis. Hindcasting is performed by starting a simulation at the photo date of the oldest-available wetlands data, running it through the present day, and comparing the output to present-day NWI data. The primary goals of hindcasting are to assess the predictive capacity of a model and potentially to improve model predictions through calibration. In the case of SLAMM, hindcasting is used to determine whether or not the model is correctly predicting the effect of the observed sea level signal on the wetland types in a given study area.

As with all environmental models, uncertainty within input data and model processes limit model precision. Some error within hindcast results are likely caused by the relative simplicity of the SLAMM model. Additionally, the NWI data may introduce error due to lack of horizontal precision or misclassified land coverage, while the DEM may introduce error due to limitations in LiDAR accuracy. Additional error is encountered when the DEM data and NWI data were collected during different time-periods (not temporally synoptic). Lack of precise tidal data or other spatial coverages may further reduce model accuracy.

Historical NWI data are known to be less accurate, interpretation-wise, than data collected after 1980. This is a result of more field work and other local data sources being incorporated in the NWI mapping efforts starting in the 1980’s (Dick 2010).

Historical elevation data is usually not used since older technology generally produced low-vertical- resolution data. In the application of SLAMM to Galveston Bay, the lack of historical DEM data was overcome by using the elevation pre-processor. The elevation pre-processor is a module of SLAMM which estimates elevation ranges as a function of tide ranges and estimated relationships between wetland types and tide ranges (Clough et al. 2010).

There are two sources of error in these hindcasting analyses, the first being the minor differences predicted by SLAMM even without a sea-level-rise signal. A simulation with no SLR signal is referred to as a “time-zero” simulation in which SLAMM tries to predict the current condition. Due to local factors, DEM and NWI uncertainty, and simplifications within the SLAMM conceptual model, some cells inevitably fall below their lowest allowable elevation category and are immediately converted. These cells represent outliers on the distribution of elevations for a given land-cover type. Generally, a threshold tolerance of up to 5% change is allowed for in major land cover categories in our analyses. The second source of error would be errors in SLAMM predictions given a SLR signal. This second source of error is generally more interesting and important. We need to determine if SLAMM can accurately predict growth or loss trends for major land-cover categories.

There are several metrics that may be employed to judge the accuracy of the model based on the hindcast results. One possibility would be to see whether SLAMM can predict a given cell’s wetland type correctly. Because of the high potential for local uncertainty the method used for evaluation focuses on assessing the capability of the model to predict overall trends in land-cover over time.

Prepared for Gulf of Mexico Alliance 26 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Additionally, assessing model accuracy on a cell-by-cell basis is a metric that has been shown to be sensitive to the cell size utilized (Chen & Pontius 2010).

The primary metric used to evaluate SLAMM hindcast results in this study is the percent of the land cover lost during the model simulation for the primary wetland/vegetation types. The percentage loss predicted by the model is compared to the percentage of the actual land cover lost determined by comparing the historical and contemporary NWI datasets. In addition, if there are significant spatial trends visible when comparing historical data to present-day data (e.g. the majority of marsh loss occurs in the upper-right quadrant) maps of hindcasting predictions are assessed to ensure that these trends are also captured within the model results itself.

Prepared for Gulf of Mexico Alliance 27 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

A quantitative summary of the comparative results between the calibrated SLAMM model and the observed land coverage in 2004 is illustrated in Table 5. Inland-fresh marsh poses some problems within the SLAMM hindcast at this site. According to historical NWI maps, in 1979 there was a small amount of inland-fresh marsh in the Galveston Bay area, and a predominance of irregularly- flooded marsh. In 2004, a much larger quantity of inland-fresh marsh was present in the study area, particularly in the area north of East Bay as shown in Figure 18 and Figure 19. As noted above, it is likely that this increase in wetland acreage is due to changes in source data and mapping techniques. SLAMM does predict a 20% gain in tidal fresh marsh over the hindcast through the use of the soil saturation algorithm, but is not able to predict the 267% increase in inland-fresh marsh suggested by a comparison of the 1979 and 2004 datasets. In addition, discussions with local experts suggest the inland-fresh marsh north of East Bay is mischaracterized and is more of a saline prairie or transitional marsh (Walther 2011).

On the other hand, SLAMM predicts losses in dry land and regularly-flooded marsh categories that are less pronounced that the observed losses of these land covers. When looking at the results for the entire site, 4% of dry land is predicted to be lost as compared with 9% observed. However, the observed loss is the result of vast conversions of dry land to inland-fresh marsh, which may primarily be the result of changes in mapping techniques. Looking at all salt marshes, SLAMM predicts a 5% loss as compared to a 22% loss. Similarly, there are vast tracts of salt marshes that are predicted to convert to inland-fresh marsh which confounds this metric. For this reason, the “goodness of fit” of this hindcast effort was evaluated on the basis of subsite data in which these vast observed changes to inland fresh marsh did not occur. More details about these metrics may be found below.

Table 5: Hindcast results.

Predicted by Observed % Predicted % 1979 NWI 2004 NWI Land Category Hindcast change 1979- change 1979- (Acres) (Acres) (Acres) 2009†† 2009††

All Dry Land 158,395 128,802 155,249 -18.7% -2.0% All Salt Marsh† 41,296 31,925 37,639 -22.7% -8.9% All Marsh 53,442 66,580 51702 24.6% -3.3% Undeveloped Dry Land 126,988 111,975 124,182 -11.8% -2.2% Developed Dry Land 31,407 16,827 31,067 -46.4% -1.1% Irregularly Flooded Marsh 23,503 23,731 21,694 1.0% -7.7% Regularly Flooded Marsh 17,519 8,110 14,823 -53.7% -15.4% Inland Fresh Marsh 8,866 32,488 10,774 266.4% 21.5% Tidal Flat 4,005 5 5,548 -99.9% 38.5% Tidal Fresh Marsh 3,281 2,168 3,288 -33.9% 0.2% Swamp 1,378 3,989 1,360 189.5% -1.3% † All salt marsh category includes regularly-flooded and irregularly-flooded marsh †† Negative values indicate losses and positive values are gains

Prepared for Gulf of Mexico Alliance 28 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Inland Open Water Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp Tidal Flat Tidal Flat Figure 16. Galveston Bay, 1979, NWI (Hindcast Initial Condition)

Prepared for Gulf of Mexico Alliance 29 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Inland Open Water Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp Tidal Flat Tidal Flat Figure 17. Galveston Bay, 2004, Hindcast Prediction

Prepared for Gulf of Mexico Alliance 30 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Inland Open Water Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp Tidal Flat Tidal Flat Figure 18. Galveston Bay, Forecast Initial Condition (NWI date = 2004)

Prepared for Gulf of Mexico Alliance 31 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 19 illustrates the Smith Point area of Galveston Bay in which inland-fresh marsh expansion is most prevalent. The observed conversion from regularly-flooded or irregularly-flooded marshes to inland-fresh marsh may not be an actual ecological change, but a consequence of changes in protocol by the USFWS approach to wetland coding. As discussed previously, prior to 1980 less ground-truthing of aerial photos was conducted which adds a larger amount of uncertainty to historical NWI data. Understandably, the SLAMM model will not predict a conversion of salt marsh into inland fresh marsh as a function of sea-level rise.

Figure 19. Comparison of wetland coding at Smith Point in 1979 (left) and 2004 (right) Blue and orange areas are regularly- and irregularly-flooded marshes, whereas green represents inland-fresh marshes.

The hindcast results for the complete study area give a different impression than the results of the output sites shown in Figure 15. These subsites were reviewed in closer detail in order to make the process of calibrating the model based on hindcast results more tractable and effective.

Prepared for Gulf of Mexico Alliance 32 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Based on the final hindcast results, SLAMM is performing reasonably well in areas where minimal inland-fresh marsh expansion is observed. Table 6 presents the percent change in Salt Marsh (regularly- and irregularly-flooded marshes) predicted by the hindcast as compared to the percent change observed, broken down by subsite.

• In the West Bay the predicted percentage of salt marsh lost (12%) is quite similar to the percentage observed (14%). • In the East Bay the metrics are confounded by the issues discussed above (Figure 19). • In the Trinity Delta the hindcast results suggest a 1% gain of salt marsh, while the NWI data suggest a gain of only 0.5%. The predicted increase is due to small fringes of low-lying swamps converting to irregularly flooded marsh (as shown by the yellow arrow in the center image of Figure 21). However, mapped spatial distributions of salt marsh (orange and blue regions) are very similar in the observed and predicted maps (Figure 21). Note that SLAMM will not predict the observed conversion of regularly-flooded into irregularly-flooded marsh on the basis of sea-level rise. • In the Houston Ship Channel salt marsh comprises less than 2% of the land coverage and is highly affected by anthropogenic actions (deposition of dredge spoil). In this subsite SLAMM predicted a 9% gain of salt marsh while a 24% loss was observed.

Table 6. Hindcast results for Salt Marsh (regularly-flooded, irregularly-flooded, and transitional marsh) by subsite. Negative values indicate losses and positive values are gains Predicted Observed Full Study Area 9% 23% West Bay 12% 14% East Bay 10% 37% Trinity Delta -1% -0.5% Houston Ship Channel -9% 24%

Specific adjustments were made through examination of output sites during the hindcasting process in order to calibrate the overall Galveston Bay SLAMM model. At the Trinity River, the tidal flat and beach sedimentation rates were increased to 3 mm/year. This choice is supported by studies that have documented higher sediment loading rates in this area of the bay (Santschi et al. 1999; Williams 2003). In the Houston Ship channel, NWI maps indicate addition of dry land or addition of marsh from open water in several areas. This difference between the observed and model predicted land cover may be attributed to dredge and fill activities. As shown in Figure 22, the USACE has placed dredge spoil in the Houston Ship Channel. SLAMM cannot account for anthropomorphic effects, so these observations are not predicted.

Overall, hindcast results for Galveston Bay were considered sufficiently predictive to apply the calibrated SLAMM model to forecast simulations.

Prepared for Gulf of Mexico Alliance 33 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Inland Open Water Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Figure 20. West Bay Hindcast. 1979 NWI (top), Model Predicted 2004 (middle), 2004 NWI (bottom).

Prepared for Gulf of Mexico Alliance 34 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Inland Open Water Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Figure 21. Trinity River Delta Hindcast. 1979 NWI (top), Model Predicted 2004 (middle), 2004 NWI (bottom).

Prepared for Gulf of Mexico Alliance 35 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 22. USACE Dredge and fill activities in the Houston Ship Channel (2004). Designated dredge fill areas shown in orange (http://www.swg.usace.army.mil/OD/PMaps/2004/Houston%20Area/Houston-Ship- Channel.pdf)

Prepared for Gulf of Mexico Alliance 36 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Forecast

SLAMM predicts that Galveston Bay will be impacted by sea level rise, with the largest fractional losses occurring in the tidal-fresh and inland-fresh marsh categories. Conversely, SLAMM predicts gains in regularly-flooded marsh in every SLR scenario simulated (a negative loss value indicates a gain in land-cover type). A maximum of 32% of undeveloped dry land and 13% of developed dry land is predicted to be lost across all SLR scenarios. Spatially, the most severe effects of SLR are predicted to occur in the East and West bays, with significant loss of wetlands also predicted in the Trinity River delta.

Predicted Change of Land Categories by 2100 Given Simulated Scenarios of Eustatic Sea Level Rise Negative rates indicate a percent reduction from the initial condition SLR by 2100 (m) 0.39 0.69 1 1.5 2 Undeveloped Dry Land -14% -17% -21% -27% -32% Inland Fresh Marsh -9% -22% -31% -42% -52% Developed Dry Land -3% -4% -6% -10% -13% Irregularly Flooded Marsh -25% -47% -67% -74% -82% Swamp 2% -2% -5% -10% -14% Regularly Flooded Marsh 50% 94% 126% 143% 182% Tidal Swamp -47% -72% -84% -90% -93% Tidal Fresh Marsh 19% 10% -26% -78% -85% Inland Shore -5% -7% -9% -11% -15% Estuarine Beach -69% -81% -84% -87% -91% Ocean Beach -1% 29% 61% 108% 91% Cypress Swamp -6% -9% -19% -23% -26%

Soil Saturation

SLAMM modeling of the Galveston Bay included the optional soil-saturation model. Within this model, the local fresh water table is estimated based on the maximum elevation of nearby freshwater wetlands. Within 6 km of open ocean the fresh water table is assumed to be influenced by increases in sea-level, with differences most pronounced nearest to the coastline (Clough et al. 2010) using data from Carter et al, 1973).

The soil saturation algorithm within SLAMM, however, searches only in one direction to determine the local fresh-water table. This simplification may lead maps of SLAMM results to include green stripes of fresh marsh moving through what was previously dry land (red).

Despite this minor loss of geographic realism on output maps, the algorithm may provide useful information about the susceptibility of certain dry lands to water saturation due to changes in the water table. For example, the use of the soil saturation module improved the overall hindcast results for Galveston bay predicting conversion of dry land to inland frash marsh in the East Bay/Smith Point region.

Prepared for Gulf of Mexico Alliance 37 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Time Zero

SLAMM can be used to simulate a “time zero” step, in which the results for the initial NWI photo date are produced. As there is no sea level rise, accretion, or erosion imposed in this time step, conversions in land cover types are based solely on comparisons between land elevations and the SLAMM conceptual model. Due to local factors, DEM and NWI uncertainty, and simplifications within the SLAMM conceptual model, some cells inevitably fall below their lowest allowable elevation category and are immediately converted. These cells represent outliers on the distribution of elevations for a given land-cover type. Generally, a threshold tolerance of up to 5% change is allowed for in major land cover categories in our analyses.

Table 7 presents the changes noted between the initial condition and the time zero for the entire Galveston study area, (for all land categories comprising greater than 0.1% of the initial condition). The largest change was noted in inland-fresh marsh in which 7% was initially lost; all other land categories have a change of less than 5% between the initial condition and time zero.

The seven percent loss in inland-fresh marsh at time-zero can be attributed to a conversion of inland-fresh marsh to transitional marsh, predominantly in the Smith Point area (shown in Figure 19). Local National Wildlife Refuge staff indicated that the NWI maps of this area do not accurately describe the current salinity of the marshes in this area, which are around 10 ppt, making these marshes more likely to be transitional salt marshes rather than inland fresh (Walther 2011). Therefore conversion to transitional marsh at time zero in the Smith Point area was allowed to occur.

Table 7. Changes in dominant land cover types at time zero Acres at Percent Initial Acres Change Time Zero

Undeveloped Dry Land Undeveloped Dry Land 363,732.7 353,413.1 -3%

Inland Fresh Marsh Inland Fresh Marsh 125,962.0 117,450.8 -7%

Developed Dry Land Developed Dry Land 108,823.7 108,198.0 -1%

Irregularly Flooded Marsh Irregularly Flooded Marsh 62,876.6 60,044.7 -5%

Swamp Swamp 22,567.6 22,733.1 1%

Regularly Flooded Marsh Regularly Flooded Marsh 21,922.3 21,487.0 -2%

Tidal Swamp Tidal Swamp 5,769.6 5,957.8 3%

Tidal Fresh Marsh Tidal Fresh Marsh 5,617.4 5,571.5 -1%

Because of the large size of the study area it may be helpful to focus discussion of the predicted land-cover changes within individual output sites (Figure 15). In the following section, results of the individual output sites are discussed for the “mid-range” scenarios examined: A1B maximum (0.69 m SLR by 2100) and 1 m by 2100, along with results of the full study area for all SLR scenarios examined.

Prepared for Gulf of Mexico Alliance 38 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Galveston Bay IPCC Scenario A1B-Mean, 0.39 m SLR Eustatic by 2100

Results in Acres Initial 2025 2050 2075 2100

Open Ocean Open Ocean 861591.5 861621.1 861646.4 861671.0 861695.5

Estuarine Open Water Estuarine Open Water 398903.3 408859.1 415961.9 429327.2 443255.9

Undeveloped Dry Land Undeveloped Dry Land 363732.7 339103.1 331780.3 323363.5 313218.7

Inland Fresh Marsh Inland Fresh Marsh 125962.0 126096.9 124100.0 119051.3 114176.4

Developed Dry Land Developed Dry Land 108823.7 108031.2 107651.0 106971.0 106006.1

Irregularly Flooded Marsh Irregularly Flooded Marsh 62876.6 59633.1 58742.3 55627.6 47375.1

Estuarine Open Inland Open Water 26281.9 22101.9 21512.7 21011.0 20644.2

Swamp Swamp 22567.6 23866.8 23661.4 23284.2 22914.1

Regularly Flooded Marsh Regularly Flooded Marsh 21922.3 30387.6 25571.1 24952.4 32939.1

Tidal Swamp Tidal Swamp 5769.6 5822.7 5273.2 4210.4 3077.4

Tidal Fresh Marsh Tidal Fresh Marsh 5617.4 5774.0 6325.9 6697.2 6667.1

Inland Shore Inland Shore 5040.3 4886.6 4865.3 4826.4 4783.3

Estuarine Beach Estuarine Beach 3922.6 3069.2 2479.5 1744.9 1219.5

Riverine Tidal Riverine Tidal 3641.3 1738.5 1510.4 1319.7 1160.4

Ocean Beach Ocean Beach 1055.8 976.7 907.1 933.9 1048.7

Transitional Salt Marsh Transitional Salt Marsh 240.0 11361.3 14959.2 20216.3 23951.5

Cypress Swamp Cypress Swamp 97.7 95.1 93.5 92.6 91.8

Tidal Flat Tidal Flat 12.8 4634.0 11017.8 12758.6 13834.2 Total (incl. water) 2018059.0 2018059.0 2018059.0 2018059.0 2018059.0

Prepared for Gulf of Mexico Alliance 39 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, Initial Condition

Prepared for Gulf of Mexico Alliance 40 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2025, Scenario A1B Mean

Prepared for Gulf of Mexico Alliance 41 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2050, Scenario A1B Mean

Prepared for Gulf of Mexico Alliance 42 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2075, Scenario A1B Mean

Prepared for Gulf of Mexico Alliance 43 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2100, Scenario A1B Mean

Prepared for Gulf of Mexico Alliance 44 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Galveston Bay IPCC Scenario A1B-Max, 0.69 m SLR Eustatic by 2100

Results in Acres

Initial 2025 2050 2075 2100

Open Ocean Open Ocean 861591.5 861628.4 861669.5 861716.0 861814.2

Estuarine Open Water Estuarine Open Water 398903.3 409519.0 418660.9 440976.8 461311.6

Undeveloped Dry Land Undeveloped Dry Land 363732.7 338617.2 329263.4 315822.4 300732.6

Inland Fresh Marsh Inland Fresh Marsh 125962.0 124350.5 117347.9 106672.6 98291.5

Developed Dry Land Developed Dry Land 108823.7 107972.1 107336.2 105998.7 104110.8

Irregularly Flooded Marsh Irregularly Flooded Marsh 62876.6 59180.5 54795.1 40184.1 33521.5

Inland Open Water Inland Open Water 26281.9 22050.8 21333.7 20724.8 20226.8

Swamp Swamp 22567.6 23743.0 23307.6 22732.8 22092.8

Regularly Flooded Marsh Regularly Flooded Marsh 21922.3 31369.2 25616.2 41956.7 42626.7

Tidal Swamp Tidal Swamp 5769.6 5729.7 4650.8 2885.3 1586.8

Tidal Fresh Marsh Tidal Fresh Marsh 5617.4 5879.8 6774.2 7126.3 6169.9

Inland Shore Inland Shore 5040.3 4881.5 4851.0 4783.1 4696.1

Estuarine Beach Estuarine Beach 3922.6 2963.5 2049.0 1184.1 757.4

Riverine Tidal Riverine Tidal 3641.3 1729.2 1443.6 1184.9 991.5

Ocean Beach Ocean Beach 1055.8 985.5 960.5 1138.5 1362.5

Transitional Salt Marsh Transitional Salt Marsh 240.0 12212.9 18939.1 25844.1 25350.3

Cypress Swamp Cypress Swamp 97.7 94.8 93.0 91.8 89.2

Tidal Flat Tidal Flat 12.8 5151.5 18967.2 17036.0 32326.7

Total (incl. water) 2018059.0 2018059.0 2018059.0 2018059.0 2018059.0

Prepared for Gulf of Mexico Alliance 45 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, Initial Condition

Prepared for Gulf of Mexico Alliance 46 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2025, Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 47 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2050, Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 48 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2075, Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 49 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2100, Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 50 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

In the A1B maximum scenario, significant effects of SLR were observed in each of the study area subsites. In the West Bay, irregularly-flooded marsh and inland-fresh marsh are predicted to be converted to regularly-flooded marsh and tidal flat. Both North and South Deer Island are predicted to lose acreage, with nearly complete loss of South Deer by 2100. The protection of dry land was not included in these simulations; therefore, SLAMM predicts effects in developed dry land areas, including the land surrounding the runways at Scholes International Airport. In East Galveston Bay inland-fresh marsh north of the East Bay is predicted to become regularly-flooded marsh and open water, while a significant amount of the marsh on the barrier island is predicted to be permanently inundated, converting to either tidal flat or open water. Near the mouth of the Trinity River SLAMM predicts the land cover richness in the river delta to be reduced. Tidal fresh marsh and swamp are predicted to transition to irregularly-flooded marsh and, to a lesser extent, regularly- flooded marsh. In addition, nearly all the estuarine beach bordering the bay is predicted to be lost by 2100. In the Houston Ship Channel (HSC) soil saturation is predicted to affect the developed area near the interchange of TX 225 and 146. The same mechanism appears to lead to the increased amount of swamp predicted in the vicinity of the Armand Bayou Nature Center. At the northern end of the HSC in the San Jacinto River delta, much of the freshwater swamp and marsh is predicted to transition to regularly-flooded marsh and open water.

Prepared for Gulf of Mexico Alliance 51 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

West Galveston Bay, 2004 (top) and 2100 (bottom), Scenario A1B Maximum

.

Prepared for Gulf of Mexico Alliance 52 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

East Galveston Bay, 2004 (top) and 2100 (bottom), Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 53 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Trinity River, Galveston Bay, 2004 (top) and 2100 (bottom), Scenario A1B Maximum

Prepared for Gulf of Mexico Alliance 54 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Housto n Ship Channel, Galveston Bay, 2009 (left) and 2100 (right), Scenario A1B Maximum

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Prepared for Gulf of Mexico Alliance 55 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Galveston Bay 1 m Eustatic SLR by 2100

Results in Acres

Initial 2025 2050 2075 2100

Open Ocean Open Ocean 861591.5 861637.9 861694.3 861794.0 862110.7

Estuarine Open Water Estuarine Open Water 398903.3 410376.2 421230.5 448796.1 481934.1

Undeveloped Dry Land Undeveloped Dry Land 363732.7 338036.2 325891.8 307460.4 288339.8

Inland Fresh Marsh Inland Fresh Marsh 125962.0 122228.9 110341.3 96733.5 86553.0

Developed Dry Land Developed Dry Land 108823.7 107902.7 106920.0 104768.3 101951.4

Irregularly Flooded Marsh Irregularly Flooded Marsh 62876.6 58562.4 45841.2 31186.1 20772.5

Inland Open Water Inland Open Water 26281.9 21997.8 21159.9 20466.5 19920.9

Swamp Swamp 22567.6 23603.2 22978.6 22159.6 21328.5

Regularly Flooded Marsh Regularly Flooded Marsh 21922.3 32239.9 33001.5 45039.0 49486.8

Tidal Swamp Tidal Swamp 5769.6 5598.7 3909.5 1918.7 918.1

Tidal Fresh Marsh Tidal Fresh Marsh 5617.4 6025.3 7262.7 6723.9 4156.1

Inland Shore Inland Shore 5040.3 4877.0 4823.6 4727.4 4608.7

Estuarine Beach Estuarine Beach 3922.6 2836.1 1642.0 861.5 614.9

Riverine Tidal Riverine Tidal 3641.3 1719.8 1384.8 1083.8 884.9

Ocean Beach Ocean Beach 1055.8 994.2 1047.0 1327.0 1703.0

Transitional Salt Marsh Transitional Salt Marsh 240.0 13338.6 25090.1 33973.5 31591.3

Cypress Swamp Cypress Swamp 97.7 94.4 92.5 90.8 78.7

Tidal Flat Tidal Flat 12.8 5989.6 23747.6 28949.1 41105.6

Total (incl. water) 2018059.0 2018059.0 2018059.0 2018059.0 2018059.0

Prepared for Gulf of Mexico Alliance 56 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, Initial Condition

Prepared for Gulf of Mexico Alliance 57 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2025, 1 m

Prepared for Gulf of Mexico Alliance 58 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2050, 1 m

Prepared for Gulf of Mexico Alliance 59 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2075, 1 m

Prepared for Gulf of Mexico Alliance 60 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2100, 1 m

Prepared for Gulf of Mexico Alliance 61 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

In the 1-m SLR scenario, West Bay irregularly-flooded marsh and inland-fresh marsh are predicted to be lost, with the majority of these areas converting to open water. A small extent of these areas is predicted to persist as regularly-flooded marsh and tidal flat in 2100 under this scenario. Conversely, much of the dry land behind the marshes in the northern part of the West Bay appears to be resilient. Similar to the A1B Max scenario, land surrounding the runways at Scholes International Airport is predicted to convert to regularly-flooded marsh. In East Galveston Bay a significant amount of the marsh around the East Bay is predicted to be permanently inundated, converting to either tidal flat or open water. The only areas predicted to persist as inland-fresh marsh are those designated as diked. Near the Trinity River, tidal fresh marsh and swamp are predicted to transition to irregularly-flooded marsh and regularly-flooded marsh, with an increase in open water, potentially creating new flow paths and connection between the Trinity River and Bay. In the Houston Ship Channel, soil saturation is again predicted to affect developed areas and much of the freshwater swamp and marsh in the San Jacinto River delta is predicted to transition to open water and regularly-flooded marsh. However, the majority of dry land in this subsite appears resilient to inundation.

Prepared for Gulf of Mexico Alliance 62 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

West Galveston Bay, 2004 (top) and 2100 (bottom), Scenario 1 m

Prepared for Gulf of Mexico Alliance 63 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

East Galveston Bay, 2004 (top) and 2100 (bottom), Scenario 1 m

Prepared for Gulf of Mexico Alliance 64 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Trinity River, Galveston Bay, 2004 (top) and 2100 (bottom), Scenario 1 m

Prepared for Gulf of Mexico Alliance 65 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Housto n Ship Channel, Galveston Bay, 2004 (left) and 2100 (right), Scenario 1 m

Open Ocean Open Ocean Estuarine Open Water Estuarine Open Water Undeveloped Dry Land Undeveloped Dry Land Inland Fresh Marsh Inland Fresh Marsh Developed Dry Land Developed Dry Land Irregularly Flooded Marsh Irregularly Flooded Marsh Estuarine Open Inland Open Water Swamp Swamp Regularly Flooded Marsh Regularly Flooded Marsh Tidal Swamp Tidal Swamp Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat Prepared for Gulf of Mexico Alliance 66 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Galveston Bay 1.5 m Eustatic SLR by 2100

Results in Acres

Initial 2025 2050 2075 2100

Open Ocean Open Ocean 861591.5 861655.2 861742.7 862087.3 863041.1

Estuarine Open Water Estuarine Open Water 398903.3 411209.6 425366.6 458358.1 503756.2

Undeveloped Dry Land Undeveloped Dry Land 363732.7 336966.5 319109.4 293582.2 267016.7

Inland Fresh Marsh Inland Fresh Marsh 125962.0 118433.6 99933.9 84201.5 72851.8

Developed Dry Land Developed Dry Land 108823.7 107775.2 106039.8 102485.4 98057.9

Irregularly Flooded Marsh Irregularly Flooded Marsh 62876.6 56805.3 33604.3 20716.4 16200.6 Inland Open Water 26281.9 21931.7 20966.4 20120.0 19520.1

Swamp Swamp 22567.6 23376.0 22437.8 21285.3 20265.4

Regularly Flooded Marsh Regularly Flooded Marsh 21922.3 33482.9 43439.9 57171.7 53236.4

Tidal Swamp Tidal Swamp 5769.6 5330.1 2941.9 1163.9 587.2

Tidal Fresh Marsh Tidal Fresh Marsh 5617.4 6341.9 7504.3 4312.2 1208.5

Inland Shore Inland Shore 5040.3 4870.8 4784.9 4629.8 4486.2

Estuarine Beach Estuarine Beach 3922.6 2608.5 1161.7 651.9 519.3

Riverine Tidal Riverine Tidal 3641.3 1706.3 1299.3 966.5 788.8

Ocean Beach Ocean Beach 1055.8 1011.0 1227.5 1626.3 2194.1

Transitional Salt Marsh Transitional Salt Marsh 240.0 15934.8 37908.8 44255.1 40683.4

Cypress Swamp Cypress Swamp 97.7 93.8 91.8 79.2 75.3

Tidal Flat Tidal Flat 12.8 8525.8 28498.1 40366.0 53570.1

Total (incl. water) 2018059.0 2018059.0 2018059.0 2018059.0 2018059.0

Prepared for Gulf of Mexico Alliance 67 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, Initial Condition

Prepared for Gulf of Mexico Alliance 68 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2025, 1.5 m

Prepared for Gulf of Mexico Alliance 69 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2050, 1.5 m

Prepared for Gulf of Mexico Alliance 70 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2075, 1.5 m

Prepared for Gulf of Mexico Alliance 71 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2100, 1.5 m

Prepared for Gulf of Mexico Alliance 72 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Galveston Bay 2 m Eustatic SLR by 2100

Results in Acres

Initial 2025 2050 2075 2100

Open Ocean Open Ocean 861591.5 861673.2 861818.2 862667.8 864278.8

Estuarine Open Water Estuarine Open Water 398903.3 411636.3 428595.5 464936.4 525128.9

Undeveloped Dry Land Undeveloped Dry Land 363732.7 335754.2 311729.8 278868.3 247710.9

Inland Fresh Marsh Inland Fresh Marsh 125962.0 114486.1 92148.4 74811.3 59861.5

Developed Dry Land Developed Dry Land 108823.7 107626.0 104970.4 99861.4 94269.8

Irregularly Flooded Marsh Irregularly Flooded Marsh 62876.6 53564.5 25764.2 19296.2 11511.5 Inland Open Water 26281.9 21879.2 20793.0 19841.6 19202.3

Swamp Swamp 22567.6 23149.5 21909.7 20519.6 19392.2

Regularly Flooded Marsh Regularly Flooded Marsh 21922.3 35354.0 52511.5 62543.8 61791.1

Tidal Swamp Tidal Swamp 5769.6 5018.8 2230.4 911.1 379.6

Tidal Fresh Marsh Tidal Fresh Marsh 5617.4 6678.2 7050.1 2238.2 862.4

Inland Shore Inland Shore 5040.3 4864.3 4736.6 4545.5 4280.1

Estuarine Beach Estuarine Beach 3922.6 2354.1 882.2 606.7 343.4

Riverine Tidal Riverine Tidal 3641.3 1694.2 1237.5 883.0 752.3

Ocean Beach Ocean Beach 1055.8 1030.2 1379.6 2017.8 2017.0

Transitional Salt Marsh Transitional Salt Marsh 240.0 20167.3 48724.7 53958.8 47095.2

Cypress Swamp Cypress Swamp 97.7 93.4 91.0 76.4 72.0

Tidal Flat Tidal Flat 12.8 11035.6 31486.3 49475.0 59110.1 Total (incl. water) 2018059.0 2018059.0 2018059.0 2018059.0 2018059.0

Prepared for Gulf of Mexico Alliance 73 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, Initial Condition

Prepared for Gulf of Mexico Alliance 74 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2025, 2 m

Prepared for Gulf of Mexico Alliance 75 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2050, 2 m

Prepared for Gulf of Mexico Alliance 76 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2075, 2 m

Prepared for Gulf of Mexico Alliance 77 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Open Ocean Open Ocean

Estuarine Open Water Estuarine Open Water

Undeveloped Dry Land Undeveloped Dry Land

Inland Fresh Marsh Inland Fresh Marsh

Developed Dry Land Developed Dry Land

Irregularly Flooded Marsh Irregularly Flooded Marsh

Estuarine Open Inland Open Water

Swamp Swamp

Regularly Flooded Marsh Regularly Flooded Marsh

Tidal Swamp Tidal Swamp

Tidal Fresh Marsh Tidal Fresh Marsh

Inland Shore Inland Shore

Estuarine Beach Estuarine Beach

Riverine Tidal Riverine Tidal

Ocean Beach Ocean Beach

Transitional Salt Marsh Transitional Salt Marsh

Cypress Swamp Cypress Swamp

Tidal Flat Tidal Flat

Galveston Bay, 2100, 2 m

Prepared for Gulf of Mexico Alliance 78 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Erosion Map

A map of marsh erosion was generated for the A1B-Maximum (0.69 m s by 2100) SLR scenario. This map indicates where the erosion threshold was exceeded and the amount of marsh erosion that occurred from 2004 to 2100. The marsh and swamp erosion rates applied only varied from 0.77 to 2.44 horizontal m per year (as shown in Table 4); however, the observed erosion was spatially variable. SLAMM predicts the greatest amount of marsh and swamp erosion to occur near the Trinity River Delta and on the marshes separating Christmas Bay from West Bay.

Prepared for Gulf of Mexico Alliance 79 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Figure 23. Swamp and Marsh Erosion in Galveston Bay under 1 m SLR by 2100 scenario.

Sensitivity Analysis

“Sensitivity” refers to the variation in output of a mathematical model with respect to changes in the values of the model inputs (Saltelli 2002). In this study, several of the input parameters were examined to determine the model sensitivity to finite variations of their values. Sensitivity analysis was completed using the A1B-Maximum (0.69 meters by 2100) scenario and varying input parameters by 15%.

Prepared for Gulf of Mexico Alliance 80 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay When interpreting the tornado diagrams presented below, the vertical line at the middle of the diagram represents the deterministic model result. Red lines represent model results when the given parameter is reduced by 15% while blue lines represent a positive change (of 15%) in the parameter.

The average sensitivity statistic, shown on the left of the diagram before each parameter has been calculated such that when a 15% change in the parameter results in a 15% change in the relevant model result, the sensitivity is calculated as 100%.

Sensitivity analysis indicated that the inland-fresh marsh category, which makes up 17% of the land in the overall study area, was most sensitive to the rate of SLR applied, followed by the historic trend, the accretion rate, and the “salt elevation” parameters. Figure 24 indicates that a 15% increase in the SLR, historic trend and salt elevation parameters would result in less inland-fresh marsh predicted in 2100 than observed in the baseline result while an increase in the accretion rate would result in more inland-fresh marsh.

Sensitivity of Inland-Fresh Marsh to 15% change in tested parameters

29.5% - Sea Level Rise by 2100 (m)

19.8% - Historic trend (mm/yr)

18.6% - Inland-Fresh Marsh Accr (mm/yr)

15.5% - Salt Elev. (m above MTL)

0.0142% - GT Great Diurnal Tide Range (m)

86,000 88,000 90,000 92,000 Inland-Fresh Marsh

Figure 24. Sensitivity results for inland-fresh marsh

Prepared for Gulf of Mexico Alliance 81 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Irregularly-flooded marsh makes up the majority of salt marsh in the Galveston Bay study area (9% of land area). As shown in Figure 25, this landcover type showed a sensitivity pattern similar to that of inland-fresh marsh, but was much more sensitive to changes in SLR and historic trend parameters. Accretion feedbacks were used for irregularly-flooded marsh, therefore the maximum and minimum accretion rates were examined in the sensitivity analysis. Results show both the irregularly-flooded marsh category is not very sensitive to the maximum or minimum accretion rates applied when varied individually. For example, a 15% reduction in the maximum accretion value lead to an average result of 32,105 acres compared to the baseline result of 32,497 acres (a difference of 392 acres).

Sensitivity of Irreg.-Flooded Marsh to 15% change in tested parameters

60.3% - Sea Level Rise by 2100 (m)

34.6% - Historic trend (mm/yr)

7.99% - Irreg Flood Max. Accr. (mm/year)

5.25% - Salt Elev. (m above MTL)

1.51% - Irreg Flood Min. Accr. (mm/year)

0.561% - Marsh Erosion (horz. m /yr)

0.194% - Tidal Sw amp Accr (mm/yr)

0.155% - Tidal Fresh Max. Accr. (mm/year)

30,000 31,000 32,000 33,000 34,000 35,000 Irreg.-Flooded Marsh

Figure 25. Sensitivity results for irregularly-flooded marsh

Prepared for Gulf of Mexico Alliance 82 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Regularly flooded marsh was found to be sensitive to several model parameters. This category has the highest sensitivity to the parameters that directly define the tidal frame (salt elevation, tide range) and land elevation in general (Sea-Level Rise). Like the results of the irregularly-flooded marsh, regularly-flooded marsh does not appear to be extremely sensitive to the maximum and minimum accretion rates applied. Since accretion feedbacks were applied to this category, average accretion rate was not changed in the sensitivity analysis – rather the maximum and minimum rates were changed, effectively changing the shape of the curves in Figure 11 (on page 19). Therefore this analysis may not capture the full sensitivity of the model to the accretion rate applied.

Sensitivity of Regularly-Flooded Marsh to 15% change in tested parameters

25.1% - Salt Elev. (m above MTL) 18% - GT Great Diurnal Tide Range (m)

12.5% - Sea Level Rise by 2100 (m)

8.16% - Inland-Fresh Marsh Accr (mm/yr)

7.16% - Irreg Flood Max. Accr. (mm/year)

5.2% - Historic trend (mm/yr)

3.63% - Reg Flood Min. Accr. (mm/year)

3.09% - Irreg Flood Min. Accr. (mm/year)

2.3% - Reg Flood Max. Accr. (mm/year)

0.66% - Tidal Sw amp Accr (mm/yr)

0.245% - Marsh Erosion (horz. m /yr)

0.0605% - Sw amp Accretion (mm/yr)

41,000 41,500 42,000 42,500 43,000 43,500 44,000 Regularly-Flooded Marsh

Figure 26. Sensitivity results for regularly-flooded marsh

Prepared for Gulf of Mexico Alliance 83 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Tornado diagrams can also be reversed so that an individual parameter is illustrated in each diagram along with its effects on each type of land cover. After SLR, the SLAMM model appeared to be most sensitive to changes in the historic trend parameter. The historical SLR parameter provides SLAMM with the relationship between local and global SLR rates caused by local subsidence or other factors. Since local subsidence in the Galveston area has changed a great deal over the last several decades, the effect of the uncertainty of the subsidence rate is of particular interest. Figure 27 shows tidal swamp, tidal flat, ocean beach, irregularly-flooded marsh, and estuarine beach to be the most sensitive land cover types to changes in the historic trend parameter, which represents the subsidence rate.

Effect of 15% change in "Historic trend (mm/yr)" parameter

107% - Tidal Sw amp 82% - Tidal Flat 42.2% - Ocean Beach 34.6% - Irreg.-Flooded Marsh 33.5% - Estuarine Beach 33.3% - Tidal-Fresh Marsh 23.2% - Riverine Tidal 19.8% - Inland-Fresh Marsh 19.4% - Cypress Sw amp 16% - Trans. Salt Marsh 7.15% - Estuarine Open Water 7.03% - Undeveloped Dry Land 6.32% - Sw amp 5.2% - Regularly-Flooded Marsh 3.15% - Inland Open Water 3.06% - Developed Dry Land 2.75% - Inland Shore 0.0241% - Open Ocean -15 -10 -5 0 5 10 15 Percent Change

Figure 27. Effect of changes in Historic trend parameter

Prepared for Gulf of Mexico Alliance 84 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Elevation Uncertainty Analysis

An elevation uncertainty analysis was performed for this model application in order to estimate the impact of terrain uncertainty on SLAMM outputs. According to the vertical accuracy reported in the Meta-data associated with the DEM layer, the root mean squared error (RMSE) for these LiDAR data is 18 cm (Texas Water Development Board 2010). The uncertainty associated with the VDATUM correction specific to the Galveston area (RMSE = 10.1) was also included (NOAA 2011).

The means of evaluating elevation data uncertainty was the application of a spatially autocorrelated error field to the existing digital elevation map in the manner of Heuvelink (1998). This approach uses the normal distribution as specified by the RMSE for the dataset and applies it randomly over the entire study area, but with spatial autocorrelation included (Figure 28). Since elevation error is generally spatially autocorrelated (Hunter and Goodchild 1997), this method provides a means to calculate a number of equally-likely elevation maps given error statistics about the data set. A stochastic analysis may then be run (running the model with each of these elevation maps) to assess the overall effects of elevation uncertainty. Heuvelink’s method has been widely recommended as an approach for assessing the effects of elevation data uncertainty (Darnell et al. 2008; Hunter and Goodchild 1997). In this analysis, it was assumed that elevation errors were strongly spatially autocorrelated, using a “p-value” of 0.24954.

Figure 28. An example of a spatially autocorrelated error field used in in the elevation uncertainty analysis. Minimum (blue) = -0.27 meters, Maximum (red) = 0.27 meters

4 A p-value of zero is no spatial autocorrelation and 0.25 is perfect correlation (i.e. not possible). P-values must be less than 0.25.

Prepared for Gulf of Mexico Alliance 85 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay For the uncertainty analysis, 100 iterations were run for the study area with 0.69 m of eustatic SLR by 2100 simulated in each iteration.

Results indicated the effects of LiDAR elevation uncertainty within this modeling analysis were limited. As shown in Table 8, the coefficient of variance (CV) remained below 2% for the majority of land categories with the exception of categories with very low initial acreage, such as ocean beach and tidal swamp.

Table 8. Elevation uncertainty results (acres) Variable Name Min Mean Max Std. Deterministic CV Undeveloped Dry Land 311,873 313,101 314,253 506.5 314,249 0.16% Developed Dry Land 104,420 104,622 104,824 90.5 104,707 0.09% Inland-Fresh Marsh 89,648 91,154 92,476 559.6 88,934 0.61% Regularly-Flooded Marsh 41,339 42,344 43,546 435.7 43,786 1.03% Irreg.-Flooded Marsh 31,904 32,685 33,435 298.2 32,427 0.91% Tidal Flat 26,334 27,196 27,961 374.0 32,317 1.38% Trans. Salt Marsh 24,811 25,805 26,617 361.9 25,953 1.40% Swamp 17,946 18,149 18,417 97.9 17,967 0.54% Tidal-Fresh Marsh 5,480 5,786 6,037 115.0 6,040 1.99% Inland Shore 4,392 4,409 4,428 7.8 4,406 0.18% Tidal Swamp 1,474 1,698 1,939 98.4 1,505 5.79% Ocean Beach 757 847 933 30.1 829 3.55% Estuarine Beach 803 860 919 25.9 787 3.01% Cypress Swamp 71 75 80 2.3 77 3.08%

Examining histograms for regularly-flooded and irregularly-flooded marsh (Figure 29 and Figure 30), it is apparent that the final predicted acreages of these categories vary only slightly with when LiDAR and vertical datum error are accounted for. More generally, the SLAMM model does not appear to be particularly sensitive to errors on this scale. For example, predicted regularly-flooded marsh losses by 2100 (given 0.69 m of eustatic SLR) range from 20% to 24% and irregularly-flooded marsh losses range from 78% to 79% when accounting for LiDAR elevation uncertainty. The tidal swamp category had the largest coefficient of variation; however, loss rates for the category still had a narrow range (86% to 90%, deterministic was 89%).

These model results suggest that a uniform application of the 95th percentile error across an entire digital elevation map is likely an overly conservative approach to evaluating elevation errors. In other words, reported elevation errors for a dataset (RMSE) pertain to an individual cell within the digital elevation map, and are not a measure of the bias of the entire elevation map.

Prepared for Gulf of Mexico Alliance 86 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

"Regularly-Flooded Marsh" 20 15 Frequency 10 5 0

41500 42000 42500 43000 43500

"Regularly-Flooded Marsh"

Figure 29. Frequency Distribution of SLAMM predictions of regularly-flooded marsh by 2100 (in acres)

"Irreg.-Flooded Marsh" 30 25 20 Frequency 15 10 5 0

32000 32500 33000

"Irreg.-Flooded Marsh"

Figure 30. Frequency Distribution of SLAMM predictions of irregularly-flooded marsh by 2100 (in acres)

Prepared for Gulf of Mexico Alliance 87 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

"Tidal Swamp" 10 8 6 Frequency 4 2 0

1500 1600 1700 1800 1900

"Tidal Swamp"

Figure 31. Frequency Distribution of SLAMM predictions of tidal swamp by 2100 (in acres)

Prepared for Gulf of Mexico Alliance 88 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay

Conclusions

Application of a computational model is only as strong as the input data available. In the case of this analysis, model calibration efforts were somewhat confounded by significant changes in NWI coverages (e.g., large areas classified as inland fresh marsh that had previously been designated regularly-flooded marshes or dry lands). There is a high degree of uncertainty regarding which of these changes reflect new NWI data-analysis techniques and which are actual changes in land cover. These uncertainties lead to difficulties in model calibration. Because of the uncertainties in NWI data, marsh accretion rates and beach sedimentation rates were calibrated in areas where NWI coverages appeared to be more “stable.” SLAMM was more successful in predicting extent of marsh losses in these regions, as shown in the analysis of the West Bay output site in particular.

Inland-fresh marsh makes up a large percentage of the marshlands in the 2009 NWI (2004 photo date) data set. A large part of this inland-fresh marsh is located in the area near the Moody and Anahuac National Wildlife Refuges. Based on discussions with refuge staff, the inland fresh marsh designation of these areas is not completely accurate and instead these wetlands are more like a salty prairie or transitional marsh (Walther 2011). This contributes significant uncertainty when evaluating numerical model hindcast results.

Model forecasts suggest Galveston Bay will be significantly impacted by SLR and that different land categories will be susceptible depending on the rate of sea-level rise. Compared to other categories, dry lands suffer the most significant losses under lower SLR scenarios with 17% of undeveloped dry land predicted to be lost by 2100 under the A1B max (0.69 m) SLR scenario. However, marsh losses become more extreme in higher SLR scenarios. For example, at 1m SLR by 2100, 67% of irregularly flooded marsh is predicted to be lost, culminating with a loss of 82% under the 2 m scenario. Conversely, regularly flooded marsh areas are predicted to increase in each scenario, which may be attributed to the conversion of other land cover types to regularly flooded marshlands.

The analyses presented here are based on elevation data which is subject to spatial uncertainty. The LiDAR data used had a RMSE of 18 cm while the VDATUM RMSE is estimated at 10.1 cm. An elevation-uncertainty analysis indicated that model results are not particularly sensitive to errors on this scale.

Sensitivity analysis suggested that the model is most sensitive to the parameters that directly define land elevations (historical sea level rise), as well as those that define the extent of the tidal frame (salt elevation, great diurnal tide range). Tidal swamp, tidal flat, ocean beach, irregularly-flooded marsh, and estuarine beach were shown to be the most sensitive land cover types to changes in the historic trend parameter, which represents the rate of land subsidence.

Marsh accretion rates are a considerable source of uncertainty. Although the best data available were incorporated into this model application, the variability of accretion rates in a Bay as large as Galveston may be substantial. Future model applications will benefit from the continued collection of site-specific data.

Prepared for Gulf of Mexico Alliance 89 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay References

Buckley, S. M., Rosen, P. A., Hensley, S., and Tapley, Byron D. (2003). “Land subsidence in Houston, Texas, measured by radar interferometry and constrained by extensometers.” Journal of Geophysical Research, 108, 13 PP.

Chen, H., and Pontius, R. G. (2010). “Sensitivity of a Land Change Model to Pixel Resolution and Precision of the Independent Variable.” Environmental Modeling & Assessment.

Chen, J. L., Wilson, C. R., and Tapley, B. D. (2006). “Satellite Gravity Measurements Confirm Accelerated Melting of Greenland Ice Sheet.” Science, 313, 1958-1960.

Clark, P. U. (2009). Abrupt Climate Change: Final Report, Synthesis and Assessment Product 3. 4. DIANE Publishing.

Clough, J. S., and Larson, E. C. (2009). SLAMM Analysis of Southern Jefferson County, TX.

Clough, J. S., Park, R. A., and Fuller, R. (2010). “SLAMM 6 beta Technical Documentation.”

Council for Regulatory Environmental Modeling. (2008). Draft guidance on the development, evaluation, and application of regulatory environmental models. Draft, Washington, DC.

Craft, C., Clough, J. S., Ehman, J., Joye, S., Park, R. A., Pennings, S., Guo, H., and Machmuller, M. (2009). “Forecasting the effects of accelerated sea-level rise on tidal marsh ecosystem services.” Frontiers in Ecology and the Environment, 7(2), 73-78.

Darnell, A. R., Tate, N. J., and Brunsdon, C. (2008). “Improving user assessment of error implications in digital elevation models.” Computers, Environment and Urban Systems, 32(4), 268-277.

Dick, J. (2010). Personal Communication “More Questions Regarding Galveston NWI Codes.”

Gabrysch, R. K., and Coplin, L. S. (1990). “Land-surface subsidence resulting from ground- water withdrawals in the Houston-Galveston region, Texas, through 1987.” Report of Investigations, (90-01).

Galbraith, H., Jones, R., Park, R., Clough, J., Herrod-Julius, S., Harrington, B., and Page, G. (2002). “Global Climate Change and Sea Level Rise: Potential Losses of Intertidal Habitat for Shorebirds.” Waterbirds, 25(2), 173.

Gibeaut, J. C., Waldinger, R., Hepner, T., Tremblay, T. A., and White, W. A. (2003). “Changes in Bay Shoreline Position, West Bay System, Texas.” Austin: Texas General Land Office.

Glick, P., Clough, J., and Nunley, B. (2007). Sea-level Rise and Coastal Habitats in the Pacific Northwest: An Analysis for Puget Sound, Southwestern Washington, and Northwestern Oregon. National Wildlife Federation.

Prepared for Gulf of Mexico Alliance 90 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Grinsted, A., Moore, J. C., and Jevrejeva, S. (2009). “Reconstructing sea level from paleo and projected temperatures 200 to 2100 AD.” Climate Dynamics, 34(4), 461-472.

Heuvelink, G. B. M. (1998). Error propagation in environmental modelling with GIS. CRC Press.

Hunter, G. J., and Goodchild, M. F. (1997). “Modeling the uncertainty of slope and aspect estimates derived from spatial databases.” Geographical Analysis, 29(1), 35–49.

IPCC. (2001). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom, 881.

IPCC. (2007). Climate Change 2007: The Physical Science Basis. Cambridge University Press, Cambridge, United Kingdom.

Lee, J. K., Park, R. A., and Mausel, P. W. (1992). “Application of geoprocessing and simulation modeling to estimate impacts of sea level rise on the northeast coast of Florida.” Photogrammetric Engineering and Remote Sensing, 58(11), 1579-1586.

Lester, J. C., and Gonzalez, L. C. (2002). “The State of the Bay: a Characterization of the Galveston Bay Ecosystem.”

Michel, T. (2010). Personal Communication “Phone conversation.”

Monaghan, A. J., Bromwich, D. H., Fogt, R. L., Wang, S.-H., Mayewski, P. A., Dixon, D. A., Ekaykin, A., Frezzotti, M., Goodwin, I., Isaksson, E., Kaspari, S. D., Morgan, V. I., Oerter, H., Van Ommen, T. D., Van der Veen, C. J., and Wen, J. (2006). “Insignificant Change in Antarctic Snowfall Since the International Geophysical Year.” Science, 313(5788), 827-831.

Moorhead, K. K., and Brinson, M. M. (1995). “Response of Wetlands to Rising Sea Level in the Lower Coastal Plain of North Carolina.” Ecological Applications, 5(1), 261-271.

NOAA. (2011). “VDatum: Estimation of Vertical Uncertainties in VDatum - Last revised: March 2011.” (Jun. 17, 2011).

National Wildlife Federation and Florida Wildlife Federation. (2006). An Unfavorable Tide: Global Warming, Coastal Habitats and Sportfishing in Florida.

Park, R. A., Trehan, M. S., Mausel, P. W., and Howe, R.C. (1989). “The Effects of Sea Level Rise on U.S. Coastal Wetlands.” The Potential Effects of Global Climate Change on the United States: Appendix B - Sea Level Rise, U.S. Environmental Protection Agency, Washington, DC, 1-1 to 1-55.

Park, R. A., Lee, J. K., Mausel, P. W., and Howe, R. C. (1991). “Using remote sensing for modeling the impacts of sea level rise.” World Resources Review, 3, 184-220.

Prepared for Gulf of Mexico Alliance 91 Warren Pinnacle Consulting, Inc.

Application of the Sea-Level Affecting Marshes Model to Galveston Bay Park, R. A., Lee, J. K., and Canning, D. J. (1993). “Potential Effects of Sea-Level Rise on Puget Sound Wetlands.” Geocarto International, 8(4), 99.

Pfeffer, W. T., Harper, J. T., and O’Neel, S. (2008). “Kinematic Constraints on Glacier Contributions to 21st-Century Sea-Level Rise.” Science, 321(5894), 1340-1343.

Rahmstorf, S. (2007). “A Semi-Empirical Approach to Projecting Future Sea-Level Rise.” Science, 315(5810), 368-370.

Ravens, T. M., Thomas, R. C., Roberts, K. A., and Santschi, P. H. (2009). “Causes of salt marsh erosion in Galveston Bay, Texas.”

Saltelli, A. (2002). “Sensitivity Analysis for Importance Assessment.” Risk Analysis, 22(3), 579- 590.

Santschi, P. H., Allison, M. A., Asbill, S., Perlet, A. B., Cappellino, S., Dobbs, C., and McShea, L. (1999). “Sediment transport and Hg recovery in Lavaca Bay, as evaluated from radionuclide and Hg distributions.” Environ. Sci. Technol, 33(3), 378–391.

Texas Geographic Society. (n.d.). “Texas Hazard Mitigation Package - Coastal Erosion.” (Jan. 26, 2011).

Texas Water Development Board. (2010). “2006 Texas Water Development Board (TWDB) Lidar: Galveston County.” (Jun. 27, 2011).

Titus, J. G., Park, R. A., Leatherman, S. P., Weggel, J. R., Greene, M. S., Mausel, P. W., Brown, S., Gaunt, C., Trehan, M., and Yohe, G. (1991). “Greenhouse effect and sea level rise: the cost of holding back the sea.” Coastal Management, 19(2), 171–204.

Vermeer, M., and Rahmstorf, S. (2009). “Global sea level linked to global temperature.” Proceedings of the National Academy of Sciences, 106(51), 21527.

Walther, P. (2011). Personal Communication “Galveston historical marsh types question.”

White, William A., Morton, R. A., and Holmes, C. W. (2002). “A comparison of factors controlling sedimentation rates and wetland loss in fluvial-deltaic systems, Texas Gulf coast.” Geomorphology, 44(1-2), 47-66.

Williams, H. (2003). “Modeling Shallow Autocompaction in Coastal Marshes Using Cesium- 137 Fallout: Preliminary Results from the Trinity River Estuary, Texas.” Journal of Coastal Research, 19(1), 180-188.

Yeager, K. M., Santschi, P. H., Rifai, H. S., Suarez, M. P., Brinkmeyer, R., Hung, C. C., Schindler, K. J., Andres, M. J., and Weaver, E. A. (2007). “Dioxin Chronology and Fluxes in Sediments of the Houston Ship Channel, Texas: Influences of Non-Steady-State Sediment Transport and Total Organic Carbon.” Environ. Sci. Technol, 41(15), 5291– 5298.

Prepared for Gulf of Mexico Alliance 92 Warren Pinnacle Consulting, Inc.