Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers

G Model

ARTICLE IN PRESS

ECOENG-4557; No. of Pages 9

Ecological Engineering xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers Edward P. Glenn a,∗ , Pamela L. Nagler b , Patrick B. Shafroth c , Christopher J. Jarchow b a b c

Dept. of Soil, Water and Environmental Science, University of Arizona, Tucson, AZ 85721, United States U.S. Geological Survey, Sonoran Desert Research Station, University of Arizona, 85721, United States U.S. Geological Survey, Fort Collins Science Center, Fort Collins, CO 80526, United States

a r t i c l e

i n f o

Article history: Received 7 June 2016 Received in revised form 9 January 2017 Accepted 10 January 2017 Available online xxx Keywords: Environmental flows Riparian restoration Tarim River Colorado River Tamarix Populus Salix

a b s t r a c t Environmental flows have become important restoration tools on regulated rivers. However, environmental flows are often constrained by other demands within the river system and thus typically are comprised of smaller water volumes than the natural flows they are meant to replace, which can limit their functional efficacy. We review environmental flow programs aimed at restoring riparian vegetation on four arid zone rivers: the Tarim River in China; the Bill Williams River in Arizona, U.S.; the delta of the Colorado River in Mexico; and the Murrumbidgee River in southern Australia. Our goal is to determine what worked and what did not work to accomplish restoration goals. The lower Tarim River in China formerly formed a “green corridor” across the Taklamakan Desert. The riparian zone deteriorated due to diversion of surface and groundwater for irrigated agriculture. A massive restoration program began in 2000 with release of 1038 million cubic meters of water over the first three years. Groundwater levels rose but the ecological response was less than expected politically, socially and within the scientific community. However, releases continued and by 2015 portions of the original iconic Populus euphratica (Euphrates poplar) forest were reestablished. The natural flow regime of the Bill Williams River was disrupted by construction of a dam in 1968, dramatically reducing peak flows along with associated fluvial processes. As a result, the channel narrowed and riparian vegetation expanded and was comprised largely of an introduced shrub species (Tamarix spp.). Environmental flow releases including small, managed floods and sustained base flows have been implemented since the mid 1990’s to promote establishment and maintenance of native riparian trees (cottonwoods and willows) and have been successful, although in a “downsized” portion of the valley bottom. Experience from the Bill Williams was used to help design the Minute 319 environmental flow in the delta of the Colorado River in 2014. Water was released as a short, one-time pulse during spring with the intent of starting new cohorts of cottonwood and willow. However, fluvial disturbance was limited by the relatively small magnitude pulse, low flows did not continue throughout the growing season in some reaches, native tree recruitment was low, and most of the new plants recruited were Tamarix. The inundated portion of the floodplain did respond with a temporary increase in greenness as measured by satellite vegetation indices, however. The Murrumbidgee River in Australia is a tributary in the Murray-Darling River Basin, which supports iconic red gum (Eucalyptus camaldulensis) forests that depend on near-yearly floods for maintenance. During the recent Millennial Drought (2000–2010) environmental flows were provided on an experimental basis to small portions of the Yanga National Forest to see how much water was needed. As with the Colorado River delta, gains in vegetation vigor as measured by satellite vegetation indices following the flows were temporary. Environmental flows in the Bill Williams were able to restore enough overbank flooding and fluvial disturbance to promote some establishment of new cohorts of trees, but on the Colorado and Murrumbidgee Rivers larger volumes of total flows released over longer periods and targeted restoration will be needed to restore the ecosystems. A measure of

∗ Corresponding author. E-mail addresses: [email protected], [email protected] (E.P. Glenn), [email protected] (P.L. Nagler). http://dx.doi.org/10.1016/j.ecoleng.2017.01.009 0925-8574/© 2017 Elsevier B.V. All rights reserved.

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

2

success in restoring the Euphrates poplar forest on the Tarim and germinating new chorts of willows on the Bill Williams has been achieved after 15–20 years of environmental flows, but the Colorado River delta and Murrumbidgee Rivers have only received one or two flows. Success in enhancing native trees in the Colorado delta has been achieved in restoration plots, but the Murrumbidgee will require large overbank flows on a continuing schedule to rejuvenate the red gum forest. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Managing flow releases from dams for downstream environmental benefits (“environmental flows”) was historically directed at maintaining fish habitat (Orth and Maughan, 1982), but since the 1990s environmental flows have been used as restoration tools for a variety of purposes (Nilsson and Berggren, 2000; Arthington, 2012; Konrad et al., 2011, 2012). Environmental flows sometimes target riparian vegetation, including objectives such as: to promote germination and establishment of native trees (Rood et al., 2003, 2007); to recharge the alluvial aquifer to support existing, desirable vegetation (Stromberg et al., 2007); to wash salts from the riverbank to favor mesic native trees over salt-tolerant invasives (Gross, 2003); and to scour the riverbed to remove invasive species and create new bars for establishment of native trees (Richardson et al., 2007; Wilcox and Shafroth 2013). Environmental flows are typically designed to restore elements of the unregulated (natural) flow regime (Merritt et al., 2010; Poff et al., 1997), but planning and implementing such flows can be challenging (Olden et al., 2014; Konrad et al., 2011, 2012). One complication is that the amount of water allocated for environmental flows is often much less than was provided by natural flow events. Additionally, natural ecosystem dynamics are often a product of complex sequences of flows. For example, riparian ecosystems are typically not climax communities but are dynamic mosaics of habitats in different states of succession (Standford et al., 2005), and attempting to replicate the hydrological factors that lead to desirable restoration outcomes can be difficult. Therefore, success is not guaranteed, and strategies for environmental flows should be based on determining suitable magnitude, frequency, duration, timing, and rates of change of flows to meet specific objectives at specific sites (Richter et al., 2003). Several papers have reviewed the effectiveness of large-scale environmental flows in a variety of biomes and results have been mixed (e.g., Olden et al., 2014; Davies et al., 2014; Konrad et al., 2011, 2012). Therefore, each flow event should be evaluated for successes, failures and lessons learned. Olden et al. (2014) recommended that environmental flows should be analyzed by asking three questions: why was the flow experiment conducted?; what knowledge was gained from the flow and its outcomes?; and what challenges remain in meeting restoration goals? Shafroth et al. (2010) advocated combining existing hydrologic and ecologic information and models with empirical relationships determined during flow events to produce predictive hydrology-ecology models, with particular attention paid to determining threshold levels of volume and duration of flows needed to produce desired outcomes. Davies et al. (2014) and Konrad et al. (2011, 2012) recommended treating each flow event as a manipulative experiment, and advocated using the term “large-scale flow experiments” rather than environmental flows. This study reviews four cases in which environmental flows have been used to attempt restoration of riparian vegetation on four dryland rivers: the Tarim River in China (Hou et al., 2007); the Bill Williams River in Arizona (Shafroth et al., 1998, 2010; Wilcox and Shafroth 2013); the delta of the Colorado River in the U.S.

and Mexico (Jarchow et al., 2016; Pitt and Kendy, 2016; Kendy et al., 2017); and the Murrumbidgee in New South Wales, Australia (Baldwin et al., 2013; Doody et al., 2015; Nagler et al., 2016). A discussion of the successes, failures and lessons learned is presented. A major theme running through the paper is the need to balance environmental versus societal demans as a constraint on many environmental flow programs, thereby limiting their ecological effectiveness.

2. Tarim River, China Prior to development of large-scale irrigation projects starting in the 1950s, the lower Tarim River formed a 320 km “green corridor” of riparian and wetland vegetation as much as 5–10 km wide between the now-existing Daxihaizi Reservoir and Taitemar Lake in the Taklimakan Desert in northeastern China (Yongbo and Chen, 2007; UNESCO, 2010). Water in the river arises from snow melt in mountains outside the basin. The river supports iconic Populus euphratica (Euphrates poplar) forests. These trees have been called “living fossils”, having originated 60 million years ago with the uplifting of the Tibetan Plateau. Individual trees can live as long as 1000 years (UNESCO, 2010). By 1972 flows were reduced to only 20% of former volumes due to upstream diversions for agriculture; portions of the river were seasonally dry; and the water table dropped below the tolerance limit of the P. euphratica trees that formed the base of the ecosystem (Chen et al., 2003; Yongo and Chen, 2007). Taitemar Lake was dry. By the 1990s the water table had dropped from 3 to 5 m to 8–12 m deep and 47% of the remaining P. euphratica trees were dead (Chen et al., 2003). The forested area had decreased from 5.4 × 104 ha to 0.5 104 ha over 50 years and was in danger of extinction (Ling et al., 2015). In 2000 the Chinese government undertook a major effort to restore the ecosystem through the Ecological Water Development Project (Bao et al., 2017; Yongbo and Chen, 2007; Zhu et al., 2016). The Ministry of Water Resources allocated the equivalent of $2 billion US dollars for a massive restoration effort to raise the water table to support the existing P. euphratica trees and create a hydrologic regime to recruit new trees to the floodplain. Over the first three years several releases of water were made from upstream reservoirs, totaling 1038 mcm of water. The water table responded positively and water once again flowed into Taitemar Lake (Yongbo and Chen, 2007; Chen et al., 2009) but the rejuvenation of the P. euphratica forest that was desired was slow to develop, and at first the effort did not meet political and social expectations (Hou et al., 2007). However, releases have continued, averaging about 200 mcm yr−1 (Yan et al., 2014; Chen et al., 2014; Aishan et al., 2015; Zhu et al., 2016). A restored forest has developed but it is much narrower than the original forested area. The floodplain forest has made an excellent recovery within 200 m of the riverbed with medium results extending up to 800 m from the riverbed (Aishan et al., 2015). Importantly, survival and recruitment of new cohorts of P. euphratica have been documented (Aishan et al., 2015). The restoration effort is now regarded as successful, although recovery of the riparian zone has been slower and over a narrower width than originally

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

anticipated (Zhang et al., 2013). Bao et al. (2017), using satellite imagery, showed that fractional cover of natural vegetation in the riparian corridor increased by a factor of 1.5 between 2001 and 2013. Increases in cover occurred for forests, grasslands and scrublands. Depth to water has decreased from a mean of 12.6 m in 2000 to 5.5–6.2 m in 2015 (Zhu et al., 2016). A remaining problem is the low recruitment of new trees on the floodplain (Zhu et al., 2016). Three factors are responsible: (1) in dry years the flows are not sufficient to keep the roots of juvenile trees within reach of the water table; (2) upstream irrigation projects discharge agricultural return flows into the river, making it saline; and (3), regulatory dikes and other flood control structures reduce overbank flooding, which is needed to expand the forested area beyond the main channel. Other challenges include the disruption of the soil biological crust by grazing animals and tourism, and the possibility of reduced flows due to climate change (Zhu et al., 2016). Zhu et al. (2016) recommend that active restoration efforts such as planting trees should be undertaken.

3. Bill Williams River, Arizona The Bill Williams River is a 58 km long tributary of the Colorado River in western Arizona. It is formed at the junction of two free-flowing rivers, the Santa Maria and Big Sandy. The riparian corridor receives little human use and is valued primarily as a natural area, still supporting riparian plants and animals that have become scarce on the highly regulated Lower Colorado River. In 1968 a dam was constructed on the Bill Williams, creating Alamo Lake, and changing the river from free-flowing to regulated. While mean annual river flows have actually increased since the dam was built due to increased precipitation in some years, peak flows are much diminished and their seasonal timing has changed (House et al., 2006). In the pre-dam era, floods with a 10-year recurrence interval were nine times the magnitude of the maximum controlled release from Alamo Dam (∼200 m3 /s), indicating that flooding to create sites favorable for germination and establishment of new trees has decreased dramatically (Fig. 1) (Wilcox and Shafroth, 2013). There is evidence that the altered flow regime has favored non-native Tamarix shrubs over the establishment of native cottonwood, willow and mesquite trees (Shafroth et al., 2002). In recent decades, the delta of the river where it enters Lake Havasu has supported a gallery forest of mature cottonwoods with low recruitment of new replacement cohorts. In the early 1990’s, largely in response to these changes to the riparian forests, the Army Corps of Engineers (which operates Alamo Dam) in collaboration with land and water managers along the river, began to evaluate ways to manage flow releases differently, particularly to promote the expansion of native riparian trees. By 2003, a formal change to Alamo Dam’s operating rules was in place that specified “fish and wildlife restoration” as a purpose of the dam (USACE, 2003), and in 2004, the Bill Williams River became one of eight U.S. rivers that are part of the Sustainable Rivers Project, a collaboration between The Nature Conservancy and the U.S. Army Corps of Engineers that emphasizes environmental flow implementation (Konrad et al., 2012). A particularly wet winter in 2004–2005 and another in 2010 provided large inflows to Alamo Lake and set the stage for multiple years of experimental flow releases on the Bill Williams River (Shafroth et al., 2010). Although not formalized prior to this, releases from Alamo Dam had already been managed in some respects to promote native tree recruitment and survival in years following large inflows in 1993 and 1995 (Shafroth et al., 1998). In years with exceptionally high runoff upstream of Alamo Dam (e.g., 1993 and 2004–2005), high flow releases were largely designed to keep Alamo Lake from filling too rapidly and were not tightly con-

3

trolled. However, the maximum discharge was released with the intention of producing bare, moist sites conducive to germination of native cottonwood and willow seeds. In 2006, 2007, 2008, and 2010, smaller magnitude peaks were released, again to promote establishment of cottonwood and but also to examine differential mortality of tamarisk vs. cottonwood and willow seedlings associated with flooding (Wilcox and Shafroth, 2013). The flow recession was controlled following the peak in 2005, 2006, and 2010 to allow newly germinated seedlings to maintain access to soil moisture (Mahoney and Rood, 1998). In all years, as part of Alamo Dam operations, base flows were released at high enough rates to continue to support relatively shallow water tables along the river corridor, which promotes survival and growth of the riparian vegetation. The effects of experimental flows on riparian seedling recruitment have been studied from 1993 to the present, including measurements of vegetation response, sediment transport and flood dynamics in upstream and downstream sites before, during and after releases (Shafroth et al., 1998, 2002, 2010; Wilcox and Shafroth, 2013). Results from the early 1990’s revealed significant seedling establishment following high flow releases in 1993 and 1995, and indicated that successful seedling establishment was predictably associated with aspects of system hydrology delivered by the environmental flow releases (Shafroth et al., 1998). Significant new establishment was also observed following high flows and controlled recessions in 2005 and 2006 (Wilcox and Shafroth, 2013; Shafroth, unpublished data). Relatively little new establishment was observed following releases in 2010 (Shafroth, unpublished data). In 2006, woody plants in study plots were mainly seedlings established by the 2004–2005 floods. Ninety-percent of the seedlings were Tamarix and 9% were Gooding’s willow, while herbaceous plants accounted for 6–16% of vegetation cover. Hence, at the start of the study Tamarix greatly outnumbered willow seedlings. The fate of these seedlings was quantified following both 2006 and 2007 releases (Shafroth et al., 2010; Wilcox and Shafroth, 2013). Both floods caused substantial seedling mortality, but mortality was much higher for Tamarix than willow seedlings. 85% of Tamarix seedlings died at both test sites, while willow mortality were 64% and 26% at the upstream and downstream study site, respectively. Plants greater than 50–70 cm in height were resistant to removal by flood waters. Cottonwoods and willows tend to outcompete saltcedar in the establishment stage (Sher et al., 2000, 2002) By 2007, willow seedlings were much taller than Tamarix and had grown sufficiently above and below-ground to stabilize the sand bars on which they germinated. By the time large flows were released in 2010, willows not only dominated the sand bars but were many meters tall and, along with cattails (Typha domingensis), were able to stabilize the bars despite higher flows. The study concluded that even relatively small environmental flows can tip the balance towards willow over Tamarix under certain circumstances. However, success was only achieved on a down-sized portion of the valley bottom, a pattern that has also been observed at the scale of the entire river (Kui et al., in review) and on other regulated rivers such as the Tarim and Colorado River delta (Hall et al., 2011).

4. The Colorado River delta minute 319 flows Environmental flow deliveries to the Colorado River delta in 2014 were referred to as the “Minute 319 Pulse Flow” and were negotiated as an amendement (“Minute”) to the U.S./Mexico Water Treaty. The agreement was to provide a one-time release of approximately 130 million cubic meters (mcm) of water from the U.S. to Mexico via the Colorado River,followed by smaller annual base flows (IBWC 2000, 2012; Glenn et al., 2013; Pitt and Kendy, 2016). The volume of water was not negotiated based on ecosystem needs

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9 4

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

Fig. 1. Left: Flow on the Bill Williams River before and after construction of Alamo Dam; Right: managed and natural flow events, 2005–2010. From Wilcox and Shafroth (2013).

Fig. 2. History of Colorado River flows at the International Boundary Commission’s gage at the Southerly International Boundary between the U.S. and Mexico, representing water that flows below the last diversion point for water at Morelos Dam and enters the Colorado River delta. Data from IBWC, 2016. See also Mueller et al., 2017 for further information on historical flows.

but on what was available in storage above anticipated human water demands in the release year. However, the timing and hydrograph were based on the success achieved on the Bill Williams River. The objectives were broadly similar: rework the alluvium by an initial high flow rate sustained for several days to provide new areas for potential establishment of native trees; then provide a recessional limb of lower flow rates to expose sediment and promote germination and establishment of native seeds according to the “recruitment box” model for cottonwoods (Mahoney and Rood, 1998; Rood et al., 2007). The water was released from March to May, 2014, into the often-dry Colorado River channel in the delta region of the river in Mexico. The timing was meant to coincide with the main period of seed release by cottonwoods and willows (Pitt and Kendy, 2016). The pulse flow was supplemented with much smaller base flows, intended to be used in tree-planting initiatives in areas selected for active restoration. An intensive hydrological and ecological monitoring program was conducted before, during and after the release. For monitoring purposes the riparian zone was divided into seven sections or reaches, contained between flood control levees. The recent history flood flows into the delta was used to justify the Minute 319 Pulse Flow (Mueller et al., 2017). The release followed several decades of observations on the impacts on the riparian zone and estuary of a series of so-called “waste spills” of water from the U.S. to Mexico (Fig. 2) (Glenn et al., 1996, 2001; Cohen et al., 2001; Hinojosa-Huerta et al., 2002; Hinojosa-Huerta et al., 2013; Nagler et al., 2005, 2009; Mueller et al., 2017). Mueller et al. (2017) have documented the geomorphic and sediment trans-

port changes that have occurred during previous floods as well as the impacts of the Minute 319 Flow. From 1950–1963 the gage at the Southerly International Boundary between the U.S. and Mexico (below Morelos Dam, the last diversion point for water from the river) recorded peak flows in the range of 200–600 m3 s−1 every 2–3 years (IBWC, 2016). Mean annual flows were 87 m3 s−1 , or 2744 mcm yr−1 . These flows originated as waste spills from Lake Mead behind Hoover Dam. They resulted from the inherent variability in the annual snowpack in the Rocky Mountains due to El Nino/Soiuthern Oscillation cycles and other meteorological factors. In years when Lake Mead was at or near its holding capacity, excess water from snow melt was released to flow to the sea. Water releases to the delta entered a new phase after 1963. In 1964 Glen Canyon Dam upstream from Hoover Dam was completed, creating Lake Powell, the second largest reservoir on the river. From 1964–1979 waste spills to Mexico averaged only 6.4 m3 s−1 (202 mcm yr−1 ) because excess water in wet years contributed to the filling of Lake Powell. The next phase began in 1980 when Lake Powell reached full volume. Waste spills to the delta increased dramatically, averaging 105 m3 s−1 (3311 mcm yr−1 ) from 1980 to 2000. Specific events during this period included major ENSO cycles in 1983 and 1997 that led to multi-year releases from Lake Mead to Mexico, and a 1993 ENSO cycle that produced flooding on the Gila River, a tributary stream that enters the Colorado River at Yuma, Arizona. Ground and remote sensing studies from 1990 to 2000 documented the impact of the floods on the delta ecosystems (Nagler et al., 2005, 2009). Native cottonwood and willow trees germinated following the floods and by 1999 represented 10% of vegetation cover, compared to less than 1% along the Lower Colorado River in the U.S. Trees could be traced to specific flow events through treering dating (Nagler et al., 2005), and by 2000 ranged from large trees started in the 1980s to seedlings and juveniles germinated by more recent floods. Avian diversity and abundance in the delta were as much as ten times higher than on the U.S. portion of the river below Lake Mead (Hinojosa-Huerta, 2006). The commercial shrimp catch in the northern Gulf of California was positively correlated with river flows during the 1980s and 1990s (Galindo-Bect et al., 2000), and the intertidal portion of the river had reduced salinities that favored survival of larval and juvenile fish and crutaceans during flow events (Glenn et al., 2007). Since 2000, the watershed has been in drought and dams are below capacity, making waste spills rare events. This set the stage for development of the Minute 319 environmental flows. During this time, flows past Morelos Dam averaged only 1.1 m3 s−1 (34.7 mcm yr−1 ) (IBWC, 2016). The riparian vegetation underwent a slow deterioration, with native trees decreasing from 10% of vegetation cover to <2% by 2008 (Hinojosa-Huerta et al., 2013). Vegetation greenness and evapotranspiration (ET), as determined by satellite imagery, also decreased. Surprisingly, however, total

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

riparian ET averaged three times greater than surface flows over this period (Jarchow et al., 2016). Piezometer data showed that the riparian corridor supported a low-salinity, shallow aquifer, apparently from underflows of water from the U.S., into which riparian phreatophytes were rooted (Nagler et al., 2005). This led to the hypothesis that riparian ET is mainly supported by groundwater underflows from the U.S. to Mexico and local agricultural return flows, and that surface flows were needed in the riparian zone to germinate new cohorts of trees lost to fire, cutting for wood and senescence. This led ultimately to the Minute 319 release of water to the delta in March to May, 2014, with the goals of rejuvenating the population of native trees and recharging the riparian aquifer. However, many factors other than a desire to restore the riparian zone entered into the agreement. (For a history of negotiations and circumstances leading up to the pulse release see Pitzer, 2014). The March to May period was selected for the release to coincide with seed release of cottonwood and willow seeds, and to exclude as much as possible saltcedar (Tamarix ramosissima) seeds, which tend to be released later, based on data from the Bill Williams River (Shafroth et al., 1998). The Minute 319 pulse event was quite small in comparison to earlier releases, both in volume and duration (Fig. 2). The peak flow rate of 120 m3 s−1 was at the low end of the range of natural flows and was maintained at this rate for only 3 days. Furthermore, much of the water infiltrated to groundwater in Reaches 2 and 3, where the soil is sandy and the water table is deeper than in other reaches. The degree of scour and sediment transport, needed for establishment of new trees, was much smaller than previous floods (Mueller et al., 2017. Nevertheless, riparian vegetation in the zone of inundation showed a marked greenup in 2014, with NDVI values as much as 200% higher in August, 2014, compared to August, 2013 (Jarchow et al., 2016). The greenup extended through all reaches, and extended beyond the zone of inundation, suggesting that the groundwater system was involved in the greenup. Over the whole riparian corridor NDVI and ET estimated from NDVI on Landsat imagery increased 16% over 2013 baseline levels. However, by 2015 most of the increase in NDVI had faded back to 2013 levels (Jarchow et al., 2016). Germination and establishment of new cohorts of cottonwood and willows was scant, and most of the newly recruited plants were tamarisk (Shafroth et al., 2017). However, in the active restoration areas germination and growth of native tree seeds was more successful. The overall conclusion was that given the limited amount of water likely to be available for future flows, the emphasis should be on using the water for active restoration of trees at sites with a high water table (Reaches 1 and 4) (Schlatter et al., 2017), and extending the restoration efforts into downstream areas (Reaches 5 and 7) where the greenup after the pulse flow had the greatest extent (Kendy et al., 2017). Negotiations are currently underway to determine if, when and how much a repeat flow event will occur. An important finding was that only about 5% to 10% of the 130 mcm of the pulse flow could be accounted for as enhanced ET in 2014 and 2015. The volume of the pulse flow was of the same magnitude as the annual ET (120–200 mcm yr−1 ), but most of the ET requirement was met by underflows of groundwater from the U.S. and agricultural return flows in Mexico, as also occurred in the years before the pulse flow (Jarchow et al., 2016). An obvious question, therefore, was where did the bulk of the water go (Kendy et al., 2017)? Some of the pulse flow likely entered the Gulf of California as surface flows (Daessle et al., 2016) and groundwater underflows into the subterranean estuary, as the major part of the greenup in 2014 occurred in Reach 7, and most of this greenup was outside the zone of inundation, indicating that the groundwater system above the intertidal zone could have been recharged by the pulse (Jarchow et al., 2016). Some of the most important fisheries in the marine zone (e.g., shrimp, totoaba, corvina) require brackish water for lar-

5

val and juvenile stages (Calderon-Aguilera and Flessa, 2009). An alternative possibility is that the pulse water entered the regional aquifer, most of which is recovered for irrigation in the San Luis and Mexicali irrigation districts. The pulse flow demonstrated the need to test the conceptual model of the hydrology in the riparian corridor and its interaction with the regional aquifer in its flow towards the sea (Kendy et al., 2017). If the primary goal is to improve habitat, then releases of lower volume and longer duration would perhaps be more efficient than a brief, high-volume pulse. However, longer duration flows would allow more time for evaporation to take place. Instead of relying on natural seed release in combination with environmental flows to start new cohorts, future water releases might be better directed to more intensively managed restoration areas (Kendy et al., in review; Schlatter et al., 2017).

5. Murrumbidgee River, New South Wales The Murrumbidgee River is a tributary of the Murray-Darling River system in southern Australia, in a semiarid, subtropical environment (Bren, 1988; Green et al., 2011; CSIRO, 2012; Baldwin et al., 2013; Doody et al., 2015; Nagler et al., 2016). The Murray-Darling basin supports a unique wetland ecosystem of Eucalyptus trees, grasses, sedges, reeds and aquatic vegetation, which in turn supports a unique assemblage of animal species, many found nowhere else in Australia or the world (Colloff, 2014; Colloff and Baldwin, 2010; Colloff et al., 2014). The Murray-Darling River and its tributaries have undergone a decrease in flows over the past 100 years due to diversion of water for agriculture and to climate change (Bren, 1988; Kingsford and Thomas, 2001, 2002, 2004). Lower base flows and longer periods between floods have led to deterioration of the ecosystem. Of particular concern has been the loss of red gum trees (Eucalyptus camaldulensis), an iconic species that depends on overbank floods at 2–3 year intervals to recharge the vadose zone and aquifer with low-salinity water (Catelotti et al., 2015). The gradual loss of red gum area in the river system has been of concern for some time, but was especially acute during the 2000–2010 Millennial Drought (Mac Nally et al., 2011; Sims and Colloff, 2012; Wen et al., 2009; Wen and Saintilan, 2014). In response, the Murray Darling Basin Authority released experimental environmental flows into selected areas of the Yanga National Park, New South Wales, on the Murrumbidgee River, which supports about 38,000 ha of continuous red gum forest near its juncture with the Murray Darling River (Baldwin et al., 2013; Doody et al., 2015). The whole forest received natural overbank flooding in 2000, and again in 2010 with the return of rains. Three test sites received supplemental flooding in 2005 and three more sites received supplemental flooding in 2005 and again in 2008. The purpose of the flow releases was to determine the minimum amount of water needed to keep the red gum trees in good condition, based on survival, leaf area index, transpiration rates and other tree characteristics. Doody et al. (2015) conducted a space-for-time experiment to determine the impacts of the 2005 and 2008 flows. In 2010–2011, they conducted ground surveys of trees in plots receiving different flood treatments both before and after the return of natural floods in 2010. They found that trees in plots receiving supplemental floods were generally more vigorous than those in control plots, but that red gum trees were able to exercise stomatal control of transpiration during drought conditions, hence mortality rates were low. Nagler et al. (2016) conducted a follow-up study using satellite imagery and river gage records to project the results over wider areas of the forest and further back in time. Flow records from 1936 to 2012 into the red gum forest measured at Maude Weir, below the last diversion point for water for human use before

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

6

Fig. 3. History of flows into Yanga National Forest on the Murrumbidgee River in Australia. Red bars show environmental flows released in 2005 and 2008 into portions of the forest. Data are from the gage at Maude Weir, below the last diversion point for water, representing water that flows into the Yanga National Forest. From Nagler et al., 2016. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the river enters the forest, indicate that water flowing into the forest decreased markedly (Fig. 3). While flows are variable both seasonally and inter-annually, mean annual flows have decreased significantly (P < 0.001) at a rate of 0.56 m3 s−1 yr−1 , for a loss of 55% of flow volume over 76 years of management for human use of river water. The 2005 flood inundated the test sites from July to November while the 2008 flood was from January to April (Fig. 3). The flow rate is shown as 231 m3 s−1 , which is the amount needed to produce overbank flooding throughout the forest. However, in actuality much smaller flows were released from canals and weirs within the park to produce local flooding of target areas. Nagler et al. (2016) found that the 2005 supplemental flood did stimulate greening, as measured by vegetation indices, but similar to the Colorado River Delta Minute 319 pulse flow, the greening response disappeared quickly. However, they also found that differences between plots were pre-existing. Plots receiving supplemental flooding in 2005 and 2008 were already greener than control plots as early as 1995, as shown by high NDVI values on Landsat images. Therefore, the assumption of equal starting conditions built into space-fortime experiments (Baldwin et al., 2013) was not met. Possibly, these

sites were selected for supplemental flooding because they were low-lying areas that were easier to flood. Nagler et al. (2016) also found that NDVI values for the study plots as well as the whole forest were highest in 1995, following 5 years of near-yearly peak flows of 230 m3 s−1 , sufficient to flood the forest floor, with a flood duration of 3–6 months (annual flow = 1800–3600 mcm). The conclusion, also reached by the Murray-Darling Authority, was that flows of no less than 2000 mcm were needed every other year to keep the forest in good condition (MDBA, 2010, 2013). This is a large amount of water that greatly exceeds the experimental flows of 2005 and 2008. However, Nagler et al. (2016) also reported that only a fraction of the flooding requirement was consumed in floodplain ET, similar to the situation in the Colorado River delta. A large pulse of water was needed to overtop the river banks and inundate the floodplain terraces. However, most of this water subsequently drained back to the Murray River. Comparison of flow rates at river gages upstream and downstream of Yanga National Park showed that during wet years as much as 2000 mcm entered the forest, of which 1800 exited at the downstream end and only 200 mcm was consumed by forest vegetation as ET (Nagler et al., 2016). The outflow would be available to meet downstream human or environmental water demands and would not be counted as a loss from the system. By contrast, in dry years over half of the water entering the forest would be consumed as ET, and the flow would not be sufficient to keep the vegetation in vigorous condition. Kingsford (2016) concluded that the relatively large volumes of water required to maintain the red gum forests should be subjected to rigorous monitoring and evaluation to demonstrate to the public and other water users that ecological goals are actually achieved.

6. Comparison of outcomes Some common themes in outcomes can be seen by comparing these four case studies (Table 1). First, environmental flows for all four rivers were much smaller than flows formerly experienced under a natural flow regime, resulting in narrower restoration zones compared to historic conditions. In the case of the Murrumbidgee very large flows will be needed to create enough overbank flooding to recharge the aquifer. Second, water quality issues arising from outside the restoration area, particularly excess salinity due to agricultural return flows entering river, impact the

Table 1 Goals, successes, failures and lessons learned from environmental flows on four arid zone rivers. River:

Tarim

Bill Williams

Colorado Delta

Murrimbidgee

Main goals

Raise water table to restore the poplar forest; restore water to Taitemar Lake

Water table has been raised and a narrow forest of trees has been established; lake is now seasonally restored.

Failures

Recruitment and survival of new seedlings has been slow and spatially limited

The restored portion of the floodplain is spatially limited

Lessons Learned

Active restoration of trees and better control of the hydrology during releases is needed

Continued water releases will be needed to keep the native trees and wetland areas in good condition

Stimulate germination and establishment of native tree seeds in the natural floodplain and in restoration plots; recharge the groundwater Groundwater recharge and recruitment of new trees into restoration plots have been achieved; local residents have responded positively to the flow. Germination and establishment of native trees was low outside of active restoration areas; most new plants are non-native saltcedar Future releases should be managed to expand areas of active restoration

Maintain the health of the existing red gum forest by providing overbank flooding to recharge the aquifer

Successes

Rework the channel and floodplain to enable establishment of native trees; reduce the abundance of non-native saltcedar Cottonwoods, willows and emergent plants have established on bars in the channel and have created new wetland and native tree habitat

The flooding requirements for keeping the red gum forest in good condition have been defined for the Yanga National Forest The test floods did not adequately restore the red gum stands; beneficial effects were only temporary The Murray Darling Management Authority needs to define the minimum water needs of the red gum forests and demonstrate success through monitoring of results.

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

rivers as well. A third theme is that successful restoration might require more time than originally expected. In the cases of the Tarim and Bill Williams, where releases have occurred over 15–20 years, a degree of success has been achieved for each. The Colorado River delta and Murrumbidgee River have had only recently experienced environmental flows and immediate success has not been achieved. All four rivers illustrate that time and patience on the part of NGOs, resource managers and scientists are needed to achieve success.

7

imental and adaptive management approach is needed to achieve both social and ecological benefits on a specific river system. Kingsford (2016) concluded that accurate monitoring of results is needed to demonstrate to stakeholders that ecological goals are actually achieved. Expectations need to match the amount of water available, which is frequently less than natural flow regimes (Hall et al., 2011).

Acknowledgements 7. Improving the science behind environmental flows The four case studies in this paper illustrate some of the problems encountered in planning and evaluating environmental flows. Davies et al. (2014) reviewed 156 papers on environmental flows in Australia published between 1992 and 2012. They reported that most of the papers presented models for evaluating the success of environmental flows, while less than a fifth explored site-specific flow-ecology relationships, which they regarded as the key to successful design of environmental flow regimes. The majority of papers that addressed flow-ecology relationships were presented as “natural experiments” in which a flow was implemented then a description of ecological responses was undertaken. Only two of the papers presented a manipulative experiment designed to explicitly test hypotheses about flow-ecology relationships. Yet, they argued that each opportunity for environmental flows should be treated as an experiment designed to test pre-existing hypotheses about outcomes, to maximize the information gained from the release. They described two main limitations to existing environmental flow deliveries: lack of long-term monitoring and evaluation; and lack of fundamental flow-ecology knowledge about target ecosystems. Konrad et al. (2011, 2012) and Olden et al. (2014) also argued for treating large-scale flow events as experiments rather than operational management tools. Konrad et al. (2011) analyzed managed flows conducted on 40 river systems. They described five challenges for large-scale flow experiments. First, most large-scale flows are designed to achieve ecological outcomes rather than serve as learning opportunities. Hence, results might be inconclusive and might not inform water-management decisions. Second, experimental treatments and responses span multiple time scales and are difficult to control. In the case of the Murrumbidgee River releases into the Yanga National Forest, the apparent success of supplemental flows in 2005 and 2008 turned out to be related to pre-existing conditions rather than to immediate impacts of the supplemental flows. Third, environmental flows are embedded in river networks and impacts can extend much further downstream than anticipated. Much of the Minute 319 flow might have entered the intertidal zone as groundwater flow, and/or might have flowed into agricultural well fields, both of which were outside the monitoring zones. Most of the water that inundates the floodplain of the Murrumbidgee River is not retained in the forest, but exits into the Murray River with unknown downstream impacts. Fourth, achieving ecological goals depends not only on the flow event but on the antecedent conditions of the riparian zone, including presence of propagules. Sufficient recruitment of native trees was achieved on the Bill Williams River but not in the Colorado River Delta; hence, more active restoration is often needed in conjunction with environmental flows. Finally, ecological responses are taxa-specific. For example, in the case of the Bill Williams and Colorado River Delta, Tamarix recruitment accompanies the recruitment of more desirable native taxa. All four cases show that large-scale flow experiments are inseparable from their social context, including the many competing uses for water. Konrad et al. (2011) concluded that a long-term, exper-

Funding was provided by the U.S. Geological Survey and the U.S. Bureau of Reclamation.

References Aishan, T., Halik, U., Kurban, A., Cyffka, B., 2015. Eco-morphological response of floodplain forests (Populus euphratica Oliv.) to water diversion in the lower Tarim river northwest China. Environ. Earth Sci. 73, 533–545. Arthington, A.H., 2012. Environmental Flows: Saving Rivers in the Third Millennium. University of California Press, Berkeley, CA, USA. Baldwin, D.S., Rees, G.N., Wilson, J.S., Colloff, M.J., Whitworth, K.L., Pitman, T.L., Wallace, T.A., 2013. Provisioning of bioavailable carbon between the wet and dry phases in a semi-arid floodplain. Oecologia 172, 539–550. Bao, A., Huang, Y., Ma, Y., Guo, H., Wang, Y., 2017. Assessing the effect of EWDP on vegetation restoration by remote sensing in the lower reaches of Tarim River. Ecol. Indic. 74, 261–275. Bren, L.J., 1988. Effects of river regulation on flooding of a riparian red gum forest on the River Murray, Australia. Regul. Rivers 2, 65–77. CSIRO, 2012. Assessment of the Ecological and Economic Benefits of Environmental Water in the Murray–Darling Basin. CSIRO Water for a Healthy Country National Research Flagship: Canberra. Calderon-Aguilera, L.E., Flessa, K.W., 2009. Just add water? Transboundary Colorado River flow and ecosystem services in the upper Gulf of California. In: López-Hoffman, L., McGovern, E.D., Varady, R.G., Flessa, K.W. (Eds.), Conservation of Shared Environments: Learning from the United States and Mexico. University of Arizona Press, Tucson, pp. 154–169. Catelotti, K., Kingsford, R.T., Bino, G., Bacon, P., 2015. Inundation requirements for persistence and recovery of river red gums (Eucalyptus camaldulensis) in semi-arid Australia. Biol. Conserv. 184, 346–356. Chen, Y., Cui, W.L., Chen, Y., Zhang, H., 2003. Utilization of water resources and ecological protection in the Tarim River. Acta Geogr. Sin., 2003–2002. Chen, Y.J., Chen, Y.N., Liu, J.Z., Zhang, E.X., 2009. Influence of intermittent water releases on groundwater chemistry at the lower reaches of the Tarim River, China. Environ. Monit. Assess. 158, 251–264. Chen, Y.W., Xu, C., Ye, Z., Chen, Y., 2014. Desert riparian vegetation and groundwater in the lower reaches of the Tarim River basin. Environ Earth Sci. 73, 547–558. Cohen, M.J., Henges-Jeck, C., Castillo-Moreno, G., 2001. A preliminary water balance for the Colorado River delta, 1992–1998. J. Arid Environ. 49, 35–48. Colloff, M.J., Baldwin, D.S., 2010. Resilience of floodplain ecosystems in a semi-arid environment. Rangeland J. 32, 305–314. Colloff, M.J., Ward, K.A., Roberts, J., 2014. Ecology and conservation of grassy wetlands dominated by spiny mud grass Pseudoraphis spinescens in the southern Murray-Darling Basin, Australia. Aquat. Conserv. 24, 238–255. Colloff, M.J., 2014. Flooded Forest and Desert Creek: Ecology and History of the River Red Gum. CSIRO Publishing, Melbourne. Daessle, L.W., Orozco, A., Struck, U., Camacho-Ibar, V.F., van Geldern, R., Santamari-delAngle, E., Barth, J.A.C., 2016. Sources and sinks of nutrients and organic carbon during the 2014 pulse flow of the Colorado River into Mexico. Ecolog. Eng., in press, corrected proof available online 10 March. Davies, P.M., Naiman, R.J., Warfe, D.M., Pettit, N.E., Arthington, A.H., Bunn, S.E., 2014. Flow-ecology relationships: closing the loop on effective environmental flows. Marine Freshw. Res. 65, 133–141. Doody, T.M., Colloff, M.J., Davies, M., Koul, V., Benyon, R.G., Nagler, P.L., 2015. Quantifying water requirements of riparian river red gum (Eucalyptus camaldulensis) in the Murray-Darling Basin, Australia—implications for the management of environmental flows. Ecohydrology 8, 1471–1487. Galindo-Bect, M.S., Glenn, E.P., Page, H.M., Fitzsimmons, K., Galindo-Bect, L.A., Hernandez-Ayon, J.M., Petty, R.L., Garcia-Hernandez, J., Moore, D., 2000. Penaeid shrimp landings in the upper Gulf of California in relation to Colorado River freshwater discharge. Fish. Bull. 98, 222–225. Glenn, E.P., Lee, C., Felger, R., Zengel, S., 1996. Effect of water management on the wetlands of the Colorado River delta, Mexico. Conserv. Biol. 10, 1175–1186. Glenn, E.P., Zamora-Arroyo, F., Nagler, P.L., Briggs, M., Shaw, W., Flessa, K., 2001. Ecology and conservation biology of the Colorado River delta, Mexico. J. Arid Environ. 49, 5–15. Glenn, E.P., Flessa, K.W., Cohen, M.J., Nagler, P.L., Rowell, K., Zamora-Arroyo, F., 2007. Just add water and the Colorado River still reaches the sea. Enivorn. Manage. 40, 1–6.

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9 8

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

Glenn, E.P., Flessa, K.W., Pitt, J., 2013. Restoration potential of the aquatic ecosystems of the Colorado River Delta, Mexico : Introduction to special issue on Wetlands of the Colorado River Delta. Ecol. Eng. 59, 1–6. Green, D., Petrovic, J., Moss, P., Burrell, M., 2011. Water Resources and Management Overview: Murrumbidgee Catchment. NSW Office of Water Supply, Sydney. Gross, K.F., 2003. Environmental flows, river salinity and biodiversity conservation: managing trade-offs in the Murray-Darling basin. Aust. J. Bot. 51, 619–625. Hall, A.A., Rood, S.B., Higgins, P.S., 2011. Resizing a river: a downscaled, seasonal flow regime promotes riparian restoration. Restor. Ecol. 19, 351–359. Hinojosa-Huerta, O., Nagler, P.L., Carrillo-Guererro, Y.K., Glenn, E.P., 2013. Effects of drought on birds and riparian vegetation in the Colorado River Delta, Mexico. Ecol. Eng. 59, 104–110. Hinojosa-Huerta, O., 2006. Conservation of Birds in the Colorado River Delta, Mexico. Disseration from the University of Arizona Dissertation, Tucson AZ. Hou, P., Beeton, R., Carter, X., Dona, X., Lia, X., 2007. Response to environmental flows in the lower Tarim River, Xinjiang, China: ground water. J. Environ. Manage. 83, 371–382. House, P.K., Shafroth, P.B., Beauchamp, V.B., 2006. Chapter 2, hydrology and geomorphology. In: Shafroth, P.B., Beauchamp, V.B. (Eds.), Defining Ecosystem Flow Requirements for the Bill Williams River, Arizona. U.S. Department of Interior, U.S. Geol. Surv., Reston, VA, 135 p. Pages 9–30. IBWC, 2000. Minute No. 306. Conceptual Framework for United States?Mexico Studies for Future Recommendations Concerning the Riparian and Estuarine Ecology of the Limitrophe Section of the Colorado River and Its Associated Delta. International Boundary and Water Commission, El Paso, TX, http://www.ibwc.gov/Files/Minutes/Min306.pdf IBWC, 2010. Minute 316. International Boundary and Water Commission (IBWC), 2012. Minute No. 319. Interim International Cooperative Measures in the Colorado River Basin Through 2017 and Extension of Minute 318 Cooperative Measures to Address the Continued Effects of the April 2010 Earthquake in the Mexicali Valley, Baja California. International Boundary and Water Commission, El Paso, TX http:// www.ibwc.state.gov/Files/Minutes/Minute 319.pdf. International Boundary and Water Commission (IBWC), 2016. Stream gage data: 09-5222.00 Colorado River at Southerly International Boundary. http://www. ibwc.state.gov/wad/DDQSIBCO.HTM. Jarchow, C., Nagler, P.L., Glenn, E.P., Ramirez-Hernandez, J., Rodriquez-Burgeenoc, E.R., 2016. Evapotranspiration by remote sensing: an analysis of the Colorado River delta before and after the minute 319 pulse flows to Mexico. Ecolog. Eng., in press. Kendy, E., Flessa, K.W., Schlatter, K.J., de la Parra, C.A., Hinojosa-Huerta, O.M., Carrillo-Guerrero, Y.K., Guillen, E., 2017. Leveraging environmental flows to reform water management policy: lessons learned from the 2014 Colorado River delta pulse flow. Ecol. Eng. Kingsford, R.T., Thomas, R.F., 2001. Changing water regimes and wetland habitat on the Lower Murrumbidgee floodplain of the Murrumbidgee River in arid Australia. Report to Environment Australia, April 2001. New South Wales National Parks and Wildlife Service, P.O. Box 1967, Hurstville, NSW 2220, Australia. Kingsford, R.T., Thomas, R.F., 2002. Use of satellite image analysis to track wetland loss on the Murrumbidgee River floodplain in arid Australia. Water Sci. Technol., 450–463. Kingsford, R.T., Thomas, R.F., 2004. Destruction of wetlands and waterbird populations by dams and irrigation on the Murrumbidgee River in arid Australia. Environ. Manage. 34, 383–396. Kingsford, R.T., 2016. Flow regulation and its effects on Australia’s riverine vegetatgion. In: Capon, S., James, C., Reid, M. (Eds.), Vegetation of Australian River Landscapes: Biology, Ecology and Management. CSIRO Publikshing, Clayton South, VIV, Australia, pp. 261–272. Konrad, C.P., Olden, J.D., Lytle, D.A., Melis, T.S., Schidt, J.C., Bray, E.N., Freeman, M.C., Gido, K.B., Hemphill, N.P., Kennard, M.J., McMullen, L.E., Mims, M.C., Pyron, M., Robinson, C.T., Williams, J.G., 2011. Large-scale flow expreiments for managing river systems. Bioscience 61, 948–959. Konrad, C.P., Warner, A., Higgins, J.V., 2012. Evaluating dam re-operation for freshwater conservation in the sustainable rivers project. River Res. Appl. 28, 777–792. Ling, H., Zhang, P., Xu, H., Zhao, X., 2015. How to regenerate and protect desert riparian Populus euphratica forest in arid areas. Sci. Rep. 5, 15418. Murray-Darling Basin Authority (MDBA), 2010. Guide to the Proposed Basin Plan, Volume 1, Overview. MDBA Publication No. 60/10. Murray-Darling Basin Authority, Canberra. Murry-Darling Basing Authority (MDBA), 2013. Preliminary Overview of Constraints to Environmental Water Delivery in the Murray-Darling Basin. Murray-Darling Basin Authority, Canberra. Mac Nally, R., Cunningham, S.C., Baker, P.J., Horner, G.J., Thompson, J.R., 2011. Dynamics of Murray-Darling floodplain forests under multiple stressors: the past, present, and future of an Australian icon. Water Resour. Res. 47, W00G05. Mahoney, J.M., Rood, S.B., 1998. Streamflow requirements for cottonwood seedling recruitment—an interactive model. Wetlands 18, 634–645. Merritt, D.M., Scott, M.L., Poff, N.L., Auble, G.T., Lytle, D.A., 2010. Theory, methods and tools for determining environmental flows for riparian vegetation: riparian vegetation-flow response guilds. Freshw. Biol. 55, 206–225. Mueller, E.R., Schmidt, J.C., Topping, D.J., Ramirez-Hernandez, J., Rodriguez-Burgueno, J.E., Shafroth, P.B., Grams, P.E., 2017. Geomorphic change and sediment transport during a small artificial flood in a transformed post-dam delta: The Colorado River delta, United States and Mexico. Ecol. Eng., in press.

Nagler, P.L., Hinojosa-Huerta, O., Glenn, E.P., Garcia-Hernandez, J., Romo, R., Curtis, C., Huete, A.R., Nelson, S.G., 2005. Regeneration of native trees in the presence of invasive saltcedar in the Colorado River delta, Mexico. Conserv. Biol. 19, 1842–1852. Nagler, P.L., Glenn, E.P., Hinojosa-Huerta, O., 2009. Synthesis of ground and remote sensing data for monitoring ecosystem functions in the Colorado River Delta, Mexico. Remote Sens. Environ. 113, 1473–1485. Nagler, P.L., Doody, T.M., Jarchow, C.J., Barreto-Munoz, A., Didan, K., 2016. Wide-area estimates of evapotranspiration by red gum (Eucalyptus camaldulensis) and associated vegetation in the Murrayp-Darling River Basin, Australia. Hydrol. Proc. 30, 1376–1387. Nilsson, C., Berggren, K., 2000. Alterations of riparian ecosystems caused by river regulation. BioScience 50, 783–792. Olden, J.D., Konrad, C.P., Melis, T.S., Kennard, M.J., Freeman, M.C., Mims, M.C., Bray, E.N., Gido, K.B., Hemphill, N.P., Lytle, D.A., McMullen, L.E., Pyron, M., Robinson, C.T., Schmidt, J.C., Williams, J.G., 2014. Are large-scale flow experiments informing the science and management of freshwater systems? Front. Ecol. Environ. 12, 176–185. Orth, D.J., Maughan, O.E., 1982. Evaluation of the incremental methodology for recommending instream flows for fishes. Trans. Am. Fish. Soc. 111, 413–445. Pitt, J., Kendy, E., 2016. Shaping the 2014 Colorado River delta pulse flow: rapid environmental flow design for ecological outcomes and scientific learning. Ecol. Eng., in press. Pitzer, G., 2014. Stubborn personalities and good fortune. The C0olorado River Pulse Flow. River Report, Water Eduction Foundation, Summer, 2014, pp. 1-3, http://content.yudu.com/A30zo7/RiverReportSummer14/resources/index. htm?referrerUrl=https%3A%2F%2Fwww.google.com%2F. Poff, N.L., Allan, J.D., Bain, M.B., Karr, J.R., Prestegaard, K.L., Richter, B.D., Sparks, R.E., Stromberg, J.C., 1997. The natural flow regime: a paradigm for river conservation and restoration. BioScience 47, 769–784. Richardson, D.M., Holmes, P.M., Eslers, K.J., Galatowitsch, S.M., Stromberg, J.C., Kirkman, S.P., Pysek, P., Hobbs, R.J., 2007. Riparian vegetation: degradation, alien plant invasions, and restoration prospects. Divers. Distrib. 13, 126–139. Richter, B.D., Mathews, R., Wigington, R., 2003. Ecologically sustainable water management: manging river flows for ecological integrity. Ecol. Appl. 13, 206–224. Rood, S.B., Gouirley, C.R., Ammon, E.M., Heki, L.G., Klotz, J.R., Morrison, M.L., Mosley, D., Scoppettone, G.G., Swanson, S., Wagner, P.L., 2003. Flows for floodplain forests: a successful riparian restoration. Bioscience 53, 647–656. Rood, S.B., Goater, L.A., Mahoney, J.M., Pearce, C.M., 2007. Floods, fire and ice: disturbance ecology of riparian cottonwoods. Botany 85, 1019–1032. Schlatter, K.J., Grabau, M.R., Shafroth, P.B., Zamora-Arroyo, F., 2017. Integrating Active Restoration with Environmental Flows to Improve Native Riparian Tree Establishment in the Colorado River Delta. In: Ecol. Eng. Shafroth, P.B., Auble, G.T., Stromberg, J.C., Patten, D.T., 1998. Establishment of woody riparian vegetation in relation to annual patterns of streamflow Bill Williams River, Arizona. Wetlands 18, 577–590. Shafroth, P.B., Stromberg, J.C., Patten, D.T., 2002. Riparian vegetation response to altered disturbance and stress regimes. Ecol. Appl. 12, 107–123. Shafroth, P.B., Wilcox, A.C., Lytle, D.A., Hickey, J.T., Anderson, D.C., Beauchamp, V.B., Hautzinger, A., McMullen, L.E., Warner, A., 2010. Ecosystem effects of environmental flows: modelling and experimental floods in a dryland river. Freshw. Biol. 55, 68–85. Shafroth, P.B., Schlatter, K.J., Gomez-Sapiens, M., Grabau, M.R., Ramirez-Hernandez, J., Rodriguez-Burgueno, E., Flessa, K.W., 2017. A large-scale streamflow experiment for riparian restoration in the Colorado River delta. Ecol. Eng. Sher, A.A., Marschall, D.L., Gilbert, S.A., 2000. Competition between native Populus deltoides and invasive Tamarix ramosissima and the implications for reestablishing flooding disturbance. Conserv. Biol. 14, 1744–1754. Sher, A.A., Marshall, D.L., Taylor, J.P., 2002. Establishment patterns of native Populus and Salix in the presence of invasive nonnative Tamarix. Ecol. Appl. 12, 760–772. Sims, N.C., Colloff, M.J., 2012. Remote sensing of vegetation responses to flooding of a semi-arid floodplain: implications for monitoring ecological effects of environmental flows. Ecol. Indic. 18, 387–391. Standford, J.A., Lorang, M.S., Hauser, F.R., 2005. The shifting habitat mosaic of river ecosystems. Int. Ver. Theor. Angew. Proc. 29, 123–136. Stromberg, J.C., Beauchampt, V.B., Dixon, M.D., Lite, S., Paradzick, C., 2007. Importance of low-flow and high-flow characteristics to restoration of riparian vegetation along rivers in arid southwestern United States. Freshw. Biol. 52, 651–667. United Nations Educational, Scientific and Cultural Organization (UNESCO), 2010. Taklimakan Desert—Populus euphratica Forests. World Heritage Tentative Nominations List. http://whc.unesco.org/en/tentativelists/5532/. United States Army Corps of Engineers (USACE), 2003, Water Control Manual: Alamo Dam and Lake, Colorado River Basin, Bill Williams River Basin: USACE, Los Angeles District, Reservoir Regulation Section. Wen, L., Saintilan, N., 2014. Linking local ecological outcomes with basin-wide planning:a case study of Yanga National Park, an important Australian inland forsted wetland. Hydrol. Sci. J. 59, 904–915. Wen, L., Ling, J., Saintilan, N., Rogers, K., 2009. An investigation of the hydrological requirements of river red gum (Eucalyptus camaldulensis) forest, using classification and regression tree modelling. Ecohydrology 2, 143–155. Wilcox, A.C., Shafroth, P.B., 2013. Coupled hydrogeomorphic and wood seedling responses to controlled flood releases in a dryland river. Water Resour. Res. 49, 2843–2860.

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

G Model ECOENG-4557; No. of Pages 9

ARTICLE IN PRESS E.P. Glenn et al. / Ecological Engineering xxx (2017) xxx–xxx

Yan, H., Wang, Y.J., Wang, Y.Q., 2014. The influence of 10 years of water conveyances on groundwater and juvenile Populus euphratica of the lower Tarim River. Environ. Earth Sci. 71, 4091–4096. Yongbo, L., Chen, Y., 2007. Saving the green corridor Recarging groundwater to restore riparian forest along the Lower Tarim River, China. Ecol. Restor. 25, 112–117.

Zhang, X., Chen, Y., Li, W., Yu, Y., Sun, Z., 2013. Restoration of the lower reaches of the Tarim River in China. Reg. Environ. Change 13, 1021–1029. Zhu, Y., Chen, Y., Liang, R., Lu, H., Zhao, W., Fei, Y., Min, W., 2016. Ecosystem restoration and conservation in the arid inland river basins of Northwest China: Problems and strategies. Ecol. Eng. 94, 629–637.

Please cite this article in press as: Glenn, E.P., et al., Effectiveness of environmental flows for riparian restoration in arid regions: A tale of four rivers. Ecol. Eng. (2017), http://dx.doi.org/10.1016/j.ecoleng.2017.01.009

9