Levee Failure in the Sacramento-San Joaquin Delta: Implications for Agriculture, Habitat, and Infrastructure LA221: Quantitative Methods in Environmental Planning Spring 2005 Jubilee Daniels^1 , Kathryn Gaffney^1 , Josh Pollak^2 , and Mat Rogers^3 Clients: Robert Twiss, Ph.D.^4 , Curt Schmutte^5 , and Ken Devore^6 Landscape Architecture and Environmental Planning^1 , College of Natural Resources^2 , Department of Civil and Environmental Engineering^3 , UC Berkeley CALFED Bay Delta Authority^4 , California Department of Water Resources^5 , California Department of Fish and Game^6 /Contents:/ * Abstract <#abstract> * Introduction <#intro> * Background <#back> * Methods <#methods> * Results <#result> * Discussion <#disc> * Conclusions <#conclusion> * References <#refs> __ ------------------------------------------------------------------------ __ _Abstract_ Eight westernmost islands of the Sacramento-San Joaquin Delta were analyzed using Geographic Information Systems (GIS) for landscape changes due to flooding. Agricultural land use, economic value, vegetation types, habitat, species of concern, and infrastructure were modeled for existing conditions. Levee failure would result in the loss of 96% of island land area. The remnants of islands would pose little value to agriculture, resulting in a loss of more than $86 million. Vegetation loss ranged from 65% to 100%. Similar effects were observed in species habitat. The eight islands in our analysis collectively contain 543x106 m3 of accomodation space, which would require more than three days of maximum flow from the Sacramento and San Joaquin Rivers. In the event of sudden, multiple levee failure the risk of saltwater intrusion into the Delta and possible contamination of drinking water supplies is plausible. Action plans must be developed to develop appropriate actions in the event of levee failure. _Introduction_ *Overview:* The Sacramento and San Joaquin River systems drain 40 percent of California’s land area including the western Sierra Nevada, eastern Coast Ranges, and Central Valley. The Sacramento-San Joaquin Delta (Figure 1) is the confluence of these two great rivers and San Francisco Bay. Prior to American settlement, the Delta was an estuarine wetland system of over 1,150 square miles (Ingebritson et al. 2000). Agricultural development in the Delta during the period 1880-1930 created a highly managed system of constructed islands, levees, and canals (Figure 2). Today the primary importance of the Delta is as a drinking water source to two thirds of California’s population (20 million). Land use in the last century has resulted in subsidence of islands to elevations up to 26 feet (8 m) below Mean Sea Level (MSL). Levees surrounding subsided islands are prone to failure. Island flooding caused by levee failure would result in saltwater intrusion into the Delta, posing a serious threat to drinking water quality. In its current state, the integrity of the Delta levee system is not sustainable and will ultimately result in one of two steady states: i) permanent deep water flooding of one or many islands or ii) the managed rebuilding of Delta islands to above MSL. No regulatory levee integrity action plan currently exists. To protect and manage the resources of the Delta, CALFED, the California Bay-Delta Authority was created pursuant to a 1994 federal-state accord. CALFED is a consortium of 12 state and 13 federal agencies with regulatory authority over water and resource management in the Bay-Delta region, including the California Department of Water Resources and California Department of Fish and Game. *Figure 1* *Site Location:* Our clients have identified the eight westernmost Delta islands, Bethel Island, Bradford Tract, Holland Tract, Jersey Tract, Orwood Tract, Palm Tract, Sherman Island, and Twitchell Island (Figure 1) as those where levee failure would pose the greatest threat to water quality. *Problem Statement:* The effect of Delta levee failure and island flooding on landscape is not well understood. To develop an appropriate regional response plan, our clients required modeling of abrupt landscape changes resulting from planned breaches or levee failure. *Research Questions:* Our team explored the impact of levee failure on eight islands in the west Delta. Specifically, we addressed the following questions: 1. What existing habitat for threatened or endangered species in the Delta would be affected by levee failure? 2. What existing agricultural lands would be affected by flooding? Which specific crops are affected? 3. What is the amount of wetland and deep-water habitat that is lost or gained due to flooding? 4. What is the economic value of agricultural land that would be lost due to flooding? 5. What infrastructure (e.g., natural gas wells and highways) is lost to flooding? Research questions were answered by creating a Digital Terrain Model (DTM) of the west Delta area and evaluating the land surface area below MSL and the volume of water that would fill islands in the event of levee failure. Existing vegetation, wildlife habitat, species of concern, agricultural land use, economic value of agriculture, extractive resources, and infrastructure were evaluated on each island. Changes in the above conditions were assessed in the event of flooding each island. *Project Objectives:* Research questions one through five above were answered through the completion of the following objectives: Phase 1: EXISTING CONDITIONS A Build DTM B Cross Sections C 3-D Fly through D Vegetation by Island E Key Species Range by Island F Habitat by Island G Ag Types by Island H Value of Ag by Island Phase 2: POST-FAILURE CONDITIONS I Cut/Fill Analysis J Accomodation Space and Area Lost K Area of Habitat lost/gained by Island L Area of Ag Type/Value lost by Island M Lost Roads and Extractive Resources ------------------------------------------------------------------------ _Background_ The Sacramento and San Joaquin Rivers, California’s largest, drain 40 percent of the state’s land area. The Sacramento-San Joaquin Delta formed over 6000 years ago at the confluence of the Sacramento and San Joaquin Rivers with San Francisco Bay. At the time of American settlement of California, the Delta was a 1,150 square mile tidal estuarine marsh system (Ingebritson et al., 2000). Human activities from the late nineteenth through the twentieth century converted Delta wetlands to the existing network of 90 islands. Between 1880 and 1930 the Delta marshes were reclaimed, largely for agricultural use, with a system of more than 1100 miles (1700 km) of levees (Mount and Twiss 2004). Development began in the late 1850s with passage of the Swamp and Overflow Land Act. The California Board of Swamp and Overflowed Land Commissioners managed “swamps” as reclamation districts, selling land to private entities, with proceeds used to reclaim more wetlands. By 1869, substantial levees had been constructed on Sherman Island and Twitchell Island, mainly by Chinese laborers (Corbley, 2002). Construction of upstream dams in the twentieth century also deprived Delta wetlands of inorganic sediments and allochthonous organic material. The Sacramento-San Joaquin Delta is the center of water resources in contemporary California, conducting 50 percent of the state’s freshwater flows. Federal and State water projects, the California Aqueduct and Delta-Mendota Canal, export up to 7.5 million acre-feet of water from the Delta for agricultural and drinking water use south of the Delta. Agriculture diverts 80 percent of water exported to the Delta. Two-thirds of California’s population gets part of its drinking water from the Delta (Ingebritson et al. 2000). Although greatly reduced, the delta also provides habitat to a variety of aquatic, avian, and terrestrial species. Some of the islands support other uses important to the state and resource management including: natural gas extraction wells, highway infrastructure (specifically State Highways 4, 12, and 160), power transmission lines, airfields, and limited urbanization. Hydrology of the Delta is complex; freshwater flows from the Sacramento and San Joaquin Rivers, movement of salt water through tidal action, and channel bathymetry and configuration effect hydraulic mixing processes. Saline and freshwater are kept in carefully managed balance to protect water quality. Levee breaching would alter hydraulic mixing and saltwater movement into the delta, possibly halting the export of water from the Delta, causing major disruption to the state of California. Historically, levees were constructed using readily available materials and construction methods of the day (Twiss 2005). Construction on hydric wetland soils with improper shoring materials created structurally inadequate levees prone to failure. Failure mechanisms include seepage, boils, and rodent damage. Tidal and boat-generated waves also erode the fine grain mud-silt bottom materials on the channel side of Delta levees (Bauer et al. 2002). Levees are held in private ownership, with scarce state funds for improvement or reconstruction. The integrity of Delta levees was made more perilous by subsidence of the Delta islands the levees surround. Microbial oxidation of organic wetland sediments, dewatering and compression of wetland sediments, burning peat, and poor agricultural practices contributed to land subsidence. Gaseous CO2 fluxes from microbial oxidation accounted for the majority of subsidence in recent years, with subsidence rates of 0.46 to 1.06 cm/yr (Deverel and Rojstaczer 1996). Some Delta islands are now more than 26 feet (8 m) below mean sea level (Mount and Twiss 2004). In addition to threats from structural deficiencies, Delta levees could be ruptured by earthquakes. The Sacramento-San Joaquin Delta is in close proximity to a number of major faults, including the San Andreas, Hayward, and Concord-Green Valley Faults. There is a 62% probability of at least one magnitude 6.7 or greater earthquake in the San Francisco Bay Region by 2032 (US Geological Survey, 2005). Uncompacted sands, silts, clays, and peat soils forming Delta levees and levee foundations would likely liquefy during a seismic event (Boulanger et al., 1998). Subsequent levee failure and island flooding would have profound impacts on island agriculture, wetland habitats, and Delta water quality. Delta levees are also threatened by flooding from high winter or spring streamflows which could overtop or crumble levee systems. Tidal or wind-driven waves and seiches could further erode levees. To address the multiple challenges facing the Sacramento-San Joaquin Delta, the CALFED Bay-Delta Authority, was created in 1994 as a consortium of state and federal agencies and stakeholder groups with legal power and mission to solve water supply and quality issues in the Sacramento-San Joaquin Delta. The objectives of the program are to address the four areas of interest in the delta including: • Providing and maintaining good water quality for all uses, • Improving fish and wildlife habitat, • Balancing water supply and projected demand, and • Reducing risks from deteriorating levees. To meet goals for water supply, water quality, and ecosystem restoration, CALFED has made maintaining the current hydraulic integrity of the delta a priority. The levee program is currently focused on maintaining current island configuration but has no comprehensive plan for maintaining levee integrity. There is no plan for appropriate responses to abrupt changes to the Delta landscape as a result of levee failure. Nor has there yet been a comprehensive assessment of the consequences of Delta island flooding from levee failures. CALFED and member agencies require an assessment of the consequences of levee failure for a single island and for multiple islands. A thorough review of literature yielded no studies that used GIS as a tool to show changes in land use due to levee failures. Hammersmark et al. (2002) used GIS analysis to describe flooding scenarios on the McCormack-Williamson Tract, a Delta island. Levee failure risk assessment and land subsidence processes have been modeled with GIS (Pistocchi et al. 2002; Zhou et al. 2003). Liu et al. (2004) modeled changes in wetland habitats due to levee building. GIS has been used to map floodplain vegetation with a hierarchical ranking into classes of concern (Townsend, 1995). Narumalani (1995) utilized GIS to make management decisions along the Missouri river; imagery for pre- and post flood conditions were compiled to evaluate the environmental effects of flooding. Riha (1997) used GIS for river flood protection modeling by investigating the hydrologic conditions in potentially flooded areas. Beyond GIS analysis, plans for Delta levee and island management include those drafted by CALFED (CALFED 2004; CALFED 2003) and alternate plans drafter by non-government organizations (NGOs). The Delta management plan developed by the NGO Natural Heritage Institute (NHI) includes ambitious recommendations to avoid construction of another water supply canal bypassing the Delta. Among other actions, the plan calls for the decommissioning Friant Dam to restore flows to the San Joaquin River and rebuilding Delta islands with Tule ponds and additions of dredge spoils and rice straw (NHI 1998; 2002). Drastic measures such as these should be considered in GIS analysis to create a steady state solution to stopgap measures currently employed in the Delta to protect water quality. *Figure 2: Jones Tract flooding in 2004.* ------------------------------------------------------------------------ _Methods_ *Data Preparation:* Most of our data given provided by our Client was obtained from the Department of Water Resources and the Department of Fish and Game. Some of the data did not have its projection defined. First we had to deduce what the projection the data was. We guessed the data’s projection defined it in Arctoolbox and brought it into a map with projected data. If the data lined up we assumed we had it in the correct projection. The un-projected data turned out to be in TealeAlbers. We went through and defined the data in TealeAlbers and then re-projected it in NAD 1983 UTM Zone 10N. We set the transformation to NADCON. We then needed to create a shapefile of the eight most western delta islands. We used the shapefile of the reclamation districts and selected the eight most delta islands and saved them as a new shapefile. To complete our pre-processing of data we intersected our data sets from our client and obtain for public sources with our eight-island data file. This allowed us to easily calculate what agricultural, vegetation and infrastructure existing on each island within out study site (Table 1). *Table 1: Data Table* *Creating a surface model:* A surface model of the eight islands was an integral part of determining the effect of widespread levee failure. We used the surface model to estimate the accommodation space and for a cut/fill analysis of the site to determine which parts of the islands would be affected by levee failure. A digital elevation model of the site was obtained from the USGS seamless data distribution system (Table 1). The data is at 1/3 arc-second resolution, which is approximately 10 meter resolution. First, we created a surface model of the present conditions. In order to do this, we created 1 foot contour lines of the site. DEM of Eight Islands Contours of Eight Islands Then, we overlaid polygons of the eight western islands and clipped the contour lines of the site to create contours of each island. Western Islands Island Contours The island contours were then used to create a triangulated irregular network (TIN) of each island. The TINs extended beyond the boundaries of the islands, so the island polygons were used to clip the TINs to a suitable extent. The combination of these TINs comprised a digital terrain model (DTM) of the site. Unclipped TINs Islands Overlaying Unclipped TINs In order to do a cut/fill analysis, we then needed to create a surface model of the islands after the levee failure. We assumed that once the levees failed, the accommodation space would be filled with water. Only areas above sea level would be preserved after levee failure. Using this reasoning, we created a set of contours for each island that included only those elevations above sea level. The contours were used to create a DTM using the same process outlined above. Final Clipped TINs TINs with Elevations Above Sea Level We focused our analysis on Sherman island, the westernmost island, which is important for water quality and agricultural uses. Sherman Final TIN Sherman TIN with Elevations above Sea Level We used the two different surfaces representing pre and post levee failure for a cut/fill analysis. The output of the cut/fill analysis was a raster grid indicating areas that were unaffected, or lost to levee failure. The gird was then overlaid with agricultural uses, habitat information, vegetation information, gas and oil fields, and roads. *Fly-Through:* We downloaded 1m resolution geo-referenced aerial photos for our study site from teraserver. You can down load the photos through Arcmap’s terra server tool or download them directly from the terra server website. We did both. When the aerial phots are downloaded through arcmap they are not permanent unless you make them so. We riggt-clicked on the photo layer and saved it as permanent. If you download another photo or do anything else before doing this, you will loss the photo. The aerial photos as well as the vegetation layer and agricultural layers were used to create a 3D visualization of our study site in arcscene. We set the base height of the aerial photos from our DTM. Under symbology we used stretch standard deviation n=2. Under display we used bileaner interpolation (for continuous data). We then used the bird tool and the animation toolbar in arcscene to create a flythrough, and exported the flythrough as an avi file. An important trick to making the fly-through work is that when you want to stop the recoding you hit escape, this pause the fly-through. You then hit stop on the animation toolbar. *Agricultural Land Use:* Surveys of agricultural land use were obtained from the California Department of Water Resources, Division of Planning and Local Assistance (California Deprtment of Water Resources (DWR), 2005) . The nine west delta islands of concern are within Sacramento and Contra Costa counties, where surveys were conducted in 2000 and 1995 respectively. In the land use survey, each parcel of land was assigned a “class” to codify the type of land use (e.g. field crops, orchards, urban, native vegetation) and a “subclass” denoting the specific crop grown (e.g. asparagus, pears, corn). Urban and native vegetation were removed from that database and only agricultural land was further analyzed. Class and subclass codes were combined to form a “use” field, which identified each crop type with a unique code. At the time of the most current surveys, there were 24 land use classes on the west delta islands, including farmsteads. Crop values were calculated in 1995 and 2000 dollars for each of the 23 crop categories with commodity values from the (United States Department of Agriculture National Agricultural Statistics Service (USDA-NASS) California, 2005). For land uses where only a generic use was specified (e.g. miscellaneous field crops) and for crop types where no values were given for that crop (e.g. sorghum) the average value for the commodity type was used. The DWR land use survey combines melons, squash, and cucumbers into one use value, T9. The average value of watermelon, cantaloupe, honeydew, squash, and fresh market cucumbers was used to value use T9. The value of turf grass was taken as the average retail price of nine varieties of sod taken from two Northern Californian sod companies. Pasture land was given no crop value, as the fields would not be actively hayed if used as pasture and the density of livestock grazed on pastureland could not be determined. Values for the land itself were taken for rural residential, rangeland, vineyards, orchards, field crops, and vegetable crops from the American Society of Farm Managers and Rural Appraisers California Chapter (2003). The present value of crop values from 1995 and 2000 were calculated and applied to parcels in Contra Costa and Sacramento Counties respectively. Crop values were combined with land values to determine the total value of agricultural plots and displayed in GIS. Results of agriculture analysis are shown in Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, and Table 2 and Table 3. *Vegetation:* We intersected the vegetation data file with the delta islands file so that the output contained information from both files. This vegetation file contains aggregate data on agriculture. Since we have more detailed agriculture layers from DWR, we did not include agriculture in this analysis. We then added an area field to the output file in square meters and exported the data as a .dbf file. We opened the .dbf file in MS Excel, converted the areas field to acres, and summarized the area of vegetation type by each island. This is the existing conditions for vegetation in the west delta islands. Next, we repeated the above process but used the flooded and unflooded scenarios to overlay on the vegetation layer. Again, we added and area field, exported the data, converted to acres, and summarized. Next, we calculated the percent change in area of each type of vegetation per island. A map showing existing vegetation is shown in Figure 9 and Figure 10. A map showing vegetation post-failure is shown in Figure 11 and Figure 12. The resulting data is shown in Table 4. *Extractive Resources:* We intersected the gas and oil (resource extraction) data file with the delta islands file so that the output file contained information from both layers. We then added an area field to the new layer in square meters and exported the data set as a .dbf file. We opened the .dbf file in MS Excel, converted area to acres, and summarized the area of resource extraction by each island and resource producer. This is the existing conditions for extraction resources in the west delta islands. Next, we repeated the above process but used the flooded and unflooded scenario files to overlay on the extraction resource data instead of the original eight islands file. Again, we added and area field, exported the data, and summarized. Next, we calculated the percent change in area of resource extraction. A map showing existing resources and post-failure resources is shown in Figure 13. The resulting data is shown in Table 5. *Habitat:* To analyze existing habitat and quality of habitat on the eight west delta islands, we used two data sources: FRAP and CWHR. This both datasets are available for free online and the FRAP data comes ready to load into ArcMAP. FRAP (Fire and Resource Assessment Program) compiled by the California Department of Fire and Forestry Protection. This data source contains land cover information for the state of California. It includes broad categories for urban and agriculture, and more detailed land cover information for vegetation communities, such as Valley Foothill Riparian and Riparian, for the rest of the state. The FRAP dataset comes with land cover values and their associated habitat description. We began this analysis using the FRAP data. We downloaded data for Contra Costa and Sacramento counties and opened the files in ArcMAP. It was set to show the “VALUES” field in discrete colors. The “VALUES” field provides an aggregate value for habitat type, habitat size, habitat density. After loading the files, we exported the data from each county as a .dbf file for use in joining back to specific islands (described in the following paragraphs). We also used this data to create a .txt file containing four fields: value (habitat type code), Whrtype (abbreviated habitat type), Whrsize (tree size class code describes life stage of habitat), and Whrdensity (tree density class code). These fields were grouped by the Value field and saved as a .dbf file for later use in the CWHR BioView. We then overlayed the FRAP data with the eight islands file and evaluated the FRAP data to only show the land cover grid for the eight islands. Next, we selected each of the eight islands individually and exported them as their own shapefile with the same coordinate system as the FRAP data (Albers). Once the islands were isolated, we set the analysis mask in Spatial Analyst to the first of the islands. Then we used the raster calculator in Spatial Analyst to select only the land cover types (or “VALUES”) for the selected island. The output of this process was a data table containing a field ID, “VALUE” (land cover type), and “VALUE COUNT” (how many cells contain the value type). We then joined the resulting data table to the exported FRAP data so that the values had associated land cover types. Again, we exported the data from the resulting joined file. We repeated this process for each of the remaining nine islands. The next step was to summarize the island land cover types according to FRAP in a single table. This involved cutting and pasting the appropriate information from the data exported during the individual island evaluation (described above) into a single spreadsheet. We then summarized the land cover values and count of values (area) by island. Using the flooded data file, we repeated the above steps. We then summarized the existing conditions and post-failure conditions of FRAP data into a single file and calculated the percent change. The existing conditions of land-cover and post –failure conditions are shown in Figure 14 and Figure 15. This information is presented in Table 6. The next step in assessing habitat was to cross the FRAP data with the CWHR database. The CA DFG California Wildlife Habitat Relationships (CWHR) database is a database with a front-end query tool designed to cross-walk with FRAP data. The CWHR database is used to identify what terrestrial species (not including plants or aquatic species) are likely to be found in a specific habitat type. We began by executing a single condition query with the CWHR software using the following parameters: *County: *Contra Costa and Sacramento *Habitats:* Annual Grassland, Estuarine, Freshwater Emergent Wetland, Perennial Grassland, Riverine, Unknown Conifer Type, Unknown Shrub Type, Urban, Valley Foothill Riparian, Water, Deciduous Orchard, Vineyard, Irrigated Grain Crops, Dryland Grain Crops, Irrigated Row and Field Crops, Irrigated Hayfield. (Note: Excepting the Agriculture options, these habitat types were taken from the FRAP analysis of the nine islands described under the FRAP methods section. Habitat types in FRAP that did not appear in CWHR include Unknown Conifer Type, Unknown Shrub Type, and Water. Also, FRAP data only specifies “Agriculture” whereas the CWHR database breaks this category down into such agriculture types as Deciduous Orchard, Evergreen Orchard, Vineyard, etc. Based on our agriculture analysis completed for this report, we included the agriculture categories listed above.) *Habitat Stages:* All habitat stages *Elements to Exclude: *No elements excluded *Select Species: *Included all default species Special Status Category: Federally Threatened, Federally Endangered, California Threatened, California Endangered, California Protected, California Fully Protected, California Species of Concern *Seasonality Pattern: *All Season Categories for Location, All Season Categories for Habitat We saved this query as CWHR2RUN.dbf for use in the BioView analysis discussed below. The next step was to load the .txt file (with the four fields value, Whrtype, Whrsize, and Whrdensity) that we developed earlier using FRAP into the BioView function of the CWHR software. We uploaded the .txt file into BioView. We then loaded the CWHR2RUN.dbf query and included all species from the query. We then ran the BioView analysis. The result of this analysis was a .dbf file for each species that could be present given the parameters of the query (92 species total). This output includes the “VALUE” field of the original FRAP data plus five other fields containing the results of a complex analysis executed by BioView. One field contained the arithmetic mean assessing the quality of habitat that each habitat type offers. We compiled the mean value column from each individual species file into a single spreadsheet and averaged the habitat value for each habitat type across all species. This average only included species that received a value greater than zero for habitat quality. In other words, if a species received a “0” value, meaning that habitat value is not at all suitable for the given species, it was not included in the overall average value of the given habitat type. We also averaged the habitat value across the four subgroups of amphibians, birds, mammals, and reptiles. We saved this spreadsheet as a .dbf file and uploaded it into ArcMAP, joined it on the “VALUES” field to the eight island FRAP data (both existing and post-failure scenarios developed above), and set the display to show the habitat value of the islands. Maps showing habitat values derived from FRAP and CWHR for existing condition and post-failure are shown in Figure 16 and Figure 17. The results of this analysis are shown in Table 7. *Species:* In addition to a habitat analysis, we did a species analysis using the CNDDB database. The California Natural Diversity Database (CNDDB) includes specific sightings of species of concern including plants and animals. The CNDDB database is available from the CA DFG for $300 for a commercial license. It is a software program and database that comes ready to display in ArcMAP. We intersected the CNNDB shapefile with the eight islands shapefile and then exported the data in a .dbf file. We summarized the existing conditions by species richness (number of species sighted) for each island (see Figrue 18). We then created an index using species richness and Global Rank. Global Rank is a score assigned by the CNDDB biology staff that reflects overall condition (rarity and endangerment) of a given species throughout its range. We created a table out of species richness and the index score and uploaded it to ArcMAP, joined it to the shapefile for the eight islands, and graphically displayed the information. We then repeated the analysis using the post-failure scenario. These images can be seen in Figure 19 and Figure 20. *Ranking:* In order to create a priority ranking of the eight islands, we first ranked each of the islands according to habitat, species, and agricultural value. On each list, the most valuable island was given a 1, the second most valuable was assigned 2, and so on to 8. The values for each island were averaged to create an overall ranking. The island with the highest value was the one with the lowest overall ranking. ------------------------------------------------------------------------ _Results_ *Land Remaining Post-Failure:* As Figure 3 shows, very little land remains above sea level post levee failure. The following Figures and Tables show the implications of flooding on agriculture, vegetation, habitat, and species of concern. In certain cases, a single island map is provided (Sherman Island) to provide a more detailed example of how the data looks. *Figure 3* *Agriculture:* Agricultural land lost to flooding totaled 96 percent. Farmsteads and salt-intolerant crops such as melons and cucumbers suffered lower relative losses (Table 2). These land uses were likely sited on higher ground to avoid occasional temporary flooding. The relatively high value of farmsteads and truck crops, which avoided flooding, resulted in a lower percentage economic loss, $85 million (Table 3). Delta growers would be unlikely to continue to use the highly fragmented, remnant islands that survive flooding. Tidal and wave action could likely erode remnant islands. *Figure 4* *Figure 5* *Figure 6* *Figure 7* *Figure 8* *Table 2* *Table 3* *Vegetation:* Figure 9 shows exisitng vegetation, not including agriculture. As is shown, most of the existing vegetation is grass. The client for this project indicated an interest specifically in freshwater and tidal emergent vegetation. The figure below indicates that there are two types of freshwater emergent vegetation (permanently flooded palustrine [of marsh or wetlands] emegents and seasonally flooded palustrine emergents). There is no tidal emergent vegetation. This is likely due to the selected area of study for this project. We specifically focused on the portion of the island inside the levees, also defined as the reclamaiton district. This area did not extend out into the delta channels where tidal emergent vegetation would be likely to occur. Perhaps future studies should include a buffer zone around the islands in attempts to capture vegetation on the outskirts of the islands. *Figure 9* *Figure 10* Again, very little land is left post-failure and so very little vegetation is left. The areas of emergent vegetation are all but extinguished. *Figure 11* *Figure 12* Table 4 shows the vegetation types, existing acres, acres remaining post-flooding, and the percent change in each type by island. Twitchell Island is most affected by flooding, losing 100 percent of its vegetation. Palm Tract is least affected, only losing 65 percent of its vegetation. Twitchell Island and Bethel Island loose all emergent vegetation. *Table 4: Change in Vegetation by Island* *Extraction Resources:* Figure 13 shows the location of extractive resources on the islands both pre- and post-failure while Table 5 provides a summary. As Table 5 shows, Twitchell Island loses all of its existing resources. Palm Tract is least impacted losing only 75 percent. Given that gas and oil can be extracted underwater, levee failure may not totally undermine the ability to continure extracting resources from these locations. However, the existing infrastructure would be inundated and new infrastructure specific for underwater locations would be needed. Furthermore, there may be environmental impacts and permits associated with underwater extraction that are not an issue with the current situation. *Figure 13* *Table 5: Gas and Oil Resources by Island and Owner* ** *Habitat:* Habitat information was derived using FRAP land cover data cross-walked with the CWHR database. Figure 14 shows existing habitat and the related habitat score for each habitat value. As is shown, most existing habitat is located on Bethel Island. This is largely due to the fact that the urbanized areas on Bethel Island are good habitat for many bird species. As is seen in Table 7, birds score a 70.68 for the urban values. *Figure 14* *Figure 15* *Table 6: Average Habitat Score by Island* *Note: *As defined for CWHR, a habitat score of 66 or greater is considered to be of very good habitat quality. *Table 7: Average Habitat Score by Land Cover Value* *Note: *As defined for CWHR, a habitat score of 66 or greater is considered to be of very good habitat quality. Based on the CWHR query, a list of animals was derived that could potentially be found in the delta islands. The list below covers all of the animals associated with the delta vegetation that received a score greater than zero. In other words, if the CWRH query indicated any level of habitat value, it is included in the list below. Species that received an average greater than 66 across all habitat values are listed below. In the CWHR rating scheme, a score of 66 or greater indicates good habitat. *Species Associated With Delta Vegetation* *Species Associated With Delta Vegetation with a Habitat Score Greater than 66* Figure 16 shows remaining habitat location and value. Very little habitat of any value remains post-flooding. *Figure 16* *Figure 17* Table 8 shows existing acres, remaining acres, and percent change by island. The data in this table indicates some error in the assessment of flooded and non-flooded areas. For example, on Bethel Island, no freshwater emergent wetland is included in the existing acres analysis, however it does show up in the remaining acres anaylsis. Remaining acres were calculated by subtracting existing acres from acres flooded due to levee failure. In some cases like with Bethel Island, the flooded acres were greater than the existing acres, leading to a negative number for remaining acres. This discrepancy is likely due to error from using a vector-polygon to overlay on a raster grid dataset. (The work was double checked.) *Table 8: Change in Habitat Type by Island* ** *Species:* The following Figures show the scores associated with the CNDDB data set. Due to the sensitive nature of this data and that unique sightings of ceratin species should not be singled out or identified, analysis of the CNDDB was conducted on an aggregate level. Table 8 lists the species present in the study area currently and post levee failure. Figure 18 shows species richness, or the number of different species sighted, for each of the eight islands. *Figure 18* *Table 8: Species of Concern Sighted in Eight Islands* Figure 19 and Figure 20 show provide the results of an index developed for the this report, described in the Methods section. Table 7 provides a summary of the scored islands. As can be seen in the maps and table, there is not much change in the species score between current and flooded conditions. This is likely due to the fact that most of the habitat on the islands is on the levees and not in the centers of the islands where effects from flooding will be greatest. This could also be due to the fact that species sightings are not location specific, but are assessed throughout a range. So while the exact location of a species sighting may be inundated, the range of the species may remain partially in tact. *Figure 19* *Figure 20* *Table 7: Relative Significance of Islands based on Species Sightings* *Richness: *Overall number of sighted species with a range including island. *Global Rank: *Global Rank reflects overall condition (rarity and endangerment) of an element throughout its range. Ranks are assigned by the CNDDB biology staff following review of all available information. Average Global Ranks are an average for unique sightings of species per island. *Score: *Number of species per island multiplied by the average Global Rank of each species. *Island Ranking:* Sherman island was by far the highest overall island ranking, indicating that it should be the highest on a priority protection list out of the eight islands we examined. *Figure 21* *Accomodation Space:* The total amount of accommodation space, which is equivalent to the volume that would be lost in levee failure, was determined using our cut/fill analysis described in the methods. *Table 11: Accomodation Space by Island* *Figure 21* Our cut/fill analysis found that the majority of each island would be submerged in the event of levee failure. Most of the unaffected areas are unconnected, and they occur mostly on the edges of the islands. We found that levee failure can lead to significant salt-water intrusion, which would affect water quality for drinking supplies, and for agricultural uses. Based on the maximum flow of the rivers that feed into the Delta, we found that it would take about three days to fill the accommodation space. At minimum flow, it would take about 21 days to fill the accommodation space. In both cases, it would take much longer than the calculated time to lower the salinity of the Delta to acceptable levels. ------------------------------------------------------------------------ _Discussion_ *Limitations and Sources of Error:* The greatest limitation that we faced in this project was access to data. In developing the DTM and trying to complete an analysis for future subsidence, we were not able to obtain necessary information. We had hoped to obtain a DTM derived from X-radar data but this information was not made available to us in a timely manner. Similarly, we were not able to locate subsidence values that were alluded to in the Mount and Twiss paper (Mount and Twiss 2004). The data that we did have access to and used for this project came from disparate sources including the clients, the USGS, and FRAP. Much of the data had no projection assigned and had no metadata. A major source of error in our analysis came from using values from the digital elevation model (DEM) to create contours. The main issue with this is that the edges of the islands as defined by the reclamation district boundaries were not represented by contour lines. The 10-meter resolution of the DEM was, in most cases, not detailed enough to capture the levees. Since the levees were not captured in the DEM, the extent of the contour lines was slightly less than the extent of the islands. Sherman Island DEM with Island Extent Sherman Island Contour Lines and Island Extent Sherman TIN and Island Extent As a result, the DTM generated from these contours does not include all space within the actual reclamation district boundaries. We tried to address this issue by using a set of contour lines that included the levees, but this data was not available for our study. Our methods, however, can be easily replicated to create a surface model if detailed contour information for the Delta becomes available. *Figure 22* The discrepancy in our DTM island area coverage translated to the analysis for agriculture. We used the cut/fill analysis (based on the DTM) to derive vector-polygon files for flooded and non-flooded area of each island. It is these vector-polygon files that we overlayed on the agriculture, vegetation, FRAP land cover, and CNDDB datasets to derive existing, flooded, and remaining land conditions. The result is that certain areas of each island were not included in the analysis. Table 12 shows that Jersey Island was reduced in size by 5.8 percent (201 acres) and Sherman Island was reduced by 4.2 percent (427 acres). This lost space could possibly support other types of vegetation that are not identified in this analysis. *Table 12* *Analysis:* Despite our difficulties with accessing data and developing a DTM, our analysis revealed a number of interesting points. Foremost, levee failure in our study islands would lead to almost a complete loss of the islands and their existing resources. Very little land remains above sea-level and wave action may destroy what is left. We found, as expected, that these islands mainly support agriculture and there is little existing habitat. What habitat does exist is not of particularly good quality. None of the islands contained habitat rated as 100 through the CWHR query. It is interesting to note that the land cover types with the highest habitat value were the urban areas due to a high suitability for birds. Agriculture is the main land use in our eight-island study area. A variety of crops are grown in the islands from asparagus to turf grass. The approximate value of the land is 98 million dollars, representing less than one percent of California’s agricultural industry income for 2004 (California Department of Food and Agriculture, 2004). When doing an overall summary of the islands, including agriculture, habitat, and species values, Sherman Island is indicated as the island of highest value. Sherman accounts for over one-third of the overall value from agriculture. Additionally, it has the most varied habitat of all the islands; 12 habitat types (not including agriculture and urban) as defined by FRAP are identified. This is likely part of the reason that Sherman also has the highest number of species sightings. Perhaps the most intriguing information developed through our analysis was the extent of accommodation space in the Delta. Currently, the eight west Delta islands account for 544 million cubic meters. The average flow for the San Joaquin River in 2004 was 1,863 cubic feet per second (cfs). The Sacramento River flow was 25,321 cfs (U.S. Geological Survey, 2004). At this rate, it would take approximately three days of river flow to fill the space in the islands created by subsidence. At minimum flows, it would take approximately 21 days. It is important to bear in mind that this timeframe only account for eight of the approximately 90 islands that make up the Delta. If an earthquake or large storm event were to result in large-scale levee failure, the time to fill accommodation space with fresh river water could be considerably longer. ------------------------------------------------------------------------ _Conclusions_ Agricultural development in the Delta has created a highly managed system of constructed islands, levees, and canals. The agricultural land use practice in the last century leads to oxidation of organic material and has resulted in the subsidence of delta islands to elevations up to 26 feet (8 m) below Mean Sea Level (MSL). Levees surrounding the subsided islands are prone to failure. Island flooding caused by levee failure would result in saltwater intrusion into the Delta, posing a serious threat to the states drinking water supply. Water for 20 million Californians passes passing through the delta. The levee system is currently unstable and unless the stability of the levee system is addressed the levees will continue to break. There is currently no plan for re-enforcing and maintaining the delta levees integrity. Our analysis of the delta’s current and post levee-failure agricultural and habit value has lead to the following conclusions. A levee breach in any of the islands would lead to almost full submergence of the island due to the massive amount of accommodation space in the eight islands and the “bath tub” profiles of the islands. It would take 3 days of maximum stream flow of the Sacramento and San Joaquin Rivers to fill up the eight delta islands after a levee failure, resulting in salt-water intrusion into the delta. In addition, our choice of looking only at the reclamation district boundary of the islands and not buffering each islands, lead to a constructed view of what species and vegetation exist. In future studies the islands should be buffered to also look at the area surrounding each of the islands. In addition the study should be expanded to include all the Delta islands. Other islands have both urban and agricultural land use practices and would include assessing the value of urban land use. Using more detailed then 10m hypsography data for the creation of our surface models would also approve the project. Some of the data provided by our client was not available to the public but was provided by public agencies. However, some of the data had undefined projections and did not include metadata information. In order to facilitate the longevity and usefulness of the data, these agencies should begin to keep rigorous metadata with their data. It is especially important to include the date the data was collected, by whom and the potential sources of error. The data is very valuable information, and CALFED members would benefit if they were able to access each other’s data. To facilitate data sharing we recommend a creation of a Data Center. These data center would facilitate the sharing of the massive amounts of information that exists on the delta. The overall vulnerability of the delta levee system, and the importance of the delta integrity to the water supply of California, make it essential that CALFED develop a plan of action for maintaining the salt and fresh water balance in the delta, even when levee failure occurs. Our assessment of the agricultural and habit value of the eight delta islands should help in accessing the value of maintaining the current levee configuration in the delta, or seeking an alternative strategy for the Sacramento-San Joaquin Delta. *Work Products for the Client:* * The Website (the report) * A CD which includes: o The power-point presentation o All the Map documents o All the data o The poster o The fly-through ------------------------------------------------------------------------ _References_ American Society of Farm Managers and Rural Appraisers, California Chapter (2003). 2003 Trends in Ag Land and Lease Values.Retrieved April 19, 2005 from http://calasfmra.com/landvalues/. Bauer, B. O., Lorang, M. S. and Sherman, D. J. (2002). Estimating boat-wake-induced levee erosion using sediment suspension measurements. Journal Of Waterway Port Coastal And Ocean Engineering-Asce 128(4): 152-162. Boulanger, R. W., Arulnathan, R., Harder, L. F., Torres, R. A. and Driller, M. W. (1998). Dynamic properties of Sherman Island peat. Journal Of Geotechnical And Geoenvironmental Engineering 124(1): 12-20. California Bay Delta Authority website. March 2005. http://calwater.ca.gov/Regions/DeltaRegion_RPI.shtml. California Department of Food and Agriculture. Californian Agricultural Statistics 2003. October 2004. CALFED (California Bay-Delta Authority) (2003). Levee System Integrity Program, Multi-year Program Plan (years 4-7). Sacramento, CA: 13. CALFED (California Bay-Delta Authority) (2004). Ecosystem Restoration Multi-Year Program Plan (year 4-7). Sacramento, CA: 29. California Deprtment of Water Resources (DWR). (2005). Land Use Survey. Retrieved from http://www.landwateruse.water.ca.gov/. Corbley, K. (2002). Before the Levee Breaks. http://www.pobonline.com/CDA/ArticleInformation/Article/1,9169,74573,00.html. Deverel, S. J. and Rojstaczer, S. (1996). Subsidence of agricultural lands in the Sacramento San Joaquin Delta, California: Role of aqueous and gaseous carbon fluxes. Water Resources Research 32(8): 2359-2367. Hammersmark, C. T., Schladow, S. G. and Fleenor, W. E. (2002). Habitat Enhancement on the McCormack Williamson Tract: and Evaluation of Proposed Scenarios with a Hydraulic Model and GIS. American Geophysical Union, Fall Meeting 2002. Ingebritson, S. E., Ikehara, M. E., Galloway, D. L. and Jones, D. R. (2000). Delta Subsidence in California:The Sinking Heart of the State. U.S. Geological Survey. FS-005-00 Liu, H. Y., Zhang, S. K., Li, Z. F., Lu, X. G. and Yang, Q. (2004). Impacts on wetlands of large-scale land-use changes by agricultural development: The Small Sanjiang Plain, China. Ambio 33(6): 306-310. Mount, J. and Twiss, R. (2004). Subsidence, Sea Level Rise, Seismicity in the Sacramento-San Joaquin Delta: Report to the Levee Integrity Subcommittee of the California Bay-Delta Authority Independent Science Board. NHI (Natural Heritage Institute) (1998). An Environmentally Optimal Alternative for the Bay-Delta: A Response to the CALFED Program, Natural Heritage Institute, San Francisco, CA. NHI (Natural Heritage Institute) (2002). Subsided Island Restoration Design in the Sacramento-San Joaquin Delta: A Solution for Levee Fragility and Water Supply Vulnerability in the Delta, Natural Heritage Instutute, Berkeley, CA. Narumalani, S., Zhou, Z., Waltman, W.J., 1995. Evaluating Flood Impact on the Missouri River Corridor Using Remote Sensing and GIS. 1995 AAG Annual Meeting. Pistocchi, A., Zani, O. and Gottardi, G. (2002). GIS-Based Predictive Modelling for Levee Failure Risk Assessment in the Romagna River Basins. EGS XXVII General Assembly, Nice. Riha, J., Drab, A., 1997. GIS, Groundwater and Floodplain Flow Modelling – the Tool for Flood Protection Management. Water Structures Institute, Brno University of Technology. Townsend, P.A., Walsh, S.J., 1995. A Hierarchical Classification Scheme for Mapping Floodplain Vegetation through GIS and A Remote Sensing Time-Series. 1995 AAG Annual Meeting Twiss, R. (2005). Personal Communication. United States Department of Agriculture National Agricultural Statistics Service (USDA-NASS) California.(2005).California Historic Commodity and Principal Crops Data.Retrieved April 19, 2005 from http://www.nass.usda.gov/ca/. US Geological Survey (2005). Major Quake Likely to Strike San Francisco Bay Region Between 2003 and 2032. Accessed March 10, 2005 from http://quake.wr.usgs.gov/research/seismology/wg02/ Zhou, G. Y., Esaki, T. and Mori, J. (2003). GIS-based spatial and temporal prediction system development for regional land subsidence hazard mitigation. Environmental Geology 44(6): 665-678.