Aquatic Pesticide Application Plan for the San Francisco Estuary Invasive Spartina Project

This plan addresses herbicide application activities undertaken by the coalition of ISP partner agencies in the effort to eradicate non-native, invasive Spartina from the San Francisco Estuary.

Annual update prepared by Drew Kerr 2612-A 8th Street Berkeley, CA 94710 [email protected]

Under contract to Olofson Environmental, Inc. Berkeley, California

for the State Coastal Conservancy 1330 Broadway, 13th floor Oakland, Ca 94612-2530

June 2013

Current funding for the San Francisco Estuary Invasive Spartina Project comes from the California State Coastal Conservancy and grants from the California Wildlife Conservation Board. Table of Contents Table of Contents ...... i List of Figures ...... ii List of Tables ...... ii Appendices ...... ii 1. BACKGROUND...... 1 2. STATEMENT OF PURPOSE AND NEED ...... 1 3. DESCRIPTION OF THE WATER BODY SYSTEM ...... 2 Ecology ...... 2 Natural Processes Affecting Water Quality ...... 3 Water Quality ...... 4 Sediment Quality ...... 7 4. DESCRIPTION OF TARGET SPECIES ...... 9 Native Pacific Cordgrass (Spartina foliosa) ...... 9 Atlantic Smooth Cordgrass (Spartina alterniflora) and its Hybrids ...... 10 English Cordgrass (Spartina anglica) ...... 13 Chilean Cordgrass (Spartina densiflora) ...... 13 Salt-Meadow Cordgrass (Spartina patens) ...... 14 5. CONTROL TOLERANCES ...... 15 6. DESCRIPTION OF HERBICIDE ...... 16 Environmental Fate of Herbicides ...... 20 Potential Biological and Ecological Effects ...... 25 Non-Target Aquatic Plants and Algae ...... 27 Aquatic and Benthic Invertebrates ...... 28 Fish ...... 30 Birds ...... 31 Mammals ...... 31 Receiving Water Monitoring Triggers ...... 32 7. DESCRIPTION OF ALTERNATE, NON-CHEMICAL CONTROL METHODS .. 33 ISP-selected Non-chemical Control Methods ...... 33 Hand-pulling and manual excavation ...... 33 Mechanical excavation and dredging ...... 34 Covering/tarping ...... 35 Other Non-chemical Methods Evaluated ...... 36 Mowing, burning, pruning, and flaming ...... 36 Crushing and mechanical smothering ...... 37 Flooding and draining ...... 37 8. HERBICIDE APPLICATION AREAS FOR 2013 ...... 38 Site 1 – Alameda Flood Control Channel, Alameda County ...... 38 Site 2 – Bair & Greco Island Complex, San Mateo County ...... 40 Site 3 – Blackie’s Pasture, Marin County ...... 40 Site 4 – Corte Madera Creek Complex, Marin County ...... 41 Site 5 – Coyote Creek & Mowry Slough Area, Alameda & Santa Clara Counties ...... 41 Site 6 – Emeryville Crescent, Alameda County...... 42 Site 7 – Oro Loma Marsh, Alameda County ...... 43 Site 8 – Palo Alto Baylands, Santa Clara County ...... 43 Site 9 – Tiscornia Marsh (formerly Pickleweed Park), Marin County ...... 44 Site 10 – Point Pinole Regional Shoreline, Contra Costa County ...... 44 Site 11 – Southampton Marsh, Solano County ...... 45 Site 12 – Southeast San Francisco, San Francisco County ...... 45 Site 13 – Whale’s Tail / Old Alameda Creek Complex, Alameda County ...... 46 Site 15 – South San Francisco Bay Marshes, Santa Clara County ...... 47 Site 16 – Cooley Landing, San Mateo County ...... 47 Site 17 – Alameda Island / San Leandro Bay Complex, Alameda County ...... 48 Site 18 – Colma Creek / San Bruno Marsh Complex, San Mateo County ...... 49

San Francisco Estuary Invasive Spartina Project i. 2013 Aquatic Pesticide Application Plan Site 19 – West San Francisco Bay, San Mateo County ...... 50 Site 20 – San Leandro / Hayward Shoreline Complex, Alameda County ...... 50 Site 21 – Ideal Marsh, Alameda County ...... 51 Site 22 – Two Points Complex, Alameda and Contra Costa Counties ...... 51 Site 23 – Marin Outliers, Marin County ...... 52 Site 24 – Petaluma River, Sonoma County ...... 53 Site 26 – North San Pablo Bay, Solano County ...... 53 9. WATER QUALITY MONITORING PLAN (WQMP) ...... 54 Objective ...... 54 Monitoring Site Selection ...... 54 Sampling Design ...... 55 Field Sampling Procedures...... 56 Equipment Calibration ...... 57 Field Data Sheets ...... 57 Sample Shipment ...... 57 Field Variances ...... 57 Sample Analysis ...... 58 Assessment of Field Contamination...... 58 Lab QC & Data Quality Indicators ...... 58 Monitoring Site Descriptions ...... 58 10. APPLICABLE WATER QUALITY BMPS ...... 62 11. REFERENCES ...... 64

List of Figures Figure 1: Locations and mean discharges for municipal wastewater treatment plants in South San Francisco Bay...... 4 Figure 2: Conceptual Model of Possible Exposure of Biological Organisms to Herbicide Mixture Used by the Spartina Control Program ...... 26 Figure 3: Baywide map of 2013 Spartina Control Program treatment sites ...... 38

List of Tables Table 1: Dissolved concentrations of trace metals in water samples ...... 7 Table 2: Total concentrations of trace metals in water samples ...... 7 Table 3: Ranges of trace pollutants in San Francisco Bay sediments ...... 8 Table 4a: Imazapyr herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Estuary ...... 21 Table 4b: Glyphosate herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Estuary ...... 21 Table 5: Summary of water quality monitoring sites for the 2013 season...... 54

Appendices 1. Chemical properties, degradation rates, environmental fate, and toxicity of imazapyr, glyphosate, and aquatic surfactants evaluated for Spartina control 2. Field Data Collection Form 3. Chain of Custody form 4. Laboratory Quality Assurance Plan, Pacific Agricultural Laboratory

San Francisco Estuary Invasive Spartina Project ii. 2013 Aquatic Pesticide Application Plan 5. General Site Safety & Materials Handling Guidelines and Procedures for Spartina Control Projects in the San Francisco Estuary

San Francisco Estuary Invasive Spartina Project iii. 2013 Aquatic Pesticide Application Plan 1. BACKGROUND The San Francisco Bay Estuary (San Francisco Estuary or Estuary) supports the largest and most ecologically important expanses of tidal mudflats and salt marshes in the contiguous western United States. This environment naturally supports a diverse array of native plants and animals, but over the years many non-native species of plants and animals have been introduced to the Estuary, and some now threaten to cause fundamental changes in the structure, function, and value of the Estuary's tidelands. Among these threatening invaders are several species of salt marsh cordgrass (genus Spartina). In the 1970’s, non-native cordgrasses were introduced to the Estuary and began to spread, slowly at first and then much more rapidly as their populations reached critical mass. Though valuable in their native settings, these introduced cordgrasses are highly aggressive in their new environment, and routinely become the dominant plant species in areas they invade. One of the non-native cordgrass species, Atlantic smooth cordgrass (Spartina alterniflora), was rapidly spreading throughout the Estuary at the beginning of ISP’s control efforts, particularly in the South San Francisco Bay (South Bay). Atlantic smooth cordgrass and its hybrids (formed when this species crosses with the native Pacific cordgrass, S. foliosa) threatened the ecological balance of the Estuary. Based on international studies of comparable cordgrass invasions, these hybrids were likely to eventually cause the extinction of native Pacific cordgrass, while choking tidal creeks, dominating newly restored tidal marshes, and displacing thousands of acres of existing shorebird habitat. Once established in this estuary, invasive cordgrasses are able to disperse on the tides to other estuaries along the California coast through seed and/or propagules dispersal. At the start of the baywide implementation of ISP’s control program in 2005-2006, non-native invasive cordgrasses dominated approximately 1,500-2,000 acres (806 net acres) of the San Francisco Estuary in seven counties — on State, Federal, municipal, and private lands— and were spreading at an alarming rate. The California State Coastal Conservancy (Conservancy) initiated the ISP in 2000 to stave off the invasion of non-native cordgrass and its potential impacts. The ISP is a regionally coordinated effort of Federal, State, and local agencies, private landowners, and other interested stakeholders, with the ultimate goal of eradicating non-native cordgrasses from the San Francisco Estuary. The geographic focus of the ISP includes the nearly 50,000 acres of tidally-influenced marshes, mudflats and brackish channels that comprise the estuarine shorelines of the nine Bay Area counties, including Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma Counties.

2. STATEMENT OF PURPOSE AND NEED The purpose of the Spartina control program is to arrest and reverse the spread of invasive non-native cordgrass species in the San Francisco Estuary to preserve and restore the ecological integrity of the Estuary's intertidal habitats and estuarine ecosystem. The Spartina control program is needed to prevent further degradation and loss of the natural ecological structure and function of the San Francisco Estuary. In the absence of any coordinated and wide-ranging control program, within decades one-quarter to one- half (up to 10,000 acres) of the existing intertidal flats were likely to be replaced with

San Francisco Estuary Invasive Spartina Project 1. 2013 Aquatic Pesticide Application Plan dense, invasive cordgrass marsh, and much of the diverse native salt marsh vegetation replaced with nearly homogeneous stands of non-native cordgrass. This ecological conversion would have altered the structure and function of the Estuary, affecting fisheries, migratory shorebirds and waterfowl, marine mammals, endangered fish, wildlife, and plants, as well as tidal sediment transport and the rate, pattern, and magnitude of tidal flows. In addition, invasive cordgrasses would have impeded the plans of the South Bay Salt Pond Restoration Project to restore up to 15,100 acres of diked baylands to native tidal systems. To avoid these consequences, the ISP is implementing a regionally coordinated, long-term management program.

3. DESCRIPTION OF THE WATER BODY SYSTEM

Ecology Like most Pacific estuaries, the majority of the intertidal zone of the San Francisco Estu- ary naturally consists of unvegetated mudflats. Native California tidal marsh vegetation is limited to the upper intertidal zones, above mean sea level in San Francisco and San Pablo Bays. Below mean sea level, waves erode and redeposit the upper layers of bay mud with each tidal cycle. Rich deposits of fine silt and clay from the Sacramento-San Joaquin Delta have accumulated in the Estuary to form highly productive mudflats, with abundant benthic invertebrates. The mudflats provide a critical source of nutrition and energy for resident and migratory shorebirds and waterfowl, with more than one million shorebirds using the Estuary's mudflats and salt ponds during annual migration, and over half of the west coast’s migratory diving ducks making this estuary their winter home. At elevations above mean sea level (in areas that have not been diked and removed from tidal action), are the Estuary's tidal salt marshes. Pacific salt marsh vegetation is more diverse in plant species than its Atlantic counterparts. Until recent decades, the native Pa- cific cordgrass exclusively occupied the lower reaches of the Estuary's tidal salt marshes. At slightly higher elevations exists a tidal marsh plain dominated by low-growing, mostly perennial plants such as pickleweed (Sarcocornia pacifica), saltgrass (Distichlis spicata), and other salt-tolerant herbs. The tidal marsh plain may also be punctuated by salty shallow ponds (pans) harboring specially-adapted invertebrate species, and may be dissected by irregular tidal channels. The high marsh ecotone at the uppermost edge of the marsh supports an even greater diversity of plant species. Many endemic plant and animal species, including some that are rare or endangered, survive only in the Estuary's remaining tidal marshes. They remain at risk of extinction because of the severe decline over the past century in the abundance, distribution, and quality of tidal marshes. Over 90% of the Estuary’s tidal marshes have been destroyed to ac- commodate salt evaporator ponds as well as residential and commercial development. Most of the Estuary's rare species have narrow habitat requirements, and the health of their populations is normally sensitive to structural changes in their habitats - particularly the condition of the marsh vegetation. Degradation of a healthy, diverse native plant as- semblage by a single dominant invader can push rarer species to local extinction.

San Francisco Estuary Invasive Spartina Project 2. 2013 Aquatic Pesticide Application Plan Natural Processes Affecting Water Quality Water quality within the San Francisco Estuary is connected to and affected by complex natural processes at both a regional and local scale. Hydrologic relationships between the Pacific Ocean, the Estuary, and the many freshwater tributaries (including the Sacramen- to-San Joaquin River system) govern salinity levels in different portions of the Estuary and along the Bay margins. Variable natural factors such as tidal cycles, local winds, basin bathymetry, and salinity gradients interact with river flows and affect the circulation of Estuary waters through channels and bays distributing nutrients, suspended solids, and also pollutants. The major processes affecting water quality are described below. Tidal Cycles. The Estuary has two low tides and two high tides every 24.8 hours. During each tidal cycle, an average of about 1.3 million acre-feet of water, or 24 percent of the Bay and Delta’s volume, moves in and out of the Golden Gate. On the flood (incoming) tide, ocean water moves through the Golden Gate and into the Estuary’s southern and northern reaches, raising the water level at the end of the South Bay by more than eight feet, and raising the height of the Sacramento River at the upstream edge of the Estuary by about three feet. It takes about two hours for the flood tide to reach the end of the South Bay and eight hours to reach Sacramento. Sub-regional Conditions. The Suisun and North Bay sub-regions receive the majority of freshwater input from the Sacramento and San Joaquin River system. In the open bays, density-driven currents show ebb dominance of the surface water and flood dominance of the bottom water. Waters in these embayments are well oxygenated, with low- to moderate-salinity and high-suspended solids concentrations. Water residence time affects the abundance and distribution of many estuarine organisms, the amount of primary production by phytoplankton, and some of the chemical and physical processes that influence the distribution and fate of pollutants. During low flow periods of the year (late summer), the residence time of freshwater moving from the Delta to the ocean can be relatively long (on the order of months) compared to periods when outflow is very high (winter), when freshwater can move from the Delta to the ocean in days. The Central Bay sub-region is influenced by ocean waters that are cold, saline, and lower in total suspended sediment. Water quality parameters fluctuate less than in other sectors of the Bay due to the predominance of ocean water. Net exchanges of ocean and Bay waters depend on freshwater flow in the Bay, tidal amplitude, and longshore coastal currents. The southern part of San Francisco Bay receives less than 10 percent of the natural freshwater flow into the Bay, but the majority (>75 percent) of wastewater discharges. The largest flow is from San Jose, where approximately 120 million gallons per day (MGD) of treated wastewater are released into Artesian Slough, a tributary to Coyote Creek (Figure 1). This freshwater flow creates a local zone of brackish water in the otherwise saline tip of the South Bay. The rest of the South Bay, because it has so little freshwater input, is essentially a tidal lagoon with a relatively constant average salinity (approximately the same as ocean water, 32 parts per thousand [ppt]). South Bay waters are influenced by Delta outflow only during the winter months, when low-salinity water moves southward into the southern reach displacing the denser saline water northward. In the summer months, however, South Bay currents are largely influenced by wind stress

San Francisco Estuary Invasive Spartina Project 3. 2013 Aquatic Pesticide Application Plan on the surface; northwest winds transport water in the direction of the wind, and the displaced water causes subsurface currents to flow in the opposite direction. Currents and Circulation. Circulation patterns within the Bay are influenced by Delta inflows, gravitational currents, and tide- and wind-induced horizontal circulation. The cumulative effects of the latter three factors on net circulation within embayments tend to dominate over that of freshwater inflows except during short periods after large storm events (Smith 1987). Exchanges between embayments are influenced both by mixing patterns within embayments and by the magnitude of freshwater inflows (Smith 1987). Currents created by tides, freshwater inflows, and winds cause erosion and transport of sediments. Tidal currents are usually the dominant form of observed currents in the Bay. Tidal currents are stronger in the chan- Figure 1. Locations and mean discharges for municipal nels and weaker in the shallows (Cheng and wastewater treatment plants in South San Francisco Bay. Adapted from Schemel et al. 1999, based on Davis Gartner 1984). These processes enhance ex- et al. 1991. change between shallows and channels during the tidal cycle, and contribute significantly to landward mixing of ocean water and seaward mixing of river water. Also, the South Bay begins flooding while San Pablo Bay is still ebbing, making it possible for the South Bay to receive water from the northern reach (Smith 1987). Tides have a significant influence on sediment resuspension during the more energetic spring tide when sediment concentrations naturally increase, and particularly during the ebbs preceding lower low water when the current speeds are highest. Powell et al. (1989), however, observed no correlation between tidal cycle and suspended sediment loads or distribution in the South Bay. Their conclusion was that winds are the most important factor in resuspending sediments in the South Bay, and that sources of sediments are more important than transport of sediment resuspended from other parts of the Bay (Reil- ly et al. 1992). Wind-induced currents have a significant effect on sediment transport by resuspending sediments in shallow waters (Krone 1979; Cloern et al. 1989). An estimated 100 to 286 million cubic yards of sediments are resuspended annually from shallow areas of the Bay by wind-generated waves (Krone 1974; SFEP 1992b).

Water Quality Water quality in the San Francisco Estuary has improved significantly since the enact- ment of the California Water Quality Control Act (Porter-Cologne) in 1969 and the Clean Water Act in 1972. Nevertheless, the Estuary waters still carry significant loads of pollu-

San Francisco Estuary Invasive Spartina Project 4. 2013 Aquatic Pesticide Application Plan tants from human sources. Under Section 303(d) of the Clean Water Act, states were required to develop a list of water bodies that do not meet water quality standards; this list is referred to as the “303(d) list.” This list defines low, medium, and high priority pollutants that require immediate attention by State and Federal agencies. Portions of the Estuary have high-priority 303(d) listings for a number of pollutants, including dioxin compounds, furan compounds, PCBs, mercury, copper, nickel, and exotic (plant and animal) species. The most comprehensive information describing water quality in the Estuary comes from the Regional Monitoring Program managed by the San Francisco Estuary Institute (SFEI) and ongoing studies by the Interagency Ecological Program (IEP). In addition, numerous short-term studies that focus on specific sites, resources, or pollutants are conducted on a regular basis by researchers and entities conducting permit-specified monitoring of waste discharges. The primary water quality parameters discussed below are: temperature, salinity, dissolved oxygen (DO), pH, total suspended solids (TSS), turbidity, and pollutants. Temperature. Water temperatures in the Estuary range from approximately 10˚C to 22˚C (50˚F to 71.6˚F). Temperatures are influenced by seasonal solar cycles and variable inputs of river and coastal ocean waters. Temperatures are typically at the higher end of this range along the Estuary margin during daylight hours as the influence of solar energy warms the water. Salinity. The salinity of the Estuary varies spatially and temporally. Along the northern reach the salinity increases from the Delta to the Central Bay as an increasing percentage of the volume of water is tidal. At the mouth of the Sacramento River, for example, the mean annual salinity averages slightly less than 2 ppt, while in Suisun Bay it averages about 7 ppt and at the Presidio in Central Bay it averages about 30 ppt. In the South Bay, salinities remain at near-ocean concentrations (32 ppt) during much of the year, except in the vicinity of the San Jose wastewater outfall at Artesian Slough, where salinities are less concentrated. During summer months in dry years, high evaporation rates may cause salinity in the South Bay to exceed that of ocean water. Seasonal changes in the salinity distribution within the Estuary are controlled mainly by the exchange of ocean and Estuary water, and by river inflow. River inflow has the greater influence on salinity distribution throughout most of the Estuary because inflow varies widely, while variations in ocean inputs are relatively small. In winter, high flows of freshwater from the Delta lower the salinity throughout the Estuary’s northern reach. High Delta flows also intrude into South Bay, lowering salinity there for extended periods. In contrast, during the summer, when freshwater inflow is low, saline water from the Bay intrudes into the Delta. The inland limit of salinity intrusion varies greatly from year to year. In addition, channel dredging can increase gravitational circulation and enhance salinity intrusion (Nichols and Pamatmat 1988). This salt water intrusion into the Delta is predicted to increase substantially with future sea level rise. Dissolved Oxygen. Oxygen concentrations in estuarine waters are increased by the mixing action of wind, waves, and tides, the photosynthesis of phytoplankton and other aquatic plants, and high DO in freshwater inflow. DO concentrations are reduced by plant and animal respiration, chemical oxidation, and bacterial decomposition of organic matter. The Estuary’s waters are generally well oxygenated, except during summer in the extreme southern end of the South Bay where concentrations are reduced by poor tidal mixing and high water temperature. Typical concentrations of DO range from 9 to 10 milli-

San Francisco Estuary Invasive Spartina Project 5. 2013 Aquatic Pesticide Application Plan grams per liter (mg/l) throughout the Estuary during periods of high river flow, 7 to 9 mg/l during moderate river flow, and 6 to 9 mg/l during the late summer months when flows are the lowest. Unlike the 1950s and 1960s, when inadequately treated sewage and processing plant wastes depleted oxygen in parts of the Bay and Delta, today there are few reports of places in the Estuary where low oxygen concentrations adversely affect beneficial uses. Today, the lowest concentrations in the Estuary are typically observed in the extreme South Bay but, in some instances, DO levels in semi-enclosed embayments such as Richardson Bay can be much lower than in the main water body (SFEI 1994). DO will also fluctuate with the tides and temperature such that shallow ponded areas at low tide in full UV exposure may be lower than the associated open water adjacent to the site. pH. The pH of the water in San Francisco Bay is relatively constant and typically ranges from 7.8 to 8.21. Total Suspended Solids (TSS) and Turbidity. Turbidity and TSS are generally used as measures of the quantity of suspended particles. The distinction between the two terms lies mainly in the method of measurement. In general, higher TSS results in more turbid water. Regions of maximum suspended solids occur in the North Bay in the null zone2 (generally 50 to 200 mg/l, but as high as 600 mg/l TSS). The specific location of the null zone changes depending upon freshwater discharge from the Delta. TSS levels in the Estuary vary greatly depending on the season, ranging from 200 mg/l in the winter to 50 mg/l in the summer (Nichols and Pamatmat 1988; Buchanan and Schoellhamer 1995). TSS also varies with tidal stage and depth (Buchanan and Schoellhamer 1995). Shallow areas and channels adjacent to shallow areas have the highest suspended sediment concentrations. The Central Bay generally has the lowest TSS concentrations; however, wind-driven wave action and tidal currents, as well as dredged material disposal and sand mining operations cause elevations in suspended solids concentrations throughout the water column. Pollutants. Pollutant loading to San Francisco Bay has long been recognized as one of many factors that has historically stressed aquatic resources. Pollutants enter the aquatic system through atmospheric deposition, runoff from agricultural and urbanized land, and direct discharge of waste to sewers and from industrial activity. The Bay’s sediment can be both a source and a sink for pollutants in the overlying water column. The overall influx of pollutants from the surrounding land and waste discharges can cause increases in sediment pollutant levels. Natural resuspension processes, biological processes, other mechanical disturbances, dredging, and sediment disposal can remo- bilize particulate-bound pollutants. Metals. Ten trace metals in the aquatic system and in waste discharged to the Bay are monitored on a regular basis. Total and dissolved fractions are sampled three times a year at Regional Monitoring Program (RMP) stations throughout the Estuary. Tables 1 and 2 pre-

1 Water or solutions that are acidic have a pH of less than 7.0, and basic or alkaline water have a pH greater than 7.0. A pH of 7.0 is considered neutral. 2 The null zone is area or region of an estuary where the bottom, high-density and surface, low-density currents have equal and opposite effects. It is defined as the zone where the mean near-bottom speed is zero. The actual location of the null zone migrates in response to changes in river discharge. It is important because it is typically characterized by high concentrations of suspended particulate matter and rapid sediment accumulation.

San Francisco Estuary Invasive Spartina Project 6. 2013 Aquatic Pesticide Application Plan sent dissolved and total trace metal concentration ranges in Bay waters during 1998 (SFEI 1998). Organic Pollutants. Three general types of trace organic contaminants, polycylic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), and pesticides are measured in San Francisco Bay water on a regular basis. Water column concentrations of dissolved and total PAHs in 1998 ranged from 2.1 to 46 parts per trillion (pptr) and from 20 to 300 pptr, respectively (SFEI 1998). Total PCB concentrations in Bay waters during 1998 ranged from 70 to 7,000 parts per quadrillion (ppq), and were below the U.S. Environmental Protection Agency (U.S. EPA) 4-day (chronic toxicity) water quality criteria of 30 pptr. (SFEI 1998). Dissolved PCB concentrations ranged from 12 to 930 ppq. Bay waters also contained measurable concentrations of chlorinated pesticides, including chlordanes and DDTs. Total chlordane concentrations ranged from 21 to 5,700 ppq, while total DDT concentrations ranged from 190 to 9,900 ppq (SFEI 1998). A recent review of historical data from several sources found several previously unidentified organic contaminants in the San Francisco Estuary (SFEI 2002). In this study, p-nonylphenol, a common constituent in detergents and other household products, agricultural surfactants, and many industrial products, was identified in Sacramento and San Joaquin River water (at 19 ng/L and 5 ng/L, respectively), but it was not detected in Estuary water.

Sediment Quality Sediment quality in the Estuary varies greatly according to the physical characteristics of

Table 1. Dissolved Concentrations of Trace Metals in Water Samples (SFEI 1998)

Ag As Cd Cr Cu Hg Ni Pb Se Zn µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L Minimum 0.0002 0.83 0.003 0.09 0.37 0.0003 0.56 0.002 ND 0.07 Maximum 0.006 4.8 0.09 3.8 3.5 0.015 7.2 0.40 6.1 22.5

WQ Criteria 1.9 69 42 1100 5 74 210 90 1-hour WQ Criteria 36 9.3 50 3.1 8.2 8.1 81 4-day

ND – Not detectable at laboratory limits

Table 2. Total Concentrations of Trace Metals in Water Samples (SFEI 1998)

Ag As Cd Cr Cu Hg Ni Pb Se Zn µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L Minimum 0.002 ND 0.009 0.29 0.42 0.0006 0.63 0.05 ND 0.77 Maximum 0.20 9.4 0.36 101 20 0.73 49.0 15.8 6.8 98.6 WQ Criteria 2.3 69 43 1100 2.1 140 58 1-hour WQ Criteria 36 9.3 50 0.025 7.1 5.6 4-day

ND – Not detectable at laboratory limits

San Francisco Estuary Invasive Spartina Project 7. 2013 Aquatic Pesticide Application Plan the sediment, proximity to historical waste discharges, the physical and chemical condition of the sediment, and sediment dynamics that change with location and season. Gen- erally, the level of sediment contamination at a given location will vary depending on the rate of sediment deposition, which varies with seasons and tides (Luoma et al. 1990). Chemical contaminant dynamics in an estuary are closely associated with the behavior of suspended and deposited sediments. The physical and chemical characteristics of sediments, and the bioavailability and toxicity of sediment-associated chemicals to aquatic organisms, are particularly important in determining their potential impact on environmental quality. While pollutant loading to the Estuary from point and non-point sources has declined dramatically over the past two decades, and surface sediment contamination may be de- clining from historical highs, Bay sediments are still an important source and sink of pollutants. Much of the data documenting concentrations of trace metals and organics in Bay sediments are found in the historical summary of Long and Markel (1992) and in the more recent monitoring efforts by the State’s Bay Protection and Toxic Cleanup Program (BPTCP) (SFBRWQCB 1994) and Regional Monitoring Program (SFEI 1994 and 1998). Concentrations of Metals and Organic Pollutants in Sediments. Mean concentrations of trace metals and organics in sediments vary according to grain size, organic carbon con- tent, and seasonal changes associated with riverine flow, flushing, sediment dynamics, and anthropogenic inputs. Anthropogenic inputs appear to have the greatest effect on sediment levels of copper, silver, cadmium, and zinc, as well as several chlorinated and petroleum hydrocarbons (SFBRWQCB 1994). Ranges in sediment metals and trace organic

Table 3. Ranges of Trace Pollutants in San Francisco Bay Sediments (SFEI 1998)

SEDIMENT SAMPLES (MG/KG) EFFECTS LEVELS (MG/KG) Minimum Maximum ER-L ER-M Arsenic 3.1 19 8.2 70 Cadmium 0.1 2.1 1.2 9.6 Chromium 63 216 81 370 Copper 8.5 76 34 270 Lead 5.4 65 46.7 218 Mercury 0.03 0.82 0.15 0.71 Nickel 68 228 20.9 51.6 Selenium 0.06 0.52 Silver ND 2.0 1.0 3.7 Zinc 64 256 150 410

TOTAL PAHS 0.033 6.30 4.022 44.792 Total PCBs ND 0.26 0.0227 0.18 Total DDTs No Data 0.00158 0.0461 Total Chlordanes ND 0.0099 0.0005 0.006 Key: Concentrations bolded exceed the Lowest Observable Effects Level (ER-L) Concentrations bolded and underlined exceed the Median Observable Effects Level (ER-M) ND – Not detectable at laboratory limits

San Francisco Estuary Invasive Spartina Project 8. 2013 Aquatic Pesticide Application Plan concentrations during 1998 are listed in Table 3. The table also compares measured concentrations to effects range-low (ER-L) and effects range-median (ER-M) values, which are levels that are rarely associated with adverse effects to benthic organisms from exposures to sediment-associated contaminants and levels that are frequently associated with adverse impacts, respectively (Long et al., 1995). For most pollutants, ranges in measured concentrations exceed the respective ER-L values but are below the corresponding ER-M values. The exceptions are mercury, nickel, total PCBs, and total chlordanes, which exceed the ER-M values at one or more locations in the Bay. Some sites within San Francisco Bay, such as Lauritzen Canal, the Port of Oakland near San Leandro Bay, and Richmond Harbor, which have been greatly affected by historical contamination, contain sediment pollutant levels which are considerably higher than those measured by the Regional Monitoring Program.

4. DESCRIPTION OF TARGET SPECIES There are one native and four non-native species of cordgrass in the San Francisco Estu- ary. Pacific cordgrass (S. foliosa), the native species, is avoided during ISP’s control efforts and is actually conserved by controlling the invasive species. The non-native species are Atlantic smooth cordgrass (S. alterniflora), English cordgrass (S. anglica), Chilean cordgrass (S. densiflora), and salt-meadow cordgrass (S. patens). Both the non-native At- lantic smooth cordgrass and Chilean cordgrass hybridize with the native Pacific cordgrass, and their offspring (referred to as hybrid S. alterniflora or hybrid S. densiflora) are highly invasive. Key aspects of the cordgrass species found in the Estuary are con- trasted below. The roles these species play in their native habitats give ecologists an indi- cation of their potential to alter the salt marsh ecosystem of San Francicso Bay.

Native Pacific Cordgrass (Spartina foliosa) The historic range of Pacific cordgrass was confined to estuaries from Point Reyes to Ba- ja along the California coast, with large gaps in between (it was historically absent from Monterey Bay and Morro Bay). Most of the Pacific cordgrass population exists in San Francisco and San Pablo Bays, with its northern limit now being Bodega Bay, a small and recently-established natural population. It even more recently established in Tomales Bay, where its population surged following major flood and depositional events of the mid-1990s. Pacific cordgrass is a perennial, salt-tolerant marsh grass, which spreads both sexually by seed dispersal, and asexually by long, creeping rhizomes (underground stems, or runners) that propagate small clusters of leafy shoots. Clonal (asexual) growth of rhizomes allows individual plants to form extensive colonies without being pollinated by another plant. A colony thus formed is referred to as a "clone." The slender leafy shoots seldom exceed five feet in height including seed heads, and most shoots range from about one to three feet tall. The height of the cordgrass plant is related to how well it tolerates submersion in tidewaters, and thus how low in the intertidal zone it can grow. The relatively short stature of Pacific cordgrass corresponds with its limited occupation of lower elevations within the intertidal zone. Pacific cordgrass is genetically very similar to Atlantic smooth cordgrass, but the two species also have significant differences. In size, growth rate, pollen and seed production, culm (stem) density and ecological tolerances, Pacific cordgrass is much less robust than

San Francisco Estuary Invasive Spartina Project 9. 2013 Aquatic Pesticide Application Plan Atlantic smooth cordgrass (Smart and Barko 1978, Callaway 1990, Boyer, Callaway and Zedler 2000). Pacific cordgrass grows more luxuriantly in clayey mud than sand, but it naturally grows in substrates ranging from sand and mud to peat. Its leaves and stems wither in fall and are shed in winter, as the clones die back to the mud substrate. The sparse remains of Pacific cordgrass stands in winter are relatively ineffective in trapping sediment. Pacific cordgrass is generally restricted to a narrow portion of the intertidal zone, between an elevation just above mean sea level and an elevation near the level of the average higher daily tide (mean higher high water, "MHHW"). Although pickleweed is generally the dominant plant of the marsh plain, which approaches the elevation of the MHHW in California estuaries, mature marshes often have short, low-density S. foliosa scattered amongst the pickleweed. The modest range in tidal elevation of Pacific cordgrass restricts the species to the sloping banks and benches of tidal creeks and the gently sloping upper elevation of mudflats. It is the marcrophyte that grows at the lowest elevation in the intertidal zone, leaving vast acreages of Pacific tidal flats below mean sea level entirely free of emergent vegetation in natural historic conditions. Early experiments with Pacific cordgrass demonstrated that its slender, widely spaced leafy shoots and rhizomes are not as effective at stabilizing sediment compared with At- lantic smooth cordgrass, especially under exposed conditions at the bay's edge (New- combe et al. 1979). Seedlings of Pacific cordgrass are seldom found in established marshes, and appear only intermittently in sheltered upper mudflats. Pacific cordgrass is particularly valued as habitat for the endangered California clapper rail, which spends most of its time foraging for food within, or close to, the protective canopy of cordgrass. Rails can move within Pacific cordgrass stands, and spend most of their time under cover of the cordgrass foliar canopy, usually selecting prey items such as invertebrates inhabiting the cordgrass stands and their edges. In contrast to the clapper rail of southern California tidal marshes, San Francisco Bay clapper rails generally do not construct "floating nests" in Pacific cordgrass; instead, they tend to build nests in gumplants or pickleweed in the higher marsh. However, the clapper rails generally use S. foliosa to construct their nest platforms.

Atlantic Smooth Cordgrass (Spartina alterniflora) and its Hybrids Smooth cordgrass is a closely related sibling to Pacific cordgrass that is native to both the Atlantic and Gulf Coasts of the United States (Gleason and Cronquist 1991). It is unique among the world's cordgrass species in terms of its growth potential and ecological breadth, and it is the parent species of the other most invasive cordgrass species of hybrid origin, English cordgrass (S. anglica; Adam 1990). The San Francisco Estuary population of Atlantic smooth cordgrass was introduced from Maryland in the mid-1970s in an ex- periment in dredge spoils stabilization for one of the first tidal marsh restoration projects on the West Coast. Atlantic smooth cordgrass is a coarse perennial grass that, like its Pacific relative, spreads both by seed dispersal and by creeping rhizomes that form extensive clonal colonies. In parts of the San Francisco Estuary, the rate of lateral spread by rhizomes averages between 3.3 and 6.6 feet per year, in contrast with native Pacific cordgrass, which spreads only 0.6 to 2.4 feet per year in the same marshes (Josselyn et al. 1993). Similar rates of lateral spread of this species and its hybrids have been recorded more recently in Cog-

San Francisco Estuary Invasive Spartina Project 10. 2013 Aquatic Pesticide Application Plan swell Marsh on the Hayward Shoreline (K. Zaremba, M. Taylor, pers. comm.), and have been witnessed by ISP biologists at restoration sites around the Estuary including Pond B3 (Middle Bair Island) and Cooley Landing. The size range of Atlantic smooth cordgrass is wide and highly variable, depending on its local genetics and environment. In nutrient-rich, well-drained marsh sediment, such as along tidal creek banks and on newly colonized tidal flats, extensive dense stands can exceed eight feet in height. On poorly drained marsh flats, its vegetation is typically sparse and short, but its dense root and rhizome network maintains pure stands and effectively binds marsh sediments. The "tall form" and "short form" of this species were so strikingly different that they were long assumed to be distinct varieties, rather than variations based on local environmental conditions. Modern research indicates that factors related to marsh drainage, such as waterlogged soil chemistry (especially accumulation of toxic soil sulfides), excessive salinity, and nutrient deficiency interact to cause the dramatic differences in growth-forms of Atlantic smooth cordgrass (Bradley and Dunn 1989, Mendels- sohn and Seneca 1980, Valiela et al. 1978, Smart and Barko 1978). Genetic variations in height forms of Atlantic smooth cordgrass also has been defined in San Francisco Bay (Daehler et al. 1999). However, much of the variation in phenotype displayed around San Francisco Bay is also a result of multiple generations of backcrossing of the hybrids with S. foliosa. In the salt marshes of the Atlantic coastal plain, Atlantic smooth cordgrass is dominant over most of the intertidal zone. Depending on local tidal range, it can grow to and below mean low water (McKee and Patrick 1988), and it can occupy, and even dominate, the marsh plain and the low marsh. Vast, homogeneous stands of Atlantic smooth cordgrass are the characteristic signature of the Atlantic region's tidal marshes (Dame et al. 2000, Adam 1990, Chapman 1964, 1977). In contrast with Pacific cordgrass, Atlantic smooth cordgrass freely establishes in relatively exposed shorelines with significant wave action, including estuarine sand beaches. It is planted in its native range to stabilize shorelines and to trap and accumulate sediments, and the high density of its tall stems is highly effective at reducing estuarine wave energy (Gleason et al. 1979, Knutson and Woodhouse 1988, Knutson et al. 1990). Atlantic smooth cordgrass is also highly resilient in regards to other environmental tolerances. It can survive in salinity over 45 parts per thousand (well above ocean salinity of 32 ppt), and grow luxuriantly in dilute brackish water. If buried, it can regenerate from up to one foot of burial by deposited sediment. Atlantic smooth cordgrass, like other wetland species, can supply oxygen to its roots in anoxic, waterlogged mud, by using porous air- filled chambers (aerenchyma) linking its foliage to roots and rhizomes. Atlantic smooth cordgrass can also tolerate the severe waterlogging and hypersalinity that develops in poorly drained depressions in the salt marsh, including salt marsh pans. Salt marsh pans were frequent, well-developed features of historic San Francisco Estuary marshes, and important habitat for migratory waterbirds (Goals Project 1999). Along the Hayward shoreline of San Francisco Bay, Atlantic smooth cordgrass has colonized many pre- existing pans, converting them to solid cordgrass marsh. In the San Francisco Estuary, Atlantic smooth cordgrass has displayed many of the ecological traits typical of its role in its native salt marsh habitat, and some highly novel phenomena as well. Most colonies in the San Francisco Estuary are young, often forming

San Francisco Estuary Invasive Spartina Project 11. 2013 Aquatic Pesticide Application Plan nearly circular, discrete, expanding colonies, which merge into irregular patterns, resembling mold colonies in a petri dish. After numerous clones of the invader develop, they eventually coalesce into single-species meadows that exclude native marsh vegetation and alter the ecology for the animals that depend on these systems for habitat. The edges of the colonies are tall and robust, while the centers often exhibit early symptoms of die- back or "short form" growth habits. The "donut" shape of colonies is one of the species' signatures for identification in aerial photographs of San Francisco Bay. This trait is not typical of mature Atlantic salt marshes. In the mild Pacific winters, Atlantic smooth cordgrass shoots tend to retain green leaves and persistent dead leaves through much of the winter. This is an important contrast with native Pacific cordgrass: combined with the invader's much greater stem size and shoot density, year-round dense foliage gives Atlantic smooth cordgrass exceptionally high potential to accumulate and trap estuarine sediment during winter storms or floods. The San Francisco population of Atlantic smooth cordgrass has generated some unusual growth forms with strikingly atypical appearance. The dwarf form develops a profusion of short lateral shoots instead of a tall main stem, forming pure stands with complete ground cover of dense, low turf-like ankle-high vegetation on the marsh plain. The growth rate of the dwarf form is, however, vigorous. The dwarf form is genetically induced; it occurs in the same local environments that support luxuriant, tall stands of At- lantic smooth cordgrass, often contiguous with the dwarf patches. It has established at multiple locations in San Francisco Bay (Daehler et al. 1999). A comparable dwarf form of its hybrid daughter species, English cordgrass, independently evolved in Britain and New Zealand (Bascand 1970). Hybridization of Atlantic smooth cordgrass with native Pacific cordgrass. Perhaps the most novel and significant phenomenon of the San Francisco population of Atlantic smooth cordgrass is the rapid evolution of an aggressively expanding hybrid swarm formed by cross pollination with the native Pacific cordgrass (Daehler and Strong 1997). The hybrid swarm includes first-generation crosses between Atlantic smooth cordgrass and Pacific cordgrass with both species acting as pollen-parents and seed parents. Be- cause the two species' pollination periods don’t normally overlap, first-generation crosses are infrequent. Hybrids, however, have a wide range of flowering times, and act as an effective reproductive bridge between the species. The hybrids produce pollen in much greater abundance (up to 21 times greater) and with higher fertility than the native Pacific cordgrass. Superior hybrid pollen production and fertility so overwhelm populations of Pacific cordgrass ("pollen swamping") that native stands of cordgrass produce mostly hybrid back-cross seeds in the presence of flowering hybrid colonies (Ayres et al. 1999, Antilla et al. 2000). This process alone, called hybrid assimilation, can result in the extinction of the invaded species (Levin et al. 1996, Rhymer and Simberloff 1996). Genetic analysis has revealed that numerous large populations that were presumed to be Atlantic smooth cordgrass in the Estuary were predominantly hybrids and back-crosses (introgressants). The ecologically invasive traits of Atlantic smooth cordgrass appear to be prevalent in the hybrid swarm. "Pure" Atlantic smooth cordgrass is now a minority in most of the rapidly evolving hybrid swarms, and trends suggest that hybrids would eventually replace both parent species, as the hybrid-origin species English cordgrass did in Britain (see English Cordgrass, below). This recently discovered threat of genetic extinction to a native cordgrass from an alien cordgrass invasion is unique to the San Francisco

San Francisco Estuary Invasive Spartina Project 12. 2013 Aquatic Pesticide Application Plan Estuary. No native cordgrasses existed where Atlantic smooth cordgrass invaded Wash- ington and Oregon estuaries, and the cordgrasses native to Europe are genetically isolated from their hybrids. Hybrid S. alterniflora was well established and widely distributed in the Central and South Bay at the start of ISP’s control program, but it was detected early and controlled by ISP in the North Bay (and have not yet been detected in Suisun, despite intensive surveys). The northern limit of its distribution in 2005 on the east bay was at Giant Marsh (Point Pinole), and on the west bay was a clone north of Miller Creek mouth (north of China Camp). Over the years there have been infestations detected on the Petaluma Riv- er, Sonoma Creek, Mare Island, and a single clone at Sonoma Baylands. Pioneering infestations of hybrid S. alterniflora have also been managed at Drakes Estero, Limantour Estero and Bolinas Lagoon on the Point Reyes peninsula. The abundance of Atlantic smooth cordgrass and hybrids remains greatest in San Leandro Bay, Robert’s Landing, and outer Bair Island, but has been reduced by 94% baywide by ISP partners since 2006.

English Cordgrass (Spartina anglica) English cordgrass is an aggressive invader of mudflats and salt marshes in Britain, New Zealand, Australia, and the Pacific Northwest, and thrives in cool temperate climates. It originated in Britain as a fertile hybrid derived from introduced Atlantic smooth cordgrass and common cordgrass (S. maritima), a small, slow-growing creeping cordgrass native to European coasts, now greatly reduced in abundance. Within a century after its origin, English cordgrass became the dominant salt marsh grass in Britain (Lee and Partridge 1983, Gray et al. 1990). It is shorter than Atlantic smooth cordgrass and more grayish in appearance, but partly shares other traits of its parent, such as vigorously spreading rhizomes, ability to transform mudflats into vast stands of low marsh vegetation, and ability to dominate and displace associated plant species. It was introduced to the San Francisco Estuary at Creekside Park along Corte Madera Creek in Marin County, along with Chilean cordgrass, in 1976. Unlike Atlantic smooth cordgrass and Chilean cordgrass, this species failed to disperse from its point of introduction to expand the infestation beyond Creekside Park. It may be at or near its southern climatic limit on the Pacific Coast in the San Francisco Estuary.

Chilean Cordgrass (Spartina densiflora) Chilean cordgrass (also called dense-flowered cordgrass) is a distinctive cordgrass species native to South America. It has a bunchgrass growth habit, forming tight clumps or tussocks with short creeping rhizomes, and narrow, firm, in-rolled leaves (Spicher 1984), resembling European beachgrass (Ammophila arenaria). It is generally restricted to the middle marsh plain and high marsh zones where pickleweed, saltgrass, jaumea, and Grindelia stricta (gumplant) otherwise prevail. It does not spread into the low marsh where Pacific cordgrass and mudflats naturally dominate the Estuary (Kittleson and Boyd 1997). Chilean cordgrass lacks well-developed tissues specialized for transporting oxygen from foliage to roots (Spicher 1984), a feature common to cordgrasses adapted to low marsh environments that can thereby withstand greater regular tidal inundation. Chilean cordgrass, along with other South American coastal species, was introduced to Humboldt Bay, California by ship ballast containing seeds from South American ports

San Francisco Estuary Invasive Spartina Project 13. 2013 Aquatic Pesticide Application Plan that traded lumber (Spicher 1984). For most of the 20th Century, Chilean cordgrass was erroneously treated as an "ecotype," or minor geographic variation, of the native Pacific cordgrass, despite the obvious lack of diagnostic traits matching this species. In the late 1970s, the presumed native "Humboldt Bay form" of Pacific cordgrass was deliberately transplanted to a salt marsh restoration at Creekside Park along Corte Madera Creek in Marin County. Within the salt marshes fringing Corte Madera Creek, it became a locally dominant component of the middle and high salt marsh vegetation, displacing even robust pickleweed. A second population of Chilean cordgrass spontaneously established across the Bay from Creekside Park in the ancient marsh plain at Point Pinole (Whittell Marsh), Contra Costa County. The Point Pinole population was discovered in the mid-1990s, and is very close to eradication. A population was also introduced (anonymously) to Sanchez Marsh along the West Bay in Burlingame, and more recently to a brackish drainage area in Redwood City. Both of these infestations did not spread to other marshes and are down to manual removal of a handful of plants each year until the seedbank is exhausted (which takes at least three years). A single, large, individual clump of Chilean cordgrass established in a very young restored tidal marsh (breached 1995) at the former Salt Pond 2A, Napa Marsh. That pioneering plant was quickly eradicated. In 2007-2008, ISP biologists found a couple of instances of a new form of S. densiflora. It quickly became evident that this species had also successfully hybridized with native S. foliosa to produce a novel cordgrass. Within a year, ISP biologists found hybrid S. densiflora at many of the marshes in Marin where the two parents grew in close proximity. This “species” appears to spread vegetatively rather than by seed, but little is known about it and it has been actively treated since its discovery. If not for the vigilance of ISP, conducting annual monitoring at all the marshes of Marin and the rest of the Estuary, this invader would likely have gone undetected and uncontrolled, providing the opportunity for significant impacts to the ecosystem.

Salt-Meadow Cordgrass (Spartina patens) Salt-meadow cordgrass is another rhizome-forming creeping cordgrass of Atlantic salt marshes, but unlike Atlantic smooth cordgrass, it has fine stems with narrow, soft, in- rolled leaves, and is intolerant of waterlogged mud. It is naturally confined to the well- drained high salt marsh and relatively moist sandy depressions at or above tidal influence. Two distinctive geographic and ecological types have been recognized, and in the past have been treated as distinct taxonomic varieties. In peaty high salt marshes of the northeastern Atlantic coast, a relatively low form with lax, slender stems forming dense matted turfs with "cowlicks" was once treated as S. patens var. monogyna. These dense salt marsh turfs are often nearly pure stands of salt-meadow cordgrass crowding out most other species that occupy gaps in the cover caused by winter ice or drifted wracks. In sandy marshes associated with large barrier beaches and wash-over fans from Cape Cod through the Atlantic coastal plain, a coarser, erect type, formerly recognized as S. patens var. juncea, is prevalent. Intermediate forms are common. Between the 1959 publication of A California Flora (Munz and Keck 1959) and its 1970 supplement, salt-meadow cordgrass was reported in Southampton Bay, Benicia, Solano County. The time and mode of introduction is unclear. At the initiation of treatment by

San Francisco Estuary Invasive Spartina Project 14. 2013 Aquatic Pesticide Application Plan ISP and State Parks, salt-meadow cordgrass at Southampton occupied large, discrete patches in pure and exceptionally thick stands compared with its native marshes. One large stand spread into an area that supports a population of an endangered annual plant, soft bird's-beak (Cordylanthus mollis ssp. mollis now Chloropyron molle molle). This overlap with an endangered plant has greatly complicated the final eradication efforts for S. patens. The Southampton Bay cordgrass population appears to match the type description of "variety monogyna," the fine-stem type of northeastern Atlantic marshes.

5. CONTROL TOLERANCES As previously mentioned, the non-native Spartina invasion of the San Francisco Estuary is especially threatening to native marsh systems because of hybridization of Atlantic smooth cordgrass and the native Pacific cordgrass. This ability to hybridize, and the documented expansion rates of the population of hybrid forms throughout the Estuary, defines the need for a zero tolerance threshold on invasive Spartina in San Francisco Bay. The Invasive Spartina Project is a regionally-coordinated eradication effort that will ul- timately be successful only if all infestations are effectively controlled and monitored to eradication. A single small, expanding clone of hybrid Spartina within an otherwise native S. foliosa matrix has the capability of ‘swamping’ S. foliosa flowers with hybrid pollen, effectively converting the native stand into a hybrid-producing population. Within a couple of growing seasons, the majority of new seedlings establishing in the area will be of hybrid origin, resulting in the eventual extirpation of the native S. foliosa from the stand. Re- peated throughout the Estuary on various scales, this progression threatens the population stability of native Pacific cordgrass stands. Therefore, where hybrid forms of Spartina are identified, efforts must be directed at removing all of the plants in the area. There is no acceptable level of hybrid presence in an otherwise native marsh, as the inevitable result of even a small amount of hybrid presence will be the relatively rapid conversion to a non-native stand capable of infesting adjacent marshlands. Spartina densiflora has also been shown to have rapid expansion rates in the Estuary. In addition, this plant invaded Humboldt Bay, CA to such a degree that approximately 80% of the mudflats and marshlands have been dominated by it resulting in an infestation of 2000 acres by 2010. Fortunately, in the San Francisco Estuary we have only isolated populations of S. densiflora because of the efforts of the ISP and its partners, so as in the case with the more prevalent hybrid S. alterniflora, there is a need for zero tolerance and an excellent opportunity for the eradication of this species before it becomes widespread. For the other two invasive Spartina species in the Estuary, S. anglica and S. patens, each inhabits only a single infestation site and in a relatively small remnant of the initial infestation in 2005. However, S. anglica is known to be the most invasive species of Spartina in the world, itself a fertile hybrid that has dominated introduction sites on several conti- nents. With these types of threats, the fertile environment of the Estuary, and the current size of the infestations, a zero tolerance policy is appropriate. In addition, with a couple effective seasons of treatment, each of these species could be eradicated from the Estu- ary. The current Integrated Pest Management (IPM) strategy developed for S. patens re-

San Francisco Estuary Invasive Spartina Project 15. 2013 Aquatic Pesticide Application Plan lies completely on tarping the last stands and manual removal of the remaining individuals, so herbicide will no longer be required for this species.

6. DESCRIPTION OF HERBICIDE Herbicides have proven highly effective in controlling populations of non-native cordgrasses (Spartina spp.). The aquatic formulation of imazapyr (Habitat or Polaris™) was registered for use in the State of California on August 30, 2005. San Francisco Bay land managers that were engaged in their own independent Spartina control efforts prior to the inception of the ISP Control Program used aquatic glyphosate-based herbicides (Aquamaster, Rodeo). These the only two herbicide formulations registered by US EPA for use in sensitive estuarine systems; all ISP partners have since switched to imazapyr. Since imazapyr is such a slow-acting herbicide, it is often difficult to assess the efficacy of an application until the following spring. There are a number of qualities that make imazapyr the ISP’s preferred choice over the previous alternative, glyphosate. Glyphosate tends to strongly adsorb to sediment and salt particles accumulated on the Spartina, rendering the herbicide inactive. It is common for the tides to deposit abundant sediment from the turbid San Francisco Bay onto the invasive Spartina in the adjacent salt marshes. Glyphosate also requires significantly longer dry times to fully penetrate the cuticle of the plant and begin translocation. Imazapyr does not have these issues that can reduce its efficacy. In addition, imazapyr is applied at lower concentrations and the applicator does not need to “spray to wet” as with glyphosate. This greatly reduces the amount of herbicide that will enter the environment as a result of the Spartina Control Program’s efforts, and allows for low volume applications such as aerial helicopter work. This herbicide delivery system was essential to the ISP’s initial efforts because of the vast acreage of Spartina infesting hazardous and difficult to access marshes at that time. Aquatic herbicide formulations such as those used by ISP must be combined with a suitable surfactant to facilitate uptake by the plant and translocation of the herbicide down into the rhizomes. A harmless, inert marker dye or colorant is often added to the tank mix to assist the applicator at achieving full coverage while not over-applying to any areas. The following discussion addresses imazapyr, glyphosate, breakdown products, and typical surfactants and colorants. Detailed descriptions of the chemical properties, degradation rates, environmental fate, and toxicity of imazapyr, glyphosate, and all of the aquatic surfactants evaluated for the Spartina Control Program are provided in Appendix 1. ® Imazapyr. Habitat or Polaris™ are solutions of 28.7% isopropylamine salt of imazapyr in water, equivalent to 22.6% imazapyr acid equivalents (a.e.) or 2 lbs. acid per gallon, and contain a small amount of an acidifier. Because Habitat® is purportedly a similar formulation as Arsenal® and this product contains acetic acid, the acidifier in Habitat® is likely also acetic acid (Leson & Associates 2005.) No information has been found in the published literature on manufacturing impurities associated with imazapyr. Because virtually no chemical synthesis yields a totally pure product, technical grade imazapyr most likely contains some impurities. However, to some extent, concern for impurities in technical grade imazapyr is reduced by the fact that most existing toxicity studies on imazapyr were conducted with the technical grade product and encompass the toxic potential

San Francisco Estuary Invasive Spartina Project 16. 2013 Aquatic Pesticide Application Plan of the impurities (SERA 2004). A generic version of this aquatic imazapyr formulation is now available from Nufarm under the product name Polaris™. Imazapyr inhibits an enzyme (acetolactate synthase [ALS]) in the biosynthesis of the three branched-chain aliphatic amino acids valine, leucine, and isoleucine. Because animals do not synthesize branched-chained aliphatic amino acids but obtain them from eat- ing plants, the engineered mechanism for plant toxicity, i.e. the interruption of protein synthesis due to a deficiency of the amino acids valine, leucine, and isoleucine, is not generally relevant to birds, mammals, fish or invertebrates. Any toxicity to these receptors occurs through different mechanisms (Entrix 2003). Imazapyr is relatively slow acting taking several weeks for the plants to show lethal effects. However, plants cease growth within 24 hours of a successful application (J. Smith, pers. comm. 2006). On Spartina, it takes 2-4 weeks after treatment to see visible effects such as yellowing of the leaves, and complete plant death can take several months. In the San Francisco Estuary, with the relatively late season applications on invasive Spartina (mainly because of endangered species issues that effect access to the marshes), the treated plants may not re- veal much of a response before natural senescence, but will simply not emerge in the spring of the following year if fully impacted by the treatment. Glyphosate. Aquamaster and Rodeo are aqueous solutions containing 53.8% glyphosate in its isopropylamine salt form or 4 lbs. acid per gallon, and contain no inert ingredients other than water. The primary decomposition product of glyphosate is aminomethylphosphonic acid (AMPA), and the commercial product contains an impurity, 2,4-nitrosoglyphosate (NNG). The potential effects of AMPA and NNG are encompassed by the available toxicity data on glyphosate and glyphosate formulations (SERA 1996). Glyphosate inhibits an enzyme needed to synthesize an intermediate product in the biosynthesis of the aromatic amino acids, essential for protein synthesis and to produce many secondary plant products such as growth promoters, growth inhibitors, phenolics, and lignin. Animals do not synthesize these aromatic amino acids and glyphosate therefore has low toxicity to these receptors (Schuette 1998). In general, glyphosate herbicides are somewhat faster acting than imazapyr herbicides. On Spartina, complete brown-down occurs within 7 to 21 days (K. Patten, pers. comm. 2004). Surfactants. For most foliar applications of aquatic herbicide formulations, adjuvants must be added to spray solutions to improve the performance and minimize the variabil- ity of herbicide efficacy. Surfactants are designed to improve the spreading, dispers- ing/emulsifying, sticking, absorbing, and/or pest-penetrating properties of the spray mixture (Tu et al. 2001). The pure herbicide formulation mixed with water will stand as a droplet on the waxy leaf surface and the small area of contact therefore provides little potential for uptake of the active ingredient into the foliage. Water droplets containing a surfactant will spread in a thin layer over a waxy leaf surface and improve herbicide uptake by maximizing herbicide distribution and forcing the fluid into the plant. As men- ® tioned above, both Habitat and Polaris™, as well as the glyphosate herbicides Aq- uamaster® and Rodeo®, require the addition of surfactants for post-emergent applications such as the control of invasive Spartina. Imazapyr. The specimen label recommends a variety of different spray adjuvants for use on post-emergent vegetation. For non-ionic surfactants the label recommends a rate of 0.25% v/v or higher, preferably of a surfactant with a hydrophilic to lipophilic ra-

San Francisco Estuary Invasive Spartina Project 17. 2013 Aquatic Pesticide Application Plan tio between 12 and 17 and with at least 70% surfactant in the formulated product (BASF 2003). Alternately, the label recommends the use of methylated seed oils or vegetable oil concentrates at the rate of 1.5 to 2 pints per acre. For spray volumes greater than 30 gallons per acre, the surfactant should be mixed at a rate of 1%. The label further indicates that these oils may aid in imazapyr deposition and uptake by the plants under mois- ture or temperature stress. Silicone-based surfactants, which reduce the surface tension of the spray droplet to an even greater degree, allowing greater spreading on the leaf surface as compared to conventional non-ionic surfactants, are also recommended. However, the manufacturer points out that some silicone-based surfactants may dry too quickly, especially in the heat of the summer, limiting herbicide uptake (BASF 2004). One study from Washington State concluded that the esterified seed oil surfactant tested, Competitor®, performed better than the other surfactants tested, i.e. Agri-Dex®, a crop oil-based surfactant, and R-11®, a non-ionic surfactant. This finding is also supported by Patten (2002) in which the author recommended using a methylated seed oil surfactant for aerial applications and for unfavorable conditions such as less than six hours of drying time before tidal inundation, or also on moist leaves. The experience of the ISP from since 2005 has shown that a lecithin (soybean) product, Liberate®, has also been highly effective with imazapyr. In addition, this product acts as a drift retardant which may help in helicopter treatments as well as other high-pressure hose applications, to ensure full coverage on the target Spartina and minimize drift onto non-target marsh plants. Glyphosate. The Aquamaster® and Rodeo® specimen labels recommend the use of a non-ionic surfactant containing at least 50% active ingredient at a rate of 2 or more quarts per 100 gallons of tank mix (0.5% v/v). With glyphosate it is also important to balance the pH of the tank water to ensure effectiveness, and some adjuvants are designed with this purpose in mind, namely LI-700®. Not all surfactants provide the same effectiveness and surfactant costs vary widely. In general, non-ionic surfactants and crop oil concentrates are the least expensive of the surfactant classes, followed by esterified seed oils and organo-silicates (Miller & Westra 2004). The ISP identified a number of potential surfactants for use with imazapyr and glyphosate at the beginning of the Control Program in 2005. They include the non-ionic surfactants Liberate® and LI-700®, the crop-oil concentrate Agri-Dex®, the esterified seed oil Competitor®, and the organo-silicones Dyne-Amic® and Kinetic®. Based on the efficacy experienced by the ISP since 2005, and their superior relative toxicities to animals, the ISP expects to continue to exclusively use Competitor® and Liberate® in the Spartina Control Program. LI-700® may be used to balance the pH of the tank if/when glyphosate is employed. As of the 2012 Treatment Season, there are no glyphosate applications planned on invasive Spartina in the foreseeable future. Cygnet Plus® was evaluated and originally included in the ISP list of surfactant choices for 2005, but it was shown to be ineffective with imazapyr and has been removed from the list available to the ISP partners. Competitor (Wilbur-Ellis Company) is a modified vegetable oil containing a non-ionic emulsifier system. The ingredients include ethyl oleate, sorbitan alkyl polyethoxylate ester, and dialkyl polyoxy-ethylene glycol. Toxicity studies classified this surfactant as a toxicity category of 3-4 (Caution signal word).

San Francisco Estuary Invasive Spartina Project 18. 2013 Aquatic Pesticide Application Plan Liberate (Loveland Industries, Inc.) is a non-ionic, low foam penetrating adjuvant. Its active ingredients are lecithin (phosphatidylcholine, which is a naturally occurring lipid derived from soybeans that biodegrades readily), methyl esters of fatty acids, and alcohol ethoxylate. In a 1% solution, the pH is an almost neutral 6.8. Toxicity studies classified this surfactant as a toxicity category of 3-4 (Caution signal word). It improves deposition and retards drift by producing a more uniform spray pattern. Dyne-Amic (Helena Chemical Company) is a proprietary blend of non-ionic organosilicone surfactants and a methylated vegetable oil. Toxicity studies classified this surfactant as a toxicity category of 3-4 (Caution signal word). Kinetic (Helena Chemical Company) is a non-ionic wetting agent that allows for the rapid spreading and absorption of herbicide sprays into the target vegetation, and is especially effective with water-based herbicide formulations. Its active ingredients include organosilicone and polyoxypropylene-polyoxyethylene copolymer. Toxicity studies classified this surfactant as a toxicity category of 3-4 (Caution signal word). Agri-Dex (Helena Chemical Company) is a non-ionic surfactant consisting of a paraffin base petroleum oil, polyol fatty acid esters, and polyethoxylated derivatives of the fatty acid esters. Toxicity studies classified this surfactant as a toxicity category of 3-4 (Cau- tion signal word).3. Biodegradation of this adjuvant is presumed to be rapid. LI-700 (Loveland Industries, Inc.) contains phosphatidylcholine (lecithin), which is a naturally occurring lipid derived from soybeans that biodegrades readily. It also contains methylacetic acid and alkyl polyoxyethylene ether. Toxicity studies classified this surfactant as a toxicity category of 1 (Danger signal word) because of corrosive properties to the skin and/or eyes. Biodegradation of this adjuvant is presumed to be rapid because of the natural lecithin ingredients. Colorant. There are several colorants suitable for use in the marsh environment, all of which are similar in composition and performance. Blazon Spray Pattern Indicator (Mil- liken Chemical), a typical colorant, is a water-soluble polymeric product. As with most colorant products, the active ingredients in Blazon are proprietary; the Material Safety Data Sheet (MSDS) indicates that it is non-hazardous and non-toxic. The product information sheet reports that the product is non-staining to the skin or clothing. A literature survey on the toxicity of color indicators done for the U.S. Department of Agriculture reports “most commercial indicators are blue … and most often a form of Acid Blue 9…” (McClintock 1997 and Zullig 1997 cited in SERA 1997b). Acid Blue 9 is a disodium salt classed chemically as a triphenylmethane color (SERA 1997b). The Cosmetic, Toiletry, and Fragrance Association (CTFA) name for certified batches of Acid Blue 9 is FD&C blue No. 1. Herbicide application. Impacts to water quality from herbicide application depend on environmental fate, degradation rates of active agents and decomposition products of the herbicides. The primary route by which herbicide solution may contact water during invasive Spartina treatment is by overspray directly onto the adjacent water surface or onto areas that will be covered by water on the next tide. Herbicide may also be washed off plants by subsequent tidal inundation, or potentially by precipitation (although the

3 Toxicity categories are determined by the U.S. EPA for human health affects. See http://www.epa.gov/oppfead1/labeling/lrm/chap-08.htm for more information on pesticide label requirements.

San Francisco Estuary Invasive Spartina Project 19. 2013 Aquatic Pesticide Application Plan Spartina treatment occurs during the dry season for San Francisco Bay where there is little measurable rainfall). Imazapyr will be applied as a spray to Spartina foliage for control of this invasive plant. Spray mixtures may be dispersed from manually transported tanks (backpack sprayers) or spray equipment mounted on trucks, amphibious tracked vehicles, airboats, or helicopters. Application rates will be consistent with the product labels (Table 4). Applications from backpack sprayers, conventional spray truck, or airboat entails workers walking along the high marsh ecotone or through the marsh, applying herbicide directly to target plants with limited overspray to surrounding plants or water surfaces. Spot application from amphibious tracked vehicles or boats entails vehicles moving through the marsh or adjacent waterway applying herbicide with hand-held equipment to target vegetation with limited overspray. Aerial application is conducted by helicopter (the imazapyr label does not allow for fixed-wing aircraft to be used) from a boom sprayer (a horizontal pipe with spray nozzles along its length, mounted to the bottom of the helicopter). Broad- cast aerial application involving a boom sprayer results in a wider dispersion of herbicide, with greater potential for overspray onto non-target areas or the water surface. However, for aerial treatment on smaller patches of invasive Spartina, the boom may be shortened up and the apparatus turned on only when over the target vegetation. Aerial application will be used primarily at large areas of dense cordgrass infestations, particularly in locations where little native cordgrass and other non-target plants are nearby, and where the work is more than a quarter mile from sensitive receptors.

Environmental Fate of Herbicides Herbicide residues may be indirectly discharged to surface waters by tidal action or rainfall that rinses the herbicide solution from the plants. Rainfall is unlikely to occur during the planned application season (summer to early autumn in the San Francisco Bay region), and herbicide applications would be postponed if significant rainfall was predicted, but tidal inundation is inevitable on a regular cycle. Applications to invasive Spartina along slough or creek banks or at the Bayfront on fringe marshes and mudflats will result in a small percentage of the herbicide entering the water column as a residue. Food-web scale exposures become significant only with chemicals that have a tendency to bioaccumulate or biomagnify. The adverse effects associated with bioaccumulative chemicals relate to their propensity to transfer through the food web and accumulate preferentially in adipose or organ tissue. Basic routes for organism exposure to bioaccumulative compounds are the transport of dissolved contaminants in water across biological membranes, and ingestion of contaminated food or sediment particles with subsequent transport across the gut. For upper-trophic-level species, ingestion of contaminated prey is the predominant route of exposure, especially for hydrophobic chemicals.

San Francisco Estuary Invasive Spartina Project 20. 2013 Aquatic Pesticide Application Plan Table 4a: Imazapyr herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Estu- ary (Leson & Associates 2005).

Application Method Spray Volume Formulation Active Ingredient1 Surfactant2 Colorant 0.25% v/v NIS with ≥70% a.i.; High volume 0.52-0.75% solution 100 gal/acre 1-1.5 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 3 qt/100 gal handheld sprayer 4-6 pints/100 gal SBS according to label 0.25% v/v NIS with ≥70% a.i.; Low-volume directed 0.75-1.5% solution 20 gal/acre 0.3-0.6 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 3 qt/100 gal sprayer 1.2-2.4 pints/20 gal SBS according to label 0.25% v/v NIS with ≥70% a.i.; Broadcast sprayer/ 2.5-7.5% solution 10-30 gal/acre 0.5-1.5 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 0.5-1.5 qt/acre Aerial application 6 pints/10-30 gal SBS according to label

1 Active ingredient in Habitat® and Polaris is imazapyr isopropylamine salt; values expressed as imazapyr acid equivalent 2 NIS = non-ionic surfactant; MSO = methylated seed oil; ESO = esterified seed oil; VOC = vegetable oil concentrate, SBS = silicone-based surfactant, %v/v = percentage based on volume by volume

Table 4b: Glyphosate herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Es- tuary (Leson & Associates 2005).

Application Method Spray Volume Formulation Active Ingredient1 Surfactant2* Colorant

High volume 1-2% solution 100 gal/acre 4-8 lb a.e./acre ≥0.5% v/v NIS with ≥50% a.i. 3 qt/100 gal handheld sprayer 1-2 gal/100 gal

Low-volume di- 1-8% solution 25-200 gal/acre 1.35-10.8 lbs a.e./acre ≥0.5% v/v NIS with ≥50% a.i. 3 qt/100 gal rected sprayer 1-8 gal/100 gal

Broadcast sprayer/ 7-40 gal/acre/ 4.5-7.5 pints/acre 2.25-3.75 lb a.e./acre ≥0.5% v/v NIS with ≥50% a.i. 0.5-1.5 qt/acre Aerial application 7-20 gal/acre

1 The active ingredient in Rodeo® and Aquamaster® is glyphosate isopropylamine salt; values are expressed as glyphosate acid equivalent

2 NIS = non-ionic surfactant, %v/v = percentage based on volume by volume

San Francisco Estuary Invasive Spartina Project 21. 2010 Aquatic Pesticide Application Plan U.S. EPA’s Hazardous Waste Identification Rule (USEPA 1999) identifies compounds that are recognized as having a low, medium or high potential for bioaccumulation. For bioaccumulation in aquatic systems, rankings were determined using bioaccumulation factors in fish, which are indicated in laboratory tests as having low octanol-water parti- tioning coefficient (or log Kow)) values for organic compounds. Bioaccumulation potential is defined as follows: Bioaccumulation Bioaccumulation Factor potential (BAF) log Kow

High BAF >= 10,000 log Kow >= 4.0

Medium 10,000 > BAF >= 100 4.0 > log Kow >= 2.0

Low BAF < 100 log Kow < 2.0 Imazapyr. Under typical environmental conditions of pH 5-9, imazapyr is ionized and is therefore highly soluble in water. Because of its high solubility, imazapyr has an inher- ently low sorption potential with a low soil organic carbon sorption coefficient (Koc) of 8.81 (log Koc), suggesting very high mobility in soil and little adsorption to suspended solids and sediment. Its octanol/water partition coefficient (Kow) has been reported at 0.22 (log Kow), reflecting its high solubility in water and low solubility in lipids, and hence low propensity to bioconcentrate. A low bioaccumulation factor (BAF) of 3 was calculated for imazapyr, which suggests a low potential for bioconcentration in aquatic organisms (Leson & Associates 2005). U.S. EPA considers compounds with a BAF less than 100 to have low bioaccumulation potential as shown in the table above. Imazapyr is relatively mobile in soils because it only weakly adsorbs to soils and sediments. Adsorption increases with decreasing pH. Above a pH of 5, imazapyr is ionized and does not adsorb to soil. Aerobic degradation in soils occurs primarily by very slow microbial metabolism with quinoline as the main metabolite (Entrix 2003). Conditions in sediments differ substantially from those in soils, both in terms of the regular exchange of waters within the sediment pore water, and in the degree of oxygenation in sediments that affect microbial metabolism. Because the pH of sediment surfaces and sediment pore water in intertidal mudflats is above neutral (pH >7), imazapyr will be entirely in its ionized form. Thus, adsorption to sediments is expected to be minimal (Entrix 2003). Microbial metabolism in sediments has been determined to be insignificant. The degradation of imazapyr when applied directly to water largely mimics the pathway by which the herbicide would be mobilized at high tide after application to Spartina during low tide. Residual imazapyr on the plants that have not completely dried or did not penetrate the leaf cuticle will be inundated by the incoming tide and presumably solubil- ized. Aquatic degradation studies under laboratory conditions demonstrated rapid initial photolysis of imazapyr with reported half-lives ranging from 3 to 5 days (SERA 2004). The two primary photodegradation products were rapidly degraded with half-lives less than or equal to 3 days and eventual mineralization to carbon dioxide (Entrix 2003). Degradation rates in turbid and sediment-laden waters, common to estuarine environments, are expected to be lower than those determined under laboratory conditions. In controlled field dissipation studies in two freshwater pond systems with application of 1.5 lb imazapyr a.e./acre, imazapyr rapidly dissipated from the water with first-order half- lives of 1.9 days and 12.8 days. No detectable residues of imazapyr were found in the wa-

San Francisco Estuary Invasive Spartina Project 22. 2013 Aquatic Pesticide Application Plan ter and sediment after 14 and 59 days, respectively (Entrix 2003). The ISP’s NPDES water quality monitoring at treatment sites over the past several years has found a standard reduction in imazapyr in the adjacent surface water of 92-99% just one-week post- treatment over the amount present in the adjacent surface water immediately after the application. In estuarine systems, dilution of imazapyr by the incoming tide will contribute to its rapid dissipation and removal from the area where it has been applied. Studies in estuaries in Washington State examined the fate of imazapyr applied at a standard rate of 1.5 lb imazapyr a.e./acre directly to sediment. The study design was conservative because imazapyr was applied to bare mudflats with no algal or emergent vegetation intercepting the herbicide. The study measured immediate maximum concentrations of imazapyr in intertidal waters and sediment less than 3 hours after application and short-term concentrations between 24 and 72 hours after application. Sediment samples collected 3 hours after application were retrieved immediately after the first tidal wash over the area. Maximum concentrations in water and sediment were detected at 3.4 mg/L and 5.4 mg/kg, respectively. Measurable concentrations of imazapyr declined exponentially in both water and sediment, approaching the zero-asymptote at 40 and 400 hours with half-lives of <0.5 and 1.6 days, respectively. Water collected 20 and 200 feet outside the spray zone with the first incoming tide was 99% lower than the maximum water concentration at the edge of the spray zone (Leson & Associates 2005). Application of the same amount of herbicide to a stand of 5.5-foot tall Spartina resulted in a 75% reduction in concentrations in sediment through interception by the canopy (Patten 2003). In sum, this research suggests that imazapyr quickly dissipates in estuarine environments. In addition, Patten observed that native vegetation rapidly colonizes the plots treated with imazapyr after the Spartina plants have died, which supports the conclusion of very low persistence of imazapyr in estuarine environments. The ISP has routinely observed this phenomenon of rapid native plant colonization of treated areas at many sites around San Francisco Bay since 2005, usually involving passive revegetation by either annual pickleweed (Salicornia europaea), perennial pickleweed (Sarcocornia pacifica), or Jaumea carnosa. Glyphosate. Under typical environmental conditions of pH 5-9, glyphosate is ionized. Glyphosate and its salts are readily soluble in water with a solubility of about 12,000 mg/L. Its interactions with soil and sediment are primarily ionic, rather than hydrophobic and pH dependent. Laboratory and field studies indicate that glyphosate is strongly and irreversibly adsorbed by soil, sediment, and suspended sediment. Glyphosate is inactivated through soil adsorption. Because glyphosate adheres strongly to particles, it does not readily leach to waters (Sprankle et al., 1977 cited in Albertson, 1998), and potential movement of glyphosate to groundwater is unlikely. Due to its negligible vapor pressure (7.5×10-8 mmHg) and its ionic state in water, glyphosate is not expected to volatilize from water or soil. Glyphosate’s Kow has been reported at 0.00033, indicating its high solubility in water, low solubility in lipids, and thus low potential to bioaccumulate. All reported bioaccumulation factor values for glyphosate in aquatic organisms are well below 100 (Ebasco 1993; Heyden 1991; Wang et al. 1994). The highest bioaccumulation factor of 65.5 was reported for tilapia (a species of fish) in freshwater (Wang et al. 1994). Other studies report much lower bioaccumulation factors in the range of 0.03 to 1.6 for fish (Ebasco 1993). Most studies report rapid elimination and depuration from aquatic

San Francisco Estuary Invasive Spartina Project 23. 2013 Aquatic Pesticide Application Plan organisms after exposure stops (Ebasco 1993). Therefore, bioaccumulation of glyphosate is considered to be low and food-web transfer is not considered to be a significant exposure route. Soil studies have determined glyphosate half-lives ranging from 3 to 130 days. The soil field dissipation half-life averaged 44 to 60 days (Leson & Associates 2005). In the soil environment, glyphosate is resistant to chemical degradation, is stable in sunlight, is relatively non-leachable, and has a low tendency to runoff (except as adsorbed to colloidal matter). It is relatively immobile in most soil environments as a result of its strong adsorption to soil particles. Glyphosate is rapidly and strongly adsorbed to sediment, which appears to be the major sink for glyphosate in aquatic systems. Like in soils, the herbicide is inactivated and biodegraded by microorganisms. Several studies indicate that glyphosate is stable in water at pH ranging from 3 to 6. The photolytic half-life of glyphosate in deionized water exposed outdoors to sunlight was approximately 5 weeks at 100 ppm and 3 weeks at 2000 ppm. Glyphosate shows little propensity toward hydrolytic decomposition. Its hydrolysis half-life is greater than 35 days. It is also stable to photodegradation under visible light but photolyzes when exposed to UV radiation. Glyphosate’s loss from water occurs mainly through sediment adsorption and microbial degradation. The rate of microbial degradation in water is generally slower because there are fewer microorganisms in water than in most soils. Studies conducted in a forest ecosystem found that glyphosate dissipated rapidly from surface water ponds high in suspended sediment, with first order half-lives ranging from 1.5 to 11.2 days. In streams, residues were undetectable within 3 to 14 days. Other studies using water from natural sources determined glyphosate’s half-life ranging from 35 to 63 days (Leson & Associates 2005). For all aquatic systems, sediment appears to be the major sink for glyphosate residue. A review of the literature on glyphosate dissipation applied under estuarine conditions suggests that 24 to 48 hours after applications, glyphosate concentrations in water were reduced by more than 60-fold. Energetic tidal cycles and tidal currents effectively disperse bound (adsorbed) glyphosate and surfactants and dilute them in microbially active suspended sediment. Studies of the fate of glyphosate and surfactants applied in tidal marshes and mudflats have reported that concentrations of both substances dropped below detection levels as soon as two tidal cycles (one day) to seven days (Kroll 1991, Paveglio et al. 1996) after application. The initial tidal submergence of sprayed surfaces disperses a large fraction of applied glyphosate and surfactant. Research conducted for the California Department of Food and Agriculture (Trumbo 2002) studied the environmental fate and aquatic toxicity of Rodeo and R-11 in three locations, including a Sacramento-San Joaquin Delta slough, a riverine area, and a no- outlet pond. This study measured glyphosate, aminomethylphosphonic acid (AMPA; glyphosate’s primary metabolite), nonylphenol ethoxylate, and nonylphenol at treated sites one hour, two days, and eight days after application. The study also tested for toxicity using 96-hour toxicity tests with the fish species fathead minnow (Pimephales promelas). The study found that concentrations of the tested constituents at slough and river sites (with moving water) was below detectible levels for all tests, and that there was no significant mortality of test fishes. The pond site, however, showed detectable residues of glyphosate, nonylphenol ethoxylate, and nonylphenol at one hour and two days after treatment, but all constituents were below detection limits by day eight. The one-hour

San Francisco Estuary Invasive Spartina Project 24. 2013 Aquatic Pesticide Application Plan pond samples experienced 30% mortality of test fishes, which, because of the relatively low concentrations of glyphosate (which is known to be non-toxic at the detected level), was attributed to effects caused by nonylphenol ethoxylate and nonylphenol. The two- and eight-day tests showed no significant mortality to test fishes. Patten (2002) compiled data on the fate of glyphosate in water and sediment following applications in estuarine environments. Data are presented as geometric means for immediate maximum concentration (<3 hrs after application) and short-term concentration (between 24 hrs and 48 hrs after application). For use rates between 8 and 16 kg/ha (7-15 lbs/acre), the immediate maximum geometric mean glyphosate concentrations were 0.174 mg/L (174 µg/L) in water and 2 mg/kg in sediment. The short-term geometric mean glyphosate concentrations were 0.003 mg/L (3 µg/L) in water and 1.9 mg/kg in sediment. These independent lines of research in the fate of imazapyr or glyphosate combined with a surfactant in tidal (and other) habitats suggest that potential impacts to water quality and beneficial uses of waters of the State caused by spraying these herbicide mixtures in intertidal environments are likely be small and temporary. Therefore, controlled applications (i.e., following FIFRA label instructions) of herbicides registered for use in the estuarine environment are not expected to degrade water quality, except to a limited tem- poral and spatial extent. In summary, the use of imazapyr or glyphosate combined with a surfactant to treat infestations of non-native cordgrass results in less than significant impacts on water quality due to the rapid degradation rate and controlled application of herbicides to the target plants. Since application of herbicide would take place during low tide and low wind conditions, the herbicide would likely be absorbed by plants for a minimum of several hours (up to several weeks in high marsh) following application resulting in less than significant quantities of imazapyr, glyphosate or surfactants entering the water.

Potential Biological and Ecological Effects The known properties of the herbicides, potential methods of application, and the ecological characteristics of the Estuary were evaluated to develop a conceptual model and identify likely receptors and exposure pathways. The conceptual model (Figure 2) includes identification of primary and secondary herbicide sources, illustrates the links between sources, release and transport mechanisms, affected media, exposure routes, and potentially exposed ecological receptors. For effects to occur, a receptor and a complete exposure pathway must be present. An exposure pathway is only considered complete when all four of the following elements are present: project-related source of a chemical, a mechanism of release of the chemical from the source to the environment, a mechanism of transport of the chemical to the ecological receptor, and a route by which the receptor is exposed to the toxic nature of the chemical. The exposure routes associated with the complete pathways include direct contact with the herbicide mixture during and immediately after application, ingestion of contaminated surface water and sediments, direct contact with contaminated surface water and sediments, and food-web exposure. Although several complete exposure pathways may exist, not all pathways are comparable in magnitude or significance. The significance of a pathway as a mode of exposure depends on the identity and nature of the chemicals in-

San Francisco Estuary Invasive Spartina Project 25. 2013 Aquatic Pesticide Application Plan Figure 2. Conceptual Model of Possible Exposure of Biological Organisms to Herbicide Mixture Used by the Spartina Control Program

volved and the magnitude of the likely exposure dose. For birds and mammals, ingestion is generally the most significant exposure pathway. Dermal contact is expected to be insignificant and unquantifiable due to the nature of the application site, and frequent movement, ranging habits, and furry or feathery outer skin of most wildlife species.

San Francisco Estuary Invasive Spartina Project 26. 2013 Aquatic Pesticide Application Plan Because ISP applications of herbicides normally occur only once or (at most) twice a year at a given infestation site, and compounds in the herbicide mixture are not expected to persist in significant concentrations, chronic exposure is not likely. In addition, since the aquatic formulation of imazapyr was a newly registered herbicide in California as of 2005, there are few other sources for its introduction to the environment to add to those of the ISP. Therefore, this evaluation focuses on acute toxicity, which would occur when the compounds are present at relatively high concentrations during and immediately following application. Herbicide solutions have the potential to affect organisms that live in the water column, including algae, non-target plants, fish and aquatic invertebrates. While some other receptors such as mammals and birds may spend a considerable portion of their time in the water, they are generally more likely to be affected by other exposure routes, primarily dermal contact during application and incidental ingestion of residues in the sediment during foraging.

Non-Target Aquatic Plants and Algae Due to their engineered mechanisms of action, imazapyr and glyphosate are toxic to a wide variety of plants. Native salt marsh plants, aquatic macrophytes, and algae in the Estuary waters where the herbicides would be applied could be negatively affected. However, both imazapyr and glyphosate are ineffective for treating submersed aquatic vegetation. Imazapyr. The species most sensitive to technical grade imazapyr and an herbicide/surfactant mixture appear to be freshwater aquatic macrophytes, with reported EC25 values for duckweed (Lemna gibba) of 0.013 mg/L for growth and for common water milfoil (Myriophyllum sibiricum) of 0.013 mg/L for shoot growth and 0.0079 mg/L for root growth (Hughes 1987; Roshon et al. 1999; both in SERA 2004). Aquatic algae appear to be substantially less sensitive. The most sensitive species of algae tested was a unicellular green algae (Chlorella emersonii) with an EC50 of about 0.2 mg/L for growth. Some algal species appear to be stimulated rather than inhibited by imazapyr concentrations of up to 100 mg/L (Hughes 1987 in SERA 2004). Some species of plants, including aquatic plants, may develop resistance to imazapyr. Bioassays conducted on Chlorella emersonii indicated that resistant strains may be less sensitive by a factor of 10. (Land- stein et al. 1993 in SERA 2004.) Due to the infrequent application of imazapyr for control of the perennial Spartina (i.e. once per year), development of resistance to imazapyr is unlikely. Recent studies conducted in Washington State also document the potential for imazapyr to impact non-target vegetation. Effects of imazapyr application on non-native Japanese eelgrass were compared to glyphosate application. For both herbicides, the eelgrass canopy was killed if herbicide was applied on dry eelgrass at low tide with imazapyr being more toxic. Application onto an eelgrass bed with a thin overlying film of water did not result in toxic effects. Within 12 months, all treated eelgrass beds had recovered. Persis- tence was not reported from the sediment underlying these eelgrass beds (Patten 2003). ISP partners have not experienced a situation where eelgrass beds are growing near the target Spartina in the San Francisco Estuary, so no collateral damage is expected to this species. Glyphosate. In laboratory growth inhibition studies with submersed aquatic plants no adverse effects on the growth of elodea (Elodea canadensis), water milfoil (Myriophyllum

San Francisco Estuary Invasive Spartina Project 27. 2013 Aquatic Pesticide Application Plan spicatum), and wild celery (Valisneria americana) were found with glyphosate concentrations of up to 1 mg/L (Forney & David 1981 in WS FEIS 1993). These results are consistent with the findings of other investigators who report that submerged plants are either resistant or affected only by very high glyphosate concentrations (Evans 1978; Peverly & Crawford 1975; both in WS FEIS 1993). A large number of studies with a variety of green algae, blue-green algae, diatoms, and periphyton indicate that glyphosate is slightly toxic to practically non-toxic to most algae. Most algae tolerate concentrations of glyphosate greater than 1 mg/L (WS FEIS 1993). Species of algae vary in their sensitivity to glyphosate in terms of population growth (EXTOXNET, Giesy 2000). Field studies indicate the least toxicity to phytoplankton (microscopic floating algae), possibly because of dilution and adsorption to particulates in open water and flooded marshes. Few data are available on effects to marine algae, as most toxicity tests have been performed on freshwater species. Giesy et al. (2000) reviewed the data available on glyphosate toxicity to micro-organisms, and found that acute toxicity EC50 values ranged from 2.1 to 189 mg/L. NOECs ranged from 0.73 to 33.6 mg/L. Giesy et al. (2000) also reviewed the data available on glyphosate toxicity to aquatic macrophytes, and found that acute toxicity EC50 values ranged from 3.9 to 15.1 mg/L. It should be noted that these studies included tests on the non-aquatic Roundup formulation as well as other forms of glyphosate. The formulated product known as Roundup, which includes the terrestrial surfactant polyethoxylated tallowamine (POEA) is known to be more toxic than the technical grade glyphosate found in aquatic formulations such as Aquamaster. For studies conducted on microorganisms using glyphosate tested as isopropylamine salt, EC50 values ranged from 72.9 to 412 mg/L, and NOEC values ranged from 7.9 to 26.5 mg/L (Giesy et al. 2000). The lowest of these NOEC values (0.73 mg/L) is well above the maximum concentration of 0.026 mg/L reported by Paveglio et al. (1996) and the immediate maximum geometric mean glyphosate concentration of 0.174 mg/L reported by Patten (2002). Therefore, these data indicate that impacts to non-target submerged aquatic plants or algae are not likely. Impacts in estuarine conditions with high concentrations of suspended sediment, which interfere with glyphosate activity, would be even less likely.

Aquatic and Benthic Invertebrates Imazapyr. Imazapyr has been found to have low toxicity to aquatic invertebrates. A study where Daphnia was exposed to an imazapyr formulation produced a 48-hour EC50 concentration of 373 mg imazapyr a.e./L (Cyanamid 1997 in Entrix 2003.). Another study with Arsenal® (similar to Habitat® but only labeled for terrestrial use) with an unspecified surfactant determined a 48-hour LC50 of 350 mg Arsenal/L (79.1 mg imazapyr a.e./L) and a NOEC of 180 mg Arsenal/L (40.7 mg imazapyr a.e./L) for the freshwater flea (Daphnia magna), highlighting the potential effects of surfactants on aquatic toxicity. Other studies also reported 24 and 48-hour LC50 concentrations of greater than 100 mg/L, the highest dose tested (“HDT”), in static tests conducted with newly-hatched Daphnia (Kintner & Forbis 1983 in SERA 2004). Chronic studies reported no adverse effects on survival, reproduction or growth of 1st generation Daphnia after 7, 14 and 21-days of exposure at concentrations up to 97.1 mg/L, the HDT (Manning 1989 in SERA 2004). Test- ing with other invertebrate species that exhibit alternative life cycles has been limited to survival of pink shrimp (Penaeus duorarum) and growth studies with the Eastern oyster (Crassostrea virginica). Acute toxicity to pink shrimp was determined at LC50 >132 mg San Francisco Estuary Invasive Spartina Project 28. 2013 Aquatic Pesticide Application Plan imazapyr a.e. /L, the HDT, which was also the NOEC. The EC50 for growth inhibition of the Eastern oyster was established at a concentration greater than 132 mg imazapyr a.e./L, with the NOEC set at this concentration, the HDT (Mangels & Ritter 2000 in SERA 2004) A recent microcosm study analyzing benthic macroinvertebrates in a logged pond con- firmed the low toxicity of imazapyr to benthic freshwater macroinvertebrates. The study analyzed macroinvertebrate community composition, chironomid deformity rate, and chironomid biomass and concluded that imazapyr did not affect the macroinvertebrate community at the concentrations tested. The NOEC was determined to be greater than 18.4 mg/L (Fowlkes et al. 2003). Glyphosate. Glyphosate is only slightly toxic to practically non-toxic to marine and freshwater aquatic invertebrates. Acute toxicity for freshwater invertebrates varies from 545 to 780 mg/L for water flea (Daphnia magna), to 673 mg/L for mosquito 4th instar (Anopheles quadrimaculatus), to 1,157 mg/L for a leech (Nephaelopsis obscura). Acute toxicity for marine invertebrates was reported as greater than 10 mg/L for Atlantic oyster larvae (Crassostrea virginica), 281 mg/L for grass shrimp (Palaemonetes vulgaris), and 934 mg/L for fiddler crab (Uca pugilator) (ExToxNet 2005; Henry 1992, Heydens 1991; both in SERA 2004). The wide variation in the aquatic toxicity of glyphosate has been attributed to the dilution water, temperature, formulation, and the amount of suspended sediment in the water. Toxicity appears to increase with temperature, and decrease with elevated pH and suspended sediment (Schuette 1998). Giesy et al. (2000) reviewed the data available on glyphosate toxicity to aquatic invertebrates. Few data were available for marine species, and those studies that did use marine species were conducted with glyphosate acid, not salt. Acute toxicity EC50 values for five marine species ranged from 281 mg/L to greater than 1000 mg/L, and NOEC values ranged from 10 to 1000 mg/L. Data compiled by Ebasco (1993) include mortality tests on two marine species, for which EC50 values were found to be 281 mg/L and greater than 1,000 mg/L. Grue et al. (2002) conducted laboratory studies to evaluate reproductive effects of exposure to Rodeo mixed with four different surfactants, including R-11, LI 700, and Agri- dex, on Pacific oysters. The EC50 for glyphosate alone was 68.1 mg/L, the EC50 for the tank mix including Rodeo and R-11 surfactant was 29.9 mg/L, and the EC50 for the R-11 surfactant alone was 1.0 mg/L.

The lowest of these NOEC and LC50 values (10 mg/L) for glyphosate or glyphosate/surfactant mixtures is well above the maximum glyphosate concentration of 0.026 mg/L reported by Paveglio et al. (1996) and the immediate maximum geometric mean glyphosate concentration of 0.174 mg/L reported by Patten (2002). Therefore, these data indicate that impacts to aquatic invertebrates due to post-application water concentrations of glyphosate are unlikely in experimental conditions. Impacts in estuarine conditions with high concentrations of suspended sediment, which interfere with glyphosate activity, would be even less likely. Kubena et al. (1997) conducted sediment and water toxicity studies on marine invertebrates (oysters and amphipods). The LC50 values for Rodeo and surfactant in water ranged from 200 to 400 mg/L, and the LC50 values in sediment ranged from 1000 to 6000

San Francisco Estuary Invasive Spartina Project 29. 2013 Aquatic Pesticide Application Plan mg/kg. These LC50 values are well above the highest measured geometric mean sediment concentrations of 2.3 mg/L reported by Kilbride et al. (2001) and Patten (2002). Field studies of glyphosate/surfactant applications to tidal mudflat invertebrate communi- ties in Willapa Bay, Washington, agree with laboratory tests, which indicate low potential for adverse impacts to benthic invertebrates. Sampling of benthic invertebrates in mudflats up to 199 days after glyphosate/surfactant (X-77) applications revealed no short- term or long-term effects. Short-term laboratory tests of amphipods exposed to glyphosate and surfactants did not affect survival even at high concentrations relative to post- spray field conditions (Kubena 1996).

Fish Imazapyr. As detailed in both the 2003 Entrix and 2004 SERA reports, a number of standard bioassays submitted to the U.S. EPA in support of the registration of imazapyr indicate very low toxicity to fish with 96-hr LC50 values greater than 100 mg/L in most studies. According to U.S. EPA’s ecotoxicity classification for aquatic organisms (see Table A-6), these values classify imazapyr as practically non-toxic, the lowest category for addressing acute risk to aquatic organisms from exposure to chemicals (U.S. EPA 2005). A recent study suggests that Habitat® has relatively low toxicity to juvenile rain- ® bow trout. The LC50 determined for Arsenal (a terrestrial formulation similar to Habi- tat® that did not contain any surfactants) was determined at 22,305 mg imazapyr a.e./L (King et al. 2004).

One study reported much lower 96-hr LC50 values of 4.7 mg/L for Nile tilapia (Tilapia nilotica) and 2.7 mg/L for silver barb (Barbus genionotus) (Supamataya et al. 1981 in SERA 2004). Although the herbicide used was not specified, it is likely that a formulation was used rather than the technical grade active ingredient. Historically imazapyr herbicides contained surfactants beause they were used in terrestrial systems for forestry, and a formulation that removed the surfactant was only developed in 1992. The use of an herbicide containing surfactants might explain the considerably lower LC50 values. The 2004 SERA report used the lowest LC50 value from this study, 2.7 mg/L, for their risk assessment despite some reservations about the study due to the fact that they only had access to its abstract and because the species studied were not native to the U.S. Never- theless, the 2004 SERA report assumed that, even though the study was not well documented, the response of these apparently sensitive species may well encompass the response of other sensitive species native to the U.S. (SERA 2004, p. 4-22). This conclusion is supported by a study that examined the comparative sensitivity of eight ESA- listed fish species to standard test organisms exposed to five different pesticides or metals in order to validate the use of surrogate species as a predictive tool in toxicological assessments. Based on their findings, the authors concluded that a safety factor of two would provide a conservative estimate in risk assessments for listed cold-water, warm- water and euryhaline fish species (Sappington et al. 2000 in Entrix 2003, p. 49). Glyphosate. Acute toxicity studies with warm and cold water fish indicate that technical glyphosate is slightly to practically non-toxic (U.S. EPA 1993). Acute toxicity LC50 values were reported at 86 mg/L in rainbow trout, 120 mg/L in bluegill sunfish, and 168 mg/L in harlequin (ExToxNet 2005). Chronic toxicity studies with a terrestrial formulation of glyphosate, Roundup®, found no significant adverse effects on growth, carcino- genicity, feeding, and agonistic behavior in rainbow trout fingerlings. The authors con-

San Francisco Estuary Invasive Spartina Project 30. 2013 Aquatic Pesticide Application Plan cluded that sublethal levels of the formulation are relatively non-toxic (Morgan & Kice- niuk 1992 in WS FEIS 1993). A recent study with the aquatic formulation Rodeo® determined the LC50 for juvenile rainbow trout at 782 mg glyphosate a.e./L. Giesy et al. (2000) reviewed the data available on glyphosate toxicity to fish. Although some data were available for anadromous species, it appears that all tests were conducted using freshwater test methods. Acute toxicity LC50 values for glyphosate tested as isopropylamine salt ranged from 97 to greater than 1,000 mg/L and NOEC values ranged from <97 to 1,000 mg/L. Data compiled by Ebasco (1993) on one-day acute toxicity tests indicate EC50 values ranging from 12.8 mg/L to 240 mg/L.

The lowest of these NOEC and LC50 values (12.8 mg/L) for glyphosate or glyphosate/surfactant mixtures is well above the maximum glyphosate concentration of 0.026 mg/L reported by Paveglio et al. (1996) and the immediate maximum geometric mean glyphosate concentration of 0.174 mg/L reported by Patten (2002). Therefore, these data indicate that impacts to fish due to maximum post-application water concentrations of glyphosate are unlikely in experimental conditions. Impacts in estuarine conditions with high concentrations of suspended sediment, which interfere with glyphosate activity, would be even less likely.

Birds Imazapyr. Only a few toxicity studies exist for birds. No adverse effects were noted at imazapyr concentrations of up to 5,000 ppm in the diet. Based on the highest doses tested and the U.S. EPA ecotoxicity categories, these results suggest that imazapyr is practically non-toxic to birds through the oral pathway. No data exist for the potential toxicity of imazapyr to shorebirds (Fletcher 1983a,b,c,d in SERA 2004). However, research in Washington State (Patten & Stenvall 2002) indicates that shorebirds do not forage within stands of non-native Spartina, which reduces or eliminates their exposure via the ingestion pathway. No studies exist on toxicity to raptors or on preening or inhalation exposure potentials. Glyphosate. Effects of glyphosate on birds have been tested on mallard ducks (dabbling ducks which ingest wetland sediment along with seeds, insects, and vegetation) and bobwhite quail. Glyphosate is no more than slightly toxic to birds. Several single-dose acute oral studies indicate that glyphosate is practically non-toxic to upland birds and only slightly toxic to waterfowl (U.S. EPA 1993). As with mammals, very high dietary concentrations of glyphosate (a 4,640 mg/kg dietary concentration) resulted in no adverse reactions such as weight loss or mortality (Ebasco 1993). Chronic exposure studies with glyphosate determined a NOEC of 1,000 ppm in the diet (Heydens 1991 in WS FEIS 1993). Little data is available on toxicity of surfactants to birds.

Mammals Imazapyr. Based on U.S. EPA ecotoxicity criteria, imazapyr is considered practically non-toxic to mammals via oral or dermal administration based on acute and chronic studies conducted with a variety of mammalian species. For example, the reported acute oral LD50 for technical imazapyr in rats is greater than 5,000 mg/kg body weight (b.w.). Rats were observed to rapidly excrete imazapyr in urine and feces with no residues detected in their liver, kidney, muscle, fat, or blood. No observable effect was noted for any formulation of imazapyr administered dermally. Very few inhalatory studies were performed and

San Francisco Estuary Invasive Spartina Project 31. 2013 Aquatic Pesticide Application Plan none tested concentrations high enough to determine acute toxicity. Inhalatory effects at sublethal concentrations (<5 mg/L aerosol) were found with technical grade imazapyr resulting in slight nasal discharge and congested lungs. Technical grade imazapyr and imazapyr isopropylamine salt were both found to be moderately irritating to rabbit eyes with complete recovery within 7 days. Technical grade imazapyr is reported as mildly irritating to rabbit skin. Commercial formulations of imazapyr appear to be less toxic by dermal exposure (Entrix 2003, p. 42-44). Chronic and subchronic toxicity studies with imazapyr utilizing dogs, mice, and rats as test subjects did not suggest any systemic toxic or carcinogenic effects (SERA 2004). Glyphosate. Glyphosate has been determined to be practically non-toxic to mammals by ingestion with an acute oral LD50 of 5,600 mg/kg b.w. in rats. The NOEL for chronic toxicity to rats has been determined at 362 mg/kg b.w./day (8,000 ppm) and LOEL at 940 mg/kg b.w./day (20,000 ppm) (USDA 1981; Monsanto 1983; both in WS FEIS 2003). The reported acute LD50 values for dermal effects range from >5,000 to 7,940 mg/kg for rabbits. Subchronic oral toxicity studies of glyphosate with rats and dogs indicate that oral does of up to 2,000 ppm do not significantly affect behavior, survival, or body weight. Laboratory studies of the chronic effects of glyphosate show that it is slightly to practically non-irritating to rabbits’ eyes. No significant reproductive, teratogenic, mutagenic, or carcinogenic effects from exposure to concentrations of up to 300 ppm were reported in 20-year laboratory studies with rats, dogs, rabbits, and mice. Little is known about potential interactive effects between applied imazapyr or glyphosate/surfactant solutions and cumulative loads of herbicides, insecticides, detergents, per- fume agents, and many other organic contaminants in the San Francisco Estuary. It is reasonable to assume that cumulative, interactive effects occur in organisms of the Estuary, but the complexity of multiple interactions in uncontrolled field conditions makes defini- tive research difficult. In practice, total dosages of imazapyr or glyphosate/surfactant solutions applied in field conditions (amount and concentration of solution applied, and the number of subsequent applications to eradicate survivors) depends on many factors which are independent of the physiology of imazapyr, glyphosate or the surfactants themselves. The physiological activity and health of the plant, and the interference of spray coverage by persistent dead leaves or sediment films, all can affect the percent kill of vegetation, and the ability of regenerative buds to survive and re-establish the population. Regeneration requires re- application of herbicide or other eradication methods.

Receiving Water Monitoring Triggers In the proposed revisions to the Statewide General NPDES Permit for Aquatic Weed Control released June 27, 2012, the State Water Resources Control Board describes new Receiving Water Monitoring Triggers. In the absence of Receiving Water Limitations or other adopted criteria, objectives or standards for imazapyr, the State Water Board used data from the U.S. EPA Office of Pesticides Ecotoxicity Database to develop Receiving Water Monitoring Triggers to protect all beneficial uses of the receiving water. Toxicity studies were reviewed and a monitoring trigger was set at one-tenth of the lowest LC50 (lethal concentration that killed 50% of a test species in laboratory toxicity tests) for the most sensitive freshwater aquatic species. The most sensitive species in this review of the literature by the Water Board was rainbow trout, with a 96 hour LC50 of 112 mg/L; con-

San Francisco Estuary Invasive Spartina Project 32. 2013 Aquatic Pesticide Application Plan sequently, the monitoring trigger was set at 11.2 mg/L. However, according to the lan- guage in the proposed permit, exceeding the monitoring trigger does not constitute a vio- lation of this General Permit but rather requires the Discharger to perform certain specified actions including additional investigations and BMP’s. Based on the past seven years of water quality monitoring, the coalition of ISP partners does not anticipate that applications of imazapyr to eradicate invasive Spartina will ever reach this new trigger. The highest concentration found in a treatment event sample in 2011 was 0.38 mg/L, almost two full orders of magnitude below this new standard.

7. DESCRIPTION OF ALTERNATE, NON-CHEMICAL CONTROL METHODS The Spartina Control Program implements a number of non-chemical control methods in situations where other forms of treatment could be more effective, environmental impacts could be reduced, sites are appropriate for volunteer efforts and educational opportunities, or where local ordinances dictate. These methods are especially appropriate for small, newly establishing infestations, and are most appropriate for Spartina densiflora since it is a discretely-rooted bunchgrass. Non-chemical control methods are also a great way to en- gage the community around a site because they foster volunteer stewardship efforts and public outreach. During the development of the Site-Specific Plans for control of invasive Spartina around the Bay, the Spartina Control Program evaluated every known control method that could be used. A number of criteria were used in the evaluation including efficacy at controlling the Spartina, human health and safety, damage to the marsh habitat and/or other aspects of the environment, impacts on water quality, feasibility of applying the method in the salt marsh, cost, etc. Several non-chemical methods (hand-pulling, manual excavation and covering/tarping) continue to be incorporated into the most recent Site-Specific Plans. The remainder of the methods that were evaluated were found to have significant limitations, and are not part of ISP’s current plans. However, some of these methods may be used in conjunction with the selected control methods at a later date. The entire set of possible control methods that were evaluated are discussed below, starting with the methods that were selected and incorporated into the plans.

ISP-selected Non-chemical Control Methods

Hand-pulling and manual excavation Manual removal methods are the simplest technology for removal of cordgrass. Manual removal includes pulling cordgrass seedlings out of soft marsh sediments or using hand tools such as spades, mattocks, or similar tools to excavate larger plants. Manual removal methods are effective primarily at removing aboveground plant parts, or the discrete root system of Spartina densiflora, but are much less effective at removing rhizomes (the horizontal underground stem that sends out roots and shoots from buds) from species like hybrid Spartina alterniflora that rapidly regenerate shoots. Unless digging removes the entire marsh soil profile containing viable rhizomes and buds, its effect is equivalent to pruning (see Mowing, burning, pruning, and flaming, below). The vigor with which remaining rhizomes resprout and regrow is often proportional to the severity of the disturbance. Frequent re-digging and maintenance is needed to attempt to exhaust rhizome reserves of energy and nutrients, and reduce the population of buds capable of resprouting.

San Francisco Estuary Invasive Spartina Project 33. 2013 Aquatic Pesticide Application Plan For hybrid Spartina alterniflora, the main target plant of ISP around the Estuary, manual removal is only effective on isolated seedlings, or very young discrete clones (asexually reproducing colonies of cordgrass). Manual excavation in tidal marshes is extremely la- bor-intensive. Most cordgrass colonies occur in soft mud, where footing needed for digging is impossible or hazardous, even for workers on platforms, mats, or snowshoe-like boots adapted for walking on mudflats. Dug plants with roots left in contact with moist soil may retain viability and regenerate in place or disperse on high tides to establish new populations, so heavy bags of the removed vegetation and substrate must be hauled manually from the site for proper disposal. As the invasive Spartina populations are reduced around San Francisco Bay, the ISP has increased its use of manual removal methods as part of an Integrated Pest Management (IPM) strategy and a way to reduce the need for herbicide where appropriate. Manual removal is particularly useful on Spartina densiflora, the second most common invasive cordgrass species in the Estuary which is a discretely-rooted bunchgrass as opposed to the deeply-rooted rhizomatous clones of the main problem species, hybrid Spartina alterniflora. The infestations of S. densiflora occur mainly in Marin County with the excep- tion of a few outlier populations in Contra Costa and San Mateo Counties. Many tons of S. densiflora have been removed manually over the years, and as of 2011 virtually all of the S. densiflora infestations that remain around the Bay are managed with purely manual methods to complete the eradication. Disposal of manually removed material, especially root/rhizome systems, is problematic. On-site disposal in marshes may cause additional marsh disturbance and may result in spread of invasive cordgrass by regeneration of viable roots. Where manual removal occurs next to levees, salt ponds, or other non-tidal environments, local disposal may be feasible.

Mechanical excavation and dredging Mechanical removal in marshes uses equipment specially designed for working in semi- terrestrial, semi-aquatic wetland environments. Excavation and dredging are accomplished using (1) amphibious dredges fitted with excavators, clamshells, or "cutterhead" dredges, or (2) excavators working from mats (large wood pile supports placed flat on geotextile fabric placed over the marsh surface). Some locations could allow use of conventional shallow-draft, barge-mounted dredging equipment working within reach of marsh from the margins of navigable channels, particularly at high tide. Where cordgrass colonies lie within the limited reach of track-mounted excavators working from levees, mechanical removal could be performed without entry of equipment to aquatic or wetland environments. Mechanical excavation working to the full depth of the rhizome system (up to three feet) in tidal marshes has the potential to be significantly more effective than manual excavation. Similarly, maceration techniques that almost completely destroy both aboveground and belowground living mass of cordgrass have high potential effectiveness. Both techniques also have significant limitations in the San Francisco Estuary, however. Excava- tors working from levees have an inherent limitation of short reach or access distance, usually a working distance of less than 20 feet for the size equipment that typical levees could bear. Floating barges with clamshell or cutterhead dredges, in contrast, would need to work at high tides within about 70 feet of the leading edge of cordgrass vegetation.

San Francisco Estuary Invasive Spartina Project 34. 2013 Aquatic Pesticide Application Plan Excavators have sufficient reach to dispose of excavated marsh soil and biomass in non- wetland areas, on levees, or in aquatic habitats such as salt ponds, which lack vegetation. Heavy equipment often is used within the Estuary's tidal marshes for purposes other than eradication of cordgrass, including removal of large debris hazards and contaminated materials, and construction or maintenance of ditches or canals. Most of this work is done on mats, to distribute the weight of equipment and protect underlying vegetation. These actions are usually aimed at operations that are highly localized (points or narrow alignments) in the marsh, and usually on the relatively firm marsh plain. Even there, equipment may become mired in soft spots, and removal of mired equipment can damage the marsh. In contrast to this maintenance work, removal of invasive cordgrass involves a randomized, mosaic pattern, and occurs most often in the low marsh and mudflats which do not easily support mats and geotextile fabrics. Thus, control methods based on excavators working on mats would be most applicable to localized, large patches of invasive cordgrass on the marsh plain. Some tidal flats invaded by cordgrass occur on sandy deltas with intertidal sand bars (e.g., San Lorenzo Creek) where equipment could be staged, but this situation is unusual. The feasibility of using mechanical excavation or dredging methods at a particular location would be determined based on site-specific conditions. Aside from the feasibility of these scenarios listed above, excavation or maceration with heavy equipment has been evaluated by ISP partners as simply too damaging to the sensitive estuarine ecosystem, resulting in long-term alterations and scars. If the marsh plain is excavated down to the full depth of the rhizome system of hybrid S. alterniflora, the elevation of that area has been reduced for many years until sufficient sediment can accrete and native plants can establish at an appropriate level of tidal inundation. The disturbance from excavation is also likely to increase erosion by removing the anchoring effects of the vegetation, and can have short-term water quality impacts from the suspended sediment released. When large-scale invasive Spartina control is evaluated for the most appropriate method, the use of aquatic herbicide presents a far lower impact than excavation or maceration.

Covering/tarping This technique is intended to exhaust the reserves of energy and nutrients in cordgrass roots and rhizomes and increase the environmental and disease stress on the plants. Covering typically involves securing opaque geotextile fabric completely over a patch of cordgrass. This excludes light essential to photosynthesis and "bakes" the covered grass in a tent of high temperature and humidity. This technique may be used for discrete clones where the geotextile fabric can be fas- tened to the marsh surface securely with stakes for a sufficiently long period of time. High tides, high winds, and tide-transported debris common in tidal marshes often make this method difficult or impossible to implement in some situations. Care must be taken to cover the entire clone to a distance sufficient to cover all rhizomes, as well as a buffer around the entire clone to account for vegetative expansion. If rhizomes spread beyond the reach of the blanketing cover, rhizome connections to exposed, healthy stems can translocate (pipe) foods to the stressed, starving connected portions of the clone under the fabric, and increase overall survival. Staking geotextile on soft mudflats is very difficult, and may make this method infeasible for most situations at this elevation.

San Francisco Estuary Invasive Spartina Project 35. 2013 Aquatic Pesticide Application Plan Other Non-chemical Methods Evaluated

Mowing, burning, pruning, and flaming Cordgrasses (as well as most other grasses) are well adapted to disturbances that "crop" or otherwise remove aboveground biomass because they have evolved with a variety of herbivores. A single event that removes living aboveground cordgrass biomass generally stimulates cordgrass growth, and as soon as a cordgrass stand refoliates, it begins to "re- charge" its roots and rhizomes with new food reserves. If vegetation is removed with frequency, roots and rhizomes may be prevented from regenerating reserves of energy and nutrition and eventually cordgrass may be killed as its organs of regeneration and storage become exhausted, however this could take many applications throughout the growing season to be successful. If the cordgrass is mown close to the mud surface, it also severs the connections in the leaves (the aerenchyma) that transport oxygen down to roots growing in anoxic, waterlogged sediment, and this could further stress the plant. Repeated close mowing may be used to increase physiological stress to a point that cordgrass cannot regenerate; frequent burning would have similar effects. The use of pruning, burning, and mowing for cordgrass eradication in open mudflats and marshes would require very frequent treatment of all aboveground growth until the cordgrass rhizome/root systems become exhausted. For robust stands of hybrid S. alterniflora, mowing has never been found to be an effective eradication tool, so there is no way to know how many times the technique would have to be implemented to reach the goal. Howev- er, mowing has been effectively implemented by ISP as part of their IPM strategy for the eradication of S. densiflora. This species of cordgrass does not lose all of its aboveground biomass each year during senescence, so stands that were previously treated with herbicide but did not die fully are restricted from producing enough new green growth to conduct another application the following year. Mowing was used to stress the plants and to remove the standing biomass so the status of each plant could be evaluated and an appropriate follow-up treatment implemented where necessary. This method was implemented from 2008-2010 and was so successful that it could be discontinued entirely in 2011 in favor of manual removal. Controlled burning could be used in some situations to remove vegetation prior to other treatments, or to prevent pollen and seed dispersal in founder colonies invading new sites. Burning would be used only in suitable locations, and only during periods of low-wind conditions (normally early morning), when fire hazards in succulent vegetation of tidal pickleweed marshes would be manageable. Ignition, however, may be difficult in cordgrass stands on mudflats, and there is likely to be significant collateral damage to other marsh vegetation and potentially to endangered species that call these systems home. Selective pruning (partial mowing with "weed-whackers" or flaming with hand torches) may be used to remove flowerheads and seedheads of discrete clones of hybrid S. alterniflora to prevent flow of pollen from contaminating seed production of native cordgrass, and to prevent seed production within founding colonies. However, pruning would have little or no effect on the clone's growth rate and must be followed up with other methods to control spread.

San Francisco Estuary Invasive Spartina Project 36. 2013 Aquatic Pesticide Application Plan Mown vegetation without viable seeds or propagules may be left in place or removed from the site. Vegetation containing viable seeds or propagules would require removal from the treatment site and disposal in a suitable area not conducive to cordgrass growth.

Crushing and mechanical smothering This method uses an amphibious tracked vehicle to trample new plant shoots and stems, and cover them with a layer of sediment. The objective is to smother the plant by preventing the use of stems to transport oxygen to its roots and rhizomes. The method would typically be used in the fall, and ideally a period of time after mowing, when young shoots and stems have developed. This method has been used with some success in Washington State, but has never been used in the San Francisco Estuary.

Flooding and draining Flooding and draining techniques entail constructing temporary dikes or other structures (or in some cases simply closing existing flood control structures) to impound standing water or remove water to kill emergent vegetation. Cordgrasses are intolerant of permanent flooding as well as dry conditions, and are generally absent in the diked nontidal salt marshes of the Estuary. Salt evaporation ponds, managed waterfowl ponds, and completely diked pickleweed marsh exclude cordgrasses, native and non-native alike. Atlan- tic smooth cordgrass and English cordgrass are capable of invading tidal marsh pools (salt pans) subject to irregular tidal influence (Campbell et al. 1990, P. Baye, pers. ob- serv.), but they are not likely to survive in typical diked wetlands. When tidal marshes are diked and drained rather than flooded, they undergo rapid physical and chemical changes. Organic matter decomposes when microbes are exposed to air; clays shrink when dewatered; and sulfides formed in oxygen-free mud transform to sul- fates forming strong acids (Portnoy 1999). Therefore, diking and draining, although con- ceivably effective for killing cordgrass, would adversely impact marsh soil chemistry and structure, and the longer salt marsh soils are diked and drained the more difficult these adverse soil changes would be to reverse. For these reasons, diking and draining only would be used in critical situations where no other method is feasible, and only after care- ful evaluation and planned mitigation. Diked salt marsh soils that remain permanently flooded undergo relatively slower and less significant changes. Diked flooded salt marshes would eliminate existing standing vegetation, but are readily re-colonized by young salt marsh vegetation if the diking is brief. Isolating the treatment area for flooding or draining may be accomplished by constructing temporary dikes or by closing openings in existing dikes. Temporary constructed dikes need not be large to accomplish this form of treatment. Low earthen berms (about one foot above marsh plain elevation), constructed using low-ground pressure amphibious excavators, could be built around large colonies of cordgrass within open marsh plains. Alternatively, water-filled geotextile tubes ("inflatable dams"), analogous with inflatable cofferdams used in aquatic construction/dewatering operations, may be used. Upon com- pletion of treatment, berms would be graded down to marsh surface elevation, and inflatable dams removed. Temporary dike structures may be difficult to construct in tidal mudflats. Mudflat sediments are usually too soft to "stack" into berms, and firmer material placed on fluid or plastic muds simply subsides into the flats. Similarly, inflatable dams may not be feasible for softer tidal flats.

San Francisco Estuary Invasive Spartina Project 37. 2013 Aquatic Pesticide Application Plan Many populations of non-native cordgrasses have invaded marshes restored by breaching dikes. In these situations, a dike-enclosed tidal marsh could be temporarily re-closed ("choked") by placing a sheetpile barrier in the existing breach, thus creating a temporary lagoon and effecting mass cordgrass eradication. Water control structures (adjustable tidegates) may be installed to enable marsh managers to maintain water depths lethal to cordgrass, suitable diving duck habitat, and adequate water quality. Marsh recolonization is expected to proceed rapidly following restoration of tidal flows. An alternative form of treatment, intermediate between flooding and draining, would be to combine impoundment of water with deliberate solar evaporation, creating hypersaline lagoons. Hypersaline conditions would make the habitat transformation even more rapidly lethal for invasive cordgrass. Restoring tidal flows to temporary salt ponds, however, may require dilution of brines, which could increase the already high cost of these methods making them infeasible for ISP.

8. HERBICIDE APPLICATION AREAS FOR 2013 Twenty-four sites containing 153 sub-areas and less than 39 acres of non-native Spartina are slated for treatment during the 2013 Spartina Treatment Season (Figure 3). Approxi- mately 141 of the sub-areas will potentially be treated with herbicide for at least a portion of their control work, while 12 sub-areas will utilize only manual methods (on Spartina densiflora near eradication) and six sub-areas will receive no treatment pursuant to the requirements of USFWS in ISP’s Biological Opinion for 2013. These treatment sites are receiving follow-up applications and many are approaching eradication, hence the significant reduction in the baywide treatment area from over 800 net acres in 2006. In the cases where it will be utilized, aquatic herbicide was determined to be the most efficacious and least impactful method for Spartina control following a thorough evaluation of the available control methods described above. Each site has been evaluated taking into account many factors including the biology and phenology of the target plants, site access, endangered species issues, habitat sensitivity, the efficacy of non-chemical control methods for specific infestations, partner involvement, stakeholder input, and other factors. A brief description of each site follows including the herbicide delivery systems(s) to be used and an identification of the ISP partner involved in treatment work on the site. Detailed maps of each site can be found at http://www.spartina.org/control/sites.htm on the ISP website.

Site 1 – Alameda Flood Control Channel, Alameda County The Alameda County Flood Control Channel (ACFCC) is a large, unlined, trapezoidal channel that runs from east to west through Hayward, Alameda County, draining a nearly 800 mi2 watershed into the San Francisco Bay. The levees on both sides of the ACFCC are topped with multi-use public trails that are part of the San Francisco Bay Trail and Coyote Hills Regional Park. Downstream from Ardenwood Blvd., beyond the levees to the north, are both active and inactive commercial salt ponds, with an industrial facility and parking lot along the furthest upstream levee. To the south are active salt ponds, seasonal wetlands, and Coyote Hills Regional Park. There are currently no housing units, schools or other similar facilities within this lower reach. The combined infestation of the six sub-areas of the Alameda Flood Control Channel comprised one of the largest hybrid Spartina alterniflora infestations in San Francisco

San Francisco Estuary Invasive Spartina Project 38. 2013 Aquatic Pesticide Application Plan Figure 3. Location of 2013 Spartina Treatment Sites within the San Francisco Estuary. Each treatment Bay.site isThe comprised ISP's 2004 of 1-23 mapping sub-areas, effort which estimatedare identified a bytotal letters of roughly(a through 200 z) in contigproject uplansous acresand do ofc- S.uments alterniflora / hybrids on this site spread over approximately 470 acres (43%) of salt Bay marsh and tidal mudflats prior to treatment. The treatments from 2005 -2012 have reduced the infestation by at least 99% and less than 1 acre remains scattered across the sub-areas in 2013. The treatment method at this site is aquatic herbicide, which will be applied by backpack sprayers since the infestation has now been reduced below the level of broadcast aerial or even amphibious tracked vehicles. Partners on this site include the Alameda County Flood Control District and the California Wildlife Foundation. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

San Francisco Estuary Invasive Spartina Project 39. 2013 Aquatic Pesticide Application Plan Site 2 – Bair & Greco Island Complex, San Mateo County The Bair & Greco Island complex is located in the southwest portion of the San Francis- co Bay Estuary. The northern edge of the complex is at Belmont Slough on the border of Foster City and Redwood City. The southern border of the complex is the Union Pacific railroad line adjacent to the SF2 South Bay Salt Ponds restoration site just south of the Dumbarton Bridge. The site includes marsh islands, active and inactive commercial salt ponds, six large sloughs with numerous smaller channels, and other bayfront marsh that is part of the San Francisco Don Edwards National Wildlife Refuge (DENWR). The Bair & Greco Island complex contains many different marsh systems, all of which are impacted to varying degrees by Spartina alterniflora hybrids. Of the roughly 3,060 acres of Baylands within the complex, there were approximately 216 acres (7%) infested with non-native Spartina in 2005. Since much of this site complex is comprised of island sub-areas with significant mobilization challenges, it has been difficult to get complete treatment in any given year, but five years of effective herbicide treatments in this complex leave less than 10 net acres to treat/retreat in 2013. USFWS is requiring through the Biological Opinion that portions of one sub-area (2c) just receive seed suppression in 2013 to minimize potential impacts to clapper rails from the loss of hybrid Spartina cover. Seed suppression involves the application of a dilute solution of aquatic herbicide that will arrest the development of the plant to stop seed dispersal. Since full mortality of the treated plants will not be achieved, most of the aboveground biomass will be preserved for another year to help clapper rails transition. The treatment method at this site is aquatic herbicide, which will be applied by airboat, and backpack sprayer, and a helicopter will be used for the seed suppression application. Partners on this site include the U.S. Fish and Wildlife Service, Don Edwards National Wildlife Refuge, the San Mateo County Mosquito & Vector Control District, and Cali- fornia Wildlife Foundation. Note that this is a complex site composed of multiple sub- areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 3 – Blackie’s Pasture, Marin County Blackie’s Pasture is a small City of Tiburon park located along the shoreline of Richard- son Bay, adjacent to Tiburon Boulevard. With the ample parking provided at the park, it is heavily used by the public for passive recreation. The park is comprised of a 0.7-acre pasture, a small creek channel (“Blackie’s Creek”) along the eastern edge of the pasture, and a shoreline area that includes the channel mouth, a small open mudflat (fed by sediment delivered by the creek), and landscaped pathways and picnic areas. The total area of non-native Spartina at the Blackie’s Pasture site was around one acre in 2005, divided between two sub-areas. This has been reduced to less than about 450 ft2 that will be treated in 2012. The upper portion of this site is represented by the Blackie’s Creek Channel (Sub-Area 3a), wherein Spartina alterniflora hybrids dominated what was previously an open channel. Spartina densiflora and Spartina alterniflora hybrids dominated the mouth of Blackie’s Creek (Sub-Area 3b), readily hybridizing with the native Spartina foliosa stand there. Some excavation of the upper channel occurred in 2005 as part of a public works project, but only the invasive Spartina in the center of the creek was actually removed; the remainder was treated for the first time in 2006. The primary treatment method at this site is manual removal and offsite disposal of the remaining seedlings of S. den-

San Francisco Estuary Invasive Spartina Project 40. 2013 Aquatic Pesticide Application Plan siflora, and a small application of aquatic herbicide on the hybrid S. alterniflora that cannot be effectively removed by digging. The imazapyr will be applied by backpack sprayer. Partners on this site include the City of Tiburon Public Works, Tiburon Audubon So- ciety, and California Wildlife Foundation.

Site 4 – Corte Madera Creek Complex, Marin County The Corte Madera Creek complex is located on the west side of the North San Francisco Bay in Marin County, south of the San Quentin peninsula and San Rafael Bay. The complex consists of a wide corridor stretching upstream from the mouth of Corte Madera Creek to the uppermost point of non-native Spartina growth in this watershed at about 2.2 miles from the mouth. The entire Corte Madera Creek corridor is heavily developed with both residential and commercial facilities. This ISP site complex is comprised of 12 distinct sub- areas that occur on both public and private lands, each of which requires different treatment and public outreach approaches. The infestation at the Corte Madera Creek complex was still in a relatively early stage of establishment at the inception of ISP, with approximately 12 acres scattered over 318 acres, or 4% of the total marsh area. The infestation at this complex contains some unique features, including the only infestation of Spartina anglica in the Estuary, as well as the first documented case of Spartina densiflora hybridizing with the native Spartina foliosa (which became a relatively common occurrence by 2008-2009 where both species existed in close proximity). Most of the sub-areas have tiny S. densiflora infestations remaining that allow for purely manual control to be used to complete the eradication. Several sub-areas also have hybrid Spartina alterniflora, bringing the overall site total to four of the five invasive cordgrass species currently found in the Bay. The Conservation Corps North Bay (formerly Marin Conservation Corps), Friends of Corte Madera Creek and ISP conducted a great deal of manual removal throughout the winters of 2005-2012, and this method had proven highly effective to help complete the eradication work begun with imazapyr, leaving less than 0.1acre of invasive Spartina in this watershed. All of the sites that shifted from chemical to manual control will be revis- ited this season to manually pull/dig the few plants that have resprouted or emerged from the seedbank, while any healthy Spartina densiflora plants will be treated with imazapyr at the larger infestations at Creekside Park and the left bank of the Corte Madera Creek mouth. Partners on this site include the Friends of the Corte Madera Creek Watershed, California Department of Fish and Game, California State Lands Commission, College of Marin, City of Larkspur, Golden Gate Bridge Highway and Transportation District, Mar- in County Parks and Open Space District, and the Marin County Flood Control and Wa- ter Conservation District. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 5 – Coyote Creek & Mowry Slough Area, Alameda & Santa Clara Counties The area encompassed by this Site-Specific Plan includes approximately 3,652 acres of marshland within the San Francisco Don Edwards National Wildlife Refuge that lie between Coyote Creek to the south and the Dumbarton Bridge to the north. The site is surrounded entirely by marsh and salt ponds, and there is very little public access, except for

San Francisco Estuary Invasive Spartina Project 41. 2013 Aquatic Pesticide Application Plan a portion of the Bay Trail along part of Newark Slough (sub-area 5c) and LaRiviere Marsh (sub-area 5d). Within this site complex the sub-areas include restored tidal marshes as well as highly complex and diverse historic marsh habitats that include well- developed channels, high marsh, mudflats, pans, and thin strip marshes. The pioneering infestation of hybrid Spartina alterniflora in the Coyote Creek and Mowry Slough complex was limited in its distribution at the inception of ISP, although some cryptic hybrid morphologies were still evolving within stands of native S. foliosa that weren’t detected until 2008. In sum, these infestations covered approximately 15.3 acres in 2005 scattered over this very large marshland complex, which is equal to just 0.4% of the area. Much of this area was treated for the first time in 2006 using a directed herbicide application tool (spray ball) that the ISP developed with PJ Helicopters of Red Bluff, CA. The results from 2007 treatment were less efficacious and several of these areas need to be treated by broadcast aerial application in 2008 after they had expanded. In 2009, a more aggressive, ground-based treatment strategy was initiated throughout the Refuge in response to increasing discoveries of cryptic hybrids in this extremely valuable habitat. Imazapyr was applied by airboat and backpack sprayer, and these methods will be utilized again in 2013 on the remaining acre widely dispersed throughout the habitat types described above. The partners on this site are the U.S. Fish and Wildlife Service, Don Edwards National Wildlife Refuge and the California Wildlife Foundation. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 6 – Emeryville Crescent, Alameda County The Emeryville Crescent marsh is a 103.5-acre, fringing mixed pickleweed (Sarcocornia pacifica) marsh shoreline between Powell Street in Emeryville and the eastern landfall of the San Francisco Bay Bridge. Two sub-areas, Emeryville Crescent East (6a) and Em- eryville Crescent West (6b), have been delineated due to the historical ownership and maintenance of the site. The site abuts an extremely heavily developed area on the east side of the Bay, with Interstate 80/580 directly adjacent to the east and the approach to the Bay Bridge adjacent to the south. Local anglers frequently use the marshlands included in this site. Illegal activities such as dumping and littering, unauthorized camping, and public inebriation also occur along the edges of, and sometimes within, the marshlands of this site. The non-native Spartina infestations at Emeryville Crescent were in the early stages of establishment on the site in 2005, at less than 2.6 acres or 2.5% of the combined marsh acreage. The infestation was located mainly along the bay edge of the marsh, adjacent to the open mudflats on the outer edge of the site, with a few scattered clones that were establishing on the interior portions of the marsh within sub-area 6b. Several new hybrid areas appeared since 2008 that were developing within the extensive S. foliosa band and were hard to detect until they reached critical mass. Treatment in 2013 will be required on less than 0.1 acre after several years of thorough applications with high efficacy. The primary treatment method at this site is aquatic herbicide, which will be applied by backpack sprayer. Partners on this site include the California Department of Parks and Recrea- tion and the East Bay Regional Parks District.

San Francisco Estuary Invasive Spartina Project 42. 2013 Aquatic Pesticide Application Plan Site 7 – Oro Loma Marsh, Alameda County Oro Loma Marsh is a large, 324-acre, recently restored salt pond located on the eastern shore of the San Francisco Bay Estuary adjacent to the town of San Lorenzo, about 1.5 miles south of the Oakland International Airport. The marsh is surrounded by levees, with Bockmann Channel and Sulfur Creek bordering the marsh to the north and south, respectively. The San Francisco Bay Trail, a multi-use public recreational pathway, uti- lizes the levee to the west of Oro Loma, and the Southern Pacific Railroad borders the marsh to the east. The surrounding area includes various industrial and commercial developments to the north and south including a sewage treatment plant, electrical substa- tion, and capped landfill. Beyond the railroad to the east are residential developments, the Skywest Golf Course, and Hayward Municipal Airport, with I-880 approximately 0.5 mile from the marsh edge. The entirety of the Oro Loma Marsh site was infested with hybrid Spartina alterniflora across both sub-areas prior to successful helicopter treatment in 2005, totaling approximately 100 acres (31% of the marsh). This was one of the largest infestations of hybrid S. alterniflora in the eastern Central Bay, and was adjacent to several other large infestations including sites along the San Leandro/ Hayward shoreline. Successful helicopter treatments also occurred in 2006 & 2007, and scattered stands of Spartina throughout the site were treated from 2008-2011 by Hydrotraxx and airboat, leaving less than one acre in 2013. The primary treatment method for 2013 at this site is aquatic herbicide, which will be applied to the remaining stands and scattered plants by airboat and backpacks. The partner on this site is the East Bay Regional Parks District.

Site 8 – Palo Alto Baylands, Santa Clara County The Palo Alto Baylands is a 1,030-acre nature preserve and park complex owned by the City of Palo Alto and located on the bayfront south of the Dumbarton Bridge. The site includes all marshland from San Francisquito Creek south to Charleston Slough. The site is part of a large expanse of intact native Spartina foliosa and pickleweed (Sarcocornia pacifica)/gumplant (Grindelia stricta) marsh habitat. The park has high visitation on the established pathways through the marsh, and serves as an educational center for bird- watchers, naturalists, local schools, and the public. Other recreational users of the preserve include hikers, kayakers, anglers, bikers, joggers and many others. The pioneering infestation of hybrid S. alterniflora on this site covered approximately one acre in 2005, scattered along the bayfront, mostly south of the Baylands Nature In- terpretive Center, but discovery of numerous cryptic hybrid forms over the years indicates the original infestation was significantly larger but had not yet developed the critical mass to be detected amongst so much S. foliosa. Successful treatment from 2005- 2012 has eliminated many of the original infestation areas, and kept new discoveries from expanding into significant problems. There are several main areas of infestation: the southeast side of the "Yacht Club Bay" near Mayfield Slough, clones near Hooks Point and Hooks Island, and scattered plants along the bayfront from Mayfield Slough to Charleston Slough. The treatment method at this site is aquatic herbicide, which will be applied to about 0.3 acre of hybrid Spartina by backpack sprayer, with one area only accessible by boat. Partners on this site include the City of Palo Alto and Palo Alto High School.

San Francisco Estuary Invasive Spartina Project 43. 2013 Aquatic Pesticide Application Plan Site 9 – Tiscornia Marsh (formerly Pickleweed Park), Marin County Tiscornia Marsh is a 10-acre Marin Audubon site (formerly a City of San Rafael Park) located on the edge of San Rafael Bay in the northwestern San Francisco Bay Estuary. It is bounded to the north by San Rafael Creek and to the south by East Canal Street, with the community center at Pickleweed Park on the western border. The marsh is comprised of a thin band of high marsh pickleweed/gumplant habitat, which grades abruptly from a 4-5 foot escarpment to an extensive mudflat extending bayward. This band of marshland tapers as it extends southward along the park boundary, and becomes very thin as it curves eastward along the riprap of a levee surrounding an area filled for development. There is an east/west wooden service walkway through the marsh that provides access to the Pacific Gas and Electric (PG&E) power line tower adjacent to the site. At the inception of ISP’s efforts in this area in 2004, there was a well-established population of Spartina densiflora within the fringing marshlands of the park. These plants were at their greatest density in the marsh near the outlet of San Rafael Creek and formed a dense stand on the Bay edge surrounding the escarpment near an electrical tower. Scat- tered plants extended southward from this main area along the Bay edge, and then east and southeast to Shoreline Park. The total infestation size of S. densiflora on the site was approximately 0.05 acre. In spring 2005, a pioneering infestation of hybrid S. alterniflora was also discovered just south of the PG&E walkway. Treatment from 2005-2012 has been highly effective; inventory has not yet been conducted for 2013 but there may not be any hybrid Spartina left to be treated. The primary treatment method at this site is manual removal of all S. densiflora (none was found in the spring 2013 survey) with very limited aquatic herbicide application by backpack sprayer if hybrid S. alterniflora re- growth is found. Partners on this site include Marin Audubon, City of San Rafael, and California Wildlife Foundation.

Site 10 – Point Pinole Regional Shoreline, Contra Costa County Point Pinole Regional Shoreline is a 2,315-acre park owned by the East Bay Regional Parks District. It is located on the western edge of the city of Richmond in Contra Costa County. Point Pinole opened to the public in 1973 after the property was acquired from Bethlehem Steel. Bethlehem had acquired the land in the early 1960s from Atlas Powder Co., one of several firms that had manufactured gunpowder and dynamite there for almost 100 years. The park occupies a short peninsula composed of a main upland core with open, grassy parklands interspersed with predominantly eucalyptus woodlands. Along the northern portion of the park is the historical Whittell Marsh (sub-area 10a), characterized by an extensive pickleweed marshplain and Grindelia-lined channels. Southern Marsh (sub- area 10b) on the other end of the park is a narrow band of tidal marsh grading quickly over a 10-20 meter span from high marsh pickleweed to sandy mudflat. Sub-area 10c was added in 2008 with EBRPD’s acquisition of Giant Marsh, which is comprised of a thin pickleweed marshplain criss-crossed by tidal ditches created during a prior era. The non-native Spartina infestations at Point Pinole were in the early stages of establishment at the onset of treatment. The plants are scattered throughout the sub-areas, but their overall acreage is very small. Giant Marsh was first treated in 2008. The primary treatment method for the 0.01 acre of Spartina remaining at this site is aquatic herbicide, which will be applied by backpack sprayer. All S. densiflora (primarily in Whittell Marsh

San Francisco Estuary Invasive Spartina Project 44. 2013 Aquatic Pesticide Application Plan that does not have any hybrid S. alterniflora) will be removed manually and disposed of off-site. The partner on this site is the East Bay Regional Parks District.

Site 11 – Southampton Marsh, Solano County Southampton Marsh is the largest extant marsh within the Carquinez Strait. Its roughly 175 acres are located within the 720-acre Benicia State Recreation Area, Solano County. Highway 780 borders the park on the north and east, Southampton Bay on the south, and to the west stands residential development within the City of Vallejo. Cyclists, runners, walkers and roller skaters use the park’s 2.5 miles of road and bike paths, which circle the perimeter of the park. The marsh lies in the central portion of the park, and consists mostly of high marsh pickleweed/gumplant habitat, with a deep main channel and several smaller channels throughout. Cordylanthus mollis ssp. mollis (soft bird’s-beak), an endangered annual plant species, can be found in some areas of the marshplain and along several channel banks. Access to the marsh is restricted to park personnel and researchers to protect these rare plant populations and other ecological aspects of the site from potential damage. Southampton Marsh contains the only known population of Spartina patens in the San Francisco Estuary. Several large clones were scattered throughout the southern and western portions of the marsh prior to the initiation of treatment, with multiple smaller clones peppered throughout the area including several adjacent to the main channel. There was less than an acre of S. patens at this site in 2005, which has been reduced by herbicide applications in the areas that are not contiguous with Cordylanthus. Some of the areas occupied by S. patens are directly adjacent to, or interspersed with individuals or patches of Cordylanthus. In 2012, State Parks and ISP developed a new strategy to make more progress on this lingering infestation, but implementation has been hampered by rail ex- clusion zones that overlap with the S. patens. There are also several cryptic hybrid S. alterniflora forms that were first identified in 2008, and these have been added to the necessary control work. The primary treatment method at this site is aquatic herbicide for S. patens and for the trace amount of hybrid S. alterniflora, which will be applied by backpack sprayer, with some digging/tarping of S. patens possible on the marshplain. The partner on this site is the California Department of Parks and Recreation.

Site 12 – Southeast San Francisco, San Francisco County The Southeast San Francisco complex includes a scattered group of remnant marshlands totaling 56.7 acres within a heavily industrialized landscape on the western shores of the San Francisco Bay Estuary. The complex is bounded by Mission Creek to the north, and the San Francisco County and City boundaries to the south. The Southeast San Francisco complex is adjacent to an inactive naval shipyard, shipping container facilities, and the San Francisco 49er’s stadium (formerly Candlestick Park), as well as the Bayview residential neighborhood of San Francisco. Within this area there are a number of marshland habitats including intertidal mudflats, native Spartina foliosa stands, riprap shoreline, marshland fill, and pickleweed-dominated tidal marsh plain. The nine sub-areas of the Southeast San Francisco complex contained approximately 8.2 acres of hybrid Spartina alterniflora in 2005. This represents 14.5% of the area of these fragmented remnant marshlands. The individual patches of non-native Spartina within this area represented localized ‘stepping stones’ in the available marsh habitat of the area

San Francisco Estuary Invasive Spartina Project 45. 2013 Aquatic Pesticide Application Plan to the open waters of the relatively lightly-infested North Bay. Treatment occurred at most of these sites from 2006 to 2012, reducing the area needing treatment in 2013 to less than 0.05 acre. The treatment method at this site is application of aquatic herbicide by backpack. Partners on this site include the California Department of Parks and Recrea- tion, Literacy for Environmental Justice (LEJ), Port of San Francisco, Golden Gate Audubon, City of San Francisco Recreation & Parks, California Wildlife Foundation, and the U. S. Navy. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 13 – Whale’s Tail / Old Alameda Creek Complex, Alameda County The Whale’s Tail and Old Alameda Creek Complex is a 564-acre site on the eastern shores of the San Francisco Bay Estuary, south of the San Mateo Bridge and bordered to the east by Union City. This area includes remnant historical marshland patches that pre- date alterations to the site for salt production, channelized flood control structures, restored salt pond marshland, small sinuous channels, high marsh flats, mudflats, eroding scarp, sand/shell beach, small depositional deltas, and other habitats. The areas included within this complex are almost entirely restricted from public access and are either managed by California Department of Fish and Game (CDFG) as wildlife habitat (sub-areas 13d, 13e, 13f & 13i-l), or by Alameda County Flood Control District for flood control purposes (Sub-Areas 13a, 13b, 13c, 13g, 13h). On the northern side of the main channel, formerly diked salt ponds are undergoing restoration activities to convert them to tidally influenced marshlands. To the south of the main channel, Cargill Corporation maintains active salt-producing evaporator ponds. The invasive Spartina at the Whale’s Tail and Old Alameda Creek Complex is one of the oldest infestations of non-native cordgrass in the San Francisco Estuary. The marshlands of this site contained a total of 82 acres of hybrid Spartina alterniflora representing 14.5% of the area. Applications from 2005 -2012 have been very effective, reducing the area to be treated in 2013 to less than 0.25 acre, mostly scattered in channels and other low elevation areas. The infestations are approaching eradication throughout the site in a wide variety of marsh habitats including high marsh pickleweed (Sarcocornia pacifica)/saltgrass (Distichlis spicata), lower marsh Spartina foliosa/mudflat areas, channel banks, edges of salt pans, and bayfront scarps and mudflats. Since the Cargill Mitigation Marsh was a former salt evaporator pond, it was largely unvegetated with native salt marsh species when tidal action was partially restored in 1995. Without any biotic resistance to invasion, the marsh became infested with large, coalesc- ing clones of invasive Spartina. On the eastern portion of the site, these clones had coa- lesced into meadows. The Cargill Mitigation Marsh sub-area contained approximately 19 acres of S. alterniflora hybrids, representing 38.8% of this restoration site in 2005. Treatment over the past six years has reduced the infestation substantially, with less than 0.1 acre scattered over the site left to treat in 2013. Sub-areas 13i-13l have been added in recent years as they have been breached and once again open to tidal exchange, and these sub-areas are similar to Cargill Mitigation Marsh in that the absence of biotic resistance makes them susceptible to Spartina invasion and rapid widespread establishment in the absence of treatment.

San Francisco Estuary Invasive Spartina Project 46. 2013 Aquatic Pesticide Application Plan The treatment method at this site is aquatic herbicide, which will be applied by backpack (successful treatment has eliminated the need for helicopter application as of 2009 as well as amphibious tracked vehicle utilized 2008-2010). Partners on this site include the Ala- meda County Flood Control District, California Department of Fish and Game, and the California Wildlife Foundation. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 15 – South San Francisco Bay Marshes, Santa Clara County The South Bay Marshes are located at the far southern tip of the San Francisco Bay in Santa Clara County, with both San Mateo and Alameda Counties bordering to the northwest and northeast, respectively. The area includes over 100 miles of shoreline, and encompasses some 1,750 acres of marshland. This highly diverse area includes extensive current and former salt ponds, restoration marshes, creek channels and sloughs, bay fill, intact remnant salt marsh, brackish marsh areas, slough edge marshes, pans, islands, mudflats, sand/shell beaches and other marsh habitats. Included within this area are Gua- dalupe Slough, Coyote Creek, Alviso Slough, Permanente Creek, outer Charleston Slough and San Francisquito Creek. There is a high degree of complexity in the South Bay Marshes that will be enhanced significantly by the work of the South Bay Salt Ponds Restoration Project, which will convert sizable portions of former salt-making ponds to various types of marsh habitat. The infestation was approximately 8 acres in 2005, but it had expanded rapidly until 2008 despite the effectiveness of imazapyr, largely due to previously-undetected cryptic hybrids and post-September 1 treatment. The majority of the infestation occurs in the area of Knapp Tract, on Alviso and Guadalupe Sloughs, and along the bayfront adjacent to the DENWR newly acquired ‘salt ponds’ A1 and A2W. Both Ponds A1 and A2W marshes are composed of large expanses of intact native Spartina foliosa (California cordgrass) and pickleweed (Sarcocornia pacifica), interspersed with gumplant (Grindelia stricta) and tule (Scirpus sp.) Both marshes lie adjacent to compacted earthen levee roads, which are gated, locked, and require DENWR permission to access. The other portions of the site are patchy and scattered throughout the far South Bay extending from Charleston Slough in the northwest to the Coyote Creek tidelands in the northeast. Until 2008 the marshes of Faber Laumeister were thought to be free from hybrid Spartina, but surveys and genetic testing showed that this area required treatment for the first time in 2009. Both Faber Laumeister and Stevens Creek Tidal Marsh (in Shoreline Regional Park) have been effectively treated and their infestations are approaching eradication. The primary treatment method at this site is aquatic herbicide, which will be applied by conventional spray truck and backpack sprayer to approximately 0.5 net acre. Partners on this site include the Santa Clara Valley Water District, City of Mountain View, U.S. Fish and Wildlife Service, Don Edwards National Wildlife Refuge, and the California Wildlife Foun- dation.

Site 16 – Cooley Landing, San Mateo County Cooley Landing is a 165-acre salt marsh restoration site located on the western shoreline of the Estuary in East Palo Alto, south of the Dumbarton Bridge and adjacent to the point where the Hetch-Hetchy Aqueduct makes landfall on the western shore at Menlo Park.

San Francisco Estuary Invasive Spartina Project 47. 2013 Aquatic Pesticide Application Plan The site is a former salt production evaporator pond that is undergoing restoration to tidal marsh. Initial restoration activities were completed between September and December of 2000, and included the excavation of two breaches through the east levee at locations of historic tidal channels. Revegetation of the former salt pond was expected to occur through natural colonization as opposed to active planting. Performance criteria for the restoration of Cooley Landing required 70 percent cover of salt marsh vegetation and less than five percent cover of non-native vegetation by the tenth year following restoration. Cooley Landing is part of the Ravenswood Open Space Preserve. Prior to opening Cooley Landing to tidal action in 2000, just five adult clones of invasive Spartina covering a total of 0.1 acre were present along the levees outboard of the restoration area. However, since some hybrid S. alterniflora was known to occur both north and south of the restoration area, and restored salt ponds lack the biotic resistance in the form of an established native plant community, the infestation spread rapidly and already covered 20 acres of the restoration site or 13 % of this large area by 2005. Treatment from 2006-2012 has reduced the infestation by 85%, but a significant amount of hybrid Spartina will be need to be treated in 2012 and this invader is still present to varying degrees throughout the site. USFWS Sacramento required in ISP’s Biological Opinion that a majority of this site not be treated in 2011, so the remaining infestation in that area was allowed to rebound and expand. The site was effectively treated in 2012, leaving approximately 2 net acres of hybrid for 2013 treatment. The treatment method at this site is aquatic herbicide, which will be applied by airboat, backpack and truck. Partners on this site include S.S. Papadopoulos Associates, Midpeninsula Regional Open Space District, and the California Wildlife Foundation.

Site 17 – Alameda Island / San Leandro Bay Complex, Alameda County This ISP site complex includes all marshlands of the Alameda and San Leandro Bay Area extending from the western tip of Bayfarm Island and San Leandro Channel in the west, to east of Interstate 880 and the Oakland Coliseum in the east. The northern boundary of the site is the Port of Oakland shipping terminals, and the southern edge is 98th Ave on San Leandro Creek. This area supports many diverse habitat types despite the fact that it is within such a highly developed land matrix. Within this area there are recently restored tidal marshes, freshwater ponds and upland islands, highly complex and diverse historic marsh habitats that include channels, high marsh, mudflats and pans, thin strip marshes along rip-rapped shoreline, public parks and trails, open mudflats, creek channels and mouths, sandy beach areas, marinas, private residences, commercial areas, industrial manufacturing facilities, shipping, and many other land use types. The Spartina infestations within this site are distributed throughout the habitat types described above and were very well established at the initiation of treatment. In 2005, the shoreline of this site contained 88.5 acres of non-native Spartina targeted for control by ISP partners. Arrowhead Marsh, MLK New Marsh and Fan Marsh supported the largest infestations of Spartina in the Alameda and San Leandro Bay Complex and none of these were permitted for treatment since 2010 pursuant to ISP’s Biological Opinion (BO) from USFWS; of these sub-areas, only the western half of Arrowhead Marsh is permitted for treatment in the 2013 BO and the infestation at Damon Marsh on the eastern shoreline will also be allowed to rebound and expand.

San Francisco Estuary Invasive Spartina Project 48. 2013 Aquatic Pesticide Application Plan Treatment from 2005 -2010 has reduced the overall infestation significantly throughout the complex, but some sites have only received seed suppression treatments and consequently are still heavily infested. The treatment method for 2013 at this site complex is aquatic herbicide, which will be applied by conventional spray truck, backpack sprayer, and airboat to approximately one net acre of hybrid Spartina. Partners on this site include the Alameda County Flood Control District, East Bay Regional Parks District, City of Alameda, City of Oakland, Port of Oakland, U.S. Coast Guard, Save the Bay, and Cali- fornia Wildlife Foundation. Note that this is a complex site composed of multiple sub- areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 18 – Colma Creek / San Bruno Marsh Complex, San Mateo County This ISP site complex includes all of the marshlands north of the San Francisco Interna- tional Airport, up to the northern tip of the outlet of Colma Creek. Within this area there are broad mudflats and fringe marsh along the shoreline of the industrial fill of South San Francisco, pickleweed and gumplant-lined flood control channels, and upland marsh eco- tones. Much of this area is highly developed with light industrial, commercial, and business facilities, and a portion of the Bay Trail runs through the northern portion of the site. There were an estimated 100.3 acres of marshland within this site in 2005, a substantial portion of which was created and dominated by hybrid Spartina alterniflora. The Spartina infestations within this site are distributed throughout the habitat types described above. In sum, the marshes of this site contained about 60 acres of non-native Spartina at the inception of control efforts, representing 60% of the possible habitat. The infestation of this ecosystem engineer had rapidly expanded onto the open mudflats on the western portion of this site, accreting sediment and raising the elevation to a more suitable level within the tidal regime, as well as constricting the flood control channels on the northern and southern portions. Non-native Spartina had dominated most of the available marshland habitat, with only scattered populations of native tidal marsh plant species remaining in the area. Because of the unusually large and high density population of clapper rail in this site complex, treatment was phased over several years, with only the upper channels receiving applications in 2006. The entire site was treated in 2007, with approximately 60% just “chemically mowed” with a low, sub-lethal concentration of imazapyr to stop seed production while maintaining the above-ground biomass for rail refugia. In 2008, all remaining hybrid Spartina was treated with the full concentration of imazapyr. This strategy was remarka- bly successful and less than one net acre remains scattered over the complex as of 2013, with the western portion of San Bruno Marsh now largely devoid of Spartina after just five treatments. The treatment method for 2013 at this site complex is aquatic herbicide, which will be applied by airboat and backpack sprayer. Partners on this site include the San Mateo County Mosquito & Vector Control District, San Mateo County Flood Con- trol District, City of South San Francisco, and the San Mateo County Transit District (SamTrans). Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

San Francisco Estuary Invasive Spartina Project 49. 2013 Aquatic Pesticide Application Plan Site 19 – West San Francisco Bay, San Mateo County This ISP site complex includes all marshlands of San Mateo County extending south from the San Francisco/San Mateo County line in the north to the San Mateo-Hayward Bridge in the South. Many of the sub-areas for this site are small marshes or mudflat areas bordered by light or heavy industrial development, riprap shoreline, Highway 101, the San Francisco International Airport, or other intensive uses. Several are partially restored marshes, but all sites are fragmented from the historical marsh matrix. Only a few of these sub-areas support a diverse marsh, and hybrid Spartina came to dominate each one. The infestations of non-native Spartina that constitute the West San Francisco Bay Com- plex are scattered along the shoreline in many types of habitats. Spartina can be found along the rip-rap of shoreline development, in remnant or newly formed pickleweed marsh, along channels emptying into the bay, amongst sand beaches, within large established marsh, in wide lagoons, on shallow mudflats, and in small coves and sheltered crannies all along the Bay shoreline. In all sub-areas, hybrid Spartina was rapidly expanding into the existing available habitat. Out of an estimated 350 acres of marsh habitat covered by this complex, there were 85 net acres (24%) of hybrid Spartina requiring control in 2005. Years of successful treatment by SMCMVCD has reduced this site complex to approximately 1.5 net acres to be treated in 2013. The primary treatment method at this site is aquatic herbicide, which will be applied primarily by backpack sprayer, with airboat utilized for a couple of mudflat areas. Site 19 included two sites (Sanchez and Burlingame Lagoons) that contained small infestations of Spartina densiflora that had been transplanted anonymously at some point in the past. Successful imazapyr treatment in 2008 had reduced this S. densiflora infestation to the point of purely manual removal by 2009, and this method will be used again in 2013 to continue the eradication if any plants of this species remain. Partners on this site include the San Mateo County Mosquito & Vector Control District, San Francisco Inter- national Airport, Oyster Point Marina, City of Brisbane, City of Burlingame, City of San Mateo, County of San Mateo, City of Foster City, California State Lands Commission, City of South San Francisco, U.S. Coast Guard Reservation, and a number of individual commercial property owners. Note that this is a complex site composed of multiple sub- areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 20 – San Leandro / Hayward Shoreline Complex, Alameda County This ISP site complex includes the marshlands of the San Leandro and Hayward shoreline, Alameda County, extending south from the Metrolinks Golf Course and Oakland International Airport in the north to the San Mateo-Hayward Bridge in the south. The marshland areas in this site complex range from large, complex restored marsh systems to channel-bank fringe marsh. They line the east shore of the Bay, providing a natural border between the highly urbanized and developed areas of the cities of San Leandro, San Lorenzo, and Hayward and the open waters of the Bay. Much of this area is regularly used for passive recreational activities along portions of the Bay Trail and within EBRPD lands. The infestations of non-native Spartina that constitute the San Leandro and Hayward Shoreline Complex are located in many types of habitats. Invasive Spartina can be found

San Francisco Estuary Invasive Spartina Project 50. 2013 Aquatic Pesticide Application Plan along the rip-rap of shoreline fill and levees, in remnant or newly formed pickleweed marsh, along channels emptying into the bay, amongst sand/shell beaches, within large established marsh restoration sites, on mudflats, and in small coves and sheltered marsh areas along the Bay edge. Of 580 acres of marsh habitat within this complex, there were an estimated 204 net acres of non-native Spartina requiring control in 2005, representing 35% of the area. Treatment from 2005 -2012 has reduced this infestation by over 90%, leaving approximately 7 acres by 2013. Several large sites were not treated in 2011 or 2012 pursuant to ISP’s Biological Opinion from USFWS and this continues to be the permit status for these sites in 2013. This includes much of the Robert’s Landing complex (Sites 20d, 20f, & 20g) as well as two of the three sections of Cogswell Marsh (20n & 20o). The treatment method at this site complex is aquatic herbicide, which will be applied by airboat, conventional spray truck, and backpack sprayer. Partners on this site include the Alameda County Flood Control District, California Wildlife Foundation, East Bay Regional Parks District, City of San Leandro, City of Oakland, and the Oro Loma Sanitary District. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site-Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 21 – Ideal Marsh, Alameda County Ideal Marsh is a 179-acre wetland restoration site located on the eastern shore of the San Francisco Bay Estuary that was allowed to naturally restore to unrestricted daily tidal exchange after a levee was breached in a storm. The shoreline marshes of this site stretch for approximately 2.5 miles south of the Alameda Flood Control Channel to a point about a mile north of the Dumbarton Bridge where a levee cuts back to the shoreline. Levees along the eastern edge of this site separate it from current salt evaporator ponds, with the Coyote Hills located just over one mile to the east. The two sub-areas of Ideal Marsh (North and South) were being rapidly colonized with an expanding infestation of invasive Spartina, with a total of 98 acres of hybrid Spartina alterniflora in 2006 representing 54% of the marshlands at Ideal Marsh. These plants occupied all habitat types present in these marshes, including open mudflats, sand/shell beaches, eroding marsh edge, pickleweed plain, levee edge and other areas. Helicopter treatment from 2005-2008 reduced the infestation by 85% or more. The area was treated by backpack for the first time in 2009 with a supporting helicopter application only along the bayfront. The site now contains just small patches and individual hybrid S. alterniflora plants scattered across the marsh plain and in a few of the channels, with a slightly more extensive infestation still remaining along the bayfront (possibly related to tidal inundation and dry time). The treatment method at this site is aquatic herbicide, which will be applied to approximately one net acre primarily by backpack sprayer in 2013, although an airboat will also be utilized along the bayfront. The partners on this site are the U.S. Fish and Wildlife Service, Don Edwards National Wildlife Refuge and California Wild- life Foundation.

Site 22 – Two Points Complex, Alameda and Contra Costa Counties The Two Points Complex is located on the east side of north San Francisco Bay and southeast San Pablo Bay, along the shoreline of the Cities of Richmond, Pinole, Hercules and Rodeo. It includes areas both north and south of the Point Richmond peninsula and the

San Francisco Estuary Invasive Spartina Project 51. 2013 Aquatic Pesticide Application Plan Richmond-San Rafael Bridge. The shoreline adjacent to these sites is heavily developed with land use including commercial, industrial, residential, and park or marina areas. The Bay Trail provides recreational access along the upland edge of many of the marshes in the complex. The tidal systems of this complex include several large, intact historical marshes, many fringing marshes with mixed pickleweed/Spartina vegetation assemblages, mudflats and flood control channels. The infestations at the Two Points Complex were still in the early stages of establishment at the inception of ISP control efforts, with approximately 5 acres scattered over the 598 acres of the six sub-areas, or less than 1% of the total marsh area. The establishing clonal patches within this area were located mostly along the bayfront amongst native Spartina stands or along the banks of small channels within the marshes, and several areas had begun to coalesce into meadows. The channels and marsh plain of San Pablo Marsh were being rapidly invaded until treatment began in 2005 on the western portion and 2007 for the eastern. The easternmost portion of San Pablo Marsh was treated for the first time in 2008 because the plants senesced very early ahead of treatment in both 2006 & 2007. The treatment method at this site for 2013 is aquatic herbicide, applied by airboat and backpack sprayers to approximately 2 net acres of hybrid Spartina. Partners on this site include the State of California Lands Commission, California Wildlife Foundation, Repub- lic Services, Chevron/Texaco, Richmond Rod and Gun Club, and East Bay Regional Parks District. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 23 – Marin Outliers, Marin County This ISP site complex includes selected marshlands within eastern Marin County extending south from Novato to Sausalito. Many of the sub areas are small marshes or mudflat areas bordered by residential or commercial development, while several are partially restored marshes. In sum, these areas represent an important patchwork of small marsh areas within Marin County. The infestations of non-native Spartina that constitute the Marin Outliers complex are scattered along the shoreline in many types of habitats. Spartina can be found along the rip-rap of shoreline development, in remnant or newly formed pickleweed marsh, along channels emptying into the bay, within large established marsh, on shallow mudflats, and in small coves and sheltered crannies along the Bay edge. Non-native Spartina was in the relatively early stages of expansion into the existing habitat throughout the site complex in 2005 at the inception of treatment; of an estimated 130 acres of marsh habitat within the complex there were 2.6 net acres of non-native Spartina representing 2% of the marsh area. Most of these infestations are now down to only a few hundred square feet or less after effective treatments from 2005 -2012. The primary treatment method at a number of sites in this complex is manual removal of S. densiflora after the reductions realized over the past several control seasons using imazapyr. Other sites will still depend on aquatic herbicide treatment (because of the rhizomatous nature of hybrid S. alterniflora), which will be applied by backpack sprayer to less than 0.2 acre. Partners on this site include Marin Audubon, County of Marin, California Department of Fish & Game, California Wildlife Foundation, Loch Lomond Marina, Strawberry Recreation District, City of San Rafael,

San Francisco Estuary Invasive Spartina Project 52. 2013 Aquatic Pesticide Application Plan City of Mill Valley, City of Novato, California State Lands Commission, California State Parks, McNear Brick and Block, Paradise Cay Yacht Club, City of Sausalito, and a number of smaller residential and commercial landowners. Note that this is a complex site composed of multiple sub-areas. Please refer to the Site-Specific Plan for detailed descriptions of the sub-areas and the associated treatment methods.

Site 24 – Petaluma River, Sonoma County This area includes approximately 4,500 acres of marshland and riparian habitat within the Petaluma River Watershed stretching from the City of Petaluma, at the confluence of the Petaluma River and Lynch Creek in the north to San Pablo Bay in the south. This site consists of a complex mosaic of historic tidal marsh habitat, developed shoreline, brackish tidal riparian edge zones, restoration sites, light industrial facilities and urban development. The pioneering infestation of hybrid Spartina alterniflora in the Petaluma River complex was discovered in November 2006, and was still very limited in its distribution. The majority of the infestation is located along the banks of the river adjacent to a dredging and barge dock facility just downstream of the Highway 101 crossing south of Petaluma, with scattered infestations located upstream and downstream of this central core. In sum, these infestations covered less than 0.1 acre scattered over this very large marshland complex, which is equal to less than 0.01% of the area. The first treatment at this site occurred in 2007 and follow up work from 2008-2011 has kept this pioneering infestation in check. Several areas of cryptic hybrids continued to develop within the extensive stands of native S. foliosa until they reached critical mass and could be detected during inventory monitoring. The treatment method at this site for 2013 is aquatic herbicide, which will be applied by backpack sprayer to less than 500 ft2. The primary ISP partner on this site is the Friends of Petaluma River, a group that has existing relationships with the community of the watershed and the landowners along the river, with additional support from Cali- fornia Wildlife Foundation.

Site 26 – North San Pablo Bay, Solano County There are two main portions of this sub area: the restoration site on the east side of the Napa River called White Slough in Vallejo, and the marshes along the northern shoreline of San Pablo Bay including Mare Island, the mouth of Sonoma Creek, and Sonoma Bay- lands. White Slough marsh is a roughly 135-acre restored tidal marsh that lies to the east of Highway 37 and west of Sonoma Boulevard in the City of Vallejo. The marsh is a sparse- ly vegetated tidal marsh in the initial stages of colonization. The majority of the area is open mudflat with tidally inundated low sections. The periphery of the marsh is composed of scattered pickleweed (Sarcocornia pacifica), Spartina foliosa, and alkali bulrush (Bolboschoenus maritimus). ISP considers hybrid S. alterniflora to have been eradicated from this site since none was found or treated for three consecutive years (2010-2012). The other marsh areas included in this sub-area are the massive mudflat and shoreline area of San Pablo Bay National Wildlife Refuge on the west side of Mare Island, the mouth of Sonoma Creek, and the restored marsh of Sonoma Baylands. This large area of restored and historic tidal marsh, developed shoreline, industrial and decommissioned

San Francisco Estuary Invasive Spartina Project 53. 2013 Aquatic Pesticide Application Plan military facilities is only lightly infested with non-native Spartina, including both S. densiflora and hybrid S. alterniflora. All S. densiflora on Mare Island will be removed manually and disposed off-site, while any necessary hybrid treatment in 2013 will be conducted by backpack sprayer to a very limited footprint under 500 ft2.

9. WATER QUALITY MONITORING PLAN (WQMP)

Objective Conduct water quality monitoring for ISP’s coalition partners sufficient to achieve compliance with NPDES Statewide General Permit requirements.

Monitoring Site Selection Up to 141 sites (sub-areas) of invasive Spartina throughout the San Francisco Bay Estu- ary are slated for treatment with aquatic herbicide during the 2013 control season. Ac- cording to the Statewide General Permit and current NPDES requirements, ten percent (10%), or 14 of the 141 sites, must be monitored for water quality. These sites are to be selected at representative locations, and are to include all of the herbicides being applied. The WQMP has been designed to answer the two key questions articulated by the Permit: 1. Does the residual aquatic herbicide discharge cause an exceedance of receiving water limitations? 2. Does the discharge of residual aquatic herbicide, including active ingredients, inert ingredients, and degradation byproducts, in any combination cause or contribute to an exceedance of the “no toxics in toxic amount” narrative toxicity objective? To assist with sampling site selection, ISP identified four different treatment site types, as follows: I. Tidal Marsh, Microtidal Marsh, Former Diked Bayland, Backbarrier Marsh II. Fringing Tidal Marsh, Mudflats, and Estuarine Beaches III. Major Tidal Slough, Creek or Flood Control Channel IV. Urbanized Rock, Rip-Rap, Docks, Ramps, etc. The ISP has selected representative sites of each of these marsh site types to be sampled for water quality. Type IV infestation sites are usually very small, sparse, and adjacent to large bodies of water with constant flushing that will serve to quickly dilute any herbicide incidentally entering the water column. For these reasons, along with the fact that ISP has either eradicated or reduced its Type IV sites to tiny infestations of widely-scattered individual plants, fewer sites of this type will be monitored for 2013. Site Types I and II are considered to be the sites most likely to develop detectible levels of herbicide in the water column. For 2013, the final sampling plan list includes four Type I sites, three Type II sites, four Type III sites, and three Type IV sites chosen to represent the range of herbicide delivery systems and marsh dynamics present in our work program. Imazapyr is the only herbicide that will be utilized by ISP’s coalition partners for Spartina treatment in 2013. Table 5 provides a summary of the sites, and the sites are described in more detail in a subsequent section.

San Francisco Estuary Invasive Spartina Project 54. 2013 Aquatic Pesticide Application Plan Sampling Design The sampling events are designed to characterize the potential risk involved with imazapyr applications relative to adjacent surface waters. Consistent with permit requirements, the monitoring program will include background/pre-treatment sampling up to 24 hours prior to the application, application event monitoring immediately post-treatment, and one-week post-application event monitoring. During background sample collection, the point will be recorded using GPS to aid ISP staff in locating the point for future sampling events. The application event samples will be collected immediately adjacent to the treatment area after sufficient time has elapsed such that treated water will have entered the adjacent area on the incoming tide. Since it is standard protocol for the ISP partners to treat Spartina on a low or receding tide whenever possible, application event samples will often be taken 0.5-5 hours post-treatment when the tide has again flooded the site. Final- ly, the one-week post-treatment monitoring will be conducted when sufficient water is present at the site on the seventh day after the application. To enhance quality assurance, the ISP will be submitting three to four duplicates to the lab over the course of the season for the corresponding 45 total base samples taken. These will be added to either the treatment event or one-week post-treatment event since the herbicide levels in the pre- treatment samples are usually ND (not detected). It is standard for the lab to include blanks as part of their quality control, but the ISP will send additional trip blanks consisting of distilled water on a regular basis.

San Francisco Estuary Invasive Spartina Project 55. 2013 Aquatic Pesticide Application Plan Table 5. Summary of Water Quality Monitoring Sites for the 2013 Treatment Season

Site Treatment Sites Marsh Type Application Number Date Corte Madera Creek 4j III 6/24/13 Imazapyr – Backpack Mouth Creekside Park 4g I 6/24/13 Imazapyr – Backpack

Coliseum Channels 17i IV 7/9/13 Imazapyr – Backpack Sonoma Creek 26c III TBD Imazapyr – Backpack Oakland Inner Harbor 17f IV TBD Imazapyr – Backpack Blackie's Creek Mouth 3b II TBD Imazapyr – Backpack Imazapyr – Airboat, Newark Slough 5c III TBD Backpack Imazapyr – Airboat, Belmont Slough 2a III TBD Backpack Imazapyr – Airboat, SFO 19h IV TBD Backpack

Palo Alto Baylands 8 II TBD Imazapyr – Backpack Imazapyr – Airboat, TBD San Pablo Marsh 22b I Backpack Imazapyr – Airboat, TBD Greco Island South 2h I Backpack

Giant Marsh 10c II TBD Imazapyr – Backpack Imazapyr – Airboat, TBD Seal Slough 19p I Backpack

Field Sampling Procedures In 2013, the Invasive Spartina Project will conduct its water quality monitoring program as a coalition for its partners around the Estuary as it has since 2005. Water samples will be collected using a sampling rod and pre-cleaned amber glass 1-liter bottles. To collect the sample, the bottle is attached to the sampling rod with a clamp, extended out over the water at the application site, and lowered to approximately 50% of the water depth. When the bottle is full it is pulled back out of the water and the cap is affixed to the mouth of the bottle. The sample is labeled in permanent ink with the sample ID number, date, time, and initials of the sampler. The sample ID number is determined by the following protocol: a four-letter code unique to the site, followed by the site visit number (e.g., 01 for pre-treatment, 02 for treatment, or 03 for one-week post-treatment), followed by the time since the application (e.g., “pre” for the baseline sample, the number of hours since the application for the treatment sample, or “1w” for the one-week post-treatment).

San Francisco Estuary Invasive Spartina Project 56. 2013 Aquatic Pesticide Application Plan Equipment Calibration Temperature, electrical conductivity, salinity, and dissolved oxygen will be measured in the field with a portable YSI Model 85 (Yellow Springs Instruments Inc., Ohio, USA), while pH will be measured with an Oakton waterproof pHTestr1 (Oakton Instruments, Illinois, USA). To assure accurate and reliable temperature, electrical conductivity, salinity, and dissolved oxygen measurements, the YSI Model 85 meter will be calibrated, operated, and maintained in accordance with the manual specifications found at http://www.ysi.com/media/pdfs/038503-YSI-Model-85-Operations-Manual-RevE.pdf. To assure accurate and reliable pH measurement, the pHTester 1 meter will be calibrated, operated, and maintained in accordance with the manual specifications found at http://www.4oakton.com/Manuals/pHORPIon/WPpHTestr1_2mnl.pdf.

Field Data Sheets At each sampling location, the sample ID number, the time of the sampling, the sample depth, and the water temperature, pH, dissolved oxygen, conductivity, and salinity measurements, will be entered on a Field Data Collection Form (“FDCF”, Appendix 2). Also recorded on the FDCF will be site information, including the site ID number, the station location (application point, upstream, downstream), station type (reference, treated), wind conditions, tidal cycle, water color, and the type of herbicide and surfactant that might be present. Any other unusual conditions or concerns will be noted, and any fish, birds, or other wildlife present will be recorded. The FDCFs will be dated and numbered consecu- tively for each site on that date. Data from these field forms will be entered into an elec- tronic spreadsheet for processing, and the FDCFs will be compiled into a data log and kept permanently in the ISP office.

Sample Shipment Following collection, water samples will be stored on ice packs and shipped for priority overnight delivery to the Pacific Agricultural Laboratory in Portland, OR. If samples are not shipped until the following day, they will be stored in a cooler on ice until they can be transferred to a refrigerator, and subsequently transferred back into a cooler for shipping.

Field Variances The ISP usually selects and plans to monitor one to two more sites each season than is necessary for compliance with the NPDES Permit, to allow for failed sampling events or analyses. If a situation should arise that precludes being able to collect a water quality sample at a designated point suitably close to the specified times (within 24 hours prior to herbicide treatment, within several hours post-treatment, or one week post-treatment), the Water Quality Monitoring Manager (WQMM) will determine whether (1) sampling at the site type is needed to complete the sampling events required by the NPDES permit, or (2) sampling at the site type is not needed for permit compliance, and the site/event can be dropped. If the site type is needed, then the WQMM will consider whether surrogate sampling of some sort (e.g., sampling at a point reasonably nearby the initial point, or at a later or earlier time) would provide an acceptable substitute. If so, the variation will be carefully documented and justified on the Field Data Sheet. If the WQMM determines that surrogate sampling would not be suitable, then an alternate, similar site will be selected and sampled as a complete replacement for the initial site. Data from samples al-

San Francisco Estuary Invasive Spartina Project 57. 2013 Aquatic Pesticide Application Plan ready collected at the initial site will be kept and reported, along with an explanation of the reason for substitution. Any significant problems with sampling events that cannot be remedied in such a way, or any other significant water quality issues that should arise, will be reported to the US EPA Region 9 Project Manager, or their designee.

Sample Analysis The samples will be analyzed within the appropriate holding times for imazapyr (extracted within seven days, analyzed within 21 days of extraction). Results are reported as parts per billion (ppb), equivalent to µg/L. The analytical method used for imazapyr is EPA 8321B in which the extracts are analyzed using liquid chromatography with mass spectroscopy (LC/MS) detection, with a Limit of Quantitation (LOQ) of 0.1 ppb (the minimum detectable level of the analytical method). The lab runs one blank each time it con- ducts an analysis (minimum of one sample tested per batch, maximum of three). Results will be reported at the end of the season to the San Francisco Bay Regional Water Quality Control Board and placed on the ISP’s website for public viewing.

Assessment of Field Contamination Field Blanks. To help assess contamination from field equipment, ambient conditions, sample containers, transit, and the laboratory, one field blank will be collected and submitted for analysis on a regular basis. Field blank samples will be obtained by pouring distilled water into a sampling container at the sampling point.

Lab QC & Data Quality Indicators Each season, the contracted analytical laboratory (“lab”) is required to provide a Quality Assurance Plan (*QAP”) that meets USEPA standards prior to initiating analysis. The lab plan must specify the method of analysis to be used, and describe any variations from a standard protocol. The WQMM will review the lab QAP and determine if it is adequate.

At a minimum, the following DQIs will be required for the lab: Criteria Method Indicator Goal Accuracy of measure- Analyze matrix spikes and 1 matrix spike per 10 samples (10%) ment spike duplicates > 65% @ 2.0 ug/L

Agreement between Analyze lab duplicates and/ Relative percent difference < 25% measurements or matrix spike duplicates Completeness Percent of usable data 95% return (completed/submitted) Comparability of re- Standard reporting units All data reported in micrograms per liter (ug/L) sults or parts per billion (ppb) Use of standardized analy- Standard method used if possible, any modifi- sis methods cations identified, described, and supported. Detection Limits Limit of Quantitation LOQ

Monitoring Site Descriptions Following are brief descriptions for each of the monitoring sites. Additional site information was included previously in Section 8 of this plan.

San Francisco Estuary Invasive Spartina Project 58. 2013 Aquatic Pesticide Application Plan Corte Madera Creek Mouth. The main infestation area at this Type III site is an island at the mouth of Corte Madera Creek is Type III located behind the Larkspur Ferry Termi- nal, north across the creek channel from Greenbrae Boardwalk and the Corte Madera Ecological Reserve. This island contains the largest infestation of hybrid S. alterniflora in the watershed as a result of seeds being strained out of the incoming tide as propagules enter Corte Madera Creek. This site also contains an infestation of S. densiflora that is one of the largest remaining in the Estuary. This portion of the site was not permitted for treatment in 2011, and consequently both infestations were allowed to rebound and expand; both were treated in 2012 and the S. densiflora was subsequently removed manually in the autumn. The invasive Spartina at the site will be treated with imazapyr in 2013; the application to flowering S. densiflora (before seed set) is intended to arrest plant development to eliminate any further seed bank accumulation in the substrate (or dispersal to other marshes). After clapper rail breeding season is over on September 1, Conserva- tion Corps North Bay will manually remove the remaining S. densiflora for offsite disposal. Creekside Park. This Type I site is estimated to contain approximately 21 acres of marshland habitat adjacent to the upper portion of Corte Madera Creek just upstream of Bon Air Road. The park was restored in 1976, when a new channel system was excavated, upland areas were graded to intertidal elevations, and central islands were constructed from the channel dredge spoils as upland refugia. As part of this initial restoration effort, both Spartina densiflora and Spartina anglica were planted, as native marsh plants failed to establish within the first year of restoration. These plants were imported from Hum- boldt Bay and England respectively. Efficacy from imazapyr applications conducted from 2006 to 2008 controlled the infestation and stopped most seed production, but had not succeeded in full mortality on the meadow areas of S. densiflora plants. Beginning in winter 2008-2009, dense areas of the infestation were mowed to bare ground to allow the remaining live plants to sprout fresh healthy growth that could either be retreated with herbicide or dug. In 2013, Creekside Park will receive an imazapyr treatment to arrest the development of S. densiflora seed until manual removal can be conducted after clapper rail breeding season. The application to other Spartina species on the site, including S. anglica, should be the only tool needed to continue eradication efforts. Coliseum Channels. This Type IV site includes the upper portions of the flood control channels that drain into San Leandro Bay. To differentiate them from the downstream mouths of the channels, the western boundary of these areas is defined as Interstate 880, which runs perpendicular to these channels and west of the Oakland Coliseum, with the eastern edge defined by where tidal marsh plant species are no longer present. These channels are typically steep-sided and degraded, often choked with sediment and copious litter from Coliseum events, and overgrown along their edges with weedy upland species. The infestation here has been reduced to widely scattered patches and individual plants that will be treated by backpack sprayer to complete the eradication. Sonoma Creek. This Type III site at the mouth of Sonoma Creek was discovered late in 2008 and eluded treatment in 2009 due to very early senescence. The banks of the creek are lined with continuous stands of native Spartina foliosa that stretch for miles, and these clones were not observed standing out from the crowd until they reached critical mass and an airboat survey was able to detect them. The infestation consisted of two linear stands of very large plants at the mouths of two smaller channels off the eastern bank of

San Francisco Estuary Invasive Spartina Project 59. 2013 Aquatic Pesticide Application Plan Sonoma Creek along the northern shoreline of San Pablo Bay. These stands have been greatly reduced by ISP treatment, and backpack sprayer will be used to continue eradication efforts in 2013. Oakland Inner Harbor. This Type IV site consists of the armored shoreline and marinas of the Oakland Inner Harbor from the old Alameda Naval Air Station south to the High Street Bridge. Treatment throughout these scattered sites on the Alameda and Oakland shorelines began in 2007, with the majority of the work done from a shallow-bottom boat with support from backpacks in areas with no access from the water. This same approach will be used for 2013 treatment on the remaining small patches and scattered plants. Blackie’s Creek Mouth. Blackie’s Pasture is a small City of Tiburon park co-managed by the City of Tiburon and Tiburon Audubon. The park is located along the shoreline of Richardson Bay, adjacent to Tiburon Boulevard. The mouth of Blackie’s Creek, a Type II site, was heavily infested with both hybrid S. alterniflora and S. densiflora when control efforts began. After several years of manual and chemical control work, the infestation now contains a handful of S. densiflora seedlings sprouting from the seed bank, and just scattered small patches of hybrid S. alterniflora. Treatment in 2013 will involve purely manual removal of seedlings of S. densiflora and retreatment of the remaining hybrid S. alterniflora by backpack sprayer. Newark Slough. This Type III site encompasses roughly 400 acres of marsh and creek channel bank stretching from Thornton Avenue and Hickory Street in the City of Newark downstream to the edge of the abandoned railroad line, 900 meters upstream of the confluence with Plummer Creek. In its upstream reach, the wide, levee-bound slough winds sinuously through the Don Edwards San Francisco Bay National Wildlife Refuge, skirt- ing the southwest edge of the large hillside that the Refuge headquarters sits atop, along Marshlands Road just south of the Hwy. 84 approach to the Dumbarton Bridge, and past some decommissioned salt ponds. The fringing marsh upstream of the Refuge headquarters is very wide on the north side of the channel, and contains an extremely high density of gumplant (Grindelia stricta) that dominates large areas of the pickleweed marsh plain. Fringing channel bank marsh habitat borders the waters of the channel along the remainder of its length, often dropping off steeply at the channel’s edge. The site will be treated mostly by airboat, possibly with some backpack sprayer work away from the channel banks. Belmont Slough. This 448-acre, type III site includes Belmont Slough, North Point Res- toration, and the northern shoreline along Redwood Shores. The sloughs are open tidal waters lined with strips of mixed pickleweed/Spartina foliosa marsh of varying widths. The shorelines and islands are comprised of thin to moderate-width open mudflats grading into native Spartina marsh, with some pickleweed/gumplant (Grindelia stricta) marsh at higher elevations. All sloughs and marshes are bordered by levees topped by access roads or the Bay Trail. A newly breached marsh at North Point on the right bank of Bel- mont Slough at the mouth has already been invaded by hybrid Spartina and will need to be included in future monitoring and treatment. Airboat will be used on this site in 2013, a welcome improvement over the previously-used amphibious tracked vehicles in the very soft mud at the mouth. Some support from backpack sprayers will likely be employed at the upper elevation zones where this method would be more efficient. SFO. This Type IV site around the perimeter of San Francisco International Airport (SFO) includes seven distinct edges with varying degrees of marsh development based on San Francisco Estuary Invasive Spartina Project 60. 2013 Aquatic Pesticide Application Plan exposure and accretion, totaling approximately 25 acres. There are two large runway strips that jut out into the bay, the longer running roughly southeast to northwest with the shorter strips running perpendicular. The largest area of marsh is adjacent to the runways along the southern shoreline of SFO, just east of Hwy. 101 in Millbrae. This protected cove has accreted substantial sediment and has prograded marsh out as much as 200m from the concrete and fill. At the Millbrae Avenue security gate to the runways, a large culvert empties a concrete flood control channel that draws stormwater from the airport complex. Two other areas of minimal pickleweed marsh have developed, one on the northeast side of the junction of the two runway strips and the other just south of Sea- plane Harbor to the northwest of the shorter runways at the end of the N. Access Road. Both of these face the open bay, and hence are subject to greater wave energy resulting in less accretion. The site will be treated by a combination of airboat and backpack sprayer in 2013. Palo Alto Baylands. This Type II site is part of a 1,940 acre nature preserve and park complex, one of the largest tracts of undisturbed marshland remaining in San Francisco Bay, owned by the City of Palo Alto and located on the western bayfront approximately 2.5 miles south of the Dumbarton Bridge. The site is located east of Hwy. 101 at the end of Embarcadero Road, and includes those areas south of San Francisquito Creek and north of Charleston Slough. Within the site, Harriet Mundy Marsh is a peninsula vegetated with pickleweed (Sarcocornia pacifica), S. foliosa, and gumplant (Grindelia stricta) that extends out to Sand Point from the main parking area. There is a restored marsh cove to the southwest of the parking area that was once home to a yacht club before it was allowed to silt in and return to marshland. The water quality sample will be taken from the most infested area of the site, Hooks Island, which is located just offshore from Mayfield Slough and contains a heavily channelized pickleweed marsh with large areas of S. foliosa. The site will be treated in 2013 by backpack sprayers (including some island areas in the southwest only accessible by boat). San Pablo Marsh. This Type I site is a 165-acre marsh at the mouth of San Pablo Creek on the City of Richmond shoreline in southeastern San Pablo Bay. San Pablo Marsh has an extensive pickleweed marsh plain, with Grindelia lining the banks of the channels and a S. foliosa fringe on the bayfront as well as in the channels. The marsh stretches east to an 11-acre pickleweed, S. foliosa and alkali bulrush (Bolboschoenus maritimus) cove bordered by levees on either side, located behind the Richmond Rod and Gun Club rifle range. The western portion of this site has been treated since 2005, but the mudflat clones, as well as the infestation encroaching up the channels, were first treated by airboat and backpack in 2008. Efficacy has been very high on the mudflat clones and in the upper channels, but the established heavy infestation along the fringe has taken longer to control, and a portion was not permitted for treatment in 2011 and subsequently was allowed to rebound and expand. Airboat and backpack sprayers will be used again at this site in 2013. Greco Island South. Greco Island is reported to be the largest remaining prehistoric tidal marsh in the South Bay with a total area of 817 acres (Greco Island North sub-area covers 556 acres while Greco Island South is 261 acres). This Type I site is located immediately southeast of Bair Island across the mouth of Redwood Creek and approximately one mile northwest of the western landfall of the Dumbarton Bridge at Ravenswood. The southern shoreline is bordered by West Point Slough and Bayfront Park in Menlo Park. The north-

San Francisco Estuary Invasive Spartina Project 61. 2013 Aquatic Pesticide Application Plan ern shore on the open bay is comprised of wide mudflats receiving flow from many small, shallow sloughs filled with native Spartina that continue up onto the pickleweed marsh plain. The southeastern lobe of Greco contains more plant diversity and a large population of clapper rails, with many sinuous channels heavily lined with Grindelia after being freed from competition with the hybrid Spartina meadows present at the inception of ISP treatment. The channels in the northwestern lobe are currently being enhanced with Grindelia by ISP to create better clapper rail habitat and increase the carrying capac- ity over the entire island. The site will be treated with a combination of airboat and backpack sprayer in 2013. Giant Marsh. Giant Marsh is a 30-acre pickleweed marsh in the far southwestern corner of Point Pinole Regional Shoreline on San Pablo Bay. The infestation at Giant Marsh is composed of scattered clones of hybrid Spartina alterniflora along the bayfront edge of the marsh and out onto the mudflats. This infestation was in the very early stages of establishment (less than an acre) before treatment began, and had not yet colonized the interior of the marsh nor established much along the network of manmade channels. This Type II site was treated for the first time in 2008 and the remaining infestation will be treated in 2013 using backpack sprayers. Since the infestation at this site is almost exclusively within the fringe of S. foliosa, identification of cryptic hybrids has been challeng- ing and has extended the length of time needed to eradicate the infestation. Seal Slough. This Type I site is located in the City of San Mateo on its eastern border with Foster City. The site begins 200m upstream of the bridge crossing of J. Hart Clinton Drive spanning the slough channel, at tide gates that restrict water exchange and transform the upstream slough into the sinuous, 6km-long Marina Lagoon that is lined with residential properties. The portion of the site downstream of the tide gates is characterized by large mudflats that have accreted in the absence of scour from the full volume of the slough. On the downstream side of the bridge to the north, the mouth of the waterway opens to a 300m-wide cove bordered by a 70-acre tidal marsh to the east and the large hillside of Shoreline Park to the west above a heavily armored bank. The marsh contains small channels, mudflats, pans, mid-marsh pickleweed (Sarcocornia pacifica) and some gumplant (Grindelia stricta) stands, sand/shell beach berms along most of the bayfront, and PG&E power line towers anchored in the western marsh edge at the mouth. In 2006, CalTrans began a mitigation project by excavating a somewhat sinuous channel to the bay in the southeastern corner of the marsh, and the fresh substrate along the banks was quickly infested with hybrid Spartina from the neighboring site. There is an upland berm running north/south in the middle of the site that serves to divide the older marsh from this mitigation site. The central portion of the marsh at the mouth of Seal Slough was not permitted for treatment in 2011, and subsequently was allowed to rebound and expand. Airboat will replace amphibious tracked vehicle as the primary treatment method at the site, while crews with backpack sprayers will complete the rest of the work in 2013.

10. APPLICABLE WATER QUALITY BMPS The following mitigations were identified in the Spartina Control Program’s Program- matic Environmental Impact Report/Statement (PEIR/S). These mitigations will be implemented at all herbicide treatment sites and verified by ISP staff. IMPACT WQ-1: Degradation of Water Quality Due to Herbicide Application

San Francisco Estuary Invasive Spartina Project 62. 2013 Aquatic Pesticide Application Plan MITIGATION WQ-1: Herbicides shall be applied directly to plants and at low or receding tide to minimize the potential application of herbicide directly onto the water surface, as well as to ensure proper dry times before tidal inundation. Herbicides shall be applied by a certified applicator and in accordance with application guidelines and the manufacturer label. The Control Program shall obtain coverage under the State NPDES Permit for the Use of Aquatic Herbicides and any necessary local permits. A monitoring program shall be implemented as part of the NPDES permit, and shall include appropriate toxicological studies to determine toxicity levels of the herbicide solutions being used. The Control Pro- gram shall use adaptive management strategies to refine herbicide application methods to increase control effectiveness and reduce impacts. IMPACT WQ-2: Herbicide Spills MITIGATION WQ-2: Herbicides shall be applied by or under the direct supervision of trained, certified or licensed applicators. Storage of herbicides and adjuvants/surfactants on-site shall be allowed only in accordance with an approved spill prevention and con- tainment plan; on-site mixing and filling operations shall be confined to areas appropriately bermed or otherwise protected to minimize spread or dispersion of spilled herbicide or surfactants into surface waters. When containers of herbicide larger than the standard 2.5 gallon are used (such as the 15 gallon containers that may be used by the helicopter contractors for aerial application), these containers must remain in the staging area(s) on a levee or other appropriate upland site. These larger containers will not be allowed into the marsh, and a spill response plan must be in place in the event of an accidental discharge, to ensure that herbicide does not reach the marsh or surface waters. IMPACT WQ-3: Fuel or Petroleum Spills MITIGATION WQ-3: Fueling operations or storage of petroleum products shall be maintained off-site, and a spill prevention and management plan shall be developed and implemented to contain and clean up spills. Transport vessels and vehicles, and other equipment (e.g., mowers, pumps, etc.) shall not be serviced or fueled in the field except under emergency conditions; hand-held gas-powered equipment shall be fueled in the field using precautions to minimize or avoid fuel spills within the marsh. Other, specific best management practices shall be specified as appropriate in project-specific Waste Discharge Requirements. In addition to these water quality mitigation measures, each partner agency and its contractors are required to have an acceptable Site Safety and Materials Handling Plan (Ap- pendix 5).

San Francisco Estuary Invasive Spartina Project 63. 2013 Aquatic Pesticide Application Plan 11. REFERENCES Adam, P. 1990. Saltmarsh Ecology. Cambridge University Press, Cambridge, UK. Anttila, C.K, R.A. King, C. Ferris, D.R. Ayres, and D.R. Strong. 2000. Reciprocal hybrid formation of Spartina in San Francisco Bay. Molecular Ecology 9: 765-770. Ayers, D.R., D. Garcia-Rossi, H.G. Davis, and D.R. Strong. 1999. Extent and degree of hybridization between exotic (Spartina alterniflora) and native (S. foliosa) cordgrass (Poaceae) in California, USA determined by random amplified polymorphic DNA (RAPDs). Mo- lecular Ecology Volume 8, 1179-1186. Bascand, L.D. 1970. The roles of Spartina species in New Zealand. New Zealand Ecological So- ciety Proceedings 17: 22-40. BASF Corporation. 2003. Habitat® Herbicide, Specimen, EPA Reg. No. 241-426, 2003. BASF Corporation. 2004. Habitat® Herbicide for Aquatic and Invasive Vegetation Con- trol, 2004. Boyer, K.E., J.C. Callaway, and J.B. Zedler. 2000. Evaluating the progress of restored cordgrass (Spartina foliosa) marshes: belowground biomass and tissue nitrogen. Estuaries 23: 711- 721. Bradley, P.M. and E.L. Dunn. 1989. Effects of sulfide on the growth of three salt marsh halo- phytes of the southeastern United States. American Journal of Botany 76: 1707-1713. Callaway, J.C. 1990. The introduction of Spartina alterniflora in South San Francisco Bay. M.A. thesis, San Francisco State University. San Francisco, CA. 50 pages. Chapman, V.J. 1977. Coastal Vegetation. Chapman and Hall, London. Cornish, P.S. and S. Burgin. 2005. Residual effects of glyphosate herbicide in ecological restoration. Restoration Ecology 13:695-702. Daehler, C.C., and D.R. Strong. 1997. Hybridization between introduced smooth cordgrass (Spartina alterniflora; Poaceae) and native California cordgrass (S. foliosa) in San Fran- cisco Bay. American Journal of Botany 84(5): 607-611. Daehler, C.C., C.K. Antilla, D.R. Ayres, D.R. Strong, J.P. Baily. 1999. Evolution of a new ecotype of Spartina alterniflora in San Francisco Bay. American Journal of Botany Volume 86, 543-544. Dame, R., M. Alber, D. Allen, M. Mallin, C. Montague, A. Lewitus, A. Chalmers, R. Gardner, C. Gilman, Bjorn Kjerfve, J. Pinckney, and N. Smith. 2000. Estuaries of the south Atlantic Coast of North America: their geographic signatures. Estuaries 23: 793-819. Dow AgroSciences LLC. 2001. Rodeo®, Specimen Label, EPA Reg. No. 62719-324, revised April 17, 2001. Ebasco. 1993a. Element F. Chemical Methods Only: Human Health Effects of Glyphosate. Final Report. Prepared for Washington State Department of Ecology by Ebasco Environmental. January 1993. Ebasco Environmental. 1993b. Prepared for Washington State Department of Ecology. Element I: Integrated weed management alternative for managing noxious emergent plants. Ebasco Environmental, a Division of Ebasco Services Incorporated. 40 pp. Ebasco. 1993c. _Final Report, Element C: Efficacy and Impacts. Prepared for Washington State Department of Ecology. Ebasco Environmental, a Division of Ebasco Services Incorpo- rated.

San Francisco Estuary Invasive Spartina Project 64. 2013 Aquatic Pesticide Application Plan Entrix, Inc. 2003. Ecological Risk Assessment of the Proposed Use of the Herbicide Imazapyr to Control Invasive Cordgrass (Spartina spp.) in Estuarine Habitat of Washington State, prepared for Washington State Department of Agriculture, October 30, 2003. ExToxNet: a cooperative effort of University of California-Davis, Oregon State University, Michigan State University, Cornell University, and the University of Idaho, Pesticide In- formation Profile for Glyphosate; http://extoxnet.orst.edu/, accessed April 5, 2005. Fowlkes, M. D., Jerry L. Michael, Thomas L. Crisman, and Joseph P. Prenger. 2003. Effects of the Herbicide Imazapyr on Benthic Macroinvertebrates in a Logged Pond Cypress Dome, Environmental Toxicology and Chemistry, vol. 22, no. 4, pp. 900–907. Giesy, J.P., S. Dobson, and K.R. Solomon. 2000. Ecotoxicological Risk Assessment for Roundup Herbicide. In Review of Environmental Contamination and Toxicology. 167:35-120. Gleason, H.A. and A. Cronquist. 1991. Manual of vascular plants of Northeastern United States and adjacent Canada, Second Edition. New York Botanical Garden. Gleason, M.L., D.A. Elmer, and J.S. Fisher. 1979. Effects of stem density upon sediment retention by salt marsh cord grass, Spartina alterniflora Loisel. Estuaries 2: 271-273, Goals Project. 2000. Baylands Ecosystem Species and Community Profiles: Life histories and environmental requirements of key plants, fish and wildlife. Prepared by the San Francis- co Bay Area Wetlands Ecosystem Goals Project. P.R. Olofson, editor. San Francisco Bay Regional Water Quality Control Board, Oakland, California. Gray, A.J. D.F. Marshall, and A.F. Raybould. 1991. A century of evolution in Spartina anglica. Advances in Ecological Research Volume 21, 1-62. Grue, C., B. Smith, N. Kohn, and J. Davis. 2002. Effects of Rodeo and R-11, LI 700, Agri-dex, and Hasten on Embryogenesis in Pacific Oysters. Presented at Spartina Conference 2002, Olympia Washington Heydens, W.F. 1991. Rodeo herbicide use to control Spartina. Impact of glyphosate on marine and terrestrial organisms. Monsanto Agricultural Company, St. Louis, MO, USA. Kilbride, K.M., and F.L. Paveglio. 2001. Long-term fate of glyphosate associated with repeated rodeo applications to control smooth cordgrass (Spartina alterniflora) in Willapa Bay, Washington. In Archives of Environmental Contamination and Toxicology, 40. 179-183. King K., C. Curran, B. Smith, D. Boehm, K. Grange, S. McAvinchey, K. Sowle, K. Genther, R. Highley, A. Schaaf, C. Sykes, J. Grassley, and C. Grue. 2004. Toxicity of Rodeo® and Arsenal® Tank Mixes to Juvenile Rainbow Trout, Third International Conference on In- vasive Spartina, San Francisco, California, November 8-10, 2004. Kittleson, P.M. and M.J. Boyd. 1997. Mechanisms of expansion for an introduced species of cordgrass, Spartina densiflora, in Humboldt Bay, California. Estuaries 20: 770-778. Knutson, P.L. and W.W. Woodhouse Jr. 1983. Shore stabilization with salt marsh vegetation. Special Report No. 9, U.S. Army Corps of Engineers, Coastal Engineering Research Cen- ter, Fort Belvoir, Virginia. Knutson, P.L., H.H. Allen, and J.W. Webb. 1990. Guidelines for vegetative erosion control on wave-impacted coastal dredged material sites. U.S. Army Corps of Engineers, Waterways Experimental Station, Dredging Operations Technical Support Program, Technical Re- port D-90-13. Department of the Army, Waterways Experimental Station,Vicksburg, MS. Kroll, R.B. 1991. Field investigations of the environmental fate of Rodeo (glyphosate) in two tidal marshes. Technical Report 115. Maryland Department of the Environment, Baltimore, MD.

San Francisco Estuary Invasive Spartina Project 65. 2013 Aquatic Pesticide Application Plan Kubena, K.M., C.E. Grue, and T.H. DeWitt. 1997. Rounding up the facts about Rodeo: development of concentration-response relationships for selected non-target species. In Proceed- ings of the Second International Spartina Conference, March 20-21. Lee, W.G. and T.R. Partridge. 1983. Rates of spread of Spartina anglica and sediment accretion in the New River Estuary, Invercargill, New Zealand. New Zealand Journal of Botany 21: 231-236. Leson & Associates. 2005. Use of Imazapyr Herbicide to Control Invasive Cordgrass (Spartina spp.) in the San Francisco Estuary: Water Quality, Biological Resources, and Human Health and Safety, prepared for the San Francisco Estuary Invasive Spartina Project, May 4, 2005. Levin, D.A, J. Francisco-Ortega, and R.K. Jansen. 1996. Hybridization and the extinction of rare species. Conservation Biology Volume 10, 10-16. McKee, K.L. and W.H. Patrick. 1988. The relationship of smooth cordgrass (Spartina alterniflora) to tidal datums: a review. Estuaries 11: 143-151. Mendelssohn, I.A. and E.D. Seneca. 1980. The influence of soil drainage on the growth of salt marsh cordgrass, Spartina alterniflora, in North Carolina. Ecology 60: 574- 584Newcombe, C.L., J.H. Morris, P.L. Knutson, and C.S. Gorbics. 1979. Bank erosion control with vegetation, San Francisco Bay, California. Miscellaneous Report No. 79-2, U.S. Army Corps of Engineers, Coastal Engineering Research Center, Belvoir, Virginia. Miller P. and P. Westra. 2004. Herbicide Surfactants and Adjuvants, Colorado State University Cooperative Extension, Bulletin no. 0.559, August 23, 2004. Monsanto Company. 2000. Aquamaster®, Complete Directions for Use in Aquatic and other Noncrop Sites, EPA Reg. No. 524-343, 2000. Munz, P.A., and D.D. Keck, 1968. A California Flora and Supplement. University of California Press. Patten K. 2002. Smooth cordgrass (Spartina alterniflora) control with imazapyr, Weed Technolo- gy, vol. 16, pp. 826-832, 2002. Patten K. 2003. Persistence and non-target impact of imazapyr associated with smooth cordgrass control in an estuary, Journal of Aquatic Plant Management, vol. 41, pp. 1-6. Paveglio, F.L., K.M. Kilbride, C.E. Grue, C.A. Simenstad, and K.L. Fresh. 1996. Use of Rodeo® and X-77® spreader to control smooth cordgrass (Spartina alterniflora) in a southwestern Washington estuary. II. Environmental Fate. Environmental Toxicology and Chemis- try 15. Rhymer, J.M. and D.S. Simberloff. 1996. Extinction by hybridization and introgression. Annual Review of Ecology and Systematics 27: 83-109. San Francisco Estuary Invasive Spartina Project. 2012. 2012-2013 Site-Specific Spartina Control Plans, San Francisco Estuary Invasive Spartina Project. Prepared by SFEISP, Berkeley, California. www.spartina.org/project_documents/. June 2012. Schuette J. 1998. California Environmental Protection Agency, Department of Pesticide Regula- tion, Environmental Fate of Glyphosate, revised November 1998. SERA (Syracuse Environmental Research Associates, Inc.). 1997. Effects of Surfactants on the Toxicity of Glyphosate, with Specific Reference to Rodeo. Prepared for USDA Animal and Plant Health Inspection Services. SERA, Fayetteville, NY. SERA (Syracuse Environmental Research Associates, Inc.). 2004. Imazapyr - Human Health and Ecological Risk Assessment – Final Report, prepared for USDA, Forest Service, Decem- ber 18, 2004.

San Francisco Estuary Invasive Spartina Project 66. 2013 Aquatic Pesticide Application Plan Shaner D, S. O’Connor. 1991. The Imidazolinone Herbicides. CRC Press, Ann Arbor, MI. Smart, R.M. and J.W. Barko. 1978. Influence of sediment salinity and nutrients on the physiological ecology of selected salt marsh plants. Estuarine and Coastal Marine Science 7: 487- 495. Spicher, D.P. 1984. The ecology of a caespitose cordgrass (Spartina spp.) introduced to San Francisco Bay. M.A. Thesis, San Francisco State University, San Francisco, California. Sprankle, P., W.F. Meggitt and D. Penner. 1975. Rapid inactivation of glyphosate in soil. Weed Sci. 23:224-228. Tu M, C. Hurd, & J.M. Randall. 2001. Weed Control Methods Handbook: Tools and Techniques for Use in Natural Areas, April 2001. U.S. Environmental Protection Agency, Technical Overview of Ecological Risk Assessment, Analysis Phase: Ecological Effects Characterization, Ecotoxicity Categories for Terres- trial and Aquatic Organisms; http://www.epa.gov/oppefed1/ecorisk_ders/toera_analysis_eco.htm#Ecotox, accessed April 2, 2005. U.S. Fish and Wildlife Service and State Coastal Conservancy. (2003) Final Programmatic Envi- ronmental Impact Statement/Environmental Impact Report, San Francisco Estuary Inva- sive Spartina Project: Spartina Control Program. Volume 1: Final Programmatic Envi- ronmental Impact Statement/Environmental Impact Report. State Clearinghouse #2001042058. USFWS, Sacramento, California/State Coastal Conservancy, Oakland, California. www.spartina.org/project_documents/eis_final.htm. September 2003. U.S. Fish and Wildlife Service and State Coastal Conservancy. (2003) Final Programmatic Envi- ronmental Impact Statement/Environmental Impact Report, San Francisco Estuary Inva- sive Spartina Project: Spartina Control Program. Volume 2: Appendices. State Clearing- house #2001042058. USFWS, Sacramento, California/State Coastal Conservancy, Oak- land, California. www.spartina.org/project_documents/eis_final.htm. September 2003. Valiela, I., J.M. Teal, W.G. Deuser. 1978. The nature of growth forms in the salt marsh grass Spartina alterniflora. American Naturalist Volume 112 (985) 461-470. Wang, Y.S., C.G. Jaw, and Y.L. Chen. 1994. Accumulation of 2,4-D and Glyphosate in Fish and Water Hyacinth. Water, Air and Soil Pollution. 74(3/4):397-403.

San Francisco Estuary Invasive Spartina Project 67. 2013 Aquatic Pesticide Application Plan

Appendix I

Chemical properties, degradation rates, environmental fate, and toxicity of imazapyr, glyphosate, and aquatic surfactants evaluated for Spartina control

Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

San Francisco Estuary Invasive Spartina Project 2013 Aquatic Pesticide Application Plan Table A-1: Chemical description; degradation rates, products, and pathways; bioaccumulation ratings; and advantages and disadvantages of imazapyr and glyphosate herbicides for estuarine use

Imazapyr Glyphosate Trade Name Habitat® (Bayer Corporation) Rodeo® (Dow Chemical Company) (Company) Aquamaster® (Monsanto Corporation)

Registration No. 81334-34-1 1071-83-6 Formulation Aqueous solution of isopropylamine salt of imazapyr plus Aqueous solution of isopropylamine salt of glyphosate; acidifier; active ingredient: 28.7% isopropylamine salt of technical formulation contains 2,4-nitrosoglyphosate imazapyr; equivalent to 22.6% imazapyr (“NNG”) impurity; active ingredient: 53.8% glyphosate isopropylamine salt; equivalent to 48.0% glyphosate Chemical name IUPAC: (RS)-2-(4-isopropyl-4-methyl-5-oxo-2-imidazolin- IUPAC: N-(phosphonomethyl)glycine 2-yl)nicotinic acid CAS: N-(phosphonomethyl)glycine CAS: 2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo- 1H-imidazol-2-yl]-3-pyridinecarboxylic acid Chemical formula

Formula C13H15N3O3 C3H8NO5P Herbicide family Imidazolinone Organophosphorus Mode of action Systemic, broad-spectrum (non-selective); Systemic, broad-spectrum (non-selective); amino acid synthesis inhibitor, specifically, inhibits amino acid synthesis inhibitor; inhibits acetohydroxyacid synthase (“AHAS”) aka acetolactase 5-enolpyruvylshikimate-3-phosphate synthase, needed by synthase (“ALS”), the first enzyme in the synthesis of plants to synthesize chorismate, an intermediate branched-chain aliphatic amino acids (valine, leucine, and metabolic product in the synthesis of aromatic amino isoleucine) and as a result inhibits protein synthesis and acids cell growth Molecular weight 261.28 g/mole imazapyr 169.08 g/mole glyphosate 320.42 g/mole imazapyr isopropylamine salt 228.22 g/mole glyphosate isopropylamine salt Table A-1 contd.: Chemical description; degradation rates, products, and pathways; bioaccumulation ratings; and advantages and disadvantages of imazapyr and glyphosate herbicides for estuarine use

Imazapyr Glyphosate Specific gravity 1.04–1.07 0.5 Minimum 1 hour 6 hours drying time Highest proposed 1.5 lb a.e./acre 10.8 lb/acre application rate

Rate of kill Very slow Relatively slow Volatility Vapor pressure = 1.8×10-11 mm Hg Extremely low vapor pressure, thus, negligible risk of Henry’s Law constant of 7.1×10-17 atm m3/mole movement through volatility No volatilization from dry soil surfaces; low volatilization of imazapyr from water or moist soil surfaces. Solubility Water: 11,272 mg/L Water: ~12,000 mg/L

Soil organic carbon Koc = 8.81 Koc = 24,000 adsorption coefficient Very low Koc indicates low sorption potential. Very high Koc indicates tight sorption to most soils, suspended solids, and sediments in the environment.

Octanol/water Kow = 0.22, 1.3 Kow = 0.0003 partition coefficient Degradation Slow anaerobic microbial degradation. No degradation Primarily degraded by microbes and fungi in soil or pathways under anaerobic conditions. Rapid photolysis in water. water, under both aerobic and anaerobic conditions. Photodegradation in water and soil are not expected to contribute significantly to glyphosate degradation. Degradation Quinolinic acid Aminomethylphosphonic acid (“AMPA”); further products degraded to carbon dioxide and phosphate.

Half-life in soil t½ = 25–141 days Average t½ = 32 days, based on 47 agricultural and forestry studies. In most cases, >90% degraded within six months after application.

Half-life in t½ = <2 to 7 days t½ = >3 to 12 months benthic sediment

Table A-1, page 2/3 Table A-1 contd.: Chemical description; degradation rates, products, and pathways; bioaccumulation ratings; and advantages and disadvantages of imazapyr and glyphosate herbicides for estuarine use

Imazapyr Glyphosate

Half-life in water No detectable degradation due to hydrolysis up to t½ = 7–14 days 30 days, pH 5-7 Average t½ = 1-4 days (photolysis) Bioaccumulation BCF = 3; BCF in fish after 10-14 day exposure period = 0.2 to 0.3 Low potential for bioaccumulation Low potential for bioaccumulation in aquatic animals; poorly absorbed when ingested by terrestrial mammals; any absorbed glyphosate is rapidly eliminated resulting in minimal tissue retention. Advantages for — Rapid photolysis in water — Low leaching potential due to strong sorption to estuarine use — Shorter minimum drying time than glyphosate soil/sediment particles — No adsorption to particles — Formulation can be mixed with salt water — Aerial applications require an order of magnitude lower spray volumes than glyphosate — Application is more cost-effective than application of glyphosate — Does not require use of non-ionic surfactants Disadvantages for — Increased adverse effects to non-target emerged — Efficacy hindered by minimum drying time estuarine use vegetation due to higher efficacy on vascular plants — Inactivated by adsorption to sediment particles — Formulation requires mixing with freshwater, which is not readily available — Aerial applications require large spray volumes, which require frequent refilling of helicopter tanks — Application is expensive — Requires use of non-ionic surfactants

Table A-1, page 3/3 Table A-2: Chemical properties, environmental fate, general toxicity rating, and toxicity of adjuvants

Adjuvant Degradation General Toxicity (Manufacturer) Ingredients1 Chemical Properties Pathways Toxicity Rating (lowest reported) Non-ionic Surfactants (“NIS”)

R-11® (surface activator) 80% octylphenoxy — soluble in lipid and Slowly biodegraded by Mammals: practically 96-hr LC50, rainbow trout 3.8 ppm2 (Wilbur-Ellis Company) polyethoxyethanol, water progressive shortening of non-toxic orally, mild 96-hr LC50, bluegill sunfish 4.2 ppm2 20% butanol and — flammable ethoxylate chain; skin irritation possible 96-hr LC50, juvenile rainbow trout 6 ppm5 compounded silicone — specific gravity = 1.0 intermediate breakdown Fish: Moderately toxic 48-hr LC50, Daphnia spp. 19 ppm3 products of polyethylene Other aquatic biota: LD50 oral, rabbit >5,840 mg/kg2 glycol (anti-freeze) and slightly toxic LD50 dermal, rabbit >5,000 mg/kg2 short-chain ethoxylates

X-77® (spreader activator) Alkylarylpoly (oxy- — soluble in lipid and Slowly biodegraded by Mammals: practically 96-hr LC50, rainbow trout 4.2 ppm2 (Valent Corp.) ethylene) glycols, free fatty water progressive shortening of non-toxic orally 96-hr LC50, bluegill sunfish 4.3 ppm2 acids, isopropyl alcohol — flammable ethoxylate chain; Fish and other aquatic 48-hr LC50, Daphnia spp. 2 ppm2 intermediate breakdown biota: moderately toxic LD50 oral, rabbit >5,000 mg/kg2 products of polyethylene LD50 dermal, rabbit >5,000 mg/kg2 glycol (anti-freeze) and short-chain ethoxylates

Liberate® (penetrating Phosphatidylcholine — emulsifiable Biodegradation presumed Mammals: practically 96-hr LC50, rainbow trout 17.6 ppm1 surfactant, deposition and (lecithin), methyl esters of — specific gravity = 0.976 rapid due to natural non-toxic orally, NOEC, rainbow trout 12.5 ppm1 drift control agent) fatty acids, alcohol lecithin ingredients moderate skin irritation 48-hr LC50, Daphnia magna 9.3 ppm1 (Loveland Industries, Inc.) ethoxylate possible NOEC, Daphnia magna 7.5 ppm1 LD50 oral, rat >5,000 mg/kg1 LD50 dermal, rat >5,000 mg/kg1

LI-700® (wetting and Phosphatidylcholine — emulsifiable Biodegradation presumed Mammals: practically 96-hr LC50, rainbow trout 17 ppm2 penetrating surfactant) (lecithin), methylacetic — not flammable rapid due to natural non-toxic orally, causes 24-hr LC50, rainbow trout 22 ppm2 (Loveland Industries, Inc.) acid, alkyl polyoxyethylene — specific gravity = 1.03 lecithin ingredients skin and eye irritation 96-hr LC50, juv. rainbow trout 700 ppm5 ether Fish and other aquatic 96-hr LC50, bluegill sunfish 210 ppm2 biota: practically non- 48-hr LC50, Daphnia spp. 170 ppm3 toxic LD50 oral, rat >5,000 mg/kg2 LD50 dermal, rat >5,000 mg/kg2 Cygnet Plus 75% d-limonene and — flammable Mammals: causes skin NOEC, Ceriodaphnia dubia 3.0 ppm4 (Cygnet Enterprises) related isomers, — specific gravity = 0.87 and eye irritation; 96-hr LC50 Ceriodaphnia dubia 6.6 ppm4 15% methylated vegetable Fish: slightly toxic NOEC, rainbow trout 30 ppm4 oil, 10% alkyl hydroxypoly Other aquatic biota: 96-hr LC50, rainbow trout 45 ppm4 oxyethylene; manufactured moderately toxic NOEC, fathead minnow 15 ppm4 from natural limonene 96-hr LC50, fathead minnow ppm4 Esterified Seed Oils (“ESOs”) or Mehylated Seed Oils (“MSOs”)

Competitor® Ethyl oleate, sorbitan alkyl — soluble in water Fish: slightly toxic 96-hr LC50, rainbow trout 95 ppm3 Wilbur-Ellis Company) polyethoxylate ester, — combustible Other aquatic biota: 48-hr LC50, Daphnia spp. >100 ppm3 dialkyl polyoxy-ethylene — specific gravity = 0.9 practically non-toxic glycol Table A-2 contd.: Chemical properties, environmental fate, general toxicity rating, and toxicity of adjuvants

Adjuvant Degradation General Toxicity (Manufacturer) Ingredients1 Chemical Properties Pathways Toxicity Rating (lowest reported) Crop Oil Concentrates (“COC”)

Agri-Dex® (wetting and Proprietary; heavy range — dispersible in water as Biodegradation presumed Mammals: practically 96-hr LC50, rainbow trout 271 ppm2 penetrating agent) paraffin-based petroleum micelles rapid non-toxic through oral 24-hr LC50, rainbow trout 386 ppm2 (Helena Chemical oil with polyol fatty acid — moderately flammable ingestion, mild skin and 96-hr LC50, juv. rainbow trout 271 ppm5 Company) esters and eye irritant; Fish and 48-hr LC50, Daphnia spp. >1,000 ppm3 polyethoxylyated other aquatic biota: LD50 oral, rat 5,010 mg/kg2 derivatives practically non-toxic LD50 dermal, rabbit >2,020 mg/kg2 Silicone-based Surfactants

Dyne-Amic® (activator, Organosilicone , Fish and other aquatic 96-hr LC50, rainbow trout 23.2 ppm3 spreader-sticker, wetting methylated vegetable oil biota: slightly toxic 48-hr LC50, Daphnia spp. 60 ppm3 and penetrating agent, buffer) (Helena Chemical Company)

Kinetic® (spreader-sticker, Organosilicone , Fish and other aquatic 96-hr LC50, rainbow trout 13.9 ppm3 wetting agent) polyoxypropylene- biota: slightly toxic 48-hr LC50, Daphnia spp. 60.7 ppm3 (Helena Chemical polyoxyethylene Company) copolymer Colorants

Blazon® Spray Pattern Proprietary; 30% non-ionic — pH = 7.0 Mammals: practically LD50 rat >5,000 mg/kg1 Indicator “Blue” polymeric colorant, — completely soluble in non-toxic orally; mild (Milliken Chemical) 70% water water skin irritant; not — specific gravity = 1.07 mutagenic — mildly acidic

1 Manufacturer specimen labels 2 Referenced in Entrix 10/03. 3 Erik Johansen, Washington State Department of Agriculture, Memorandum Re: Summary of Acute Toxicity Data for Five Spray Adjuvants, February 4, 2004. 4 Pacific Ecorisk, An Evaluation of the Acute Toxicity of “CYGNET PLUS” to Ceriodaphnia dubia (water flea), Oncorhynchus mykiss (rainbow trout), and Pimephales promelas (fathead minnow), December 10, 2004. 5 King et al. 2004.

Table A-2, page 2/2 Table A-3a: Imazapyr herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Estuary

Application Method Spray Volume Formulation Active Ingredient1 Surfactant2 Colorant 0.25% v/v NIS with 70% a.i.; High volume 0.52-0.75% solution 100 gal/acre 1-1.5 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 3 qt/100 gal handheld sprayer 4-6 pints/100 gal SBS according to label 0.25% v/v NIS with 70% a.i.; Low-volume directed 0.75-1.5% solution 20 gal/acre 0.3-0.6 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 3 qt/100 gal sprayer 1.2-2.4 pints/20 gal SBS according to label 0.25% v/v NIS with 70% a.i.; Broadcast sprayer/ 2.5-7.5% solution 10-30 gal/acre 0.5-1.5 lb a.e./acre ~1% v/v MSO, ESO, or VOC; 0.5-1.5 qt/acre Aerial application 6 pints/10-30 gal SBS according to label

1 Active ingredient in Habitat® is imazapyr isopropylamine salt; values expressed as imazapyr acid equivalent 2 NIS = non-ionic surfactant; MSO = methylated seed oil; ESO = esterified seed oil; VOC = vegetable oil concentrate, SBS = silicone-based surfactant, %v/v = percentage based on volume by volume

Table A-3b: Glyphosate herbicide mixture component concentrations and application rates for treatment of non-native Spartina in San Francisco Estuary

Application Method Spray Volume Formulation Active Ingredient1 Surfactant2* Colorant

High volume 1-2% solution 100 gal/acre 4-8 lb a.e./acre 0.5% v/v NIS with 50% a.i. 3 qt/100 gal handheld sprayer 1-2 gal/100 gal

Low-volume 1-8% solution 25-200 gal/acre 1.35-10.8 lbs a.e./acre 0.5% v/v NIS with 50% a.i. 3 qt/100 gal directed sprayer 1-8 gal/100 gal

Broadcast sprayer/ 7-40 gal/acre/ 4.5-7.5 pints/acre 2.25-3.75 lb a.e./acre 0.5% v/v NIS with 50% a.i. 0.5-1.5 qt/acre Aerial application 7-20 gal/acre

1 The active ingredient in Rodeo® and Aquamaster® is glyphosate isopropylamine salt; values are expressed as glyphosate acid equivalent 2 NIS = non-ionic surfactant, %v/v = percentage based on volume by volume Table A-4: Worst-case concentration of imazapyr herbicide dissolved in leading edge of incoming tide Assumptions

Worst-case occurs on the leading edge of lateral flow from overtopped channel through an herbicide-treated marsh Herbicide was uniformly sprayed across the entire marsh surface (but not in channels) at an application rate r = 15.6 mg a.e./sqft The herbicide applied on a unit area (1 sqft) is therefore mass m = 15.6 mg a.e. The herbicide dissolves completely in the incoming water A percentage, p, of the herbicide sticks to the vegetation canopy, and does not dissolve in the first one foot of flow depth Incoming tidal water overbanks channel and flows laterally across the surface of the marsh to a maximum distance D Water flow across marsh (after it leaves channel) has a uniform depth d = 1ft A percentage, s, of the active herbicide that was deposited onto the sediment surface dissolves into the water column The dissolved herbicide is instantly fully dissolved in the first unit volume that flows through No evaporation No rain or other input of fresh water

Application rate Habitat® label application rate: 4-6 pints per acre Label indicates 2 pounds imazapyr acid equivalents per gallon Habitat® 6 pints/acre 1.5 lb a.e./acre 2 = 0.75 gal/acre = 15.61 mg a.e./ft

Variables (p, D, and s can be varied): 2 r = 15.61 mg a.e./ft Herbicide application rate 2 m = 15.61 mg a.e. Initial mass of herbicide per unit area (per 1 ft ) p = 0% Percentage of applied herbicide that is absorbed into vegetation canopy d = 1 ft Depth of water flow across marsh (1 ft allows unit volume calculations) a D = 100 ft Distance of lateral flow across the marsh surface s = 60% Percentage of herbicide reaching the sediment that resuspends into water column 3 C = ? Concentration of herbicide in water column (mg a.e./ft )

Equationb C = m × (1-p) × D × s = (mass per unit area) × (1-percent absorbed by plant canopy) × (percent dissolved in water column) × (number of units through which water flows)

3 Computed Concentration C = m 1-p D s = 937 mg/ft 15.61 100% 100 60% 33.1 mg/liter

Notes a) Most Spartina infested marshes in the San Francisco Estuary that will become inundated by tidal water in the days following imazapyr application have a multitude of channels throughout the marsh that will transport water directly from the San Francisco Bay before overbanking and causing lateral flow across the marsh. In these marshes there would be a maximum of 100 feet of lateral flow through sprayed marsh before meeting with another flow. b) Calculation does not take into account potential decay during period of time between spraying and water inundation nor any decay that might occur in water column once the herbicide is resuspended from sediment. Table A-5: Ecotoxicity categories for acute toxicity of pesticides to wildlife1

Mammals Birds Acute Oral or Acute Inhalation Acute Oral Acute Inhalation Toxicity Category Dermal LD50 (mg/kg) LC50 (ppm) LD50 (mg/kg) LC50 (ppm) Very highly toxic <10 <50 <10 <50 Highly toxic 10-50 51-500 10-50 50-500 Moderately toxic 51-500 501-1000 51-500 501-1,000 Slightly toxic 501-2,000 1001-5000 501-2,000 1,001-5,000 Practically non-toxic >2,000 >5,000 >2,000 >5,000

Table A-6: Ecotoxicity categories for acute toxicity of pesticides to aquatic organisms1

Fish or Aquatic Invertebrates Toxicity Category Acute Concentration LC50 (mg/L) Very highly toxic <0.1 Highly toxic 0.1-1 Moderately toxic >1-10 Slightly toxic >10-100 Practically non-toxic >100

Table A-7: Ecotoxicity categories for acute toxicity of pesticides to insects1

Concentration Toxicity Category (μg/bee) Highly toxic <2 Moderately toxic 2 - 11 Practically non-toxic >11

1 U.S. EPA, Technical Overview of Ecological Risk Assessment, Analysis Phase: Ecological Effects Characterization, September 28, 2004. Table A-8: Toxicity of imazapyr to mammals

Animal Administration Testing Facility 3 Test Substance Species Route Gender LD50 or ED50 Effect (Reporting Year) ʑ >5,000 mg/kg b.w. NOEL Rat oral ʐ >5,000 mg/kg b.w. NOEL American Cyanamid ʑ >2,000 mg/kg b.w. NOEL Company (1983)1 Rabbit dermal Imazapyr technical ʐ >2,000 mg/kg b.w. NOEL ʑ >1 ppm ND Food and Drug Research Rat inhalatory >1 ppm ʐ ND Laboratories (1983)1 (analytical)

AC 243,997 (93% pure) Rat inhalation ʑ+ʐ >1.3 ppm L Voss et al. (1983)2

ʑ >10,000 ppm diet DA oral ʐ >10,000 ppm diet DA DA, B, A, S, ʑ 4,200 mg/kg b.w. CY, C, DBW intraperitoneal DA, B, A, S, Rat ʐ 3,700 mg/kg b.w. CY, C, DBW ʑ >5,000 mg/kg b.w. DA subcutaneous Imazapyr ʐ >5,000 mg/kg b.w. DA Medical Scientific isopropylamine ʑ >2,000 mg/kg b.w. NOEL dermal Research, Laboratory technical ʐ >2,000 mg/kg b.w. NOEL (1983)1 (49.3% a.i.) ʑ >10,000 mg/kg b.w. DA oral ʐ >10,000 mg/kg b.w. DA DA, B, A, S, ʑ 3,450 mg/kg b.w. CY, C, DBW Mouse intraperitoneal DA, B, A, S, ʐ 3,000 mg/kg b.w. CY, C, DBW ʑ >5,000 mg/kg b.w. DA, B, S subcutaneous ʐ >5,000 mg/kg b.w. DA, B, S Table A-8 contd.: Toxicity of imazapyr to mammals

Animal Administration Testing Facility 3 Test Substance Species Route Gender LD50 or ED50 Effect (Reporting Year) ʑ >5,000 mg/kg b.w. DA American Cyanamid Rat oral ʐ >5,000 mg/kg b.w. DA Company (1983)1 ʑ >5,000 mg/kg b.w. DA American Cyanamid Mouse oral Imazapyr ʐ >5,000 mg/kg b.w. DA Company (1986)1 isopropylamine ʑ >2,148 mg/kg b.w. NOEL American Cyanamid Rabbit dermal (25% a.i.) ʐ >2,148 mg/kg b.w. NOEL Company (1983)1 ʑ >0.2 NOEL Food and Drug Research Rat inhalatory ʐ >0.2 (analytical) NOEL Laboratories (1983)1 Hershman & Moore Arsenal® 4-AS Rat inhalatory ʑ+ʐ >4.62 ppm L (1986)2

Chopper®RTU (NOS) Rat inhalatory ʑ+ʐ >3.34 ppm L Werley (1987)2

1 cited in Entrix 10/03. 2 cited in SERA 12/04, Appendix 1 3 Acronyms: A = ataxia (loss of ability to coordinate muscular movement); B = blepharoptosis (drooping of upper eyelid); b.w. = body weight; C = convulsion; CY = cyanosis (bluish discoloration of skin and mucous membranes resulting from inadequate oxygenation of blood); DA = decreased activity; DBW = decreased body weight; ED50 = dose causing 50% inhibition of a process; L = lethality; LD50 = lethal dose, 50% kill; ND = nasal discharge; NOEL = no-observable-effect level (no toxic signs); NOS = not otherwise specified; S = sedation Table A-9: Toxicity of imazapyr to birds

Test Test Substance Species (Observed Effect) Result* >1890 mg/kg diet LD , 18-weeks dietary 50 ~200 mg/kg b.w. 1890 mg/kg HDT NOEL, 18-weeks dietary ~200 mg/kg b.w. Northern bobwhite quail >5000 mg/kg diet LD , 5-day acute dietary 50 ~674 mg/kg b.w. 5000 mg/kg HDT Arsenal® NOEL, 5-day acute dietary ~674 mg/kg b.w. (identical with >1890 mg/kg diet Habitat®) LD , 18-weeks dietary 50 ~200 mg/kg b.w. 1890 mg/kg diet NOEL, 18-weeks dietary ~200 mg/kg b.w. Mallard duck >5000 mg/kg diet LD , 5-day acute dietary 50 ~674 mg/kg b.w. 5000 mg/kg HDT NOEL, 5-day acute dietary ~674 mg/kg b.w.

* Fletcher 1983a, 1983b, Fletcher et al. 1984a, 1984b, 1984c, 1984d, 1995a, 1995b; all in SERA 12/04, Appendix 3 Table A-10: Toxicity of imazapyr and imazapyr herbicide/surfactant mixtures to fish

Test Substance + Surfactant Animal Species Test Result Reference Arsenal® Herbicide 96-hr LC 113 ppm surfactant (28.7% imazapyr) + Hasten 50 Smith et al. 20021 Arsenal® Herbicide 96-hr LC50 479 ppm surfactant (28.7% imazapyr) + Agri-Dex® Rainbow trout, juvenile Arsenal® Herbicide (Oncorhynchus mykiss) 77,716 ppm of concentrate Grue 20031 96-hr LC50 (28.7% imazapyr) 22,305 mg imazapyr a.e./L King et al. 2004 ® 43,947 ppm of concentrate Arsenal Concentrate 1 96-hr LC50 Grue 2003 (53.1 a.i. imazapyr) 23,336 mg imazapyr a.e./L AC 243,997 with Cohle & McAllister 96-hr LC >1000 mg/L isopropylamine in water 50 1984a2 Arsenal® Herbicide Bluegill sunfish Cohle & McAllister 96-hr LC 180 mg/L (22.6% purity) (Lepomis macrochirus) 50 1984b2 AC 243,997 Kintner & Forbis 96-hr LC >100 mg/L (99.5% purity) 50 1983a2 Rainbow trout (Salmo gairdneri) Channel catfish Peoples 19842 Imazapyr NOS 96-hr LC >100 mg/L (Ictaluras punctatis) 50 Gagne et al. 19942 Bluegill sunfish (Lepomis macrochirus) Arsenal® Herbicide Cohle & McAllister 96-hr LC 110 mg/L (22.6% purity) 50 1984c2 Rainbow trout Arsenal® Herbicide (Salmo gairdneri) 96-hr LC >110 mg a.e./L Drotter et al. 19952 (21.5% purity) 50 Table A-10 contd.: Toxicity of imazapyr and imazapyr herbicide/surfactant mixtures to fish

Test Substance + Surfactant Animal Species Test Result Reference NOEC 120 mg a.i./L AC 342,997 LOEC >120 mg/L Drotter et al. 19982 (purity NOS) MATC >120 mg/L Fathead minnow 28-day (Pimephales promelas) >118 mg a.i./L AC 342,997 NOEC >118 mg a.i./L Drotter et al. 19992 (99.6% purity) LOEC >118 mg a.i./L MATC Atlantic silverside AC 243,997 (marine) 96-hr LC 184 mg/L Manning 1989a2 (99.5% purity) 50 (Menidia menidia)

24-hr LC50 4,670 μg/L Nile tilapia 48-hr LC50 4,630 μg/L (Tilapia nilotica) 72-hr LC50 4,610 μg/L Supamataya et al. Imazapyr NOS 2 96-hr LC50 4,360 μg/L 1981 Silver barb 24-hr LC50 2,706 μg/L (Barbus genionotus) 96-hr LC50 2,706 μg/L

1 cited in Entrix 10/03 2 cited in SERA 12/04

Abbreviations: LC50 = lethal concentration, 50% kill; LOEC = lowest-observable-effect concentration; MATC = maximum allowable toxicant concentration; NOEC = no-observable-effect concentration (no toxic signs); NOS = not otherwise specified

Table A-10, page 2/2 Table A-11: Toxicity of imazapyr and imazapyr/surfactant mixtures to aquatic invertebrates

Test Substance Species Test Result Reference (observed effect) Arsenal® Applicator’s Freshwater benthic In-situ microcosm Concentrate >18.4 mg/L (HDT) Fowlkes et al. 2003 macroinvertebrates NOEC, (D, BM) (479 g imazapyr a.e./L) Arsenal®Herbicide NOEC 180 mg/L Forbis et al. 19842 (22.6% purity) 48-hr LC50 350 mg/L Arsenal® Freshwater water flea 48-hr LC 79.1 mg imazapyr a.e./L + unidentified surfactant (Daphnia magna) 50 Cyanamid 19971 NOEC 40.7 mg imazapyr a.e./L

48-hr EC50 (?) 373 mg imazapyr a.e./L Eastern oyster EC50 (G) >132 mg imazapyr/L Arsenal® (Crassostrea virginica) NOEC >132 mg imazapyr/L (HDT) Mangels & Ritter 20001 Pink shrimp >132 mg imazapyr/L EC50 (S) (Penaeus duorarum) >132 mg imazapyr/L (HDT) Freshwater water flea AC 243,997 24-hr LC >100 mg imazapyr a.e./L (Daphnia magna) 50 Kintner & Forbis 19832 (technical) 48-hr LC >100 mg imazapyr a.e./L (<24 hours old) 50 AC 243,997 Freshwater water flea 7, 14, 21-day NOEC 97.1 mg/L (HDT, MATC) Manning 19892 (99.5% a.i.) (Daphnia magna) (S/R/G) Grass shrimp AC 243,997 BCF <1 (not calculable) Drotter et al. 19962 (Paleomonetes pugio) (purity NOS) BCF <1 (not calculable) Drotter et al. 19962 AC 243,997 Eastern oyster EC (G) >132 mg/L Drotter et al. 19972 (99.6% purity) (Crassosstrea virginica) 50 AC 243,997 96-hr EC (G) >173 mg/L Ward 19892 (99.5% purity) 50

1 cited in Entrix 10/03 2 cited in SERA 12/04, Appendix 4

Abbreviations: BM = biomass, D = deformity, S = survival; R = reproduction; G = growth; HDT = highest dose tested; MATC = maximum allowable toxicant concentration Table A-12: Toxicity of imazapyr and imazapyr/surfactant mixtures to non-target aquatic vegetation

Test Test Substance Species (Observed Effect) Result Reference 2 Green algae EC50 (G) 71 mg/L Hughes 1987 1 (Selenastrum capricornutum) EC25 (G) 78 mg/L Mangels & Ritter 2000 Freshwater diatom EC50 (G) >59 mg/L Mangels & Ritter 20001 (Navicula pelliculosa) EC25 (G) >59 mg/L Saltwater diatom EC (G) 85 mg/L 50 Hughes 19872 Technical grade (Skeletonema costatum) EC25 (G) 42.2 mg/L imazapyr Blue-green algae EC (G) 117 mg/L 50 Mangels & Ritter 20001 (Anabaena flos-aquae) EC25 (G) 7.3 mg/L Green algae EC (G) 0.2 mg/L Landstein et al. 19932 (Chlorella emersonii) 50 Duckweed EC (G) 0.024 mg/L 50 Hughes 19872 (Lemna gibba) EC25 (G) 0.013 mg/L

EC25 (G shoots) 0.013 mg/L EC50 (G shoots) 0.032 mg/L Common water milfoil EC (# roots) 0.022 mg/L 25 Roshon et al. 19992 (Myriophyllum sibiricum) EC50 (# roots) 0.029 mg/L ® Arsenal + EC25 (G roots) 0.0079 mg/L unidentified EC50 (G roots) 0.0099 mg/L surfactant Green algae EC (G) 14.1 mg/L 50 Mangels & Ritter 20001 (Selenastrum capricornutum) EC25 (G) 8.36 mg/L LC 24 ppb Duckweed 50 Mangels & Ritter 2000 EC (G) 0.0216 mg/L (Lemna gibba) 50 Mangels & Ritter 20001 EC25 (G) 0.0132 mg/L

1 cited in Entrix 10/03. 2 cited in SERA 12/04, Appendix 4. Abbreviations: S = survival; R = reproduction; G = growth; HDT = highest dose tested; MATC = maximum allowable toxicant concentration Table A-13: Toxicity endpoints for risk quotient calculation and levels of concern for interpretation of risk quotients

Aquatic Mammals Birds Aquatic Non-endangered Endangered animals vascular plants plants plants and algae Assessment Acute EC50 or LC50 LD50 oral LD50 oral EC50 EC25 seedling EC25 seedling acute toxicity emergence and emergence and vegetative vigor vegetative vigor or NOEC Chronic NOEC early- NOEC NOEC life stage or full 2-generation 21-week life-cycle tests reproduction reproduction Levels of concern (risk quotient greater than) Acute risk 0.50.5 0.5 1.0 1.0 1.0 Acute restricted use 0.1 0.2 0.2 Acute risk endangered species 0.05 0.1 0.1 Chronic risk 1.0 1.0 1.0

U.S. Environmental Protection Agency, Technical Overview of Ecological Risk Assessment, Analysis Phase: Ecological Effects Characterization and Risk Characterization, September 28th, 2004.

Appendix II

2013 Field Data Collection Form

Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

San Francisco Estuary Invasive Spartina Project 2013 Aquatic Pesticide Application Plan Page ___ of ___ on this date Field Data Collection Form San Francisco Estuary Invasive Spartina Project, Aquatic Pesticide Application Plan, 2612-A 8th St, Berkeley, CA, 94710

Site ID (XXXX) (eg. DEAD): ______Date: ______Collected By: ______

Station Location (circle): at application point upstream downstream Station Type (circle): Reference Treated

Wind (circle): low high Tidal Cycle (circle): high low slack Water Color (circle): green green-brown brown blue (dye)

Herbicide: imazapyr Surfactant (circle): Liberate Competitor Gallons tank mix applied ______Application Time (Start/Finish): ______/______

Field Measurements Water Depth pH Dissolved Water Temp Conductivity Salinity Meter Oxygen Used Meters mg/L O C mS ppt

YSI 85

Samples Collected Sample ID (XXXX-YY-Ab)* Time Sample Depth (m) Notes

* XXXX-YY-Ab (eg. DEAD-01-pre, ZIPY-02-0.5h) = XXXX: Site, YY: site visit number (01-1st, 02-2nd, 03-3rd, etc.), A: time to application (either pre, increments thereafter in half hours – 0.5), b: time increment (h=hour, w=week (for 1 week post-treatment))

Additional Notes or Comments: ______

Wildlife presence: ______

Appendix III

Chain of Custody for Pacific Agricultural Laboratory

Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

San Francisco Estuary Invasive Spartina Project 2013 Aquatic Pesticide Application Plan Analytical Request/Chain of Custody Page _____ of ______Pacific Agricultural Laboratory, LLC PAL Project # ______12505 N.W. Cornell Rd., Ste. 4 • Portland, OR 97229 Tel 503.626.7943 • Fax 503.641.0644 • www.pacaglab.com Pacific Agricultural Laboratory

Name ______Requested Analysis Contact ______Address______City ______State ______Zip ______

Telephone ______Fax ______Requested Turnaround Time Email ______Cell Phone ______

CLIENT INFO Standard (10 working days) Project # ______Purchase Order # ______q Method of Shipment ______q Rush ______please specify

Sample Sample Sample Container No. of PAL ID Client Sample ID Date Time Type Type Containers Comments

DATE TIME DATE TIME Relinquished by: ______Received by ______SIGNATURE SIGNATURE Relinquished by: ______Received by ______SIGNATURE SIGNATURE Lab Comments: ______

All services performed by PAL are subject to the Standard Terms and Conditions on reverse side of this form. Pacific Agricultural Laboratory, LLC Standard Terms and Conditions

Unless otherwise agreed in a separate contract, services provided by Pacific Agricultural Hazardous Materials/Samples: Unused portions of samples found or suspected to Laboratory, LLC (PAL) are expressly limited to the terms and conditions stated herein. be hazardous according to state or federal laws will either be returned to the client at Submission of samples is deemed acceptance of these terms and conditions. the expense of the client, or properly disposed of as hazardous waste at the expense of the client. Samples are the property of the client at all times, even while in the Limits of Liability: All analytical services provided by PAL are made on a best possession or under the control of PAL for analysis. All other samples will be properly effort basis. Established methods of analysis will be followed whenever possible; and anonymously disposed of as nonhazardous waste after expiration of sample retention however every sample has unique properties that may require deviation or adaptation of time. Samples subject to USDA foreign soil or plant permits shall be disposed of in established methodologies. The total liability of PAL will be limited to retesting or refund accordance with applicable permit conditions. of the paid for services provided, at the option of PAL. Use of PAL constitutes acceptance of these limitations of liability. Sample Retention/Disposal: Samples submitted for analysis are retained and stored under proper conditions and security for a period of time after the issuance of the final Confidentiality: Pacific Agricultural Laboratory is zealously protective of customer report. Retention times are generally as follows: information and uses its best business practices to maintain strict and absolute customer confidentiality. Confidentiality can be obtained, to the fullest extent allowed by the law, by placing written, mutual confidentiality agreements into force, upon request. Sample Type Retention Time Information is released to third parties only upon the authorization of the customer, Surface Water, Wastewater 30 days by court order, or otherwise as required by law, and taking precautions to ensure Nonhazardous Soil and Sludge 60 days confidential transfer of information between PAL and its customers by courier or mail, Food Commodities, Plant Tissue 60 days facsimile, email (internet), and/or telephone as the customer may direct. PAL shall not Other Nonhazardous Materials 60 days be responsible for any disclosure of any information of customer unless PAL specifically Hazardous Materials 30 days agrees to keep such information confidential by separate agreement.

Billing: All fees are billed directly to the client. Third party billing will not be accepted Report and Document Retention: PAL shall retain final reports and all supporting without prior arrangement and consent by PAL and agreement of the third party to all documentation and analytical data used to generate reports for eight years following the standard terms and conditions. generation of the report, after which time PAL shall be free to destroy the information.

Payment Terms: For clients without approved credit from PAL, payment must be Sample Containers and Shipping Materials: PAL will provide appropriate sample received prior to the release of final project report. For clients with approved credit, terms containers, shipping containers and packing materials at no additional charge upon prior are net 30 days from the date of invoice unless otherwise stated on that invoice. Any arrangement. Standard shipping will be UPS Ground, and the client will be charged for changes from these terms must be agreed upon prior to sample submission. A service expedited container shipping and the shipping of samples to PAL for analysis. charge of 1 _% per month (18% annual rate) will apply for outstanding balances that are past due. In the event of default of payment for analytical services rendered, the client Analytical Service Requests: Requests for analytical Services can be made by is responsible for reasonable collection charges including court costs and attorney’s fees telephone, fax, email or in writing. The client must confirm requests for service in incurred by PAL. There is an additional $25.00 charge for any returned checks. writing, using a PAL chain of custody form prior to the commencement of work by PAL, and following directions for sampling as provided by PAL. It is very important that the Litigation: All fees and costs associated with compliance by PAL to any subpoena analytical services to be provided by PAL be clearly understood by both PAL and the for documents, for consultation in preparation for or testimony in any deposition or client prior to commencement of projects. PAL will not be responsible for delays caused proceeding, or for any other purpose relating to the work of PAL, in connection with by incomplete information provided by the client including missed hold times and work performed for a client, shall be paid by the client. Such costs include, but are delayed report generation. not limited to, fees for persons responsible for responding to subpoenas, reproduction of reports and data in support of reports, mileage and other travel expenses, attorney preparations for testimony, court testimony, attorney fees, and any other expenses associated with PAL’s participation in the litigation.

Appendix IV

Quality Assurance Plan (QAP) Pacific Agricultural Laboratory (PAL)

Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

San Francisco Estuary Invasive Spartina Project 2013 Aquatic Pesticide Application Plan SOP-AM-033 Pacific Agricultural Laboratory Extraction of Imidazolinone Herbicides in Water

1.0 Scope and Application

1.1 This procedure describes the extraction of imidazolinone herbicides from aqueous samples. This method is applicable to all types of water including, but not limited to, drinking water, storm water, surface water, and groundwater.

2.0 Summary of Method

2.1 A 500mL aliquot of sample is acidified to pH 2 and 12.5g sodium chloride is added. Sample is shaken in a separatory funnel with three 50mL portions of dichloromethane. Organic layers are drained [through acidified sodium sulfate] into a round bottom flask, and concentrated by rotary evaporation (SOP-AM-027).

3.0 Interferences

3.1 Potential interferences may include contamination from glassware and solvents, and co-extracted materials from the sample matrix. Care must be taken to avoid and/or minimize these potential interferences.

4.0 Sample Handling and Preservation

4.1 Samples should be taken in 1-L amber glass bottles with a PTFE lined cap. 4.2 Samples are taken at neutral pH, and stored at 4°C prior to extraction. 4.3 All water samples shall be extracted within seven (7) days of sampling.

Revision Number: 1 1 Revision Date: 9/8/10 SOP-AM-033 Pacific Agricultural Laboratory 5.0 Apparatus and Instrumentation

5.1 1000 mL glass separatory funnel 5.2 500 mL graduated cylinder 5.3 600 mL beaker 5.4 250 mL round bottom flask 5.5 Large glass funnel 5.6 pH meter 5.7 Top-loading balance, accurate to ±0.01g 5.8 Magnetic stir bar 5.9 Magnetic stir plate 5.10 Rotary evaporator, Rotavap; Yamato RE50

6.0 Reagents and Supplies

6.1 Organic-free water, DI H2O 6.2 Methanol (MeOH) w/0.5% Formic Acid 6.3 Pesticide-grade Dichloromethane, DCM 6.4 6 N Hydrochloric Acid, HCl 6.5 Sodium chloride, ACS grade 6.6 Glass beads 6.7 [Glass wool]

6.8 [Acidified sodium sulfate, Na2SO4]

7.0 Procedure

7.1 For each sample, the necessary glassware items (separatory funnel, 600 mL beaker, and flat-bottom flask) are obtained, rinsed with Dichloromethane if necessary, and labeled with sample number. Beakers contain a magnetic stir bar, and two glass beads are added to each flat- bottom flask. Using a graduated cylinder, measure 500 mL of organic-

Revision Number: 1 2 Revision Date: 9/8/10 SOP-AM-033 Pacific Agricultural Laboratory free water for QC and transfer to a beaker with a stir bar. Likewise, measure and transfer 500 mL of sample into a beaker with a stir bar. 7.2 [Sodium sulfate funnels are prepared by placing a small plug of glass wool into a glass powder funnel, to which ~25g acidified Sodium Sulfate is added. Funnels are rinsed with ~10mL DCM, and solvent is drained into waste. A funnel is placed on each labeled collection flask.] 7.3 Using a 500 mL graduated cylinder, a 500 mL aliquot of sample is measured and transferred to the labeled 600 mL beaker. 7.4 Method Blank (BLK) consists of 500 mL deionized water in a 600mL beaker. This sample will be the negative control (QC) for the analysis. 7.5 Lab Control Sample/Lab Control Sample Duplicate (LCS/LCSD) each consist of 500 mL DI water in a 600mL beaker. Project specific spike compounds are added to each, and the standard log number and spike volume are recorded on extraction bench sheet. These samples will be the positive control (QC) for the analysis. 7.6 The pH of each sample and QC is adjusted to 2.0 by dropwise addition of 6N hydrochloric acid. 7.7 12.5 g of sodium chloride is added to each beaker, stirring until salt is completely dissolved. 7.8 The contents of each beaker are transferred into the appropriately labeled separatory funnel. Samples and QC are extracted by shaking three times with 50mL DCM. The lower (DCM) layers are drained [through the acidified sodium sulfate funnel] into the corresponding flat-bottom round flask.

7.9 [After all solvent is collected, Na2SO4 funnels are rinsed with ~20mL Dichloromethane, to optimize recovery of analytes.] 7.10 Extracts are concentrated to ~0.5 mL using rotary evaporation (SOP-AM- 027), and remaining solvent is evaporated to dryness under a steady stream of nitrogen gas.

Revision Number: 1 3 Revision Date: 9/8/10 SOP-AM-033 Pacific Agricultural Laboratory 7.11 Extract is transferred to labeled culture tubes as per SOP-AM-XXX (Rotavap) using MeOH w/0.5% Formic acid as final solvent. Final volume is 2mL for most Imidazolinone extractions. 7.12 Extracts should be stored in refrigerator until analysis.

8.0 Calculations

8.1 N/A

9.0 Quality Control

9.1 At a minimum, batch QC will include a method blank (MB), and a Laboratory Control Sample/Laboratory Control Sample Duplicate (LCS/LCSD). Additional QC will be performed if there are project and/or method specific requirements. An extraction batch consists of a batch of 20 consecutive samples extracted within 7 days. 9.2 Spike recoveries are calculated after analysis to evaluate extraction efficiency.

10.0 Reporting

10.1 N/A

11.0 References

11.1 American Cyanamid Method 2261 11.2 American Cyanamid Method M1900

Revision Number: 1 4 Revision Date: 9/8/10 SOP AM-034 Pacific Agricultural Laboratory

Imidazolinone Herbicides in Water by EPA 8321B

1.0 Scope and Application 1.1 This procedure is used to determine the concentrations of Imidazolinone herbicides in liquid matrices.

2.0 Summary of Method 2.1 A measured volume of sample is extracted using AM-033, Extraction of Imidazolinone Herbicides in Water. 2.2 Extracts are analyzed using liquid chromatography with mass spectroscopy (LC/MS) detection.

3.0 Interferences 3.1 Potential interferences may include contaminated solvents and extraction glassware, dirty chromatographic equipment, and co-extracted materials from the sample matrix. Care must be taken to avoid and/or minimize these interferences.

4.0 Sample Handling and Preservation 4.1 Store samples at 4oC out of direct sunlight. Water samples should be extracted within 7 days of sampling and analyzed within 40 days of extraction 4.2 Personal protection measures should be taken while handling solvents and samples.

Revision Number: 1 1 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

5.0 Apparatus and Instrumentation 5.1 Analytical balance, Sartorius model CP124S, accurate to 0.0001g. Calibration of balance shall be checked daily (SOP EQ-001). 5.2 N-EVAP evaporation manifold with heated water bath 5.3 HPLC System 5.3.1 Agilent 1100 HPLC system equipped with binary pump, autosampler, solvent degasser, and single quadrapole mass spectrometer. 5.3.2 Agilent Chemstation software 5.3.3 Analytical Column – C18 reverse phase column, 100mm x 3.0mm ID, 2.5 µm particle size, Agilent Zorbax SB-C18 or equivalent.

6.0 Reagents and Supplies 6.1 Organic-free reagent water 6.2 Methanol, Chemsolve, HPLC Grade 6.3 Acetonitrile (ACN), Chemsolve, HPLC Grade 6.4 Formic Acid, EMD, ACS Grade 6.5 Luer lock tipped syringe 6.6 Screw capped tubes with Teflon lined lids 6.7 13mm 45 µm nylon syringe filters 6.8 Auto sampler vials with PTFE lined caps 6.9 Volumetric flasks, class A 6.10 Gas tight syringes with PTFE tipped plungers 6.11 HPLC/MS Tuning Standard – Aglient ES Tuning Mix G2421A

Revision Number: 1 2 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

7.0 Procedures 7.1 Sample Extraction: 7.1.1 Extract waters via the procedure outlined in Pacific Agricultural Laboratory SOP AM-033 “Extraction of Imidazolinone Herbicides in Water”. 7.1.2 Store extracts in refrigerator until analysis. 7.2 Solvent exchange of water extracts: 7.2.1 Transfer a 1 ml aliquot of the sample extract to a culture tube. Mark the meniscus of the liquid in the tube. 7.2.2 Evaporate the solvent under a steady stream of nitrogen using the N-Evap evaporation manifold. 7.2.3 Reconstitute the extract as follows: add 500 uL methanol, then 500 uL Mobile Phase A (95% organic free water, 5% ACN, 0.05% formic acid). . 7.2.4 Filter the sample extract into an autosampler vial through a 45 µm 13 mm syringe filter using a luer tipped syringe. 7.2.5 Cap the vial and label with appropriate moniker. 7.3 Preparation of HPLC mobile phase: 7.3.1 The mobile phase is contained in two reservoirs, one containing the aqueous portion (Mobile Phase A) and one containing the organic( Mobile Phase B) portion. 7.3.2 Prepare Mobile Phase A by combining 950 mL of organic free water, 50 mL ACN, and 0.5 mL formic acid. 7.3.3 Prepare Mobile Phase B by combining 950 mL of ACN, 50 mL organic free water, and 0.5 mL formic acid. 7.4 Chromatographic conditions: 7.4.1 Flow rate: 0.40 mL/minute 7.4.2 Injection volume: 10 ul 7.4.3 Column Temperature: 45 oC

Revision Number: 1 3 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

7.4.4 Solvent Gradient: Time %A %B 0.0 80 20 1.5 80 20 8.0 30 70 10 30 70 7.4.5 Re-equilibration time: 3 minutes, 80% A/20% B

7.5 Mass Spectrometer Conditions: 7.5.1 Ionization Mode: API-Electrospray o 7.5.2 Drying Gas: N2, 11.0 L/min, 250 C 7.5.3 Nebulizer Pressure: 30 psig 7.5.4 Capillary Voltage: 1500 V 7.6 Mass Spectrometer Detector settings: 7.6.1 Settings for use in MS data acquisition (SIM ions and fragmentor voltages) vary by analyte and are displayed in Table 2 of the Appendix (12.2). 7.7 If the peak areas of the sample signals exceed the calibration range of the system, dilute the extract as necessary and reanalyze the diluted extract.

Revision Number: 1 4 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

7.8 Calibration: 7.8.1 Electrospray MS System: The MS system is calibrated for accurate mass assignment, sensitivity, and resolution using the Agilent ES Tuning Mix G2421A. The following masses are calibrated in positive and negative ionization modes: MASS POSITIVE NEGATIVE 1 118.09 112.99 2 322.05 431.98 3 622.03 601.98 4 922.01 1033.99 5 1521.97 1633.95

Tune parameters are adjusted to ensure ions are present at each of the masses with counts >50000 and peak widths within the range of 0.60 – 0.70 amu. 7.8.2 Stock Standards: Individual analyte stock standards are made at concentrations between 500-1000 µg/ml by transferring 25-50 mg neat standard to a 50 mL class A volumetric flask, dissolving the neat standard in acetonitrile or methanol, and diluting to the mark with acetonitrile or methanol. Stock standards prepared from neat standards may be used for a maximum of two years. Alternatively, a solution containing 1000 µg/ml of analyte may be obtained from ChemService or other reputable manufacturer and used as a stock standard. In this case, the stock standard may be used until the expiration date provided by the manufacturer. 7.9.3 Working Standards: A 10 µg/ml working standard is made by transferring appropriate amounts, depending on initial concentrations, of stock standards to a 10 mL class A volumetric flask and diluting to the mark with methanol or acetonitrile. The

Revision Number: 1 5 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

amount of stock standard to transfer will range between 100-200 µL and is calculated using the formula: Amt. Stock Std.(µL) = [Final Conc. (10µg/ml)] x [Final Vol. (10ml) Initial Stock Conc. (µg/µL) The working standard solution is transferred to an appropriately labeled screw cap tube and may be used for a maximum of one year. 7.9.4 Preparation of external standard calibration curve: an appropriate aliquot of the working standards are added to an autosampler vial and diluted to 1 ml with Mobile Phase A. A minimum of 5 standards are prepared at the following suggested levels: 0.005 ug/ml, 0.010 ug/ml, 0.020 ug/ml, 0.05 ug/ml, and 0.10 ug/ml. The calibration range can be adjusted to meet expected levels in the samples. The calibration standards are prepared as follows: Calibration Aliquot Concentration Volume of Final level volume of aliquot(s) buffer volume 100 ng/ml 100 µl 1000 ng/ml 900 µl 1.0 ml

50 ng/ml 50 µl 1000 ng/ml 950 µl 1.0 ml 20 ng/ml 200 µl 100 ng/ml 800 µl 1.0 ml 10 ng/ml 100 µl 100 ng/ml 900 µl 1.0 ml 5 ng/ml 50 µl 100 ng/ml 950 µl 1.0 ml

7.9.5 The system is calibrated prior to the injection of a set of sample extracts. After injecting a set of standards, a linear calibration curve is prepared. Exclude the origin as a point. The R value of the generated curve should be 0.99 of better. If the calibration fails to meet these criteria, the cause of the deviation should be rectified and the system recalibrated.

7.9.6 The calibration is verified by injecting a CCV at the mid point concentration of the calibration curve after no more than twenty Revision Number: 1 6 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

samples. If the response deviates by more than +/- 15% from the initial calibration, the system should be recalibrated and the samples bracketed by the either the initial calibration or the prior passing CCV and the failing CCV should be reanalyzed. If the CCV is >15% of initial calibration, the samples bracketed by the either the initial calibration or the prior passing CCV and the failing CCV can be used if the sample contains no detectable residues.

8.0 Calculations 8.1 Water Samples:

amount f/curve (ng/ml) x final volume (ml) x dilution factor = result (ug/liter, ppb) sample volume (ml)

9.0 Quality Control 9.1 Initial Demonstration of Proficiency – the laboratory shall demonstrate initial proficiency with each sample preparation technique, by generating data of acceptable accuracy and precision for target analytes in a clean matrix. The laboratory must also repeat the demonstration whenever new staff is trained or significant changes in instrumentation are made. 9.1.1 Calculate the average recovery and the standard deviation of the recoveries of the four QC reference samples. Refer to Section 8.0 of EPA Method 8000 for procedures in evaluating method performance. 9.2 Method Reporting Limits (MDLs) 9.2.1 The MDL is defined as the minimum concentration of a substance that can be measured and reported with 99% confidence that the analyte concentration is greater than zero and is determined from analysis of a sample in a given matrix type containing the analyte.

Revision Number: 1 7 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

9.2.2 The extraction and analysis of seven replicates of a spiked sample determine the MDL. 9.2.3 Multiply the standard deviation of the seven replicate results by the one sided 99%t-statistic (3.14) to obtain the MDL for each analyte. 9.2.4 These results are kept on file and should be re-evaluated annually, when significant changes in instrumentation are made, or when new staff are added. 9.3 Sample Quality Control for Preparation and Analysis 9.3.1 The laboratory will have procedures for documenting the effect of matrix on method performance. 9.3.2 Water matrix – minimum QC samples shall include a method blank (MB), Laboratory Control Sample (LCS), and a Laboratory Control Sample Duplicate (LCSD). A matrix spike may be prepared and analyzed provided there is adequate sample. 9.4 QC Frequency – an analytical batch is defined as a set of no more than 20 samples extracted within 14 days. The QC frequency for each analytical batch is as follows: Method blank – 5% Matrix Spike/Matrix Spike Duplicate – 5% Laboratory Control Sample/Laboratory Control Sample Duplicate – 5% 9.4.1 In house method performance criteria for spike and surrogate compounds should be developed using guidance found in Section 8.0 of EPA Method 8000. 9.4.2 If the recovery data is outside acceptance limits, the samples should be re-extracted and/or the data flagged as necessary.

Revision Number: 1 8 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

10.0 Reporting 10.1 If all QC criteria have been met, the data is then compiled and a report is generated, including sample raw analytical results and QC data, and submitted to the client.

11.0 References 11.1 EPA Method 8321B, SW-846 Revision 2, December 2007. 11.2 Pacific Agricultural Laboratory Quality System Manual. 11.3 EPA Method 8000B, SW-846 Revision 2, December 1996. 11.4 SW-846, Chapter One, Revision 1, 1992.

12 Figures and Appendices 12.1 Table 1 - Analyte list and reporting limits 12.2 Table 2 – Mass Spectrometer Data Acquisition Settings

Approved:

Date:

Revision Number: 1 9 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

TABLE 1 ANALYTE LIST AND LIMIT OF QUANTITATION (LOQ) Analyte LOQ, ug/L

Imazamox 0.02 Imazapic 0.02 Imazapyr 0.02 Imazethapyr 0.02

TABLE 2 – MASS SPECTROMETER DATA ACQUISITION SETTINGS

Time SIM Ions Fragmentor Voltage Capillary Voltage 0.00 220,222,234, 200 2000 V 248,262,277, 278,290,293, 306,307

Revision Number: 1 10 Revision Date: 9/8/10

SOP AM-034 Pacific Agricultural Laboratory

TABLE 3 – SIM IONS FOR IDENTIFICATION/QUANTIFICATION

Analyte Quantification Ion Qualifier Ions Ionization Fragmentor Mode Voltage Imazamox 306 307,278 positive 200 Imazapic 293 277,220 positive 200 Imazapyr 262 234,222 positive 200 Imazethapyr 290 262,248 positive 200

Revision Number: 1 11 Revision Date: 9/8/10

Appendix V

General Site Safety & Materials Handling Guidelines and Procedures for Spartina Control Projects in the San Francisco Estuary

Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

San Francisco Estuary Invasive Spartina Project 2013 Aquatic Pesticide Application Plan

General Site Safety & Materials Handling Guidelines and Procedures for Spartina Control Projects in the San Francisco Estuary

Invasive Spartina Control Plan ATTACHMENT 1

IMPORTANT DISCLAIMER AFFECTING LEGAL RIGHTS. The San Francisco Estuary Invasive Spartina Project (ISP) compiled this document to provide a suggested set of general guidelines and procedures that are consistent with the mitigation measures required by the ISP EIR [spelled out](“PEIR”). These guidelines may be used by ISP partners and contractors as a minimal baseline, consistent with the PEIR, for planning and implementing site-specific Spartina control work in the San Francisco Estuary. These general guidelines are not intended to cover all safety procedures and precautions that may be necessary, nor are they intended to substitute for a comprehensive set of safety procedures and precautions for any particular control work or site. The ISP and the Conservancy make no warranties, assur- ances or representations of any kind with respect to the scope, extent, propriety or effectiveness of the suggested procedures that are described in these guidelines. Each ISP partner is responsible for and should independently develop and implement all appropriate safety precautions and procedures needed for the control work it undertakes. As a condition to use of these guidelines, each ISP partner and any of its contractors agree that the ISP and the State Coastal Conservancy (“Conservancy”) shall not be responsible for the acts or omis- sions of any ISP partner, or its contractors or volunteers, and agree to release and hold harmless the ISP and the Conservancy from any claims or liability, in connection with the development or implementation of site-specific safety procedures and precautions.

May 2005 Updated June 2011 Table of Contents

Table of Contents...... i

Emergency Information...... 1 Emergency Phone Numbers...... 1 Nearest Hospital...... 1

Sensitive Receptors...... 2 Listing of Sensitive Receptors ...... 3

Site Safety Protocols for Working in Marshes, Mudflats, and other Tidal areas of the San Francisco Estuary ...... 4 Teams...... 4 Channels...... 4 Mudflats ...... 5 Chemical or Physical Hazards ...... 5 Tides...... 5 Weather...... 6 Communication...... 6

Herbicide Handling, Spill Prevention and Spill Response ...... 7 Herbicide Use...... 7 Herbicide Storage...... 7 Container Disposal...... 7 Spill Response...... 8 Preventing Spills ...... 9

Spray Drift Reduction ...... 11

Petroleum Fuel Spill Prevention and Response...... 13

Herbicide Information...... 14

Invasive Spartina Control Plans i General Site Safety & Materials Attachment 1 Handling Guidelines Emergency Information

Emergency Phone Numbers In case of any emergency, call 9-1-1, and follow dispatcher instructions

Pesticide Emergency: o Call the ChemTrec (Chemical Transportation Emergency Center) emergency number, 1-800-424-9300, for instructions on how to handle any pesticide emergency o Emergency involving the BASF product (Habitat): 800-832-HELP (800-832-4357) o Emergency involving the Nufarm product (Polaris AQ): 877-325-1840 o Emergency involving Monsanto products (Aquamaster): 314-694-4000 o Emergency involving Dow products (Rodeo): 800-369-2436 or 979-238-2112

Chemical Spill into Marsh or other Waters: o Report all spills within 24 hours to the San Francisco Regional Water Quality Control Board: 510-622-2369

Nearest Hospital The following space is provided for the ISP grant recipient or its contractor(s) to provide the location and directions to the closest hospital.

Name: ______

Address: ______

______

Phone: ______

Directions to hospital: ______

______

A map should be attached or otherwise made available on site.

Invasive Spartina Control Plans 1 General Site Safety & Materials Attachment 1 Handling Guidelines Sensitive Receptors When applying herbicides for non-native Spartina control, care must be taken to protect human health, and particularly “sensitive receptors” that may be located near the application area. Sensitive receptors include hospitals, schools, and residences near the Bay margin that are in close proximity (e.g., within 0.25 mile) to areas being treated with herbicide. The potential presence of sensitive receptors must be evaluated on a site-specific basis. The Adjacent Land Uses section in the site-specific Invasive Spartina Control Plan contains some information regarding potential sensitive receptors at each sub-area. In general, sensitive receptors are most likely to occur at sites located in the Corte Madera Creek watershed in Marin County, and along the shorelines of Alameda, Hayward and San Leandro in Alameda County. Birders, bicyclists, joggers, pedestrians, and users of recreational facilities (including parks, marinas, launch ramps, fishing piers, and beaches) that sur- round the Bay also could be sensitive receptors. The ISP grant recipient(s) and their contractors are responsible for fully identifying and protecting sensitive receptors. To minimize risks to the public, mitigation measures for herbicide treatment methods must be implemented by entities engaging in treatment activities. Such measures include, but are not limited to, the following: 1. Post signs for sensitive receptors within 500 feet. At least one week prior to application, post signs informing the public of impending herbicide treatment at prominent locations within a 500-foot radius (approximately 1/10 mile) of treatment sites where homes, schools, hospitals, or businesses could be affected. Schools and hospitals within 500 feet of any treatment site should be separately noticed at least one week prior to the application. 2. Avoid aerial spraying near sensitive receptors. Do not conduct aerial spraying within 0.25 mile (1,320 feet) of a school, hospital, or other sensitive receptor location. 3. Minimize drift. Manage herbicide application to minimize potential for herbicide drift (See Spray Drift Reduction, page 11 of this document). Herbicide must not be applied when winds are in excess of 10 miles per hour or when inversion conditions exist, or when wind could carry spray drift into inhabited areas. 4. Post signs at access points. Post colored signs at and/or near any public trails, boat launches, or other potential points of site access 24 hours prior to treatment. The signs should inform the public that the area is to be sprayed with glyphosate and/or imazapyr herbicide for weed control, and that the spray is harmful if inhaled. The signs should advise “no entry” for humans and animals until eight (8) hours after treatment, and the treatment date and time should be stated. A 24-hour ISP contact number may be provided. 5. Avoid high use areas. Avoid application of herbicides near areas where the public is likely to contact water or vegetation. For example, avoid applying herbicide in or adjacent to high use areas within 24 hours of high use times, such as weekends or certain holidays. If a situation arises that makes it necessary to treat high-use areas during such times, the areas should be closed to the public before, during, and after treatment.

Invasive Spartina Control Plans 2 General Site Safety & Materials Attachment 1 Handling Guidelines Listing of Sensitive Receptors The following spaces are provided for the ISP grantee and its contractors to list sensitive receptors within 0.25 mile of the herbicide treatment site. This information should be made available in advance to herbicide application contractors. Schools Hospitals Name

Address

Contact

Phone

Name

Address

Contact

Phone

Name

Address

Contact

Phone

Residences Name

Address

Contact

Phone

Name

Address

Contact

Phone

Name

Address

Contact

Phone

Invasive Spartina Control Plans 3 General Site Safety & Materials Attachment 1 Handling Guidelines Site Safety Protocols for Working in Marshes, Mudflats, and other Tidal areas of the San Francisco Estuary Tidal lands of the San Francisco Estuary present many unique hazards to workers who must access intertidal areas during the performance of their activity. The following is a summary of some of the hazards one may encounter when accessing these areas, and suggested precautions. There is no assertion made here, either stated or implied, that this list is comprehensive of all hazards that could possibly be encountered while in intertidal areas of the Estuary. Caution should be exercised at all times while in these areas, and common-sense danger avoidance techniques should be employed.

Teams Always travel with a partner when entering or working in marshlands. There are many hidden hazards associated with marshland travel and work that are not readily discernable at first glance. A team of at least two individuals adds a necessary level of safety for any work or activity taking place in the marsh. Supervisors should provide daily worker safety briefings prior to commencing work on site. These briefings need not be exhaustive, but should include any new information the supervisor may have obtained about the work conditions on the site, weather conditions, team assignments, equipment condition, or other pertinent issues.

Channels Watch for hidden channels and holes in the marsh plain as you traverse the marsh. Often smaller channels in the marsh are obscured by vegetation. These channels can be quite deep, and may result in a sprain or pulled muscle, or possible fractures. Use a probe, like a stick or staff to check ahead of your path for unseen channels. Keep alert for vegetation changes, like Grindelia sp. (Gumplant), which grows along channel edges and may indicate a hidden channel. Channel banks can sometimes be quite soft, and the mud that lines the channel can often be unstable. It is not unusual to sink deeply into these muds. This could be dangerous during an incoming tide. Always probe the mud within channels to test its ability to support your weight before stepping forward. Channels often block direct routes through the marsh. These channels can be quite small or very large. Great caution should be observed when considering crossing these channels. It may seem worthwhile in most cases to leap across the channel to get to the other side. This should only be done on the smaller channels, when your partner is able to follow, when you have surveyed the route for alternate paths around the channel, and at low tide. Large channels should be avoided entirely, and placing driftwood bridges over these channels is not advised. Workers may fall from unstable bridges into larger channels and risk injury, drowning, hypothermia, or equipment destruction. It is best in these situations to find a way around the channel.

Invasive Spartina Control Plans 4 General Site Safety & Materials Attachment 1 Handling Guidelines Mudflats Mudflats at low tide can be quite dangerous to the unprepared. Often these flats are extremely soft, making travel over them slow and messy. Without proper footwear made for travel over mudflats (called “mudders”), workers may sink up to their thighs in mud. If stuck deeply in a mudflat or channel bottom, you can often extract yourself by spreading out your weight over the mud by, in effect, laying or crawling on the muddy surface. Rocking your boots or waders back and forth to open space around your boot can also work to extract your feet from the mud. Assistance from your partner in the marsh can be essential during these situations. If feet sink into soft mud or quicksand, do not make vio- lent movements in an attempt to get free. If boots or waders become stuck, slip one foot out gradually, rest the leg on the surface and gradually free the other leg. Lying on the surface and spreading the weight can avoid sinking. Move to firm ground using a “leop- ard crawl” (spread eagled, face down, keeping the maximum area of the body in contact with the ground at all times).

Chemical or Physical Hazards Many marshland areas have been historic sites of dumping or disposal. Many marshes have accumulated debris or wrack that contains all manner of refuse. As a result, some areas have large amounts of this waste material, and in some cases, toxic waste or hazardous chemicals. Supervisors should be made aware of any known chemical or toxic waste issues associated with a site and take appropriate precautions. Workers should be notified prior to the beginning of operations within the marsh what the condition of the marsh is relative to toxic or hazardous substances, and be appropriately equipped. Wear footwear capable of resisting puncture by sharp objects. Nails, glass, chunks of concrete, rusty metal and other debris can severely injure workers without appropriate footwear. Ideally, workers should wear hip or chest waders with reinforced soles, that are resistant to puncture, tearing or chemicals. In areas where there are known concentrations of toxic or hazardous substances, a site- specific safety plan should be prepared and an appropriately trained hazardous materials expert should supervise work. At a minimum, workers should wear protective gloves and eyewear, long-sleeve shirts, and thoroughly wash all clothing subsequent to work in the marsh. Workers should also thoroughly wash themselves with soap and water following work activities. If potentially toxic or hazardous materials are discovered during work activities, the area should be marked and reported to the appropriate authorities (the County Hazardous Ma- terials Office and/or the Regional Water Quality Control Board). The area should be avoided until the material has been assessed and/or removed from the site.

Tides All workers in the marsh shall be made aware of the tidal schedule prior to work in the marsh. Work shall commence on an ebb tide and cease on the incoming tide or earlier. Allow ample time to return to non-tidal areas before the incoming tide starts to advance

Invasive Spartina Control Plans 5 General Site Safety & Materials Attachment 1 Handling Guidelines across the work site. This general rule may be modified in higher marsh habitats where tidal action is lessened, but workers should always be alert and mindful of working in situations where the incoming tides may trap them, and allow ample time for exiting the marsh prior to an incoming tide. If in doubt, get out. Tides can rise extremely quickly in some areas, and it is possible that rising tides may outpace the ability of workers to out- run the increasing water levels, especially in soft muds or heavily vegetated marsh plains.

Weather It is always important to monitor weather conditions prior to and during work activities within the marsh. Wind, rain, fog or other inclement weather can mean the difference between a safe work site and an extremely dangerous one. Winds usually occur in the early afternoon or late morning during the summer months, though dangerous weather patterns can occur at any time of the year. Rainfall may subject workers to hypothermia if unprepared, or may result in potentially dangerous floodwaters. Winds can increase wave action, whip up salt spray or dust. Fog can decrease workers ability to communicate or dis- cern potential hazards in the marsh. It is ill advised to go into marshland terrain in bad visibility. For all work performed in the intertidal areas of the Estuary, workers or supervisors should check weather forecasts prior to commencing work on the site, should monitor weather conditions for any changes while on site and should modify work plans accordingly to insure the safety of all personnel.

Communication Open lines of communication between workers in the marsh must be maintained. When more than one team will be working in the marsh at any one time, it is advisable to have a communication link to a land base and between individual teams for safety. In the case of injury, discovery of hazardous materials, endangered species, or cultural artifacts, or for other reasons, land-based assistance can be contacted from the field for immediate help or first aid. On the ground coordination via phone or walkie-talkie when crews are spread out over the marsh can help to avoid dangerous situations.

Invasive Spartina Control Plans 6 General Site Safety & Materials Attachment 1 Handling Guidelines Herbicide Handling, Spill Prevention and Spill Response The following information and practices are to be incorporated into herbicide-based Spartina control operations associated with the San Francisco Estuary Invasive Spartina Project (ISP).

Herbicide Use All herbicides shall be applied by or under the direct supervision of trained, certified or licensed applicators and in accordance with the product label On-site mixing and filling operations shall be confined to areas appropriately bermed or otherwise protected to minimize spread or dispersion of spilled herbicide or surfactant into surface waters

Herbicide Storage Proper herbicide storage is one of the keys to using herbicides safely. Always wear rub- ber gloves when handling herbicides in storage, and review product labels for specific storage instructions. General rules for herbicide storage include: Keep all herbicides in their original containers. Store herbicides in a locked shelter away from children and animals. Store herbicides in a dry, cool and well- ventilated area. DO NOT subject herbicides to freezing or extremely high temperatures. Store herbicides separately from seed, fertilizer, insecticides and food. Make periodic inspections of storage facilities and storage containers. Check for possible leaks, spills and other similar problems. Keep appropriate absorbent material in the storage area at all times as well as a plastic container for storing damaged material. Reject any broken or leaking containers when herbicides are delivered. Do not store herbicides in office or break areas where employees congregate.

Container Disposal Empty herbicide containers must be disposed of according to government regulations or be returned to the manufacturer for disposal. Empty containers not returned to the manufacturer can be handled according to the procedures below, as long as local, state and federal laws are followed: Triple rinse containers with water at the application site. Always pour the rinse- water into an appropriate receptacle. Rinsed containers should be disposed of in a landfill approved for pesticide disposal or in accordance with applicable government procedures. Check with your

Invasive Spartina Control Plans 7 General Site Safety & Materials Attachment 1 Handling Guidelines supervisor to find out if and when herbicide containers may be handled in this manner.

Spill Response Under all circumstances, it is the responsibility of the applicator to assure that all precautions are taken prior to initiating work to assure protection of water quality and the environment. The applicator is also responsible for the provision of a Spill Response Kit that is appropriate for the work being undertaken.

The following procedures should be followed in the case of a non-petroleum chemical spill: Put on protective gloves, eyewear, a long-sleeved shirt and pants before cleanup If a container is leaking, immediately transfer the remaining herbicide to another appropriate container to prevent further spillage If the herbicide was spilled on a person, remove the contaminated clothing and rinse the product from the body. If necessary, perform appropriate first aid. Cover the spill area with an absorbent material to soak up the herbicide. Common cat litter, sawdust, soil or sand can all be used for this purpose. Consult the manufacturer for more specific clean up recommendations. Remove any contaminated items from the spill area to prevent further contamination Remove the absorbent material with a broom and or shovel after the spill has been absorbed. Make sure all contaminated soil is removed from the spill area as well. Place the contaminated soil and absorbent material into a suitable container, and dispose of the container in an approved landfill area Do not wash down the area with water using a high pressure hose. You may spread the spill and make the herbicide more difficult to contain and clean up. When a spill occurs on a site, or is large enough that you need help to contain or clean it up, contact a supervisor immediately. In case of a major spill, call the manufacturer or ChemTrec (Chemical Transportation Emergency Center), 1-800- 424-9300.

Spill Response Kit A Spill Response Kit should provided at the work site and be immediately accessible to all personnel. Some or all of the following items may be included in a Spill Response Kit. Consider site-specific conditions and the chemicals to be used to determine which of the following items are appropriate. PVC Gloves or equivalent (to mid forearm) PVC boots or equivalent Chemical resistant splash goggles Vice grip pliers

Invasive Spartina Control Plans 8 General Site Safety & Materials Attachment 1 Handling Guidelines Phillips head screwdriver (2) Shovels Brooms, dustpan Clay granules or a sawdust Activated charcoal or other appropriate absorbent material First aid kit Tyvek coveralls (2 pair) or neoprene coveralls Recovery drums DOT triangular reflector kit Source of clean water and soap In the case of refueling or mixing activities planned on open mudflats the spill response kit should include a portable wet vacuum or other pumping equipment

Preventing Spills The following procedures will help to minimize the risk of spills occurring: Keep bags and cardboard containers dry at all times Prevent or correct leaks in herbicide containers and application equipment Properly dispose of all empty pesticide containers Tie down or otherwise secure containers when transporting pesticides to prevent them from falling from a vehicle Store herbicides only in their original containers or properly labeled service containers Stay alert and attentive when handling or using herbicides Where on-site or in-field transfer of liquid chemicals (herbicide mixtures, fueling operations) is planned, the transfer will occur at an appropriate upland site (staging area) to avoid contamination of the marsh or adjacent surface waters. A closed transfer system with a dry lock is preferred for these operations.

Procedures for Liquid Spill Response The following procedures should be followed in the case of a non-petroleum spill: Put on protective gloves, eyewear, a long-sleeved shirt and pants before cleanup If a container is leaking, immediately transfer the remaining herbicide to another appropriate container to prevent further spillage If the herbicide was spilled on a person, remove the contaminated clothing and rinse the product from the body. If necessary, perform appropriate first aid or seek immediate medical attention. Cover the spill area with an absorbent material to soak up the herbicide. Common cat litter, sawdust, soil or sand can all be used for this purpose. Consult the manufacturer for more specific clean up recommendations.

Invasive Spartina Control Plans 9 General Site Safety & Materials Attachment 1 Handling Guidelines Remove any contaminated items from the spill area to prevent further contamination Remove the absorbent material with a broom and or shovel after the spill has been absorbed. Make sure all contaminated soil is removed from the spill area as well. Place the contaminated soil and absorbent material into a suitable container, and dispose of the container in an approved landfill area Do not wash down the area with water using a high pressure hose. You may spread the spill and make the herbicide more difficult to contain and clean up. When a spill occurs on a site, or is large enough that you need help to contain or clean it up, contact a supervisor immediately. In case of a major spill, call the manufacturer or ChemTrec (Chemical Transportation Emergency Center).

Under all circumstances it is the responsibility of the applicator to assure that all precautions are taken prior to initiating work to assure protection of water quality and the environment. The applicator is also responsible for the provision of a Spill Response Kit that is appropriate for the work being undertaken.

Invasive Spartina Control Plans 10 General Site Safety & Materials Attachment 1 Handling Guidelines Spray Drift Reduction

Definition of The Department of Pesticide Regulation (DPR) defines pesticide drift pesticide drift as the pesticide that moves through the air and is not deposited on the target area at the time of application. Drift does NOT include movement of pesticide and associated degradation compounds off the target area after application (e.g., translocation, volatilization, evaporation, or the movement of pesticide dusts or pesticide residues on soil particles that are windblown after application.)

The pesticide Pesticide drift, particularly from agricultural fields, has been known to drift issue impact adjacent residential areas, cause damage to non-target crops, and contaminate the environment.

How does Low levels of pesticide drift may occur from all types of pesticide ap- pesticide drift plications. Pesticide drift becomes unacceptable when pesticides are occur? applied by imprecise methods or under environmental conditions that prohibit the applicator from maintaining control over the path the pesticide takes once it leaves the application equipment.

The San Francisco Estuary Invasive Spartina Project (ISP) has identified the use of herbicide as a critical component of its Spartina Control Program. The herbicide used for Spartina Control is imazapyr (Polaris™ or Habitat®) a product with exceptionally low toxicity, approved by U.S. EPA and the State of California for use in sensitive aquatic and estuarine environments. The human health risks associated with imazapyr are very low, and it requires no special personal protection measures for handling and application beyond those on the FIFRA label. In any case, it is desirable to minimize exposure of humans or non-target plants to pesticide drift. The ISP requires that all herbicide application under the Control Program be managed to minimize spray drift to protect human health and the environment. Application of herbicide and surfactants in accordance with product labels (including the Supplemental La- beling for Aerial Application in California) will minimize spray drift. In addition, the ISP recommends the following: 1. For ground application of herbicide mixture by vehicle-mounted or towed ground equipment: a. Herbicide should be applied only when wind speed is 10 miles per hour or less at the application site, as measured by an anemometer positioned four feet above the ground. b. Discharge should start after entering the target site; discharge should be shut off before exiting the target site. 2. For aerial application of herbicide mixtures: a. Application should be by helicopter; no airplane application should be used. b. Nozzle orifices of broadcast sprayers should be directed backward. c. Flow of liquid from each nozzle should be controlled by a positive shutoff system.

Invasive Spartina Control Plans 11 General Site Safety & Materials Attachment 1 Handling Guidelines d. Spray nozzles should be adjustable to allow control of droplet size. Use up to 1500 microns for windy conditions. e. Boom pressure should not exceed the manufacturer’s recommended pressure for the nozzles being used. f. Herbicide should be applied only when wind speed is three to 10 miles per hour at the application site, as measured by an anemometer positioned four feet above the ground. g. Discharge should start only after entering the target site; discharge height should not exceed 10-15 feet above the target vegetation; discharge should be shut off whenever necessary to raise the equipment over obstacles; discharge should be shut off before exiting the target site.

Invasive Spartina Control Plans 12 General Site Safety & Materials Attachment 1 Handling Guidelines Petroleum Fuel Spill Prevention and Response Spills of gasoline or other petroleum products, required for operation of motorized equipment, into or near open water could degrade water quality, with potential for bioaccumulation of contaminant toxicity. Several types of equipment used for treatment of Spartina may present opportunities for petroleum spills. Equipment used in Spartina control activities include: Amphibious tracked vehicles Spray trucks Water-based excavators (e.g. Aquamog) Gas-powered mowers (e.g. Weed-Whackers) Air boats and outboard motor boats Fueling Fueling of amphibious tracked vehicles, spray trucks or land-based excavators should be done offsite at fueling stations or suitable staging areas. A suitable staging area shall be equipped with sufficient protection to prohibit a petroleum spill from migrating beyond the immediate fueling area (e.g., an impermeable plastic tarp set between raised berms, a catch basin or similar portable device). Water-based excavators, airboats and outboard motor boats shall be fueled offsite at commercial fueling stations or designated locations such as equipment maintenance yards. When fueling is done on or adjacent to treatment sites, a spill prevention and response plan must be prepared and implemented. A copy of this plan shall be provided to the Invasive Spartina Project at [email protected]. Gas powered, hand held machinery (e.g., weed whackers) shall be refueled on a non- absorbent tarp or mat placed under machinery to catch any spills. In addition to spills during refueling operations, small amounts of oil or fuel may leak from improperly maintained equipment. Before using any equipment in the marsh, check to make sure that it is in good working order with no signs of leakage or corrosion that might indicate the potential for inadvertent spills on the work site. Transport vessels and vehicles, and other equipment (e.g., mower, pumps, etc.) shall not be serviced or fueled in the field except under emergency conditions. Under all circumstances, it is the responsibility of the applicator to assure that all precautions are taken prior to initiating work to assure protection of water quality and the environment. The applicator is also responsible for the provision of a Spill Response Kit that is appropriate for the work being undertaken.

Invasive Spartina Control Plans 13 General Site Safety & Materials Attachment 1 Handling Guidelines Herbicide Information This section provides product labels and Material Safety Data Sheets (MSDS) for herbicides and adjuvants that have been evaluated and approved for use in controlling non- native Spartina in the San Francisco Estuary. Product labels and MSDSs contain important information to help protect human health and the environment, and they should be included as a part your Site Safety Plan. Included in this section are the following products:

Imazapyr Herbicide: Polaris™ (imazapyr-based herbicide) – Product label and MSDS. The product label and MSDS for Habitat® are essentially identical and it is approved by ISP.

Surfactants: 1. Liberate® (non-ionic surfactant/drift retardant) - Product label and MSDS. 2. Competitor® (methylated seed oil) – Product label and MSDS.

Colorants: 3. Turf Trax® or Hi-Light®, (spray pattern indicator) - Product label and MSDS.

Please note that ONLY the aquatic herbicides and surfactants that are listed here are approved for use in Spartina control in the San Francisco Estuary. There are other drift re- tardants and anti-foaming agents that may be used, provided the ISP Partner has reviewed the product information and found the product to pose no significant risk to human health or the estuarine environment. It is the responsibility of the applicator to obtain product labels and MSDSs for any products not included in this document. It is the responsibility of the applicator to assure that the most current product labels are obtained and followed.

Invasive Spartina Control Plans 14 General Site Safety & Materials Attachment 1 Handling Guidelines