FLOODPLAIN VEGETATION AND SOILS ALONG THE UPPER SANTA ANA RIVER, SAN BERNARDINO COUNTY, CALIFORNIA

2007·9.2 MB·English

by  Jack H Burk

Madrono, Vol. 54, No. 2, pp. 126-137, 2007
FLOODPLAIN VEGETATION AND SOILS ALONG THE UPPER SANTA ANA
RIVER, SAN BERNARDINO COUNTY, CALIFORNIA
Jack H. Burk, C. Eugene Jones', William A. Ryan^, and John A. Wheeler^
Department of Biological Science, California State University, Fullerton,
CA
Fullerton, 92834-6850
Abstract
The patterns ofvegetation and soils were documented in an approximately 20 km^ area offluvial
terraces adjacent to the Santa Ana Riverin southwestern San Bernardino County, California. Within
this area there are three terraces ofdiffering elevations that were last disturbed during major flood
events. TheoldestterracesurfacesprobablywerelastdisturbedduringtheAguaMansafloodof1862,
whichdisturbedorcreatedmanycurrentgeomorphicfeaturesoftheSantaAnaRiverbasin. Themost
recent disturbance oftwo other terraces was identified based on photographs offlooding events in
1938 and 1969. Principal component analysis identified three assemblages of species whose
distribution corresponded to the three terraces that differed in elevation, soil texture, and age since
last disturbance by flooding. Canonical correspondence analysis showed that the assemblages
identified by PCA were highly correlated with changes in soil texture and organic matter. The most
reliable indicator species were Heterotheca sessiliflora and Lepidospartum squamatum for the first
assemblage (associated with the lowest terrace, and corresponding to early successional assemblages
identified by other researchers); Salvia apiana and Senecio flaccidus for the second assemblage
(associated with the intermediate terrace); and Artemisia californica, Opuntia parryi, and
Stephanomeria pauciflora for the third assemblage (associated with the highest terrace). Eriastrum
densifolium subsp. sanctorum, the rare Santa Ana River Woolly Star, was associated with the earlier
phasesofsuccession. Themost important soil factordistinguishingtheseassemblageswasthesilt/clay
content ofthe soil.
KeyWords: alluvial scrub, Eriastrum densifolium subsp. sanctorum, floodplain, SantaAna River, soil
texture, succession.
Much ofthe floodplain vegetation ofthe Santa downstream, resulting in relatively clear water
Ana River has been extensively modified by being released below the dam. Clear water tends
channehzation, urban encroachment, or water to pick up a new sediment load from the area
table lowering, and now may be modified further downstream, eroding the stream bed, cutting
by the recent construction of the Seven Oaks a deeper stream course channel, and further
Dam. This 152.4 m high dam is located where reducing the likeHhood of new sediments escap-
the Santa Ana River emerges out of the San ing the channel and rejuvenating the adjacent
Bernardino Mountains on to a large flood plain habitats in subsequent flooding events.
i
that is north ofthe City ofRedlands and south of Vegetation associations in a floodplain often
the City of Highland in San Bernardino County reflect differences in flooding frequency and
(Fig. 1). intensity. Early successional plant species tend
Baxter (1977) and Nilsson and Berggren (2000) to establish in the open areas adjacent to the river
\
have presented extensive discussions of the course that are more frequently flooded. Species
environmental effects of such dams. A major on more elevated sites, which are much less \
reported effect is that they dramatically reduce frequently flooded, tend to support mid-succes- j
j
the downstream incidence offlooding and reduce sional plant communities. Still more elevated
1
the sedimentation and scouring that are often terraces, which are only flooded during extreme
\
associated with the rejuvenation of early succes- events, tend to support late-successional or
sional habitats. Dams also trap much of the mature associations characteristic of that geo- tj
sediment that would normally be transported graphical region (Blom and Voesenek 1996).
\
In Southern California, many ofthe remaining
I
unmodified alluvial valleys, outwash fans, and
i
' Author for correspondence, email: cejones(^ riverine deposits support an open scrub contain- i
fullerton.edu ing taxa from adjacent coastal sage scrub, j'
^Present Address: Oregon Department ofTranspor- chaparral, a few species also found in the
NtaEt,ionS,alEenvmiOroRnm9e7n3t0a1l-3S8e7r1vices Section, 355 Capitol St. California deserts, and several local endemics [|\i
Present Address: Department ofBiology, University (Ingles 1929). Hanes (1984) described the vegeta- :
of Wisconsin, River Falls, 0405 Ag-Science Building, tion occupying most of the floodplain of the il
River Falls WI 54022-5001 upper Santa Ana River as alluvial scrub. This \'.
2007] BURK ET AL.: SANTA ANA RIVER RIPARIAN SUCCESSION 127
CITYOFHIGHLAND
4000
ft
^
m
/ 1000
SevenOaksDam
fir'.--'9 10 30
2a2b 35 4i^4b
5a. 1519lg§2 33
1b
la '"^^^^^ 3B
2
75 25 22 39^^...-
8^4^3
6 3
REDLANDS
5 CITYOFREDLANDS MUNICIPALAIRPORT
Street
River/Stream Channel
Pnnfu
Ocean
FBiegr.nar1d. inLoocCaotuinotnys,oCfalitfhoernsitau.dyCenatrreaalamnadpsealmepvlaetiosniteiss4i2n5fml.ooTdphleailnocvaetgieotnaotfiotnheofSetvheenSOaanktsaDAanma iRsiivnedri,caStaend
on the far right hand side,just east ofwhere Greenspot Road crosses the Santa Ana River.
vegetation also has been referred to as Riversid- Lepidospartum squamatuni (A. Gray) A. Gray,
ean alluvial fan sage scrub (in the California Asteraceae, (scalebroom) is dominant throughout
Natural Diversity Data Base/Holland classifica- the ASC and rare in CSS; 4) cover of spring
tion [Holland 1986]) and southern alluvial fan wildflowers is greater than in CSS; and 5)
scrub (Magney 1992). In Manual of California riparian woodland species such as Platanus
Vegetation (Sawyer and Keeler-Wolf 1995), it is raceniosa Nutt., Platanaceae, and Baccharts
termed the Scalebroom series. In this paper, we salicifolia (Ruiz Lopez & Pavon) Pers., Aster-
refer to this vegetation type as the alluvial scrub aceae, occur in alluvial scrub but not CSS.
community (ASC). It is considered a rare natural Early vegetation changes following fluvial
community in the State of California, with surface formation have been described for the
a Rarefind occurrence ranking of 1.1 (CNDDB ASC ofthe San Gabriel River floodplain (Smith
2005). 1980). Smith identified three seres: an open
The alluvial scrub community is distinguished pioneer stage lasting 30-37 yrs, an intermediate
from coastal sage scrub (CSS) by the following stage, and a mature stage supporting large
qualities: 1) it has a greater variety of evergreen evergreen shrubs at 40 50 yrs after flooding.
shrubs than CSS; 2) it supports chaparral and She observed that these stages tended to be
desert species (e.g., Rhamnus crocea Nutt., associated with the frequency of flooding events
Rhamnaceae, Rhus integrifolia (Nutt.) Brewer & that scour these surfaces and deposit new
S. Watson, Anacardiaceae, R. ovata S. Watson, sediments.
Anacardiaceae, Prunus ilicifolia (Nutt.) Walp., The recently completed Seven Oaks Flood
Rosaceae, Juniperus californiea Carriere, Cupres- Control Dam on the main channel of the Santa
saceae, and Yucca whipplei Torrey, Liliaceae); 3) Ana River will substantially reduce the magni-
MADRONO
128 [Vol. 54
tude and frequency of major flood events and, of Redlands on the south, the San Bernardino
thus, modify the various riparian communities Mountains on the east, the City ofHighlands on
downstream along the river (Burk et al. 1987), the north and the San Bernardino International
including the ASC. The Santa Ana River ASC is Airport on the west (Fig. 1). The study site tilts
also vulnerable because of sand and gravel mine gradually east to west at a 1-2% slope. Geo-
activity occurring within the floodplain. The morphology of the site presents a complex of
construction of the Seven Oaks Dam and the coalesced alluvial fans originating from the Santa
mining activities within the floodplain could Ana River, Mill Creek, Plunge Creek, and City
reduce theacreageavailable forearly successional Creek. The oldest ofthe geomorphic surfaces are
species and change the species composition ofthe mid- to late-Holocene in age, having been highly
ASC. modified by overland flow that moves fine
The upper Santa Ana River supports three sediment, and even gravels and cobbles (Wood
endangered species, the Santa Ana River Woolly and Wells 1996). The alluvial surface is irregu-
Star {Eriastrwn densifoJiuni (Benth.) H. Mason larly dissected by channels that have avulsed ,
subsp. sanctorum (Milliken) H. Mason, Polem- during major flood events.
oniaceae), the Slender-horned Spineflower {Do- Fluvial deposits from these frequently flooded i
decahema leptoceras (A. Gray) Rev. & Hardham, channels are sand and gravel beds that are ;
Polygonaceae), and the San Bernardino Kanga- deposited or reworked during major flood events i'
roo Rat [Dipodojnys inerriaini Mearns subsp. (Mussetter Engineering, Inc. 1999). The period
parvus (Rhoades) Elliot, Heteromyidae), all of since the surfaces were significantly reworked by
which are associated with this spatially and flood events is the time interval for succession.
temporally vulnerable alluvial scrub community. Woodruff (1980) has described the upper Santa
Therefore, it is important to understand the Ana River alluvial terrace soils as Soboba stony
changes in this vegetation type that follow flood loamy sand. This Soboba soil series consists of
disturbance and the timing of these successional nearly level to moderately sloping soils formed
changes in order to develop conservation prac- from granitic alluvium on alluvial fans. At our
tices to both preserve these species and maintain study site, the Soboba series surface is grayish-
the diversity of this sensitive community. brown stony loamy sand about 25 cm thick
The current study reconstructs successional overlying alluvium up to over 100 m in depth.
changes following major flood events by charac- These soils are excessively drained, have low
terizing vegetation patterns and edaphic changes fertility, low water-holding capacity, and very
after the formation of the surfaces of differing high permeability (Woodruff 1980).
ages (terraces) by erosion and/or deposition In most years, rainfall events do not result in
within the overall floodplain. Specifically, our flows that escape the established river channel
objective is to interpret the chronosequence of and, as a result, do not scour or deposit new
vegetation and edaphic patterns found on extant alluvium over the existing terraces. Known flood
terraces that originated following the largest events large enough to shape overbank areas
recorded flood, the Agua Mansa flood of 1862. through erosion and deposition occurred in 1862,
Specifically, we hypothesized that: 1891, 1916, 1927, 1938, and 1969 (Mussetter j
Engineering, Inc. 1999). The Agua Mansa flood
1) tsohiels,maajnodratgeerrsaicnecse olfastdidfifsetriunrgbaenlceevawtoiounlsd, ionfu1n8d6a2t,edwitmhucahn eosftitmhaeteSdan3t2a0,A0n0a0 cRfsi,veprrofblaoboldy-
have unique species assemblages;
2) cspoemcpioessitaisosnemtbolatgheossewroeuplodrtebdefosrimtihlearSainn wpliatihni,ndoiusrtusrtbiundgyaarneda.moTdhiifsyfilnogodallisocfhtahreactteerrriazceeds
Gabriel River floodplain; aasveraag1e00o-nylryeovnecnet,inthaatceinst,uroyne(UtShaAtCoEcc2u0r0s0).on jj
3) increasing canopy cover as reflected in The U.S. Army Corps of Engineers (USACE \
surface light measurements would be a fac-
2000) has simulated these historical flood over-
tor in species composition differences found
among terraces; and flows and concluded that the second largest !
overflows occurred in the similar magnitude
4) edaphic factors would be related to time floods of 1891 and 1938 (both with 100,000 cfs).
since last disturbance of the terrace and These floods created the somewhat lowerterraces ;
wspoeuclieds calosmoposbietiroenladtiefdferteoncdeisffaemroenncgesteirn- that are now considered to be the intermediate-
aged terraces. Next in magnitude were the floods
races. !
of 1916, 1927, and 1969 (two events of32,460 cfs
;
on 25 Jan. 1969 and a second one of40,495 cfson
25 Feb. 1969), which also resulted in overbank jj
Methods and Materials flows. Aerial photographs from 1938 and 1969 ||i|
clearly show that terrace surfaces that we
ii
Study Site. This investigation focuses on an consider to be intermediate- and recent-aged
i
approximately 20 km- area bounded by the City respectively were disturbed by floods in those li
ii
2007] BURK ET AL.: SANTA ANA RIVER RIPARIAN SUCCESSION 129
two years. Flow rates for the 1938 and 1969 perpendicular to this base line. In the grid-
floods were recorded on the U.S.G.S. Stream selected sample set, each of the 40 samples
Gauge #11066460 located near Arlington in the consisted of a randomly oriented 50-m line
City ofRiverside, CA (David Lovell, Department transect placed at the randomly selected grid
of Public Works, San Bernardino County, point. Orientation was constrained in order to
Personal Communication). keep the entire transect on a single fluvial surface
and to avoid structures and disturbances. In-
Sample sites. After an analysis of the surfaces tercept length to the nearest 5 cm was recorded
(i.e., terrace surfaces) within the study area, for woody plants (by species), plus herbaceous
sample sites were chosen based on their being plants, bryophytes/cryptogamic crust, litter, and
included within the three major terraces of dead shrubs. No attempt was made to distinguish
differing ages (see Burk et al. 1987 for a detailed individual herbaceous species since the majority
map ofthese terraces). Four sample site locations of the sampling was carried out during the dry
w1itahnidn5t)heaanrdea1s9c3r8ea(tseadmpblyetshiete1s9269a(nsdam4p)lfelsoiotdess csoenassiosnt.edAolfmionsttrodaullcedofantnhuealhgerrabssaecseoaunsd hceorvbes.r
(Fig. 1) were assigned by two methods. First, ten Nomenclature follows Hickman (1993) and
sites, designated by both numbers (1, 2, 3, and 5) voucher specimens are filed at the Fay A.
manadpplietntgersth(ea-dbi)s,trwiebruetisoynsotfemEartiiacsaltlryumlodceantseidfoalfituemr MacFadden Herbarium (MACF) at California
State University, Fullerton.
subsp. sanctorum (eds) (Burk et al. 1987). It
appeared that several areas that lacked eds were Soils. Overthe longtemi, asfluvial terracesage
quite similar in vegetation and substrate to their soil properties slowly tend to differentiate
adjacent areas containing eds. Therefore, in order (Jenny 1941). However, recent work on similar
to study factors associated with the distribution fluvial terraces in the Cajon Wash (a system that
of this endangered species, sites designated with empties into the Santa Ana River just west and
an ''a" were placed randomly within areas slightly south ofour study area), has showed that
occupied by eds, whereas nearby sites of the silt and organic matter rapidly accumulates
same age, designated with a "b", contained during the first 40 yrs after flooding even in the
apparently comparable vegetation but lacked soils ofthese terraces (McFadden and Weldon II
the presence of eds. These plots provided 1987). Specifically, they argue that the vast
a sampling of terrace surfaces last disturbed majority of this rapid accumulation is due to
during the 1969 and 1938 floods (what we the influx of eolian (airborne) dust. This influx
considered to be the young and intennediate- not only increases the organics but also increases
aged terraces). the water holding capacity of these soils. Such
Forty additional sample sites (grid sample changes could affect the rate of plant succession
plots) were selected in a stratified random manner in such soils and have a significant effect on
on a grid that extended from the northern the persistence of particular early successional
boundary to the southern boundary through the
species.
central section ofthe floodplain (Fig. 1, all other Therefore, soil samples were collected to
numbered sites). These sites were selected to determine the extent of silt and organic matter
thoroughly sample all terraces within the flood- accumulation at our sample sites, using separate
plain, including those formed in 1862. The grid protocols for the eds comparison sites and the
contained eight 70 ha cells, each 30 min of grid sample plots. For the eds comparison plots,
latitude and longitude on a side, located from a soil sample was obtained from a random
34 5'00" to 34 6'30"N and from 117 10'30" to location along each 30-m line transect used for
117^11'30"W. Six sites were selected per cell: one vegetation sampling for a total of three sub-
site in each of the four quarters of the cell and samples from each comparison site. For the grid
one additional site in both the north and south sample plots, soil samples were obtained by
heallivmeisnaotfedeawchhecnellt.heEyigphtrosvieteds twoerfeallsuobustesqiudeenttlhye tthaekin5g0-fimvetsruabns-escatmplsetasrtaitng10f-rmomintaervralasndaolmolnyg
fgthlleooboadflpilepladoisniwtaiirteohan.ianEgaMscayhgsetslealmmap(nlMeaGgsPeiltSelawnNaasNvalvoilcgaOatOteOidPorni,on isdeelnetcitceadlpomiantn.neSroilatsamalpllessitewse:reAob2t5a-incemd mientaanl
Inc., San Dimas, CA). cylinder (10-cm in diameter) was forced vertically
into the soil, then the upper 10-cm of soil was
Vegetation sampling. In order to characterize removed from the cylinder, and the remaining
the ASC both qualitatively and quantitatively, soil, from 15 25 cm depth, constituted the
plant species cover at each site was measured sample. All samples from each sample site were
using the line-intercept method (Barbour et al. thoroughly mixed and a sub-sample was removed
1998). In the Eriastrum comparison sites, a 100-m for analysis.
base line was randomly placed. Three randomly Soil samples were analyzed for texture, organic
positioned 30-m line transects were then placed matter, and salinity. Samples were sieved through
MADRONO
130 [Vol. 54
1.25- and 2-mm screens to separate larger plus the 3 general plant groups (herbaceous,
fractions. Approximately 100 grams of the < cryptogams, and dead shrubs). However, woody
mm
2 fractionwere analyzed fortextureusingthe species occurring in only one sample were
Bouyoucos hydrometer method (ASTM Method subsequently omitted from the analysis leaving i
D433) after removal of organic matter by 23 active "species" under consideration. Sample
mcaotmtbeursticoonnteanstouotflitnheed b<y C2omxm(19f8r5a)c.tiOorngawnaisc esiltiemsin6,ate15d, f31r,om32,an3a3l,ys4i2s,s4i6n,ceantdhe4y8pwreorveedalstoo
I
determined gravimetrically after burning a 100 g contain extreme values in either soil or cover,
oven-dried sample in a muffle furnace at 500 C leaving 42 active samples. In both PCA and i
for24 hrs. Silt, combinedwith clay, is reported as CCA, centering was focused on species. In CCA, i
"fines" and represents the < 0.05 mm fraction. rare species were downweighted and there was no
mm
Values reported as sand are for the 0.05-2 mm rescaling or detrending. Of the 11 measured
fraction and gravel is defined as the 2-13 environmental variables, four factors (fines,
fraction. organics, sand, and gravel) made significant ^
Salinity was determined using the conductance contributions to the analysis based on evaluation
of the supernatant from a mixture of 20 g of ofinflafion factors (Burk et al. 1987; Ryan 1995). '
< 2 mmsoiland 100 ccofdeionizedwaterequated Monte Carlo tests ofthe first canonical axis and
to standard solutions of NaCl mg/1 converted to all canonical axes were run with unrestricted
soil osmotic potential using Van Hoffs law (after permutations. CCA ordination plots are biplots
Burket al. 1987). Theunits used in the analysis are of species and environmental factors. No trans-
conductance units ofmillimohs per meter. formations of variables were employed in the ^
analyses.
Light measurements. Light levels were mea- \
sured at each sampUng site using a LiCor Model
170 hght meter fitted with a quantum sensor and Results ^
served two separate purposes. Ambient light
30 cm above the ground was expressed as Descriptive Statistics- Descriptive statistics for
percentage of full unobstructed light above the species cover data and for environmental vari- ,
plant canopy and provided a measure of light ables are presented in Tables 1 and 2, respective-
transmission through thatcanopy. Reflected hght ly. Eight non-perennial plant categories (i.e.,
from the soil surface, measured by turning the herbaceous ground cover and substrate cate-
sensor toward that surface, was also expressed as gories) were encountered on the transects. Of
a percentage of full unobstructed hght. It was these, grass, bare ground, and dead shrubs had
used as an indirect measure of soil coverage and the highest overall mean cover values (Table 1).
color. All light measurements were taken on days Thirty species of perennial vascular plants
with less than 25% cloud cover and never when occurred on the transects, many of which were
a cloud shaded the site. These measures were also only encountered occasionally. Only five species
taken between 1000 and 1600 hrs Pacific Stan- had mean cover values above 2%: Eriogonum
dard Time only, in order to avoid low sun angles. fasciculatum Benth., Polygonaceae, Juniperus
The maximum unobstructed light value for each californica, Eriodictyon trichocalyx, Lepidospar- \
point was taken as the maximum forthat point in tum squamatum, and Adenostoma fasciculatum
order to standardize the data for sun angle. In all Hook. & Arn., Rosaceae. Percent cover as \
measurements, care was taken to orient the reflected in light measurements did not vary |
sensor perpendicular to the soil surface. The significantly with terrace age. Ambient light (%
readings were recorded at 5 m intervals (n = 10) ofmaximum) reflected from the soil surface was
along the transect from a randomly selected quite high overall, indicating an open habitat
starting point with the sensor 30 cm above the (Table 2).
gforrouanmdbaitenetachanpdointr.efTlehcetepderlciegnhttagaet oefafcuhll lsiugbh-t Ordination. The CCA eigenvalue for Axis 1
sample was then calculated. The means ofthe ten (0.152) is three times greater than the eigenvalues
mtheeasvuarleuedslfioghrtapmebriceennttaagensd wreerfleecutsededlitgohtobletvaeilns AtfohxreestPheC1-Ao4trhaeersrepeac0xt.ei7sv1e,l(yT0.a.0bW6l8ee,3p)0r..e0sT5eh3n,et aebinogdtehn0vt.ah0le4u3ePsCfooArf |
for each sample site. and the CCA results because PCA eigenvalues |
Ordination. In order to describe the nature of are substantially higher for Axis 1, suggesting
the relationships between the vegetation and the that the relationship betweenenvironment factors
environmental factors, principal component anal- and species can be determined with CCA.
ysis (PCA) and canonical correspondence analy- Axes 1 and 2 of the PCA suggested three
suissin(gCCpAro)towceorles pienrCfoArmNeOdCoOn aflolr50Wisnadmoplwess(itteers aspsescoiceisatgiroonu(pFsiga.n2d).foEurriosgpeocniuemsfwaistchiocuutladteufimni(teirvf)e ,[i
Braak and Smilauer 1998). Thirty-six units were and Opuntia litoralis (Engelm.) Cockerell, Cacta- ii
analyzed: 33 woody plants identified to species ceae (ol) were at opposite poles ofboth axes from
i
i
2007] BURK ET AL.: SANTA ANA RIVER RIPARIAN SUCCESSION 131
cToavbelreas1.recDoersdcerdifprtoimvefoSrttayt-itswoti5c0smfotrraSnpseecctise.sMCoosvtesrpeDciaetsaw.erSepencoitesenarceoulnitsteerdeidnidneaslcletnrdainsnegctosr,detrheoreffmoreeanzer%o
values were included in the calculation ofthe means. All forty-two sample sites are included.
Species Mean ± Standard Deviation Minimum-Maximum
Grass 32.0 -H 20.5 0.5-64.3
Bare ground 14.2 -H 13.6 0.3-59.5
Dead shrubs 9.5 -1- 6.5 0.2-28.9
Eriogonumfasciculatum 6.2 -H 6.8 0.0-32.5
+
Bryophytes/cryptogamic crust 5.3 5.5 0.0-27.0
Tall herbs 5.1 3.9 0.0-15.4
Stone 3.3 4.4 0.0-25.2
Juniperus californica 3.0 6.6 0.0-22.7
Eriodictyon trichocalyx 3.0 ± 3.9 0.0-15.9
Lepidospartum squamatum 2.5 4.9 0.0-25.2
Adenostomafasciculatum 2.2 5.5 0.0-24.7
Gutierrezia californica 1.8 0.5 0.0-11.9
Yucca whipplei 1.7 2.6 0.0-13.3
Low herbs 1.6 ± 1.8 0.0-7.0
Lotus scoparius 1.4 -¥- 1.9 0.0-7.0
Bebbiajuncea 1.1 -+- 2.9 0.0-11.8
Prunus ilicifolia 0.8 -+- 4.3 0.0-26.4
Opuntia littoralis 0.8 -+- 1.3 0.0^.8
Salvia apiana 0.7 -+- 2.3 0.0-11.0
Enceliafarinosa 0.6 -H 1.9 0.0-8.0
Artemisia californica 0.5 -+- 1.5 0.0-7.1
Mirabilis californica 0.5 -+- 1.2 0.0-5.5
Rhus ovata 0.4 -+- 1.9 0.0-12.3
Stephanomeriapauciflora var.pauciflora 0.3 -H 0.8 0.0^.3
Ericameriapinifolia 0.3 -H 0.9 0.0^.4
Opuntiaparry 0.3 H- 0.7 0.0-2.8
Bunchgrass 0.2 -H 0.8 0.0-3.8
Eriastrum densifolium subsp. sanctorum 0.2 -H 0.8 0.0-5.1
Phoradendron densum 0.1 -H 0.5 0.0-3.1
Cercocarpus betuloides 0.1 -+- 0.7 0.0-4.3
Marah macrocarpus var. macrocarpus 0.1 0.5 0.0-3.0
Solanum xanti 0.1 -+- 0.4 0.0-2.2
Senecioflaccidus var. douglasii 0.1 -H 0.3 0.0-1.8
Cuscuta sp. 0.0 -H 0.3 0.0-1.8
Gnaphalium californicum 0.0 0.2 0.0-1.5
Croton californicus 0.0 0.1 0.0-0.7
Dudleya lanceolata 0.0 0.0 0.0-0.1
Calystegia macrostegia 0.0 0.1 0.0-0.4
Table 2. Descriptive Statistics for Measured Environmental Variables. All forty-two sample sites
are included.
Variable Mean ± Standard Deviation Minimum-Maximum
Litter cover (%) 56.9 26.8 0.0-99.0
Clay particulates (%) 2.4 1.0 0.2^.3
Silt particulates (%) 7.0 3.5 1.7-15.6
Sand particulates (%) - (sand) 90.5 -+- 3.5 82.3-96.2
Fine particulates (clay + silt %) (fines) 9.5 4.1 3.9-17.7
Salinity/conductance (millimohs/m) 0.4 0.3 0.1-1.0
Organic Matter (% dry weight) - (Organics) 0.6 0.6 0.2^.0
mm
2-13 soil particles (% dry weight) (gravel) 10.5 6.9 2.1-33.6
< 2 mm soil particles (% dry weight) 87.7 7.7 65.3-97.6
Ambient light (% ofmaximum) 81.3 11.2 50.6-99.6
Reflected light (% ofmaximum) 8.2 3.3 4.2-23.6
MADRONO
132 [Vol. 54
Table 3. Summary of Canonical Correspondence Analysis of Floodplain Communities of the
Santa Ana River Near Redlands, California. Input data included 42 samples, 23 species with 490
occurrencesand4environmentalvariables. ** Indicatesthataparticularcanonicalaxiswassignificantly(P< 0.01)
correlated with environmental factors.
Statistic Axis 1 Axis 2 Axis 3
Eigenvalue 0.152 0.057 0.035
Species/environment correlations 0.677** 0.5989** 0.579**
%
Cumulative variance ofspecies data 9.6 13.2 15.4
%
Cumulative variance ofspecies-environment relation 58.1 79.8 93.2
Juniperus californica (jc) and Lotus scoparius species centroids were projected perpendicular to
(Nutt.) Ottley, Fabaceae (los) and thus had quite the line representing sand, the species from PCA j
different affinities and are not included in any of Group 1 formed a group at higherconcentrations '
the three groups. (Species are identified with of sand. PCA group 2 intermingled with PCA
symbols used throughout this paper.) Group 1 on Axis 3.
The CCA correlations between species and The last two columns in Table 5 show the axis
environment ranged from 0.462 for Axis 4 to tolerance for each species on Axes 1 and 3 ofthe
0.677 forAxis 1. The relationship between species CCA ordination. A low value indicates a narrow
and the four environmental variables was highly niche breadth for the principal environmental
significant (P < 0.01). The cumulative percent of factor driving each axis and thus identifies
the variance of the species-environment relation indicator species for each PCA group. Axis 2 is
indicates that the environmental conditions not included because of a lack of interpretable
measured were probably important contributors separation on that axis.
to species distribution.
The CCA biplots of species scores and Discussion
environmental factor scores in Fig. 3A showed
that Axis 1 was related first (r = —0.61) to the Species cover data document that the upper
amount of fines and secondly to organics (r = Santa Ana River alluvial zone is an open habitat
—0.47). Axis 2 was related (r = —0.32) princi- supporting a diversity of annual and perennial
pally to organics. Axis 3 was related to gravel (r plant species including, in some, but not all sites,
= +0A9) and sand (r= -0.51) (Fig. 3B). Table 4 the rare Santa Ana River Woolly Star (eds).
presents the statistics on the majorenvironmental Measured environmental variables show that the
variables. The relationship between individual soils are characterized by a high content ofsand |
species and the environmental axes as determined or particles > 2 mm in size and that the percent i
from t-value biplots showed that multiple species ofambient light reaching the soil surface is quite
had significant correlations with the environmen- high in all terraces, again indicating that the
tal axes (Table 5). habitat is quite open. These features are expected
When species scores in Fig. 3A were projected in periodically flood-rejuvenated floodplain hab-
\
to the lines representing fines, three groupings of itats.
species were apparent. Heterotheca sessiliflora Regarding succession within alluvial scrub, our
I
(Nutt.) Shinn., Asteraceae (hf), Enceliafarinosa findings are consistent with those ofSmith (1980) •
Torrey & A. Gray, Asteraceae (ef), and Lepido- and Hanes et al. (1989) who concluded that there
spartum squamatum (Is) were consistent indica- are three serai stages apparent in southern
tors of soils with the least silt and clay content. California alluvial scrub vegetation following
i
All ofthe species that fell on the negative side of major disturbance ofterraces by flood. Our data, i
Axis 1 were associated with higher silt/clay however, suggest much longer periods of domi-
content and formed a second group that, with nance for each serai stage and redefine the two |
the exception of the outlier Croton califomicus later serai stages described by Hanes et al. (1989) j
f
Muell. (Euphorbiaceae [cc]) were the members of and Smith (1980).
PCA Group 3. The third cluster had affinity with In our data, terraces differed in plant assem- |I(
soils of intermediate silt/clay content and in- blages, soil attributes, and time since a major t|(
cluded all species between Yucca whipplei (yw) flood disturbance. The dominant plants and i
and Ericameria pinifolia (A. Gray) H. M. Hall edaphic conditions that characterized each of I
(oAfstPerCaAceaeGr[eopu])p. Th2eatnhidrd3clusspteecriewsa.s aOrmgiaxntiucrse wthaestienrdriaccaetsedarbeyasthefolplroewsse.ncTeheoffLiresptiadsoesmpbalratugme Ij
correlated with Axes 1 and 2 almost equally squamatum and/or Heterotheca sessiliflora. Ter- |p)i
and were the only important correlate ofAxis 2. races that were modified by the 1969 flood '\{
TohnerAexfiosre,2.liCttlCeAintAexripsret3abwlaessecpoarrraetliaotnedocaclumrorsetd asluspopodretpeedndtahibslyasfsoeumnbdlagoen.soEinlcselcihaarfaacrtienroiszaedwabsy :,:s1
equally with sand and gravel (Fig. 3B). When high sand, and low organic and clay content
2007] BURK ET AL.: SANTA ANA RIVER RIPARIAN SUCCESSION 133
Group 2
Fig. 2. Principal components analysis axis 1 (horizontal) and axis 2 (vertical) for upper Santa Ana River
floodplain species scores. Thedelimited groups are the PCA groups referred to in the text. Open symbols represent
species with no group affinity. Species codes are explained in Table 5. In comparison to Fig. 3, this figure is
rotated 180\
(which were characteristic of the lower elevation selected) species, probably because the washed
terraces disturbed by the 1969 flood), but was not sand substrate lacks nutrient and water holding
widely distributed. Other plants with centroids in capacity and is, therefore, nutrient limited.
the first assemblage included the endangered Eriogonumfasciculatum was not clearly associat-
Eriastrum densifolium subsp. sanctorum and the ed with this early serai stage in our data; rather it
more common Bebbiajuncea (Benth.) E. Greene proved to be an isolated species whose distribu-
(Asteraceae). Croton califomicus also had its tion was difficult to interpret. However, Smith
centroid in this group but it was generally (1980) and Hanes et al. (1989) did find that E.
distributed and, therefore, was of little value as fasciculatum was an early successional species,
an indicator. unlike the case in our study.
Hanes et al. (1989) and Smith (1980) refer to The character ofthe other two assemblages in
this first assemblage as the pioneer community. It our study differed significantly from the in-
is pioneer only in the sense of being first after termediate and mature stages described by Hanes
flood. It contains essentially no ruderal (r- et al. (1989) and Smith (1980). Malosma laurina

2007] BURK ET AL.: SANTA ANA RIVER RIPARIAN SUCCESSION 135
Table 4. Summary Statistics for Soil Features Included in the Canonical Correspondence
AnalysisOF Floodplain Vegetation alongthe upper Santa Ana River, California.
Statistic Gravel Sand Fines Organics
Mean (%) 9.2 82.0 8.5 0.4
Standard deviation 6.0 10.1 4.7 0.2
Minimum 0.7 61.3 1.4 0.1
Maximum 23.2 98.6 17.7 1.0
(Nutt.) Abrams (Anacardiaceae), Rhus integrifo- The second assemblage on the upper Santa
lia, Rihes aureuni Pursh. (Grossulariaceae), and Ana River alluvial zone supports Senecio flacci-
Salvia niellifcra E. Greene (Lamiaceae), which dus Less. (Asteraceae) and Salvia apiana Jepson
Smith (1980) considered indicators ofthe mature (Lamiaceae), which had narrow distributions on
stage, did not occur in our samples and are soil gradients, and were, therefore, good indicator
uncommon in the upper Santa Ana River species. Additional shrubs with centroids in the
floodplain. Smith indicated that her mature intermediate assemblagewere Ericameriapinifolia
stands were 47+ years old, which roughly and Yiiccca whipplei. The presence ofcryptogam-
corresponds to the age of stands on terraces last ic crust and dead shrubs was also noted in this
disturbed in 1969 in the current study. phase. Other species found in the second assem-
Table 5. Listof Species Included inthe Samples. Codes and PCA Group are included for those species in
the ordinations. A (?) indicates an unclear group affiliation, an (*) indicates a significant (P < 0.05) correlation
between the species and the environmental factor and (ns) indicates non-significance. Axis tolerance is the root
mean squared deviation for each species on axes 1 and 3.
PCA Axis 1 Axis 3
Species Code Group Fines Organics Sand Gravel Tolerance Tolerance
Adenostouui af 3 ns ns * ns 0.97 0.92
fasciculatiini
Artemisia ac 3 ns ns * * 0.76 0.90
californica
Behhiajimcea bj 1 ns ns * * 0.76 0.86
Croton cc 1 ns ns ns ns 1.63 0.78
californicus
Enceliafarinosa ef 1 * * * * 0.24 0.43
Eriastrum ed 1 * ns * * 1.68 0.85
densifolium
Ericameria ep 2 * * * * 1.24 1.20
pinifolia
Eriodictyou et 3 * * * * 1.18 0.88
trichocalyx
Heterotheca hf 1 * ns * * 0.04 0.29
sessiliflora
Juniperus jc 7 * ns * * 0.81 0.79
californica
Lepidospartuni Is 1 * * * * 0.86 1.29
squamatum
Lotus scoparius los 7 * ns ns ns 0.76 0.60
Mirahilis mc 2 * ns * ns 0.80 0.93
californica
Opuntia litoralis ol 7 ns * * 0.88 1.07
Opuntiaparryi op 3 * ns * * 0.74 0.92
Salvia apiana sa 2 * * * * 0.39 0.66
Senecio flaccidus sf 2 * ns * * 0.31 0.18
Stephanomeria sp 3 * * * * 0.54 0.97
pcmciflora
Yucca niiipplei yw 2 ns ns ns ns 0.88 1.11
Herbaceous h 3 * * * 0.89 0.98
plants
Dead shrubs ds 2 * * * * 0.99 0.97
Cryptogams br 2 * * * * 0.80 1.04