Search

Search Results

• 3311. [Article] Cruise ship disturbance to Kittlitz's murrelets (Brachyramphus brevirostris) in Glacier Bay National Park and Preserve, Alaska

  The Kittlitz's murrelet (Brachyramphus brevirostris), a small pursuit-diving seabird in the family Alcidae, occurs across much of coastal Alaska and parts of the Russian Far East. Glacier Bay National ...

  Citation × Citation Citation Abstract Copy Title: Cruise ship disturbance to Kittlitz's murrelets (Brachyramphus brevirostris) in Glacier Bay National Park and Preserve, Alaska Author: Marcella, Timothy K. The Kittlitz's murrelet (Brachyramphus brevirostris), a small pursuit-diving seabird in the family Alcidae, occurs across much of coastal Alaska and parts of the Russian Far East. Glacier Bay National Park, located in Southeast Alaska, is believed to support approximately 37% of the worldwide breeding population of Kittlitz's murrelets during the summer months. Recent concern over apparent population declines in Alaska, coupled with the Park's dual mandate of resource preservation and visitation, led to this study. Cruise ships, although not the most numerous vessel type operating in Glacier Bay, have previously been identified as the vessel type eliciting the greatest disturbance response from Kittlitz's murrelets. During the murrelet breeding seasons in 2011 and 2012, my field assistants and I collected focal observations of 4,251 Brachyramphus murrelets from the bow of cruise ships traveling through Glacier Bay. Identification of murrelets to species was hampered by both the distance at which murrelets responded to the approaching ship and the type of response to the ship (diving vs. flushing). For roughly 40% of focal observations of murrelets from cruise ships, the species of murrelet (Kittlitz's murrelet or marbled murrelet [B. marmoratus]) could not be identified. Apparent habitat partitioning by the two murrelet species in Glacier Bay resulted in 79% of identified murrelets in the upper section of the Bay (Upper Bay) being Kittlitz's murrelets, while 83% of identified murrelets in the lower section of the Bay (Lower Bay) were marbled murrelets. In the Upper Bay, cruise ships are predicted to disturb 61% of all murrelets within 850 m on either side of the cruise ship's course (i.e., elicited a flushing or diving response), whereas in the Lower Bay, cruise ships are predicted to disturb 72% of murrelets within 850 m of the ship's course. Using Cox multistate models, I demonstrated that murrelets in the Upper Bay (predominantly Kittlitz's murrelets) were more likely to dive than flush in response to approaching cruise ships, whereas murrelets in the Lower Bay (predominantly marbled murrelets) were more likely to flush than dive. Also, murrelets in the Upper Bay responded to cruise ships by flushing or diving at shorter distances from the ship compared to murrelets in the Lower Bay. Murrelets in both areas of Glacier Bay generally reacted to cruise ships at greater distances when the ship approached indirectly, presumably because of the larger profile presented by a passing ship as opposed to a directly advancing ship. Absolute distance of the cruise ship from a focal murrelet was a strong predictor of murrelet disturbance response; no other management-relevant covariates that were measured during this study (e.g., ship velocity, distance to shore, whether a cruise ship had entered the Bay earlier that day) explained a significant proportion of the variation in murrelet response. Inferences based on data collected on-board cruise ships were limited to murrelet disturbance responses that occurred within 1 km of the ship. This was because of limits to the distance from the ship at which behavioral responses could be observed and the a priori assumption that disturbance to murrelets by cruise ships was unlikely at distances greater than 1 km. Results from shipboard observations indicated that some proportion of murrelets encountered at the farthest distance we could make inferences were on occasion disturbed (point estimate at 850 m perpendicular distance from ship's course = 15-30% probability of flushing or diving). This suggests that disturbance of murrelets by cruise ships in Glacier Bay exceeded expected distance thresholds. In order to investigate the effects of cruise ships on murrelet behavior at distances greater than 1 km, my assistants and I collected a total of 643 focal observations of Kittlitz's murrelets during 181 hours of observation from land-based observation sites in the Upper Bay during the 2012 field season. By combining these data with AIS and GPS ship tracks, I was able to append distance to the nearest cruise ship to each focal murrelet observation and search for patterns in murrelet behavior. By collecting data in this manner, I was able to avoid biasing the study based on pre-conceived notions of what constituted a threshold distance for cruise ships to disturb Kittlitz's murrelets. Using a segmented regression model within a logistic regression framework, I found that Kittlitz's murrelets exhibited a disturbance threshold (defined as an increased incidence of flushing from the water) by cruise ships at distances of at least 1.6 km, and perhaps as great as 6.0 km, with a best estimate of threshold disturbance distance at 3.8 km from a cruise ship. When cruise ships were greater than 3.8 km from focal Kittlitz's murrelets, the baseline probability of murrelets flushing during a focal observation period was 12.5%. When cruise ships were less than 3.8 km from focal Kittlitz's murrelets, the probability of flushing increased logistically with decreasing distance to an estimated 48% for the closest approach distances. The unexpectedly long distances at which murrelet behavior was affected by cruise ships in Glacier Bay is most likely attributable to social facilitation by other disturbed murrelets, because similar numbers of murrelets flushed when cruise ships were approaching (n = 30) as when they were receding (n = 27). Once a Kittlitz's murrelet flushed from the water, the subsequent duration of flight did not vary with distance to the nearest cruise ship. Instead, the duration of Kittlitz's murrelet flight was associated with time of day. The strong association between the proximity of cruise ships and the probability of a murrelet flushing, even at distances of several kilometers, demonstrates that Kittlitz's murrelets in Glacier Bay are susceptible to disturbance from cruise ships at distances greater than has previously been published for any seabird. Title: Cruise ship disturbance to Kittlitz's murrelets (Brachyramphus brevirostris) in Glacier Bay National Park and Preserve, Alaska Author: Marcella, Timothy K. The Kittlitz's murrelet (Brachyramphus brevirostris), a small pursuit-diving seabird in the family Alcidae, occurs across much of coastal Alaska and parts of the Russian Far East. Glacier Bay National Park, located in Southeast Alaska, is believed to support approximately 37% of the worldwide breeding population of Kittlitz's murrelets during the summer months. Recent concern over apparent population declines in Alaska, coupled with the Park's dual mandate of resource preservation and visitation, led to this study. Cruise ships, although not the most numerous vessel type operating in Glacier Bay, have previously been identified as the vessel type eliciting the greatest disturbance response from Kittlitz's murrelets. During the murrelet breeding seasons in 2011 and 2012, my field assistants and I collected focal observations of 4,251 Brachyramphus murrelets from the bow of cruise ships traveling through Glacier Bay. Identification of murrelets to species was hampered by both the distance at which murrelets responded to the approaching ship and the type of response to the ship (diving vs. flushing). For roughly 40% of focal observations of murrelets from cruise ships, the species of murrelet (Kittlitz's murrelet or marbled murrelet [B. marmoratus]) could not be identified. Apparent habitat partitioning by the two murrelet species in Glacier Bay resulted in 79% of identified murrelets in the upper section of the Bay (Upper Bay) being Kittlitz's murrelets, while 83% of identified murrelets in the lower section of the Bay (Lower Bay) were marbled murrelets. In the Upper Bay, cruise ships are predicted to disturb 61% of all murrelets within 850 m on either side of the cruise ship's course (i.e., elicited a flushing or diving response), whereas in the Lower Bay, cruise ships are predicted to disturb 72% of murrelets within 850 m of the ship's course. Using Cox multistate models, I demonstrated that murrelets in the Upper Bay (predominantly Kittlitz's murrelets) were more likely to dive than flush in response to approaching cruise ships, whereas murrelets in the Lower Bay (predominantly marbled murrelets) were more likely to flush than dive. Also, murrelets in the Upper Bay responded to cruise ships by flushing or diving at shorter distances from the ship compared to murrelets in the Lower Bay. Murrelets in both areas of Glacier Bay generally reacted to cruise ships at greater distances when the ship approached indirectly, presumably because of the larger profile presented by a passing ship as opposed to a directly advancing ship. Absolute distance of the cruise ship from a focal murrelet was a strong predictor of murrelet disturbance response; no other management-relevant covariates that were measured during this study (e.g., ship velocity, distance to shore, whether a cruise ship had entered the Bay earlier that day) explained a significant proportion of the variation in murrelet response. Inferences based on data collected on-board cruise ships were limited to murrelet disturbance responses that occurred within 1 km of the ship. This was because of limits to the distance from the ship at which behavioral responses could be observed and the a priori assumption that disturbance to murrelets by cruise ships was unlikely at distances greater than 1 km. Results from shipboard observations indicated that some proportion of murrelets encountered at the farthest distance we could make inferences were on occasion disturbed (point estimate at 850 m perpendicular distance from ship's course = 15-30% probability of flushing or diving). This suggests that disturbance of murrelets by cruise ships in Glacier Bay exceeded expected distance thresholds. In order to investigate the effects of cruise ships on murrelet behavior at distances greater than 1 km, my assistants and I collected a total of 643 focal observations of Kittlitz's murrelets during 181 hours of observation from land-based observation sites in the Upper Bay during the 2012 field season. By combining these data with AIS and GPS ship tracks, I was able to append distance to the nearest cruise ship to each focal murrelet observation and search for patterns in murrelet behavior. By collecting data in this manner, I was able to avoid biasing the study based on pre-conceived notions of what constituted a threshold distance for cruise ships to disturb Kittlitz's murrelets. Using a segmented regression model within a logistic regression framework, I found that Kittlitz's murrelets exhibited a disturbance threshold (defined as an increased incidence of flushing from the water) by cruise ships at distances of at least 1.6 km, and perhaps as great as 6.0 km, with a best estimate of threshold disturbance distance at 3.8 km from a cruise ship. When cruise ships were greater than 3.8 km from focal Kittlitz's murrelets, the baseline probability of murrelets flushing during a focal observation period was 12.5%. When cruise ships were less than 3.8 km from focal Kittlitz's murrelets, the probability of flushing increased logistically with decreasing distance to an estimated 48% for the closest approach distances. The unexpectedly long distances at which murrelet behavior was affected by cruise ships in Glacier Bay is most likely attributable to social facilitation by other disturbed murrelets, because similar numbers of murrelets flushed when cruise ships were approaching (n = 30) as when they were receding (n = 27). Once a Kittlitz's murrelet flushed from the water, the subsequent duration of flight did not vary with distance to the nearest cruise ship. Instead, the duration of Kittlitz's murrelet flight was associated with time of day. The strong association between the proximity of cruise ships and the probability of a murrelet flushing, even at distances of several kilometers, demonstrates that Kittlitz's murrelets in Glacier Bay are susceptible to disturbance from cruise ships at distances greater than has previously been published for any seabird. Close

• 3312. [Article] Abrupt deglacial climate changes in the North Pacific and implications for climate tipping points

  Paleoclimate archives have revealed abrupt climate events that are superimposed on more gradual climate changes throughout the last glacial and deglacial periods. The underlying causes of such rapid climate ...

  Citation × Citation Citation Abstract Copy Title: Abrupt deglacial climate changes in the North Pacific and implications for climate tipping points Author: Praetorius, Summer Kate Paleoclimate archives have revealed abrupt climate events that are superimposed on more gradual climate changes throughout the last glacial and deglacial periods. The underlying causes of such rapid climate changes are still poorly understood, but the strong expression of these events in northern hemisphere records likely points to climatic mechanisms of a northern origin. A leading hypothesis for the trigger of these climate fluctuations has been changes in the strength of the Atlantic meridional overturning circulation (AMOC). However, the very rapid nature of some of the observed climate transitions (3-50 years) suggests a potential role for abrupt shifts in atmospheric circulation or nonlinear feedbacks within the climate system. Understanding the relative timing and magnitude of these events in different regions of the globe will help to identify the sources and possible amplifying mechanisms that have led to abrupt climate changes in the past, which will provide insight and constraints on the potential for abrupt climate changes in the future. This dissertation seeks to characterize climate changes occurring in the Northeast Pacific during the last deglaciation, a time period that encompasses the dynamic transition between the last ice age and the modern day interglacial period. So far, high-resolution records with precise chronologies from the North Pacific have been sparse, and paleoclimate models and proxy reconstructions disagree about the deglacial climate changes that are both predicted and observed to have occurred in this region. Marine sediment records from the Gulf of Alaska (GOA) have exceptionally high resolution (~1 cm/yr), making it possible to reconstruct climate changes in unprecedented detail for the North Pacific region. We establish new multi-decadal scale records of surface ocean variability using planktonic oxygen isotopes and sea-surface temperature (SST) estimates based on the alkenone U₃₇[superscript K'] unsaturation index, as well as regional records of ice-rafting and deglacial volcanic activity sourced from the Mt. Edgecumbe volcanic field (MEVF). The age models for these records are constrained by high-precision radiocarbon dating, tephra correlation, and "tuning" to the decadal-scale North Greenland Ice Core Project (NGRIP) oxygen isotope record. We combine new and previously published data from a depth transect of marine sites in the GOA and Northeast Pacific to place surface ocean changes in context of oceanic variability throughout the water column. These reconstructions are then used to evaluate three fundamental questions: 1) what are the timing and patterns of deglacial climate changes in the North Pacific relative to other regions, 2) what are the potential forcing mechanisms for deglacial climate variability in this region, and 3) how does the subsurface ocean respond to and influence abrupt climate change. In chapter two, we compare the timing and patterns of climate changes occurring between the North Pacific and North Atlantic regions. A major debate in the paleoclimate literature has been whether these regions operate in a synchronized or seesaw like mode. We compare the high resolution GOA and NGRIP oxygen isotope records as proxies for local temperature, and find that both synchronous and asynchronous climate patterns occur between regions throughout the past 18,000 years. The most abrupt climate transitions are preceded/accompanied by synchronous behavior, whereas times of relative climate stability exhibit asynchronous or anticorrelated (seesaw) patterns. This implies that coupling of North Pacific and North Atlantic heat transport could act as an amplifying mechanism in abrupt northern hemisphere climate change, whereas opposing oceanic regimes could act to balance northern hemisphere heat transport, and thus promote climate stability. In chapter three, we examine the timing between regional deglaciation and volcanism to evaluate potential feedbacks between climate and volcanic activity. Although volcanic eruptions have been observed to contribute to abrupt climate fluctuations with global effects in historical times, the role of volcanic forcing in climate variability of the more distant past (prior to the Holocene) has been neglected due to the very short-time scales in which volcanic events occur, and the difficulty of obtaining records with high enough resolution to capture these events and their associated climate effects. We evaluate the source and timing of a sequence of 23 tephra layers preserved in high-accumulation rate sediment cores proximal to the MEVF, and examine the regional climate response to this volcanic activity through comparison with reconstructions of sea surface temperatures, oxygen isotopes, and the δ¹⁸O of seawater. We find that the onset of enhanced volcanic activity coincides with abrupt warming at the onset of the Bølling Allerød, regional retreat of glaciers, and a period of rapid vertical land motion predicted from a model of regional isostatic rebound. These finding support the hypothesis that deglaciation may promote volcanism by removing crustal loading. The records of sea surface variability show large fluctuations during the episode of intense volcanic activity, suggesting that deglacial volcanic activity may not only respond to climate, but may also contribute to climate variance during the deglacial interval. In Chapter four, we examine the oceanographic changes that lead to two episodes of hypoxia in the GOA that lasted for millennia during the deglaciation. Similar hypoxic events have been documented across the North Pacific, indicating a widespread expansion of the oxygen minimum zone (OMZ) during the Bølling Allerød and early Holocene warm periods. These episodes have been linked to enhanced export productivity in many sites, however, the driving mechanisms for enhanced productivity and ocean deoxygenation remain elusive. Our alkenone temperature reconstructions reveal two abrupt warmings of 4-5°C that precisely coincide with the onset of increased export productivity and a sudden shift to hypoxic conditions, suggesting a strong link between ocean warming, marine productivity, and deoxygenation. Oxygen isotopes throughout the water column indicate that a transient subsurface warming of ~2°C might have accompanied the first hypoxic event during the BA. We propose that abrupt ocean warming lead to an expansion of the North Pacific OMZ through a reduction in oxygen solubility, enhanced thermal stratification, and a stimulation of marine productivity through the stabilization of the euphotic zone (related to stratification), combined with enhanced nutrient input from remobilization of iron in hypoxic shelf sediments. These studies indicate that large surface and subsurface ocean changes occurred in the North Pacific during the last deglaciation, with the potential for important feedbacks on global climate. Title: Abrupt deglacial climate changes in the North Pacific and implications for climate tipping points Author: Praetorius, Summer Kate Paleoclimate archives have revealed abrupt climate events that are superimposed on more gradual climate changes throughout the last glacial and deglacial periods. The underlying causes of such rapid climate changes are still poorly understood, but the strong expression of these events in northern hemisphere records likely points to climatic mechanisms of a northern origin. A leading hypothesis for the trigger of these climate fluctuations has been changes in the strength of the Atlantic meridional overturning circulation (AMOC). However, the very rapid nature of some of the observed climate transitions (3-50 years) suggests a potential role for abrupt shifts in atmospheric circulation or nonlinear feedbacks within the climate system. Understanding the relative timing and magnitude of these events in different regions of the globe will help to identify the sources and possible amplifying mechanisms that have led to abrupt climate changes in the past, which will provide insight and constraints on the potential for abrupt climate changes in the future. This dissertation seeks to characterize climate changes occurring in the Northeast Pacific during the last deglaciation, a time period that encompasses the dynamic transition between the last ice age and the modern day interglacial period. So far, high-resolution records with precise chronologies from the North Pacific have been sparse, and paleoclimate models and proxy reconstructions disagree about the deglacial climate changes that are both predicted and observed to have occurred in this region. Marine sediment records from the Gulf of Alaska (GOA) have exceptionally high resolution (~1 cm/yr), making it possible to reconstruct climate changes in unprecedented detail for the North Pacific region. We establish new multi-decadal scale records of surface ocean variability using planktonic oxygen isotopes and sea-surface temperature (SST) estimates based on the alkenone U₃₇[superscript K'] unsaturation index, as well as regional records of ice-rafting and deglacial volcanic activity sourced from the Mt. Edgecumbe volcanic field (MEVF). The age models for these records are constrained by high-precision radiocarbon dating, tephra correlation, and "tuning" to the decadal-scale North Greenland Ice Core Project (NGRIP) oxygen isotope record. We combine new and previously published data from a depth transect of marine sites in the GOA and Northeast Pacific to place surface ocean changes in context of oceanic variability throughout the water column. These reconstructions are then used to evaluate three fundamental questions: 1) what are the timing and patterns of deglacial climate changes in the North Pacific relative to other regions, 2) what are the potential forcing mechanisms for deglacial climate variability in this region, and 3) how does the subsurface ocean respond to and influence abrupt climate change. In chapter two, we compare the timing and patterns of climate changes occurring between the North Pacific and North Atlantic regions. A major debate in the paleoclimate literature has been whether these regions operate in a synchronized or seesaw like mode. We compare the high resolution GOA and NGRIP oxygen isotope records as proxies for local temperature, and find that both synchronous and asynchronous climate patterns occur between regions throughout the past 18,000 years. The most abrupt climate transitions are preceded/accompanied by synchronous behavior, whereas times of relative climate stability exhibit asynchronous or anticorrelated (seesaw) patterns. This implies that coupling of North Pacific and North Atlantic heat transport could act as an amplifying mechanism in abrupt northern hemisphere climate change, whereas opposing oceanic regimes could act to balance northern hemisphere heat transport, and thus promote climate stability. In chapter three, we examine the timing between regional deglaciation and volcanism to evaluate potential feedbacks between climate and volcanic activity. Although volcanic eruptions have been observed to contribute to abrupt climate fluctuations with global effects in historical times, the role of volcanic forcing in climate variability of the more distant past (prior to the Holocene) has been neglected due to the very short-time scales in which volcanic events occur, and the difficulty of obtaining records with high enough resolution to capture these events and their associated climate effects. We evaluate the source and timing of a sequence of 23 tephra layers preserved in high-accumulation rate sediment cores proximal to the MEVF, and examine the regional climate response to this volcanic activity through comparison with reconstructions of sea surface temperatures, oxygen isotopes, and the δ¹⁸O of seawater. We find that the onset of enhanced volcanic activity coincides with abrupt warming at the onset of the Bølling Allerød, regional retreat of glaciers, and a period of rapid vertical land motion predicted from a model of regional isostatic rebound. These finding support the hypothesis that deglaciation may promote volcanism by removing crustal loading. The records of sea surface variability show large fluctuations during the episode of intense volcanic activity, suggesting that deglacial volcanic activity may not only respond to climate, but may also contribute to climate variance during the deglacial interval. In Chapter four, we examine the oceanographic changes that lead to two episodes of hypoxia in the GOA that lasted for millennia during the deglaciation. Similar hypoxic events have been documented across the North Pacific, indicating a widespread expansion of the oxygen minimum zone (OMZ) during the Bølling Allerød and early Holocene warm periods. These episodes have been linked to enhanced export productivity in many sites, however, the driving mechanisms for enhanced productivity and ocean deoxygenation remain elusive. Our alkenone temperature reconstructions reveal two abrupt warmings of 4-5°C that precisely coincide with the onset of increased export productivity and a sudden shift to hypoxic conditions, suggesting a strong link between ocean warming, marine productivity, and deoxygenation. Oxygen isotopes throughout the water column indicate that a transient subsurface warming of ~2°C might have accompanied the first hypoxic event during the BA. We propose that abrupt ocean warming lead to an expansion of the North Pacific OMZ through a reduction in oxygen solubility, enhanced thermal stratification, and a stimulation of marine productivity through the stabilization of the euphotic zone (related to stratification), combined with enhanced nutrient input from remobilization of iron in hypoxic shelf sediments. These studies indicate that large surface and subsurface ocean changes occurred in the North Pacific during the last deglaciation, with the potential for important feedbacks on global climate. Close

• 3313. [Article] Restoring the Columbia River Estuary : Chinook Salmon Recovery and Invasive Species Management

  As highlighted in this study, shallow-water habitats and the prey they provide are important for juvenile salmon in the Columbia River estuary, but the spread of invasive species and large-scale changes ...

  Citation × Citation Citation Abstract Copy Title: Restoring the Columbia River Estuary : Chinook Salmon Recovery and Invasive Species Management Author: Klopfenstein, Rachael As highlighted in this study, shallow-water habitats and the prey they provide are important for juvenile salmon in the Columbia River estuary, but the spread of invasive species and large-scale changes to the estuary influence how these habitats are utilized. Reed canarygrass (Phalaris arundinacea) (“PHAR” hereafter) is a highly invasive aquatic plant species that affects a number of emergent wetland habitats in the upper Columbia River estuary, and is of concern to those trying to restore important shallow-water habitats for juvenile salmon. The presence of PHAR undoubtedly reduces plant diversity, but this study seeked to elucidate how PHAR affects juvenile Chinook Salmon rearing success relative to natural emergent vegetation. The study further evaluated the importance of wetland-deriver prey to juveniles found in shallow-water habitats along a habitat gradient (i.e., back-water channel, confluence, and main stem). We conducted research at a floodplain wetland restoration site in the tidal-fluvial portion of the estuary, where a water control structure is used to manage water levels and the spread of PHAR. The goal of the research was to compare two microhabitats within the floodplain: areas dominated by PHAR and areas dominated by natural emergent vegetation. Using hatchery-raised juvenile Chinook salmon, we designed a series of feeding experiments in artificial enclosures to determine whether invasive PHAR and natural emergent vegetation provide similar foraging and growth opportunities for juvenile salmon. We further identified differences in physical and biological parameters of each habitat (e.g., temperature, dissolved oxygen, and invertebrate community) and examined feeding characteristics (e.g., diet composition and modeled growth rates) for juveniles through a stomach content analysis. Additionally, we conducted beach seining and diet analysis for fish collected along a habitat gradient and compared the diet composition and modeled growth rates between the sample areas. In the floodplain wetland, invertebrate prey compositions from the 2015-2016 (March- June) fallout and emergence traps were similar, but prey abundances and diversity varied seasonally. Across both years, for emergence traps, the average total density (per m²) of invertebrates in the natural vegetation and in PHAR was 154±18 m² and 225±34 m², respectively, and for the fallout traps was 664±95 m² and 662±64 m², respectively. Salmon diets were mostly similar, and a large proportion of diets in both habitats were made up of Copepoda/Cladocera (60-98% composition by biomass). Growth during the net pen experiment differed significantly between the two vegetation types (Kruskal-Wallis: p<0.001), with fish growing more in the natural emergent vegetation. In 2015, juveniles grew an average 6.4 mm FL in the natural emergent vegetation, compared to 4.7 mm FL in PHAR, and consumed fewer Copepoda/Cladocera and more Diptera over the 10 days. In 2016, juveniles grew an average 9.2 mm FL in the natural emergent vegetation compared to 7.6 mm FL in PHAR. In 2016, fish from both habitats consumed a higher abundance of Copepoda/Cladocera and grew more, despite poor water quality conditions at the site. Of the subyearlings (n=170) and yearlings (n=14) collected in habitats adjacent to the floodplain wetland, 48% were of known hatchery origin. Modeled potential growth rates for subyearlings were similar on average between February and May in the back channel, confluence, and main stem, ranging from 0.067–0.07 g/g/d, but the rates varied seasonally. Growth potential was higher in the back channel habitat earlier in the sampling season (February-March), as the modeled daily growth rates were 0.08 g/g/d compared to 0.06 g/g/d in the main stem. All fish collected prior to May were relatively small and unclipped, highlighting the seasonal benefits of shallow-water habitats to support a variety of life-history traits. When comparing modeled growth rates of the juveniles reared at the floodplain wetland compared to the in adjacent habitats, estimates from the floodplain wetland were higher across all months. Restoration of shallow-water habitats is important for the overall health of the estuary, but the effects of large-scale ecosystem changes (e.g., flow regulation and spread invasive species) on salmon recovery and estuary-management decisions are uncertain. For example, water control structures to limit the spread of PHAR also reduce fish access to floodplain habitats, and fewer high-flow events further limit fish access and may facilitate the spread of invasive species. PHAR has been shown to reduce overall plant diversity and, as observed in this study, may provide habitat that is less suitable to juvenile Chinook Salmon. The high density of important prey for salmon found at the site, regardless of the presence of PHAR, and high potential growth measured at the site, emphasize the importance of restoring shallow-water habitats that are accessible to juvenile Chinook Salmon in the upper estuary. Title: Restoring the Columbia River Estuary : Chinook Salmon Recovery and Invasive Species Management Author: Klopfenstein, Rachael As highlighted in this study, shallow-water habitats and the prey they provide are important for juvenile salmon in the Columbia River estuary, but the spread of invasive species and large-scale changes to the estuary influence how these habitats are utilized. Reed canarygrass (Phalaris arundinacea) (“PHAR” hereafter) is a highly invasive aquatic plant species that affects a number of emergent wetland habitats in the upper Columbia River estuary, and is of concern to those trying to restore important shallow-water habitats for juvenile salmon. The presence of PHAR undoubtedly reduces plant diversity, but this study seeked to elucidate how PHAR affects juvenile Chinook Salmon rearing success relative to natural emergent vegetation. The study further evaluated the importance of wetland-deriver prey to juveniles found in shallow-water habitats along a habitat gradient (i.e., back-water channel, confluence, and main stem). We conducted research at a floodplain wetland restoration site in the tidal-fluvial portion of the estuary, where a water control structure is used to manage water levels and the spread of PHAR. The goal of the research was to compare two microhabitats within the floodplain: areas dominated by PHAR and areas dominated by natural emergent vegetation. Using hatchery-raised juvenile Chinook salmon, we designed a series of feeding experiments in artificial enclosures to determine whether invasive PHAR and natural emergent vegetation provide similar foraging and growth opportunities for juvenile salmon. We further identified differences in physical and biological parameters of each habitat (e.g., temperature, dissolved oxygen, and invertebrate community) and examined feeding characteristics (e.g., diet composition and modeled growth rates) for juveniles through a stomach content analysis. Additionally, we conducted beach seining and diet analysis for fish collected along a habitat gradient and compared the diet composition and modeled growth rates between the sample areas. In the floodplain wetland, invertebrate prey compositions from the 2015-2016 (March- June) fallout and emergence traps were similar, but prey abundances and diversity varied seasonally. Across both years, for emergence traps, the average total density (per m²) of invertebrates in the natural vegetation and in PHAR was 154±18 m² and 225±34 m², respectively, and for the fallout traps was 664±95 m² and 662±64 m², respectively. Salmon diets were mostly similar, and a large proportion of diets in both habitats were made up of Copepoda/Cladocera (60-98% composition by biomass). Growth during the net pen experiment differed significantly between the two vegetation types (Kruskal-Wallis: p<0.001), with fish growing more in the natural emergent vegetation. In 2015, juveniles grew an average 6.4 mm FL in the natural emergent vegetation, compared to 4.7 mm FL in PHAR, and consumed fewer Copepoda/Cladocera and more Diptera over the 10 days. In 2016, juveniles grew an average 9.2 mm FL in the natural emergent vegetation compared to 7.6 mm FL in PHAR. In 2016, fish from both habitats consumed a higher abundance of Copepoda/Cladocera and grew more, despite poor water quality conditions at the site. Of the subyearlings (n=170) and yearlings (n=14) collected in habitats adjacent to the floodplain wetland, 48% were of known hatchery origin. Modeled potential growth rates for subyearlings were similar on average between February and May in the back channel, confluence, and main stem, ranging from 0.067–0.07 g/g/d, but the rates varied seasonally. Growth potential was higher in the back channel habitat earlier in the sampling season (February-March), as the modeled daily growth rates were 0.08 g/g/d compared to 0.06 g/g/d in the main stem. All fish collected prior to May were relatively small and unclipped, highlighting the seasonal benefits of shallow-water habitats to support a variety of life-history traits. When comparing modeled growth rates of the juveniles reared at the floodplain wetland compared to the in adjacent habitats, estimates from the floodplain wetland were higher across all months. Restoration of shallow-water habitats is important for the overall health of the estuary, but the effects of large-scale ecosystem changes (e.g., flow regulation and spread invasive species) on salmon recovery and estuary-management decisions are uncertain. For example, water control structures to limit the spread of PHAR also reduce fish access to floodplain habitats, and fewer high-flow events further limit fish access and may facilitate the spread of invasive species. PHAR has been shown to reduce overall plant diversity and, as observed in this study, may provide habitat that is less suitable to juvenile Chinook Salmon. The high density of important prey for salmon found at the site, regardless of the presence of PHAR, and high potential growth measured at the site, emphasize the importance of restoring shallow-water habitats that are accessible to juvenile Chinook Salmon in the upper estuary. Close

• 3314. [Article] Factors affecting within-season and between-season breeding dispersal of burrowing owls in California

  Dispersal is integral to our understanding of the life history and population biology of many vertebrates, but difficulties in detecting long distance movements have complicated its study. Moreover, studies ...

  Citation × Citation Citation Abstract Copy Title: Factors affecting within-season and between-season breeding dispersal of burrowing owls in California Author: Catlin, Daniel H. Dispersal is integral to our understanding of the life history and population biology of many vertebrates, but difficulties in detecting long distance movements have complicated its study. Moreover, studies of factors affecting dispersal are often unable to determine the relative contributions of variables such as nesting success, mate fidelity, and nest site fidelity. I examined the effects of nest depredation on dispersal in comparison to successful nests and nests that failed for other reasons. Additionally, I investigated a suite of biological factors affecting within-season and between-season breeding dispersal by burrowing owls (Athene cunicularia) in California, attempting to partition the effects of these covariates and to deal with long distance detectibility issues. For both types of dispersal, I divided dispersal into two components; dispersal probability and dispersal distance. I used experimental and observational approaches to investigate within-season dispersal in two contrasting environments; a large grassland and an agricultural landscape. I found that the factors affecting dispersal probability and dispersal distance were different, supporting my decision to examine each separately. Of the factors investigated, dispersal probability was influenced most by study area, mate fidelity, and nesting success. The proportion of individuals dispersing tended to be greater for owls that lost their mate due to death or dispersal (60%, 6 of 10) and owls whose nests were depredated (50%, 10 of 20) than for owls that did not lose their mates (33%, 6 of 18) and owls whose nests were successful (17%, 1 of 6), respectively. The results from an experiment where we removed eggs from pairs of owls to simulate nest depredation were consistent with the observational results, suggesting that owls whose nests were depredated may have been more likely to disperse than control owls. The reactions of owls from depredated nests, however, did not appear to differ from those whose nests failed for other reasons. In contrast, owl dispersal distance was most affected by owl gender, and to a lesser degree by study area and nesting success. Dispersal distance was greater for female owls (median = 1575 m, n = 13) than male owls (median = 417 m, n = 11), greater for owls from the grassland area (median= 939 m, n = 9) compared to owls from the agricultural area (median = 829 m, n = 15), and greater for owls whose nests had failed (median = 1018 m, n = 17) than for owls that successfully bred (median 475 m, n = 7). Nest depredation, however, did not appear to increase dispersal distance. The geometric models performed poorly at approximating within-season dispersal distance, indicating that many owls disperse farther than predicted by a "first is best" model. I speculate that the distribution of within-season dispersal distances by burrowing owls is related to the densities of suitable territories and mates, which are more variable than predicted by a geometric model within a breeding season. I used data from band resightings and nesting success (1998-2003) to examine factors related to between-season breeding dispersal by burrowing owls in an agricultural environment. Of the factors investigated, nesting success appeared to have the greatest effect on burrowing owl dispersal. The proportion of individuals dispersing was greater for owls whose nests had failed (68%, 28 of 41) than owls whose nests were successful (27%, 58 of 212). Similarly, dispersal distance was greater for owls whose nest failed (mean = 745 ± 175 m, n = 28) than owls with successful nests (mean = 340 ± 36 m,n = 58). The owls exhibited high rates of nest site and mate fidelity between breeding seasons. There was evidence that previous experience at a breeding site may have reduced dispersal probability and that unpaired owls may have been more likely to disperse and dispersed slightly greater distances than those that retained their mates. Nesting success, however, appeared to be the major factor contributing to burrowing owl breeding dispersal after controlling for nest site and mate fidelity, particularly for male owls. Despite the complexity of the dispersal process, a geometric model provided a reasonably good fit to the distribution of between-season breeding dispersal distances at relatively short distances, but failed to predict a small percentage of long distance dispersals. Geometric models appeared to be a better fit for the distribution of between-season breeding dispersal distances than within-season breeding dispersal distances. Factors affecting within-season dispersal were generally similar to those affecting between-season dispersal. Both within-season and between-season breeding dispersal were affected by nesting success and mate fidelity, but the effects of these factors differed between the two types of breeding dispersal, suggesting that time constraints and competition play a larger role in within-season dispersal than between season dispersal. In addition, both studies supported a difference in dispersal behavior, in which the factors that affected dispersal probability were distinct from those that affected dispersal distance. These results help determine the relative contributions of nesting success, mate fidelity, and nest site fidelity to avian dispersal, offer some evidence that the effects of nest depredation are not distinct from the effects of nest failure in general, and provide further support for the division of dispersal into dispersal probability and dispersal distance. Title: Factors affecting within-season and between-season breeding dispersal of burrowing owls in California Author: Catlin, Daniel H. Dispersal is integral to our understanding of the life history and population biology of many vertebrates, but difficulties in detecting long distance movements have complicated its study. Moreover, studies of factors affecting dispersal are often unable to determine the relative contributions of variables such as nesting success, mate fidelity, and nest site fidelity. I examined the effects of nest depredation on dispersal in comparison to successful nests and nests that failed for other reasons. Additionally, I investigated a suite of biological factors affecting within-season and between-season breeding dispersal by burrowing owls (Athene cunicularia) in California, attempting to partition the effects of these covariates and to deal with long distance detectibility issues. For both types of dispersal, I divided dispersal into two components; dispersal probability and dispersal distance. I used experimental and observational approaches to investigate within-season dispersal in two contrasting environments; a large grassland and an agricultural landscape. I found that the factors affecting dispersal probability and dispersal distance were different, supporting my decision to examine each separately. Of the factors investigated, dispersal probability was influenced most by study area, mate fidelity, and nesting success. The proportion of individuals dispersing tended to be greater for owls that lost their mate due to death or dispersal (60%, 6 of 10) and owls whose nests were depredated (50%, 10 of 20) than for owls that did not lose their mates (33%, 6 of 18) and owls whose nests were successful (17%, 1 of 6), respectively. The results from an experiment where we removed eggs from pairs of owls to simulate nest depredation were consistent with the observational results, suggesting that owls whose nests were depredated may have been more likely to disperse than control owls. The reactions of owls from depredated nests, however, did not appear to differ from those whose nests failed for other reasons. In contrast, owl dispersal distance was most affected by owl gender, and to a lesser degree by study area and nesting success. Dispersal distance was greater for female owls (median = 1575 m, n = 13) than male owls (median = 417 m, n = 11), greater for owls from the grassland area (median= 939 m, n = 9) compared to owls from the agricultural area (median = 829 m, n = 15), and greater for owls whose nests had failed (median = 1018 m, n = 17) than for owls that successfully bred (median 475 m, n = 7). Nest depredation, however, did not appear to increase dispersal distance. The geometric models performed poorly at approximating within-season dispersal distance, indicating that many owls disperse farther than predicted by a "first is best" model. I speculate that the distribution of within-season dispersal distances by burrowing owls is related to the densities of suitable territories and mates, which are more variable than predicted by a geometric model within a breeding season. I used data from band resightings and nesting success (1998-2003) to examine factors related to between-season breeding dispersal by burrowing owls in an agricultural environment. Of the factors investigated, nesting success appeared to have the greatest effect on burrowing owl dispersal. The proportion of individuals dispersing was greater for owls whose nests had failed (68%, 28 of 41) than owls whose nests were successful (27%, 58 of 212). Similarly, dispersal distance was greater for owls whose nest failed (mean = 745 ± 175 m, n = 28) than owls with successful nests (mean = 340 ± 36 m,n = 58). The owls exhibited high rates of nest site and mate fidelity between breeding seasons. There was evidence that previous experience at a breeding site may have reduced dispersal probability and that unpaired owls may have been more likely to disperse and dispersed slightly greater distances than those that retained their mates. Nesting success, however, appeared to be the major factor contributing to burrowing owl breeding dispersal after controlling for nest site and mate fidelity, particularly for male owls. Despite the complexity of the dispersal process, a geometric model provided a reasonably good fit to the distribution of between-season breeding dispersal distances at relatively short distances, but failed to predict a small percentage of long distance dispersals. Geometric models appeared to be a better fit for the distribution of between-season breeding dispersal distances than within-season breeding dispersal distances. Factors affecting within-season dispersal were generally similar to those affecting between-season dispersal. Both within-season and between-season breeding dispersal were affected by nesting success and mate fidelity, but the effects of these factors differed between the two types of breeding dispersal, suggesting that time constraints and competition play a larger role in within-season dispersal than between season dispersal. In addition, both studies supported a difference in dispersal behavior, in which the factors that affected dispersal probability were distinct from those that affected dispersal distance. These results help determine the relative contributions of nesting success, mate fidelity, and nest site fidelity to avian dispersal, offer some evidence that the effects of nest depredation are not distinct from the effects of nest failure in general, and provide further support for the division of dispersal into dispersal probability and dispersal distance. Close

• 3315. [Article] Distribution and movements of Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin

  Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin and other large river systems in western Alaska have declined dramatically since the late 1990s. This continuing trend has ...

  Citation × Citation Citation Abstract Copy Title: Distribution and movements of Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin Author: Eiler, John H. Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin and other large river systems in western Alaska have declined dramatically since the late 1990s. This continuing trend has raised concerns over the future status of the returns, and severely impacted commercial and subsistence fisheries within the drainage. Management is further complicated by the mixed-stock composition of the run, the presence of other temporally similar salmon species, and the need to equitably allocate harvests between the numerous fisheries and user groups scattered throughout the basin. Detailed information is needed on Chinook salmon run characteristics to better understand and manage the returns, and facilitate conservation efforts. However, this goal is exacerbated by the massive size and remote nature of the basin, the large number of highly mobile fish, and the compressed timing of the run. To address these challenges, radio telemetry was used to determine the stock composition and spawning distribution of the returns, and the migratory characteristics of the fish. The migratory patterns exhibited by returning salmon provide a number of insights into the status of the run. Since the Yukon River is essentially free-flowing (i.e., not regulated), this study also presented an opportunity to document the distribution and upriver movements of large returns of wild Chinook salmon under natural conditions. During 2002-2004, returning adult Chinook salmon were captured in the lower Yukon River (approximately 300 km upriver from the river mouth), tagged with radio transmitters, and tracked upriver using remote tracking stations located on important migratory routes and major spawning tributaries. Aerial tracking surveys were used to locate fish in spawning areas and between stations. The fish responded well to the capture and handling procedures, with most (2,790, 98%) resuming upriver movements. Although the fish initially displayed a negative tagging response, with slower migration rates observed immediately after release, the duration of this response was relatively short (several days) and less severe as the fish moved upriver. Independent measures indicated that the swimming speeds and timing of the fish upriver from the tagging area were comparable to untagged fish, suggesting that the tagging methods used were relatively benign. Fish returned to spawning areas throughout the basin, ranging from several hundred to over 3,000 km from the tagging area. Distribution patterns were similar across years, suggesting that the principal components of the run were identified. Most spawning fish were clustered in a number of key tributaries, with smaller numbers of fish located in other spatially isolated areas. The fish typically returned to clear water tributaries that were relatively entrenched, had moderate gradients, and were associated with upland areas. Fish were largely absent in lowland reaches characterized by meandering, low gradient, highly alluvial channels often associated with main river floodplains. There was suggestive evidence of mainstem spawning in reaches of the Upper Yukon. The status of fish remaining in other mainstem areas was less certain, and may represent local spawning activity or fish that died while in-transit to upriver areas. Although Chinook salmon spawned throughout the basin, the run was dominated by two regional components (Tanana and Upper Yukon), which annually comprised over 70% of the return. Substantially fewer fish returned to other areas ranging from 2-9% of the return, although the collective contribution of these stocks was appreciable. Most regional returns consisted of several principal stocks and a number of small, spatially isolated populations. Regional and stock composition estimates were similar across years even though differences in run abundance were reported, suggesting that these abundance differences were not related to regional or stock-specific differences. Run timing was relatively compressed compared to rivers in the southern portion of the range, with most stocks passing through the lower river over a 6-week period, ranging from 16 to 38 d. Run timing was generally earlier for stocks traveling farther upriver, although exceptions were noted. Lower basin stocks were primarily later run fish. Pronounced differences were observed in the migration rates (km/d) exhibited by regional stocks. Substantially slower swimming speeds were observed for fish returning to terminal tributaries in the lower basin ranging from 28-40 km/d compared to 52-62 km/d for upper basin stocks. The migratory patterns (migration rates in sequential reaches) of the fish also showed distinct regional differences. Average migration rates through the lower river were remarkably similar for the different stocks, ranging from 57-62 km/d, with most stocks exhibiting a general decline as the fish moved farther upriver. Tanana River stocks displayed a pronounced reduction in swimming speed after leaving the Yukon River main stem, with migration rates declining to 24 km/d on average as the fish approached their terminal tributaries. Conversely, upper basin stocks exhibited a relatively gradual (but variable) overall decline in migration rate even though these fish were traveling substantially greater distances upriver. Average migration rates for upper basin stocks ranged from 43-61 km/d as the fish approached their terminal tributaries. There was substantial variation in the migratory patterns exhibited by individual fish, although these patterns tended to be similar to the patterns exhibited by the regional stocks, particularly as the fish moved farther upriver from the tagging area. The dominant source of variation among fish reflected the average migration rate, with individual fish traveling slower in the lower basin exhibiting consistently slower migration rates as they moved upriver compared to their faster moving counterparts. This migratory pattern was consistent across stocks, and on average explained 74% of the within-stock variation in migration rate represented by the multivariate data. The second source of variation in migration rate reflected a shift in the relative swimming speeds of the individual fish as they progressed upriver. Although movement rates declined for nearly all of the fish during the migration, differences were observed in the pattern of the decline. Fish with faster migration rates in the lower river exhibited a pronounced decline in swimming speed as they moved upriver, whereas fish moving slower in the lower river displayed a more gradual decline in migration rate. On average, this migratory pattern explained 22% of the within-stock variation in migration rate represented by the multivariate data. Most fish (98%) exhibited continuous upriver movements and strong fidelity to the rivers they entered. However a small number of fish (n = 66) deviated from this pattern. Some of these individuals initially passed their final destination and continued upriver for varying distances before reversing direction, swimming back downstream, and entering their terminal tributary. Although most of these excursions were relatively short (< 30 km), there were several instances where fish traveled hundreds of kilometers out of their way. Thirty-four fish tracked to terminal tributaries subsequently left these rivers, and traveled to other terminal tributaries within the basin (n = 31) or were harvested in upriver fisheries (n = 3). Although most of these incidents involved nearby tributaries, major diversions were also observed, with several fish traveling over 300 km to natal rivers after leaving the initial tributary. Chinook salmon returns to the Yukon River typically consisted of a series of distinct and sizable increases in the number fish entering the river over the course of the run, commonly referred to as pulses. A large number of fish (n = 251) were radio tagged over a 4-day period during a pulse in 2003 to provide information on the progression of the pulse as it moved upriver. The time taken by the pulse to move past subsequent upriver locations increased as the fish moved farther upriver from the tagging area, with the fish passing sites located 580 and 800 km upriver over a span of 14 and 21 d, respectively. Although not surprising considering the extensive variation in migration rates observed among individual fish, this finding does suggest that these pulses do not represent cohesive aggregates of fish moving upriver. Unlike the well established methods used to estimate other life history characteristics, the development of quantitative methods for analyzing and modeling fish movements has lagged noticeably behind, due in part to the complexity associated with movement data and (prior to the advent of telemetry) the difficulty of collecting this type of information on free-ranging individuals. Two fundamentally different analytical approaches, hierarchical linear regression models and multivariate ordination, were used during this study to evaluate factors thought to influence the upriver movements of the fish. In spite of the inherent differences, both methods provided strikingly similar results, indicating that the study findings were not dependent on the approach used, and suggesting that the results were plausible based on the information available and the weight of evidence. Both analytical methods had advantages, and provided complementary information. With hierarchical linear models, it was possible to simultaneously evaluate a wide range of explanatory variables (in our case, both biological and environmental), which provided standardized comparisons and simplified the interpretation of the results. Since both fixed and random effects were incorporated in the models, it was possible to account for sources of variation when insufficient information was available to identify the underlining factors – an important consideration since few field studies provide comprehensive data. With multivariate ordination, separate analyzes were needed to examine the relationships between the migration rates and the biotic and physical variables. In addition to being cumbersome, this limitation made it more difficult to compare the relative influence of the different factors and interactions between factors. However, ordination was very useful as an exploratory tool. Although compartmentalized by stock, across fish comparisons were simple and relatively straightforward. Because the explanatory variables were evaluated separately in relation to the ordination score assigned to the fish, it was possible to examine and compare highly correlated variables. Ordination was also able to identify overall patterns within the data and assess the relative importance. While this can be accomplished within the framework of linear regression using mixture models to determine whether multiple distributions exist within the data, the process is much simpler with ordination. The migratory patterns of the fish were influenced by a wide range of factors, with evidentiary support for complex, multi-faceted relationships. Physical features of the basin demonstrated stronger explanatory power, accounting for over 70% of the observed variation in migration rate compared to 18% for the biological characteristics of the fish. Parameter estimates associated with the steepness of the migratory route and remaining distance the fish had to travel to reach their natal rivers were most strongly correlated with migration rate, with consistent relationships observed across stocks. Migration rates were also noticeably slower in extensively braided reaches of the basin. The weaker relationships between migration rate and biotic factors may reflect stabilizing selection on long-distance migrants. Smaller fish exhibited minimally faster swimming speeds on average than larger individuals. This relationship was stronger in highly braided reaches. Run timing was positively related to migration rate for most stocks. Surprisingly, upper basin stocks traveling farther upriver displayed progressively negative relationships, suggesting that late-run fish were moving slower. Ancillary information suggests that this decline may relate to deteriorating fish condition later in the season. Title: Distribution and movements of Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin Author: Eiler, John H. Chinook salmon, Oncorhynchus tshawytscha, returning to the Yukon River basin and other large river systems in western Alaska have declined dramatically since the late 1990s. This continuing trend has raised concerns over the future status of the returns, and severely impacted commercial and subsistence fisheries within the drainage. Management is further complicated by the mixed-stock composition of the run, the presence of other temporally similar salmon species, and the need to equitably allocate harvests between the numerous fisheries and user groups scattered throughout the basin. Detailed information is needed on Chinook salmon run characteristics to better understand and manage the returns, and facilitate conservation efforts. However, this goal is exacerbated by the massive size and remote nature of the basin, the large number of highly mobile fish, and the compressed timing of the run. To address these challenges, radio telemetry was used to determine the stock composition and spawning distribution of the returns, and the migratory characteristics of the fish. The migratory patterns exhibited by returning salmon provide a number of insights into the status of the run. Since the Yukon River is essentially free-flowing (i.e., not regulated), this study also presented an opportunity to document the distribution and upriver movements of large returns of wild Chinook salmon under natural conditions. During 2002-2004, returning adult Chinook salmon were captured in the lower Yukon River (approximately 300 km upriver from the river mouth), tagged with radio transmitters, and tracked upriver using remote tracking stations located on important migratory routes and major spawning tributaries. Aerial tracking surveys were used to locate fish in spawning areas and between stations. The fish responded well to the capture and handling procedures, with most (2,790, 98%) resuming upriver movements. Although the fish initially displayed a negative tagging response, with slower migration rates observed immediately after release, the duration of this response was relatively short (several days) and less severe as the fish moved upriver. Independent measures indicated that the swimming speeds and timing of the fish upriver from the tagging area were comparable to untagged fish, suggesting that the tagging methods used were relatively benign. Fish returned to spawning areas throughout the basin, ranging from several hundred to over 3,000 km from the tagging area. Distribution patterns were similar across years, suggesting that the principal components of the run were identified. Most spawning fish were clustered in a number of key tributaries, with smaller numbers of fish located in other spatially isolated areas. The fish typically returned to clear water tributaries that were relatively entrenched, had moderate gradients, and were associated with upland areas. Fish were largely absent in lowland reaches characterized by meandering, low gradient, highly alluvial channels often associated with main river floodplains. There was suggestive evidence of mainstem spawning in reaches of the Upper Yukon. The status of fish remaining in other mainstem areas was less certain, and may represent local spawning activity or fish that died while in-transit to upriver areas. Although Chinook salmon spawned throughout the basin, the run was dominated by two regional components (Tanana and Upper Yukon), which annually comprised over 70% of the return. Substantially fewer fish returned to other areas ranging from 2-9% of the return, although the collective contribution of these stocks was appreciable. Most regional returns consisted of several principal stocks and a number of small, spatially isolated populations. Regional and stock composition estimates were similar across years even though differences in run abundance were reported, suggesting that these abundance differences were not related to regional or stock-specific differences. Run timing was relatively compressed compared to rivers in the southern portion of the range, with most stocks passing through the lower river over a 6-week period, ranging from 16 to 38 d. Run timing was generally earlier for stocks traveling farther upriver, although exceptions were noted. Lower basin stocks were primarily later run fish. Pronounced differences were observed in the migration rates (km/d) exhibited by regional stocks. Substantially slower swimming speeds were observed for fish returning to terminal tributaries in the lower basin ranging from 28-40 km/d compared to 52-62 km/d for upper basin stocks. The migratory patterns (migration rates in sequential reaches) of the fish also showed distinct regional differences. Average migration rates through the lower river were remarkably similar for the different stocks, ranging from 57-62 km/d, with most stocks exhibiting a general decline as the fish moved farther upriver. Tanana River stocks displayed a pronounced reduction in swimming speed after leaving the Yukon River main stem, with migration rates declining to 24 km/d on average as the fish approached their terminal tributaries. Conversely, upper basin stocks exhibited a relatively gradual (but variable) overall decline in migration rate even though these fish were traveling substantially greater distances upriver. Average migration rates for upper basin stocks ranged from 43-61 km/d as the fish approached their terminal tributaries. There was substantial variation in the migratory patterns exhibited by individual fish, although these patterns tended to be similar to the patterns exhibited by the regional stocks, particularly as the fish moved farther upriver from the tagging area. The dominant source of variation among fish reflected the average migration rate, with individual fish traveling slower in the lower basin exhibiting consistently slower migration rates as they moved upriver compared to their faster moving counterparts. This migratory pattern was consistent across stocks, and on average explained 74% of the within-stock variation in migration rate represented by the multivariate data. The second source of variation in migration rate reflected a shift in the relative swimming speeds of the individual fish as they progressed upriver. Although movement rates declined for nearly all of the fish during the migration, differences were observed in the pattern of the decline. Fish with faster migration rates in the lower river exhibited a pronounced decline in swimming speed as they moved upriver, whereas fish moving slower in the lower river displayed a more gradual decline in migration rate. On average, this migratory pattern explained 22% of the within-stock variation in migration rate represented by the multivariate data. Most fish (98%) exhibited continuous upriver movements and strong fidelity to the rivers they entered. However a small number of fish (n = 66) deviated from this pattern. Some of these individuals initially passed their final destination and continued upriver for varying distances before reversing direction, swimming back downstream, and entering their terminal tributary. Although most of these excursions were relatively short (< 30 km), there were several instances where fish traveled hundreds of kilometers out of their way. Thirty-four fish tracked to terminal tributaries subsequently left these rivers, and traveled to other terminal tributaries within the basin (n = 31) or were harvested in upriver fisheries (n = 3). Although most of these incidents involved nearby tributaries, major diversions were also observed, with several fish traveling over 300 km to natal rivers after leaving the initial tributary. Chinook salmon returns to the Yukon River typically consisted of a series of distinct and sizable increases in the number fish entering the river over the course of the run, commonly referred to as pulses. A large number of fish (n = 251) were radio tagged over a 4-day period during a pulse in 2003 to provide information on the progression of the pulse as it moved upriver. The time taken by the pulse to move past subsequent upriver locations increased as the fish moved farther upriver from the tagging area, with the fish passing sites located 580 and 800 km upriver over a span of 14 and 21 d, respectively. Although not surprising considering the extensive variation in migration rates observed among individual fish, this finding does suggest that these pulses do not represent cohesive aggregates of fish moving upriver. Unlike the well established methods used to estimate other life history characteristics, the development of quantitative methods for analyzing and modeling fish movements has lagged noticeably behind, due in part to the complexity associated with movement data and (prior to the advent of telemetry) the difficulty of collecting this type of information on free-ranging individuals. Two fundamentally different analytical approaches, hierarchical linear regression models and multivariate ordination, were used during this study to evaluate factors thought to influence the upriver movements of the fish. In spite of the inherent differences, both methods provided strikingly similar results, indicating that the study findings were not dependent on the approach used, and suggesting that the results were plausible based on the information available and the weight of evidence. Both analytical methods had advantages, and provided complementary information. With hierarchical linear models, it was possible to simultaneously evaluate a wide range of explanatory variables (in our case, both biological and environmental), which provided standardized comparisons and simplified the interpretation of the results. Since both fixed and random effects were incorporated in the models, it was possible to account for sources of variation when insufficient information was available to identify the underlining factors – an important consideration since few field studies provide comprehensive data. With multivariate ordination, separate analyzes were needed to examine the relationships between the migration rates and the biotic and physical variables. In addition to being cumbersome, this limitation made it more difficult to compare the relative influence of the different factors and interactions between factors. However, ordination was very useful as an exploratory tool. Although compartmentalized by stock, across fish comparisons were simple and relatively straightforward. Because the explanatory variables were evaluated separately in relation to the ordination score assigned to the fish, it was possible to examine and compare highly correlated variables. Ordination was also able to identify overall patterns within the data and assess the relative importance. While this can be accomplished within the framework of linear regression using mixture models to determine whether multiple distributions exist within the data, the process is much simpler with ordination. The migratory patterns of the fish were influenced by a wide range of factors, with evidentiary support for complex, multi-faceted relationships. Physical features of the basin demonstrated stronger explanatory power, accounting for over 70% of the observed variation in migration rate compared to 18% for the biological characteristics of the fish. Parameter estimates associated with the steepness of the migratory route and remaining distance the fish had to travel to reach their natal rivers were most strongly correlated with migration rate, with consistent relationships observed across stocks. Migration rates were also noticeably slower in extensively braided reaches of the basin. The weaker relationships between migration rate and biotic factors may reflect stabilizing selection on long-distance migrants. Smaller fish exhibited minimally faster swimming speeds on average than larger individuals. This relationship was stronger in highly braided reaches. Run timing was positively related to migration rate for most stocks. Surprisingly, upper basin stocks traveling farther upriver displayed progressively negative relationships, suggesting that late-run fish were moving slower. Ancillary information suggests that this decline may relate to deteriorating fish condition later in the season. Close

• 3316. [Article] The effects of habitat, climate, and Barred Owls on long-term demography of Northern Spotted Owls

  To the best of our knowledge, one or more authors of this paper were federal employees when contributing to this work. This is the publisher’s final pdf. The article is copyrighted by Cooper Ornithological ...

  Citation × Citation Citation Abstract Copy Title: The effects of habitat, climate, and Barred Owls on long-term demography of Northern Spotted Owls Author: Franklin, Alan B., Diller, Lowell V., Horn, Rob B., Schwarz, Carl J., McCafferty, Christopher, Gremel, Scott A., Biswell, Brian L., Clark, Darren A., Herter, Dale R., Doherty, Paul F. Jr, Hines, James E., McDonnell, Kevin, Bailey, Larissa, Higley, J. Mark, Yackulic, Charles B., Hobson, Jeremy, Burnham, Kenneth P., Dugger, Katie M., Carlson, Peter C., Clement, Matthew J., Davis, Raymond J., Rockweit, Jeremy, Huyvaert, Kathryn P., Andrews, Lawrence S., Forsman, Eric D., Augustine, Benjamin, Saenz, Jessica, Glenn, Elizabeth M., Blakesley, Jennifer, Olson, Gail S., Ackers, Steven H., McDonald, Trent, Nichols, James D., Green, Adam, Ruiz, Viviana, Sovern, Stan G., White, Gary C., Reid, Janice A. To the best of our knowledge, one or more authors of this paper were federal employees when contributing to this work. This is the publisher’s final pdf. The article is copyrighted by Cooper Ornithological Society and published by Central Ornithology Publication Office. It can be found at: http://www.aoucospubs.org/loi/cond Title: The effects of habitat, climate, and Barred Owls on long-term demography of Northern Spotted Owls Author: Franklin, Alan B., Diller, Lowell V., Horn, Rob B., Schwarz, Carl J., McCafferty, Christopher, Gremel, Scott A., Biswell, Brian L., Clark, Darren A., Herter, Dale R., Doherty, Paul F. Jr, Hines, James E., McDonnell, Kevin, Bailey, Larissa, Higley, J. Mark, Yackulic, Charles B., Hobson, Jeremy, Burnham, Kenneth P., Dugger, Katie M., Carlson, Peter C., Clement, Matthew J., Davis, Raymond J., Rockweit, Jeremy, Huyvaert, Kathryn P., Andrews, Lawrence S., Forsman, Eric D., Augustine, Benjamin, Saenz, Jessica, Glenn, Elizabeth M., Blakesley, Jennifer, Olson, Gail S., Ackers, Steven H., McDonald, Trent, Nichols, James D., Green, Adam, Ruiz, Viviana, Sovern, Stan G., White, Gary C., Reid, Janice A. To the best of our knowledge, one or more authors of this paper were federal employees when contributing to this work. This is the publisher’s final pdf. The article is copyrighted by Cooper Ornithological Society and published by Central Ornithology Publication Office. It can be found at: http://www.aoucospubs.org/loi/cond Close

• 3317. [Article] Contemporary regional forest dynamics in the Pacific Northwest

  Recent climatic warming trends and increases in the frequency and extent of wildfires have prompted much concern regarding the potential for rapid change in the structure and function of forested ecosystems ...

  Citation × Citation Citation Abstract Copy Title: Contemporary regional forest dynamics in the Pacific Northwest Author: Reilly, Matthew J. (Matthew Justin), 1975- Recent climatic warming trends and increases in the frequency and extent of wildfires have prompted much concern regarding the potential for rapid change in the structure and function of forested ecosystems around the world. Episodes of mortality in wildfires and insect outbreaks associated with drought have affected large areas and altered landscapes, but little is known about the cumulative effects of these disturbances at the regional scales. I used data from two different forest inventories in the Pacific Northwest to develop a framework for tracking regional forest dynamics and examine variation in tree mortality rates among vegetation zones that differ in biophysical setting as well as recent and historical disturbance regimes. In the second chapter I developed an empirically based framework for tracking regional forest dynamics using regional inventory data collected from 2001 to 2010. I characterized the major dimensions of forest structure and developed a classification incorporating multiple attributes of forest structure including biomass, size, and density of live trees, the distribution and abundance of dead wood, and the cover of understory vegetation. A single dimension related to live tree biomass accounted for almost half of the variation in a principal components analysis of structural attributes, but dimensions related to density and size of live trees, dead wood, and understory vegetation accounted for as much additional variation. Snags and biomass of dead and downed wood were related to multiple dimensions while understory vegetation acted independent of other dimensions. Results indicated that structural development is more complex than a monotonic accumulation of live biomass and that some components act independently or emerge at multiple stages of structural development. The hierarchical classification reduced the data into three “groups” based on live tree biomass, followed by eleven "classes" that varied in density and size of live trees, and finally twenty-five structural types that differed further in the abundance of dead wood and cover of understory vegetation. Most structural types were geographically widespread but varied in age of dominant trees by vegetation zone indicating that similar structural conditions developed in environments with different biophysical setting, climate, and disturbance/successional histories. Low live biomass structural types (<25 Mg/ha) differed in live tree density and the abundance of live and dead legacies, demonstrating that the variation in early developmental stages depends on the rate of tree establishment and the nature and severity of recent disturbance. Forests in early developmental stages made up less than 20% of most vegetation zones and diverse types with live or dead legacies associated with wildfires were rare. Moderate live biomass structural types (25-99 Mg/ha) represented multiple mid, mature, and late developmental stages, some of which lack analogs in existing conceptual models of structural development such as lower density woodlands with big trees. These structural types included two that have high densities of snags indicative of recent episodes of mortality; together these made up as much as 10% of some dry vegetation zones. Several high live biomass structural types (100->300 Mg/ha) were identified and substantiated the diversity and relative dominance of mature and later developmental stages, particularly in wet vegetation zones. The relative abundance and make up of structural types varied widely by vegetation zone. Most forests in wet vegetation zones had moderate to high live biomass and were in mid and mature developmental stages, while diverse early developmental stage stages were extremely rare. Dry forests had a far greater range of variation in the relative abundance of structural types which is partially attributable to the greater range of climatic conditions they included, but also to the occurrence of recent episodes of mortality associated with wildfires and insects. In the third chapter I examined variation in tree mortality rates using a different regional inventory that occurred from the mid-1990s to the mid-2000s. I compared the distribution of rates among stands in different vegetation zones and stages of structural developmental. I developed a simple framework based on changes in live tree density and mean tree size and examined trends in structural change associated with disturbances at different levels of mortality across all stages of structural development. Most plots were within the range of "background" mortality rates reported in other studies (<1.0 %/yr) and extremely high "stand-replacing" levels of mortality (>25%/yr) were rare. Approximately 30% of plot mortality rates occurred at intermediate levels (>1%/yr and <25%/yr) as result of insects and fire, highlighting the importance of conceptualizing mortality as a continuum as opposed to just “background” or “stand-replacement” to fully represent dynamics at a regional scale. The distributions of mortality differed among many vegetation zones. Levels of mortality were primarily <2.5%/yr in western hemlock, silver fir, and mountain hemlock vegetation zones where fires were rare and insects and pathogens occurred predominantly at endemic levels. Rates were highest in subalpine forests and higher elevation grand fir and Douglas-fir forests as a result of fire and insects. Mortality rates in ponderosa pine, the hottest driest forest vegetation zone, were surprisingly low, and there was little to no mortality in plots with no evidence of disturbance. Mortality rates varied among developmental stages in all vegetation zones but few consistent patterns emerged. Levels of mortality were often lowest in early developmental stages but varied in later stages where they were lowest in wet vegetation zones and highest in subalpine and dry vegetation zones. Application of a simple framework indicated that multiple trajectories of structural change were common at levels of mortality <2.5%/yr, but structural change at higher levels was predominantly associated with a “thinning” trajectory defined by decreases in density and increases in mean tree size. Results indicated that the rate and magnitude of mortality related change during the study period varies widely across the region. Rapid change has occurred in subalpine, grand fir/white fir, Douglas-fir, and ponderosa pine vegetation zones where disturbances such as insects and fire were widespread. However, these disturbances have potentially restored some aspects of historical structure by reducing overall density and increasing the dominance of bigger trees. In western hemlock, silver fir, and mountain hemlock vegetation zones where higher levels of mortality related to disturbances were rare, wildfires have increased landscape diversity by creating diverse early successional habitats and most change was more subtle but may be manifest oevr longer periods if current trends continue. This examination of short-period mortality rates and associated structural change across a broad geographic provides context for understanding trends from localized studies and potential ecological consequences of mortality, but there is still a great deal of uncertainty as to how the effects of a changing climate and disturbance regimes will manifest themselves over longer time scales. This dissertation is one of the first field based assessments of recent forest dynamics at a regional scale. The results of both chapters, each based on a different dataset, told a similar story. The abundance of structural types in various vegetation zones estimated during the mid-2000s was consistent with the cumulative effects of tree mortality during the preceding decade. It was evident that wildfire effects and recent mortality were small relative to the regional extent of the study and have contributed to structural diversity and restoration of historic structure in stands where fire exclusion and past logging has increased total stand density and decreased the dominance of big trees. However, the rate of change and cumulative effects of recent forest dynamics varied widely by geographic location and vegetation zone and there was greater variability and uncertainty regarding the effects of mortality at smaller landscape scales where individual events like large wildfires have the potential to rapidly alter the landscape structure and composition. Assessing this variability and the scales at which trade-offs (e.g. losses of old-growth and creation of diverse early developmental stages) occur will be an important next step in understanding the cumulative ecological effects of recent wildfires and tree mortality on Pacific Northwest forests. Title: Contemporary regional forest dynamics in the Pacific Northwest Author: Reilly, Matthew J. (Matthew Justin), 1975- Recent climatic warming trends and increases in the frequency and extent of wildfires have prompted much concern regarding the potential for rapid change in the structure and function of forested ecosystems around the world. Episodes of mortality in wildfires and insect outbreaks associated with drought have affected large areas and altered landscapes, but little is known about the cumulative effects of these disturbances at the regional scales. I used data from two different forest inventories in the Pacific Northwest to develop a framework for tracking regional forest dynamics and examine variation in tree mortality rates among vegetation zones that differ in biophysical setting as well as recent and historical disturbance regimes. In the second chapter I developed an empirically based framework for tracking regional forest dynamics using regional inventory data collected from 2001 to 2010. I characterized the major dimensions of forest structure and developed a classification incorporating multiple attributes of forest structure including biomass, size, and density of live trees, the distribution and abundance of dead wood, and the cover of understory vegetation. A single dimension related to live tree biomass accounted for almost half of the variation in a principal components analysis of structural attributes, but dimensions related to density and size of live trees, dead wood, and understory vegetation accounted for as much additional variation. Snags and biomass of dead and downed wood were related to multiple dimensions while understory vegetation acted independent of other dimensions. Results indicated that structural development is more complex than a monotonic accumulation of live biomass and that some components act independently or emerge at multiple stages of structural development. The hierarchical classification reduced the data into three “groups” based on live tree biomass, followed by eleven "classes" that varied in density and size of live trees, and finally twenty-five structural types that differed further in the abundance of dead wood and cover of understory vegetation. Most structural types were geographically widespread but varied in age of dominant trees by vegetation zone indicating that similar structural conditions developed in environments with different biophysical setting, climate, and disturbance/successional histories. Low live biomass structural types (<25 Mg/ha) differed in live tree density and the abundance of live and dead legacies, demonstrating that the variation in early developmental stages depends on the rate of tree establishment and the nature and severity of recent disturbance. Forests in early developmental stages made up less than 20% of most vegetation zones and diverse types with live or dead legacies associated with wildfires were rare. Moderate live biomass structural types (25-99 Mg/ha) represented multiple mid, mature, and late developmental stages, some of which lack analogs in existing conceptual models of structural development such as lower density woodlands with big trees. These structural types included two that have high densities of snags indicative of recent episodes of mortality; together these made up as much as 10% of some dry vegetation zones. Several high live biomass structural types (100->300 Mg/ha) were identified and substantiated the diversity and relative dominance of mature and later developmental stages, particularly in wet vegetation zones. The relative abundance and make up of structural types varied widely by vegetation zone. Most forests in wet vegetation zones had moderate to high live biomass and were in mid and mature developmental stages, while diverse early developmental stage stages were extremely rare. Dry forests had a far greater range of variation in the relative abundance of structural types which is partially attributable to the greater range of climatic conditions they included, but also to the occurrence of recent episodes of mortality associated with wildfires and insects. In the third chapter I examined variation in tree mortality rates using a different regional inventory that occurred from the mid-1990s to the mid-2000s. I compared the distribution of rates among stands in different vegetation zones and stages of structural developmental. I developed a simple framework based on changes in live tree density and mean tree size and examined trends in structural change associated with disturbances at different levels of mortality across all stages of structural development. Most plots were within the range of "background" mortality rates reported in other studies (<1.0 %/yr) and extremely high "stand-replacing" levels of mortality (>25%/yr) were rare. Approximately 30% of plot mortality rates occurred at intermediate levels (>1%/yr and <25%/yr) as result of insects and fire, highlighting the importance of conceptualizing mortality as a continuum as opposed to just “background” or “stand-replacement” to fully represent dynamics at a regional scale. The distributions of mortality differed among many vegetation zones. Levels of mortality were primarily <2.5%/yr in western hemlock, silver fir, and mountain hemlock vegetation zones where fires were rare and insects and pathogens occurred predominantly at endemic levels. Rates were highest in subalpine forests and higher elevation grand fir and Douglas-fir forests as a result of fire and insects. Mortality rates in ponderosa pine, the hottest driest forest vegetation zone, were surprisingly low, and there was little to no mortality in plots with no evidence of disturbance. Mortality rates varied among developmental stages in all vegetation zones but few consistent patterns emerged. Levels of mortality were often lowest in early developmental stages but varied in later stages where they were lowest in wet vegetation zones and highest in subalpine and dry vegetation zones. Application of a simple framework indicated that multiple trajectories of structural change were common at levels of mortality <2.5%/yr, but structural change at higher levels was predominantly associated with a “thinning” trajectory defined by decreases in density and increases in mean tree size. Results indicated that the rate and magnitude of mortality related change during the study period varies widely across the region. Rapid change has occurred in subalpine, grand fir/white fir, Douglas-fir, and ponderosa pine vegetation zones where disturbances such as insects and fire were widespread. However, these disturbances have potentially restored some aspects of historical structure by reducing overall density and increasing the dominance of bigger trees. In western hemlock, silver fir, and mountain hemlock vegetation zones where higher levels of mortality related to disturbances were rare, wildfires have increased landscape diversity by creating diverse early successional habitats and most change was more subtle but may be manifest oevr longer periods if current trends continue. This examination of short-period mortality rates and associated structural change across a broad geographic provides context for understanding trends from localized studies and potential ecological consequences of mortality, but there is still a great deal of uncertainty as to how the effects of a changing climate and disturbance regimes will manifest themselves over longer time scales. This dissertation is one of the first field based assessments of recent forest dynamics at a regional scale. The results of both chapters, each based on a different dataset, told a similar story. The abundance of structural types in various vegetation zones estimated during the mid-2000s was consistent with the cumulative effects of tree mortality during the preceding decade. It was evident that wildfire effects and recent mortality were small relative to the regional extent of the study and have contributed to structural diversity and restoration of historic structure in stands where fire exclusion and past logging has increased total stand density and decreased the dominance of big trees. However, the rate of change and cumulative effects of recent forest dynamics varied widely by geographic location and vegetation zone and there was greater variability and uncertainty regarding the effects of mortality at smaller landscape scales where individual events like large wildfires have the potential to rapidly alter the landscape structure and composition. Assessing this variability and the scales at which trade-offs (e.g. losses of old-growth and creation of diverse early developmental stages) occur will be an important next step in understanding the cumulative ecological effects of recent wildfires and tree mortality on Pacific Northwest forests. Close

• 3318. [Article] Trophic cascades and large mammals in the Yellowstone ecosystem

  Reintroduction of wolves to Yellowstone National Park (YNP) in 1995-96 provided a rare opportunity to observe the response of an ecosystem to the return of a top predator, including possible reversal of ...

  Citation × Citation Citation Abstract Copy Title: Trophic cascades and large mammals in the Yellowstone ecosystem Author: Painter, Luke E. Reintroduction of wolves to Yellowstone National Park (YNP) in 1995-96 provided a rare opportunity to observe the response of an ecosystem to the return of a top predator, including possible reversal of decades of decline of aspen, cottonwood, and tall willows suppressed by intensive herbivory on elk winter ranges. To investigate changes in aspen stands in northern Yellowstone since the return of wolves, I compared browsing intensity and heights of young aspen in 87 randomly selected stands in 2012 to similar data collected in the same stands in 1997-98. I also measured the spatial density of elk and bison scat piles as an index to relative population densities, and used annual counts of elk to calculate trends in elk density. In 1998, browsing rates averaged 88%, heights were suppressed, and no tall saplings (≥200 cm) were found in sampling plots. In 2012, browsing rates in 2012 were much lower averaging 44%, and 28% of plots had at least one sapling ≥200 cm, tall enough to escape browsing and therefore more likely to survive to replace dying overstory trees. Heights of young aspen were inversely related to browsing intensity, but not significantly related to leader length, suggesting that differences in height were primarily due to differences in browsing, not factors related to productivity. Aspen recovery was patchy, possibly due in part to locally high elk or bison densities in some parts of the winter range. These results of reduced browsing with increased sapling recruitment were consistent with a trophic cascade from wolves to elk to aspen resulting in a widespread and spatially variable recovery of aspen stands. There was wide variation in browsing intensity and aspen height between sectors of the Yellowstone northern ungulate winter range (northern range). The east sector generally had lower rates of browsing and more stands with tall saplings than the central and west sectors, a pattern that matched recent trends in elk population densities. Only a small minority of stands in the west sector had tall saplings, consistent with higher elk densities in the west. Densities of elk in winter on the northern range recently have been highest in the northwest sector outside the park boundary, where elk benefit from lower wolf densities and milder winters. Aspen stands did not recover at a comparable range-wide elk density when elk were culled in the park in the 1950s and 1960s, suggesting that the influence of wolves may be an important factor in the recent redistribution and reduction of herbivory impacts by elk. To examine the relationship between elk and aspen outside of YNP, I assessed browsing intensity and sapling recruitment in 43 aspen stands in the Shoshone National Forest east of the park, compared to data collected in the same stands in 1997-98. As in northern YNP, results were consistent with a trophic cascade with reduced browsing and increased recruitment of aspen saplings, but aspen recovery was patchy. Elk densities were moderate to high in most of the area, suggesting that the partial aspen recovery may involve a behavioral response to predation or other factors resulting in local variation in browsing impacts. Livestock may also have limited aspen recruitment. Recovery of some aspen stands in the Shoshone National Forest may provide some of the first evidence of a trophic cascade from wolves to elk to aspen outside of a national park, a trophic cascade possibly weakened by the influence of another large herbivore (cattle). Like cattle, bison in northern Yellowstone may have an effect on woody browse plants. Bison have increased in number and may prevent recovery of some aspen stands in places of high bison density. I also examined browsing impacts of bison on willow and cottonwood in the Lamar Valley. To distinguish the effects of bison from those of elk, I compared browsing at different heights on tall willows, below and above the reach of bison. Because elk were absent from the area in summer when bison were present at high density, I also measured browsing that occurred in the summer. I found high rates of summer browsing, and growth of willows and cottonwoods was suppressed in the Lamar Valley. Above the reach of bison (>100 cm), growth was not suppressed and browsing rates were low, suggesting that these plant species have been released from suppression by elk but bison have compensated for some of the reduction in elk browsing. This study provided the first evidence of significant herbivory by bison of woody browse plants in Yellowstone, and revealed some of the complexity of the Yellowstone food web. In summary, these research results support the hypothesis of a trophic cascade resulting from large carnivore restoration and subsequent changes in elk population densities and distribution. The return of wolves may have combined with other factors such as changes in hunting and land ownership, and increased predation by bears, to result in large-scale shifts in the distribution of elk in northern Yellowstone and greatly reduced elk densities in some areas. If these trends continue, the result may be a new alternative state with lower elk densities, and potential for enhanced biodiversity through reduced herbivory of woody browse species. Title: Trophic cascades and large mammals in the Yellowstone ecosystem Author: Painter, Luke E. Reintroduction of wolves to Yellowstone National Park (YNP) in 1995-96 provided a rare opportunity to observe the response of an ecosystem to the return of a top predator, including possible reversal of decades of decline of aspen, cottonwood, and tall willows suppressed by intensive herbivory on elk winter ranges. To investigate changes in aspen stands in northern Yellowstone since the return of wolves, I compared browsing intensity and heights of young aspen in 87 randomly selected stands in 2012 to similar data collected in the same stands in 1997-98. I also measured the spatial density of elk and bison scat piles as an index to relative population densities, and used annual counts of elk to calculate trends in elk density. In 1998, browsing rates averaged 88%, heights were suppressed, and no tall saplings (≥200 cm) were found in sampling plots. In 2012, browsing rates in 2012 were much lower averaging 44%, and 28% of plots had at least one sapling ≥200 cm, tall enough to escape browsing and therefore more likely to survive to replace dying overstory trees. Heights of young aspen were inversely related to browsing intensity, but not significantly related to leader length, suggesting that differences in height were primarily due to differences in browsing, not factors related to productivity. Aspen recovery was patchy, possibly due in part to locally high elk or bison densities in some parts of the winter range. These results of reduced browsing with increased sapling recruitment were consistent with a trophic cascade from wolves to elk to aspen resulting in a widespread and spatially variable recovery of aspen stands. There was wide variation in browsing intensity and aspen height between sectors of the Yellowstone northern ungulate winter range (northern range). The east sector generally had lower rates of browsing and more stands with tall saplings than the central and west sectors, a pattern that matched recent trends in elk population densities. Only a small minority of stands in the west sector had tall saplings, consistent with higher elk densities in the west. Densities of elk in winter on the northern range recently have been highest in the northwest sector outside the park boundary, where elk benefit from lower wolf densities and milder winters. Aspen stands did not recover at a comparable range-wide elk density when elk were culled in the park in the 1950s and 1960s, suggesting that the influence of wolves may be an important factor in the recent redistribution and reduction of herbivory impacts by elk. To examine the relationship between elk and aspen outside of YNP, I assessed browsing intensity and sapling recruitment in 43 aspen stands in the Shoshone National Forest east of the park, compared to data collected in the same stands in 1997-98. As in northern YNP, results were consistent with a trophic cascade with reduced browsing and increased recruitment of aspen saplings, but aspen recovery was patchy. Elk densities were moderate to high in most of the area, suggesting that the partial aspen recovery may involve a behavioral response to predation or other factors resulting in local variation in browsing impacts. Livestock may also have limited aspen recruitment. Recovery of some aspen stands in the Shoshone National Forest may provide some of the first evidence of a trophic cascade from wolves to elk to aspen outside of a national park, a trophic cascade possibly weakened by the influence of another large herbivore (cattle). Like cattle, bison in northern Yellowstone may have an effect on woody browse plants. Bison have increased in number and may prevent recovery of some aspen stands in places of high bison density. I also examined browsing impacts of bison on willow and cottonwood in the Lamar Valley. To distinguish the effects of bison from those of elk, I compared browsing at different heights on tall willows, below and above the reach of bison. Because elk were absent from the area in summer when bison were present at high density, I also measured browsing that occurred in the summer. I found high rates of summer browsing, and growth of willows and cottonwoods was suppressed in the Lamar Valley. Above the reach of bison (>100 cm), growth was not suppressed and browsing rates were low, suggesting that these plant species have been released from suppression by elk but bison have compensated for some of the reduction in elk browsing. This study provided the first evidence of significant herbivory by bison of woody browse plants in Yellowstone, and revealed some of the complexity of the Yellowstone food web. In summary, these research results support the hypothesis of a trophic cascade resulting from large carnivore restoration and subsequent changes in elk population densities and distribution. The return of wolves may have combined with other factors such as changes in hunting and land ownership, and increased predation by bears, to result in large-scale shifts in the distribution of elk in northern Yellowstone and greatly reduced elk densities in some areas. If these trends continue, the result may be a new alternative state with lower elk densities, and potential for enhanced biodiversity through reduced herbivory of woody browse species. Close

• 3319. [Article] Physical-Environmental Effects of Wave and Offshore Wind Energy Extraction: A Synthesis of Recent Oceanographic Research

  The ocean deployment of arrays of Wave Energy Converters (WEC arrays) appears likely in the near future, and deployment of offshore wind turbines has already started. These technologies tap into a potential ...

  Citation × Citation Citation Abstract Copy Title: Physical-Environmental Effects of Wave and Offshore Wind Energy Extraction: A Synthesis of Recent Oceanographic Research Author: Özkan-Haller, Tuba, Haller, Merrick C. The ocean deployment of arrays of Wave Energy Converters (WEC arrays) appears likely in the near future, and deployment of offshore wind turbines has already started. These technologies tap into a potential renewable energy resource but also involve complex systems with uncertain environmental consequences that will likely scale with the size of their ocean footprint. This synthesis talk will concentrate on the potential physical effects of these array technologies. Both WEC arrays and offshore wind farms consist of sizable structures placed in the water column; hence, their mere presence is a potential environmental stressor. Possible effects on the physical environment include wave scattering and wave shadowing; added drag on the coastal current field; modifications to sediment transport (by way of the aforementioned changes to the wave and current forcing); and changes to local sediment characteristics (due to anchors and pilings). In many ways, these effects are similar to those caused by other ocean structures that have been studied for some time (e.g., offshore platforms). However, there are additional potential effects of WECs and wind turbines that require further attention. For example, extraction of wave energy by WECs could have additional environmental consequences. Similarly, offshore wind farms can alter the local wind field, in turn altering locally-generated waves. We will address effects due to wave or wind installations on the wave field, on local ocean circulation, and on sediment transport characteristics. Because WECs partially extract and scatter incident wave energy, they cause significant modifications in the near-field. In fact, if device performance can be optimized at field scales, then by definition the near-field effects will be maximized, i.e., if energy extraction is maximized the potential physical effects of WECs are also maximized. Over the past decade a sizable number of studies have applied theoretical principles using varying assumptions and simplifications to the problem of WEC-wave interactions. Some of these assumptions (e.g., “optimal” motions, monochromatic wave conditions, etc.) have now been shown to be unrealistic, and there has been a convergence toward classes of models that appear to produce reasonable estimates. While recent model studies have managed to bound the problem, significant uncertainties remain. The primary cause for the remaining uncertainties is the lack of observational studies, particularly data sets that provide spatial information about the wave field in the vicinity of in situ devices. Nonetheless, a few studies have undertaken scaled laboratory testing, and these data sets are beginning to lend confidence to the available numerical model results and shed light on the dominant processes. Once near-field effects are understood, far-field effects can be assessed. Far-field effects influence the wave field near beaches, which, in turn, influences the sand transport processes that govern the morphodynamics of the beach face. Fortunately, hydrodynamic modelling of large-scale wave propagation processes in the absence of structures is highly advanced, i.e., if given accurate incident wave conditions in the lee of an installation and bathymetry for the model domain, models can well-simulate local wave conditions, wave-driven currents and sediment transport patterns. Therefore, once near-field WEC/wave dynamics are understood, expanding our understanding to the far-field will be relatively straightforward. Nonetheless, observational studies of far-field beach modifications shoreward of an installation will help to further solidify our understanding of beach behaviour. Offshore wind farms can also potentially influence the local wind field around them. Previous studies of such modifications at land-based wind farm installations serve as a reasonable basis for predictions offshore . Any changes to offshore winds will also influence the local wave field, especially where local winds are the dominant source of waves. Such effects will be minimal near coasts where the local wave climate is dominated by incident swells generated at large distances (e.g., the U.S. West Coast). In contrast, locally generated waves are a more important component of the wave climate on the East Coast of the U.S Modification to ocean currents by an array of structures can be assessed by considering the additional frictional effects (“form” drag) caused of the array. If the drag caused by a dense of array of structures is large, circulation will be altered, which might result in reduced current velocities or the diversion of currents toward an area of less drag. Note that ocean currents already experience drag due to bottom friction; hence, the question hinges on the relative magnitude of the drag induced by structures versus the pre-existing frictional drag. Finally, any near-field modifications to the wave and circulation field (due to either WEC arrays or wind farms) will necessarily result in changes in sediment transport. Any local reduction in flow velocities can result in a reduction of the sediment carrying capacity of circulation leading to sediment accumulation at the site. Small-scale modification to a current will also likely cause bumps and holes around the pilings or anchors. These effects are similar to those observed around existing offshore structures and pilings, and can be accounted for in the design of the structures. Far-field modification of waves and associated changes in wave-induced currents can also result in changes in sediment transport patterns near beaches. Although some recent studies exist, questions regarding far-field effects on beaches are still relatively poorly addressed. Title: Physical-Environmental Effects of Wave and Offshore Wind Energy Extraction: A Synthesis of Recent Oceanographic Research Author: Özkan-Haller, Tuba, Haller, Merrick C. The ocean deployment of arrays of Wave Energy Converters (WEC arrays) appears likely in the near future, and deployment of offshore wind turbines has already started. These technologies tap into a potential renewable energy resource but also involve complex systems with uncertain environmental consequences that will likely scale with the size of their ocean footprint. This synthesis talk will concentrate on the potential physical effects of these array technologies. Both WEC arrays and offshore wind farms consist of sizable structures placed in the water column; hence, their mere presence is a potential environmental stressor. Possible effects on the physical environment include wave scattering and wave shadowing; added drag on the coastal current field; modifications to sediment transport (by way of the aforementioned changes to the wave and current forcing); and changes to local sediment characteristics (due to anchors and pilings). In many ways, these effects are similar to those caused by other ocean structures that have been studied for some time (e.g., offshore platforms). However, there are additional potential effects of WECs and wind turbines that require further attention. For example, extraction of wave energy by WECs could have additional environmental consequences. Similarly, offshore wind farms can alter the local wind field, in turn altering locally-generated waves. We will address effects due to wave or wind installations on the wave field, on local ocean circulation, and on sediment transport characteristics. Because WECs partially extract and scatter incident wave energy, they cause significant modifications in the near-field. In fact, if device performance can be optimized at field scales, then by definition the near-field effects will be maximized, i.e., if energy extraction is maximized the potential physical effects of WECs are also maximized. Over the past decade a sizable number of studies have applied theoretical principles using varying assumptions and simplifications to the problem of WEC-wave interactions. Some of these assumptions (e.g., “optimal” motions, monochromatic wave conditions, etc.) have now been shown to be unrealistic, and there has been a convergence toward classes of models that appear to produce reasonable estimates. While recent model studies have managed to bound the problem, significant uncertainties remain. The primary cause for the remaining uncertainties is the lack of observational studies, particularly data sets that provide spatial information about the wave field in the vicinity of in situ devices. Nonetheless, a few studies have undertaken scaled laboratory testing, and these data sets are beginning to lend confidence to the available numerical model results and shed light on the dominant processes. Once near-field effects are understood, far-field effects can be assessed. Far-field effects influence the wave field near beaches, which, in turn, influences the sand transport processes that govern the morphodynamics of the beach face. Fortunately, hydrodynamic modelling of large-scale wave propagation processes in the absence of structures is highly advanced, i.e., if given accurate incident wave conditions in the lee of an installation and bathymetry for the model domain, models can well-simulate local wave conditions, wave-driven currents and sediment transport patterns. Therefore, once near-field WEC/wave dynamics are understood, expanding our understanding to the far-field will be relatively straightforward. Nonetheless, observational studies of far-field beach modifications shoreward of an installation will help to further solidify our understanding of beach behaviour. Offshore wind farms can also potentially influence the local wind field around them. Previous studies of such modifications at land-based wind farm installations serve as a reasonable basis for predictions offshore . Any changes to offshore winds will also influence the local wave field, especially where local winds are the dominant source of waves. Such effects will be minimal near coasts where the local wave climate is dominated by incident swells generated at large distances (e.g., the U.S. West Coast). In contrast, locally generated waves are a more important component of the wave climate on the East Coast of the U.S Modification to ocean currents by an array of structures can be assessed by considering the additional frictional effects (“form” drag) caused of the array. If the drag caused by a dense of array of structures is large, circulation will be altered, which might result in reduced current velocities or the diversion of currents toward an area of less drag. Note that ocean currents already experience drag due to bottom friction; hence, the question hinges on the relative magnitude of the drag induced by structures versus the pre-existing frictional drag. Finally, any near-field modifications to the wave and circulation field (due to either WEC arrays or wind farms) will necessarily result in changes in sediment transport. Any local reduction in flow velocities can result in a reduction of the sediment carrying capacity of circulation leading to sediment accumulation at the site. Small-scale modification to a current will also likely cause bumps and holes around the pilings or anchors. These effects are similar to those observed around existing offshore structures and pilings, and can be accounted for in the design of the structures. Far-field modification of waves and associated changes in wave-induced currents can also result in changes in sediment transport patterns near beaches. Although some recent studies exist, questions regarding far-field effects on beaches are still relatively poorly addressed. Close

• 3320. [Article] Vegetation and Small Mammal Responses to Western Juniper (Juniperus occidentalis) Control in Eastern Oregon

  Shrub-steppe ecosystems of western North America provide habitat for many wildlife species, are important components of public and private rangelands, and offer recreational opportunities for millions ...

  Citation × Citation Citation Abstract Copy Title: Vegetation and Small Mammal Responses to Western Juniper (Juniperus occidentalis) Control in Eastern Oregon Author: Morozumi, Corinne N. Shrub-steppe ecosystems of western North America provide habitat for many wildlife species, are important components of public and private rangelands, and offer recreational opportunities for millions of people. They are some of the most vulnerable ecosystems in the United States and have been altered by human activities such as livestock grazing, active fire suppression, conversion to agriculture, and urbanization of the west. Since the late 1800s woody encroachment of piñon-juniper species has also contributed to the loss of shrub-steppe habitat. Presently, land managers remove woody tree species in order to recover shrub-steppe although it is unclear how responses to these management activities differ due to site-specific conditions and existing woodland development. I studied post juniper-thinning responses in eastern Oregon at a wildlife area important as winter range for mule deer (Odocoileus hemionus). I investigated vegetative responses to western juniper (Juniperus occidentalis) thinning across a woodland development gradient and tested for interactive effects of juniper cutting and cattle exclusion. In addition, I explored plant and small mammal successional dynamics after juniper thinning and examined plant community responses within microhabitats created by the felled trees. I compared vegetative responses to juniper thinning as well as cattle exclusion among sites where juniper were subordinate (Phase I), co-dominant (Phase II), and dominant to shrubs and grasses (Phase III, sensu Miller et al. 2005). At the Phase I site, thinning did not increase herbaceous biomass while thinned plots at the Phase II site had 2.71 times more median herbaceous biomass (99.17% CI: 1.37 to 5.37 times more biomass) than unthinned plots. Conversely, herbaceous biomass at the Phase III juniper woodland site was 0.36 times lower in thinned plots when compared to unthinned plots (99.17% CI: 0.17 to 0.78 times less biomass). Unfortunately, many of the responses were driven by exotic species release. Where juniper were subordinate (Phase I), mean percent cover by exotic grass increased by 24.58 percentage points in thinned plots as compared to unthinned plots (98.30% CI: 0.27 to 48.90 percentage points higher). Thinning at the Phase II site increased mean exotic grass cover by an estimated 28.47 percentage points as compared to mean exotic cover in the same plots before treatment (98.30% CI: 4.15 to 52.79 percentage points higher cover). Median native bunchgrass cover at this site was 5.06 times greater after juniper treatment (99.7% CI: 1.78 to 14.35 times higher percent cover). I found few main or interactive effects of cattle exclosure after one year of treatment. Responses to grazing exclosure may take longer to develop. These results indicate that sites within the wildlife area respond differently to juniper management and that exotic grass control will be key to successful shrub-steppe recovery. In addition, I used a time-since-juniper thinning chronoseries consisting of plots cut in 2008, 2009, 2010, 2012, and an adjacent uncut control to explore how shrub-steppe flora and fauna are responding to juniper treatment through time. Shrub cover and seedling density were low in each plot. I recorded the highest seedling abundance (mean of 0.25 seedlings) in the uncut control plot. Median grass cover in the uncut control was 9.50% while in the most recently treated plot (2012) it was 26.75%. Small mammal relative abundance and diversity was low across all time-since-treatment plots though highest in the plot with the greatest time-since-treatment. Deer mouse (Peromyscus maniculatus) was the most abundant species in all plots and accounted for 70-95% of all unique captures. Least chipmunk (Tamias minimus) were present in the plot with greatest time-since-treatment indicating the potential recovery of key native shrubs since these small mammals perform an important seed dispersal role. I also investigated how potential microhabitats created by the felled juniper might support different plant communities. I assessed plant responses within zones created by 1) the felled tree (canopy zone), 2) the original duff zone, and 3) the between - tree interspace zone. The three zones had different plant community compositions as analyzed multivariately with non-metric multidimensional scaling (NMS). Canopy and duff zones were forb - dominated and had less exotic grass invasion while interspace zones were heavily invested with medusahead (Taeniatherum caput-medusae) and ventenata (Ventenata dubia). Exotic herbaceous productivity was lowest under the standing trees of the control plot and high in the interspace of all plots with juniper thinning. In the 2012 cut plot, grass cover was 27.90 percentage points less in the canopy zone compared to the interspace zone (98.75% CI: 17.64 to 38.16 percentage points less grass cover). In the uncut control, median native biomass was 2.08 times greater under the standing trees compared to the interspace (96.25% CI: 1.09 to 3.97 times more biomass). Results from the time-since-treatment chronoseries indicated that exotic grass dominance might be limiting shrub-steppe recovery. Active restoration of shrubs and native grasses may be necessary to address the dominance of exotic grass after juniper thinning. Positive signs of habitat recovery included small mammal responses though I was unable to calculate population estimates due to plot size. Title: Vegetation and Small Mammal Responses to Western Juniper (Juniperus occidentalis) Control in Eastern Oregon Author: Morozumi, Corinne N. Shrub-steppe ecosystems of western North America provide habitat for many wildlife species, are important components of public and private rangelands, and offer recreational opportunities for millions of people. They are some of the most vulnerable ecosystems in the United States and have been altered by human activities such as livestock grazing, active fire suppression, conversion to agriculture, and urbanization of the west. Since the late 1800s woody encroachment of piñon-juniper species has also contributed to the loss of shrub-steppe habitat. Presently, land managers remove woody tree species in order to recover shrub-steppe although it is unclear how responses to these management activities differ due to site-specific conditions and existing woodland development. I studied post juniper-thinning responses in eastern Oregon at a wildlife area important as winter range for mule deer (Odocoileus hemionus). I investigated vegetative responses to western juniper (Juniperus occidentalis) thinning across a woodland development gradient and tested for interactive effects of juniper cutting and cattle exclusion. In addition, I explored plant and small mammal successional dynamics after juniper thinning and examined plant community responses within microhabitats created by the felled trees. I compared vegetative responses to juniper thinning as well as cattle exclusion among sites where juniper were subordinate (Phase I), co-dominant (Phase II), and dominant to shrubs and grasses (Phase III, sensu Miller et al. 2005). At the Phase I site, thinning did not increase herbaceous biomass while thinned plots at the Phase II site had 2.71 times more median herbaceous biomass (99.17% CI: 1.37 to 5.37 times more biomass) than unthinned plots. Conversely, herbaceous biomass at the Phase III juniper woodland site was 0.36 times lower in thinned plots when compared to unthinned plots (99.17% CI: 0.17 to 0.78 times less biomass). Unfortunately, many of the responses were driven by exotic species release. Where juniper were subordinate (Phase I), mean percent cover by exotic grass increased by 24.58 percentage points in thinned plots as compared to unthinned plots (98.30% CI: 0.27 to 48.90 percentage points higher). Thinning at the Phase II site increased mean exotic grass cover by an estimated 28.47 percentage points as compared to mean exotic cover in the same plots before treatment (98.30% CI: 4.15 to 52.79 percentage points higher cover). Median native bunchgrass cover at this site was 5.06 times greater after juniper treatment (99.7% CI: 1.78 to 14.35 times higher percent cover). I found few main or interactive effects of cattle exclosure after one year of treatment. Responses to grazing exclosure may take longer to develop. These results indicate that sites within the wildlife area respond differently to juniper management and that exotic grass control will be key to successful shrub-steppe recovery. In addition, I used a time-since-juniper thinning chronoseries consisting of plots cut in 2008, 2009, 2010, 2012, and an adjacent uncut control to explore how shrub-steppe flora and fauna are responding to juniper treatment through time. Shrub cover and seedling density were low in each plot. I recorded the highest seedling abundance (mean of 0.25 seedlings) in the uncut control plot. Median grass cover in the uncut control was 9.50% while in the most recently treated plot (2012) it was 26.75%. Small mammal relative abundance and diversity was low across all time-since-treatment plots though highest in the plot with the greatest time-since-treatment. Deer mouse (Peromyscus maniculatus) was the most abundant species in all plots and accounted for 70-95% of all unique captures. Least chipmunk (Tamias minimus) were present in the plot with greatest time-since-treatment indicating the potential recovery of key native shrubs since these small mammals perform an important seed dispersal role. I also investigated how potential microhabitats created by the felled juniper might support different plant communities. I assessed plant responses within zones created by 1) the felled tree (canopy zone), 2) the original duff zone, and 3) the between - tree interspace zone. The three zones had different plant community compositions as analyzed multivariately with non-metric multidimensional scaling (NMS). Canopy and duff zones were forb - dominated and had less exotic grass invasion while interspace zones were heavily invested with medusahead (Taeniatherum caput-medusae) and ventenata (Ventenata dubia). Exotic herbaceous productivity was lowest under the standing trees of the control plot and high in the interspace of all plots with juniper thinning. In the 2012 cut plot, grass cover was 27.90 percentage points less in the canopy zone compared to the interspace zone (98.75% CI: 17.64 to 38.16 percentage points less grass cover). In the uncut control, median native biomass was 2.08 times greater under the standing trees compared to the interspace (96.25% CI: 1.09 to 3.97 times more biomass). Results from the time-since-treatment chronoseries indicated that exotic grass dominance might be limiting shrub-steppe recovery. Active restoration of shrubs and native grasses may be necessary to address the dominance of exotic grass after juniper thinning. Positive signs of habitat recovery included small mammal responses though I was unable to calculate population estimates due to plot size. Close