Stella 04 TWS Merced Veg

8/6/2019 Stella 04 TWS Merced Veg

1/13

302

INTRODUCTION

One of the attributes of a natural alluvial riversystem is a self-sustaining, diverse ripariancorridor (Trush et al. 2000). Riparian zones,defined by Gregory et al. (1991) as "three dimensionalzones of direct interaction betweenterrestrial and aquatic ecosystems," providemultiple benefits to instream and terrestrialecosystems and are widely recognized as centersof biodiversity and corridors for dispersal ofplants and animals in the landscape (Gregory etal. 1991; Johansson et al. 1996). Riparian forestsprovide leaf litter to instream food webs, largewoody debris and shading for fish habitat, andforage, cover, nest sites and migratory corridorsfor wildlife. Physical effects of riparianprocesses include regulating instream

temperatures, filtering nutrients and agriculturalchemicals from runoff, stabilizing banks,facilitating sediment accretion on floodplains,and providing cooler, more humid, and lesswindy microclimates (Gregory et al. 1991;Mitsch and Gosselink 1993; Malanson 1993;Bayley 1995; Naiman and Descamps 1997).

In natural river systems, physical forces such asflooding, erosion, and sediment depositionstrongly influence the character and developmentof riparian vegetation. As rivers meander andfloodplains form, old vegetation stands aredestroyed and new ones are created (Figure 1).

These cycles of destruction and creation sustaina mosaic of vegetation stands along river banksthat are diverse in species composition, age, andcanopy structure (Fonda 1974; Strahan 1984).

On the Merced River, as with other rivers inCalifornias Central Valley, water resourcedevelopment and human land uses have greatlyaltered alluvial river processes and riparianvegetation conditions (Vick 1995; StillwaterSciences 2001, 2002). Vegetation clearing foragricultural development and mining for goldand gravel have greatly reduced the extent offloodplain riparian forests. Dams and associatedflow regulation and diversion for irrigation andflood control have altered seasonal flow patterns,and reduced flood peaks and sediment supply.Bank revetment, combined with flow regulation,

has reduced river migration. These physicalchanges have resulted in changes in riparianvegetation structure, species composition, andsuccessional processes. Similar effects of flowregulation on riparian ecosystems have beendescribed for other alluvial river systems(Decamps et al. 1988; Rood and Mahoney 1990;Johnson 1992; Auble et al. 1994; Ligon et al.1995; McBain and Trush 1997; McBain andTrush 2000).

Stillwater Sciences assessed riparian vegetationspatial patterns and successional processes in thelower 52 miles of the Merced River from fall

Riparian Vegetation Dynamics on the Merced River

JOHN STELLA1,2, JENNIFER VICK AND BRUCE ORR1

1Stillwater Sciences, 2532 Durant Ave, Berkeley, CA 94704

2

Department of Environmental Science, Policy and Management, University of California,Berkeley, CA 94720

ABSTRACT. As part of a larger restoration planning effort for the lower 52 miles of the MercedRiver corridor,riparian vegetation conditions, spatial patterns, and successional processes were assessed relative to conditionsthat existed prior to the construction of Exchequer Dam in 1926. We mapped riparian vegetation along theriver corridor and conducted pilot level studies at floodplain sites to: (1) document current plant speciescomposition, vegetation structure, and distribution relative to hydrogeomorphic factors; (2) assess vegetationshifts in response to the postdam hydrologic regime; and (3) evaluate cottonwood seedling recruitment andestablishment patterns. The site evaluations included analysis of time-series aerial photographs, field surveysof channel and floodplain topography, belt transects documenting vegetation composition and structure, surveysof riparian seedling recruitment and survival, and analysis of seedling survival and site-specific hydrologic datausing a conceptual model of riparian seedling establishment. Results indicate that: (1) riparian zone area hasdecreased since dam construction in 1926 and is currently confined by land uses and current flow conditions to19 percent of the pre-dam floodplain; (2) encroachment of riparian vegetation into the former active channelsince construction of Exchequer Dam has resulted in a confined and simplified channel and loss of thebare alluvial surfaces necessary for recruitment of native pioneer tree species; and (3) flow regulation has createdartificially stable conditions that induce riparian seedlings to recruit lower on banks than occurred historically,leading to high scour and inundation mortality from flows later in the year. These conditions have contributed tothe decline of cottonwood-dominated forest stands throughout the river corridor.


2/13

303

RIPARIAN VEGETATION DYNAMICS ON THE MERCED RIVER

STUDY AREA AND BACKGROUND

The Merced River, a tributary to the San JoaquinRiver, drains a 1,276-square-mile watershed onthe western slope of the Sierra Nevada Rangeand joins the San Joaquin River in the CentralValley approximately 90 miles south ofSacramento. The lower 52 miles of the MercedRiver, which comprises the restoration planningand study reach, flows through a sequence ofnested Quaternary alluvial fans (Harden 1987).At the upstream end of the study reach, the rivervalley broadens from the confined bedrockvalleys in the upper watershed and the riverbecomes a highly dynamic, multiple channel(anastomosing) system. Historically, these

channels occupied the entire valley floor (up to4.5 miles wide) in the Snelling (RM 48.0)vicinity. Downstream of the Dry Creekconfluence (RM 31.7), the valley width narrowsand the historical channel was a single-thread,meandering system.

The watershed experiences a Mediterraneanclimate, having wet winters and dry summers;mean annual precipitation ranges from 55 inchesin the Yosemite headwaters to 1014 inches onthe Central Valley floor (USDA Forest Service1997). Similar to other rivers originating from

the west side of the Sierra Nevada Range, naturalflow patterns in the Merced River are typified bylate spring and early summer snowmelt, fall andwinter rainstorm peaks, and low summer baseflows.

Flow in the Merced River is controlled byseveral mainstem dams owned by the MercedIrrigation District (Merced ID) (Figure 2). NewExchequer Dam, located at RM 62.5, controlsrunoff from 81 percent of the basin and createsLake McClure, the largest storage reservoir inthe system. This dam replaced the originalExchequer Dam, which was completed in 1926and which had a reservoir capacity of 281,200

1999 to summer 2000 as part of an evaluation ofenvironmental baseline conditions conducted for

the Merced River corridor restoration planningprocess (Stillwater Sciences 2001). Thevegetation studies used a combination ofcorridor-scale vegetation mapping and pilotstudies at four floodplain sites (Figure 2). Rivercorridorpatterns were assessed by mappingvegetation along the Merced River from MercedCountys eastern boundary at Hornitos Road(River Mile [RM] 55.5) to the San Joaquin Riverconfluence (RM 0) using a combination ofremote sensing and field observations. Theobjectives of the mapping were to document thelocation, extent, and general composition ofcurrent riparian vegetation in the corridor, assessthe degree of invasion by non-native woodyspecies, and prioritize reaches for preservationand restoration. The objectives of the sitespecificpilot studies were to: (1) documentcurrent vegetation species composition, canopystructure, and distribution relative tohydrogeomorphic factors; (2) assess changes invegetation extent and condition since flowregulation began in 1926; and (3) evaluaterecruitment patterns of pioneer riparian trees.Study sites were selected in areas that hadexperienced relatively little disturbance fromagriculture, urban development, and gold and

gravel mining. To the extent feasible, sites werechosen that exhibited active channel bars, mixedagestands of riparian vegetation, and actual orpotential for riparian vegetation recruitment.The site-specific analyses pilot studies werelimited in scope and designed to provide initialdocumentation of ecological patterns and trendsand generate hypotheses describing howhistorical and present-day ecological processesaffect the diversity and integrity of ripariancommunities along the lower Merced River.

FIGURE 1. Conceptual model of riparian succession.


3/13

304

JOHN STELLA, JENNIFER VICKI AND BRUCE ORR

acre-feet3. New Exchequer Dam was constructedin 1967 and has a maximum reservoir storagecapacity of 1,024,600 acre-feet. McSwain Dam,completed in 1966, is located at RM 56.0 and isoperated as a re-regulation reservoir for LakeMcClure and as a hydroelectric facility.

Downstream of McSwain Dam, two low damsdivert flow from the river into the Merced IDirrigation system: Merced Falls at RM 55.0 andCrocker-Huffman Dam at RM 52.0. Crocker-Huffman Dam represents the upstream limit ofthe restoration planning and study reach.

Post-dam flow regulation has changed the timingand magnitude of peak flows on the MercedRiver. Whereas historically the river experiencedsharp winter flow peaks from winter storms anda prolonged snowmelt pulse in spring, today theriver downstream of the dams experience muchlower winter peak flows, no spring pulse, and

elevated summer flow to provide water forriparian diverters. Mean annual peak flowbefore 1926 was 16,161 cfs for the period ofrecord (1902-1925). From 1926 through 1967,when New Exchequer Dam was constructed,mean annual peak flow was 7,059 cfs, and wasfurther reduced to 4,551 for the post-1967 period(Stillwater Sciences 2001). Reduction in peakflows has reduced the scale of geomorphic processes suchas erosion, deposition, and channel meandering.Elimination of the snowmelt pulse has disruptedpopulations of key organisms such as salmon and ripariantrees which have life history processes that rely onannual elevated flow conditions in spring.

In conjunction with flow regulation, bank revetment hasbeen used extensively throughout the Merced River to cor-rect and prevent bank erosion. Revetment, using rock or con

crete rubble, limits channel migration and hinders theestablishment of native riparian vegetation. Non-native,invasive vegetation species, such as giant reed(Arundo donax), are often associated with bank revetment.

Mining for gold and gravel have further altered

the Merced River corridor. From 1907 through1952, the lower Merced River channel andfloodplain were dredged for gold and coarsetailings were deposited on the floodplains to aheight of approximately 15-20 feet. The dredgertailings, which currently cover 7.6 square milesof floodplain in the Snelling vicinity, have replaced theonce diverse and productive floodplains and riparian forestswith barren piles of cobbles and boulders, which currentlyconfine the river channel and floodplain to a narrow corridor.

Large-scale aggregate mining began in the Merced Rivervalley in the 1940s and continues today. Two types ofmining, in-channel mining and floodplain mining, occurred

in and along the river. At in-channel mines, operatorsexcavated sediment directly from the river channel,leaving behind large in-channel pits. At floodplain mines,operators excavated pits in the floodplain adjacent to theriver channel. These pits were typically separated from thechannel by narrow, unengineered berms. Many of theseberms have since failed, resulting in capture of the riverchannel by the pits. Until recently, in-channel andcaptured pits occupied 7.3 miles (or 40 percent) of thegravel-bedded reach of the river (from approximately RM25.5 to 52.0).

METHODS

Vegetation Mapping

Vegetation mapping was conducted by StillwaterSciences and the Geographic Information Center(GIC) at California State University Chico using

FIGURE 2. Location of study sites on theMerced River.


4/13

305


a combination of aerial photograph interpretationand field verification. The vegetation classificationsystem used for the vegetation mapping corresponds toclassification systems used throughout the Sacramentoand San Joaquin basins (CSUC 1998, McBain and Trush2000, Jones and Stokes Associates 1998) and is based

on Holland-type classifications (Holland 1986).Primary criteria used in designating each vegetationtype included its areal abundance within the river corridor,the uniqueness of its color infrared signature, andassociation with hydrologic processes and geomorphiclandforms. To aid in management and restoration efforts,non-native exotic trees were mapped as separate cover typesif they were identifiable from the aerial photographs.

Vegetation assemblages were identified andmapped from color infrared aerial photographs(scale 1:24,000) taken in May 1999. Photographs werescanned, orthorectified, and enlarged electronically toapproximately 1:6,000, and boundaries of cover type

polygons were digitized on-screen. Natural color stereoaerial photographs taken in 1993 (scale 1:6,000, U.S.Bureau of Reclamation) were used to map geomorphicsurfaces (current active channel, current floodplain,former floodplain and terrace) and to aid in identificationof vegetation types that could not be clearly identifiedfrom the infrared photographs.

Field verification of the vegetation maps wasconducted between fall 1999 and spring 2000and in June 2000. Of the 3,008 total polygonsmapped, a sample of 693 polygons (23%) wasfield-verified, primarily those adjacent to the

channel in the lower portion of the river corridor(RM 2.032.5). A randomized polygonverification method was not possible due toproperty access and field logistical issues.

Intensive Site Analyses

Four sites, named Cuneo, Snelling, ShafferBridge, and Stevinson, were chosen for analysisof existing vegetation dynamics, hydrogeomorphic

processes, and vegetation changes since construction ofExchequer Dam in 1926 (Figure 2). The most detailedanalyses were conducted at the Snelling (RM 48.2) andShaffer Bridge (RM 32.5) sites and results arepresented below. All analyses conducted are listed inTable 1 and described fully in the baseline

conditions report (Stillwater Sciences 2001).

Coarse-scale changes in the extent and conditionof riparian vegetation since construction ofExchequer Dam were assessed by analyzingaerial photograph series taken in 1937, 1950,1965, 1979, 1998 and 1999. Because the earliestseries was taken eleven years after constructionof Exchequer Dam in 1926, pre-dam conditionswere inferred from vegetation density, tree size,and recent geomorphic activity apparent in the1937 photographs. In addition, flood maps(USACE, unpublished data) outlining areasinundated by the January 1997 flood (8,279 cfs

peak flow as measured at the Merced IDCrocker-Huffman gauge) were reviewed. Threeperiods of differing hydrologic conditions wereconsidered in the analyses: (1) pre-ExchequerDam (before 1926); (2) the inter-dam period (19261967);and (3) post-New Exchequer Dam (1967present).

Vegetation species composition and structurewere documented at the Snelling and ShafferBridge sites using vegetation transects and relevsurveys in conjunction with topographic surveysand flow monitoring. At each site, 4-7 crosssections across the active channel and floodplain

were surveyed. Within the bankfull channel,points were surveyed at every slope break or atfive-foot intervals, whichever distance wasshorter. Outside the bankfull channel, pointswere surveyed at every slope break or at ten-footintervals, whichever distance was shorter. Permanentendpins were placed at both ends of each cross section,and the location of the cross section and end pins wasmapped onto aerial photographs.

TABLE 1. Summary of study site field surveys and analyses used to evaluate riparian vegetation functionalrelationships.

Field Surveys and Analyses Stevinson(RM 2.2)

ShafferBridge

(RM 32.5)

Snelling(RM 48.2)

Cuneo(RM 50.8)

1. aerial photograph analysis X X X X

2. channel and floodplain cross section surveys X X X

3. transect(s) of vegetation cover type distribution,canopy structure, and geomorphic position

X X X X

4. species lists for common vegetation cover types X

5. relevs (species composition and cover bycanopy strata) for common vegetation covertypes

X X

6. seedling recruitment surveys X

7. water stage monitoring X X

8. hydraulic and sediment transport modeling X X


5/13

306


Twenty-foot-wide belt transects were centeredon the topographic cross sections on both left-and right-bank floodplains. Individual stems ofwoody species were censused for species,diameter size class, and location along the surveytape. The ground surface elevation at the

location of each stem was estimated by crossindexingthe plants position on the survey tapewith the elevation surveyed from the topographiccross section. This method assumes thatelevation across the width of the belt transect isuniform, which was not always the case in thefield. In most cases, however, the elevationaldifferences between the center line and edges ofthe belt transect were minimal, and no systematictopographic bias was evident in the field.

Three hydrologic parameters were calculated forstems of each species encountered in the belttransects: mean elevation above summer

baseflow, inundation duration, and floodfrequency. Inputs into the calculations includedraw ground surface elevation data, flow durationand flood frequency curves (developed fromstreamflow records), and stage-discharge ratingcurves developed at each cross section usingHEC-RAS, a one-dimensional, step-backwaterhydraulic model. Inputs to the model includedthe surveyed cross sections and channel slope,hourly flow data from a nearby flow gauge,roughness coefficients for both the channel bed(0.045) and floodplain (0.07), and a boundarycondition rating curve. The boundary condition

rating curve was developed using hourly waterstage data from a continuous water level recorder(Global Water WL-14 WaterLogger) deployed atthe downstream end of the site and data from thesame nearby gauge.

River-wide recruitment and establishment ofFremont cottonwood (Populus fremontii) andwillows (Salix spp.) on alluvial bars and floodplainswere assessed by boat surveys conducted in fall 1999and spring 2000. Areas where seedlings (age < 1 year) hadrecruited in the same year and where saplings (age 1 to 5years) had established in prior years were recordedonto aerial photographs to provide a qualitative

description of spatial patterns of recruitment.

At the Snelling site, seedling surveys wereconducted on gravel bars to quantify seedlingrecruitment and overwinter survival. Six gravelbars were surveyed in October 1999 to documentrecruitment. Surveys were conducted using a10.8-ft2 (1.0-m2) plot frame laid contiguouslyalong a transect that extended from the channelmargin to the upland edge of the bar. Withineach plot, seedling or sapling species and age,substrate texture, and surface moisture conditionwere recorded; age was determined based onstem buds scars. The transects were monumented

with rebar for later resurveying. Follow-upsurveys were conducted in June 2000. Only onegravel bar had a sufficient number of seedlingsin both years to evaluate between-yeardifferences in recruitment patterns.

RESULTS AND DISCUSSION

As described for other river systems throughoutthe Central Valley (Strahan 1984, McBain andTrush 2000) natural floodplain processes thatpromote riparian forest sustainability have beenaltered or eliminated, and many mature riparianstands along the Merced River may be relicts ofpre-dam hydrologic conditions. The pre-damriparian zone supported dense, multistoriedstands of deciduous trees, including valley oak(Quercus lobata), Fremont cottonwood, willow,Oregon ash (Fraxinus latifolia), box elder (Acernegundo), and other species (as described

generally for Central Valley rivers in Thompson1961, 1980; Roberts et al. 1980; Conard et al.1980). These riparian forests varied greatly inwidth, from a narrow strip in confined reaches toseveral miles wide on broad alluvial floodplains(Thompson 1961). Local historical accounts ofthe Merced River describe a rich aquatic andterrestrial fauna supported by riparian forest andmarsh habitats (Edminster 1998). Katibah(1984) estimates that the Merced River and thelower San Joaquin River (from the Mercedconfluence to Stockton) supported over 90,000acres of riparian forest, approximately one tenthof the Central Valleys pre-settlement riparian

forest. Figure 3 illustrates a generalized crosssection of the Merced River prior to damconstruction (1926).

Reduction in Riparian Vegetation Extent

Currently, the riparian corridor downstream ofCrocker-Huffman Dam is narrower and morefragmented than under historical conditions as aresult of gold dredging, gravel mining,agricultural development, and flow regulation(Stillwater Sciences 2002). These effects mirrorthe consequences of land use on Central Valleyrivers as a whole; of the 900,000 acres of presettlement

riparian forest in the entire CentralValley, only 100,000 acres (11 percent) remain,of which half are in a "disturbed and/ordegraded" condition (Katibah 1984).Currently, riparian vegetation covers 3,928 acresof floodplain downstream of Crocker-HuffmanDam and was mapped as 13 cover types (Table2). The total vegetated area comprises 19.4percent of the pre-dam floodplain (Table 3)which was mapped from existing topographicrelief and is the area assumed to inundatefrequently under the pre-dam hydrologic regimeand support riparian vegetation communities1.


6/13


An additional 4,027 acres of dredger tailing andother disturbed areas were mapped, comprising21.6 percent of the historical floodplain; theseareas contain some small patches of native andexotic vegetation. As described in historicalaccounts (Edminster 1998), the pre-damfloodplain supported a mosaic of riparianwoodland, oak woodland, grassland, slough, andwetland habitats. Current vegetation conditionsvary along the river from a thin band one treecanopy wide throughout most of the river tolarge patches of relatively intact floodplain forest

near the confluence with the San Joaquin River.Non-native woody vegetation comprises lessthan one percent of the existing riparianvegetation. Non-native herbaceous and aquaticplants, which were not mapped separately fromnative species, are both numerous andubiquitous, especially within the herbaceous andmarsh cover types.

Vegetation Encroachment

Vegetation encroachment into the active channelis one of the most widespread and potentiallyintractable effects of flow regulation on alluvial

river systems (Johnson 1994; Scott et al. 1996).In natural alluvial river systems, geomorphicprocesses such as flooding, erosion, andsediment deposition maintain an equilibriumchannel shape and cross section width. Throughthese processes, the river constructs andmaintains a multi-staged channel (whichincludes the low flow, active, and bankfullcomponents) and a floodplain (Figure 3). Withreduced flow magnitude, scour of alluvial bars inthe active channel is reduced, allowing ripariantrees to become established in the former (predam)active channel.

Throughout much of the Merced River studyreach, including the pilot study sites, vegetationhas encroached onto formerly active bars,causing the river channel to become narrowerand eliminating the multi-staged form of thechannel (Figure 4). The resulting channel issimple in cross section, with the current activeand bankfull channel limited to the pre-dam lowflow channel. In an earlier study, Vick (1995)assessed the effects of flow regulation onchannel width in the Merced River based onreview of aerial photographs from 1937 and

1993, and concluded that vegetationencroachment had reduced the active channelwidth from Crocker-Huffman Dam to RM 15.02

by one-third of the mean 1937 channel width.

In addition to narrowing the active channel,vegetation encroaching on regulated riverscommonly establishes lower on the banks thanunder pre-dam conditions (Pelzman 1973;Strahan 1984). Belt transect data for adult treesat the Snelling and Shaffer Bridge sites indicatethat the mean establishment elevation for mostspecies is


7/13

308


TABLE 2. Riparian vegetation cover types and distributions within the Merced River corridor.

Cover Type Description and Habitat Total Vegetation Coverage in

the Merced River Corridor

Total Area

(acres)

Percent of Total

Vegeta-

tion Area

Herbaceous Cover Contains herbaceous plant communities, includinggrassland terraces, tailing transitional areas, and some

seasonal wetlands.

1,362 35

Mixed Riparian Forest Riparian hardwood forest often including Oregon ash,white alder (Alnus rhombifolia), box elder, valley oak, andwillow.

881 22

Cottonwood Forest Contains >50 percent crown canopy Fremont cottonwood

and various subcanopy species combinations.

439 11

Mixed Willow Contains almost exclusively willow, including narrow leaf

willow (S. exigua), Gooddings black willow (S.gooddingii), arroyo willow (S. lasiolepis), and red willow(S. laevigata)

406 10

Valley Oak Forest Contains >50 percent crown canopy valley oak, occurs onterraces, and younger stands have established on former

floodplains that are no longer frequently inundated

342 9

Riparian Scrub Contains early seral stage vegetation (shrubs and smalltrees) of various species that may indicate some form of

regular disturbance or scour

297 8

Marsh Areas with surface water supporting emergent plants;

occurs in some backwater channels and in some dredgertailing swales

66 2

Blackberry Scrub Contains >50 percent crown canopy Himalayan

blackberry(Rubus discolor) or California blackberry(R.ursinus). Occurs commonly adjacent to disturbed areas.

47 1

Eucalyptus Contains >50 percent crown canopy eucalyptus

(Eucalyptus spp). Occurs commonly in monospecificstands on heavily modified banks

45 1

Box Elder Contains >50 percent crown canopy box elder (Acernegundo). Occurs commonly as monospecific stands inthe lower river corridor.

21 1

Giant Reed Contains clonal monospecific standsof giant reed (Arundodonax). Occurs commonly on revetted or otherwisedisturbed banks

12 50 percent crown canopytree of heaven(Ailanthus altissima) , an invasive exotic tree species

10


8/13

309


four feet above summer base flow include whitealder (Alnus rhombifolia), button-willow(Cephalanthus occidentalis), Oregon ash, arroyowillow (Salix lasiolepis), and box elder. Thiselevation range corresponds to a post-daminundation duration of 10 to 27 percent and flood

recurrence interval of 1.2 to 2.5 years. Specieswith mean rooting elevations occurring betweenfour and 8.5 feet above base flow includeEucalyptus (Eucalyptus spp.), Fremontcottonwood, narrow-leaf willow (Salix exigua),Gooddings black willow (Salix gooddingii), andvalley oak. This elevation zone ranges from 1 to10 percent post-dam inundation duration andflood recurrence intervals from 2.7 to 12.3 years.

Vegetation Recruitment and Establishment

Another common effect of flow regulation is aloss of young cohorts of pioneer riparian species

resulting from a lack of suitable establishmentsites (Scott et al. 1996). Willows andcottonwoods require bare, moist alluvial bars togerminate. These surfaces become limited whenencroached vegetation occupies and immobilizesexisting alluvial bars, and flow conditions andlimited sediment supply prevent new bars fromforming. During surveys conducted from RM2.0 to 32.5 in spring 2000, few youngcottonwood saplings (age 2+ years) wereobserved, and of those found, their size and

location and associated flood debris suggestedthat they all established following the January1997 flood event. Cottonwood seedlings (age 4 for Shaffer Bridge, and n>8 for Snelling) was not available at that site for that species.2 Baseflow was defined as 205 cfs, which was calculated as the average of the mean monthly flows for July, August

and September (251, 150, and 214 respectively) for the post-New Exchequer period (1968 to present).3 Based on actual annual instantaneous peak flows from the CDWR gauge Merced River below Snelling(MSN), WY 1968-1997. Recurrence interval flows were fit using the Log Pearson III distribution and differ

somewhat from the distribution calculated by CDWR, which uses a linear interpolation between actual data points.

TABLE 4. Mean elevation above baseflow, inundation duration, and flood recurrence interval for adult treessurveyed at the Shaffer Bridge and Snelling sites.

3Baseflow for the Merced River was defined as 205 cfs, which was

calculated as the average of the mean monthly flows at Snelling forJuly, August and September (251, 150, and 214 cfs respectively) forthe post-New Exchequer period (1968 to present).


9/13

310


When the transect was resurveyed in June 2000,fewer seedlings were documented than in theprevious year. The maximum density was 2.3seedlings/ft2. Seedling survival from the previousyear was 4 percent or less for Fremontcottonwood, arroyo willow, and Californiabutton-willow, and 67 percent for silver maple,which were located farther from the channel

edge (Figure 5). Almost all of the seedlingsdocumented in the June 2000 survey were lessthan one year old (i.e., germinated in spring 2000);78 percent of the Fremont cottonwoods, 40 percent ofthe arroyo willows, and all of the Californiabutton-willows were new recruits (Figure 6).

The pattern of seedling recruitment and mortalityat the Snelling site suggests that Fremontcottonwood, arroyo willow, and Californiabutton-willow readily germinated on bars in theactive channel but did not survive beyond thefirst year. Seedling mortality was likely causedby scour, sediment deposition, or prolonged

inundation. The seedling surveys conducted atthe Snelling site support the river-wideobservation that cottonwood and willow seedlingsreadily recruit along the Merced River but do not surviveto reproductive maturity. Because of the limited scope ofthe pilot surveys, these data should be interpreted assuggestive, rather than definitive, of conditions elsewhereon the river.

Recruitment Box Model Analysis

Riparian tree recruitment depends on localhydrologic conditions during the seed releaseperiod. Early successional species, such ascottonwood and willow, release many seeds thatare viable for a short time, typically 23 weeks(Braatne et al. 1996) and require bare, moistsubstrates to germinate. Seedling recruitment,therefore, occurs on the surfaces that happen tobe moist and bare during the seed release period.Mahoney and Rood (1993, 1998) describe thiswindow of optimal conditions as the

FIGURE 4. Representative cross section of the currentMerced River riparian zone.

FIGURE 5. Seedling survival between October 1999and June 2000, Snelling cross section 13+95.Percent survival for each species is indicated inparentheses.


10/13

311


recruitment box, which is defined bytopographic elevation with respect to river stageand period of seed release and viability (seeannotations A and B in Figure 7). Therecruitment box is constrained at the higherelevations by the seedling's ability to maintain

contact with the receding water table followingfloods, and at lower elevations by inundation andscour the following winter. Within therecruitment box, a further constraint is themaximum survivable rate of water table decline(see annotation C in Figure 7). Numerous studiesreport that seedlings and cuttings of variouscottonwood species survive water table declinesof 11.5 inches/day (McBride et al. 1989;Mahoney and Rood 1991, 1998; Segelquist et al.1993). Flow reduction at rates that exceedpotential root growth generally results inseedling mortality from desiccation.

Recruitment box conditions at the Snelling siteare shown in Figure 8. The cottonwood seedrelease period is estimated from April 15 throughJune 15 (Stillwater Sciences, unpublished data).The vertical axis reflects flow for water years1999 and 2000. Using the rating curve generatedby the Snelling site hydraulic model, the

elevation of the recent cottonwood seedlingcohorts is also plotted on this axis, as well as theroot crown elevations for several mature orsenescent cottonwoods on the site that areconsidered to have established prior toconstruction of Exchequer Dam in 1926, based

on the review of historical aerial photographs.For comparison to actual flow conditions, atheoretical maximum survivable rate of watertable decline of 1.5 inches/day is plotted onFigure 8 during the cottonwood seed release period.

Analysis of the Snelling site data using therecruitment box and hydraulic models indicatesthat the 1999 seedling cohort was not viable forseveral reasons. The seedlings recruited withinthe current bankfull channel, where they wereinundated by flows less than 500 cfs, well belowthe post-dam1.5-year flood (1,338 cfs) level and

the elevation range of mature cottonwoods. Apotential causal factor is the rate of river stagedecline, which was steeper than seedlingtolerance limits (1.5 inches/day) during most ofthe 1999 cottonwood seed release period, andleveled out only when flow was below 500 cfs.Before May 20, it is likely that seedlings would

FIGURE 7. The Recruitment Box concept (redrawn fromMahoney and Rood, 1998). The recruitment box is aspace defined in elevation and time in whichseedlings of riparian plant species are likely tobecome successfully established based on streamflow conditions. The graph represents the relation-ship between the stream hydrograph and the timingof (A) the seed release and viability period for a par-ticular riparian plant species; (B) the range of bankelevations (or stream stages) at which successfulseedling recruitment is likely to occur for thatspecies; and (C) the survivable rate of stream stagedecline determined by the seedlings ability tomaintain functional contact with the receding watertable through root elongation.

FIGURE 6. Annual variability in recruitment betweenthe 1999 and 2000 cohorts at Snelling crosssection 13+95.


11/13

312


not have survived desiccation because the bankdewatering rate exceeded seedling root growthrates. Finally, flows in winter/spring 2000 werehigh enough to submerge the 1999 seedlingcohort for several months and may explain thehigh seedling mortality from 1999 to 2000. It is

also possible that the high water table duringsummer 1999 prevented seedlings fromdeveloping deep root systems, thus making themvulnerable to being uprooted by subsequentwinter flows. Recruitment conditions in 2000were similar to those in 1999. The river stagedecline was very rapid during the seed releaseperiod and leveled out near summer baseflowlevels, resulting in seedlings germinating at lowbank elevations.

The recruitment box concept may also be used todevelop hypotheses of linkages between riverregulation, vegetation encroachment, and species

composition shifts. Reduction in peak flowsduring the spring seed-release period caused byflow regulation may facilitate encroachment byshrub species that form thickets on formerlyactive channel bars and banks. The reduction inspring peak flows, which began afterconstruction of Exchequer Dam in 1926, are notconducive to establishment of trees such asFremont cottonwood and Gooddings blackwillow because seed release for these speciescoincides with spring snowmelt floods, whichprovide suitable germination sites and watertable conditions. Smaller spring peak flows

combined with elevated summer baseflow maypromote establishment of shrub species such asnarrow-leaf willow, which release seeds later inthe summer and propagate vigorously throughvegetative growth.

CONCLUSION

These investigations indicate that riparianprocesses in the Merced River corridor areimpaired in several ways. Riparian vegetation isconfined by land uses and current flow

conditions to only 19 percent of the pre-damfloodplain. Encroachment of riparian vegetationinto the former active channel is widespreadthroughout the river corridor and has resulted ina confined and simplified channel. Inconjunction with a lack of large floods, channelmigration, and alluvial bar formation, thisvegetation encroachment onto formerly activebars prevents establishment of pioneer riparianspecies and arrests natural vegetationsuccessional patterns.

Flow regulation induces cottonwood seedlings torecruit lower on banks than occurred historically,

leading to high mortality from prolongedinundation or scour later in the year. Currently,spring peak flows are insufficient for cottonwoodcohorts to establish on sites such as high-flowchannels and high floodplains that are safe fromsubsequent winter scour and flooding. Theseconditions contribute to the decline ofcottonwood-dominated forest stands throughoutthe river corridor.

Despite these impaired processes, someconditions provide key opportunities forrestoration in the Merced River corridor. Naturalrecruitment of cottonwoods has been observedon riparian sites that have been artificiallygraded or mined for gravel (Stella, personalobservations), and these sites can provide data onedaphic and hydrologic requirements forvegetation reestablishment on reconstructedfloodplains. Also, seed source and dispersal

FIGURE 8. Recruitment Box analysis of the Snellingsite. Flow data is from the Merced River atCrocker-Huffman gage, WY 1999-2000 (MercedIrrigation District)


12/13

313


ability for most tree species within the MercedRiver corridor do not appear to limitregeneration of riparian forest stands. For winddispersedspecies, such as willows and Fremont cottonwood,seed source is abundant and dispersal appears to bewidespread throughout the river corridor. Valley oak,

box elder, and Oregon ash, which have larger seedsand more limited dispersal ability, are well-distributedthroughout the river corridor and are currentlyestablishing naturally on post-dam floodplains.However, good seed source alone does notensure that a desired species mix will occurnaturally on restoration sites; many projects mayrequire active revegetation.

These opportunities and constraints suggestcertain implications for restoration. Restorationstrategies in which stream flow processes aremanipulated to stimulate germination andestablishment of riparian vegetation on

floodplains through natural seedfall wouldrequire channel meandering with formation ofnew alluvial surfaces to re-establish successionalprocesses for pioneer species (Figure 1). In thisapproach, the river system would be essentiallyscaled down, with a flow regime and channelmorphology that mimic pre-dam conditions, butwith reduced flow magnitude and bankfullchannel size. In combination with this river-wideapproach, site-specific interventions can beconducted in which floodplains are reconstructedor graded to favor desired riparian vegetationtypes, which can be planted directly or allowed

to establish through natural recruitment. For allrestoration efforts, monitoring and adaptivemanagement would be required to ensure thatrestoration objectives are ultimately achieved.

ACKNOWLEDGEMENTS

This work was funded by a CALFED grantadministered by the National Fish and WildlifeFoundation. John Bair and Scott McBain ofMcBain and Trush participated in the design andimplementation of all field work conducted forthis project, analysis of vegetation data, andreview. Ralph Boniello and Jeff Oppermanassisted in data collection, methods refinement,and general troubleshooting. Beth Hendricksonof California Department of Water Resourcesprovided field assistance and review. Additionalhelpful review and comments were provided byDr. Joe McBride (University of California atBerkeley Department of EnvironmentalScience, Policy, and Management), Dr. WilliamTrush (McBain and Trush), and Dr. RichardHarris (University of California Extension).

LITERATURE CITED

Auble, G.T., J.M. Friedman, and M.L. Scott. 1994.Relating riparian vegetation to present and futurestreamflows. Ecological Applications 4:544-554.

Bayley, P.B. 1995. Understanding large riverfloodplain

ecosystems: significant economicadvantages and increased biodiversity andstability would result from restoration ofimpaired systems. BioScience 45:153-158.

Braatne, J.H., S.B. Rood and P.E. Heilman. 1996. Lifehistory, ecology, and conservation of ripariancottonwoods in North America. Pages 57-86 in:R.F. Stettler, H.D. Bradshaw, Jr., P.E. Heilman,and T.M. Hinckley (Editors). Biology of Populusand its implication for management andconservation, NRC Research Press, NationalResearch Council of Canada, Ottawa.

Conard, S.G., R.L. MacDonald, and R.F. Holland.1980. Riparian vegetation and flora of the

Sacramento Valley, in riparian forests inCalifornia: their ecology and conservation.Institute of Ecology Publication No. 15,Agricultural Sciences Publications, University ofCalifornia, Berkeley.

CSUC (California State University at Chico). 1998.Sacramento River riparian vegetation maps at1:100,000 scale. Geographic Information Center,CSUC, Chico, CA.

Edminster, R.J. 1998. The Merced River basin:geographic and ecological considerations ofnatural wetlands in Merced County, California.Los Banos, CA.

Decamps, H., M. Fortune, F. Gazelle, and G. Pauou.1988. Historical influence of man on the ripariandynamics of a fluvial landscape. LandscapeEcology 1:163-173.

Fonda, R.W. 1974. Forest succession in relation toriver terrace development in Olympic NationalPark, Washington. Ecology 55:927-942.

Gregory, S.V., F.J. Swanson, W.A. McKee, and K.W.Cummins. 1991. An ecosystem perspective ofriparian zones. Bioscience 41:540-551.

Harden, J.W. 1987. Soils developed in graniticalluvium near Merced, California. In: J.W.Harden (Editor). Soil Chronosequences in theWestern United States, U.S. Geological Survey

Bulletin 1590-A.Holland, R.F. 1986. Preliminary descriptions of theterrestrial natural communities of California.California Department of Fish and Game, nongameHeritage Program, Sacramento.

Johansson, M.E., C. Nilsson, and E. Nilsson. 1996. Dorivers function as corridors for plant dispersal?Journal of Vegetation Science 7:593-598.Johnson, W.C. 1992. Dams and riparian forests: casestudy from the upper Missouri River. Rivers 3:229-242.

Johnson, W.C. 1994. Woodland expansion in thePlatte River, Nebraska: patterns and causes.Ecological Monographs, 64(1):45-84.


13/13

314


Katibah, E.F. 1984. A brief history of riparian forestsin the Central Valley of California. Pages 23-29.in: R.E. Warner and K.M. Hendrix (Editors).California Riparian Systems: Ecology,Conservation, and Productive Management.University of California Press, Berkeley, CA.

Jones and Stokes Associates. 1998. Historical riparianhabitat conditions of the San Joaquin RiverFriant Dam to the Merced River. Prepared forSan Joaquin River Riparian Habitat RestorationProgram, U.S. Bureau of Reclamation, Fresno, CA.

Ligon, F.K., W.E. Dietrich, and W.J. Trush. 1995.Downstream ecological effects of dams: ageomorphic perspective. BioScience 45:183-192.

Mahoney, J.M. and S.B. Rood. 1991. A device forstudying the influence of declining water table onpoplar growth and survival. Tree Physiology 8:305-314.

Mahoney, J.M. and S.B. Rood. 1993. A model forassessing the effects of altered river flows on therecruitment of riparian cottonwoods. In: B.

Tellman, H.J. Cortner, M.G. Wallace, L.F.DeBano, and R.H. Hamre (Editors). RiparianManagement: Common Threads and SharedInterests. General Technical Report RM-226,USDA Forest Service.

Mahoney, J.M. and S.B. Rood. 1998. Streamflowrequirements for cottonwood seedlingrecruitmentan integrative model. Wetlands 18:634-645.

Malanson, G.P. 1993. Riparian landscapes. CambridgeUniversity Press, Cambridge, England.

McBain & Trush. 1997. Trinity River maintenanceflow study, final report. Prepared for the HoopaValley Tribe Fisheries Department, with

assistance from R. Franklin, Hoopa Valley Tribe,G. Matthews, K. Barnard. McBain & Trush,Arcata, CA. November 1997.

McBain & Trush. 2000. Habitat restoration plan forthe Lower Tuolumne River corridor, final report.Prepared for the Tuolumne River TechnicalAdvisory Committee, with assistance fromUSFWS Anadromous Fish Restoration Program.McBain & Trush, Arcata, CA.

McBride, J.R., N. Sugihara, and E. Norberg. 1989.Growth and survival of three riparian woodlandspecies in relation to simulated water tabledynamics. Prepared for Pacific Gas and ElectricCompany, Department of Research and

Development, San Ramon, CA.Mitsch, W.J. and J.G. Gosselink. 1993. Wetlands. Van

Nostrand Reinhold Company, New York, NY.Naiman, R.J. and H. Descamps. 1997. The ecology of

interfaces: riparian zones. Annual Review ofEcology and Systematics 28:621-658.

Pelzman, R.J. 1973. Causes and possible prevention ofriparian plant encroachment on anadromous fishhabitat. Administrative Report 73-1, CaliforniaDepartment of Fish and Game, Region 1.

Roberts, W.G., J.G. Howe, and J. Major. 1980. Asurvey of riparian flora and fauna in California.Pages 3-19 in: A. Sands (Editor). RiparianForests in California: Their Ecology and

Conservation. Agricultural Sciences Publications,University of California, Berkeley.

Rood, S.B. and J.M. Mahoney. 1990. Collapse ofriparian poplar forests downstream from dams inwestern prairies: probable causes and prospectsfor mitigation. Environmental Management

14:451-464.Scott, M.L., J.M. Friedman, and G.T. Auble. 1996.

Fluvial processes and the establishment of bottomlandtrees. Geomorphology 14:327-339.

Segelquist, C.A., M.L. Scott, and G.T. Auble. 1993.Establishment of Populus deltoides undersimulated alluvial groundwater decline. TheAmerican Midland Naturalist 130:274-285.

Stella, J.C. Personal observations of dredger tailingsites, upper Merced River, 1999-2000.

Stillwater Sciences. 2001. Merced River CorridorRestoration Plan, Baseline Studies. Volume II:Geomorphic and Riparian VegetationInvestigations Report. Berkeley, CA. Available at

[http://www.stillwatersci.com/publications.htm].Stillwater Sciences. 2002. Merced River Corridor

Restoration Plan. Berkeley, CA. Available at[http://www.stillwatersci.com/publications.htm].

Stillwater Sciences. Unpublished data on seed releaseperiods collected along the Tuolumne and SanJoaquin Rivers, Spring 2002.

Strahan, J. 1984. Regeneration of riparian forests ofthe Central Valley. Pages 58-67 in: R.E. Warnerand K.M. Hendrix (Editors). California RiparianSystems: Ecology, Conservation, and ProductiveManagement. University of California Press, Berkeley.

Thompson, K. 1961. Riparian forests of the

Sacramento Valley, California. Annals of theAssociation of American Geographers 51:294-315.Thompson, K. 1980. Riparian forests of the

Sacramento Valley, California. In: Sands(Editor). Riparian Forests in California: Their Ecology andConservation, Priced Publication 4101. Division ofAgricultural Sciences, University of California.

Trush, W.J., S.M. McBain, and L.B. Leopold. 2000.Attributes of an alluvial river and their relation towater policy and management. Proceedings of theNational Academy of Sciences 97:11858-11863.

USDA Forest Service. 1997. Ecological Subregions ofCalifornia. S.R. Miles and C.B. Goudey(Compilers). R5-EM-TP-005. Pacific Southwest

Region, San Francisco, CA.Vick, J.C. 1995. Habitat rehabilitation in the lower

Merced River: a geomorphological perspective.Master's thesis, University of California, Berkeley.

Documents

Stella 04 TWS Merced Veg