Screening Analysis of Delta Island Reclamation Alternatives for Ecosystem Restoration and Salinity Reduction Callie Harrison, Chris Enright, Kamyar Guivetchi * DWR Suisun Marsh Branch* May 2000 *Executive Summary* The Department of Water Resources Suisun Marsh Branch conducted a reconnaissance level modeling analysis to evaluate impacts of island reclamation alternatives for ecosystem restoration and salinity control in the western Delta. The study focuses on alternative configurations for Sherman Lake, Big Break and Franks Tract ( Map 1 ). The analysis was requested by DWR Central District after recent modeling analysis indicating salinity reduction in the Delta from selective levee breaches within the Suisun Marsh. Central District requested the alternatives evaluation as a screening measure to assist and guide development of a CALFED Category III proposal. Modeling scenarios were crafted and revised to identify alternatives that provide both ecosystem restoration and salinity reduction potential. The DWR DSM1 (Suisun Marsh version) model was used. A base model run was made simulating historical conditions in drought water years 1991 and 1992. All alternative configuration simulations and salinity impacts analyses are for water year 1992. System wide salinity impacts are reported for one day in July 1992 representing the critically dry period between June and September 1992. *Sherman Lake Scenarios* The analysis initially focused on Sherman Lake including the remnant channel on the east end of the lake ( Scenarios 1-10 <#scenario1>). Scenarios ranged from complete reclamation (no levee breaches) to permutations of small breaches on the Sacramento, Broad Slough, and San Joaquin Rivers side of Sherman Lake. Additionally, alternative configurations of the remnant east side channel were investigated. The range of salinity increase or decrease from the base case in Delta locations was small (less than 6% in the western Delta, less than 3% in the south Delta). Given the hydrodynamic influence of Three Mile Slough on Delta salinity, we investigated the impact of widening Three Mile Slough in combination with Sherman Lake scenarios ( Scenarios 11-13 <#scenario11>). Widening Three Mile Slough alone results in higher salinity (1-6%) in the lower Sacramento River and lower salinity (2-8%) in the lower San Joaquin River. South Delta salinity is marginally reduced. Closing Sherman Lake or opening an east Sherman Lake channel between the Sacramento and San Joaquin Rivers reduces west Delta salinity between 3 and 10%, and reduces south Delta salinity between 1 and 3%. *Franks Tract Scenarios* Alternative configurations investigated for Franks Tract ranged from complete reclamation (no levee breaches), to permutations of small breaches on the east and west side ( Scenarios 14-18 <#scenario14>). Percentage salinity reductions in the Delta for all scenarios were significant: 10-35% in the central Delta and 15-30% near Clifton Court Forebay while Suisun Bay and western Delta salinity increased between 1 and 5%. *Big Break Scenarios* Alternative configurations investigated for Big Break included complete reclamation, and complete reclamation with a 100 foot breach on the north side remnant Dutch Slough channel ( Scenarios 19 and 20 <#scenario19>). Percentage salinity reductions in the Delta for both scenarios were significant (10-15% in the central Delta, 7-12% near Clifton Court Forebay) while Suisun Bay and western Delta salinity increased marginally. *Combination Big Break and Franks Tract Scenarios* Given the significant salinity impacts of Big Break and Franks Tract scenarios, a few combination permutations were investigated ( Scenarios 21-23 <#scenario21> ). Complete reclamation of both islands results in salinity reductions of 10-45% in the central Delta, about 30% near Clifton Court Forebay. This represents greater salinity reductions than each scenario alone. The addition of small breaches on both islands will likely cause less salinity reduction. Results indicate significant sensitivity to breach size and location. The physical mechanisms controlling the observed salinity response are discussed in the main body of this report. *Corroborative Modeling Analysis* An independent verification of the results was desired acknowledging that Delta levee breach modeling challenges the limits of the DSM1 models' capabilities. Two independent verification efforts have thus far been completed using 1) a two-dimensional finite element Bay-Delta model (RMA 2 and 11) and, 2) the DWR DSM2 model. The consulting firm Resource Management Associates (RMA) modeled the east Sherman Lake channel ( Scenario 8 <#scenario8> ) and the Big Break reclamation ( Scenario 19 <#scenario19>). DSM1 and RMA predict similar salinity values in the south Delta for the east Sherman Lake scenario and similar salinity trends for the Big Break scenario. However, DSM1 predicts salinity reductions in the south Delta of approximately 5 times greater than the RMA results. The reasons for this difference should be investigated further. The DWR DSM2 model was applied to the Franks Tract reclamation scenario using an alternative modeling approach. A CALSIM base study provided monthly average Delta inflows and exports for the 16-year period from October 1975 through September 1991. Salinity reductions in the range of 15-20% were observed for Old River near Rock Slough, and 10-15% near Clifton Court Forebay. We believe that these results substantially corroborate DSM1 results. *Conclusions* Application of the DSM1 model as a screening tool effectively guided the search for feasible Delta island reclamation alternatives for further investigation. Combinations of island reclamation and engineered breaches were identified that provide opportunities for ecosystem restoration and central and south Delta salinity reduction. The investigation indicates that some Sherman Lake alternatives provide marginal salinity reduction. Big Break and Franks Tract alternatives show potential for significant central and south Delta salinity reduction. All of the participants in the modeling effort so far agree that further field and modeling analysis should be pursued to elucidate the mechanisms controlling the observed hydrodynamics and salinity response. Participants recommend that a multiple agency, university, and stakeholder team of hydrodynamics and water quality experts should be convened and afforded resources to improve modeling techniques, understand physical mechanisms, and identify optimal reclamation and levee breach configurations for ecosystem reclamation and salinity reduction benefits. *I. Background * The Department of Water Resources Suisun Marsh Branch conducted a reconnaissance level modeling analysis to evaluate water quality impacts of island reclamation alternatives for ecosystem restoration and water quality control in the western Delta. The analysis was requested by DWR Central District after recent modeling analysis indicating water quality benefits in the Delta from selective levee breaches within the Suisun Marsh. Central District requested the alternatives evaluation as a screening measure to assist and guide development of a CALFED Category III proposal. Modeling scenarios were crafted and revised to identify alternatives providing both salinity reduction and ecosystem restoration potential. The study focuses on alternative configurations for Sherman Lake, Big Break and Franks Tract ( Map 1 ). The goal of the evaluation was to screen many alternative Delta island reclamation scenarios for water quality and ecosystem restoration potential. Therefore, the level of technical analysis is thus far limited. DWR Suisun Marsh Planning staff recommends that additional in-depth analysis should be pursued. Additional analysis of some screened alternatives is being conducted by a consultant using a two-dimensional hydrodynamics, water quality, and sediment transport model. This report may also be accessed from the DWR Suisun Marsh Brach web site at www.iep.water.ca.gov/suisun <../../index.html>. *II. Alternatives Investigated * Twenty-three individual historical simulations were conducted based on water year 1992. Each simulation tested a unique combination of levee rehabilitation, engineered small breaches, restoration of remnant channels, and channel widening. Attachment 1 <#attachment1> contains a complete listing and description of each scenario modeled. Scenario numbers in Attachment 1, and in the text of this report, correspond directly to the numbering of spatial salinity plot results included in Attachment 2 <#attachment2>. *III. Modeling Approach * The DWR Delta Simulation Model Suisun Marsh Version (DSM1) was used for the analysis. The downstream tidal boundary for DSM1 is at the Golden Gate, thus removing the influence of the boundary on the solution. The model was calibrated in 1997 under the auspices of the CALFED Storage and Conveyance Committee. The calibration included a revision and improvement of the model geometry based on up-to-date Delta and Suisun Bay bathymetry data. Complete documentation of the calibration process is available from the web site at www.iep.ca.gov/suisun/dsm1recal/status.html <../dsm1recal/status.html>. A historical condition simulation (base case) is used to provide a basis for comparison to the reclamation scenarios. The base case study was conducted for the time period between October 1990 and September 1992. This period was selected to allow evaluation of dry/critical period effects of the reclamation scenarios. The base condition was simulated for a two-year period in order to minimize initial condition effects on salinity output during the second year. All simulations were made using historical tides, hydrology, facilities operations, and agricultural consumptive use. All reclamation scenarios were simulated for the period between October 1991 and September 1992. The initial salinity condition provided to the model was estimated using the base case salinity at the end of water year 1991. To test the effects of running a one-year versus two-year model run, a two-year test run for water years 1991-92 was made for the Big Break scenario to compare to the 1992 only scenario. No significant difference in daily average salinity output was observed. Since this was the primary output used to guide this study, we were satisfied with the use of a one-year historical (water year 1992) run. Levee breaches in all cases are simulated as hydraulic weirs 100 feet wide and 20 feet deep with flow coefficients in the upstream and downstream direction of 0.9. Other breach configurations are indicated in scenario descriptions ( Attachment 1 <#attachment1>). *IV. Results * Spatial salinity plots were generated for each scenario that guided formulation of subsequent scenarios. The plots show the spatial percentage salinity change from the base case for Suisun Bay, Suisun Marsh, and the Delta (spatial salinity plots are in Attachment 2 <#attachment2>). Time series tidal day average salinity plots for the water year 1992 were also generated which verify that salinity differences between alternatives and the base case remain nearly constant for the summer and fall period of 1992. System-wide salinity on July 29, 1992 was selected for comparative analysis of spatial salinity. 1. Sherman Lake Scenarios The analysis initially focused on Sherman Lake including the remnant channel on the east end of the lake. Generally, the range of salinity increase or decrease compared to the Base Case in Delta locations for all scenarios was less than about 2% (Scenarios 1-4). A narrow breach on either the Sacramento ( Scenario 1 ) or San Joaquin River ( Scenario 2 ) results in slight salinity increases persisting into the south Delta. Hydraulic connection through a narrow breach from Broad Slough ( Scenario 3 ) results in slightly decreased salinity in the south Delta, but slightly increased salinity in the upper reaches of the Sacramento River. Three narrow breaches (along the Sacramento River, Broad Slough, and Mayberry Slough) scenario ( Scenario 4 <3narr.GIF>) results in nearly negligible change from the Base Case in areas away from the Sherman Lake area. Complete reclamation of Sherman Lake ( Scenario 5 ) causes a marginal salinity increase on the downstream end of the San Joaquin River, while salinity decreases on the lower Sacramento River below Emmaton, and increases above Emmaton. Changes elsewhere were small. Variation in tidal prism can indicate changes in the amount of salinity mixing in the Delta. Expanded tidal prism causes greater tidal excursion and increases the potential for mixing ocean water upstream. Tidal range increases in scenarios 1, 2, 4, and 5 on the Sacramento River in the vicinity of Steamboat Slough. This could explain the increased salinity observed in this area. Another set of scenarios focused on reclaiming Sherman Lake and restoring the former channel on the east side connecting the Sacramento River with Mayberry Slough (Scenarios 6-10). The Delta salinity impact from these scenarios is minimal with decreases in salinity in the south Delta of less than one percent. Restoration of the channel at the approximate historical width of 292 feet ( Scenario 6 ) results in slight salinity decreases in the south Delta and slight increases in the vicinity of the restored East Sherman Lake channel. Halving the width of the channel ( Scenario 7 ) increases salinity marginally throughout most of the Delta. Conversely, doubling the width of the channel ( Scenario 8 ) results in slight salinity decreases throughout most of the Delta. However, further widening of the channel to 1000 feet ( Scenario 9 ) does not appear to further decrease salinity in the Delta. Varying the width of the East Sherman Lake channel (146', 292', and 584') does not appear to significantly change the tidal range throughout the system. Imparting a more direct connection between the Sacramento and San Joaquin Rivers by extending the 1000 foot channel to the San Joaquin via Mayberry Cut ( Scenario 10 ) results in insignificant (< 1%) increase in salinity throughout most of the central and south Delta. 2. Three Mile Slough Scenarios The known hydrodynamic and salinity transport importance of Three Mile Slough led to the next series of scenarios focusing on the effect of channel width. Widening Three Mile Slough to 1000 feet ( Scenario 11 <3mile.GIF#scenario11>) results in small decreases in salinity in the south Delta while salinity in the eastern Delta slightly increased. However, the percent changes are small and are in areas with initially low salinity, thus changes are nearly negligible. Widening Three Mile Slough in combination with closing Sherman Lake ( Scenario 12 <3mile-nobr.GIF> ) and creating a 1000 foot East Sherman Lake channel ( Scenario 13 <3mile-esherm.GIF>) both result in greater salinity decreases (up to about 3%) in the south Delta and up to 9% in the lower San Joaquin River. Slight increases persist up the Sacramento River in both cases due to the increased net flow from the Sacramento the San Joaquin River. 3. Franks Tract Scenarios To assess the potential effects of restoration further east into the Delta, scenarios at Franks Tract were developed. Closing Franks Tract ( Scenario 14 ) results in significant decreases in salinity in the south Delta (up to about 30%) with only slight increases resulting elsewhere in the system. Similar decreases in salinity are seen with breaching Franks Tract on the east side near Mandeville Island ( Scenario 15 ). Sensitivity to east side breach size is small ( Scenario 16 ). 2000 square foot and 100 square foot reservoir (orifice equation) openings were simulated in separate runs. Results show similar salinity decreases of approximately 30%, with the smaller opening having slightly greater decreases. The east side Franks Tract breach was also simulated as a narrow (100 foot) weir opening on a wide one dimensional channel with volume characteristics matching the open water area ( Scenario 17 ). This approach allows hydrodynamic routing and salinity dispersion in one-dimension compared to the zero-dimensional "reservoir" approach typically used in the DSM1 model. Slightly smaller decreases in salinity in the south Delta result compared to Scenario 15. Since the reservoir approach allows mixing to occur instantaneously in reservoirs, there could be a reduction in the salinity differential between the reservoir and the adjacent channel. This would tend to decrease mixing (or "salt trapping"), resulting in the slightly less saline conditions experienced. These mechanisms should be investigated further. Recent levee breach studies have shown that breaches located along less energetic channels, such as the channel along the east side of Franks Tract, are not as sensitive to breach size as those along higher energy channels (CALFED Suisun Marsh Levee Investigation report in preparation). Breaching Franks Tract on both the east and west sides ( Scenario 18 simulated as two 2000 sq. ft. openings) results in salinity decreases of between 15 - 20% in the south Delta. In response to review comments, an additional scenario simulating a narrow weir opening on the west side of Franks Tract was made ( Scenario 24 ). The geometry of the breach and channel were identical to that of Scenario 17, the east side breach on Franks Tract. Resulting salinity reduction patterns and values of the east and west side breaches on Franks Tract are similar. This is likely due to the similarity of geometry and flows of the channels near the breach locations in both of these scenarios. 4. Big Break Scenarios Reclaiming Big Break ( Scenario 19 ) results in salinity decreases of approximately 10 - 12% in the south Delta with decreases of 5 - 8% persisting upstream on the Sacramento River. Breaching the mid-point of the remnant channel on the north side of Big Break ( Scenario 20 ) results in salinity decreases of about 8% near Clifton Court Forebay and also tends to slightly decrease salinity throughout most of Suisun Marsh. 5. Franks Tract - Big Break Combination Scenarios V.  Verification with RMA’s 2-D Results and DWR's DSM2 An independent verification of the results was desired acknowledging that the Delta breach scenarios modeled in this study are near the limits of DSM1 models' capabilities. Two independent verification efforts have thus far been completed using 1) a two-dimensional finite element Bay-Delta model (RMA 2 and 11) and, 2) the DWR DSM2 model. Resource Management Associates (RMA) used their two-dimensional finite element Bay-Delta model to simulate a historical four-month period between June and September 1992 using period average hydrology and a repeating June 1992 tide for two of the scenarios. This approach is necessary due to the high computational cost of using the RMA model. For purposes of comparison, the DSM1 model was executed in as nearly the same fashion as possible. Both the DSM1 and RMA 2-D output for the wide East Sherman Lake channel scenario ( Scenario 8 <#scenario8> ) result in similar daily average salinity values in the south Delta. However, the percent changes are small (generally less than 2% decrease in salinity). For the Big Break scenario ( Scenario 19 <#scenario19>), both models exhibit similar general salinity trends. However, DSM1 predicts salinity reductions in the south Delta approximately 5 times greater than the RMA results. When DSM1 is run in the same mode as the RMA model, even greater salinity reduction in the south Delta results. The reasons for this difference should be investigated further. In addition to the two-dimensional model verification runs, an independent modeling analysis was conducted within DWR using the DSM2 model and an alternative modeling approach to corroborate salinity trends. A CALSIM base study was used to provide monthly average Delta inflows and exports for the 16-year period from October 1975 through September 1991. The complete Franks Tract reclamation case was tested for comparison. Salinity reductions in the range of 15-20% were observed for Old River near Rock Slough, and 10-15% near Clifton Court Forebay. We believe that these results substantially corroborate those found using DSM1. *VI. Physical Mechanisms Causing Salinity Modification * The salinity response to island reclamation and engineered levee breaches discussed in this report results from the interplay of opposing hydrodynamic phenomena. First, changing the volume of the Delta by reclaiming Delta islands modifies the way tidal energy is dissipated resulting in modified tidal range and tidal excursion. Figure 25 shows an example how tidal prism is modified by a 100 foot San Joaquin River side Sherman Lake breach. In this case, tidal range is increased about 1 inch on the lower Sacramento River, and decreased less then an inch on the lower San Joaquin River. Complete reclamation alternatives (Scenarios 14, 19, and 21) increase tidal range 2 to 3 inches in the central Delta. Figure 26 shows the impact on tidal range of complete reclamation of both Big Break and Franks Tract (Scenario 21). In general, increased tidal range leads to increased tidal excursion (the distance the tide moves back and forth) and mixes more salt upstream. Second, the geometry of levee breaches and differences in tidal exchange on inundated lands compared to adjacent channels affects salt mixing. Salt can be "trapped," or more efficiently mixed, when water bodies containing different salinity concentrations mix together. This occurs when the ebb and flood of the tide is out of phase between inundated islands and the adjacent channel. The influence of this mechanism is reduced for scenarios that largely or completely reclaim formerly extensively breached islands (like Big Break and Franks Tract). Reclamation reduces salt trapping resulting in reduced channel salinity. Each mechanism exerts differential influence depending on unique site conditions, breach size and location, inundated land volume, and overall system geometry. *VII. Conclusions * Application of the DSM1 model as a screening tool effectively guided the search for feasible Delta island reclamation alternatives for further investigation. Combinations of island reclamation and engineered breaches were identified that provide opportunities for ecosystem restoration and central and south Delta salinity reduction. The investigation indicates that some Sherman Lake alternatives provide marginal salinity reduction. Big Break and Franks Tract alternatives show potential for significant central and south Delta salinity reduction. All of the participants in the modeling effort so far agree that further field and modeling analysis should be pursued to elucidate the mechanisms controlling the observed hydrodynamics and salinity response. Participants recommend that a multiple agency, university, and stakeholder team of hydrodynamics and water quality experts should be convened and afforded resources to improve modeling techniques, understand physical mechanisms, and identify optimal reclamation and levee breach configurations for ecosystem reclamation and salinity reduction benefits. ATTACHMENT 1 Alternatives Descriptions Sherman Lake Narrow Breach Scenarios: 1. Breach on Sacramento River Side – Sherman Lake closed with one 100 foot breach located on the Sacramento River side of Sherman Lake. 2. Breach on San Joaquin River Side – Sherman Lake closed with one 100 foot breach located on the San Joaquin River side of Sherman Lake. 3. Broad Slough – Sherman Lake closed with one 100 foot breach located on the Broad Slough side of Sherman Lake. 4. Three Narrow Breaches – Three 100 foot breaches located along the Sacramento River side, Broad Slough, and Mayberry Slough upstream of Mayberry Cut. Sherman Lake No Breach Scenario: 5. Closing Sherman Lake – Reclaiming Sherman Lake. East Sherman Lake Channel Scenarios: 6. East Sherman Lake Channel – Channel along the east side of Sherman Lake from the Sacramento River to Mayberry Cut. The channel is not connected to the Lake. Channel width of 292 feet. 7. Narrow East Sherman Lake Channel – Channel configuration as described above with channel width decreased to 146 feet. 8. Wide East Sherman Lake Channel – Channel configuration as described above with channel width increased to 584 feet. 9. "Ultra-Wide" East Sherman Lake Channel – Channel configuration as described above with channel width increased to 1000 feet. 10. "Ultra-Wide and Long" East Sherman Lake Channel – Channel configuration as described above and included widening Mayberry Cut to 1000 feet. Three Mile Slough Scenarios: 11. Three Mile Slough – The width of Three Mile Slough was increased to 1000 feet with no change to the existing configuration of Sherman Lake. 12. Widening Three Mile Slough and Closing Sherman Lake – Combination of 11 and 5 above. 13. Widening Three Mile Slough and Creating 1000 foot E. Sherman Lake Channel – Combination of 11 and 9 above. Franks Tract Scenarios : 14. Franks Tract – Reclaim Franks Tract. 15. Franks Tract Narrow Breach (East side only, modeled as reservoir) – Simulated 2000 sq. ft. breach on east side of Franks Tract near Mandeville Island. Franks Tract modeled as a reservoir. 16. Franks Tract 100 sq. ft. Breach (East side only, modeled as reservoir) – Simulated 100 sq. ft. reservoir opening on east side of Franks Tract near Mandeville Island. 17. Franks Tract Narrow Breach (East side only, modeled as channel) – Simulated 100 foot breach on east side of Franks Tract near Mandeville Island. Franks Tract modeled as wide, shallow channel. 18. Franks Tract Double Breach (East and West sides) – Simulated two 2000 sq. ft. breaches on east side as described above, and on west side near Bethel Island. Franks Tract modeled as a reservoir. 24. Franks Tract Narrow Breach (West side only, modeled as channel) (Additional run made in response to review comments, June 7, 2000) – Simulated 100 foot breach on west side of Franks Tract near Bethel Island. Franks Tract modeled as wide, shallow channel. Big Break Scenarios: 19. Reclaiming Big Break – Remove Big Break channels. 20. Big Break Narrow Breach – One 100 foot breach at midpoint of historic channel on north side of Big Break. Franks Tract & Big Break Combinations: 21. Closing Franks Tract and Big Break – Combination of runs 14 and 19 above. 22. Franks Tract Double Breach and Big Break Narrow Breach – Combination of runs 18 (Franks Tract breached at two locations with 2000 sq. ft. openings) and 20 (Big Break simulated with a narrow, 100 foot wide breach) above. 23. Franks Tract 100 sq. ft. Breaches and Big Break Narrow Breach – Reservoir openings on node 232 and 219 set to 100 sq. ft. opening size. All other Franks Tract reservoir openings closed as in run 14. Big Break narrow breach (100 foot wide) as described in run 20. ATTACHMENT 2 Spatial Salinity Plots Figure Number _Spatial Salinity Plot Scenario_ 1. Breach on Sacramento River Side 2. Breach on San Joaquin River Side 3. Breach on Broad Slough Side 4. Three Narrow Breaches: Sacramento River Broad Slough, and Mayberry Slough <3narr.GIF> 5. Close/Reclaim Sherman Lake 6. East Sherman Lake Channel; 292 feet . 7. East Sherman Lake Channel; 146 feet . 8. East Sherman Lake Channel; 584 feet. 9. East Sherman Lake Channel; 1000 feet. 10. East Sherman Lake Channel; 1000 feet including Mayberry Cut. 11. Three Mile Slough width 1000 feet; existing Sherman Lake configuration. <3mile.GIF> 12. Three Mile Slough width 1000 feet; Close Sherman Lake. <3mile-nobr.GIF> 13. Three Mile Slough width 1000 feet with 1000 foot E. Sherman Lake Channel <3mile-esherm.GIF>. 14. Franks Tract complete reclamation . 15. Franks Tract east side 2000 s.f. breach near Mandeville Island . 16. Franks Tract east side 100 s.f. breach near Mandeville Island . 17. Franks Tract east side 2000 s.f. breach near Mandeville Island; modeled channel . 18. Franks Tract east and west side 2000 sq. ft. near Mandeville and Bethel Islands . 19. Big Break complete reclamation . 20. Big Break 100 foot breach at midpoint of Dutch Slough channel on north side . 21. Franks Tract and Big Break complete reclamation . 22. Franks Tract Double 2000 sq. ft. breach and Big Break 2000 sq. ft. breach . 23. Franks Tract 100 sq. ft. Breaches and Big Break 2000 sq. ft. breach. 24. Franks Tract west side 2000 s.f. breach near Bethel Island, modeled channel. Tidal Range Plots 25. 28-Day Average Tidal Range, San Joaquin River Breach to Sherman Lake 26. 28-Day Average Tidal Range, Big Break and Franks Tract Reclamation