Northern Prairie Wildlife Research Center
Until about 12,000 years ago, boreal spruce forests extended from the eastern United States to the foothills of the Rocky Mountains and dominated the vegetation of the Great Plains (Wright 1970). During this time, however, much of the Nebraska sandhills remained treeless because the Tertiary alluvium in that area was unconsolidated and strong northern periglacial winds created a system of transverse dunes (Rand 1973). A climatic warming trend in the post-Wisconsin period resulted in the demise of the boreal forest, the establishment of widespread grasslands on the western plains, and the development of deciduous forests in the eastern United States.
There has been considerable speculation about the factors responsible for the lack of trees on the midwestern plains (Borchert 1950; Bryson 1966; Wells 1970). Wells (1970) has suggested that fires which swept unhindered for miles across the prairie, and the arid climate in the plains, have both played important roles in limiting the advance of eastern deciduous forest species into the prairie biome. Fluctuations in the post-pleistocene climate have resulted in wetter periods favoring forest advancement and drier periods favoring grassland dominance. According to Rand (1973), the present climatic period favors grasslands and has remained in effect during the past 900 years. Initial expansion of eastern forests into the plains occurred along the major river courses during more mesic climatic periods when winter rainfall was greater (Kellogg 1905; Gleason 1922). In post-settlement times, the decrease in the frequency and severity of prairie fires and an increase in controlled grazing, encouraged forest development (Bessey 1899).
In Nebraska, the advancement of forests along major river systems such as the Platte was generally confined to a narrow strip along the river channel because the shifting streambed and high scouring action of uncontrolled flows inhibited forest invasion on the wide alluvial bottomlands (Kellogg 1905; Gleason 1922). Kellogg (1905) noted that over long stretches of the Platte timber growth was either "wholly absent" or consisted only of scattered cottonwood and willow. Red cedar was also present in localized areas of floodplain, but this species was most abundant along bluffs in western Nebraska.
According to Kellogg (l905) the range of red cedar expanded along the Platte from the foothills of the Rocky Mountains towards the east at the same time that deciduous forest species were expanding westward. He noted that, although cottonwood and willow were the dominant deciduous species in the advancing forest, other eastern species such as American elm, green ash, hackberry, and box elder were also present, but were generally minor components of the vegetation and their numbers progressively declined towards the West. Along with problems in establishment, much of the forest advance was hindered by the cutting of cottonwood for use in railroad and bridge building, and the cutting of red cedar for fence posts.
Since 1909, climatic conditions in central and western Nebraska have remained fairly uniform, but peak and mean annual discharge in the Platte and North Platte Rivers have-decreased significantly (Williams 1978). In many instances, declines in peak and mean annual discharge correspond to the development of the following reservoirs in Wyoming: Pathfinder, 1909; Guernsey, 1927; Alcova, 1938; Seminoe, 1939; Glendo, 1957; and Lake McConaughy in Nebraska, 1941 (Williams 1978). These reductions in discharge have decreased scouring and shifting of the alluvium on the streambed and have allowed extensive forest development on the floodplain since 1930 (Frith 1974).
Although the expansion of deciduous forest species (i.e., cottonwood, green ash, American elm, diamond willow) from the east and red cedar from the west provided a substantial number of seeds for forest development on the Platte River floodplain, planted tree claims also served as important seed sources and provided seeds from a greater variety of species. The Timber Culture Act of 1873 encouraged the planting of such tree claims, because homesteaders who planted 40 acres of trees were entitled to an additional quarter section of land (Bose 1977). Virtually all the arboreal species which established on the Platte River floodplain, including cottonwood, red cedar, Russian olive, green ash, slippery elm, American elm, red mulberry, box elder, silver maple, hackberry, and Siberian elm, were planted in tree claims (Albertson and Weaver 1945; Read 1958). Cottonwood, green ash, box elder, silver maple, American elm, and slippery elm all have airborne seeds which were easily carried to the river channel, whereas birds were probably responsible for the transport of red cedar, Russian olive, and red mulberry seeds to the floodplain (Kellogg 1905; Albertson and Weaver 1945; McVaugh 1957). Sandbar willow, indigo bush, rough-leaf dogwood, and red-osier dogwood, the dominant shrubs on the Platte River floodplain, probably developed from seed sources brought during the expansion of eastern deciduous forest to the West, whereas the range of buffaloberry and skunk bush, species which are only occasionally encountered, probably expanded eastward from the foothills of the Rocky Mountains (Kellogg 1905). Twelve major vegetation types have been delineated for the Platte River floodplain (Fig. 22). A more detailed summary of the vegetation types on the floodplain is given in Appendix M. Average percentage cover values for the principal species in each type are provided in Appendix N.
Because of the significant negative impact of woody vegetation on use of the river channel by sandhill and whooping cranes, studies were undertaken to determine forest age structure and period of forest development on the Platte River floodplain and to identify factors conducive to establishment of woody vegetation.
Mean annual discharges and associated change of channel widths for the North Platte River in western Nebraska and the Platte River in south central Nebraska are presented in Figs. 23 and 24. All available data on discharge rates for the period from 1910 to 1979 have been plotted.
Although, on the average, mean discharge in the North Platte River (42.63 m3/sec) was considerably less than in the Platte River (56.28 m3/sec) between 1910 and 1977, the general trends in hydrology were the same for the two rivers. In both instances, peak and mean annual discharge have declined substantially since 1909 (Williams 1978). Although the decline in discharge levels has been gradual since 1910, the major decline occurred after 1930. Average pre-1930 discharge levels at North Platte (86.0 m3/sec) and Overton (107.1 m3/sec) were nearly three times the average post-1930 discharge levels at these locations (24.6 m3/sec, and 40.1 m3/sec). Major increases in peak and mean annual discharge also occurred in 1971, 1973, and 1974, as a result of unusually large runoff into the North Platte and Platte Rivers from snowmelt in the Rocky Mountains (Williams 1978).
The age of trees and shrubs growing along the Platte and North Platte Rivers provides insight into factors contributing to the establishment of woody vegetation. The average and earliest age of cottonwood, red cedar, willow, and certain other species in riparian woodlands was determined by counting the annual growth rings on increment cores taken from trees growing along 44 transect lines sampled during a survey of the vegetation (Appendix O). Because of the high frequency of false annual rings in cottonwood and willow and the difficulties in discerning initial growth rings in all species, these age estimates are only approximate, and probably fall within 5 years of the actual dates of establishment (R. Q. Landers, pers. comm.). The development of false rings is probably minimized along the Platte, however, because the majority of trees are located on the floodplain where the water table is relatively high, and there is probably adequate soil moisture throughout the entire growing season.
The oldest trees encountered during the survey were located on uplands, old river shorelines, and raised river islands. Cottonwoods cored near Maxwell, Brady, Kearney, and Mormon Island, with establishment dates of 1900, 1894, 1881, and 1890, respectively, were the oldest trees encountered during the survey. The oldest willow and red cedar were near Overton and Kearney; both were established in 1916 (Appendix O).
Although the earliest dates of establishment provide information about the pre-l909 advancement of woody species along the Platte, the average dates of establishment are more indicative of the period when most forest development occurred. The average age of trees is shown in relation to the study sites (site 1 at Chapman to site 406 at Lake McConaughy) in Fig. 25. Trees that established along the Platte during the relatively high water discharge period before 1930 were generally found on upland sites or along the original (1890's) river shoreline. Trees growing on upland sites have been circled in Fig. 25. Exceptions to this general trend occurred at locations 229-5, 156-5, and 156-8, where diamond willow trees, established in 1926, 1925, and 1916, were growing on lowlands along the major channels. Red cedar in the understory of many of the upland sites has occurred since 1930. In one upland site (44-13) near Mormon Island, hackberry has become the predominant overstory species since 1950.
Most trees became established on the alluvial bottomlands of the North Platte and Platte River floodplains following the major decline in peak and mean annual discharge associated with the completion of Kingsley Dam and the impoundment of water to form Lake McConaughy (Fig. 24). Since 1950, establishment has been confined primarily to the understory red cedar, willow, and Russian olive in the woodlands west of Overton. Red cedar, willow, hackberry, and catalpa have developed in the understory between Overton and Chapman since 1950, but new forest establishment of cottonwood and willow has also occurred.
Cottonwood and willow were the pioneer arboreal species at all locations except at site 208-3 (Darr), and, on the average, developed 19-21 years before red cedar (Appendix O, Fig. 26). Catalpa established only 4 years after cottonwood at site 4-1, near Chapman, but because only one catalpa was cored during the study (catalpa is relatively rare along the Platte), this value may not accurately reflect the average period of establishment for this species. Estimates for establishment of Russian olive were limited to one increment core taken from a tree in a dense stand of Russian olive at site 337-3, near Hershey. Although there was widespread development of Russian olive along the North Platte, and in isolated areas along the Platte, attempts to core additional trees were not made because Russian olive has particularly tough wood and is difficult to core without damaging the increment borer. The Russian olive core was taken from a tree in what appeared to be the oldest Russian olive stand along the North Platte, and indicated that establishment occurred about 1952, approximately 15 years after cottonwood establishment. This establishment date coincides with observations made by local landowners that the majority of Russian olive establishment occurred in the late 1940's and early 1950's (Mrs. C. Summers, pers. comm. and J. Boyle, pers. comm.).
Cottonwood was the only species for which there was a large enough sample of trees growing before 1930 that growth between the 1920-29 period when mean annual discharge was relatively high (95.6 m3/sec on the North Platte, 112.5 m3/sec on the Platte) could be compared to the 1930-39 period when mean annual discharge was substantially lower (43.9 m3/sec on the North Platte, 47.2 m3/sec on the Platte). Total incremental growth of cottonwood in the 1920-29 period averaged 53.18 mm, and declined to 46.09 mm (adjusted for senescence) during the 1930-39 period, but the difference was not statistically significant.
The effects of increased mean annual discharge on growth during 1971, 1973, and 1974 were also investigated by comparing total incremental growth of cottonwood, willow, and red cedar during 1967-70 (mean discharge 17.25 m3/sec on the North Platte; 38.0 m3/sec on the Platte) to total incremental growth of these species during 1971-74 (mean discharge 41.25 m3/sec on the North Platte; 86.25 m3/sec on the Platte). Although mean water discharge in 1972 was not particularly high, growth during this year was included with 1971, 1973, and 1974, to increase the size of the total increment, and thus reduce measurement errors. Because of the potential differences in the growth response of upland and bottomland trees to changes in hydrology, t-test comparisons were made for each species, and independently for trees on upland sites, and trees on lowland (bottomland) sites. Only total and bottomland comparisons were made for willow because only three trees of this species were cored on upland sites.
Radial growth in cottonwood declined on upland sites, increased on bottomland sites, and increased overall (total) between 1967-70 and 1971-74. None of these changes in cottonwood was large, nor was any statistically significant (P<0.05, Table 12). Radial growth in red cedar, however, declined significantly (P<0.05) on uplands, on bottomlands, and overall between 1967-70 and 1971-74. In contrast, radial growth in willow was significantly greater (P<0.01) during 1971-74 on both bottomlands and overall. These data indicate no significant change in the growth of cottonwood with changes in hydrology, whereas red cedar growth was inhibited and willow growth was enhanced during periods of high mean discharge.
Because willow generally establishes and develops in substantially wetter areas of the floodplain than red cedar, it is reasonable to expect that it would respond negatively to periods of low discharge, whereas red cedar would respond positively. Climatic factors such as rainfall and temperature could be partially responsible for these observed changes in radial growth. However, Johnson et al. (1976) and Everitt (1968) concluded that flood frequency and groundwater supplies, rather than climate, were the major environmental variables controlling the radial growth of floodplain trees because of nearly continuous moisture from groundwater and periodic flooding.
The characteristics of the soils found in 37 areas chosen as representative of the major vegetation types along the North Platte and Platte Rivers are presented in Appendix P. Grassland areas sampled along the Platte generally are characterized by well developed lowland prairie soil dominated by silty loam, and occur principally within the Leshara, Grigston, and Hobbs soil types. Except for the poorly drained sandy loam Wann soils on Mormon Island, these grasslands have moderately well-drained soils. Forest soils are usually very well drained whether they are located on uplands or on the floodplain bottomlands. Russian olive was the only woodland type found on poorly drained soils characterized by silty clay loam (Lamo) and sandy loam/clay (Las).
Russian olive establishes primarily on wet grassland sites and is usually associated with intensive grazing of wetland pastures where Lamo and Las soils dominate. In time, however, these grassland sites probably become drier because alluvial deposition and degradation of the river channel effectively elevates these sites in relation to the channel. At site 337-1 near Hershey, for example, where the oldest Russian olive stands along the Platte and North Platte were found, Russian olive is not currently establishing, and the soils are well drained and have been classified as McGrew.
Cottonwood forests were found on numerous soil types, ranging from riverwash alluvium to silty loam and clay loam. Cottonwood apparently establishes only on coarse sandy alluvium along the river channel. The subsequent soil development into a number of soil types is dependent upon the localized environmental conditions at a particular site. The appearance of most young cottonwood and willow trees on relatively undeveloped soils such as Platte and Boel and the known germination and establishment requirements of these species (Moss 1938; McLeod and McPherson 1973) support the hypothesis that these species develop only on bare alluvium. The poorly drained Lamo soils were the only soil type on which cottonwood was not abundant.
Red cedar was typically encountered on upland, well drained soils dominated by silty loam (i.e., Leshara, Grigston, Haverson) and sandy loam (i.e., Darr and Cass) soils. Red cedar was not found on the coarse-textured Platte or Boel soil types at any location except Keystone (site 404-2). At the Keystone site, however, it is suspected that red cedar originally became established on more highly developed upland soils that have subsequently been buried under layers of alluvial overwash. Nevertheless, red cedar establishes late in the vegetational sequence, generally after the development of fine textured forest soils. The upland location and the relatively dry moisture regime that characterize these sites are probably more important factors leading to red cedar establishment than the characteristics of the soils.
The following idealized vegetation development scheme was based on the classification of contemporary plant communities (Appendix M), the age structure of the forest stands, and vegetation-soil relationships. The schematic diagram in Fig. 26 illustrates the major steps in the vegetation development sequence which includes five phases. The first three phases involve sandbar and river island vegetation development and the final two stages are concerned with the vegetation dynamics in advanced stages of forest development. The vegetation sequence described in the final two stages is entirely hypothetical, because these stages in forest development have yet to be reached anywhere within the study area.
Sandbar phase (Phase I)
Exposed sandbars are rapidly colonized by annual species such as lovegrass, nutsedge, and barnyard grass during the initial phase of vegetation development. (Percentage cover values for the principal species in each habitat type are provided in Appendix N). Herbaceous perennial and biennial species such as spikerush, American bulrush, cocklebur, loosestrife, prairie cordgrass, and white sweet clover also develop at this time, but their numbers are usually small because they are not adapted for short-term, high volume seed production, and therefore they generally are not rapid colonizers. Seedlings of woody species, including cottonwood, sandbar willow, diamond willow, peach-leaf willow, American elm, and green ash, also establish during this phase of vegetation development. Hosner and Minckler (1963) and Noble (1979) observed American elm, green ash, box elder, and silver maple seedlings during the initial phase of vegetation development. These four woody species have airborne seeds which could easily be blown to the river floodplain, but the majority of seedlings which establish on the sandbars probably developed from seeds which have been carried in the river.
It is difficult to assess how long this phase of vegetation development persists. In many situations, flooding and scouring removes the Phase I vegetation and causes a reversion back to exposed sandbars. In other situations, the Phase I vegetation may persist for 1-5 years. During the 4-year period, annual species decline and perennials mature to a size at which major flooding and scouring is necessary to remove them.
Shrub phase (Phase II)
As the Phase I vegetation begins to stabilize the sandbars, they become elevated above the river channel from alluvial deposition on the sandbar, alluvial degradation of the river channel, or a decline in average river stage levels. Several authors (Everitt 1968; Hefley 1937; McVaugh 1957; Nickel 1978) indicate that this elevational change is primarily the result of overbank deposition. Nickel (1978) found twigs, seeds, and organic matter indicative of surface sediments, buried under several layers of alluvial sediments in soil profiles taken from Platte River islands in central Nebraska. This elevational change allows perennial species to become better established and enables them to resist removal by flooding and scouring action.
Although species composition in Phase I and II is often very similar, herbaceous species tend to dominate in Phase I, and arboreal and shrub species dominate in Phase II. The prominent Phase II shrubs include sandbar willow, indigo bush, red-osier dogwood, and rough-leaf dogwood; only sandbar willow was found in Phase I. The delayed development of indigo bush and rough-leaf dogwood is probably the result of both a limited seed source and the presence of environmental conditions that do not favor seed germination and seedling establishment of these species under Phase I.
Investigations of seedbanks from Phase I sandbars along the Platte River (Currier, unpublished data) indicated that although viable indigo bush seeds were present in the substrate, rough-leaf dogwood seeds were not. Indigo bush produces tiny indehiscent fruits which are transported via the river to sandbar sites, whereas rough-leaf dogwood seeds develop in fleshy fruits and are distributed primarily by birds. Although indigo bush seeds floating in the river should have nearly the same chance of landing on Phase I or Phase II sandbars, germination and establishment of this species generally occurs on the drier, slightly elevated Phase II sandbars. Birds carrying rough-leaf dogwood seeds, on the other hand, may preferentially land on Phase II sandbars in response to the taller, more substantial vegetation. Conditions for indigo bush and rough-leaf dogwood establishment appear to be ideal along the edges of the raised Phase II sandbars where the soil moisture is intermediate between Phase I and Phase II and large numbers of wind-rowed seeds may be deposited as river stage levels decline.
Trees such as green ash, American elm, and hackberry may occasionally become established during this vegetation phase, but these species require a higher nutrient substrate than cottonwood and willow (Shull 1944; Rand 1973; Johnson et al. 1976) and may be confined to areas where seeds are deposited in alluvial organic material. Weaver (1960) indicates that most floodplain soils are low in nitrogen and organic matter but usually have adequate supplies of other macro-nutrients. Green ash, American elm, hackberry, and other hardwood species may require a higher nitrogen content in the soil before they can establish. Inputs of organic matter through river overwash and forest litter should increase with forest age, and thus the establishment of these species should be favored at later phases in the sequence of vegetation development.
Phase II vegetation develops during the period from 5 to 20 years after initial vegetation establishment. During this 15 year span, shrub vegetation invades, expands, matures, and then begins to decline. Sandbar willow, the pioneer shrub species along the Platte, apparently matures and degenerates after 15 to 20 years. Estimates of sandbar willow longevity range from 20 to 35 years (Wilson 1970; Noble 1979).
Young forest phase (Phase III)
Vegetation development during this phase can follow a number of patterns, depending upon local environmental conditions. Although cottonwood develops as the dominant overstory in most locations, occasionally hackberry, green ash, American elm, red cedar, or Russian olive may become the dominant arboreal species in areas where cottonwood either is not present or is relatively sparse. As a result of continued overbank deposition and degradation of the river channel, these forest sites are generally much drier than Phase I and Phase II sandbars. Sandbar willow, indigo bush, and red-osier dogwood are probably unable to reproduce under these drier conditions, and stands of these species begin to thin as individuals reach maturity and die. The present distribution of rough-leaf dogwood along the Platte, however, suggests that it is able to grow under drier conditions than indigo bush, sandbar willow, and red osier dogwood, and in many areas may replace these species.
In forests with intermittent flooding and saturated soils, Russian olive often occurs in the forest understory, whereas red cedar and red mulberry tend to dominate the understory on drier sites. These species develop late in the vegetation sequence primarily because their seeds are dispersed by birds which roost on young cottonwoods and other trees. Wind-borne seeds of hackberry, green ash, and American elm may also germinate and establish in this forest understory, these species generally occur in low-lying areas in the forest where soil moisture is relatively high and there is an accumulation of organic matter (Shull 1944, Rand 1973; Johnson et al. 1976). Robertson et al. (1978) suggest that differences in micro-relief are responsible for the distribution of arboreal species on relatively flat (i.e., less than 3 m total relief) river floodplains.
In some areas, there is no arboreal or shrub development in the forest understory. In these situations meadow vegetation develops beneath the cottonwood canopy. Such open forests may be the result of intensive grazing, poor initial establishment of shrubs, or the removal of the understory vegetation by continual deposits of alluvial overwash. Forests developing on the floodplain north of Jeffreys Island are representative of such alluvial overwash sites.
Phase III vegetation develops during the period from 20 to 50 years after initial vegetation establishment. Near the end of this phase, cottonwood and willow (primarily diamond willow) are nearing maturity, and the understory species are replacing them in dominance.
Cottonwood maturity phase (Phase IV)
The low moisture conditions in most mature floodplain forests prevents regeneration of cottonwood and willow. Therefore, when cottonwood and willow mature and die, red cedar, Russian olive, green ash, American elm, red mulberry, and other common understory species become the dominants. Only a few areas along the Platte have reached this phase of vegetation development. These areas include the red cedar/cottonwood forests near Kearney (site 106) and Maxwell (site 277), but because few of the overstory trees have actually died, the Kearney and Maxwell sites are considered to be at the initial stages of Phase IV. The length of this vegetation phase is dependent upon the longevity of cottonwood and willow trees. Cottonwood mortality has been estimated at 30 to 75 years after establishment on a number of midwestern floodplains (Shelford 1954; Read 1958; Everitt 1968). Increment tree cores from the Kearney and the Maxwell sites indicate that the oldest living cottonwood trees ranged in age from 79 to 98 years old (Appendix O). Based on these estimates of cottonwood mortality and the actual tree ages from the Platte, this phase of vegetation development has been estimated to occur at 50 to 100 years after initial vegetation development.
Maturity of understory species (Phase V)
Because there are no areas along the Platte where the forest vegetation (other than cottonwood forest) is known to be older than 100 years, vegetation development in this final phase is hypothetical. Red cedar, green ash, American elm, hackberry, red mulberry, and other species probably will dominate after cottonwood and willow have matured and died. As these species mature, they will most likely be replaced by younger individuals of the same species and therefore continue to dominate these sites. Hosner and Minckler (1963) suggested that box elder, American elm, silver maple, and green ash probably would regenerate indefinitely on floodplains along rivers in southern Illinois. Rand (1973) hypothesized a similar perpetuation of box elder, green ash, American elm, and red mulberry along the Republican River in south central Nebraska. Major shifts in the dominant species at particular sites occurs only with substantial changes in the conditions for germination and establishment of these species.
Cottonwood and willows are the principal as well as the pioneer woody species invading sandbars of the Platte and North Platte Rivers. Their presence on crane staging areas poses a serious threat to continued use of the River by sandhill and whooping cranes. Therefore, studies were undertaken to document the environmental conditions under which seed germination, seedling establishment, and seedling mortality occur.
In 1979, cottonwoods released seeds from 3 June to about 26 July on riparian lands along the Platte and North Platte Rivers. The majority of the seeds, however, were released between 5 June and 20 June. Almost all of the seeds collected between 7 June and 20 June were viable. Collections after 20 June were inadequate to test viability of ripened cottonwood seeds. Previous studies indicate that cottonwood and willow release virtually all of their viable seed during May, June, and July (Ware and Penfound 1949; Kapustka 1972).
Willow seed release occurred somewhat later than among cottonwoods. The peak period of willow seed release occurred between 12 June and 20 July, but continued until 30 July. Viability of seeds rapidly decreased from 70-80% on 12 June, to 60% one week later, to 0% by 30 July. Willow seed was released for nearly 2 months, but viability of the seeds during the latter part of the seed release period undoubtedly was quite low. Although the seed release period is relatively short, extremely large quantities of seed are produced. Kapustka (1972) estimates that a 15 cm DBH (diameter at breast height) cottonwood tree annually releases 340,000 to 650,000 seeds. Most of these seeds probably are viable but because of their short viability and seed release periods, and the limited availability of continuously moist, unvegetated substrates, few ever germinate and develop into seedlings.
Seed viability curves for cottonwood and willow under experimental treatment regimes are shown in Fig. 27. The curves represent total percentage germination over a 2 week period for seeds subsampled from each treatment at intervals of 3, 7, 10, 14, 17, and 34 days. Under all treatment regimes, there was a rapid deterioration of seed viability with age. Willow seed remained viable for less than 2 weeks, whereas cottonwood seed retained its viability for nearly 3 weeks.
Although among-treatment differences in seed viability were observed, they were generally small. The largest differences were between high temperature (20-30 degrees C) and low temperature (10-15 degrees C) wet treatments. The high temperature wet treatments had initial (day 3) viabilities close to 100%, but because nearly all the seeds stored under this treatment had germinated by day 3, it was impossible to determine viability for longer periods of storage. Hosner (1958) observed a similar germination response after soaking seeds of cottonwood and willow in water for 4 days. The cottonwood treatment in which the seeds were stored with their pappus hairs was the only high temperature wet treatment in which the seeds did not all germinate within 3 days. Seed viability under this latter treatment, however, declined more rapidly than for all other cottonwood treatments (i.e., from 100% at day 3, to 80% at day 7, less than 10% at day 14, and 0% by day 17).
Wet treatments at the lower temperature regime (10-15 degrees C) had much lower initial (day 3) viabilities for cottonwood (75%) and willow (30%) than under the higher temperature regime. Because seeds stored under wet conditions at 10-15 degrees C did not all germinate within 3 days, viability determinations could be conducted for a longer period of time than under the higher temperature regime. Viability data generally indicate that the lower temperature regime slowed both the initial germination response and the decline in viability over time.
The seed viability curves for the dry treatments in both cottonwood and willow are intermediate between those for the high and low temperature wet treatments. Although initial (day 3) viability for willow under the dry treatment was nearly 95%, it dropped to less than 10% by day 7, and was 0% by day 14. Cottonwood viability declined less rapidly under the dry treatment regime, from an initial (day 3) 90-95%, to 80-90% at day 7, 17% at day 14, 3% at day 17, and 0% by day 34.
These findings are in agreement with the results of previous studies. Although cottonwood has been estimated to remain viable for up to 7 weeks (Horton et al. 1960), most studies place its viability at 2 to 4 weeks (Moss 1938; Ware and Penfound 1949; Kapustka 1972). Willow has been shown to have a shorter period of seed viability of only 1 to 3 weeks (Moss 1938; Ware and Penfound 1949; McLeod and McPherson 1973). By storing seeds at 10% relative humidity, Moss (1938) was able to extend the limited viability periods of cottonwood and willow to 8-10 weeks.
Although 71 permanent 1x1 m quadrants on the floodplain were monitored for seedling establishment and mortality, seedlings were encountered in only 17 (Table 13). A total of 307 mature seedlings (i.e., present at the initiation of the study) were recorded, but only 26 new seedlings were established in these 17 quadrants during the study. The mature seedlings were found on sandbars, in upland shrubland, and in upland woodlands, whereas new seedlings became established exclusively on sandbars (Fig. 28). Two-thirds of all the mature and new seedlings were found in quadrants adjacent to a narrow river channel at the Mormon Island site. Most of the new seedling establishment occurred in quadrants where mature seedlings were already present.
Seedling mortality was almost exclusively confined to sandbar sites, where the majority of seedlings were located. Most seedling mortality occurred between 17 June and 16 July when river stage levels were high. All new seedlings became established after stage levels declined in mid-July, and most survived until the study was terminated on 30 September.
Seedling establishment is seldom, if ever, limited by seed availability because a large volume of viable cottonwood and willow seed is transported via the river. Sandbar availability, however, undoubtedly is very crucial to cottonwood and willow seedling establishment. For example, although the Jeffrey Island site was located closest to a cottonwood forest (Fig. 28), and thus within close proximity to a potentially large seed source, sandbars were limited, and seedling establishment was the lowest of three study locations (Table 13). At the Audubon and Mormon Island sites, forests were less abundant, but sandbars were more extensive and substantially greater seedling establishment occurred (Fig. 28, Table 13).
Seedling relationships to environmental parameters
New seedling establishment was the only dependent variable for which statistically significant (P<0.05) relationships could be shown with the environmental parameters based on multiple regression analyses. The best model is
where Y' is the number of new seedlings, X1 is percentage fine sand, X2 is percentage silt, and X3 is critical percentage soil moisture (Appendix Q). Positive relationships exist between seedling establishment and percentage fine sand and percentage soil moisture during the critical establishment period, and a negative relationship between percentage silt and seedling establishment. Critical moisture is the most highly significant (F = 4.79, P = 0.0322) variable in the model.
Whereas the regression analysis was used to model the environmental conditions under which seedling establishment took place, a reciprocal averaging technique was designed to identify groups of quadrants with similar environmental characteristics, and then to investigate the relationships between seedling and environmental variables in each group. The reciprocal averaging ordination of the 71 permanent quadrants monitored during the study is based on 12 environmental parameters (Appendix Q). The quadrants were divided into three broad habitat types based on their positions in the ordination: upland, raised sandbar, and sandbar.
The distribution of seedlings along the moisture gradient confirms previous observations (Table 13) that mature seedlings were found on both upland and sandbar sites, but new seedlings became established only on sandbars. Among the sandbar sites, however, the raised sandbar supported the greatest number of new seedlings.
The 71 permanent quadrants were grouped into 5 types of sites, based on the 3 broad habitat types (i.e., upland, raised sandbar, sandbar) and the presence of new seedlings. These five groups are upland (U), raised sandbar (R), raised sandbar with new seedling establishment (RS), sandbar (S), and sandbar with new seedling establishment (SS).
Means for all the seedling and environmental parameters in each of the five types of sites are presented in Table 14 and definitions of the environmental parameters are presented in Appendix Q. Analysis of variance indicated that, except for percentage sand, there were significant differences (P<0.05) in the means for all parameters among the five types of sites.
Because the majority of seedling establishment and seedling mortality occurred at RS sites, comparisons were made between the environmental conditions at these sites and those at all other sites. In many instances, the environmental conditions at the RS sites did not differ significantly (P<0.05) from those which characterized the U and R sites. For example, average percentage soil moisture, critical percentage soil moisture, percentage coarse gravel, percentage sand, percentage fine sand, days of substrate exposure, average water depth during flooding, and critical maximum water depth did not differ significantly among the RS, R, and U sites. Although exposure, average water depth, and critical maximum water depth at the RS sites did not differ significantly from the values for these parameters at the U sites, the sites differed fundamentally because the U sites were never flooded. The values for total maximum water depth in the RS and R sites were substantially greater than in the U sites and thus these differences between the U and the R and RS sites were statistically significant. The similarity of the water regime parameters for the R and RS sites was not unexpected because these sites are located at approximately the same elevation on the floodplain (raised sandbars) and were separated only on the basis of the presence of new seedlings.
Soil texture in the RS sites did not differ from that in the R sites. The U sites, however, had a significantly greater percentage of silt than either the R or RS sites. The percentage gravel in the R sites also was significantly greater than that in the U sites, but there was no significant difference between the percentage gravel in the RS sites and that in either the R or the U sites.
All three soil moisture parameters were highest in the RS sites, intermediate in the R sites, and lowest in the U sites. Minimum percentage soil moisture, however, was the only soil moisture parameter which was significantly greater in the RS than in the R or U sites. Soil moisture is critical for the survival of embryonic cottonwood and willow seedlings (Moss 1938), consequently this difference in minimum percentage soil moisture probably is responsible for the significantly greater seedling establishment and numbers of mature seedlings found in the RS sites.
Although there were many similarities in the environmental characteristics of the RS, R, and U sites, there were major differences in the environmental characteristics of the RS and the S and SS sites. In general, the S and SS sites are characterized by shorter periods of substrate exposure, greater water depths, a greater percentage of coarse sediments, and lower percentage soil moisture than in the RS sites.
Comparisons between the RS and the U, R, S, and SS sites indicate that the relatively high percentage soil moisture, fine textured soils, long period of substrate exposure, and moderate water depth during flooding make the RS sites particularly favorable for seed germination and seedling establishment. The number of mature seedlings in the RS sites was also significantly greater than in all other types of sites, suggesting that the RS sites have served as successful habitats for seed germination and establishment in the past. Although mature seedling mortality was significantly greater in the RS sites than elsewhere, this is most likely the result of larger seedling numbers in the RS sites rather than actual differences in mortality rates.
The establishment of seedlings in SS sites on two separate quadrants is more difficult to explain than seedling establishment at the RS sites. The low percentage soil moisture, the coarse, porous sediments, and the relatively short exposure period at the SS sites are not particularly favorable conditions for seedling establishment. However, the greater water depth and more frequent flooding at these sites may compensate for the poor moisture-retaining capacity of the soils and provide an adequate moisture regime for seed germination and seedling establishment. The lack of mature seedlings and the relatively deep water regime at the SS sites suggest that, although seedling establishment can occur, the probability of seedling survival is quite low.