Jonathan M. Friedman

FLOOD DEPENDENCY OF COTTONWOOD ESTABLISHMENT ALONG THE MISSOURI RIVER, MONTANA, USA

Flow variability plays a central role in structuring the physical environment of riverine ecosystems. However, natural variability in flows along many rivers has been modified by water management activities. We quantified the relationship between flow and establishment of the dominant tree (plains cottonwood, Populus deltoides subsp. monilifera) along one of the least hydrologically altered alluvial reaches of the Missouri River: Coal Banks Landing to Landusky, Montana. Our purpose was to refine our understanding of how local fluvial geomorphic processes condition the relationship between flow regime and cottonwood recruitment. We determined date and elevation of tree establishment and related this information to historical peak stage and discharge over a 112-yr hydrologic record. Of the excavated trees, 72% were established in the year of a flow >1400 m3/s (recurrence interval of 9.3 yr) or in the following 2 yr. Flows of this magnitude or greater create the necessary bare, moist establishment sites at an elevation high enough to allow cottonwoods to survive subsequent floods and ice jams. Almost all cottonwoods that have survived the most recent flood (1978) were established >1.2 m above the lower limit of perennial vegetation (active channel shelf). Most younger individuals were established between 0 and 1.2 m, and are unlikely to survive over the long term. Protection of riparian cottonwood forest along this National Wild and Scenic section of the Missouri River depends upon maintaining the historical magnitude, frequency, and duration of floods >1400 m3/s. Here, a relatively narrow valley constrains lateral channel movement that could otherwise facil- itate cottonwood recruitment at lower flows. Effective management of flows to promote or maintain cottonwood recruitment requires an understanding of locally dominant fluvial geomorphic processes.

Fluvial process and the establishment of bottomland trees

The effects of river regulation on bottomland tree communities in western North America have generated substantial concern because of the important habitat and aesthetic values of these communities. Consideration of such effects in water management decisions has been hampered by the apparent variability of responses of bottomland tree communities to flow alteration. When the relation between streamflow and tree establishment is placed in a geomorphic context, however, much of that variability is explained, and prediction of changes in the tree community is improved. The relation between streamflow and establishment of bottomland trees is conditioned by the dominant fluvial process or processes acting along a stream. For successful establishment, cottonwoods, poplars, and willows require bare, moist surfaces protected from disturbance. Channel narrowing, channel meandering, and flood deposition promote different spatial and temporal patterns of establishment. During channel narrowing, the site requirements are met on portions of the bed abandoned by the stream, and establishment is associated with a period of low flow lasting one to several years. During channel meandering, the requirements are met on point bars following moderate or higher peak flows. Following flood deposition, the requirements are met on flood deposits ;high above the channel bed. Flood deposition can occur along most streams, but where a channel is constrained by a narrow valley, this process may be the only mechanism that can produce a bare, moist surface high enough to be safe from future disturbance. Because of differences in local bedrock, tributary influence, or geologic history, two nearby reaches of the same stream may be dominated by different fluvial processes and have different spatial and temporal patterns of trees. We illustrate this phenomenon with examples from forests of plains cottonwood (Populus deltoides ssp. monilifera) along meandering and constrained reaches of the Missouri River in Montana.

Relating riparian vegetation to present and future streamflows

The intense demand for river water in arid regions is resulting in widespread changes in riparian vegetation. We present a direct gradient method to predict the vegetation change resulting from a proposed upstream dam or diversion. Our method begins with the definition of vegetative cover types, based on a census of the existing vegetation in a set of 1 ° 2 m plots. A hydraulic model determines the discharge necessary to inundate each plot. We use the hydrologic record, as defined by a flow duration curve, to determine the inundation duration for each plot. This allows us to position cover types along a gradient of inundation duration. A change in river management results in a new flow duration curve, which is used to redistribute the cover types among the plots. Changes in vegetation are expressed in terms of the area occupied by each cover type. We applied this approach to riparian vegetation of the Black Canyon of the Gunnison National Monument along the Gunnison River in Colorado. We used TWINSPAN to cluster plots according to species occurrence. This analysis defined three vegetative cover types that were distinct in terms of inundation duration. Quantitative changes in the extent of cover types were estimated for three hypothetical flow regimes: two diversion alternatives with different minimum flows and a moving average modification of historical flows. Our results suggest that (1) it is possible to cause substantial changes in riparian vegetation without changing mean annual flow, and (2) riparian vegetation is especially sensitive to changes in minimum and maximum flows. Principal advantages of this method are simplicity and reliance on relatively standard elements of plant community ecology and hydrologic engineering. Limitations include use of a single environmental gradient, restrictive assumptions about changes in channel geometry, representation of vegetation as quasi—equilibrium cover types, and the need for model validation.

Dominance of non-native riparian trees in western USA

Concern about spread of non-native riparian trees in the western USA has led to Congressional proposals to accelerate control efforts. Debate over these proposals is frustrated by limited knowledge of non-native species distribution and abundance. We measured abundance of 44 riparian woody plants at 475 randomly selected stream gaging stations in 17 western states. Our sample indicates that Tamarix ramosissima and Elaeagnus angustifolia are already the third and fourth most frequently occurring woody riparian plants in the region. Although many species of Tamarix have been reported in the region, T. ramosissima (here including T. chinensis and hybrids) is by far the most abundant. The frequency of occurrence of T. ramosissima has a strong positive relation with the mean annual minimum temperature, which is consistent with hypothesized frost sensitivity. In contrast the frequency of occurrence of E. angustifolia decreases with increasing minimum temperatures. Based on mean normalized cover, T. ramosissima and E. angustifolia are the second and fifth most dominant woody riparian species in the western USA. The dominance of T. ramosissima has been suspected for decades; the regional ascendance of E. angustifolia, however, has not previously been reported.

EXTREME FLOODS, CHANNEL CHANGE, AND RIPARIAN FORESTS ALONG EPHEMERAL STREAMS

The geomorphic effectiveness of extreme floods increases with aridity and decreasing watershed size. Therefore, in small dry watersheds extreme floods should control the age structure and spatial distribution of populations of disturbance-dependent riparian trees. We examined the influence of extreme floods on the bottomland morphology and forest of ephemeral streams in a semiarid region. Along six stream reaches on the Colorado Piedmont we examined channel changes by analyzing a rectified sequence of aerial photographs spanning 56 yr, and we investigated the spatial distribution of different-aged patches of forest by aging 189 randomly sampled cottonwood trees. Channel change in these ephemeral sand-bed streams is dominated by widening, which occurs over a span of hours during infrequent floods, and postflood narrowing, which occurs over decades between floods. Narrowing is accelerated where reliable moisture increases the density and growth rate of vegetation on the former bed. Reproduction of cottonwood trees has occurred mostly in former channel bed during periods of channel narrowing beginning after floods in 1935 and 1965 and continuing for as long as two decades. Thus cottonwood establishment is related to low flows at the time scale of a year, but to high flows at the time scale of decades. At sites that have not experienced major floods in the last 80 yr, little channel change has occurred, cottonwood reproduction has been limited, tree density has declined, and succession to grassland is occurring. Because channel change and tree reproduction in this region are driven by infrequent local events, channel width and tree age distributions vary greatly over time and among sites. For the same reason, riparian forests along these ephemeral streams can be as wide as forests along perennial rivers with much higher mean discharge.

The role of vegetation and bed-level fluctuations in the process of channel narrowing

Abstract A catastrophic flood ire 1965 on Plum Creek, a perennial sandbed stream in the western Great Plains, removed most of the bottomland vegetation and transformed the single-thalweg stream into a wider, braided channel. Following eight years of further widening associated with minor high flows, a process of channel narrowing began in 1973; narrowing continues today. The history of channel narrowing was reconstructed by counting the annual rings of 129 trees and shrubs along a 5-km reach of Plum Creek near Louviers, Colorado. Sixty-three of these plants were excavated in order to determine the age and elevation of the germination point. The reconstructed record of channel change was verified from historical aerial photographs, and then compared to sediment stratigraphy and records of discharge and bed elevation from a streamflow gaging station in the study reach. Channel narrowing at Plum Creel: occurs in two ways. First, during periods of high flow, sand and fine gravel are delivered to the channel, temporarily raising the general bed-level. Subsequently, several years of uninterrupted low flows incise a narrower channel. Second, during years of low flow, vegetation becomes established on the subaerial part of the present channel bed. In both cases, surfaces stabilize as a result of vegetation growth and vertical accretion of sediment.

Mortality of riparian box elder from sediment mobilization and extended inundation

To explore how high flows limit the streamward extent of riparian vegetation we quantified the effects of sediment mobilization and extended inundation on box elder (Acer negundo) saplings along the cobble-bed Gunnison River in Black Canyon of the Gunnison National Monument, Colorado, USA. We counted and aged box elders in 144 plots of 37.2 m2, and combined a hydraulic model with the hydrologic record to determine the maximum shear stress and number of growing-season days inundated for each plot in each year of the record. We quantified the effects of the two mortality factors by calculating the extreme values survived during the lifetime of trees sampled in 1994 and by recounting box elders in the plots following a high flow in 1995. Both mortality factors can be modeled as threshold functions; box elders are killed either by inundation for more than 85 days during the growing season or by shear stress that exceeds the critical value for mobilization of the underlying sediment particles. Construction of upstream reservoirs in the 1960s and 1970s reduced the proportion of the canyon bottom annually cleared of box elders by high flows. Furthermore, because the dams decreased the magnitude of high flows more than their duration, flow regulation has decreased the importance of sediment mobilization relative to extended inundation. We use the threshold functions and cross-section data to develop a response surface predicting the proportion of the canyon bottom cleared at any combination of flow magnitude and duration. This response surface allows vegetation removal to be incorporated into quantitative multi-objective water management decisions. Copyright

Restoration of riparian forest using irrigation, artificial disturbance, and natural seedfall

In interior western North America, many riparian forests dominated by cottonwood and willow are failing to reproduce downstream of dams. We tested the hypothesis that establishment is now prevented by absence of the bare, moist substrate formerly provided by floods and channel movement. Along Boulder Creek, a dammed stream in the Colorado plains, we tested the effects of disturbance (sod removal), irrigation, and addition of seed on the establishment of seedings of plains cottonwood (Populus deltoides subsp.monilifera) and peachleaf willow (Salix amygdaloides). In unirrigated, undisturbed plots, mean cottonwood density was 0.03 seedlings/m2. Irrigation or disturbance alone produced mean cottonwood densities of 0.39 and 0.75 seedlings/m2. Plots that were both irrigated and disturbed produced a mean cottonwood density of 10.3 seedlings/m2. The effects of irrigation and disturbance on cottonwood establishment were significant (P<0.005); added seed had no significant effect (P=0.78). The few cottonwood seedlings in unirrigated plots were in low positions susceptible to scour by future moderate flows. We conclude that cottonwood establishment along Boulder Creek is limited by the scarcity of bare, moist sites safe from future scour. Establishment of peachleaf willow was significantly affected only by disturbance; daily sprinkler irrigation did not provide sufficient moisture to increase survival of this species. Our results demonstrate the feasibility of restoring plains cottonwood forests using natural seedfall, even where only widely scattered adult trees are present. Because use of natural seedfall conserves the genetic makeup of the local population, this method may be preferable to the use of imported cuttings.

Latitudinal variation in cold hardiness in introduced Tamarix and native Populus

To investigate the evolution of clinal variation in an invasive plant, we compared cold hardiness in the introduced saltcedar (Tamarix ramosissima, Tamarix chinensis, and hybrids) and the native plains cottonwood (Populus deltoides subsp. monilifera). In a shadehouse in Colorado (41°N), we grew plants collected along a latitudinal gradient in the central United States (29–48°N). On 17 occasions between September 2005 and June 2006, we determined killing temperatures using freeze‐induced electrolyte leakage and direct observation. In midwinter, cottonwood survived cooling to −70°C, while saltcedar was killed at −33 to −47°C. Frost sensitivity, therefore, may limit northward expansion of saltcedar in North America. Both species demonstrated inherited latitudinal variation in cold hardiness. For example, from September through January killing temperatures for saltcedar from 29.18°N were 5–21°C higher than those for saltcedar from 47.60°N, and on September 26 and October 11, killing temperatures for cottonwood from 33.06°N were >43°C higher than those for cottonwood from 47.60°N. Analysis of nine microsatellite loci showed that southern saltcedars are more closely related to T. chinensis while northern plants are more closely related to T. ramosissima. Hybridization may have introduced the genetic variability necessary for rapid evolution of the cline in saltcedar cold hardiness.

USE OF INDIVIDUALISTIC STREAMFLOW-VEGETATION RELATIONS ALONG THE FREMONT RIVER, UTAH, USA TO ASSESS IMPACTS OF FLOW ALTERATION ON WETLAND AND RIPARIAN AREAS

We analyzed the transverse pattern of vegetation along a reach of the Fremont River in Capitol Reef National Park, Utah, USA using models that support both delineation of wetland extent and projection of the changes in wetland area resulting from upstream hydrologic alteration. We linked stage-discharge relations developed by a hydraulic model to a flow-duration curve derived from the flow history in order to calculate the inundation duration of 361 plots (0.5 × 2 m). Logistic regression was used to relate plant species occurrence in plots to inundation duration. A weighted average of the wetland indicator values of species was used to characterize plots as Aquatic, Wetland, Transitional, or Upland. Finally, we assessed how alterations in the flow duration curve would change the relative widths of these four zones. The wetland indicator values of species and the wetland prevalence index scores of plots were strongly correlated with inundation duration. Our results support the concept that plants classified as wetland species typically occur gradient study reach of the Fremont River that satisfied the vegetation criterion for a regulatory wetland was narrow (2 m wide). Both the unvegetated Aquatic zone (7.8 m) and the Transitional zone (8 m) were substantially wider. The Transitional zone included the maxima of several species and was, therefore, not merely a combination of elements of the Wetland and Upland zones. Multiplicative increases or decreases in streamflow regime produced a wetter, or drier, bottomland vegetation, respectively. Systematic reductions in flow variability reduced the width of both the Wetland and Transitional zones and increased the width of the Upland zone. Our approach is widely applicable to inform water management decisions involving changes in flow regime.