2003 0722 CSUMB Report BCurry.pdf - The Pajaro River Watershed

California State UniversityRobert Curry, Research DirectorWatershed InstituteEarth Systems Science & PolicyCSU Monterey BaySeaside, CALIF. 93955Bob_curry@csumb.eduWatershed Restoration Class – Spring, 2003Pájaro River WatershedFlood Protection PlanWm BodensteinerLani CloughSuzanne GilmorePaul HuntingtonJoy LarsonApril McMillanSteve MackC. Andrew MauckSerena PringEmily RothAmy ThistleMelanie VincentDRAFT OF July 22, 2003 A1 Public Copy

Executive SummaryBecause of the unique geologic and hydrologic setting of the Pájaro River inits dynamic watershed, traditional approaches to flood control may not beeffective and will require constant expensive maintenance. The river thatnow flows through it did not create the lower Pájaro Valley and it is notpossible to “restore” such a system to stability because there is no evidenceof any past stable Pájaro River channel in the lower valley. An artificial floodcontrol channel was constructed by early residents and was upgraded by theU.S. Army, and later by the Corps’ of Engineers to try to minimize propertylosses associated with large floods in this watershed of about 1300 squaremiles. Historically the Pájaro watershed system has carried runoff fromSanta Clara, San Benito, Santa Cruz, and Monterey counties into MontereyBay through various channels in Monterey County. The river is nowartificially confined to join Corralitos Creek to enter the ocean along the SantaCruz/Monterey County border.We find that a substantial area of on-channel storage of floodwater has beenlost in the upper watershed areas of San Benito and Santa Clara counties.Some of this lost storage can be recovered for little or no public cost toreduce flood heights (on the order of 4 feet) in the artificial floodway channelof the lower river. Redesign of that lower channel may accommodate addedflood capacity to provide a working flood channel that carries a generouslyestimated 100-year flood volume. Such redesign, coupled with upstreamchannel restoration that is part of a flood storage enhancement project, willhave very substantial wildlife and water quality habitat benefits.DRAFT 7/22/032Pájaro Watershed Flood Management

Table of ContentsExecutive SummaryTable of ContentsChapter 1Introductory ContextThe Pájaro Watershed System DynamicsThe WatershedLower Watershed – Santa Cruz and Monterey CountiesUpper Watershed – San Benito and Santa Clara CountiesThe River SystemStable Channel AlternativesThis Project ReportCoordination with otherRaines, Melton, CarellaPhilip William AssociatesU. S. Army, Corps of EngineersiiiChapter 2Design Flood AnalysesPurposeMethodsData CollectionData AnalysisResultsRegional AnalysesDiscussionStorm Patterns and HistoryChapter 3Upper Basin In-channel Flood Storage and Restoration OpportunitiesBasic ConclusionsTheoryMethodsFindingsChannel IncisionChannel DiversionsRestoration of Channel FunctionsSuggested Restoration Options for the San Benito RiverAggregate Mining Company OpportunitiesSuggested Enhancement Options for the Upper Pájaro RiverConclusionsReferences Cited and Historical Materials ConsultedDRAFT 7/22/033Pájaro Watershed Flood Management

FIGURES AND PLATESFig 1 Map of Lower Pájaro River and Elkhorn areas showing historical changesFig 2 Historical aerial photo of Pájaro Valley showing landslidesMap A Rancho Vega del Rio Pájaro and Northern Monterey County 1875Map B Land Ownership in Vega area about 1910Fig 3 Historical (1938) aerial photo showing past breakout areasFig 4 Historical (1938) aerial photo showing meander wavelength in lower valleyFig 5 Map of Pleistocene Lake San Benito (from Jenkins)Fig 6 Calculated flood discharge frequency/magnitude plot at ChittendenFig 7 Monterey Herald photograph of 1938 Lower Pájaro floodingMap C FEMA 100-year flood map and map of areas considered for flood storageaugmentation in this reportFig 8 Soil profile in active overbank storage areasFig 9 Lake San Benito soil profileFig 10 Example of area that can be restored to flood storageFig 11 Effects of channel incision on channel stability and habitatFig 12 Cienega Road HouseFig 13 Llagas Creek Channel photosFig 14 Gabion Basket representationFig 15 Stream Barb representation with Gabion BasketsFig 16 Detailed topography of a portion of Lower San Benito River near Highway 101Fig 17 San Benito River Channel topography near Mitchell RoadAPPENDICESApp 1. Note on higher recorded 1998 flow at Highway 156 than downstream atChittenden gauge from L. Freeman, USGSApp 2. Analysis of 1998 and 1995 storm conditions with map of precipitation stationsused in the analysisApp 3. Streambank property owners in San Benito County (separate file)App 4. In preparationApp 5. Example of Historical Changes in Lower San Benito River (Figs 18-21)App 6. Topographic detail for a cross section in the App 5. Historical Change areaApp 7. Mines in the San Benito County permit files (separate file)App 8. Economics and Socioeconomic SettingsDRAFT 7/22/034Pájaro Watershed Flood Management

Pájaro River Watershed FloodManagement AlternativesA study by the CSUMB WatershedRestoration Class, Spring 2003CHAPTER 1Introductory ContextThis report is a group effort of 12 upper level students who have focused much of theireducation on Watershed Science through the Earth Systems Science and PolicyProgram at California State University Monterey Bay. Some participants had alreadygraduated from CSUMB or UC Santa Cruz; while most were finishing seniors witheducations that included advanced hydrology, water law, and riparian ecology. CSUMonterey Bay stresses an “outcomes-based” education with active, applied learning.This work is not financially supported, but a small anonymous donation of $500helped with copying and telephone costs. The Santa Clara Valley group “People forLivable and Affordable Neighborhoods” supported a detailed watershed map madeespecially for this effort by Eureka Cartography in Berkeley. San Benito County andthe Graniterock Company contributed map and data resources.This report follows the theme of our educational program and treats the PájaroWatershed as a physical and biological system. We take the position that it is notpossible to isolate the processes and problems in the lower watershed from thecausal mechanisms in the upper watershed. We look at the watershed as a completesystem with material and energy flows that support living ecosystems and organisms.We assess the causes of dysfunction, which in this particular case focuses onresponses of humans to flooding and sediment transport, and evaluate potentialsolutions utilizing fundamentals of fluvial geomorphology and restoration ecology.This particular study was undertaken in the context of significant fundamentaldisagreements between residents, agencies, and government entities. Following theCalifornia Supreme Court finding that upheld lower court’s rulings against Countygovernments for causing flooding in 1995 through lack of required maintenance,Monterey and Santa Cruz counties requested that the U.S. Army Corps of Engineersconsider a new flood control project to protect the downstream areas from 100-yearreturn-period floods. This class effort focused on the opportunities to reducedownstream flood hazards through upstream flood detention and through design of astable channel alternative in the artificially constrained lower reaches of what is calledthe Pájaro Valley.DRAFT 7/22/035Pájaro Watershed Flood Management

The Corps’ had been requested to evaluate protection options for only the lower riversystem working with only the lower county governments, and further constrained bylimited budgets and the necessity to work with a set of “stakeholders” whorepresented diverse and often contradictory viewpoints. With this impossible set ofconstraints, landowner, environmental, and agency views that had seemed in conflictwith each other soon refocused on conflict with the Corps’ themselves, who ultimatelywere left to represent only the county governments who had brought them into theproject.This report is now presented simultaneously with the Corps of Engineers flood controlproposals and with a citizens’ sponsored and funded set of alternative flood protectionsolutions produced by the renown hydrologic consulting firm of Philip WilliamsAssociates. It is hoped that this university effort can help to expand the very limitedscope of the many other ongoing and recent studies to create viable alternatives inthis very complex watershed system.The Pájaro Watershed System DynamicsThe Watershed: At present the Pájaro River watershed drains an area ofapproximately 1300 square miles. The watershed primarily drains the counties of SanBenito, and Santa Clara, with some added contribution from Santa Cruz County. Verysmall areas of Fresno and Monterey counties are also within the watershed butcontribute very little to the runoff. About 91 percent of the watershed is in NorthAmerica while the outlet in the Lower Pájaro Valley and the Corralitos andWatsonville Slough tributaries are on the Pacific Plate. Due to active faulting within thewatershed boundaries, the rivers’ coarse is continuously changing and has notstabilized in a valley of its own construction. The San Benito River is now 51% of theentire Pájaro drainage area but contributes only about 25% of the runoff at Chittenden(an average of 49 ac-ft/an/sq.mi.) The Pájaro above the San Benito junction (atSargent) contributes about 180 ac-ft/an/sq.mi from 39% of the basin.Corralitos/Salsipuedes tributary is only about 3% of the watershed but contributes onthe order of 435 ac-ft/sq.mi, or nearly 10% of the total discharge of the Pájaro system.Constructed reservoirs have a maximum capacity of 42,680 ac-ft (Hernandez:18,500;Uvas:9950; Chesbro:8090; and Pacheco:6140). We estimate that about 60,000 ac-ftof near-channel flood storage also exists in areas that are subject to overbank or inchannelflood storage or were 50 years ago. About 24,000 ac-ft of lost storage can bereadily restored at little or no public cost.A map of the watershed that incorporates the detailed findings of this report isavailable on-line in a medium-resolution 10 MB and low resolution 700 KB version athttp://home.csumb.edu/c/currybob/world/Pajaro/ where this report itself and some ofits graphics is also available. This watershed map utilizes the existing left-bank leveeof the lower river as the watershed divide between Elkhorn Slough and the Pájarowatersheds.Lower Watershed, Santa Cruz and Monterey Counties: The Pájaro Watershed isunusual. Traditional engineering solutions must accommodate the unique geology andhydrologic character of the basin. The headwaters of the basin are in North Americabut the primary plate boundary represented by the Calavaras and San Andreas Faultzones separates the mouth of the present river from its historic source areas. Activetransform faulting has repeatedly and progressively modified the course of the riverDRAFT 7/22/036Pájaro Watershed Flood Management

that today bears the name Pájaro. The unusual shape of the watershed itself, with along source area far south of the outlet is the result of continual stretching of thewatershed by active faulting that pulls the lower river northwestward, farther andfarther from its headwaters.Much of the lower river, west of the San Andreas Fault Zone, does not flow in a valleyof its own making. The original course of Corralitos Creek in Santa Cruz County (seeFig. 1) and its alluvial aquifer have now been taken over by the Pájaro River system.The ancestral Pájaro River has been repeatedly offset northward by right-lateral faultoffset, sometimes emptying to the coast through Elkhorn Slough at Moss Landing,and other times commingling with Corralitos Creek as it does today. California’s StateGeologist, Olaf Jenkins (1973) postulated that landslides near Chittenden Gap,forming Lake San Benito and later Lake San Juan that repeatedly spilled and scouredoverflow channels in the Carneros Creek/Elkhorn Slough area, might have repeatedlydammed the main river. Even today, during flood stage, the lower river flows to thesea at Moss Landing. Jenkins reasoned that these changes are geologicallycontemporary, having occurred in the last few thousand to 20,000 years at the most.Fundamental evidence for the very young character of this lake and its overflow is thefact that the lake shorelines are evidently not evidently tilted or deformed, despitebeing astride two active faults, and finding that the lake sediments contain a fullycontemporary local flora and fauna.Fig 1DRAFT 7/22/037Pájaro Watershed Flood Management

Figure 1 represents a slight modification of the original Jenkins map (Curry, 1996) witha series of name changes to better reflect the geologic evolution of the present lowerPájaro River as it spilled through Chittenden Gap to overwhelm any preexisting localwatercourses. It is critical to appreciate that Corralitos Creek and its presumedtributary Aromas Creek did not capture the Pájaro River, but instead a great lakedammed by faulting and/or landslides spilled catastrophically into what we now callthe Pájaro Valley. This explains the lack of terraces and floodplain deposits in thelower Pájaro Valley, and the massive Lake San Benito silts that now blanket the lowervalley to support its agriculture.Because the river that now flows through it did not form the lower Pájaro Valley, thewatercourse is inherently unstable. Fluvial geomorphology recognizes this conditionas “overfit”, with the natural watercourse being too big for its channel. Coupled to thisinherent instability is the fact that the lower Pájaro Valley is traversed by the SanAndreas Fault and the subsidiary Zayante-Vergeles fault system (R. Anderson, 1990).These are all among the most active terrestrial fault systems on the North Americancontinent. The 1989 Loma Prieta earthquake apparently deformed the Pájaro Riverlevee system (personal survey notes). Today the lowest point in the Pájaro Valley isnot the Pájaro River but is a small overflow watercourse along the extreme south sideof the lower valley. Based on undercutting of the hillsides at the south edge of thepresent Pájaro Valley and preserved cutoff meanders there, the southernmost edge ofthe valley has been the lowest point for at least several hundred to several thousandyears (see Fig. 2).It is thus perplexing that the present river course and levee system coincide with thelower Corralitos Creek channel. Based on the early maps made shortly afterstatehood in 1850 and local place names, a grazing wetland commons existed in theMexican Ranchero period in the area still known as the Vega (see Map A, RanchoVega del Rio Pájaro, Map B). The vega meadows here were apparently flood irrigatedregularly to constrain land use and thus provided a grazing Mexican land grant untilStatehood and private (Porter) ownership. The Vega is adjacent to a spot on theoriginal river (see Map A) where the river was straightened after the boundarybetween Santa Cruz and Monterey counties was established (California HistoricalSurvey, 1923). An alluvial thalweg (central river channel) is now buried beneath thelevee system and has been the locus of flood outbreaks from at least the 1930sthrough 1995 (see Fig 3). All of the positions of today’s levees crossing the 1854channel position are sites of piping and passage of river water under the levees duringhigh water as seen in 1995 and 1998 (personal observation, R. Curry and landownerdiscussions).DRAFT 7/22/038Pájaro Watershed Flood Management

Map B. 1908 Parcel Map of a portion of the Lower Pájaro Valley showing the historicVega area and dot-dashed County boundary as it exists today.DRAFT 7/22/039Pájaro Watershed Flood Management

Figure 2 -- 1939 Photo of Lower Pájaro Valley. Watsonville in lower right. Thelandslides are readily seen at the position of Highway 1 today, near the center left ofthe photo. Also visible are the flow lines from past floods that impinge against the left(south) side of the valley.DRAFT 7/22/0310Pájaro Watershed Flood Management

Map A 1875 based on 1854 land surveyIt may be that in the Ranchero and early statehood period, the lower Pájaro River waschannelized to try to restrict regular overbank flow in distributary channels so that landuse could be made more efficient. Looking at the 1939 and earlier aerial photos, westill see clear evidence of those distributaries (cf Fig. 3). The earliest detailedtopographic map (Capitola Quadrangle, 1912) shows “Watsonville Creek” that flowsfrom the left bank of the Pájaro River across that river from Salsipuedes Creek inWatsonville, directly south near Salinas Road and into Elkhorn Slough. That channelis still there and still carries rainfall and flood overflow runoff to Moss Landing. Runofffrom a major part of the townsite of Pájaro does not enter the Pájaro River today butflows via “Watsonville Creek” to Elkhorn Slough. The confusing topography wascommented on by William Brewer in his diary in 1864 that noted that the flat valleylooked like "an old lake filled in as is shown by the terraces around its sides."(Farquhar, 1930). Olaf Jenkins identifies a “Lake Pájaro” and “Lake Aromitas” in theold lower Pájaro Valley (1973).DRAFT 7/22/0311Pájaro Watershed Flood Management

Upper Watershed, San Benito and Santa Clara Counties: The upper watershed of thePájaro system is at least as complex as that of the portion west of the San Andreas Fault.There is indirect geologic evidence that Santa Clara Valley from San Jose southwardthrough Morgan Hill and Gilroy may have been the course of a major river carrying coarsegravels southward toward the present Pájaro River and that a lake in San Benito Countylater spilled northward along Coyote Valley into San Francisco Bay (Iwamura, 1995). Anopen and porous alluvial gravel characterizes the near surface substrate beneath both thenorth-flowing Coyote Creek and the south-flowing Llagas and Uvas Creek valleys. A verylow gradient “watershed divide” near Morgan Hill has southward flow in a shallowsubsurface aquifer, presumably recharged by Santa Clara Water District facilities fromCalifornia Water Project sources (Anderson Reservoir) and from locally captured anddiverted watercourses. Where this shallow gravel aquifer is exposed in the bank of thePájaro River, along the westernmost Santa Clara -- San Benito County border, manycubic feet per second of water flow continuously into the Pájaro River. These high watertables were recognized long before the San Luis Project brought Mt. Shasta water intosouthern Santa Clara and northern San Benito counties. The high groundwater levels arerecognized as a particular agricultural problem in San Benito County (Jones & Stokes,1998) where some are saline.The thick uniform silt deposits of Northern San Benito and Southern Santa Claracounties are themselves enigmatic (see Fig 5 from Jenkins). Jenkins refers to themas “Pleistocene” meaning of Pleistocene age (greater than 10,000 years ago) anddraws parallels with glacial age origin silts. Indeed, the surface deposits of lakebedsilts are remarkably uniform fine sandy silt similar to glacial origin rock flour in bothtexture and lack of chemical weathering. But calling upon an ancestral San JoaquinRiver system to deposit these silts from the Sierra Nevada is, at present, notdemonstrated. Jenkins hypothesizes that the silts may be derived locally from theolder Purisima Formation (locally now called the Etchegoin Formation east of the SanAndreas Fault). Subsurface deposits of northern San Benito County are characterizedby localized sands and gravels that appear to be river deposits embedded in siltsformed in shallow ephemeral lakes (Stanley, et al, 2002; Jones & Stokes, 1998).These are then buried by the more uniform overlying silt lakebeds. It is these surfacelake silt unit(s) that have been transported downstream to blanket the lower PájaroRiver Valley. It is not clear that they are being eroded from agricultural fieldsupstream, and may simply be carried in flood flows from upstream bank erosion.DRAFT 7/22/0313Pájaro Watershed Flood Management

Figure 5 Jenkin’s map Lake San Benito with its tectonic settingThe Calavaras, San Andreas, and Sargent fault zones define much of the course ofthe present tributaries of the upper Pájaro River system. These right-lateral strike-slipplate-bounding fault systems essentially lengthen the headwaters of the Pájaro River,repeatedly moving the upper river system southward 10’s of kilometers relative to thePacific Plate. The Old San Juan Stage Road between Salinas and San Juan Bautistaappears to follow an abandoned course of what is now called the San Benito Riverafter that river was pulled northward on the west side of the faults to join the upperPájaro River. All of this may have happened during as little as a few hundred orthousand year period of lakes being dammed and spilling before the river ultimatelybroke through the Chittenden water gap to spill westward rather than southward. It isinteresting to note that this rare example of a true water gap in western United Statesis actually called “Chittenden Pass”. A water gap is a pass through a mountain rangeor ridge cut by water. These are generally found in places like the Appalachianswhere a very old river is able to keep flowing while mountains are arched upwardbeneath it or while erosion lowers the river across a buried bedrock feature.Chittenden Pass is indeed a narrow part of the new river valley but cut bycatastrophically spilling water.The River System: No other reasonably large North American river drains awatershed that is as complex or as geologically active as the Pájaro . Only in theDRAFT 7/22/0314Pájaro Watershed Flood Management

Himalaya and Alaska are there possibly watersheds of greater than 1000 squaremiles with an equal level of active watercourse displacement and contemporarychanges in drainage area, and those rely in part on glaciers to block and divert thefaulted landscapes. The Pájaro is unique in that geologic activity must be factored into an understanding of the dynamics of flood hazard evaluation in a populated area.Ongoing geologic deformation renders constructed features like levees and channelsvery impermanent. Stream gradients and streambed elevations are changing by feetper century from non-anthropogenic causes (cf, 1906 earthquake and loss ofnavigability of Elkhorn Slough to the commercial steamer carrying Watsonville cargoto Moss Landing, Loma Prieta earthquake, creep on the Calavaras fault). Traditionalapproaches to flood hazard mitigation must accommodate this constant change.Stable Channel Alternatives: Stable channel concepts are almost a tautology in aconstantly changing watershed system. But because we have 65 year-old or olderaerial photos of almost the entire watershed, we can find evidences of thecharacteristics of river channels and flood patterns preserved from the time beforelaser leveling and powerful tractors. Many of the historic areas of lowland floodingand lake silts throughout the watershed were initially farmed as orchards. Uplandswere used for hay and barley. The lower Pájaro Valley was noted for its apples andthe upper valleys for walnuts (Crosetti, 1993). These seasonal crops were tolerant ofwinter flooding, seasonal root saturation, and some aggradation. Access to farmlandswith mechanized equipment and safety of grazing animals led to efforts to straightenchannels and, as elsewhere in the world, to shorten channels and cut off meanderloops. The 1854 Coast and Geodetic Survey mapping, later expanded in the 1870’sto include more inland areas through the U.S. Lands Office, showed that the Pájarohad been altered by the time of statehood. The 1854 survey, at a scale of 1:10,000, isaccompanied by survey notes (Wm. M. Johnson, 1854) that state: “Extending from themouth of the Pájaro River to the Salinas River is a range of low sand hills betweenwhich and the older formation lay several ponds. These mark the former bed of thePájaro, it having evidently at one time, found its way to the ocean through thischannel, but by an accumulation of its waters, during the winter months, it burst thenarrow strip of beach which separates it from the sea, and thus formed itself a newmore direct outlet”. By 1909, the Coast and Geodetic Survey report noted that thePájaro River “has low but well-defined banks and there is no evidence of recentchanges in its course” (1910 C&GS survey notes). Those coastal surveys generallyextended only 2.5 miles inland.Maps of Santa Cruz and of Monterey Counties were prepared in the 1870’s and areon file in the University of California Santa Cruz map library (see list in ReferencesCited). An example is shown as Map A. It is important to appreciate that the riverplan form shown in these early commercial maps was based on earlier US LandOffice plat maps and the Coast and Geodetic surveys. It is the County boundarymaps that show accurately the changes in position of the Pájaro River and that mustbe used for the actual position of the river (California Historical Survey Commission,1923). Based on that definitive reference, the channel of the Pájaro had beenstraightened shortly after Statehood and continued to be altered through the late1800’s.Based on geomorphic understanding of the relationships between a river and itsnatural floodplain, one can establish a channel geometry that, for a given gradient andsediment load, can approximate the shape of a channel that is self-maintaining (Curry,DRAFT 7/22/0315Pájaro Watershed Flood Management

1981; Riley, 2003). Of course, the lower Pájaro River does not have a floodplain inthe normal sense of a surface of deposition and transportation of sediment and waterthat exceeds the effective dominant discharge of the river system. The lower PájaroValley land surface is a flood-deposit, but not one formed through an equilibriumrelationship between its river and its flood regime (see Whiting, 1998). Thus, use ofstandard hydrologic relationships between flood frequency and magnitude to estimateideal channel dimensions and form may be limited in applicability. Not only is the riverchanging in length because of human channel shortening, but also the seaward limitof the river mouth has moved inland many 10’s of meters since the first 1854 survey(1910 C&GS survey notes). Further, tectonic deformation may be tilting the wholelower Pájaro Valley and surroundings southward. Still further, changed drainageareas in the upper watershed and incision of watercourses are apparently increasingthe ratios of runoff to rainfall.But use of historic aerial photos to interpret pre-channelization or flood-time flowpatterns can provide clues to the “natural” channel form that the Pájaro would take ifunconstrained. As pointed out by outside Corps of Engineers project review teammembers (USCofE, 1998), the current levee-constrained channel may not reflect astable channel configuration. British work, funded through the US Army Corps ofEngineers, has concluded that, as a general rule in sand-bed rivers, the mean annualdischarge and the bankfull discharge form lower and upper bounds, respectively, tothe range of effective discharge, while the 2-year flow is an upper bound to the rangeof bankfull discharge (Soares, cite).Ron Copeland provided a contribution to the Corps’ Project Review Team report forthe lower Pájaro Project (USCoE, 1998). He suggested that use of a channel-formingdominant discharge with a probability of recurrence of 1.5 to 2.0 years could permitestimation of ideal bankfull width and meander wavelength for a given gradient,roughness, and sediment load regime. That is the same approach as described byRosgen in his Fig 1 (see next) (Rosgen, 1996). It has real merit. Copeland includedFig 4 from Akers and Charlton, 1970, in his contribution to the Pájaro review teamreport (figure follows). Using a calculated (Fig 6) discharge for a 2.0-year returnperiod at Chittenden, we calculate that the dominant channel-forming flow that shouldequate to bankfull discharge in a stable channel is about 3500 cfs. Using that value inthe Ackers and Charlton figure yields a stable meander wavelength for a channelunconstrained laterally by levees with a value of 1000 to 1500 feet. That is what wesee in the historic overflow channels on the old aerial photos (Fig 4), and in the earlyhistoric maps of the river platform. Thus there is a corroboration of theory andsystems function in the lower Pájaro River channel, despite the unusual nature of therelationships between the watershed and the areas subject to flooding.DRAFT 7/22/0316Pájaro Watershed Flood Management

DRAFT 7/22/0317Pájaro Watershed Flood Management

99% = 1.010 yr95% = 1.053 yr50% = 2 yr5% = 20 yr1% = 100 yr0.2% = 500 yr10000010000Magnitude1000100Pajaro River at Chittenden 1940-2000101-3 -2 -1 0 1 2 3Standard normal deviate of probability of excedanceFigure 6: Plot of actual peak floods (X‘s) versus LogPearson Type III calculatedvalues (open Circles). This is not calculated using the required methodology, as donein Chapter 2.This Project ReportWhen the original U.S. Army Engineers flood control project was begun in 1943 andcompleted in 1948, all 4 counties in the watershed signed off on an agreement toaccept responsibility for maintenance of the flood control works in accord with adetailed maintenance plan prepared by the Army (Secretary of War, 1944). In the1960’s the upstream counties, under the organization of Santa Clara County,requested a Congressional exemption from the earlier agreement (Secretary of theArmy, 1965), and it was granted. This political context prevented several efforts todevelop a watershed-based joint powers authority to manage the watershed after theMarch, 1995 floods that took one life in Pájaro and caused many millions of dollars oflosses in the Lower Valley.Congressional efforts in response to landowner concerns following the 1995 and 1998floods lead to appropriations for, and efforts by the Corps’ to review and revise theflood control project. Because of the failures of prior efforts to solicit cooperation fromupstream counties, it was deemed politically necessary to restrict the scope of floodcontrol efforts to a downstream project that simply rebuilt the original 1948 projectwithin the same reaches of the Lower Pájaro River that had been the subject ofstructural efforts in the past (Congressman Sam Farr, personal communication).DRAFT 7/22/0318Pájaro Watershed Flood Management

The current project attempts to rectify the inability of the government efforts toconsider solutions that most effectively and economically deal with the river systemrather than only the lower river reach. While this is a suitable context for investigationby an academic institution, it also provides a public service outside the context ofpolitical limitations of elected and regional persons and bodies. Because the Corps’must complete an environmental impact statement and analysis for their proposedlower river project, the opportunity to think outside of the artificial box can be requiredthrough § 102.2.c of the National Environmental Policy Act. This project documentseeks to provide some bases for that required analysis.We approach this task through the following primary foci:1. An analysis of the design flood magnitude and duration that must beaccommodated by any lower river protective works.2. An assessment of potential opportunities for reducing those flood flows throughenhanced upstream flood storage using natural or small-scale structuralenhancements that will increase wildlife habitat and amenities for upstreamlandowners and governments in order to encourage their implementation.3. Analysis of the unique geologic and hydrologic characteristics of the presentconfiguration of the Pájaro Watershed as they control and limit options for floodhazard reduction.4. Compilation and preparation of a comprehensive database on the watershed indigital format that can be shared by the 4 counties and the interested public.Additional analyses for the economic feasibility of combinations of upstream anddownstream flood mitigation efforts, the political economic driving forces that need tobe acknowledged and accommodated to make a watershed-wide flood controlsolution work, the roles of federal and state agencies in permitting and regulatingeffective solutions, and the environmental constraints and restoration opportunitiesafforded by a watershed-wide flood control project are also woven into the fabric ofthis report.Coordination with ongoing workRaines, Melton & Carella, Inc. (RMC) have been contracted through the PájaroRiver Watershed Flood Prevention Authority, formed through coordination of theAssociation of Monterey Bay Area Governments (AMBAG) to consider opportunitiesto increase upstream flood storage through modification of existing reservoirs orconstruction of new flood control dams. Their first report is available through AMBAGand, for a limited time, on their website: http://www.rmcengr.com/Pages/prwfpa.htm(Phase I). RMC conducted standard hydrologic modeling of effects of urbanization inthe largely rural upper watershed, and assessed costs of new or rebuilt conventionaldams that could provide some flood control benefits. The findings basicallydemonstrate that build-out in San Benito County has little net effect on countywideand watershed-wide runoff volumes, and that costs for old-style flood control damsexceed benefits. One finding of the initial RMC study became the focus of aconcurrent Phase III study looking at the ephemeral Soap Lake wetland area alongthe upper Pájaro River and lower Llagas and Uvas creeks. RMC concluded that thisDRAFT 7/22/0319Pájaro Watershed Flood Management

natural ephemeral basin provided on the order of 30,000 ac-ft of storage and that,without it, flood peaks at Chittenden would increase about 137% for the 100-yearevent. The RMC Phase II study, also available now, looks at alternatives in the lowervalley for bypass and underground floodways and compares them to the variousCorps’ proposals for levee modification.Our work also looked at Soap Lake and considered alternatives for enhancing floodstorage in a portion of that feature. We did not assume that diminished developmentpressure or conservation-flood easements could preserve all of the existingoccasionally flooded agricultural land, and thus looked at compensating alternatives toallow some levels of development and new highway construction. AMBAG and theWatershed Flood Prevention Authority are exploring flood easements for the core7900 acres of the site.Philip Williams and Associates, Ltd. (PWA) were contracted in June of 2003 by theSierra Club to investigate alternatives not considered by RMC or by the Corps’ aspublicly revealed to that date. The PWA report, being released simultaneously withthis report, considers a series of downstream flood mitigation scenarios and linkssome of them to opportunities for enhanced upstream flood detention to reducedownstream costs, environmental losses, and maintenance. The PWA studiesconsider stable channel alternatives as well as constricted high-maintenancechannelization options to provide a wider range of alternatives than have beenpublicly discussed by any entities to date. Among the options considered by PWA isone proposed by state and federal regulatory agencies to regrade the channel to a“self-maintaining” form. It is designed to transport sediment through the systemwithout mechanized assistance, and tries to meet stated goals and objectives of thesepublic agencies that must review and approve any chosen alternative.U.S. Army, Corps of Engineers (Corps’) is the lead agency for the downstream floodcontrol project. The Corps’ has been involved repeatedly following the initial projectcompletion immediately after WW II. Their charges include annual monitoring andoversight of levee and channel maintenance, repair and resurvey after the 1989 LomaPrieta Earthquake and the 1995 and 1998 floods, and design and constructionoversight of any new flood control project that modifies or replaces their originalproject. City and County governments and citizens have nearly continuouslyrequested intervention and design improvements for the Corps’ projects that protectthe City of Watsonville and the lower Pájaro flood channel. As was revealed in the1997 trial of CalTrans for ponding of flood waters associated with the 1995 floods, theState of California had always assumed that the Corps’ had responsibility for 100-yearflood protection for the entire Pájaro Valley and, thus, that highways crossing thatvalley at its lowest point need not accommodate any but local rainfall runoff beneaththe highway berm. The Corps’ has held repeated public informational meeting andtried to use a “stakeholder” process to consider concerns of the lower Pájaro Rivercommunities. A very considerable effort was initiated in 1998 by the Corps’ tocritically review past and anticipated future activities of the agency using a nationwidein-house professional team (United States Army, Corps of Engineers, 1998), but the publichas not seem much response from the Corps’ to that foundation report. The agencywill again attempt to provide a series of alternatives and choose one for final preferredevaluation during July 2003.DRAFT 7/22/0320Pájaro Watershed Flood Management

California State UniversityRobert Curry, Research DirectorWatershed InstituteEarth Systems Science & PolicyCSU Monterey BaySeaside, CALIF. 93955Bob_curry@csumb.eduWatershed Restoration Class – Spring, 2003Pájaro River WatershedFlood Protection PlanWm BodensteinerLani CloughSuzanne GilmorePaul HuntingtonJoy LarsonApril McMillanSteve MackC. Andrew MauckSerena PringEmily RothAmy ThistleMelanie VincentAPPENDICESA20

CHAPTER 2Design Flood AnalysisAnalysis of Flood Flow Frequency:San Benito River, Pájaro River, and TributariesPurposeThe purpose of this analysis is to determine the discharge of the 100-yearflow events for several gages on the San Benito River, Pájaro River, Uvas Creek, andPacheco Creek. The 100-year flow frequency events are compared both for all dataavailable and at 10-year sub-sets of the flow data. These estimates of discharge werecalculated using the Log-Pearson Type III methodology as described in Bulletin 17B:Guidelines for Determining Flood Flow Frequency by the US Water ResourcesCouncil (1982).MethodsData CollectionPeak annual flow discharge and stage heights at several gages in thePájaro River system watershed. Each gage number, name, river system, anddrainage area are summarized in Table 1. These data were obtained online from theUS Geological Survey.For the sites at Uvas Creek (11154200) and Pacheco Creek (11193000),years with zero flow in the peak record are adjusted to have a discharge of 0.01cfs.At the Hollister site (11158500), data for the water year 1957 is missing, and excludedfrom the analysis.Data AnalysisFor each gage, the peak annual flow data were ranked by peak discharge(Q), with the highest discharge of record with the rank of 1. The data does not have anormal distribution, requiring the log of the discharge to be taken for analysis. Thesample mean (Ŷ LT ), standard deviation (S yLT ), and standardized skew (g s ) are takenoff the log-transformed discharge (Q LT ).Due to the nature of flood events, and the small sample size of extremeevents, the accuracy of the sample skew is poor. An adjustment is made to thesample skew (g s ) for improved accuracy. The adjusted skew used in this analysis isadjusted by the following equation:g adj = g s * (1+(6/n))DRAFT 7/22/0321Pajaro Watershed Flood Management

where:g adj is the adjusted skewg s is the standardized skewn is the sample sizeFor several exceedence probabilities (p) ranging from 0.99 (1.01-yearrecurrence interval) to 0.01 (100-year recurrence interval), the values of thestandardized variate K were obtained using tables included in the Bulletin 17B reportfor the adjusted skew value. The Log-Pearson Type III estimates are determined fromthe following equation:K(g adj )Y LT = Ŷ LT + KS LTwhere:Y LT is the log of the estimated discharge for the exceedence probability atY LT is the mean of the log-transformed sampleK is the Log-Pearson Type III variate determined using the adjusted skewS LT is the standard deviation of the log-transformed sampleThe antilog of the Y LT values determined is the estimate of discharge atthe specific exceedence probabilities or recurrence intervals. In addition, 90% upperconfidence intervals were set for all stations at each exceedence probability.Smaller sub-sets of data from each station were analyzed for flood flowfrequency at intervals of 10 years. The sub-set analysis of the 100-year flood wasdetermined using the same methodology as the Log-Pearson Type III describedabove, including skew adjustment. No confidence intervals were estimated in thisanalysis.ResultsThe results of the flood frequency analysis for the select gaging stationsin the San Benito River, Pájaro River and tributaries are summarized in Table 2, andconfidence intervals are graphed as shown in Figure 1.Station NameStationNumberYears ofrecordCalculated100-yearflood Q cfsCalculated 50-year flood Q cfsCalculated25-yearflood Q cfsPájaro at Chittenden 11159000 62 30172.03 26759.25 22654.49San Benito at 156 11158600 31 7157.91 7052.91 6813.80San Benito near Hollister 11158500 33* 30234.73 21948.87 15034.67DRAFT 7/22/0322Pajaro Watershed Flood Management

Uvas Creek near Gilroy 11154200 35** 7009.29 6994.648 6942.47Pacheco Creek near Dunneville 11153000 43** 9187.33 9164.01 9063.46Table 2: Summary of flood flow frequency estimates* Water year 1957 missing data**Adjusted for zero flow years50000450004000035000300002500020000150001000050000Chittenden upper confidence intervalupper conf.Q estimate0.99 0.9 0.7 0.5 0.2 0.1 0.04 0.02 0.01exceedence probability160001400012000100008000600040002000San Benito at 156 upper confidence intervalupper conf.Q estimate00.99 0.9 0.7 0.5 0.2 0.1 0.04 0.02 0.01exceedence probabilityDRAFT 7/22/0323Pajaro Watershed Flood Management

70000Hollister upper confidence intervalupper conf.Q estimate60000500004000030000200001000000.99 0.9 0.7 0.5 0.2 0.1 0.04 0.02 0.01exceedence probability80007000600050004000300020001000Uvas Creek upper confidence intervalupper conf.Q estimate00.99 0.9 0.7 0.5 0.2 0.1 0.04 0.02 0.01exceedence probability20000180001600014000120001000080006000400020000Pacheco Creek upper confidence intervalupper conf.Q estimate0.99 0.9 0.7 0.5 0.2 0.1 0.04 0.02 0.01exceedence probabilityFigure 1: Log Pearson Type III results with upper confidence intervals for all gages.DRAFT 7/22/0324Pajaro Watershed Flood Management

The gages at Chittenden (11159000) and at Hollister (11158500) havesimilar 100-year peak floods, and have the highest maximum discharge of the gagesanalyzed. The gage at Chittenden (11159000) has the largest drainage area, butdoes not have the largest 100-year maximum discharge estimate, in part because itsdata set is longer and thus the confidence is better (lower interval).The results for the 100-year flow decadal analysis for each gaging stationare listed in Figure 2.140000Decadal analysis100 year flood discharge cfs120000100000800006000040000ChittendenGage at 156HollisterPacheco C.Uvas C.2000001940 1950 1960 1970 1980 1990Figure 2: 100 year flood estimates by decadeAll gages show a small increase in the estimate of the 100-year flooddischarge from the time period of 1960 to 1990. A decrease is also seen in all gagesfrom the 1950 estimate to the 1960 estimate.DiscussionThe above analysis was conducted using the standard reference as required of theCorps of Engineers. It requires a series of adjustments for extreme value rare eventsto account for their statistical rarity and for the non-symmetrical distribution ofprecipitation and runoff. An oversimplified way of looking at such data sets is that it iseither raining or it is not raining. If it is not raining, the amount of rain is 0.0 andcannot get any less. But if it is raining it can almost always rain harder and get wetter.“Dry” is a fixed value but “wet” is not. The adjustments are made using a table to fitthe data to a certain log-transform that Mr. Pearson called Type III and that fits a greatmany precipitation-related data sets.The Corps’ has chosen to design for a 40,100 cfs peak at Murphy’s Crossing, belowChittenden. That chosen value is subject to many caveats. The actual calculatedDRAFT 7/22/0325Pajaro Watershed Flood Management

value for the maximum possible 100-year flood at Chittenden is closer to 43,500 cfsusing the Corps’ methodology. But there is flood storage at Murphy’s Crossing, in theAromas area (Aromitas Lake of Jenkins) and in the Soda Lake area just below theChittenden gauge. USGS actually gauges the Pájaro during high flows at the bridgeat Aromas, not at Chittenden several miles upstream. Earlier chosen design floodswere higher, but the current value is not unreasonable. Because the lower river isformed by spillover from the upper watershed, drainage area does not increase in alinear fashion downstream. This is a unique watershed. As we shall show, thechannel capacity at Soap Lake and in the San Benito River increases in a very nonlinearfashion for flows above about 22,000 cfs as gauged at Chittenden. Thus the logplot of flows versus return period above that discharge tends to “flatten” (see the X’sor actual values in Fig 6 above versus the calculated Log-Pearson III curve). That is,high flood flows tend to be smaller than would be predicted based on the full period ofrecord because of the shape of the channels in the upper watershed and their faultdammedcharacteristics. The Corps’ design value is thus conservative in that it isabove reasonably probable values.The flow record was disaggregated into separate decades and each was assessedindividually to look for trends. In practice, one should not use a single gauging stationto predict a flood magnitude beyond two-times the length of the actual record. That is,to estimate a 100-year flood, one needs at least 33 years of peak flow record. Thus,the predictions based on 10-year periods do not reflect actual 100-year flowpredictions, but do give potential clues regarding changes in flood frequency throughtime. From this analysis we see that the 1955 Christmas storm at Chittenden in anotherwise non-remarkable decade would have forced prediction of a much larger 100-year event, but that the more frequent large events in later decades change thatpredicted value. The Christmas, 1995, flow at Chittenden was estimated at 24,000 cfsand was only exceeded there by the February 1998 event at 25,100 cfs. The March1995 event was estimated at 21,500 cfs. Thus, the 40,100 cfs figure being used bythe Corps’ for a design value is 160% of the maximum historic peak in 62 years ofinstrumental record. The February 1938 storms caused the levees to break in thelower Pájaro (Monterey Herald, 2-12-38) and flooded the Watsonville area with areported 3 feet of water. A newspaper photo of the lower Pájaro Valley below thetown of Pájaro (Fig 7) at about the location of Highway 1 today looks very much likethe 1995 conditions.DRAFT 7/22/0326Pajaro Watershed Flood Management

Figure 7 -- 1938 Lower Pájaro Valley flood. Photo point appears to be near presentHighway One.The lack of gauge record for 100 years at Chittenden does not limit our analyses backonly to its start in 1940. Study of precipitation records for the periods for which wehave gauge record permit comparison with those same records for the 40 or moreyears before stream gauge record. Where we have 100-years of record for dailyrainfall, such as at Hollister, we see that the 1955 event was by far the largestcumulative net storm rainfall. Although Hollister recorded 1.0 to 1.38 inches in singledays in 1935, 1936, and 1937, and although Watsonville and Hollister recorded morethan 1 inch per day for three consecutive days in 1937, it was the Christmas storm of1955 that set the standard for the Pájaro watershed. Beginning December 20 th , SantaCruz mountain summit areas recorded more than 10-inches a day through the 23 rd .Hollister recorded 1.93 inches on the 22 nd , 3.75 on the 23 rd , and 1.01 on the 24 th . Inthe southern Santa Clara County area the February 2-4, 1945 storm, with over 10inches in a day at Morgan Hill, provided the maximum historical rainfall period, andthat overlaps with Chittenden discharge record where flow was significantly less thanin 1955, 1998, and 1995. It is thus reasonable to postulate that the Chittenden gagehas recorded the largest Pájaro River events of the past 100 years, and that sustainedhigh flows must have been greatest in 1955, followed by 1998.There is the anomaly of the 1998 flow record at on the San Benito River that meritsfurther discussion. The official USGS gauge record indicates that the Pájaro tributarypeak flow was greater than the downstream flow at Chittenden in 1998. According tothe U.S. Geological Survey Field Office Supervisor, Larry Freeman, this may reflect areal difference where flood storage in the lower San Benito River below the Hollistergauging sites retains flow and diminishes the peak at Chittenden. However, thegauging station on the San Benito River was washed out in 1998 and the flow had tobe estimated based on water backed up at the Highway 156 bridge, rendering theestimate good only to ± 25 percent (see Appendix 1). Indirect evidence, presented inthe next chapter on flood storage, supports Freeman’s hypothesis that there is a largeflood storage volume still available in the Lower San Benito River.DRAFT 7/22/0327Pajaro Watershed Flood Management

Regional AnalysisThe trend in percent contribution to lower Pájaro flow volumes from the upper PájaroRiver versus the San Benito tributary deserves some analysis. RMC pointed out theapparent shift away from Santa Clara County contributions from the north andincreased San Benito County contributions from the south. The three major tributarieswere all dammed for water supply reservoirs about the same time in the 1960’s, andall are full during major flood events so there should be no net effect of reservoirs onrelative runoff from each of the three major Pájaro tributaries. The RMC analysis isvaluable and included here (RMC, Tech Memo 1-2-1 of October 8, 2001)“Basis of ComparisonThe Pájaro River watershed is large and the land uses are varied from denseurban to intensive agricultural to grazing lands to unused acreage. Changes in land useand management plans can affect watershed behavior. To be sure the hydrologic modelwill address the needs of decision makers and planners, three questions must beaddressed: what hydrologic parameters are necessary for comparison, where in thewatershed should these parameters be predicted, and at what exceedence frequenciesshould these parameters be predicted.Parameters to be usedThe most widespread parameter used for comparing changes to watersheds is“the annual instantaneous maximum peak discharge.” This is the discharge (rate offlow) in a stream channel and adjoining overbanks that is the greatest value at any timeduring a water year no matter how long the discharge lasts. A water year is the yearending September 30 and beginning the previous October 1. It is assigned the calendaryear corresponding to the September 30 date.The second most prevalent hydrologic parameter is the volume of flow in thestream.Generally the annual maximum 1-day average discharge value or 3-day averagedischarge is used in highlighting differences in runoff. For the Pájaro River watershedthe annual maximum 3-day average discharge is recommended because the watershedsare generally large and the 1-day average discharge is often reflective of theinstantaneous peak discharge.Two parameters are recommended – instantaneous peak discharge and 3-dayaverage discharge. Both parameters are to be annual maximum values.Parameters to be predictedDRAFT 7/22/0328Pajaro Watershed Flood Management

Shown in Table 1 are annual instantaneous maximum peak discharges fromtwo longterm stream gages – one on the San Benito River near the City of Hollisterand one on the Pájaro River at Chittenden just upstream of the end of the Corps ofEngineers Flood Control project.The San Benito River near Hollister gage had a drainage area of 586 squaremiles, whilethe current gage located at Highway 156 has a drainage area of 607 squaremiles. The drainage areas at the two gage locations are within 3.5 percent of oneanother and the combined record can be considered as one continuous record since1950. The drainage area at the San Benito stream gage is approximately half of that atthe Pájaro River at Chittenden gage. Data has been collected on the Pájaro Rivercontinuously since 1940. The four largest instantaneous peak events shown on thefollowing table are in the 1956, 1958, 1995 and 1998 water years.The ratios for the peak discharges at the Chittenden gage divided by the peakdischarges at the San Benito River gage for the four major flood years are:Water YearRatio1956 3.2171958 2.0261995 1.2871998 0.728Because the ratio of the drainage areas at the gages is approximately 2.0, onemight expect that the peak discharges maintain about that same ratio. However, the1956 event, the Christmas 1955 flood, shows much more of the peak dischargeattributable to the Soap Lake portion of the Chittenden gage’s drainage area. The April1958 flood was fairly evenly distributed. The two most recent floods, the March 1995flood and the February 1998 flood, had much more of their peak discharge comingfrom the San Benito River portion of the overall watershed at the Chittenden gage site.The following table shows the average daily discharges on the two rivers forthe four largest flood recorded at the Chittenden gage. The ratios of the sum of theaverage flows for the maximum three consecutive days are shown below:Date Chittenden San Benito Ratio12/1955 45,300 cfs-days 10,040 cfs-days 4.5124/1958 44,480 cfs-days 12,580 cfs-days 3.5363/1995 41,120 cfs-days 19,170 cfs-days 2.1452/1998 45,800 cfs-days 25,790 cfs-days 1.776DRAFT 7/22/0329Pajaro Watershed Flood Management

Interestingly, the maximum consecutive 3-day flow volume was approximatelythe same for all four major floods on the Pájaro River. The amount of volumecontributed by the San Benito River watershed, however, has grown from around aquarter in the 1950’s floods to around a half in the 1990’s floods. This means that therest of the 1,186 square mile watershed at the Chittenden gage contributed less volumein the 1990’s floods than it did in the 1950’s floods.”Based on the RMC analysis, above, it would appear that something is changing in thePájaro Watershed system. To investigate further, we looked into storm tracks for the1995 and 1998 events based on precipitation and runoff at stations to the west andsouth of the center of the Pájaro Watershed. Appendix 2 includes a map of thestations used and plots of precipitation and runoff.Storm Patterns: Appendix 2 (Storm Analysis) compares the 1995 and 1998 PájaroWatershed events based on rainfall and runoff stations on both the west and southaxes of the Pájaro Watershed (see map in that Appendix). The two flood periodswere associated with fundamentally different storm patterns. The 1995 event wasshorter and much less intense at Corralitos and Hollister than in 1998, but the 1995storm near in the middle San Benito River watershed at Pinnacles National Monumentwas more intense than in 1998. More fundamentally, all stations indicate that the1995 peak discharge was nearly synchronous with the rainfall peak; while in 1998 thefirst rainfall peak did not result in a synchronous flood peak, and the 1998 rainfall hada longer duration and second period of intensity compared to 1995. What this seemsto mean is that the 1995 storm stalled right over the centroid of the watershed nearHollister and produced an intense 48-hour flood, while the 1998 floods were the resultof more widespread rainfall for a longer time resulting in flood peaks that werepossibly near simultaneous, derived from both the upper Pájaro and San Benitosubbasins. The 1945 and 1955 events were more like 1998 based on theirwidespread rainfall patterns. Standard probability analysis does not, unfortunately,differentiate among differing causal mechanisms for standard winter rainfall floods.The fact that the 1995 flooding in the Lower Pájaro Valley had a much steeper risinghydrograph limb may partly explain why piping (flow under the levees with erosion)appears to have contributed to the levee failures in 1995 but not in 1998 even thoughthe flood stages below Murphy’s Crossing were similar. These differences arereflected in the hydrographs at Chittenden:DRAFT 7/22/0330Pajaro Watershed Flood Management

DRAFT 7/22/0331Pajaro Watershed Flood Management

California State UniversityRobert Curry, Research DirectorWatershed InstituteEarth Systems Science & PolicyCSU Monterey BaySeaside, CALIF. 93955Bob_curry@csumb.eduWatershed Restoration Class – Spring, 2003Pájaro River WatershedFlood Protection PlanWm BodensteinerLani CloughSuzanne GilmorePaul HuntingtonJoy LarsonApril McMillanSteve MackC. Andrew MauckSerena PringEmily RothAmy ThistleMelanie VincentDraft of July 22, 2003 Public Copy A31

CHAPTER 3Upper Basin In-channel Flood Storage and Restoration OpportunitiesBasic Conclusions:A very substantial volume of flood storage exists in the upper watershed. Focus todate has been on the Soap Lake subbasin of the upper Pájaro and lower Llagas andUvas tributaries. This area is part of the Lake San Benito basin and is very flat withpoorly integrated drainage. Most of the basin is underlain by hydric soils and is inagriculture. The RMC reports have tentatively outlined 30,000 ac-ft of flood storageover 7900 acres at an average depth of over 3 feet. That is the area that is subject toflooding in the 100-year flood, and approximately corresponds to a portion of theFEMA flood delineation map (see Map C for a portion of that map). Our team hasidentified a larger upper Pájaro River area subject to inundation to an average depthof 1.5 feet that gives about the same de-facto storage volume (see Map C). Our teamhas identified about 3000 acres of the RMC 7900 acres that could be excavated toenhance flood storage for an additional 7700 ac-ft of storage. The excavated materialcould be used for nearby protective berms and fill to allow some non-agricultural landuses in the areas subject to very shallow infrequent inundation of 1 foot or less. Thenet result is about 7000 ac-ft of added storage above the passive 30,000 ac-ft thatalready exists.On the San Benito River and its tributary Tres Pinos Creek, about the same 30,000ac-ft of de-facto passive flood storage exists today, but it is located along in-channeland channel margin areas along the braided channel itself. This total 60,000 ac-ft ofstorage capacity modifies the runoff characteristics of the Pájaro River today and allflood control designs assume that such storage is functional and in place. Diking ofsewage lagoons and active in-channel mining operations subtract from that storageand increase downstream peak flows. Today’s San Benito River is diked andmodified so that storm flow volumes and peaks derived from that tributary should beincreasing. We find that those channel changes have occurred progressively after thelate 1940’s and 1950’s. We estimate that opportunities exist to enhance flood storageon the lower San Benito River below Tres Pinos for an added 14,700 ac-ft withoutencroaching on areas outside of the current (1996) FEMA-defined 100-year activeflood zone. An example of an area suitable for restoration of natural overbank floodstorage is shown in Fig. 10.Thus the total potential augmentation of flood storage above Chittenden is on theorder of 22,400 ac-ft. This added volume of in-channel and near-channel storage hasa direct effect on flood peaks in the lower Pájaro River Valley, lowering the floodpeaks by about 10,000 cfs and the stage below Murphy’s Crossing by about 4 feet forthe 100-year design event (see PWA Lower Pájaro report, 2003). We believe thatthere are incentives for costs of upstream flood storage augmentation to be borne bylocal landowners. We believe that a river parkway plan can be combined with suchaugmentation to protect, stabilize, and enhance biotic and cultural resourcesupstream in a win-win situation so that costs of downstream flood protection arereduced while biologic and water quality values are increased throughout thewatershed system.DRAFT 7/22/0332Pajaro Watershed Flood Management

We also find that on the order of 6000 ac-ft of flood storage on the San Benito Riversouth of Hollister was lost before 1955, and that this cannot be recovered todaybecause housing and other structures are now located on that portion of thefloodplain.MAP C: Storage areas considered in this report. Inset: 1996 FEMA 100 year flood areaTheory:A balance between a river and its floodplain is necessary for the system to functionwithout continual artificial (human) input or damage from floods and/or bank erosion.Natural watershed systems are drained by waterways that store sediment and waterboth in the channel and on its floodplain. The floodplain is constructed by the riveritself as a self-regulating feature for storage of floodwater that exceeds the volumesthat can be carried in the natural stable channel. Sediment that cannot be carried bythe system during short flood periods is stored as bars and other deposits in thechannel and on the floodplain (Curry, 1981) awaiting the next flood flows.DRAFT 7/22/0333Pajaro Watershed Flood Management

When the balance between the volumes of water and sediment that are supplied to ariver from its tributaries and from its bed and banks is changed, the river systemattempts to rebalance itself. Rivers cannot store energy. They must use it as theygain it, dropping 300 feet in elevation, for example, from Hollister to the ocean. Ifwater flow volumes exceed sediment volumes, the river will attempt to erode sedimentfrom its bed and banks to rebalance itself and equilibrate its rate of work with thepotential and kinetic energy available to it. The river that drains a watershed isadjusted to carry the range of floods and sediment inputs that occur naturally in thatwatershed. If a period of major landslide activity occurs, for example along the faultzones in the Upper San Benito River, that sediment is stored in the channel awaitingsequential years of flood flows to move it downstream. This leads to natural channelaggradation, or build-up of sediment in the bed. After this occurs, flood flowsredistribute that sediment year after year, parceling it out for transport through lowgradientreaches downstream. The steep gradients on the depositional areas of theSan Benito River near Hollister (18-20 ft per mile) are the result of the great naturalinstability of the watershed hillslopes upstream.When a river is deprived of sediment or when flood flows exceed the volumesnecessary to carry the sediment entrained in that flood flow, the river erodes its bedand/or banks. Gravel and sand mining in the natural riverbed act to “starve” the riverof sediment, and lead to channel incision (downcutting) and/or bank cutting. Whendowncutting is severe, the river can no longer store floodwater in its floodplainbecause it cannot access its floodplain as it would naturally do every 2 to 3 years (seeFig, 11). If riverbed mining exceeds the long-term natural sediment supply, thewatershed system is said to be in disequilibrium. That is, the natural form of thewatercourse and its watershed are no longer balanced with the water and sedimentthat are moving through it. One of the most extreme examples of this imbalance is onthe Lower Russian River in California where the sediment-starved middle reacharound Healdsburg has incised as much as 20 feet and is now completely separatedfrom its floodplain. As a consequence of this loss of flood storage, floodingdownstream has increased in frequency and severity to the point that the area aroundthe town of Guerneville has become the Nation’s focus for the federal flood insurancedebacle where people repeatedly claim flood losses that cumulatively far exceed thevalues of the properties (James Witt, personal communication, 1997). At the MonteRio gauge on the lower Russian River, 5 or more “100-year floods” have occurredsince 1986.DRAFT 7/22/0334Pajaro Watershed Flood Management

Figure11 Cartoon showing how incision reduces flood storage and riparian habitatMethods:The Pájaro Watershed above the Chittenden gauge was inspected carefully todetermine where natural floodplain areas were no longer being inundated in majorfloods. The 1998 flood was of a magnitude such that it should have covered most ofthe natural floodplains that function in balance with the Pájaro River and its tributaries.If, as the probability plot for Chittenden suggests, the 1998 event was a 30-100 yearmagnitude flood at various places throughout the watershed, then the floodplainshould have carried water with sufficient depth and velocity to leave a record in thesurface soils. Surface soil characteristics on a floodplain reflect the depositional, andoccasionally erosional, passages of flood waters where they are vegetated and causeslowing of flood flows. Along the San Benito River, low-lying riverside bench landsbelow the level of the agricultural Lake San Benito land surface have very youngpoorly developed soils that are characteristic of flood deposits formed in the last fewhundred years (Fig 8). These deposits are very different from the moderatelydeveloped soils on the higher Lake San Benito and Lake San Juan surfaces (Fig 9).The current active stream channels do not have any silt-size organic-rich soildevelopment at all. Thus, it is possible to differentiate unambiguously wherefloodplains have been abandoned in the last century.Surveying of channel cross-sections and elevations was done at several places tocompare with historic data. A good plane table topographic map was made in 1917-1919 in the Hollister area (Hollister, USGS, 30-minute quadrangle) andDRAFT 7/22/0335Pajaro Watershed Flood Management

photogrammetric maps were made based on aerial photography taken between 1952(San Juan Quad) and 1955 (Hollister Quad.) Tres Pinos Quad photos were made in1953. The date of the published map is not material to the reference elevations, norare the dates of photorevisions. The topography is based on the original aerial dataexcept where specifically noted in purple overprint. For all of the upper Pájaro USGSquadrangles, published revisions in the 1970’s, 1980’s and 1990’s all specifically noteno topographic revisions after the original 7.5 minute quadrangle aerial base surveys.Field surveys in 2003 augmented a very detailed photogrammetric and 2-foot contourintervalground-based survey made privately for the San Benito River area byGraniterock Company. Those December 2000 data with very detailed aerialphotography at a scale of 1-inch = 500 feet were provided to us digitally byGraniterock. The earliest topographic surveys of 1917-1919, as published in the 1921USGS topographic map, were made on site by plane table and alidade. Although thecontour interval on those maps was only 50 feet, the surveyors clearly and definitivelynoted the heights of the stream banks with a “step” in the contour at the break inslope. By measuring the stream gradient on the map and the length of the step atmap scale, the heights of the banks can be estimated. For the upper San BenitoRiver above Hollister, those banks were 7 to 8 feet high at the time of the earlysurveys.San Benito County staff cooperated to provide access to their mining operation files,and to the plat maps and property records so that we could tabulate and attempt tocontact all property owners bordering the San Benito River below Tres Pinos. Theserecords are tabulated in Appendix 3. Those property owners were contacted wherepossible and state, federal, and local agencies were polled to try to learn of theirconcerns and interests in upper Pájaro watershed watercourses (Appendix 4). Fieldinvestigations were conducted on the Llagas, Uvas, and San Benito tributaries as wellas portions of the main stems of the Pájaro. We investigated evidences of activechannel modifications, gauging station status, riverbed and bank conditions,evidences of bed-form and plan-form erosion or change, and high-flow markers orfield evidence. These observational data are integrated into our findings andopinions.Findings:Channel Incision:We verified earlier reports (Goldner Associates, 1997, ) that the San Benito River hadbeen incising. Our findings for the thalweg elevations along the uppermost SanBenito above Hollister are as shown in this table:DATEBlossomRdHospital Rd Union Road Nash Road1917-1919 350 ft elev 320-25 ft.elev 298-99 ft.elev ~275 ft elev1955 330 ft elev ~308 ft elev ~295 ft elev 269 ft. elev2000-2002 ? 314.4 ft. elev 280.7 ft. elev 259.1 ft. elevDRAFT 7/22/0336Pajaro Watershed Flood Management

Although Hospital Road shows some aggradation after 1955, all of the other dataindicate progressive incision. The Hospital Road data may reflect the annual fillingthat takes place there for a summer road across the streambed. In this Hollistersection of the San Benito River, the natural floodplains had been abandoned by 1955and development was taking place on them. In 1995 and, especially, in 1998, theflood flows that were confined to an incised channel, cut laterally and made thechannel as much as 3 times as wide as before those floods. This is the natural waythat a watershed system works to regain equilibrium. Lacking overbank low-velocitywater storage, the deep high velocity flow undermines and cuts the easily erodedsand and gravel banks. This provides the sediment load that the high velocityconfined river is capable of moving, and it begins the process of cutting a new floodplane at the lower level of the streambed. This lateral erosion will continue until thewidth of the new deeper channel is sufficient to expend the available energy of theflowing water against the stream bed itself with little energy left for bank cutting. In thecase of the San Benito between Hospital Road and Hollister, this will be about a 0.75-mile width if no reclamation is undertaken. As this occurs, the constructed featuresand bridges will be damaged or lost, as is seen in the case of this newly-built CienegaRoad house during the 1998 floods (Fig 12):Figure 12 - House along Cienega Road south Hollister, 1998The detailed Graniterock aerial photos permitted us to investigate the entirechannel below Hospital Road to the junction with the Pájaro. We wereunable to receive landowner permissions to survey most of that channel andneeded to investigate the majority of the channel where public bridge rightsof-waydo not exist, and thus where the channel is not constricted artificially.The Graniterock aerial photos and the accompanying 2-foot contour intervalmaps are only a year old and reflect today’s conditions. These permitted usto compare the present topography of the river with that in the 1950’s asmapped by the US Geological Survey, and with sequential aerialphotographs. We borrowed and digitized the Soil Conservation Servicehistorical aerial photo enlargements of August 1959, and copied the availablecollections from the University of California Map Library and elsewhere.These included the 1931 lower Pájaro Valley, 1939 entire river, 1952 andDRAFT 7/22/0337Pajaro Watershed Flood Management

1967 flight lines and the full digital 1998 federal digital orthophoto quadrangleseries.We learned that there were three classes of change in the San Benito Riverchannel that all affect downstream flood peak heights. There were twodifferent kinds of land use changes that affect runoff timing and volume to theupper Pájaro channel derived from San Benito and Santa Clara counties. Thechanges we document can be summarized in 5 classes as follows:1. Those where direct channel incision prevents or reduces overbankflood storage onto a floodplain along the river. Rather than model thedegree of incision necessary to affect flood storage on floodplains, wesimply noted abandoned floodplains recognized by soils andvegetation. This kind of change greatly accelerates passage offloodwaters downstream, except where the channel incision interceptsthe groundwater surface and vegetation thus chokes the channel toslow water velocity.2. Those where channel widening with or without a deeper centralchannel (thalweg) effectively increase the capacity of a channel andthus reduce the height of a flood and access of those waters to theirfloodplain. This kind of change accelerates flood runoff because thewater remains in the channel and flows at a higher velocity than wouldoverbank floodplain flow.3. Those associated with a change from a multi-thread or braidedchannel to a single more efficient channel, often accompanied byreduced in-channel vegetation. This kind of change accompaniesincision and is favored where a central channel is deliberately gradedor confined to protect banks from erosion or to prevent lateralmigration of the channel, as for example where sewage lagoons orhighways are being protected. This kind of channelization changegreatly accelerates flow and reduces flood storage.4. Those associated with a straightening and cleaning of seasonal orflood-period temporary drainage channels on the floodplain. This wasobserved today only in the Soap Lake area but these sameconstructed drainage channels also are seen in 1917 mapped on thenow-abandoned floodplain south of Hollister. This class of changesreduces the time that overbank floodwater remains out of the channel,thus having a modest impact on downstream flood height.5. Those associated with dams and flood control structures and bankprotection measures that harden banks, reduce bank and bedroughness, and reduce infiltration capacity and land surface runoffdetention during intense rainfall events. Public works projects such asbridges, spillways, and highway berms tend to reduce bank and bedfriction and thus accelerate runoff. The farther upstream or fartherfrom the channel that these works are found, the less the degree ofdirect impact on peak flood heights. No matter how intense the rainfallor how long its duration. Uvas, Chesbro, and Hernandez reservoirsDRAFT 7/22/0338Pajaro Watershed Flood Management

clearly attenuate (reduce) flood peaks for events when they are notfull and spilling. The RMC report concludes that:“The three large reservoirs in the watershed – Hernandez, Uvas andChesbro – have been very effective in reducing the peak discharges ofthe more frequent events and, in the case of Hernandez Reservoir, havebeen effective in reducing peak discharges across the frequencyspectrum.” (RMC Hydro Technical Memorandum, 2000).That report concluded that, in 1937 before the three major watersupply reservoirs were constructed, the 100-year discharge atChittenden would have been about 12 percent larger than today.We disagree. That modeled value is based on observed historicalattenuation of flood peaks below those reservoirs. We investigatedthe watersheds above two of those reservoirs and did not findevidence of hillslope overland flow in the oak woodlands thatrepresent the conditions that existed in the reservoir basins beforethey were constructed. We thus disagree that the 100-year peakintensity rainfall and runoff event would be detained or attenuated bya full and spilling reservoir system. The opposite should be the casebecause a full reservoir with super-elevation at the spillways will notabsorb or detain any more rainfall and thus peak discharges at theextreme event are increased unless these water supply reservoirs arefirst drawn down. Flotsam around the shorelines of Uvas andChesbro show that they both have filled to above the elevations of thespillway inverts.Dikes along both Llagas and Uvas creeks in Santa Clara County andsignificant channelization and straightening of the primary channelshad led to loss of fish passage and high velocity channel erosion insome places. Much of this is now being repaired and channelroughness elements are being put in place to try to rebalance thesetributaries. Our impressions were that the channels themselves arenow as rough or rougher than were their natural antecedents,particularly where filled with Arundo and other plants, and tortuouslythreaded through urban areas. Thus, acceleration of runoff is minimal(Fig 13 photos are examples of Llagas conditions)DRAFT 7/22/0339Pajaro Watershed Flood Management

Fig. 13 Two views of the Llagas Creek Channel showing roughnesselements. Left image is just above Soap Lake, right in central urbanareaChannel DiversionFor the Lower San Benito River, the Graniterock aerial photos and contourmaps permitted us to establish that a local mining strategy has been toisolate various portions of the channels and to protect mining areas andchannel banks with berms. Some of these berms are built to the same heightas the natural Lake San Benito lakebed land surface. That elevation assuresthat the berms are above any historic level of the river. The berms and dikesreduce access by floodwaters to the full channel width and the incisionreduces access to adjacent floodplains so that the river is greatly constrainedand downstream flooding is increased. Figs 18-21 (Appendix 5 – HistoricalChange in the San Benito River) show an example of this kind ofmanipulation in the lowermost reaches of the San Benito River just upstreamfrom San Juan Road. Fig 18 is from the December 2000 Graniterock surveyand shows Highway 101 at its junction with Highway 129, and San JuanHighway. The “A”s are placed on abandoned floodplain remnants. A road isseen going from the sand mining operation area upstream (right) along thecrest of a constructed berm that is the same height at the Lake San Benitoagricultural lands. This berm thus isolates the present river from its floodplainremnants, some of which are used for mining equipment storage and somefor agriculture as was the case in the earlier photographs (Fig 19, taken in1939). Topographic detail can be seen in Appendix 6. Modifications from1950’s through the 70’s are shown on the 7.5 minute USGS Quadranglesshown in Fig. 20Channel diversions are found throughout the Pájaro River basin north of TresPinos. Because the natural channels in both Santa Clara and San Benitocounties were braided or wide and changing from year to year, early propertyowners confined the channels widely. Llagas Creek is now confined byberms over much of its length. Uvas is confined by major levees throughGilroy. San Benito River is confined to protect the City of Hollister, to protectvarious sewage treatment facilities, and to protect agricultural uses.Agriculture and development do not exist on most of the natural floodplainexcept above the City of Hollister where most of the floodplain is developedand where gravel mining and public works has resulted in many training andconfining dikes. Below (downstream) of Hollister the natural floodplain isused for cattle grazing and for a single sod farm. The Pacific Sod Farm(Tom Galdos, personal communication, 2003) has cooperated to protect itsprimary growing area with a low berm that was overtopped in 1998.Overbank silts are needed for the operation of this farm, where each sod cropexcavates a portion of the soil resource, and we were told that production isbecoming marginal without further sediment accumulation.RESTORATION OF CHANNEL FUNCTIONSWe estimate that an average one-fourth mile width of the 6.5-mile long lowerSan Benito River below Highway 156 could be restored to provide anDRAFT 7/22/0340Pajaro Watershed Flood Management

average 8-foot depth of water that is not now being stored at flood stagesabove that of a 25-30 year recurrence interval. To reclaim this storagevolume mid-channel levees would have to be breached, incised channelswould have to be recontoured or confined by gabion baskets or otherstructures or plantings to slow peak flood flow volumes, and overwidechannel reaches would need gabion structures or plantings to constrict flowto a central meandering channel.Non-structural solutions, primarily involving willow plantings, have beeneffective in the Carmel River for this kind of restoration of a low-flow centralchannel that supports wildlife and protects riverbanks from erosion. The SanBenito River is more problematic than the Carmel. Unlike the Carmel,aggregate mining is a primary tax base for San Benito County. Further thechannel of the San Benito (but not Upper Pájaro) has a very low base flowand is dry much of many years, thus making vegetation management moredifficult. The history of mining and degree of channel incision that hasresulted on the San Benito create a more immediate need for active solutionsthat will set the stage for raised water tables, increased in-stream vegetation,and slow aggradation of the active riverbed.Suggested Restoration options for San Benito River:Two primary restoration strategies must be used on the San Benito River.The levees and dikes that exist within the channel must be breached atsufficient points to allow ready and rapid exchange of floodwaters throughoutthe channel. This will create a floodway, or zone of active flood storage. It isimportant that this storage be “on-channel”; that is, readily able to retainfloodwater as the stage rises in the river. All of the berms need not beremoved, but the more that can be removed, the greater the storage capacityof that active channel. For sites like the Hollister sewage lagoons, the leveescannot be breached, but for sites such as shown in Appendix 5, they must bebreached. For a site like the Pacific Sod farm, where an entire meander isprotected by a berm, some accommodation can be made to allow floodingonly at flood stages of 25-year return period or greater. This is about themagnitude where these protective berms overtop today.For the overwide channels and other sites where floodplains have beenabandoned directly along the San Benito River channel, we recommendconsideration of a series of gravel-filled gabion baskets that extend from thebanks toward an optimal central channel. These structures do not cross thechannel and do not impact the low-flow channel. The serve as a series ofconfining and “training” structures that focus the flow of the river in a singlethreadcentral channel, while simultaneously creating flow velocity reductionagainst the banks and sediment deposition zones. As the central channelbecomes defined after one or more channel-forming events (see Rosgenfigure on p 16), then a second and third set of baskets are built on top of thefirst until the grade of the channel at flood stage is high enough to reach thefloodplain and restore stable channel geometry. If properly placed, thegabion basket assemblages will encourage pool and riffle geometry in thecentral channel, and will allow vegetation to become established along thebase of the present riverbanks. That vegetation is the primary tool forreducing bank erosion and for slowing the flood velocities. In effect, eachDRAFT 7/22/0341Pajaro Watershed Flood Management

“lift” or set of gabion baskets becomes a control structure for a new floodplainin the overwide channel areas,If sediment supply were very large and aggregate mining were not occurring,it would be a simple matter to allow the channel to aggrade to sequentialgabion installations until the system was returned to a condition similar to thatprior to human alteration. But sediment supply is episodic and not unlimitedas is demonstrated by the ever-widening channels (see Riley, 2003).Further, the aggregate industry owns much of the channel and its banks andis the logical entity with the capability to stabilize and restore the riverchannels.Aggregate Mining Company Opportunities:We tabulated and plotted all riverside ownerships (see Appendix 3). TheGranite Rock Company of Watsonville, California, owns or controls the majorportion of the channel between the Pájaro confluence and Tres Pinos. Theylease surface portions of their parcels for farming on the Lake San Benitosoils, and extract aggregate resources from the channel bed, usually by“skimming” the active braided bed. Their active mining operations ceased inthis area 5 years ago. There are other aggregate operators on the river, andall are in theory regulated by both the State and the County (see Appendix 7).Regulation is not consistent or effective. The State, under the Surface Miningand Reclamation Act (SMARA) requires a Reclamation Plan and financialassurances (bonding) for each operator. This program is administered bySan Benito County. The State has no authority over land use permits, so theCounty is also responsible for either issuing a Conditional Use Permit ormaking specific findings that may allow “grandfathered” projects as vesteduses. Thus, San Benito County carries the primary responsibility foroversight of an industry that provides an important part of its tax base.According to the California Department of Fish and Game (personalcommunication, Santa Rosa office, 2003), some San Benito River operatorsmay not be in compliance with their Section 404 regulations for in-channelmodifications. According to some operators, San Benito County is attemptingto limit their operations, in part because of complaints by riversidelandowners about bank erosion (such complaints were heard from manyproperty owners that we contacted). This environment restrains miningoperations, with some operations currently shut down awaiting permits. Wesee an opportunity to use aggregate mining operators, with access to heavyequipment and aggregate resources, to help provide a solution.A San Benito River Parkway Plan needs to be developed to stabilize andrestore the lower San Benito River. At the present time, public respect for theriver is very low. Both access and amenities are rare. Many residents of thatcounty only see the river from highway bridges and have no idea what isactually in the channel. Where the channel has incised to the water table andmid-channel willow thickets exist, local residents and the County complain,with some validity, that flood flows are then forced into the banks withresulting erosion. Where roads access the channel or banks, refuse isdumped to be carried away by subsequent floods. Temporary summer rivercrossings, as at Nash Road and Hospital Road, are installed seasonally withDRAFT 7/22/0342Pajaro Watershed Flood Management

culverts and fill. Other parts of the channel are used for off-road vehiclerecreation resulting in destruction of the veneers of gravel cobble bed armorleading to erosion with only minimal flow velocities in subsequent winters.Exotic vegetation in the channel provides a seed source that spreads toadjacent agricultural fields.Graniterock has shown us its willingness to discuss and promote restorationoptions, including a River Parkway. They are on record with such a proposal,and conducted the channel survey for just such a purpose. For them, theincentive is continuing County cooperation and permitting through allregulatory agencies. They want to access the aggregate resources. For theriverside landowners and the County Public Works agency, the incentive isreduced erosion and maintenance costs. For the local residents, theincentive is a potential river parkway with 10 or more miles of high-valueriparian parkway and habitat, and some public access. For the downstreamcounties, the incentive is flood storage and reduced loss of lands and costsdownstream for flood control. This is a potential win-win situation.Practically, such restoration planning and implementation takes time. Someareas must be maintained for mining if the operators are to cooperate andprovide support for the restoration. Because of the high value of theagricultural production on the Lake San Benito silt soils, mining aggregate offchannelis not practiced locally. Because mining does not take place duringflood periods or when groundwater levels are high, operators need to mineand stockpile in the dry season. A well-designed restoration plan thatattempts to integrate aggregate resource mining is not a tautology. It can bedone. The Merced River parkway, the San Joaquin River Parkway, andseveral other California examples provide models. Enhanced flood storageaccrues slowly. It may take decades to achieve the full component ofpotential enhanced flood storage. You cannot simultaneously aggrade andmine the same parts of the channel. Mining must be focused on those siteswhere there are minimal streamside potential flood storage areas that can berestored. Gabion baskets would have to be installed in areas not beingmined as well as in areas being mined. As many as three tiers of basketsmay need to be placed initially just to bring high flood flows up to floodplaingrade, but mining can continue between those tiers of baskets. We areworking to restore what is called the energy grade line of the surface of theflood flows at 25-30-year magnitude events only. We can allow all otherlesser floods to pass down a central thalweg. Fig 14 is a cartoon thatillustrates this open central channel. Unfortunately, the USDA StreamRestoration Best Management Practices web site does not provide examplesof these 1-km wide scale restoration structures, but the principles that theyillustrate are often applicable(http://www.wcc.nrcs.usda.gov/watershed/UrbanBMPs/stream.html). What isimportant is the fact that the structures are low-tech, porous, inexpensive anddo not obstruct the central channel. Like the Stream Barb structure used insmaller channels (Fig 15), the gabion basket structures slow water at theedges of the channel and are easy to install and maintain.We recommend that Graniterock and other willing San Benito Riveraggregate mining operators be invited to develop plans for a restoration/riverDRAFT 7/22/0343Pajaro Watershed Flood Management

parkway system that can be implemented with no or minimal outside funding.Graniterock has already demonstrated a willingness to propose such actionand assisted our study through their generous sharing of their aerial surveydata. If conservation easements or land trust arrangements can beimplemented for parts of the upper Pájaro watershed in conjunction withthese major landholders, this may facilitate faster completion of potentialstorage volumes. We can help to facilitate such planning andimplementation.Suggested Enhancement Options for Upper Pájaro River:The Upper Pájaro River, along the Santa Clara-San Benito County line isfundamentally different than the San Benito River. Here a channel is incisedup to 25 feet below the Lake San Benito lakebed, but because the riverbedhas historically carried a reliable supply of influent groundwater, a densefinger of riparian forest characterizes most of the channel. This matureriparian forest of cottonwood, alder, maple, and willow has a dense woodyinstream fabric of logs and mid-channel growth, with a diverse pool structure.Although only 100 m wide in places, this riparian corridor provides highquality wildlife habitat and, apparently, allows anadromous fish passage intoLlagas and Uvas creeks.Because the riparian forest is so dense and woody debris so prevalent, floodstages rise rapidly and go overbank onto the old San Benito lakebed. Locallandowners report that flooding reaches the old lakebed level at a frequencyof 10 years or less. Because of the high regional groundwater levels thatseasonally saturate up to the lakebed silt cap, the soils of the area areclassed as hydric and, unless cropped continually, revert to wetlandconditions with emergent wetland plants. Farmers have constructed drainagechannels across these lands to carry shallow groundwater and rainfall intothe Pájaro.We were able to meet with local landowners and/or farm leaseholders. Welearned that this Soap Lake area, just south of Gilroy, and situated alongHighway 25 between Gilroy and Hollister, may be a target for extensivedevelopment. An ongoing effort sponsored through the Pájaro RiverWatershed Flood Prevention Authority and AMBAG seeks to establish floodor conservation easements for the Soap Lake basin. Our sources suggestthat the opportunity costs for development are so great that contiguouseasements may be very difficult to obtain. While the site looks like marginalagricultural land used for little more than growing hay or grazing with smallareas of row crops, it is in fact being leased back to local farmers and kept inagriculture as an interim holding pattern while development options areconsidered. If some of these lands would be wetlands were it not forcontinual agricultural use, then federal regulations will make it necessary tomaintain agricultural uses or raise the lands or protect them with dikes andlevees to permit non-agricultural uses. Should this be the case, a need forlocal fill may provide an opportunity to encourage landowners to excavate the3-foot deep lake-silt cap immediately adjacent to the river. This couldincrease the flood storage.DRAFT 7/22/0344Pajaro Watershed Flood Management

Our modeling of enhanced flood storage considers opportunities for locallandowners to enhance wetland status in areas now in agriculture throughexcavation of 2 to 3 feet of native surface lake silt (see Map C). There is noassurance that the expensive regarding efforts would be cost-effective forowners wishing to develop parts of the margins of Soap Lake for housing orother non-agricultural uses. Preliminary discussions with representatives oflandowners have not discouraged us from considering three-foot excavationin about 2.5 square miles of lower Soap Lake along the Pájaro River for onchannelflood storage augmentation, yielding 4800 ac-ft of new storage inaddition to the existing Soap Lake flood volume. We have also modeled anadded 2.25 square mile area extending westward to the existing railroad bedberm, adding an additional 2880 ac-ft of new storage, and raising the landelevation between Highway 101 and the rail line above flood levels. This kindof tradeoff must be approved by all regulatory agencies. In essence, amarginal non-functional wetland area now in agriculture is converted tofunctional restored planted natural wetland in exchange for allowing fill of theedges of the Soap Lake basin that are only wet during sustained 100-yearflood events at the present time. Of the 30,000 ac-ft Soap Lake basinstorage volume, some 700 ac-ft of natural storage would be traded for about7000 ac-ft of enhanced functional wetland habitat storage. This is also a winwinsituation if a developer or regional agency can be found to champion thatlarge-scale set-aside, and if the regulatory agencies favor it.CONCLUSIONSOver 60,000 ac-ft of flood storage exists on or very near the channels of theupper Pájaro watershed. Soap Lake comprises an important part of this, butonly a part of the storage that can be modified or lost with upstreamdevelopment. Approximately 22,400 ac-ft of storage enhancement is readilypossible. Most of this storage is no longer active and no longer accessible tothe river because of stream channel incision, levees and berms, anddiversions. Restoration of this volume can reduce downstream peak floodheights by on the order of 4 feet during a 100-year flood. The cost of thisflood reduction is believed to be less than the cost of protective worksdownstream that achieve the same level of protection.DRAFT 7/22/0345Pajaro Watershed Flood Management