Instream flows and the decline of riparian cottonwoods along the Yakima River,Washington, USA

Under pre‐settlement conditions the Yakima River in Washington state, USA was characterized by multiple channels, complex aquifers and extensive riparian cottonwood forests. Subsequent implementation of headwater dams to supply irrigation water has altered river and floodplain processes critical to the cottonwoods and associated riparian vegetation. In this study, we analysed hydrology and floodplain forests and especially the dominant black cottonwoods (Populus trichocarpa) along sequential reaches of the Yakima River. Elevations were surveyed and vegetation inventoried along cross‐sectional belt transects, and cottonwood tree ring interpretations investigated historic associations between river hydrology and cottonwood establishment and growth. We analysed hydrographs relative to the apparent episodes of cottonwood recruitment and applied a quantitative model for seedling colonization that required: (1) floods, disturbance flows to produce barren nursery sites, and subsequent flows for seedling (2) establishment and (3) survival. In contrast to earlier conditions, flow patterns after the 1960s have generally been unfavourable for cottonwood recruitment although some cottonwood colonization has occurred in association with physical disturbance from gravel mining. With recent flow regimes, regulated flows along upper reaches maintain the river near bank‐full throughout the growing season, thus inundating suitable seedling recruitment sites. Downstream, irrigation withdrawals reduce the river stage, resulting in seedling establishment at low elevations that are lethally scoured by subsequent high flows. These regulated flow regimes have not hindered growth of established trees, but have reduced the recruitment of cottonwoods, and particularly disfavoured females, thus altering sex ratios and producing skewed cottonwood population age and gender structures. The cottonwood decline has also been associated with other changes in riparian plant community composition, including the encroachment of invasive weeds. Based on this ecohydrologic analysis we discuss flow adjustments that could rejuvenate cottonwood forests along the Yakima River. Copyright © 2007 John Wiley & Sons, Ltd.     

Temporal and Spatial Variation in Riparian Vegetation and Floodplain Aquifers on the Regulated Dolores River,Southwest Colorado,USA

The importance of flow variability and floodplain water table recharge for the establishment and long‐term survival of riparian vegetation has been well‐documented. However, temporal and spatial variation in floodplain aquifers has received less attention, although native species can have narrow tolerances for groundwater decline. Our observations of decreased cottonwood cover on floodplains and increased willow cover on river banks since dam completion on the Dolores River led to comparisons between three long‐term study sites above and below McPhee Dam. We summarize 5 years (2010–2014) of shallow groundwater well data from transects of three wells per site. Vegetation cover data were collected from quadrats and line‐intercept transects. In the willow zone, groundwater well levels mirror in‐channel flows and rarely drop below 0.6 m from ground surface. Willow cover and stem counts on point bars are higher at dammed sites. Wells in the cottonwood zone indicate that alluvial recharge happens only during prolonged peak discharge during spring snowmelt or dam release. Years with no dam spill reduced connectivity between surface flows and groundwater, and groundwater depth dropped to between 2 and >2.5 m. Long‐term data below the dam indicate that canopy cover of the dominant cottonwoods has declined over time (48% in 1995, 19% in 2003), especially in the wake of severe drought. Mature cottonwood cover is significantly higher at the undammed site (p = 0.025). Our results indicate that floodplain habitats below dams exist under artificially extreme drought and inform how biologically diverse riparian systems will be impacted by a drying climate. Copyright © 2016 John Wiley & Sons, Ltd.     

Factors controlling the establishment of Fremont cottonwood seedlings on the Upper Green River,USA

Declines in cottonwood (Populus spp.) recruitment along alluvial reaches of large rivers in arid regions of the western United States have been attributed to modified flow regimes, lack of suitable substrate, insufficient seed rain, and increased interspecific competition. We evaluated whether and how these factors were operating during 1993–1996 to influence demographics of Fremont cottonwood (P. deltoides Marshall subsp. wislizenii (Watson) Eckenwalder) along reaches of the Green and Yampa Rivers near their confluence in northwestern Colorado. We examined seedling establishment, defined as survival through three growing seasons, at three alluvial reaches that differed primarily in the level of flow regulation: a site on the unregulated Yampa, an upper Green River site regulated by Flaming Gorge Dam, and a lower Green River site below the Green–Yampa confluence. Seed rain was abundant in all sites, and led to large numbers of germinants (first‐year seedlings) appearing each year at all sites. The regulated flow in the upper Green River reach restricted germination to islands and cut banks that were later inundated or eroded; no seedlings survived there. Mortality at the lower Green River site was due largely to desiccation or substrate erosion; 23% of 1993 germinants survived their first growing season, but at most 2% survived through their second. At the Yampa River site, germinants appeared on vegetated and unvegetated surfaces up to 2.5 m above base flow stage, but survived to autumn only on bare surfaces at least 1.25 m above base flow stage, and where at least 10 of the upper 40 cm of the alluvium was fine‐textured. Our studies of rooting depths and the stable isotopic composition of xylem water showed that seedlings in the most favorable locations for establishment at the Yampa site do not become phreatophytic until their third or fourth growing season. Further, the results of experimental field studies examining effects of shade and competition supported the hypothesis that insufficient soil moisture, possibly in combination with insufficient light, restricts establishment to unvegetated sites. Collectively, the demographic and experimental studies suggest that, in arid regions, soil water availability is at least as important as light level in limiting establishment of Fremont cottonwood seedlings. We hypothesize that in cases where arid land rivers experience large spring stage changes, recruitment is further constrained within bare areas to those sites that contain sufficient fine‐textured alluvium, saturated during the spring flood, to provide the flood‐derived soil moisture normally necessary for late‐summer seedling survival. Copyright © 1999 John Wiley & Sons, Ltd.     

Bank stabilization,riparian land use and the distribution of large woody debris in a regulated reach of the upper Missouri River,North Dakota,USA

Large woody debris (LWD) is an important component of ecosystem structure and function in large floodplain rivers. We examined associations between LWD distribution and riparian land use, bank stabilization (e.g. riprap revetment), local channel geomorphology, and distance downriver from the dam in the Garrison Reach, a regulated reach of the upper Missouri River in North Dakota, USA. We conducted a survey of shoreline‐associated LWD in the reach during typical summer flow conditions. Reach‐wide LWD density was 21.3 pieces km^?1 of shoreline, of which most pieces (39% ) were ‘beached’ between the waterline and the bankfull level, 31% of pieces had evidence of originating at their current location (anchored), 18% of pieces were in deep water (>1 m), and 13% were in shallow water. LWD density along unstabilized alluvial (sand/silt) shorelines (27.3 pieces km^?1) was much higher than along stabilized shorelines (7.2 pieces km^?1). LWD density along forested shorelines (40.1 pieces km^?1) was higher than along open (e.g. rangeland, crop land; 9.2 pieces km^?1) or developed (e.g. residential, industrial; 7.8 pieces km^?1) shorelines. LWD density was highest overall along unstabilized, forested shorelines (45 pieces km^?1) and lowest along open or developed shorelines stabilized with a blanket‐rock revetment (5.5 pieces km^?1). Bank stabilization nearly eliminated the positive effect of riparian forest on LWD density. A predicted longitudinal increase in LWD density with distance from the dam was detected only for deep LWD (including snags) along unstabilized alluvial shorelines. Partial resurvey in the summer following the initial survey revealed a reduction in total LWD density in the reach that we attribute to an increase in summer flow between years. Changes in riparian management and land use could slow the loss of LWD‐related ecosystem services. However, restoration of a natural LWD regime in the Missouri River would require naturalization of the hydrograph and modification of existing bank stabilization and channel engineering structures. Copyright © 2004 John Wiley & Sons, Ltd.     

The influence of river regulation and land use on floodplain forest regeneration in the semi‐arid upper Colorado River Basin,USA

Flow regulation effects on floodplain forests in the semi‐arid western United States are moderately well understood, whereas effects associated with changes in floodplain land use are poorly documented. We mapped land cover patterns from recent aerial photos and applied a classification scheme to mainstem alluvial floodplains in 10 subjectively selected 4th order hydrologic units (subbasins) in the Upper Colorado River Basin (UCRB) in order to document land use patterns (floodplain development) and assess their effects on Fremont cottonwood forest (CF) regeneration. Three of the mainstem rivers were unregulated, five were moderately regulated and two were highly regulated. We classified polygons as Undeveloped (with two categories, including CF) and Developed (with five categories). We ground‐truthed 501 randomly selected polygons (4–28% of the floodplain area in each subbasin) to verify classification accuracy and to search for cottonwood regeneration, defined as stands established since regulation began or 1950, whichever is most recent. From 40% to 95% of the floodplain area remained undeveloped, but only 19–70% of the floodplain area was classified as forest. Regeneration occupied a mean of 5% (range 1–17%) of the floodplain. The likelihood of the presence of regeneration in a polygon was reduced 65% by development and independently in a complex manner by flow regulation. Our analyses indicate that floodplain forests may be in jeopardy on both regulated and unregulated rivers and that information on historical forest extent is needed to better understand their current status in the UCRB. Conservation efforts need to be coordinated at a regional level and address the potentially adverse affects of both flow regulation and floodplain development. Published in 2007 by John Wiley & Sons, Ltd.