Trends in Hydrology and Salinity in Suisun Bay and the Western Delta

Home , Bradford Island, San Joaquin River, Suisun Bay

Draft Version 1.2

June 2007

DRAFT

Trends in Hydrology and Salinity in Suisun Bay and the Western Delta

Draft Version 1.2

Executive Summary...... 2 Objective...... 2 Approach...... 2 Conclusions...... 3 Report Structure...... 7 1. Introduction...... 8 1.1. Objectives of this review ...... 8 1.2. Salinity Units...... 9 1.3. Temporal and Spatial Variability...... 13 1.4. Report Structure...... 14 2. Factors Influencing Salinity Intrusion in the Delta...... 16 2.1. Climatic variability...... 17 2.2. Physical changes to the Delta and Central Valley ...... 19 2.3. Flow Management Regimes ...... 21 3. Historical Context – The last 3,000 years...... 28 3.1. Reconstructed Salinity in Northwestern Suisun Marsh (~1000 B.C – 2000 A.D) ...... 28 3.2. Reconstructed Unimpaired Flow in the Sacramento Valley (900 A.D. – 1976 A.D.)... 29 4. Qualitative Salinity Observations (late 1700s – early 1900s)...... 30 4.1. Observations from Early Explorers ...... 30 4.2. Observations from early settlers in the Western Delta ...... 33 5. Quantitative Salinity Observations (early 1900s – present) ...... 39 5.1. Fluctuation or Movement of the Spatial Salinity Distribution...... 39 5.2. Trends in Salinity at Specific Locations ...... 52 6. Conclusions...... 62 7. References...... 67 Supplemental Figures...... 70

DRAFT: Last modified 6/9/2007 10:51 PM 1 D_report_flow_salt_trends_v1p2.doc DRAFT

Executive Summary

Objective

Prompted by recent discussions that the Sacramento-San Joaquin Delta (Delta) is currently managed as a freshwater system with significantly less fluctuation in salinity than would occur under “natural” conditions, this report examines present-day salinity levels and variability in the context of historical conditions.

The following questions are addressed in this report:

• Average salinity: How does the average present-day salinity at specific locations in the western Delta and Suisun Bay compare with average salinity at the same locations at different times in history?

• Seasonal fluctuations in salinity: How does the seasonal variability in salinity in the present day compare with historical seasonal salinity fluctuations?

• Inter-annual fluctuations in salinity: How do maximum and minimum annual salinity intrusion during wet and dry years in the present day compare with historical salinity intrusion under similar hydrologic conditions?

• Flow management: How is variability at various timescales altered by the cumulative impact of upstream diversions, reservoir operations, in-Delta diversions, and south of Delta exports?

The following questions are not addressed:

• What salinity regime should be imposed on the Suisun Bay and Delta? • Is salinity the best metric to evaluate ecosystem dynamics? What other metrics should be considered?

Approach

Hydrology and salinity in Suisun Bay and the western Delta vary extensively in both space and time. Misinterpretation of the location or time period of a given observation, or comparisons between conditions at widely separated locations can cause confusion. Confusion may be compounded and lead to conflicting conclusions when salinity is discussed in qualitative terms such as “fresh” and “brackish”. To clarify discussions, careful references to quantitative salinity levels, time periods for comparison, and locations in the Delta are made throughout this report.

The introduction to this report provides contextual information on the temporal and spatial variability of salinity in Suisun Bay and the western Delta. To establish a timeline

DRAFT: Last modified 6/9/2007 10:51 PM 2 D_report_flow_salt_trends_v1p2.doc DRAFT

of natural and anthropogenic modifications, Section 2 presents a qualitative discussion of factors that influence salinity intrusion. The balance of this report is a chronological review of salinity observations, referencing paleoclimatic reconstructions (Section 3), qualitative observations (Section 4), and quantitative measurements (Section 5).

Conclusions

Key findings for each research question are summarized here with reference to more detailed information within this report.

• Average salinity

On average, the Delta is not significantly fresher today. Paleosalinity analyses and evidence from the early 1900s both suggest that the Delta is actually saltier today than in the early 1900s.

Paleoclimatic research provides evidence of salinity variability prior to European influence. Analyses of sediment cores in northwestern Suisun Marsh indicate that the marsh has experienced centennial-scale cycles of fresh and brackish conditions. During the last century, salinity has increased substantially in comparison to the previous 600 years and has been as salty as or saltier than periods as far back as 2,500 years before the present time. (Section 3.1, Page 28)

The earliest quantitative measurements of salinity were recorded by the California & Hawaiian Sugar Refining Corporation (C&H) from 1908 to 1929; C&H measured the distance its barges had to travel upstream on the Sacramento and San Joquin rivers to obtain water suitable for their refinery and quantified the salinity of the water. From 1908 through 1917, the farthest monthly averaged distance traveled to obtain suitable water was 30 miles upstream of Crockett (below Jersey Point on the San Joaquin River). By 1920, the quality of water easily obtained by C&H barges had degraded due to increased upstream irrigation diversions, especially for newly introduced rice cultivation, to the point that C&H abandoned the Sacramento and San Joaquin Rivers during the summer and fall, replacing the water supply with an agreement with Marin County.

Comparison of the C&H observations for 1908 through 1917 (prior to significant upstream diversions) with recent data indicates that there has been a 7-fold increase in salinity in the western Delta and Suisun Bay. The location of the 50 mg/L chloride (350 µS/cm EC) isohaline observed by C&H from 1908 through 1917 is approximately the same as the location of X2 (2,640 µS/cm EC) in recent years with similar unimpaired hydrology (1995 to 2005). The distance from the C&H plant to 50 mg/L chloride water is greater in the last 10 years than it was in the early 20th century, with the largest increase in the spring. Fresh water (less than 350 µS/cm EC) has been located below the confluence of the Sacramento and San Joaquin Rivers approximately 42% of the time from 1995 to 2005, compared to 72% percent of the

DRAFT: Last modified 6/9/2007 10:51 PM 3 D_report_flow_salt_trends_v1p2.doc DRAFT

time from 1908 to 1917, indicating that the Delta is now saltier on average for a similar hydrologic period. (Section 5.1.1, Page 39)

Long-term monitoring data at fixed locations in the Bay and the Delta started as early as 1920. Previous analyses of trends in Suisun Bay and the western Delta indicate that centennial-scale trends in outflow and salinity are much smaller than seasonal and decadal trends (Enright et al, 2004; Fox, 1987). Salinity at Collinsville (near the confluence of the Sacramento and San Joaquin Rivers) gradually decreased from 1930 to about 1967 but has been increasing from 1967 to present. (Section 5.2, Page 52).

• Seasonal fluctuations in salinity

Seasonal variability has significantly changed over multiple time periods. Relative to the earliest salinity measurements (1908-1917), currently: • salinity begins to intrude in the mid-winter, instead of the early-summer; • the western Delta does not get as fresh and does not stay fresh for as long into the late-spring and early-summer; • fall salinity during normal water years is greater than any previously measured time period.

Early observations in the late 1800s and early 1900s indicate that fresh water was available for a longer period of the year in the western Delta.

Accounts from early settlers near Antioch indicate that fresh water was always available in the San Joaquin River at least at low tide from 1866 to 1917, except in a couple of months during the driest years, the drought of 1870-71. In contrast, fresh water is only available today at Antioch for approximately 8 months each year on average. (Section 4.2, Page 33).

Data collected by C&H is consistent with the qualitative reports from early Antioch residents. C&H data indicates that fresh water was present further downstream in Suisun Bay and the western Delta more often and for a longer portion of the year from 1908 to 1917 compared to recent years with similar unimpaired hydrology. (Section 5.1.1, Page 39).

Availability of fresh water (and the extent of salinity intrusion) is largely determined by the net Delta outflow (NDO), the tidally-averaged flow of fresh water leaving the Delta into the Bay. Upstream diversions and consumptive use reduce NDO; evidence of their impact on salinity was observed by Antioch residents and recorded by C&H.

After significant upstream diversions began but before large upstream storage reservoirs were built, most of the water diverted for irrigation was taken from the streams when flows were already seasonally low, leading to significant increases in salinity intrusion during the summer and fall. However, salinity in the winter and

DRAFT: Last modified 6/9/2007 10:51 PM 4 D_report_flow_salt_trends_v1p2.doc DRAFT

spring was likely unaltered by water diversions during this time period because irrigation demands were seasonally low during the wet season and the seasonal flow peaks could not be captured without significant upstream storage. According to the Department of Water Resources, “[d]uring certain years of the thirteen-year period, 1917 to 1929, the extent of saline invasion into the Sacramento-San Joaquin Delta has been greater than ever before known to have occurred.” (DPW, 1931) During the 1930s, additional upstream diversions combined with a prolonged drought further increased salinity intrusion beyond any previously known levels. (Section 5.1.2, Page 46).

Upstream storage reservoirs have altered the seasonal distribution of NDO by capturing a portion of the peak flows during high runoff and snowmelt and releasing flows during the dry season intended for downstream diversion to irrigation. Consequently, fresh periods in the winter and spring are saltier now and last for a shorter interval; the summer and fall are also saltier, but the effect is smaller than it was in the 1920s through the 1940s, before significant upstream storage was constructed.

During wet years, reservoir operations have provided slightly fresher water in the fall months than would have been available in the absence of upstream storage, possibly due to releases for flood control. However, during the last 10 years, the impact of fall reservoir releases on salinity has been substantially less than the impact from 1945 to 1975. In fact, fall salinity has increased considerably, exceeding the predicted salinity under unimpaired conditions. (Section 5.2, Page 52).

• Interannual fluctuations in salinity

Interannual salinity variability has not decreased. Salinity observations in the 18th, 19th, and early 20th centuries do not indicate that maximum salinity intrusion during droughts was significantly greater than it is today.

Peak annual salinity intrusion during the recent dry years of 1976-1977, 1981, and 1987-1992 is comparable to early observations of salinity during similar hydrology in the 18th, 19th, and early 20th centuries. (Section 4.1, Page 30) Salinity intrusion at Kentucky Point on Twitchell Island during the drought of 1870-71 is similar to intrusion during 1981, a similar water year. (Section 4.2.3, Page 37)

The misconception that the Delta is “artificially fresh” and does not experience the full range of “natural” salinity intrusions may be due to casual comparison between the maximum salinity intrusion diagrams in the Delta Atlas (DWR, 1993), comparing the 1921 to 1943 time period with the 1944 to 1990 time period. These figures illustrate that the maximum salinity intrusion during six years in the 1920s and 1930s was greater than then maximum intrusion since Shasta Dam was built in 1945. However, salinity intrusion in the 1920s and 1930s should not be considered “natural” as upstream diversions were already significantly reducing river flows

DRAFT: Last modified 6/9/2007 10:51 PM 5 D_report_flow_salt_trends_v1p2.doc DRAFT

during the irrigation season (DPW, 1931; Lund et al, 2007). Salinity intrusion during the 1920s and 1930s is discussed in Section 5.1.2 (Page 46).

• Flow management

Flow management reduces the net Delta outflow, increases average salinity, and significantly alters the seasonal variability. Flow management accounts for the observed changes in seasonal salinity variability relative to the seasonal structure of the early 1900s.

The cumulative impact of upstream diversions, reservoir operations, in-Delta diversion, and south of Delta exports is evaluated with predictive models based on net Delta outflow (NDO). This method isolates the effects of flow management from other factors that influence salinity by comparing the predicted salinity based on unimpaired hydrology with predicted salinity based on actual hydrology.

Flow management has reduced annual NDO in every year of the historical record from 1929 to 2002. Monthly average unimpaired NDO increases almost 600 thousand acre-feet (TAF) from 1921 to 2003, indicating that, overall, hydrologic conditions have been getting wetter. However, the monthly average actual NDO slightly decreases by approximately 180 TAF from 1930 to 2006. (Section 2.3, Page 21)

Flow management also redistributes the outflow seasonally, greatly reducing NDO from March through June and slightly increasing NDO during some years from September through December. Substantial changes in flow management during the period of record alter these monthly trends. For instance, from 1944 to 1967 (post- CVP and pre-SWP) NDO in September was approximately 200 TAF more than unimpaired conditions, indicating the reservoirs (largely constructed 1940 through 1975) and irrigation return flows were supplementing outflow in September. From 1967 to 1975, the amount of supplemental flow in September increased from 200 TAF to approximately 600 TAF. The drought of 1976-1977 caused actual September NDO to drop below unimpaired September NDO. From 1978 through 2003, actual September NDO remains below unimpaired September NDO most years, with supplemental flow exceeding 100 TAF in only four relatively wet years. (Section 2.3, Page 21)

The impacts of these changes in NDO are evaluated with predictive models of two distinct metrics for salinity in the western Delta, X2 (the distance from the Golden Gate to the location where near-bottom salinity is 2 psu, or 2,640 µS/cm) and electrical conductivity at Collinsville. Both models indicate that, on average, actual NDO results in greater salinity intrusion into Suisun Bay and the Delta than unimpaired conditions. The models both show a similar seasonal impact of flow management, with the greatest impact in the spring and summer.

DRAFT: Last modified 6/9/2007 10:51 PM 6 D_report_flow_salt_trends_v1p2.doc DRAFT

From April through August, the long-term trend associated with flow management activities substantially increase (more than double, on average) salinity at Collinsville over unimpaired conditions. Since 1993, a seasonal shift in flow within the system results in significant increases in salinity at Collinsville from October through December relative to unimpaired conditions. The late fall high salinity far exceeds any previous conditions except the 1930s drought. Although there is a corresponding small decrease in salinity from May through August in this period (probably due to minimum flow and X2 requirements imposed in 1995), spring and summer salinities remain much greater than unimpaired conditions would provide. (Section 5.2.3, Page 54).

Report Structure

To establish a timeline of natural and anthropogenic modifications, the report begins with a qualitative discussion of factors that influence salinity intrusion in the Delta. The remainder of the report is a chronological review of salinity data.

The report is divided into four sections:

Factors Influencing Salinity Intrusion in the Delta – categorizes factors that are known to influence salinity intrusion in the Delta, identifying time periods when factors have changed significantly, and discussing the probable net impact of the changes on salinity intrusion.

Historical Context – The last 3,000 years – summarizes paleoclimatic evidence of salinity levels and variability over the last 3,000 years to provide context for salinity observations in the last 250 years.

Qualitative Salinity Observations (late 1700s to the early 1900s) – presents a qualitative analysis of anecdotal information from historical documents, including manuscripts from early explorers and settlers and legal briefs from Town of Antioch v. Williams Irrigation District (1922, 188 Cal. 451).

Quantitative Salinity Observations (early 1900s to the present) – compares the current salinity regime with salinity measurements from the early 1900s.

DRAFT: Last modified 6/9/2007 10:51 PM 7 D_report_flow_salt_trends_v1p2.doc DRAFT

1. Introduction

The San Francisco Bay / Sacramento-San Joaquin Delta Estuary (Bay-Delta) has been the focus of extensive research for many years. This report summarizes previous studies and uses publicly available data to examine trends in hydrology and salinity in the Bay-Delta.

Intensive salinity measurements in the Delta started in 1920, with samples taken at multiple locations at least once a week during the summer and fall. Consequently, many studies compare salinity to these early measurements, often implying the salinity regime of the 1920s was “natural”. However, historical reviews indicate that upstream irrigation diversions had a significant effect on streamflow and salinity intrusion even prior to 1920.

[B]y 1870 so much water was being taken from the San Joaquin River and its tributaries that streamflow was noticeably reduced. … [¶] Irrigation developed more slowly in the Sacramento Valley but with the coming of the rice industry in 1912 diversions increased dramatically. … Since rice fields must be completely inundated, rice culture was primarily responsible for increasing the gross irrigation diversion from the Sacramento River from 1,154,000 acre-feet in 1915 to 2,300,000 acre-feet in 1919. … [S]uch significant reductions in the already low summer streamflow could not help but affect the penetration of ocean salinity. (Jackson and Paterson, 1977)

Recognizing that the effect of anthropogenic modification to the landscape and water usage was significant prior to 1920, we look to other data sources to provide historical context. Paleoclimatic studies are reviewed to examine variability over thousands of years. Observations from early explorers and settlers provide anecdotal evidence of salinity before significant anthropogenic modifications. The earliest quantitative salinity measurements, collected by California and Hawaiian Sugar Refining Corporation (C&H) from 1908 to 1929, capture the changing salinity regime in the early 1900s due to anthropogenic modifications.

Comparison of data from multiples sources requires caution and attention to detail such as salinity units, time period of sampling, any averaging of the sample data, and the location of samples associated with each source.

1.1. Objectives of this review

DRAFT: Last modified 6/9/2007 10:51 PM 8 D_report_flow_salt_trends_v1p2.doc DRAFT

“Natural” conditions are often discussed as if there is a static state of the environment before significant human intervention. However, the Delta landscape and hydrology vary continuously at multiple timescales (diurnal, seasonal, annual, decadal, centennial, and so on). Rather than defining a specific time in history as “natural”, this report summarizes available salinity information with reference to the time period of the observations, comparing each period to the present-day salinity regime.

The following questions are addressed in this report:

• Seasonal fluctuations in salinity: How does the seasonal variability in salinity in the present day compare with historical seasonal salinity fluctuations?

• Flow management: How is variability at various time-scales altered by the cumulative impact of upstream diversions, reservoir operations, in-Delta diversion, and south of Delta exports?

The following questions are not addressed:

• What salinity regime should be imposed on the Suisun Bay and Delta? This report does not evaluate what salinity regime is “best” for the ecosystem, municipal water users, agriculture, state economics, or any other stakeholder group. Similarly, we do not select a time period in history as “natural” or “desirable”. Instead, we examine historical salinity observation in context and summarize information from multiple sources.

• Is salinity the best metric to evaluate ecosystem dynamics? What other metrics should be considered?

1.2. Salinity Units

Throughout this report, salinity is quantified as either electrical conductivity (EC) in units of microSimens per centimeter (µS/cm) or concentration of chloride in milligrams of chloride per liter of water (mg/L). Table 1 presents approximate equivalents between these two units.

DRAFT: Last modified 6/9/2007 10:51 PM 9 D_report_flow_salt_trends_v1p2.doc DRAFT

Table 1 – Salinity Units and Approximate Equivalents

Electrical Chlorides Conductivity (mg/L) (µS/cm) 350 50 525 100 1,050 250 1,930 500 2,640 700 3,700 1,000 5,000 1,300 10,000 2,800

Qualitative terms such as “fresh” and “brackish” are often used to describe relative salinity. The concentration of chlorides that separates fresh from brackish water has been given as 1,000 milligrams per liter (mg/L), 700 mg/L, and 50 mg/L by various sources cited in this report. (Table 2)

Table 2 – Metrics used to describe the boundary between “fresh” and “brackish”

Salinity Value Sample timing or Description Chlorides Electrical averaging milligrams per Conductivity liter (mg/L) (µS/cm EC) Isohalines depicted Annual maximum of 1,000 mg/L 3,700 µS/cm in the Delta Atlas the daily maximum (DWR, 1993) X2 position Daily average 700 mg/L 2,640 µS/cm (or a 14-day average) Barge travel by Monthly average of 50 mg/L 350 µS/cm C&H1 the daily maximum

Comparison of data from multiple sources may lead to incorrect conclusions. For instance, Figure 1 has been presented recently as an illustration of the change in salinity intrusion over time, but the graphic is misleading because it compares the location of 50 mg/L water with the location of 1,000 mg/L water without mention of the difference in salinity. Additionally, the graphic illustrates the location of the 50 mg/L water during dry years while showing the position of 1,000 mg/L water during wet years. An attempt to illustrate data from the same sources, resolving the data inconsistencies, is shown in Figure 2.

1 The California & Hawaiian Sugar Refining Corporation in Crockett (C&H) obtained its freshwater supply from barges traveling up the Sacramento and San Joaquin Rivers, generally twice a day beginning in 1905 (DWR, 1931). DRAFT: Last modified 6/9/2007 10:51 PM 10 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 1 – Confusing Comparison of Data from Multiple Sources Caution is required when comparing multiple data sources. This recent graphic compares the location of maximum salinity intrusion from two sources that use different salinity values to distinguish intrusion. Additionally, this illustration compares dry years in one time period with wet years in another time period.

Source of Graphic: Dotted lines University of California, California Colloquium on Water, show extent April 10, 2007. Ellen Hanak, of saltwater Public Policy Institute of California, incursions "Envisioning Futures for the Sacramento- San Joaquin Delta“. Slide Title: “Hydraulic Barrier” Limits Seasonal and Dry-Year Salinity Incursions

Hydraulic barrier (since 1940s) Summers, 1908-1917

Summer* 1841 50 mg/L chlorides Dry Years 1908-1917 (DPW, 1931)

1,000 mg/L chlorides Wet Years 1944-1990 (DWR, 2007)

DRAFT: Last modified 6/9/2007 10:51 PM 11 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 2 – Spatial Variability of Salinity Comparing Multiple Data Sources Salinity intrusion for relatively dry water years with similar total annual unimpaired runoff, using 1,000 mg/L chloride concentration to distinguish the extent of intrusion. Water year 1913 (12.9 MAF SRI) experienced the least extent of intrusion, most likely because upstream diversions were significantly less than later years. Water years 1926 (11.7 MAF SRI) and 1932 (13.1 MAF SRI) were subject to extensive upstream agricultural diversions, while water years 1979 (12.4 MAF SRI) and 2002 (14.6 MAF SRI) had the benefit of the CVP and SWP to provide “salinity control”.

1926 1932 2002 1979

1913

Salinity intrusion during 1913 is estimated based on the location of peak salinity intrusion of 50 mg/L water as observed by C&H (approximately 40 miles upstream of Crockett on the San Joaquin River). To determine the corresponding location of water with 1,000 mg/L chlorides, a relationship was formed based on monitoring data from 1965 to 2005.

To clarify discussions on this topic, this report makes careful references to quantitative salinity levels.

DRAFT: Last modified 6/9/2007 10:51 PM 12 D_report_flow_salt_trends_v1p2.doc DRAFT

1.3. Temporal and Spatial Variability

Salinity in the Bay-Delta varies both in space and time. In discussions of salinity trends, clarity about time scales and locations is essential.

Considering the Bay-Delta region at large spatial scale, the overwhelming spatial variability of salinity is due to the gradient from the Pacific Ocean, which has a relatively high salinity of approximately 50,000 µS/cm, to the rivers of the Central Valley, which have a relatively low salinity of approximately 100 µS/cm. Secondary salinity sources such as agricultural drainage, urban runoff, and urban discharges are not addressed in this report.

The spatial gradient oscillates back and forth with the tides. For instance, during the summer of 1991, the position of X2 moved up and down the Sacramento River along Sherman Island, traversing between 4 and 7 miles on each phase of the tide, depending on the spring-neap cycle.

Daily salinity variability at a given location can exceed seasonal or interannual variability (Figure 3).

Figure 3 – Hourly and daily salinity variability in the San Joaquin River at Antioch The daily range of EC in the San Joaquin River at Antioch can exceed the seasonal and interannual range. For instance, during the fall of 1999, the range of hourly EC is approximately 6,000 µS/cm from 3,000 µS/cm to 9,000 µS/cm, which is comparable to the range in daily average EC from approximately 100 µS/cm to 6,000 µS/cm. Salinity on San Joaquin River at Antioch 12000 Wet WY Wet WY Wet/AN WY 35 MAF 42 MAF 27 MAF 10000

8000 Hourly Data Daily Average

S/cm] 6000 μ EC [ EC

4000

2000

0 1997 1998 1999 2000

DRAFT: Last modified 6/9/2007 10:51 PM 13 D_report_flow_salt_trends_v1p2.doc DRAFT

High tide, low tide, and daily average salinity at a given location are far from identical (Supplemental Figure 8 and Supplemental Figure 9). Early salinity measurements coordinated by the State attempted to time the sampling to capture the tidal maximum salinity. The maximum salinity intrusion depicted in the Sacramento-San Joaquin Atlas is therefore the single day annual maximum of the tidal maximum salinity.

1.4. Report Structure

Factors Influencing Salinity Intrusion in the Delta Salinity intrusion into the Bay and Delta is determined primarily by the interaction of the oscillatory tides and freshwater outflow, which are both altered by natural and anthropogenic means. This section categorizes factors that are known to influence salinity intrusion in the Delta, identifying time periods when factors have changed significantly, and discussing the probable net impact of the changes on salinity intrusion. The discussion is qualitative, providing historical context and background on estuarine dynamics.

Historical Context – The last 3,000 years This section summarizes paleoclimatic study evidence of salinity levels and variability over the last 3,000 years to provide context for salinity observations in the last 250 years.

Qualitative Salinity Observations (late 1700s – early 1900s) Qualitative, anecdotal information on salinity is available from historical documents, including manuscripts from early explorers and settlers. These eyewitness accounts are often difficult to interpret because the location of observations is not precise and the salinity is generally characterized qualitatively without reference to the desired use of the water.

Legal briefs from Town of Antioch v. Williams Irrigation District (1922, 188 Cal. 451) are also a good source of anecdotal information. These briefs are more specific as to time and location of observations, and often contrast perspectives from multiple observers, providing a more balanced description of the salinity regime.

Quantitative Salinity Observations (early 1900s – present) Quantitative salinity measurements began in 1908, recorded by the California & Hawaiian Sugar Refining Corporation (C&H). C&H obtained its freshwater supply from barges traveling up the Sacramento and San Joaquin rivers, generally twice a day from 1905 to 1920. Starting in 1921, due to unprecedented salinity intrusion, C&H began obtaining a portion of its water supply from Marin County.

Present-day monitoring data, gathered by various agencies, is available from multiple sources including the Interagency Ecological Program (IEP)2 and the Environment

2 Data available at: http://www.iep.ca.gov/dss/all/ DRAFT: Last modified 6/9/2007 10:51 PM 14 D_report_flow_salt_trends_v1p2.doc DRAFT

Protection Agency STORET database3. Contra Costa Water District also maintains salinity monitoring data at its intakes dating back to 19444.

3 Data available at: http://www.epa.gov/storet/dbtop.html 4 Data available upon request. DRAFT: Last modified 6/9/2007 10:51 PM 15 D_report_flow_salt_trends_v1p2.doc DRAFT

2. Factors Influencing Salinity Intrusion in the Delta Salinity intrusion is the result of the interaction between tides and freshwater flow. Multiple factors can influence tides, freshwater flow, or the method by which the two interact. Many of these factors have experienced gradual, rapid, or periodic change, which results in changes to the salinity regime. Information on these factors has been gathered from TBI (1998), DPW (1931), Fox (1987a), Nichols et al (1986), Conomos (1979), and Knowles (2000). The reader is directed to these original sources for a more detailed discussion on each factor.

These factors are grouped into three categories (Table 3) and discussed individually and qualitatively to provide context for observed salinity variability, which is necessarily due to the cumulative impact of all factors.

Table 3 – Factors that impact salinity intrusion Many factors impact salinity intrusion. When the impact has a definite positive or negative effect on salinity, the general effect is also listed. Category • Factor o Effect on Delta Salinity Regional Climate • Precipitation patterns Change o Loss of April-July runoff may increase salinity in the spring, summer, and fall.

• Ocean Conditions o Adds periodic variability to precipitation (via mechanisms such as the El Niño/Southern Oscillation (ENSO) or Pacific Decadal Oscillation (PDO))

• Sea level rise o Generally increases salinity Physical Changes to • Deepening, widening, and straightening of Delta channels the Central Valley and o Generally increases salinity Delta Landscape • Separation of natural floodplains from valley rivers o Varies

• Reclamation of Delta Islands o Varies (impact on salinity depends on natural marsh vegetation, depth, and location)

• Creation of canals and channel “cuts” o Varies

5 Cappiella 6 Enright et al., 2004. DRAFT: Last modified 6/9/2007 10:51 PM 16 D_report_flow_salt_trends_v1p2.doc DRAFT

Category • Factor o Effect on Delta Salinity • Deepening (erosion) of Suisun Bay (since 1887)5 o Increases salinity in Suisun Bay and the western and central Delta6 Flow Management • Decreasing net Delta outflow (NDO) by increasing upstream diversions and increasing in-Delta diversions and exports o Increases salinity

• Increasing upstream storage capacity o Generally increases salinity. Reservoir releases possibly decreased salinity during some fall months. However, this is a minor decrease and occurred primarily prior to 1980.

2.1. Climatic variability

Climate impacts both the tides (through sea-level rise and oceanic conditions) and freshwater flow (through timing and quantity of precipitation and runoff ).

2.1.1. Regional Precipitation and Runoff Precipitation and runoff characterize the system hydrology, which is often quantified as “unimpaired” flow or runoff. Knowles (2000) determined that variability in freshwater flows accounts for the majority of the Bay’s salinity variability. The spatial distribution, seasonal timing, annual magnitude, decadal variability, and long-term trends of unimpaired flow all affect the hydrology and salinity transport in the Delta.

A common measure of unimpaired flow is the “Sacramento River Index” (SRI), which is the combined flow from the Sacramento River at Bend Bridge, Feather River inflow to Lake Oroville, Yuba River at Smartville, and the American River inflow to Folsom Lake. A second measure of unimpaired flow is the “Eight River Index” (8RI), which is the combined flow from the SRI plus the San Joaquin River tributaries, namely the Stanislaus River inflow to New Melones Lake, Tuolumne River inflow to New Don Pedro Reservoir, Merced River inflow to Lake McClure, and San Joaquin River inflow to Millerton Lake.

The total annual unimpaired flow of the upper Sacramento Basin7 for water years 1906 through 2006 exhibits substantial year-to-year variability with a strong decadal oscillation in the 5-year running average (Figure 4). On average, over the last 100 years, the total annual unimpaired SRI is increasing about 0.06% or 11 thousand-acre feet (TAF) each year. However, increased total annual unimpaired flow does not necessarily reduce salinity intrusion. Knowles (2000) illustrated that the seasonal timing of runoff can significantly alter salinity intrusion without any change to the total annual runoff.

7 http://cdec.water.ca.gov/cgi-progs/iodir/WSIHIST DRAFT: Last modified 6/9/2007 10:51 PM 17 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 4 – Unimpaired runoff in the upper Sacramento River basin Provided for hydrologic context when comparing conditions at different time periods, total annual unimpaired runoff varies significantly between years with a strong decadal trend. The 21-year average has been increasing since the late 1920s (from ~14 MAF to ~18 MAF).

Water Year Total 21 year Average 5 year Average Linear (Water Year Total) 45 Unimpaired Flow Estimates (DWR) Actual Flow Estimates (IEP-Dayflow)

40 Long-term Monitoring Data at Fixed Locations in Suisun Bay and Western Delta C&H Barge Travel DSM2 Historical

35 Selected years for distance to freshwater analysis Wet years 30 Dry Years

Upper Sacramento Basin Runoff [MAF] Upper Sacramento 5

0 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

The long-term trend in spring (April-July) runoff as a percentage of the total annual runoff has been decreasing (Supplemental Figure 1) by about 0.1% each year over the last 100 years. This effect is believed to be caused by climate change: as temperatures warm, more precipitation falls as rain instead of snow, and the snowpack tends to melt earlier. This leads to higher runoff during winter months, but lower runoff in spring or summer, resulting in the potential for greater salinity intrusion. Since the total annual runoff has been slightly increasing, the impact of this shift in precipitation patterns currently may not be appreciable.

Precipitation and runoff are influenced by regional events such as the Little Ice Age (about A.D. 1300 to A.D. 1850) and the Medieval Warm Period (about A.D. 800 to about A.D. 1300). For instance, during the Little Ice Age, the winter snowline in the Sierra was generally at a lower elevation and the spring and summer nighttime temperatures were significantly lower. This temperature pattern would allow the snowmelt to last further into the summer, providing a more uniform seasonal distribution of runoff such that significantly less salinity intrusion than occurs today would be expected. This expectation is borne out by paleosalinity studies in Suisun Marsh (see Section 3.1 ).

At shorter time scales, oceanic conditions such as the Pacific Decadal Oscillation (PDO) and El Niño/Southern Oscillation (ENSO) also impact precipitation and runoff patterns. DRAFT: Last modified 6/9/2007 10:51 PM 18 D_report_flow_salt_trends_v1p2.doc DRAFT

Runoff in the upper watershed is the primary factor that determines freshwater outflow from the Delta. Anthropogenic flow management (upstream diversions, reservoir operations, in-Delta diversions, and south of Delta exports) alters the amount and timing of flow from the upper watershed (see Section 2.3). Changes to the physical landscape further alter the amount and timing of flow (see Section 2.2).

2.1.2. Sea-level Rise After tectonic and erosive forces worked together to define the underlying geology of the Central Valley over 2 million years ago, sea level fluctuations during the repeated glacial advance and retreat during the Pleistocene epoch (extending from 2 million years ago to 15,000 years ago) resulted in deposition of alternating layers of marine and alluvial sediments in the Delta (Page, 1986, as presented in TBI, 1998).

A warming trend starting about 15,000 years ago ended the last glacial advance and triggered rapid sea-level rise. At the end of this period (known as the “Halocene Transgression”), approximately 6,000 years ago, sea-level had risen sufficiently to inundate the Delta at high tide (Atwater, 1979). Thus, the Delta, as we know it today, is rather young in geologic or evolutionary timescales.

During the last 6,000 years, sea-level has risen 0.02 inches per year, on average, at the Golden Gate. However, from 1920 to 2003, sea-level rose 0.6 inches, or 0.085 inches per year. This rise in sea-level would result in increased salinity intrusion into the Delta, if all other factors remain unchanged.

2.2. Physical changes to the Delta and Central Valley

Physical changes to the geometry of the system have a significant impact on the resulting hydrodynamics and salinity transport. This section is a brief introduction to the major physical changes in the Delta and Central Valley over the last 150 years. As many of the substantial physical changes were made prior to extensive flow and salinity measurement, a qualitative discussion is presented.

2.2.1. Deepening, Widening, and Straightening Channels (early 1900s-present) The lower Sacramento River was widened to 3,500 feet and straightened (creating Decker Island) around 1910. Early investigations into the increased salinity intrusion of the 1920s concluded that the channel modifications contributed to the increased salinity intrusion.

Progressive deepening of the shipping channels began in the early 1900s. Original depths were less than 10 feet; channels were gradually dredged to depths exceeding 30 feet, with maintenance dredging continuing today.

The net impact of these changes is to increase salinity intrusion. Deepening of the river increases propagation speed of the tidal wave, increasing salinity intrusion. Similarly,

DRAFT: Last modified 6/9/2007 10:51 PM 19 D_report_flow_salt_trends_v1p2.doc DRAFT

straightening the river provides a shorter path for the tidal wave, increasing salinity intrusion. Widening of the river increases the tidal prism, also resulting in further salinity intrusion.

2.2.2. Reclamation of marshland (1850-1920)

In the Central Valley Raising and strengthening natural levees in the Central Valley effectively disconnected the rivers from their floodplains, removing natural storage from the system. Natural floodplains captured large winter flows, gradually releasing them back into the main channels through the spring and summer, resulting in a more uniform flow into the Delta (reduced peak flow and increased low flow).

The native vegetation and large floodplains would have increased evapotranspiration. Combined with more groundwater accretion, these natural floodplains would have reduced total annual inflow into the Delta.

As discussed above (see Section 2.1.1), the seasonal distribution of flow controls the extent of salinity intrusion. Although disconnecting the natural floodplains probably increased the total annual inflow into the Delta, it also reduced the spring and summer inflow into the Delta, which is likely to have led to increased salinity intrusion.

In the Delta Reclamation of Delta marshland began around 1850. By 1920, almost all land within the legal Delta had been diked and drained for agriculture (DPW, 1931). Before armoring the levees and draining the marshes, the channels were likely to have been shallower and longer (more sinuous), which would have slowed the tidal wave propagation and reduced salinity intrusion.

The natural marsh surface would have increased the tidal prism. However, the shallow marsh depth and native vegetation would have slowed the tidal wave progression. The combined effect on salinity intrusion depends on the location and depth of the marsh, the native vegetation distribution, and the dendritic channels that were removed from the tidally active system.

2.2.3. Mining debris (1860-1914) Mining debris traveled down the Sacramento River, through the Delta and into the Bay from approximately 1860 through 1914. The mining debris may have contributed to extensive flooding in 1878 and 1881. Absent information on the timing and locations of the sediment waves, it is difficult to evaluate the effects of mining debris on Delta salinity regimes.

The cessation of hydraulic mining is one of many factors leading to the erosion of Suisun Bay (Cappiella et al, 1999) from 1877 to 1990. The Suisun Marsh Branch of the DWR determined that erosion of Suisun Bay (modeled as a uniform change in depth of 0.75

DRAFT: Last modified 6/9/2007 10:51 PM 20 D_report_flow_salt_trends_v1p2.doc DRAFT

meters) has increased salinity in Suisun Bay and the western Delta by as much as 20 percent (Enright, 2004).

2.3. Flow Management Regimes

Extensive local, state, and federal projects have been built to move water around the state, manipulating flow patterns above and throughout the Delta.

For clarity in the discussion that follows, definitions and discussions of actual flow and salinity, unimpaired flow and salinity, and natural flow and salinity, are given below.

Actual flow and salinity Actual flow and salinity refer to the flow and electrical conductivity, total dissolved solids concentration, or chloride concentration that occurred in the estuary. Actual conditions have been observed or measured at various times and locations; they are now measured at monitoring stations throughout the estuary. Actual data are also used to estimate flow and water quality conditions at other locations with the following tools: IEP Dayflow calculations, X2 relationships, the G-model, and DSM2 historical simulations. The use of these tools to estimate flow and water quality is necessarily dependent upon the Delta configuration to which they were calibrated. Use of these tools in hypothetical configurations (such as pre-levee conditions, flooding of islands, etc) is subject to un-quantified error.

Unimpaired flow and salinity Unimpaired flows are hypothetical flows that would have occurred in the absence of upstream diversions and storage, but with the existing Delta and tributary configuration. Unimpaired flows are estimated by the California Department of Water Resources (DWR) for the 24 basins of the Central Valley. (The Delta is one of the 24 basins.) Additionally, DWR estimates unimpaired in-Delta use and unimpaired net Delta outflow (NDO). Unimpaired NDO estimates can be used to estimate unimpaired water quality using a salinity-outflow relationship such as the X2 and G-model equations. Since unimpaired flows assume the existing Delta configuration, the use of these tools should not violate their basic assumptions. However, the results should be taken in context. Water quality based on unimpaired flows compared to water quality based on actual flows shows how flow management operations affect water quality. Water quality based on unimpaired flows cannot be considered natural.

Natural flow and salinity Natural flow and salinity reflect pre-European settler conditions, with a virgin landscape in both the Central Valley and the Delta, native vegetation, and no diversions or constructed storage. As discussed above, the natural landscape included natural storage on the flood plains and extensive Delta marsh. Estimation of natural flow requires assumptions regarding the pre-European landscape and vegetation throughout the Central Valley. Estimation of natural salinity requires development of new models to account for pre-European Delta geometry, incorporating the estimates DRAFT: Last modified 6/9/2007 10:51 PM 21 D_report_flow_salt_trends_v1p2.doc DRAFT

of natural flow. These assumptions induce an unknown level of error. For this reason, no attempt is made in this report to calculate natural flow or the resulting salinity. Instead, paleosalinity studies are examined for historical context.

2.3.1. Trends in Flow Management The number of irrigated acres in the Central Valley (Supplemental Figure 2) has been steadily increasingly since 1880, increasing the upstream diversions. Upstream diversions first became an issue with respect to Delta salinity around 1916 (after many of the major physical landscape changes discussed above), with the growth of the rice cultivation industry (see Sections 4.1 and 5.1.2). These early (pre-1943) diversions for irrigation were particularly problematic due to the seasonality of water availability and water use.

Diversions for agriculture typically start in the spring and continue through the early fall, when river flow is already low. Combined with the decrease in spring/summer flow due to the separation of natural flood basins (Section 2.2.2), early irrigation practices removed a significant fraction of the spring and summer river flow, resulting in increased salinity intrusion. The effects of these early diversions are discussed in Section 5.1.2.

The Department of Water Resources (DWR) estimated the impact of upstream diversions and exports on the salinity in the San Joaquin River at Antioch (Supplemental Figure 3), concluding that water with less than 350 mg/L chlorides would be present at Antioch approximately 88% of the time on average “naturally”, with availability decreasing to approximately 62% due to upstream diversions by 1940. (DWR, 1960)

Construction of upstream surface storage permitted further alteration of the natural patterns of Delta inflow. Supplemental Figure 4 and Supplemental Figure 5 show the extent and rapid rise of constructed reservoirs in the Central Valley. The probable effect of the increasing reservoir storage would have been to reduce inflow to the Delta in the wet season, while releases for upstream agricultural use would have done little to increase Delta inflow in the dry season.

Another mechanism for reducing Delta outflow and increasing salinity intrusion came on line when the Central Valley Project and the State Water Project began exporting Delta water in 1951 and 1968, respectively. Supplemental Figure 6 shows the rapid rise in exports from 1951 to the mid 1970s, and the sustained exporting of an average of more than 4 million acre feet per year since then.

Flow management has been altered by the implementation of water quality standards, which regulate salinity at key locations in the Bay-Delta during certain periods of the year. For instance, the listing of delta smelt as a threatened species under the Endangered Species Act in 1993, followed by the Bay-Delta Accord in 1994, and the adoption of a new water quality control plan by the State Water Rights Control Board in 1995 changed the amount and timing of reservoir releases and south of Delta exports.

DRAFT: Last modified 6/9/2007 10:51 PM 22 D_report_flow_salt_trends_v1p2.doc DRAFT

2.3.2. Impact of Flow Management on Net Delta Outflow The net impact of anthropogenic flow management can be illustrated by comparing estimated unimpaired NDO8 with actual NDO9. Since unimpaired flow estimates assume the existing Delta configuration (reclaimed islands, no natural upstream flood storage, etc.), comparison between these metrics reveals the net impact of flow management only; this analysis does not address the impact of physical changes to the landscape or sea level rise.

Average annual NDO from 1930 to 2003 is significantly reduced in the actual case (20.6 MAF) compared to the unimpaired case (28.7 MAF). While unimpaired NDO has been increasing from 1921 to 2003, actual NDO has been decreasing slightly. The natural increase in unimpaired NDO has been offset by the effects of increased water use (Figure 5). Figure 5 – Timeseries of Monthly Net Delta Outflow While unimpaired NDO (a) has been increasing from 1921 to 2003, actual NDO (b) has been decreasing slightly. The difference (c) is due to the cumulative impact of upstream diversions, reservoir operations, in-Delta diversions, and south of Delta exports. The natural increase in unimpaired NDO has been partially offsetting the effects of increased water use.

(a) Unimpaired NDO 20000

16000

12000

TAF 8000

4000

0 1920 1930 1940 1950 1960 1970 1980 1990 2000 (b) Actual NDO 20000 Monthly Average 5-year Average 16000 Linear Trend 12000

TAF 8000

4000

0 1920 1930 1940 1950 1960 1970 1980 1990 2000 (c) Unimpaired NDO - Actual NDO 4000 2000 Excess 0

TAF Deficit -2000 -4000 -6000 -8000

-10000 water use Increased p_ts_du.m 1920 1930 1940 1950 1960 1970 1980 1990 2000 15-May-2007 dms

8 Unimpaired NDO data provided by Messele Ejeta (DWR) on March 27, 2007. This data is an update to the February 1987 DWR publication “California Central Valley Unimpaired Flow Data”. 9 Actual NDO data is determined by the Dayflow program (http://www.iep.ca.gov/dayflow/index.html). DRAFT: Last modified 6/9/2007 10:51 PM 23 D_report_flow_salt_trends_v1p2.doc DRAFT

From 1921 to 2003, long-term average monthly unimpaired NDO has been fluctuating between 1,600 and 3,000 thousand acre-feet (TAF); however, the monthly peaks have been steadily increasing from around 8,000 TAF in the 1920s to nearly 18,000 TAF in the 1990s.

The seasonal distribution of unimpaired NDO reveals a hydrograph with flow peaking in May at 4,000 TAF on average. In contrast, actual NDO peaks in February at 3,000 TAF on average (Figure 6) because a significant portion of the spring snowmelt and runoff is captured by upstream reservoirs or diverted directly for irrigation.

Figure 6 – Monthly distribution of Net Delta Outflow from 1930 to 2003 Unimpaired NDO typically peaks in the spring. However, actual NDO peaks in the winter as a significant portion of the spring snowmelt and runoff is captured by upstream reservoirs or diverted directly for irrigation. The variability between years, represented by the vertical bars and ‘+’ marks, indicates the distribution is positively skewed, which means a relatively few number of years have excessively high flows. Seasonal Distribution of Net Delta Outflow 20000 Unimpaired (1930-2002) Actual (1930-1993) Actual (1994-2006)

15000 Exceedance max 10

10000 IQR 50 median 75 min

5000 Monthly Net Delta Outflow [TAF] Outflow Delta Net Monthly

0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

For all months except September and October, median unimpaired NDO is greater than actual NDO; in September, median actual NDO is slightly larger than median unimpaired NDO (470 TAF compared with 410 TAF). Since 1993, on average, September and October actual NDO are less than unimpaired NDO. This is further explored by segregating wet and dry years (Figure 7) and examining year-to-year trends for each month (Figure 8).

The impact of upstream reservoirs is illustrated by the NDO deficit for the ten wettest and the ten driest water years between 1945 (post-Shasta) and 2003 (Figure 7). During the ten wettest years, actual NDO exceeds unimpaired NDO during February and September,

DRAFT: Last modified 6/9/2007 10:51 PM 24 D_report_flow_salt_trends_v1p2.doc DRAFT

likely due to flood control releases. During the ten driest years, the NDO deficit is the greatest, particularly March through June, indicating flow management may have a greater impact during dry years.

Figure 7 – Monthly Distribution of Change in NDO for Wet and Dry Years The impact of upstream reservoirs is illustrated by the NDO deficit for (a) the ten wettest and (b) the ten driest water years between 1945 (post-Shasta) and 2003. During the ten wettest years, actual NDO exceeds unimpaired NDO during February and September, likely due to flood control releases. During the ten driest years, the NDO deficit is the greatest, particularly March through June, indicating flow management may have a greater impact during dry years. 200 actual-unimpaired / unimpaired

Exceedance 150 10 (a) Ten Wettest Water years (1945-2006) 25 [1952,1956,1958,1969,1974,1982,1983,1986,1995,1998] 100 IQR 50

median 75

90 50

Excess 0 Deficit

Percent Change Relative to Unimpaired -50

-100 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep 200 actual-unimpaired / unimpaired

150 (b) Ten Driest Water years (1945-2006) [1947,1976,1977,1987,1988,1990,1991,1992,1994,2001] 100

Excess 0 Deficit

Percent Change Relative to Unimpaired -50

-100 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep DRAFT: Last modified 6/9/2007 10:51 PM 25 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 8 – Year-to-Year Trends in NDO Excess and Deficit Increased water usage is increasing the deficit of actual NDO (relative to unimpaired NDO) in most months of the year. In July (and August, not shown), the deficit is actually decreasing, possibly due to reservoir releases for temperature control; however, actual flow remains below unimpaired flow. In September (and October, not shown), actual flows exceeded unimpaired flows, with an increasing trend from about 1945 to 1975. Since 1975, the percent change has shown a downward trend with a deficit (actual flow less than unimpaired flow) during most years since 1975. 100 Excess January

0 Deficit -100

50 Excess March 0

-50 Deficit -100

0 Deficit May

-50

-100 Actual and Unimpaired NDO 50 Excess July 0 Deficit -50

-100

200 Excess September 100 Percent Change between (Actual NDO – Unimpaired NDO) / (Unimpaired 0 Deficit -100

200 Excess November 100

0 Deficit -100 1930 1940 1950 1960 1970 1980 1990 2000

DRAFT: Last modified 6/9/2007 10:51 PM 26 D_report_flow_salt_trends_v1p2.doc DRAFT

Increased water usage is increasing the deficit of actual NDO (relative to unimpaired NDO) in most months of the year (Figure 8). In July and August (not shown), the deficit is decreasing, possibly due to increases in the unimpaired NDO during these months or reservoir releases for temperature control; however, actual flow remains less than unimpaired flow. In September and October (not shown), actual flows exceeded unimpaired flows, with an increasing trend from about 1945 to 1970. Since 1970, the percent change has shown a downward trend with a deficit (actual flow less than unimpaired flow) during most years since 1970.

DRAFT: Last modified 6/9/2007 10:51 PM 27 D_report_flow_salt_trends_v1p2.doc DRAFT

3. Historical Context – The last 3,000 years The Delta, as we know it, is relatively young in geologic or evolutionary timescales; the Delta became tidal approximately 6,000 years ago (Atwater, 1979) with the rapid sea- level rise of the Halocene Transgression. The effect of climate change is discussed qualitatively in Section 2.1.

Using biological and chemical markers in sediment cores and living trees, scientists have reconstructed flow and salinity records for areas in California as far back as 3,000 years ago in the studies referenced below. Although these studies cannot detect seasonal variability, they do offer a low resolution timeseries of flow and/or salinity to provide historical context.

3.1. Reconstructed Salinity in Northwestern Suisun Marsh (~1000 B.C – 2000 A.D)

Starratt (2001) reconstructed historical salinity variability at Rush Ranch, in the northwestern Suisun Marsh, over the last 3,000 years by examining diatoms from sediment cores. The taxa were classified according to their salinity preference: freshwater (< 2‰), freshwater and brackish water (0‰ to 30‰), brackish (2‰ to 30‰), brackish and marine (2‰ to > 30‰), and marine (> 30‰). Based on the composition of the diatom assemblages, Starrat (2001) identified centennial-scale salinity cycles.

Table 4 – Salinity Intervals over the last 3,000 years at Rush Ranch Salinity intervals determined from the diatom populations in a sediment core in northwestern Suisun Marsh.

Approximate Years Type of Interval a 1850 A.D. – present [not classified] 1250 A.D. – 1850 A.D. fresh 250 A.D. – 1250 A.D. brackish 500 B.C. – 250 A.D. fresh 1000 B.C. – 500 B.C. brackish a Classification according to Starratt (2001)

These results correspond well to other paleoclimatic reconstructions. The most recent broad-scale freshwater interval roughly corresponds to the Little Ice Age and the most recent brackish interval corresponds to the Medieval Warm Period.

Starratt notes that the post-1850 interval indicates an increase in the percentage of diatoms that prefer brackish and marine salinities compared to the last freshwater interval, indicating an increase in salinity during the last 150 years, in comparison to the previous 600 years.

During the post-1850 period, diatoms that prefer “marine” environments constitute as much as 50% of the total diatom population, a percentage that is at or above that of any

DRAFT: Last modified 6/9/2007 10:51 PM 28 D_report_flow_salt_trends_v1p2.doc DRAFT

other period. During the most recent years, “freshwater” assemblages constitute about 20% of the total population, a percentage that is only about 10% higher than the most recent brackish interval from 250 A.D. to 1250 A.D.

3.2. Reconstructed Unimpaired Flow in the Sacramento Valley (900 A.D. – 1976 A.D.)

Annual river flow estimates from tree ring analysis (Meko, 2001)10 are used here to provide context for the hydrological conditions present during eyewitness observations of salinity in the late 1700s through the early 1900s. Comparison of the reconstructed flows with measured flows for the period of overlap indicates that the tree-ring data provides a reasonable surrogate for total annual flow (Supplemental Figure 7).

Estimation of salinity from these annual flow estimates is difficult. First, the seasonal distribution of hydrology is critical in determining salinity variability. Two years with the same total annual flow could have significantly different salinity intrusion due to the timing of the flow (Knowles, 2000). Second, drastic changes to the landscape, including disconnecting the rivers from their floodplains, deepening of rivers, and reclamation of islands, alter the hydrodynamic response to freshwater flow. The interaction of tides with freshwater flow over different landscapes will result in different salinity intrusion for the same amount of freshwater inflow to the Delta.

In fact, many precipitation (or unimpaired runoff) estimates, including this reconstruction, indicate that the last 100 years are perhaps wetter than the previous 700 years. Given the same seasonal distribution of runoff and the same physical landscape, this would imply the last 100 years are fresher than the previous 500 years. However, as shown in the paleosalinity record (Section 3.1), the last 150 years have been more saline than the previous 600 years.

The reconstructed flow data are not used here to estimate salinity intrusion. Instead, the reconstructed total annual runoff is used as an indicator of the hydrology at the time of eyewitness observations before measured flow data were available. To interpret the salinity observations, we examine years with similar unimpaired runoff in Section 4.1.

10 Data Source: Meko, D.M.. 2006. Sacramento River Annual Flow Reconstruction. International Tree-Ring Data Bank. IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series #2006-105. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA. DRAFT: Last modified 6/9/2007 10:51 PM 29 D_report_flow_salt_trends_v1p2.doc DRAFT

4. Qualitative Salinity Observations (late 1700s – early 1900s)

Definition of qualitative terms Historical documents often describe salinity in qualitative terms, such as “brackish”, “fresh”, or “sweet”. To compare historical accounts with present-day salinity measurements, we must assign a relative numeric value to the qualitative description.

The earliest written accounts of explorers were often concerned with adequate drinking water. Testimony from the 1922 legal proceeding regarding salinity at Antioch (Town of Antioch v. Williams Irrigation District, 188 Cal. 451) indicates early settlers required water with less than 100 milligrams per liter (mg/L) of chloride (approximately 525 µS/cm EC) for municipal use.11 Similarly, DPW (1931) indicates that a “noticeable” level of salinity is 100 mg/L chloride. The current secondary water quality standard for municipal and industrial use is 250 mg/L chloride (1,000 µS/cm EC) (SWRCB, 2006). The more conservative value is used in this report; the demarcation between “fresh” or “sweet” water and “brackish” water is taken as 250 mg/L chloride (1000 µS/cm EC).

4.1. Observations from Early Explorers

Observations from early explorers are often one-time observations (or perhaps observations over a single month or season). Table 5 summarizes some reported observations. These observations were made from a time period when many of the natural floodplains of the Central Valley were still connected to the river system, providing natural storage. Therefore, total annual flow (SRI from the Meko reconstruction, described in Section 3.2) from both the current and previous year are given for context.

Table 5 – Qualitative Salinity Observations from Early Explorers

Date Location Description Year / Observer Reference Reconstructed Flow [MAF] 1775 near the sweet, the same 1774 / 25.48 Canizares Britton, 1987 August Sacramento- as in a lake 1775 / 18.70 in Fox, 1987b San Joaquin confluence 1776 near Antioch very clear, fresh, 1775 / 18.70 Font Britton, 1987 April (San Joaquin sweet, and good 1776 / 9.07 in Fox, 1987b River)

11 Supplement to Respondent’s Answering Brief, p. 10. DRAFT: Last modified 6/9/2007 10:51 PM 30 D_report_flow_salt_trends_v1p2.doc DRAFT

Date Location Description Year / Observer Reference Reconstructed Flow [MAF] 1776 near the sweet 1775 / 18.70 Canizares Britton, 1987 September Sacramento- 1776 / 9.07 in Fox, 1987b San Joaquin confluence 1811 near the sweet 1810 / 19.50 Abella Britton, 1987 October Sacramento- 1811 / 22.76 in Fox, 1987b San Joaquin confluence 1796 unknown salinity 1795 / 5.90 Hermengildo Cook, 1960 “far upstream” at 1796 / 9.95 Sal in TBI, 1998 high tide 1841 Three Mile brackish 1840 – 15.67 Wilkes Britton, 1987 August Slough north (undrinkable) 1841 – 5.56 in Fox 1987b of Emmaton

4.1.1. Fresh conditions Early explorers observed “sweet” water near the Sacramento-San Joaquin confluence in August and October during relatively wet years (total annual runoff greater than 18 MAF). In September of 1776, a relatively dry year (total annual runoff of only 9 MAF), sweet water was also observed near the confluence. This suggests that either the reconstructed flow estimate for 1776 is low, or the natural floodplains and groundwater storage provided additional flow as the flood flows from the previous 5 years (all with total annual runoff above 18 MAF) reentered the channels.

Under current conditions, monthly average salinity at Collinsville (near the confluence) is less than 1,000 µS/cm EC (the interpretation of the “sweet” threshold for drinking water) from August through October when the total annual unimpaired runoff is greater than about 20 to 25 MAF (Figure 9). This indicates either the interpretation of “sweet” is set too fresh, or Collinsville is currently saltier than it was in the late 18th and early 19th centuries.

Adjusting the definition of “sweet” to 1,300 µS/cm EC and excluding the most recent years (1994-present), monthly average salinity at Collinsville is generally less than 1,300 µS/cm EC from August through October when the total annual unimpaired runoff is greater than 16 MAF. This corresponds well to the above observations, but highlights a recent increase in salinity at Collinsville during moderately wet years (with total annual runoff between 14 and 26 MAF).

DRAFT: Last modified 6/9/2007 10:51 PM 31 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 9 – Salinity at Collinsville as a function of Total Annual Sacramento Unimpaired Runoff (SRI) Salinity in the post-ESA years (post-1993, illustrated as black circles) indicates an increase in September and October salinity relative to pre-ESA water years with similar unimpaired hydrology. In comparison to anecdotal evidence, the post-1993 salinity at Collinsville during moderate flow (15- 25 MAF) is greater than observed salinity in the late 18th century during similar flow. Salinity at Collinsville (1965-2005) 3000 August September 2500 October S/cm]

μ 2000

1500

1000 Monthly Average Salinity [ Salinity Average Monthly 500

0 5 10 15 20 25 30 35 40 Total Annual Sacramento River Index [MAF] p_saldata.m 29-Apr-2007 dms

In 1997, 1999, 2000, 2003, and 2004, the water at Collinsville in October would not be considered “sweet” even under the relaxed criterion, indicating that October salinity is now greater than it was in 1811.

4.1.2. Brackish Conditions The qualitative observations of high salinity intrusion are less specific about location. However, some of these observations have been interpreted by others (Cook, 1960, in TBI, 1998; Fox, 1987b) to indicate intrusion as far upstream as Rio Vista.

The drought years of 1976-1977 and 1987-1992 can be compared to these observations. During the 1976-1977 drought, daily average salinity at Rio Vista exceeded 1,000 µS/cm for approximately six months of the year. During the 1987-1992 drought, salinity at Rio Vista at high tide often exceeded 2,000 µS/cm, particularly during the fall. This is consistent with the observations in 1796 and 1841, which report salt water extending into the western Delta.

DRAFT: Last modified 6/9/2007 10:51 PM 32 D_report_flow_salt_trends_v1p2.doc DRAFT

4.1.3. Conclusions In summary, interpretation of the above observations in the context of the reconstructed Sacramento River flows does not support the hypothesis that the present-day Delta is managed as a freshwater system in comparison with its historical salinity regime.

This analysis also highlights that salinity has increased in September and October near the confluence in recent years (post-1994).

4.2. Observations from early settlers in the Western Delta

Observations from early settlers in the western Delta provide a more complete description of salinity in the late 1800s and early 1900s than the limited observations from explorers. Since residents typically stayed in generally the same location for years, these observations often provide a more precise spatial location with more information about temporal variability as well.

4.2.1. Town of Antioch Injunction on Upstream Diverters In 1920, the Town of Antioch filed a lawsuit against upstream irrigation districts alleging that the upstream diversions were causing increased salinity intrusion at Antioch. The court decision, legal briefings, and petitions provide salinity observations from a variety of witnesses.

Although anecdotal testimony summarized in legal briefs is far from scientific evidence, a fair reading of the briefs from both parties lends support to the hypothesis that the Bay- Delta was fresher in the late 1800s and early 1900s than it is today. The record does not support a conclusion that the Delta was saltier than it is today.

Case History On July 2, 1920, the Town of Antioch filed suit in the Superior Court of the State of California (hereafter the “Antioch Case”) against upstream diverters on the Sacramento River and Yuba River. A hearing for a temporary injunction began on July 26, 1920, and lasted approximately three months. On January 7, 1921, Judge A.F. St. Sure granted a temporary injunction, restraining the defendants “from diverting so much water from the said Sacramento River and its tributaries, to non-riparian lands, that the amount of water flowing past the City of Sacramento, in the County of Sacramento, State of California, shall be less than 3500 cubic feet per second”. (Town of Antioch v. Williams Irrigation District, Supplement to Appellants’ Opening Brief, p. 13)

The defendants appealed to the Supreme Court of the State of California, who issued their opinion on March 23, 1922. The Supreme Court reversed the lower court and withdrew the injunction, declaring “[i]t is evident from all these considerations that to allow an appropriator of fresh water near the outlet of these two rivers to stop diversions above so as to maintain sufficient volume in the stream to hold the tide water below his place of diversion and secure him fresh water from the stream at that point, under the

DRAFT: Last modified 6/9/2007 10:51 PM 33 D_report_flow_salt_trends_v1p2.doc DRAFT

circumstances existing in this state, would be extremely unreasonable and unjust to the inhabitants of the valleys above and highly detrimental to the public interests besides”.

Contrary to assertions in “Envisioning Futures” (PPIC, 2007), the Supreme Court did not find that there had been “substantial salinity incursions in the era before significant upstream irrigation”. The Supreme Court did not make any comment whatsoever on the evidence of salinity intrusion prior to the upstream diversions in question. Instead, the Court indicated that their decision was based on a “policy of our law, which undoubtedly favors in every possible manner the use of the waters of the streams for the purpose of irrigating the lands of the state to render them fertile and productive, and discourages and forbids every kind of unnecessary waste thereof”. (Town of Antioch v. Williams Irrigation District (1922) 188 Cal. 451) They concluded that allowing 3,500 cubic feet per second (cfs) to “waste” into the bay to provide less than 1 cfs of adequate quality water for the Town of Antioch would constitute unreasonable use of California’s limited supply of water.

The court did not base their decision on historical evidence of salinity at Antioch, which indicates that Antioch was able to divert freshwater at low tide at all times from 1866 to 1918 except possibly for a few months in 1870.

4.2.2. Salinity at Antioch – then and now Although quantitative analysis of flow and salinity was limited prior to 1920, transcripts and legal briefs from the Antioch Case provide an anecdotal history of salinity in the area. As discussed in the introduction, salinity varies substantially with the tide, generally the greatest salinity is observed near high tide and the lowest salinity is observed at low tide.

Testimony from multiple witnesses largely indicates that fresh water was always available in the San Joaquin River at Antioch at low tide, until the years just prior to 1920. Antioch’s position was that fresh water was always available before upstream development. In cross examination of Antioch’s witnesses, the upstream irrigators demonstrated that brackish conditions did occasionally exist at high tide. To remove any possible exaggeration of how fresh the water was before upstream diversions, the following anecdotal summary of salinity information is taken from legal briefs filed by the Appellants, the upstream diverters.

Evidence was provided about the operation of pumping plants along the San Joaquin River at Antioch for domestic water supply and the quality of water obtained from the pumping plants:

1866-1878: Mr. Dodge ran a pumping/delivery operation. • Dodge pumped water into a small earthen reservoir and then hauled the water to residents in a wagon. • Cary Howard testified that while he was living in Antioch (1867-1876), the water became brackish one or two years in the fall, when they had to drive into the country to get water. This was likely during the drought of

DRAFT: Last modified 6/9/2007 10:51 PM 34 D_report_flow_salt_trends_v1p2.doc DRAFT

1870-71.

1878-1880: Mr. Dahnken bought and operated the Dodge operation. • Dahnken testified that the water became brackish at high tide every year in the late summer, and remained brackish at high tide until it rained “in the mountains”.

1880-1903: Belshaw Company provided water • Dahnken testified that Belshaw Company pumped only at low tide. Belshaw had two 50,000 gallon tanks, which allowed the company flexibility in the pumping schedule.

1903-1920+: Municipal Plant • William E. Meek (resident since 1910) testified the water is brackish at high tide every year, for some months in the year. • James P. Taylor testified that for at least the last 5 years, insufficient storage required the plant to pump nearly 24 hours per day, regardless of tidal phase. • Dr. J.W. DeWitt testified that during October of most years between 1897 and 1918, the water was too brackish to drink. Even when the city only pumped at low tide, the water was occasionally so brackish that it would be harmful to irrigate the lawns.

In summary, water was known to be brackish at high tide during certain time periods. However, at least until around 1915 (when the plant started pumping continuously, regardless of tidal stage), water at Antioch was apparently not brackish at low tide.

Recent Salinity at Antioch In recent years, salinity in the San Joaquin River at Antioch is highly variable both tidally and seasonally. For the following analysis, long-term monitoring data was obtained for the period May 1, 1983 through September 30, 200212.

12 Data Source: Interagency Ecological Program, HEC-DSS Time-Series Databases. Station RSAN007. Agency: DWR-ESO-D1485C. Measurement: 1-hour EC. Time Range: May 1, 1983 through September 30, 2002 DRAFT: Last modified 6/9/2007 10:51 PM 35 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 10 – Salinity on the San Joaquin River at Antioch (Water Year 2000, above normal) Fresh water (less than 1,000 µS/cm) was available at Antioch during low tide for only about eight months of the year during water year 2000 (an above normal water year). No freshwater was available at any stage of the tide during five months of the year. 10000 Hourly Data 9000 Daily Average 4-hr Low Tide 8000

7000

6000

S/cm] 5000 μ

EC [ EC 4000

3000

2000

1000

0 Oct99 Jan00 Apr00 Jul00 Oct00

Salinity variability for water year 2000, an above normal water year based on the Sacramento River Index, indicates that the city could pump water all day for about four and half months (early February through mid-June) and could pump for a portion of the day at low tide for another two and half months (mid-June though September) (Figure 10). For the remaining five months, water at the city intakes exceeded 1,000 µS/cm EC for the entire day, regardless of tidal stage.

The seasonal distribution for low tide salinity (daily 16th percentile, or approximately 4 hours each day) is shown in Figure 11. If Antioch were to always pump at low tide during 1983-2002 but only when salinity is 1,000 µS/cm EC, they would have to stop pumping from early June through early February in the driest 25% of the years. Antioch would have to stop pumping from late August to late-December in 50% of the years; i.e. they would have an average of eight months of low-tide pumping per year, compared to the pre-1915 average of twelve months per year.

DRAFT: Last modified 6/9/2007 10:51 PM 36 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 11 – Seasonal Distribution of Low Tide Salinity at Antioch 1983 to 2002 (the 16th Percentile of each day) Low tide salinity (salinity during the freshest 4-hours of each day) varies both seasonally and year-to- year. The seasonal distribution for water years 1983 to 2002 indicates that on average (in 50% of the water years) low tide salinity exceeds 1,000 µS/cm EC from late-August through December. During the driest 25% of the years (5 out of 20 years), low tide salinity exceeds 1,000 µS/cm EC from June through January, leaving Antioch with no fresh water for eight months of the year. 6000 Driest 10% of the water years (2 of 20 years) Driest 25% of the water years (5 of 20 years) 5000

Median of all water years (10 of 20 years) 4000

Wettest 25% of the water years S/cm]

μ 3000 (5 of 20 years) EC [

2000

1000

0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

4.2.3. Salinity at Kentucky Point on Twitchell Island – then and now The appellants in the Antioch Case, representing the upstream diverters, identified one resident of Twitchell Island who reported the water at Kentucky Landing was brackish on “one or two occasions” between 1870 and 1875 during August and September. During this time, he had to travel up the San Joaquin River to Seven Mile Slough (the eastern boundary of Twitchell Island) and sailed as far as the mouth of the Mokelumne River (approximately 2 miles further up the San Joaquin River than the Seven Mile Slough junction) to obtain fresh drinking water.

The reconstructed unimpaired flows for 1870 and 1871 indicate that the SRI was approximately 11 and 10 MAF, respectively. Monitoring data from 1981 (11.1 MAF SRI) shows similar salinity intrusion as described by the Twitchell Island resident, assuming 1,000 µS/cm EC can be taken as the salinity limit for fresh drinking water in the interpretation of the early settler’s accounts. Salinity along the San Joaquin River at Bradford Island (about 1.5 miles upstream of Three Mile Slough) exceeded 1,000 µS/cm EC (about 250 mg/L Cl) during August and September. During the same time period, salinity was around 400 µS/cm EC (about 64 mg/L Cl) approximately 5 miles upstream on the San Joaquin River between Seven Mile Slough and the Mokelumne River. This

DRAFT: Last modified 6/9/2007 10:51 PM 37 D_report_flow_salt_trends_v1p2.doc DRAFT

comparison indicates that the extent of salinity intrusion in 1981 is similar to that which occurred in 1870 and 1871.

4.2.4. Conclusions The window when Antioch is able to pump water with salinity less than 1,000 µS/cm EC, has substantially narrowed in the last 125 years. Antioch was apparently able to pump fresh water at low tide year round in the late 1800s, with the possible exception of the fall season during one or two dry years. During 10 of the 20 years between 1983 and 2002 salinity was less than 1,000 µS/cm EC at low tide for only about eight months of the year. During the driest 5 years (out of 20 years), salinity was less than 1,000 µS/cm for only about four months per year (meaning that no fresh water was available at any time of the day for about eight months of the year).

Salinity intrusion up the San Joaquin River during the dry years of 1870 and 1871 as described by a Twitchell Island resident is consistent with salinity intrusion in 1981 (a similar water year). There is no evidence that salinity intrusion during the drought of 1870-71 was more extensive than salinity intrusion during similar water years in the current salinity regime.

DRAFT: Last modified 6/9/2007 10:51 PM 38 D_report_flow_salt_trends_v1p2.doc DRAFT

5. Quantitative Salinity Observations (early 1900s – present)

Salinity in the Bay-Delta varies both in space and time. To characterize this variability, we examine time series of salinity fluctuations using two methods. In Section 5.1, we examine the movement of a specific salinity value up and down the estuary, studying the location of a specific salinity value as a function of time. In Section 5.2, focus on a specific location, evaluating the change in salinity at a specific location as a function of time.

5.1. Fluctuation or Movement of the Spatial Salinity Distribution

In this section, we summarize the movement of three metrics for salinity intrusion. In Section 5.1.1, the distance to water with 50 mg/L chloride concentration as measured in the early 1900s is compared with the distance to the same quality water today as determined by both monitoring and modeling. Section 5.1.2 reviews the annual maximum salinity intrusion as defined by the location of water with 1,000 mg/L chloride. Section 5.1.3 uses a common model to explore the impact of flow management on the location of X2, water with a near-bed salinity of 2 psu.

5.1.1. Distance to Fresh Water (50 mg/L chloride) The California & Hawaiian Sugar Refining Corporation in Crockett (C&H) obtained its freshwater supply from barges traveling up the Sacramento and San Joaquin Rivers, generally twice a day beginning in 1905 (DPW, 1931). This is the most detailed salinity record prior to the intensive salinity investigation by the State, which started in 1920.

The following analysis is a comparison of the fluctuating salinity observations of C&H with more recent monitoring data and modeling results to determine how the current, managed salinity regime compares to the regime of the early 1900s.

Data Sources and Methods

C&H C&H operations required water with less than 50 mg/L chloride concentration (one third the concentration of the strictest Delta drinking water quality standard today).

The C&H barges typically traveled up the river on flood tide and returned downstream on ebb tide. Since the maximum daily salinity for a given location in the river channels typically occurs about one to two hours after high slack tide, the distance traveled by the C&H barges represents the daily maximum distance to fresh water. DPW Bulletin 27 charts the C&H data, showing the monthly minimum, average, and maximum distance traveled (Supplemental Figure 11 and Supplemental Figure 12). For this comparison, monthly averages of the daily maximum distances were extracted from Plate IV.

DRAFT: Last modified 6/9/2007 10:51 PM 39 D_report_flow_salt_trends_v1p2.doc DRAFT

Starting in 1920, C&H obtained a portion of its water supply from Marin County. During this period, for a portion of each year, the distance to fresh water is unknown. By 1918, diversions for rice fields began altering the timing and quantity of flows upstream of the Delta (DPW, 1931). Additionally, the upstream storage capacity increased approximately 2.5 MAF from 1920 to 1930. For these reasons, the analysis focuses on the C&H data from 1908 to 1917.

Monitoring Data at Fixed Stations Long-term monitoring of electrical conductivity (EC) at multiple stations within the Bay and Delta began around 1964. For this analysis, publicly available data was downloaded from the Interagency Ecological Program data repository (Table 6). The monitoring data used in this report are daily average values.

Table 6 – Sources for Long-Term Monitoring Data, used in Distance to Fresh water Analysis All data was downloaded from the Interagency Ecological Program Data Repository (http://www.iep.ca.gov/dss/) Location Station Source Data Selby RSAC045 USGS-BAY Historical Martinez RSAC054 CDEC Real-time Benicia Bridge RSAC056 USBR-CVO Historical Port Chicago RSAC064 USBR-CVO Historical Mallard RSAC075 CDEC Real-time Pittsburg RSAC077 USBR-CVO Historical Collinsville RSAC081 USBR-CVO Historical Emmaton RSAC092 USBR-CVO Historical Rio Vista RSAC101 USBR-CVO Historical DWR-ESO-D1485C Historical Georgiana Slough RSAC123 DWR-CD-SURFWATER Historical Greens Landing RSAC139 USBR-CVO Historical Antioch RSAN008 USBR-CVO Historical Jersey Pont RSAN018 USBR-CVO Historical Bradford Point RSAN024 USBR-CVO Historical San Andreas Landing RSAN032 USBR-CVO Historical

For comparison with the C&H distance to fresh water, the location of the 350 µS/cm EC (approximately 50 mg/L chloride) isohaline was estimated by interpolation between the fixed station data. The fixed stations are separated by many miles, and irregular data gaps exist for some stations resulting in even further distances between valid data points. Limited spatial resolution contributes error in the interpolation; therefore, DSM2 modeling was explored as a supplemental method.

DSM2 Historical Simulation The DSM2 historical simulation (1989-2006) was used to provide estimates of Delta hydrodynamics and water quality to complement limited field data. Since the node spacing of the DSM2 network is denser than the spacing of the monitoring stations, DSM2 better resolves the spatial distribution of salinity in the Delta.

DSM2 results include daily average EC at each node along the lower Sacramento and San Joaquin Rivers. For comparison with C&H and fixed monitoring station data, DSM2 results were post-processed to determine the location of the 350 µS/cm EC isohaline,

DRAFT: Last modified 6/9/2007 10:51 PM 40 D_report_flow_salt_trends_v1p2.doc DRAFT

using linear interpolation between nodes. For comparison of bulk statistics with the C&H years, DSM2 results were limited to water years 1995-2006 to avoid the drought in the early 1990s, which was significantly drier than 1908-1918.

Specific Water Years Since the time period of the datasets do not overlap, upstream hydrology was examined to select specific water years for comparison. Two wet years (1911 and 1916) and two dry years (1913 and 1918) were selected from the C&H time period. To avoid biasing the results towards showing a fresher Delta under current conditions, water years representing current conditions were chosen to have greater total annual flow than the corresponding C&H water years. Selected recent wet years for comparison were 1969 and 1995; selected recent dry years were 1968 and 2002 (Supplemental Figure 13).

Daily Maximum vs. Daily Average C&H data represents the maximum daily salinity at a given location, while the current conditions are characterized by the average daily salinity. Estimates of the distance that must be traveled to reach fresh water under current conditions thus underestimate the actual distance.

Distance Traveled The C&H barges traveled up the San Joaquin River from 1908 through 1918 (DPW, 1931). The distances calculated for current conditions are on the Sacramento River because interpolation of monitoring data on the San Joaquin River is problematic. In the present day, distance to fresh water up the San Joaquin River is further than the distance up the Sacramento River, so this approach will also serve to underestimate the actual distance that C&H barges would have had to travel under current conditions. For fall periods, current conditions on the San Joaquin River side of the Delta show that salinity is so high that only a small area near the mouth of the Mokelumne River meets the C&H criterion of less than 50 mg/L chlorides.

Results and Discussion Comparison of data for both wet and dry years indicate that if barges were still traveling up into the Delta for fresh water, on average they would have to travel up to 19 miles farther today than in the early 1900s.

DRAFT: Last modified 6/9/2007 10:51 PM 41 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 12 – Distance to Fresh water for similar decades Distance above Crockett to water with less than 50 mg/L chlorides is calculated from long-term fixed monitoring data for two decades with similar hydrology to the decade of C&H barge travel shown below. The shading represents the amount of fresh water below the confluence of the Sacramento and San Joaquin rivers at Collinsville. 50 C&H Data 40 1 30 20 Saltier 10

0 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919

50 2 Monitoring Data 40 30 kett to 50 mg/L chlorides 20 Saltier 10

0 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976

50 Monitoring Data 40 30 20 Saltier 10 Distance in miles above Croc

0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

1 During August and September 1918, average water quality obtained by C&H exceeded 110 mg/L chlorides. 2 Salinity intrusion during 1966 is likely an overestimate to inadequate spatial coverage of monitoring stations. Salinity at Emmaton (28 miles) exceeded 3,000 µS/cm; the next data station is above Courland (58 miles) had a salinity near 300 µS/cm (35 mg/L chlorides). Interpolation across 30 miles is inaccurate.

Select Wet Years Salinity patterns during both C&H wet years, 1911 (26 MAF total annual runoff) and 1916 (24 MAF total annual runoff), are similar, with fresh water west of Martinez for about 4-5 months (Figure 13). In contrast, during recent wet years 1969 (27 MAF total annual runoff) and 1995 (35 MAF total annual runoff) fresh water is west of Martinez for about 6 weeks.

DRAFT: Last modified 6/9/2007 10:51 PM 42 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 13 – Distance to Fresh Water in Select Wet Years If barges were still traveling up the Sacramento River to find fresh water (50mg/L chlorides), they would have to travel farther during the fall, spring, and summer than the C&H barges traveled during similar wet water years. In 1916, fresh water retreated upstream about one month earlier than the retreat in 1911, possibly influenced by the increase in upstream diversions during the five year period. In recent years with even greater unimpaired runoff, fresh water retreats upstream two to three months earlier than 1916. Additionally, fresh water reaches Martinez for a much shorter period of time, generally less than one month in recent years compared to four to five months during 1916 and 1911, respectively.

Distance to Freshwater - Wet Years San Andreas Landing 40 Saltier conditions 1911 (26 MAF ) C&H 1916 (24 MAF ) C&H Bradford Point 35 1969 (27 MAF ) Monitoring Data Rio Vista Jersey Point 1998 (31 MAF ) Monitoring Data 30 Emmaton Antioch

] to 50 mg/L chlorides 25 Saltier conditions

Collinsville 20 Earlier Pittsburg salinity Mallard 15 intrusion

Port Chicago 10

Martinez 5 Nearby Location

Distance [miles above Crockett 0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct p_barge1.m 30-May-2007

Select Dry Years The difference between C&H barge travel distance in the dry years of 1913 and 1918, especially the extra 10 miles of distance to fresh water traveled in August and September of 1918 (Figure 14), may be partially explained by the growth of the rice cultivation industry, which began around 1912 (DPW, 1931) and increased upstream diversions when river flows were already seasonally low. The most visible difference between distance traveled in the dry years of the early 1900s and the more recent dry years is the substantial increase in distance traveled from April through June, indicating a much fresher spring during the dry years of the early 1900s, before the large capacity reservoirs were built to capture the spring runoff.

DRAFT: Last modified 6/9/2007 10:51 PM 43 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 14 – Distance to Fresh water in Select Dry or Below Normal Years During dry and below normal water years, if traveling today, barges would have to travel further during spring, summer and fall than C&H traveled in the early 20th century. Although C&H recorded relatively high salinity (greater than 110 mg/L chlorides) above Bradford Point on the San Joaquin in 1918, which is greater than observed salinity on the Sacramento River near Rio Vista in similar water years, the salinity intrusion in 1918 was significantly increased by unregulated agricultural diversions, especially due to recent introduction of rice cultivation.

Distance to Freshwater - Dry/BN Years San Andreas Landing 40

Bradford Point 35 * Rio Vista Jersey Point 30 Emmaton Antioch

] to 50 mg/L chlorides 25

Collinsville 20 Pittsburg Mallard 15

Port Chicago 10 1912 (11 MAF ) C&H 1918 (11 MAF ) C&H Martinez 5 1968 (14 MAF ) Monitoring Data 2002 (15 MAF ) Monitoring Data Nearby Location 0 Distance [miles above Crockett Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct p_barge1.m 30-May-2007 * During August and September 1918, average water quality obtained by C&H exceeded 110 mg/L chlorides. All years The seasonal distribution of distance traveled for each dataset (Figure 15) illustrates the seasonal fluctuations of the salt field as well as the variability between years for each month. During the early 1900s, the median distance traveled to fresh water from March, through June is less then 8 miles upstream of Crockett, between Martinez and Port Chicago. During March, the barges had fresh water within two miles of Crockett 25% of the time. The median distance traveled by C&H in September and October is about 25 miles, between Collinsville and Emmaton.

In contrast, the median distance traveled in September and October in more recent years, according to both the monitoring data and DSM2 modeling, is between 30 and 35 miles, 5 to 10 miles farther than the C&H barges. Likewise, the median distance to fresh water in the spring, using 40 years of monitoring data, is between 18 (in March) and 27 (in June) miles upstream of Crockett, 10 to 19 miles farther than the C&H barges traveled. Limiting the analysis to the wetter period of the DSM2 historical simulation (1995-2006), the median distance traveled in the spring ranges from 9 (in March) to 24 (in June) miles above Crockett, 1 to 16 miles farther than the C&H barges traveled.

DRAFT: Last modified 6/9/2007 10:51 PM 44 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 15 – Average Monthly Distance to Fresh Water The seasonal distribution of distance to fresh water (less than 50 mg/L chlorides) as measured by C&H is significantly shorter than the distance to fresh water as calculated by fixed station monitoring data during every month, indicating an increase in salinity in recent years relative to the early 20th century. The largest increase is evident in the spring and summer; the average distance traveled in June increased over 15 miles (from Martinez nearly to Collinsville). The dramatic increase in distance traveled in the spring and summer is partially caused by a temporal shift in seasonal salinity intrusion. From 1907 to 1917, salinity intrusion typically began in July; however, salinity intrusion during 1995 to 2005 (with similar hydrology) began in April – three months earlier than during the early 20th century. During the last ten years, distance has substantially increased in the fall, indicating saltier conditions, with a corresponding slight decrease in spring and summer. However, spring and summer salinity is still much greater than the early 20th century.

San Andreas Landing 40 Saltier fall 1995-2005 C&H Data (1907-1918) Monitoring Data (1966-1975) Bradford Point 35 Monitoring Data (1995-2005) Rio Vista Jersey Point 30 Emmaton Antioch 25 ] to 50 mg/L chlorides

Collinsville 20 Pittsburg Mallard 15 Earlier Fresher fall salinity 1965-75 Port Chicago 10 intrusion

Martinez 5 Saltier conditions

Distance [miles above Crockett 0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

Interestingly, the argument that current flow management practices limit the fluctuating nature of the salt field is verified in this analysis, with a twist. On average, the annual range of distance to fresh water up the Sacramento River is limited to approximately 15 miles today, while the early 1900s data indicate an annual range of 20 miles. However, this analysis indicates that current management does not limit salinity intrusion; rather, operations limit fresh water from reaching as far downstream as it did in the early 1900s.

This is consistent with the exceedance probabilities for distance traveled up the Sacramento River (Figure 16) for different salinity levels. During the C&H years (1908- 1917), barges had to travel above the confluence of the Sacramento and San Joaquin Rivers (approximately 22 miles above Crockett) only about 35 percent of the time for less than 350 µS/cm EC water. In contrast, during the relatively wetter DSM2 historical time period (1995-2006), barges would have had to travel above the confluence approximately 60-65 percent of the time for 350 µS/cm EC water.

DRAFT: Last modified 6/9/2007 10:51 PM 45 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 16 – Distance along the Sacramento River to Specific Salinity Values Comparison between the exceedance distribution for travel distance of the C&H barges (1907 to 1917) compared to exceedance distributions for distance to specific salinity values as predicted by DSM2 (based on actual flows during a similar hydrologic period, 1995 to 2006) indicates that the location of the fluctuating 50mg/L isohaline from 1908 to 1917 corresponds to the location of X2 (2,640 µS/cm EC, or 700 mg/L) from 1995 to 2006. Fresh water (less than 50 mg/L chlorides or 350 µS/cm EC) has been located below the confluence of the Sacramento and San Joaquin Rivers approximately 42% of the time from 1995 to 2005, compared to 72% percent of the time from 1908 to 1917, indicating that the Delta is now saltier on average for a similar hydrologic period. Spatial Variability in Salinity - Sacramento River 100 50 mg/L EC <~ 350 μS/cm (1908-1917, C&H) 90 1995 to 2006 EC = 350 μS/cm (1995-2006, DSM2) EC = 1,000 μS/cm (1995-2006, DSM2) 80 EC = 2,640 μS/cm (1995-2006, DSM2) EC = 5,000 μS/cm (1995-2006, DSM2) 70

50 mg/L 40 X2 1908 to 1917

Percent Exceedance 1995 to 2006 30

0 0 10 20 30 40 50 60 Miles upstream of Crockett p_barge2.m 16-May-2007 Location of the Confluence of the Sacramento and San Joaquin Rivers

The spatial distribution of 350 µS/cm EC water quality during the early 1900s (1908- 1917) resembles the spatial distribution of approximately 2,500 µS/cm EC water quality now (1995-2006). This indicates a 7-fold increase in salinity.

5.1.2. Annual maximum salinity intrusion (1000 mg/L Chloride isohalines)

As illustrated in the Sacramento - San Joaquin Delta Atlas (DWR, 1993), there is no doubt that the 24 years from 1920-1943 experienced greater salinity intrusion than the post-Shasta era (1944-2006), with seawater intruding further into the Delta during 6 of

DRAFT: Last modified 6/9/2007 10:51 PM 46 D_report_flow_salt_trends_v1p2.doc DRAFT

the 24 years (1920, 1924, 1926, 1931, 1934, and 1939) than in any of the 45 years in the post-Shasta era (1944-1990).

This extreme salinity intrusion was due, in part, to precipitation and runoff during these years. From 1905 to 1935, the general trend of unimpaired runoff in the upper Sacramento River Basin was decreasing from a 5-year average of 25 MAF to 10 MAF (Figure 4, Page 18). Two of these water years (1924 and 1931) have the second and third lowest annual total Sacramento and San Joaquin runoff on record, with only 1977 experiencing less total runoff. Similarly, 1920 experienced one of the driest fall-winter periods (October to March) on record, with the only drier fall-winter in the post-Shasta era in water year 1977.

However, peak salinity intrusion during the 1920s and 1930s occurred between mid- August and mid-September (Supplemental Figure 14), indicating that salinity intrusion was probably not solely due to reduced runoff, which would have led to a peak salinity intrusion in late September or October (assuming the salinity maximum lags the flow minimum by about 3 or 4 weeks). The most probable secondary factor is the quantity and timing of upstream diversions during this era. Although the State and Federal Projects had not been built, diversions for local irrigation had substantially increased since the 19th century.

The salinity investigations of the era found that the extreme salinity intrusion was larger than any previous intrusions known to local residents and concluded the intrusion was due, in part, to the extensive upstream diversions. As one report observed:

Beginning in 1917, there has been an almost unbroken succession of subnormal years of precipitation and stream flow which, in combination with increased irrigation and storage diversions from the upper Sacramento and San Joaquin River systems, has resulted in a degree and extent of saline invasion greater than has occurred ever before as far as known. (DPW, 1931, p. 15).

The same report concludes that the upstream diversions likely doubled the extent of salinity intrusion above the confluence of the Sacramento and San Joaquin Rivers.

This is illustrated by comparing salinity intrusion in similar water years at different periods of development (Figure 2, Page 12). Water year 1913 experienced the greatest salinity intrusion observed by C&H from 1905 to 1917. Salinity intrusion experienced during similar water years during the 1920s and 1930s was 12 to 20 miles further upstream on the Sacramento River and even further on the San Joaquin River.

“Salinity control” provided by the CVP and SWP limit the impact of upstream diversions such that the extreme intrusion of the 1920s and 1930s has not occurred since. However, the CVP and SWP do not provide sufficient “salinity control” to limit salinity intrusion to the observed levels of the early 1900s.

DRAFT: Last modified 6/9/2007 10:51 PM 47 D_report_flow_salt_trends_v1p2.doc DRAFT

5.1.3. Modeling X2 Variability and the Effects of Flow Management Using the Kimmerer-Monismith equation (Kimmerer and Monismith, 1992), the location of X2, the distance from the Golden Gate to the 2 ppt isohaline, measured in kilometers (km), is modeled as a function of the current NDO and the previous X2.

X2(t) = 122.2 + 0.3278*X2(t-1) – 17.65*log10(NDO(t))

This empirical fit is only valid for the existing Delta geometry; therefore, this model is used to explore the impact of actual flow management relative to unimpaired conditions.

X2 has a strong seasonal and decadal variability under both unimpaired and actual flow conditions reflecting the strong seasonal and decadal variability of NDO (Figure 17). Flow management practices do little to alter to magnitude of the seasonal variability; the primary effect is a shift in the timing of salinity intrusion and a shift landward of approximately 5 km. The 1930s are an exception when the range of X2 in a given year reached nearly 70 km (double the typical range), most likely due to significant upstream diversions. The prediction of X2 for actual conditions is likely an underestimate of actual X2; during the 1930s DAYFLOW estimates the actual NDO was negative (net flow of bay water entering the Delta) averaging over -3,000 cfs for a number of months. However, the X2 prediction does not account for negative NDO.

Actual spring conditions (April-July) are much saltier than unimpaired flows would provide (~12 km further upstream). Although the fall (September-December) is saltier more than 50% of the time, there are years when actual conditions are fresher than unimpaired conditions (Figure 18). During these “fresher” periods, the maximum decrease in X2 is approximately 5 km, within the typical tidal variability of X2 in the fall.

X2 prediction is in agreement with results from analysis of C&H barge travel. Actual flow management limits the movement of X2 particularly during dry years; however, flow management does not prevent X2 from penetrating as far into the Delta, conversely, actual flow conditions do not allow the Delta to get as fresh as it would under unimpaired conditions during the driest years (Figure 19).

DRAFT: Last modified 6/9/2007 10:51 PM 48 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 17 – Location of X2 based on Unimpaired or Actual NDO X2 has a strong seasonal and decadal variability under both unimpaired and actual flow conditions reflecting the strong seasonal and decadal variability of NDO. Flow management practices do little to alter to magnitude of the seasonal variability; the primary effect is a shift in the timing of salinity intrusion and a shift landward of approximately 5 km. The 1930s are an exception when the range of X2 in a given year reached nearly 70 km (double the typical range), most likely due to significant upstream diversions. The prediction of X2 for actual conditions is likely an underestimate of actual X2; during the 1930s DAYFLOW estimates the actual NDO was negative (net flow of bay water entering the Delta) averaging over -3,000 cfs for a number of months. However, the X2 prediction does not account for negative NDO.

(a) X2 based on unimpaired NDO 130 120 110 100 90 2-u 80 X [km] 70 60 50 40 1920 1930 1940 1950 1960 1970 1980 1990 2000 (b) X2 based on actual NDO 130 120 110

100 90 2-a

X 80 [km] 70 60 50 40 1920 1930 1940 1950 1960 1970 1980 1990 2000

50 (c) Change in X2 due to flow management

2-u 30 20 - X [km]

2-a 10 X 0 -10 1920 1930 1940 1950 1960 1970 1980 1990 2000

DRAFT: Last modified 6/9/2007 10:51 PM 49 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 18 – Seasonal Distribution of X2 for different flow management regimes X2 modeling yields a similar result as the analysis of C&H data – salinity intrusion has shifted earlier in the year. Under unimpaired conditions, X2 starts increasing in June or July, but under actual conditions, X2 starts moving inland in April. Seasonal Distribution of X 1945 to 2006 2 100 unimpaired actual

80 [km] 2 X Earlier salinity 60 intrusion

Saltier

40 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep p_seasonal_stats.m 16-May-2007

DRAFT: Last modified 6/9/2007 10:51 PM 50 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 19 – Seasonal X2 variability during Wet and Dry Years Actual flow management limits the movement of X2 particularly during dry years; however, flow management does not prevent X2 from penetrating as far into the Delta, conversely, actual flow conditions do not allow the Delta to get as fresh as it would under unimpaired conditions during the driest years. Seasonal Distribution of X - Driest 10 Years 2 100

80 [km] 2 X

unimpaired actual

40 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep p_seasonal_stats.m Seasonal Distribution of X - Wettest 10 Years 16-May-2007 2 100 unimpaired actual

80 [km] 2 X

40 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep p_seasonal_stats.m 16-May-2007

DRAFT: Last modified 6/9/2007 10:51 PM 51 D_report_flow_salt_trends_v1p2.doc DRAFT

5.2. Trends in Salinity at Specific Locations

5.2.1. Salinity at Mallard Slough Data records for CCWD’s Mallard Slough and Rock Slough intakes are long enough to use for an investigation of trends in Delta salinity.

On April 21, 1967, Contra Costa Water District (CCWD) entered into an agreement with the State of California (acting through the Department of Water Resources) memorializing the terms under which the State reimburses CCWD for the decrease in availability of usable river water (water with less than 100 mg/L chlorides) at the Mallard Slough intake “due in part to the operation of the State Water Resources Development System as defined in Section 12931 of the Water Code” ([cite]).

Recognition that the SWP would increase salinity at Mallard Slough is the basis of this agreement and other similar agreements with other Delta water users. The calculation of the water deficiency entitlement was based on pumping records from 1926 through 1967, with an average of 142 days of usable water. The average number of days of usable water13 since the agreement was signed is 122, indicating a 20 day (14%) reduction in the number of days of high quality water at Mallard Slough since the SWP was built.

5.2.2. Observed salinity at Collinsville

Collinsville, near the confluence of the Sacramento and San Joaquin Rivers, was one of the first long-term sampling locations implemented by the State. The Suisun Marsh Branch14 of the DWR compiled monthly average salinity at Collinsville from 1920 through 2002, using 4-day TDS grab samples through 1971 and continuous EC measurements from 1966 to 2002. The overlap of 4 to 5 years allowed a regression of the measurements to convert the monthly averaged 4-day TDS samples to monthly average EC. Data is missing for a portion of the winter and spring prior to 1926; additionally, the entire year of 1943 is missing.

Since measurements of salinity began in 1920, there is a tendency to assume the earliest measurements represent the most “natural” conditions. However, as shown in Section 5.1.2, the annual maximum salinity intrusion in the 1920s far exceeded the annual maximum salinity intrusion during similar unimpaired hydrology in 1913. Therefore, we investigate the observed data in the context of climatic variability and anthropogenic modifications.

Even though the peak annual of the monthly averaged salinity in the 1920s and 1930s far exceeds any subsequent salinity measurements at Collinsville, during the winter and spring of the same years, Collinsville freshened considerably (Figure 20). Monthly average salinity was observed below 350 µS/cm EC (approximately 50 mg/L chloride)

13 Data is from the USBR-CVO record of EC at Pittsburg, approximately 2 km upstream of Mallard Slough from 1965-2005. Since this station is located upstream of Mallard Slough, the number of days of usable water at Mallard Slough since the SWP was built may be overestimated. 14 Data provided by Chris Enright (DWR), personal communication 2007. DRAFT: Last modified 6/9/2007 10:51 PM 52 D_report_flow_salt_trends_v1p2.doc DRAFT

for at least one month in every year with the possible exception of 1924 (inconclusive because salinity is unavailable for November through March). A fresh winter and spring is consistent with observations by C&H throughout the 1920s (Supplemental Figure 13, Page 82). However, during the recent droughts in 1976-77 and 1987-1993, salinity at Collinsville seldom drops below 350 µS/cm EC for one month in both 1989 and 1992.

Figure 20 – Observed salinity at Collinsville Caption Salinity at Collinsville 30 Monthly Average 25 1-year Running Average 5-year Running Average

EC [mS/cm] 10

0 1920 1930 1940 1950 1960 1970 1980 1990 2000

The year-to-year trends in monthly salinity at Collinsville (Figure 21) further illustrate that winter salinity during the recent droughts of 1976-77 and 1987-93 has been greater than any previous winter salinity, including the drought of the 1930s.

DRAFT: Last modified 6/9/2007 10:51 PM 53 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 21 – Year-to-Year trends in monthly average salinity Caption 10 January 5

0 10 Monthly Average March 5-year Running Average 5 20-year Running Average

0 20 15 May 10 5

40 July 30 20 10 0 40 September 30 20 10 0 20 15 November Monthly Average Electrical Conductivity at Collinsville [mS/cm] Monthly Average Electrical Conductivity 10 5

0 1920 1930 1940 1950 1960 1970 1980 1990 2000

5.2.3. Modeling salinity at Collinsville and the effects of flow management Salinity at Collinsville, near the confluence of the Sacramento and San Joaquin Rivers, can be estimated by the G-model, a conceptual-empirical model of salinity transport along the Sacramento River (Denton and Sullivan, 1993).

DRAFT: Last modified 6/9/2007 10:51 PM 54 D_report_flow_salt_trends_v1p2.doc DRAFT

Predicted salinity at Collinsville has a strong seasonal and decadal variability under both unimpaired and actual flow conditions reflecting the strong seasonal and decadal variability of NDO (Figure 22). Flow management practices have a significant impact on the seasonal variability of salinity at Collinsville, particularly during dry years, when Collinsville experiences a much greater range of monthly average salinity under actual conditions compared to unimpaired conditions. This increase in the seasonal variability of salinity is due to the increased average salinity at Collinsville, such that the range of seasonal variability increases as the average salinity increases.

Figure 22 – Predicted salinity at Collinsville based on unimpaired and actual NDO Caption (a) EC based on unimpaired NDO 20000

16000

12000

8000

4000

0 1920 1930 1940 1950 1960 1970 1980 1990 2000 (b) EC based on actual NDO 20000 Monthly Average 5-year Average 16000 Linear Trend 12000 Conductivity at Collinsville 8000

[µS/cm] 4000

0 1920 1930 1940 1950 1960 1970 1980 1990 2000 (c) actual EC – unimpaired EC 12000 10000 8000 6000 4000 2000 0

Monthly Average Electrical -2000 -4000 1920 1930 1940 1950 1960 1970 1980 1990 2000 29

Similar to the X2 prediction, the prediction of EC for actual conditions is likely an underestimate of actual EC during the 1930s; DAYFLOW estimates the actual NDO was negative (net flow of bay water entering the Delta) averaging over -3,000 cfs for a number of months. However, the EC prediction does not account for negative NDO.

The seasonal distribution of salinity based on unimpaired NDO (Figure 23) indicates that without flow manipulation, Collinsville salinity would peak in September and October at approximately 3,000 µS/cm on average (50% of the time), yet remain below 500 µS/cm on average for eight months (December through July). Distribution of salinity based on the actual NDO shows peaks in September and October around 4,000 µS/cm; August is also shown as a peak of over 4,000 µS/cm on average, and the 75% percentile average

DRAFT: Last modified 6/9/2007 10:51 PM 55 D_report_flow_salt_trends_v1p2.doc DRAFT

salinity for August through October has substantially increased, indicating a wider distribution of salinities with a positive skew (which translates to a saltier distribution).

Figure 23 – Seasonal Distribution of Salinity at Collinsville (Gmodel)

Seasonal Distribution of EC at Collinsville for Different Flow Management Regimes 10000 unimpaired 9000 actual Exceedance 10 8000 25

7000 IQR 50

median 75 6000 90 S/cm]

μ 5000

EC [ 4000

3000

2000

1000

0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

The seasonal distribution of the percent change between salinity based on unimpaired NDO and salinity based on actual NDO is illustrated Figure 24. From April through August, actual conditions show substantially greater (more than double, on average) salinity over unimpaired conditions. From September through December, unimpaired salinity is, on average, greater than actual salinity; however, about 10% of the time, actual salinity is lower than unimpaired salinity by up to 50%.

DRAFT: Last modified 6/9/2007 10:51 PM 56 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 24 – Seasonal Distribution of Difference between Actual and Unimpaired Salinity as a percent change from Unimpaired Salinity (Gmodel)

Change in Salinity at Collinsville(1956-2003) 1500

1000

500 Percent Change (Dayflow - Unimpaired) - (Dayflow Change Percent

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

As described in previous sections, the slight freshening of the fall months is likely due to reservoir releases for flood control during wet years. As development has increased, the impact of fall reservoir releases has decreased.

Limiting the analysis to the more recent years (1993-2003), the change in salinity based on unimpaired NDO and salinity based on actual NDO is significantly different (Figure 25 and Figure 26), indicating a shift in flow management within the system. While salinity at Collinsville based on actual NDO shows an increase of over 2,000 µS/cm on average in July and August over the entire time period of record, and a minimal average increase in October and November, the post-1993 years show a less severe average increase of just 1,000 µS/cm in July and even less in August, with a notable average increase in fall salinity of over 2,000 µS/cm in October and almost 3,000 µS/cm in November.

DRAFT: Last modified 6/9/2007 10:51 PM 57 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 25 – Seasonal Distribution of Salinity at Collinsville since 1993 (Gmodel)

Seasonal Distribution of EC at Collinsville for Different Flow Management Regimes 8000 unimpaired Exceedance 7000 actual 10

6000 25

IQR 50

5000 median 75

90 S/cm]

μ 4000 EC [ 3000

2000

1000

0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

Figure 26 – Seasonal Distribution of Difference between Actual and Unimpaired Salinity as a percent change from Unimpaired Salinity since 1993 Change in Salinity at Collinsville (1994-2003) 1500

1000

500 Percent Change (Dayflow - Percent - Unimpaired) Change (Dayflow

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

DRAFT: Last modified 6/9/2007 10:51 PM 58 D_report_flow_salt_trends_v1p2.doc DRAFT

5.2.4. Fall Salinity in the Western Delta The analysis summarized in Section 5.1.1, examining the distance traveled to fresh water in recent years in comparison to C&H barge records, found that the distance traveled to fresh water is on average 5-10 miles further in the fall months of 1995 to 2006 compared to similar water years in the early 1900s and in the 1960s. Similarly, comparison of salinity based on unimpaired NDO and actual NDO (Section 5.2.3) shows a dramatic shift in the impact of flow management on October and November salinity in the years after 1993.

The increase in fall salinity is observed at multiple stations in Suisun Bay and the western Delta (Figure 27) for water years with a total annual SRI of 15 to 30 MAF, generally normal and wet years. Fall salinity at Chipps Island (Figure 28) during normal years is now comparable to fall salinity during dry and critical years prior to 1994.

DRAFT: Last modified 6/9/2007 10:51 PM 59 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 27 – Post-ESA salinity in the Suisun Bay and western Delta

DRAFT: Last modified 6/9/2007 10:51 PM 60 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 28 – Increase in Fall Salinity at Chipps Island

CHIPPS Average October, November, and December 1200 1976-1993 1993-2005 1000

8000

6000

EC (microS/cm) 4000

2000

0 Dry & Critical Normal Wet deltaSalinityTren 15-Sep-2005

DRAFT: Last modified 6/9/2007 10:51 PM 61 D_report_flow_salt_trends_v1p2.doc DRAFT

6. Conclusions

Historical trends in salinity provide perspective and indicate causality. Salinity variability as determined from a sediment core in northwestern Suisun Marsh broadly corresponds to independent climate indicators, with general agreement of higher salinity during the Medieval Warm Period and fresh conditions during the Little Ice Age. However, the recent increase in salinity since the mid-1800s does not correspond to regional climate change, but rather is primarily due to anthropogenic causes.

Anthropogenic modifications to the Bay-Delta are grouped into two primary categories: changes to the landscape and flow management. The chronology of anthropogenic changes and salinity observations are summarized in Table 7 and Table 8. Up until 1917, the most significant impact on salinity was likely due to changes to the landscape of the Central Valley and Delta. Since 1917, flow management activities have the greatest impact on observed salinity, modulating the impact of natural climatic variability.

Table 7 – Chronology of anthropogenic modifications

Era Anthropogenic Modifications Prior to 1835 Minimal alteration prior to European settlement (Pre-European) 1860-1917 Changes to the landscape of the Central Valley and Delta are significant. (Early Settlement) • Reclamation of marsh lands • Alluviation then erosion of mine-derived sediment • Deepening, widening, and straightening of Delta channels Water diversions increase throughout this period. (DPW, 1931) • By 1870, irrigation diversions noticeably reduce flow in the San Joaquin River. • Gross annual irrigation diversions from the Sacramento and San Joaquin rivers grow from 1.0 million acre-feet (MAF) in 1879 to 4.3 MAF in 1917 1918-1945 Changes to the landscape are less substantial than the previous era. (Pre-CVP) • Continued deepening of Delta channels • Continued erosion of mine tailings Water diversions continue to increase throughout this period. • Upstream storage capacity grows from 1.2 MAF in 1920 to 4.6 MAF in 1943 • Annual irrigation diversions exceed 6.5 MAF by 1945 1945-1967 Changes to the landscape continue, but not as dramatic as earlier eras. (Pre-SWP) Water diversions continue to increase with substantial increases in storage. • Shasta reservoir (4.5 MAF) completed in 1945 • Upstream storage capacity increases to 17.5 MAF in 1966 • South of Delta exports begin in 1951, exceeding 1.6 MAF by 1966

DRAFT: Last modified 6/9/2007 10:51 PM 62 D_report_flow_salt_trends_v1p2.doc DRAFT

Era Anthropogenic Modifications 1967-1993 Water diversions continue to increase with substantial increases in storage. (Pre-ESA) • Oroville reservoir (3.5 MAF) completed in 1968 • Upstream storage capacity increases to 30.4 MAF by 1979 • South of Delta exports increase to 6 MAF by 1990 Water quality, water rights, and other agreements that impact timing of reservoir releases and south of Delta exports • Water Rights Decision 1485 issued in 1978 • CVP Improvement Act approved by congress in 1992 1994-present Water quality, water rights, and other agreements that impact timing of (Post-ESA) reservoir releases and south of Delta exports • Bay-Delta Accord sets interim water quality objectives in 1994

Table 8 – Chronology of Salinity Variability

Era Salinity Characteristics Historical Context Paleosalinity analysis provides historical context and indicates centennial-scale salinity variability, which broadly mirrors climatic variability.

800 B.C. – 550 B.C. Relatively brackish period 550 B.C. – A.D. 200 Relatively fresh period A.D. 200 – A.D. 1200 Relatively brackish period (includes Medieval Warm Period) A.D. 1200 – A.D. 1860 Relatively fresh period (includes Little Ice Age) A.D. 1860 – A.D. 2000 Increasing salinity, apparently not driven by climatic variability

1860-1917 ¾ Salinity intrusion is only reported during the drought of 1870. (Early Settlement) ¾ Earliest salinity measurements (1908-1917) indicate salinity of 1,000 mg/L chloride did not move upstream of the confluence, even during dry years. 1918-1945 ¾ Salinity intrusion is greater than any other time period, likely (Pre-CVP) caused by upstream diversions and lack of precipitation. ¾ Salinity retreats and fresh water reaches the confluence of the Sacramento and San Joaquin rivers during the winter even during dry years. 1945-1967 ¾ Salinity intrusion is “controlled” by reservoir releases, limiting (Pre-SWP) the impact of upstream diversions but not returning to levels observed during the relatively unimpacted period from 1908 to 1917. ¾ Delta is generally saltier than would occur under unimpaired conditions during most months. ¾ Reservoir releases slightly freshen the Delta during February and September, primarily during wet years, likely due to flood control operations. 1967-1993 Similar to the previous era, with increased reservoir capacity further (Pre-ESA) freshening the Delta during September.

DRAFT: Last modified 6/9/2007 10:51 PM 63 D_report_flow_salt_trends_v1p2.doc DRAFT

Era Salinity Characteristics 1994-present Substantial increase in fall salinity in the western Delta during all but (Post-ESA) the wettest years.

Conceptual Models The observed and modeled changes in seasonal salinity are summarized in two conceptual models.

Conceptual Model 1: Observed changes in seasonal salinity with time.

The salinity regime changes as the level of development increases and Project operations change due to regulatory requirements. Comparison of three decades with similar hydrology in Figure 29 illustrates the changing salinity regime in Suisun Bay and the western Delta.

Monthly average salinity during the spring and summer was substantially greater from 1966 to 1975 than during the early 1900s. However, the fall and early winter were slightly fresher than the early 1900s. This reduction in salinity in the fall and early winter was likely due in part to operation of the Central Valley Project (CVP) and State Water Project (SWP) with reservoir releases for flood control purposes in the fall freshening the Delta before diversions and exports were fully developed.

In contrast, salinity during the last 10 years has exceeded salinity in the early 1900s during all months for years with similar hydrologic conditions. The dramatic increase in fall salinity is accompanied by a slight decrease in spring and summer salinity, likely due to minimum flow and X2 requirements imposed in 1995, in comparison to observed levels from 1966 to 1975; however, spring and summer salinities remain much greater than levels in the early 1900s.

The range of seasonal variability during 1966 to 1975 was greatly reduced as the Delta did not get as fresh as it did in the early 1900s. During the last decade, seasonal variability has increased such that the range of salinity observed over the course of a year is similar to the early 1900s. However, the spatial location of salinity fluctuations has moved inland relative to the early 1900s, resulting in saltier conditions in the Suisun Bay and western Delta and the length of time with fresher conditions is reduced.

DRAFT: Last modified 6/9/2007 10:51 PM 64 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 29 – Seasonal variability of salinity in Suisun Bay and the western Delta under different flow management eras

Saltier fall Pre-Project (1907-1918) 1995-2005 Early Post-Project (1966-1975) Recent Post-Project (1995-2005)

Saltier Saltier spring and summer

Earlier salinity Fresher fall intrusion 1965-75 Fresher

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

Conceptual Model 2: The impact of flow management for wet and dry years.

Flow management has the largest impact during dry years (Figure 30.a.) when the Delta stays relatively salty throughout the year with limited seasonal variability compared to unimpaired conditions. During wet years (Figure 30.b.), the Delta freshens as much as it would under unimpaired conditions, but the Delta does not stay fresh for as long.

DRAFT: Last modified 6/9/2007 10:51 PM 65 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 30 – Seasonal salinity variations in the Delta under actual conditions compared to unimpaired conditions in (a) dry years and (b) wet years

(a) Dry Years

Unimpaired Actual

Seasonal range in salinity under Saltier actual conditions

Seasonal range in salinity under unimpaired conditions Fresher

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

(b) Wet Years

Unimpaired Actual

Time period Delta is fresh under unimpaired conditions Saltier Reservoir releases Time period Delta is fresh under freshen the actual conditions Delta during wet years (flood control) Fresher

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

DRAFT: Last modified 6/9/2007 10:51 PM 66 D_report_flow_salt_trends_v1p2.doc DRAFT

7. References

[update and add citations to text]

Ateljuvich, Eli Sava. 2001. Optimal Control of Salinity in the Sacramento-San Joaquin Delta.

Atwater, Brian F, Susan G. Conard, James N. Dowden, Charles W. Hedel, Roderick L. MacDonald, Wayne Savage. 1979. History, Landforms, and Vegetation of the Estuary’s Tidal Marshes.

Atwater, Brian F and Daniel F. Belknap. 1980. Tidal-Wetland Deposits of the Sacramento-San Joaquin Delta, California.

The Bay Institute. 1998. From the Sierra to the Sea: The Ecological History of the San Francisco Bay-Delta Watershed. Novato, California.

Byrne, Roger, B. Lynn Ingram, Scott Starratt, Frances Malamud-Roam, Joshua N. Collins, Mark E. Conrad. 2001. Carbon-Isotope, Diatom, and Pollen Evidence for Late Holocene Salinity Change in a Brackish March in the San Francisco Estuary.

Conomos TJ. 1979. San Francisco Bay: the urbanized estuary. San Francisco (CA): Pacific Division, American Association for the Advancement of Science.

Denton, R.A. and G.D. Sullivan. 1993. “Antecedent Flow-Salinity Relations: Applications to Delta Planning Models.” Contra Costa Water District.

Denton, R.A. 1993. Accounting for antecedent conditions in seawater intrusion modeling – applications for the San Francisco Bay-Delta. Hydraulic Engineering 1993, vol. 1. ASCE. P448-453.

[DPW] Department of Public Works. 1931. Variation and Control of Salinity in Sacramento-San Joaquin Delta and Upper San Francisco Bay. Bulletin No. 27. State of California. Department of Public Works. Publication of the Division of Water Resources.

[DPW] Department of Public Works. 1936. Sacramento-San Joaquin Water Supervision Report for Year 1935. State of California. Department of Public Works. Publication of the Division of Water Resources.

[DWR] Department of Water Resources. 1960. Delta Water Facilities. Bulletin No. 76. State of California.

[DWR] Department of Water Resources. 1962. Salinity Incursion and Water Resources. Appendix to Bulletin No. 76. The Resources Agency of California.

[DWR] Department of Water Resources. 1987. California Central Valley Unimpaired Flow Data. State of California. The Resources Agency. Division of Planning.

DRAFT: Last modified 6/9/2007 10:51 PM 67 D_report_flow_salt_trends_v1p2.doc DRAFT

[DWR] Department of Water Resources. 1993. Sacramento-San Joaquin Delta Atlas. State of California.

[DWR] Department of Water Resources. 1994. California Central Valley unimpaired flow data. Sacramento (CA): California Department of Water Resources.

[DWR] Department of Water Resources, Bay-Delta Office. 2005. CalSim-II Model Sensitivity Analysis Study. Technical Memorandum Report. State of California.

[DWR] Department of Water Resources. 2006. Progress on Incorporating Climate Change into Planning and Management of California’s Water Resources. State of California. The Resources Agency. Department of Water Resources.

[DWR] Department of Water Resources. 2007. Sacramento-San Joaquin Delta Overview. State of California.

Division of Water Rights, Permanent Committee of the Sacramento-San Joaquin River Problems Conference, and the Sacramento Chamber of Commerce. 1924. Proceedings of the Second Sacramento-San Joaquin River Problems Conference.

Enright, Chris. 1998. Commentary on Modeling the Pre-Levee Suisun Marsh Salinity Regime.

Enright, Chris. 2004. “Long-Term Trends and Variability of Suisun Marsh Salinity” presented at the San Francisco Bay-Delta Science Consortium Suisun Marsh Workshop, March 1-2, 2004.

Fox, J.P. 1987a. Salinity and Temperature Variations in San Francisco Bay. State Water Contractors Exhibit Number 266. Bay Delta Hearings, Sacramento, CA.

Fox, J. P. 1987b. Freshwater Inflow to San Francisco Bay Under Natural Conditions. State Water Contractors Exhibit Number 262. Bay Delta Hearings, Sacramento, CA.

Greene, Sheila. 1987. Historic Daily Flows (water years 1930-1955) Sacramento-San Joaquin Delta as calculated by DAYFLOW program. Department of Water Resources.

Kimmerer, Wim and Stephen Monismith. 1992. An Estimate of the Historical Position of 2PPT Salinity in the San Francisco Bay Estuary.

Knowles, Noah. 2000. Modeling the Hydroclimate of the San Francisco Bay-Delta Estuary and Watershed. Dissertation. University of California, San Diego.

Lund, J., E. Hanak, W. Fleenor, R. Howitt, J. Mount, and P. Moyle. 2007. Envisioning Futures for the Sacramento-San Joaquin Delta. Public Policy Institute of California. San Francisco, California.

Meko, D.M. 2001. Reconstructed Sacramento River System Runoff From Tree Rings. Report prepared for the California Department of Water Resources. DRAFT: Last modified 6/9/2007 10:51 PM 68 D_report_flow_salt_trends_v1p2.doc DRAFT

Monismith S. 1998. X2 workshop notes. IEP Newsletter 11(4):6-14. Available at: http://www.iep.water.ca.gov/report/newsletter.

Monroe, M.W., and J. Kelly. 1992. State of the Estuary: A Report on Conditions and Problems in the San Francisco Bay/Sacramento-San Joaquin Delta Estuary. San Francisco Estuary Project. Oakland, California.

Nichols F, Cloern J, Luoma S, Peterson D. 1986. The modification of an estuary. Science 231:567-573.

Rosencrans. 2007. Natural, Unimpaired and Actual Salinity Intrusion in the Sacramento-San Joaquin Bay-Delta Estuary.

Smith, Susan E. and Susumu Kato. 1979. The Fisheries of San Francisco Bay: Past, Present and Future.

Starratt, S.W. 2001. “Diatoms as Indicators of Freshwater Flow Variation in Central California.” PACLIM Conference Proceedings, pp. 129-144.

State Water Contractors. 1987. Additional Evidence in Regard to Freshwater Inflow to San Francisco Bay Under Natural Conditions.

State Water Contractors. 1987. Rebuttal to David R. Dawdy Exhibit 3 in Regard to Freshwater Inflow to San Francisco Bay Under Natural Conditions.

Town of Antioch v. Williams Irrigation District (1922, 188 Cal. 451)

U.S. Geological Survey. 1999. Sedimentation and Bathymetry Changes in Suisan Bay: 1867- 1990.

Wells, Lisa E. and Michelle Goman. 1995. Late Holocene Environmental Variability in the Upper San Francisco Estuary as Reconstructed from Tidal Marsh Sediments.

Zuckerman, Thomas M. 1968. Statement Before Central Valley Regional Water Quality Control Board. Stockton, California.

DRAFT: Last modified 6/9/2007 10:51 PM 69 D_report_flow_salt_trends_v1p2.doc DRAFT

Supplemental Figures

DRAFT: Last modified 6/9/2007 10:51 PM 70 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 1 –Spring (April-June) Sacramento River Runoff in Percent of Water Year Runoff Seasonal distribution of unimpaired runoff has been shifting over the last century; the percentage of total annual runoff occurring in the spring has been decreasing. From (cite)

DRAFT: Last modified 6/9/2007 10:51 PM 71 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 2 – Irrigated Acreage in the Central Valley Irrigated acreage in the Central Valley experienced two rapid periods of growth, from 1880 to 1920 and from 1940 to 1980. Introduction of rice cultivation around 1916 significantly altered the amount and timing of upstream diversions due to the agricultural practice of flooding rice fields

DRAFT: Last modified 6/9/2007 10:51 PM 72 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 3 – Salinity on the San Joaquin River at Antioch (DWR, 1960) The Department of Water Resources examined the effects of upstream depletions and south of Delta exports on salinity in the San Joaquin River at Antioch, estimating the percent of time water that a certain quality of water (with less than 350 mg/L chlorides; or less than 1,000 mg/L chlorides) would be available in the river without reservoir releases to provide salinity control. This graphic illustrates that upstream depletions and exports had a significant effect on salinity at Antioch from 1900 through 1940 prior to the construction of large upstream reservoirs. Shasta was completed in 1943; the bars for 1960, 1980, 2000, and 2020 assume the reservoirs do not make releases for salinity control and therefore underestimate the actual quality of water during these years.

DRAFT: Last modified 6/9/2007 10:51 PM 73 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 4 – California Reservoirs

California Reservoirs

Reservoir Capacity [AF]

Year Built

DRAFT: Last modified 6/9/2007 10:51 PM 74 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 5 – Surface Reservoir Capacity Reservoir construction began in 1850. Prior to the construction of Shasta Dam (4.5 MAF capacity completed in 1945), upstream reservoirs did not provide salinity control. By 1931, cumulative capacity of all upstream reservoirs in the Bay-Delta watershed exceeded 4 MAF, capable of capturing over half of the total unimpaired runoff during the drought in water year 1931 (7.76 MAF, Eight River Index). 5.0 30 Individual Dams Shasta 4.5 Cummulative Storage Capacity 4.0 25 Oroville 3.5 20 3.0

New Melones 2.5 15 San Luis New Don Pedro 2.0 Monticello

Capacity [MAF] Capacity (Putah Creek) New Exchequer 10 1.5 Lake Almanor (Merced River) Pine Flat Folsom

Parde (Kings River) New Bullards Bar Cummulative Capacity [MAF] 1.0 (Yuba River) Frian 5 O Shaughnessy Isabella Comanch 0.5 Buena Vista Clear Lake New Hogan Indian Valley 0.0 0 1850 1870 1890 1910 1930 1950 1970 1990 Year Built

DRAFT: Last modified 6/9/2007 10:51 PM 75 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 6 – Exports from the Southern Delta Exports from the south Delta started in 1951 with the completion of the Federal CVP pumping facility near Tracy, California. The SWP pumping plant, just to the west of the Federal facility, was completed in 1967. Exports generally increased from 1951 to 1976. Following the drought of 1977, exports have remained around 5 MAF/year.

DRAFT: Last modified 6/9/2007 10:51 PM 76 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 7 – Validation of Reconstructed Flows Analysis of tree rings provides for a reconstruction of upstream hydrology in the Sacramento River basin from 901 to 1977 (Meko, 2001). A comparison between observed and reconstructed unimpaired Sacramento River flow (from 1906 to 1977) illustrates good agreement particularly at low flows. 40 Reconstructed Flow 35 1:1 (Reconstructed = Observed) Linear Regression 30

River Index (MAF) 20

10 Reconstructed Sac 4-

0 0 5 10 15 20 25 30 35 Observed Sac 4-River Index (MAF)

DRAFT: Last modified 6/9/2007 10:51 PM 77 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 8 – Tidal Variability in Salinity at Antioch (1967 to 1992) Daily maximum salinity in the San Joaquin River at Antioch can be double the daily average salinity. On average, daily maximum salinity is approximately 150% of the the daily average. Similarly, the daily minimum may by 40-75% of the daily average; on average, the daily minimum is 58% of the daily average. Due to the large tidal variability, comparison of salinity observations must be specific to tidal phase, at least accounting for tidal variability.

DRAFT: Last modified 6/9/2007 10:51 PM 78 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 9 – Tidal Variability in Salinity at Rio Vista (1967 to 1992) Daily maximum salinity in the Sacramento River at Rio Vista can be 170-400% of the daily average salinity. Similarly, the daily minimum may by 10-65% of the daily average. Due to the large tidal variability, comparison of salinity observations must be specific to tidal phase.

DRAFT: Last modified 6/9/2007 10:51 PM 79 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 10 – Temporal Variability in Spatial Distribution of Salinity Isohalines (contours of equal electrical conductivity) illustrate the tidal variability of the spatial distribution of salinity.

20-Apr-1991 16:00:00 03-Sep-1991 08:00:00

Daily Minimum Salinity Daily Minimum Salinity (April 1991) (September 1991)

02-Sep-1991 22:00:00 20-Apr-1991 08:00:00 Daily Maximum Salinity Daily Maximum Salinity (April 1991) (September 1991)

DRAFT: Last modified 6/9/2007 10:51 PM 80 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 11 – C&H Barge Travel C&H obtained its water supply from barges traveling up the San Joaquin and Sacramento rivers from 1905 through 1929 (or later). The distance traveled by the barges and the concentration of chlorine in the water supply is published by DPW (1931) and reprinted herein.

Distance [miles] from Crockett Mallard Slough 18 miles Collinsville 22 miles Antioch 26 miles Jersey Point 32 miles Emmaton 28 miles Rio Vista 34 miles

DRAFT: Last modified 6/9/2007 10:51 PM 81 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 12 – C&H Barge Travel For 10 years (1908 to 1917), C&H was able to obtain water with less than 50 mg/L chlorides within 30 miles of Crockett on average (below Jersey Point on the San Joaquin River). In 1918, the quality of water obtained by C&H barges had degraded due to a combination of upstream diversions (especially for newly introduced rice cultivation) and lack of precipitation, exceeding 60 mg/L chlorides during August and September, when the barges traveled farther upstream than any time previously recorded. Then in 1919, a wetter year than 1918, the water quality was degraded for an even longer period of time, most likely due to increased upstream irrigation, exceeding 60 mg/L chloride during July, August, and September. Beginning in 1920, C&H abandoned the Sacramento and San Joaquin Rivers during the summer and fall seasons, replacing the water supply with a contract from Marin County. However, even during the driest years of the 1920s, C&H obtained water with less than 50 mg/L chloride below the confluence of the Sacramento and San Joaquin Rivers during some portion of every year.

DRAFT: Last modified 6/9/2007 10:51 PM 82 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 13 – Hydrologic Context for Distance to Fresh Water analysis Compared to the entire period of record (1906-2006), the first decade of C&H barge travel is a relatively wet period (1908-1917). To limit influence of natural hydrologic variability, we selected specific decades (a) and select years (b) which have comparable or slightly wetter hydrology than the C&H years. Hydrology distribution for time period of each decade (a) 100 All Water Years: 1906-2006 90 C&H Years: 1908 - 1917 Post-dam 1: 1965 - 1977 80 Post-dam 2: 1995 - 2006

40 Percent Exceedance 30

20 wetter

0 5 10 15 20 25 30 35 40 Sacramento Basin Total Unimpaired Runoff [MAF] p_wycdf.m 16-May-2007 (b) Context of selected water years in the overall hydrology distribution 100 All Water Years: 1906-2006 90 C&H Years Dry Years (1913,1918) C&H Wet Years (1911,1916) 80 Post-dam Dry Years (1968,1979,2002) Post-dam Wet Years (1969,1982,1995) 70

Percent Exceedance 30

20 wetter 10

0 5 10 15 20 25 30 35 40 Sacramento Basin Total Unimpaired Runoff [MAF] p_wycdf.m 16-May-2007

DRAFT: Last modified 6/9/2007 10:51 PM 83 D_report_flow_salt_trends_v1p2.doc DRAFT

Figure 14 – Salinity Intrusion (DWR, 1962) Maximum of the peak daily salinity intrusion (measured by the location of 1,000 mg/L chlorides) before and after the completion of Shasta Reservior (1945). Six years of extreme salinity intrusion between 1920 and 1944 are influenced a combination of dry hydrology and significant upstream diversions. DWR estimated the unregulated upstream diversions doubled the salinity incursion during these years.

DRAFT: Last modified 6/9/2007 10:51 PM 84 D_report_flow_salt_trends_v1p2.doc