Sample Design

We sampled for suckers with trap nets to assess species-specific annual variability in production and growth, as well as annual and seasonal variability in survival of juvenile suckers in Upper Klamath and Clear Lakes. The timing of the sampling periods was selected based on previous catch data in Upper Klamath Lake. Specifically, we targeted age-1 suckers in early June, the ramp up of age-0 sucker catches in July, the peak of age-0 sucker catches in August, and the tail end of age-0 sucker catches in September (Burdick and Martin, 2017). In 2015, sampling was conducted over three 3-week periods simultaneously in Upper Klamath and Clear Lakes. An evaluation of the study design in 2015 indicated that with increased effort concentrated into shorter time periods, we could better describe growth and differences in catch rates between sampling periods. Starting in 2016, for each sampling month, we sampled for a week in one lake and then the next week we sampled in the other lake. A July sampling period was added in 2018 to ensure that we did not miss early age-0 suckers. Sampling occurred in the same calendar weeks each year within each lake, with the exception of July 2018 in Upper Klamath Lake when poor air quality resulting from wildfires limited access to Upper Klamath Lake and delayed the sampling effort by one week compared to timing in other years.

Given the limitations of our selected gear type, our analysis of catch data is only relevant to suckers from about 45 to 300 millimeters (mm) standard length (SL). Fish small enough to pass through the mesh of our nets, such as small age-0 suckers (<45 mm SL), have a low catchability in trap nets (Burdick and Martin, 2017). Because adult suckers (>300 mm SL) are captured at high rates in spring and fall trammel net sampling and infrequently in summer trap net sampling we presume trap nets select for smaller suckers relative to trammel nets (Hewitt and Hayes, 2013). Burdick and others (2016) did not find a length-based pattern in the proportions of PIT-tagged and released suckers 71–236 mm SL that were recaptured, indicating there was not strong size selectivity within this size range.

To reduce potential sample bias caused by apparently minor spatial heterogeneity in the densities, species, ages, size or health of suckers, we selected fixed sample sites in a variety of habitats throughout both lakes. Age-0 suckers at least 45 mm SL, the size targeted in our sampling, are not known to be distributed differentially within Upper Klamath Lake based on species or size (Hendrixson and others, 2007; Burdick and Hewitt, 2012). However, age-1 suckers are more likely to be found in shallow (<1 m deep) near-shore habitats in the spring and deep water around 2 m deep in the summer (Bottcher and Burdick, 2010). Spatial patterns among age-classes of suckers have not been identified in Clear Lake (Burdick and Rasmussen, 2012). Sample areas were either 1-km long sections of shoreline or 300 m² offshore areas. Within each area, 10 fixed sites were identified as potentially accessible given a variety of water levels. In 2020, as in past years, 8 sites at each area in Upper Klamath Lake and 7 sites in each area in Clear Lake Reservoir were sampled during each sampling period (tables 1, 2). To address the concern of inadvertent bias in our fixed-site selection, randomly determined site locations were sampled during all sample years, and it was determined that there was no significant difference between fixed and random sites (Burdick and others, 2016). In order to compare catches between years, only fixed sites are reported in this publication.

Sample sites that were shallow and near shore in low water years of 2015 and 2016 were often in more than 3 m of water and far from shore from 2017 to 2020. Because juvenile sucker catch rates with trap nets decrease at depths greater than 3 m (Burdick and Hewitt, 2012), we captured few juvenile suckers at this depth. Therefore, we adjusted sample locations slightly each year to adjust for water depth. We started by going to the 2015 and 2016 locations, then driving directly toward shore from the original site until we were in less than 3 m of water before setting the trap nets.

Fish Handling and Sampling

Sampling was conducted with rectangular trap nets with mouth dimensions of 0.61 × 0.91 m, a 10-m-lead, and three internal fykes. Weight, SL, and fork length were recorded for each captured individual. The leading left pectoral fin ray was removed at the proximal joint for aging. Fin rays were not collected from small suckers (35–60 mm SL) from Upper Klamath Lake since they were presumed to be age-0 fish based on length at date of capture (Burdick and Martin, 2017). We compared the length and number of annuli on fish with fin rays collected to length of suckers without fin rays collected to validate our length-based age assumptions. A small (about 2 mm²) piece of tissue from the caudal fin was collected for genetic identification to taxa. All suckers collected (129 from Clear Lake and 55 from Upper Klamath Lake) were measured, aged either by fin ray or presumed age-0, and identified to species via genetic analysis. Emaciation, deformities, macro parasites, and petechial skin hemorrhaging were systematically recorded. Other abnormalities and afflictions were noted when they were observed. Individuals were scanned for the presence of a PIT-tag to document recaptures from prior juvenile sampling efforts and hatchery program releases. If no tag was detected, the individual was larger than 60 mm SL, and lake conditions did not compromise sucker health, a PIT-tag was inserted into the ventral abdominal musculature anterior to the pelvic girdle (Burdick, 2011). Suckers were released at their site of capture.

Aging Juvenile Suckers

To estimate sucker age, fin rays were mounted in epoxy, sectioned, and viewed by two experienced readers under magnification using transmitted light (Quist and others, 2012). The number of annuli was first determined in blind reads by two readers, with each reader having no knowledge of the other’s annuli count. When both readers agreed on a number of annuli, that number was presumed to be the correct age and was used in analyses. If there was disagreement in the annuli count, the two readers viewed the structure together and came to a consensus or a third reader acted as a tie breaker. All the suckers from both lakes were either aged or systematically assumed to be age-0 by being 60 mm or less in standard length.

Species Identification

To identify juvenile suckers to taxa, we applied genetic identification methods described by Smith and others (2020 and 2022). Caudal fin tissue was collected from juvenile suckers from each lake, dried, and analyzed at the USFWS Abernathy Fish Health Center in Abernathy, Washington. Individuals were assigned to their reporting/species group described in Smith and others (2022) using genetic stock identification (GSI) implemented in the R package rubias (Moran and Anderson, 2019). The rubias package is a Bayesian approach to the conditional GSI model which includes a leave-one-out cross validation and simulation method to enhance GSI accuracy (Anderson and others, 2008). We selected a mean posterior probability threshold of 0.9 for accepting individual assignments. Only individuals assigning to a reporting/species group above this threshold were categorized to species for subsequent analysis. Fish that could not be assigned to species based on the above criteria were considered to be indeterminate. The current genetic technique to separate Klamath largescale from shortnose suckers was built upon known species from the Upper Klamath Basin (Smith and others, 2020), and there is always the possibility that the selected markers do not translate into proper assignment of species in the Lost River Basin; consequently, although a few shortnose suckers were genetically identified in the Lost River system, most were not separated from Klamath largescale suckers and therefore were classified as the combination of Klamath largescale suckers and shortnose suckers from the Lost River Basin. Previous genetic techniques could only separate Lost River suckers from a mix of shortnose and Klamath largescale suckers (Hoy and Ostberg, 2015). In order to compare genetics across years we included a category of combined shortnose and Klamath largescale suckers in both Upper Klamath and Clear Lakes.

Indices of Juvenile Sucker Year-Class Strength and Apparent Survival

To describe annual relative (among cohorts, species, and lakes) year-class strength and apparent age-0 sucker production, we calculated (1) the proportion of August nets to catch one or more age-0 sucker (successful age-0 nets), (2) the mean August catch per unit effort (CPUE) for age-0 suckers in successful age-0 nets, and (3) the total August age-0 CPUE as the number of suckers in each taxa divided by the number of nets set. We assessed age-0 summer apparent survival by comparing CPUE by year-class between the August and September sampling periods. To provide an index to compare between years, September age-0 CPUE was divided by August age-0 CPUE. If the September CPUE was greater than the August CPUE, >1.00 was reported indicating that suckers were recruiting to the gear at a greater rate in September than in August. We also calculated an index of overwinter apparent survival for each year-class as the ratio of age-1 CPUE in June divided by age-0 CPUE from September from the previous year. If age-0 CPUE was zero and age-1 CPUE was greater than zero, then results were reported as >1.00.

We assumed that sampling efficiency was similar between years and within year sampling periods. The presence of vegetation, substrate type, and water depth have minor effects on detection probability of juvenile suckers (Burdick and others, 2008). By using the same fixed sites throughout relatively homogenous habitat with little to no vegetation, we ensured that habitat variables were similar at sampled sites between years. Furthermore, water management in Upper Klamath Lake ensures that water depth is similar each August and therefore did not differentially affect capture probability.

Observations on External Afflictions

We summarized the prevalence and intensity of external afflictions on juvenile suckers to roughly compare the apparent health of suckers between years and lakes and to potentially identify causes of sucker mortality. We paid special attention to those afflictions that are either common or potentially associated with mortality such as Lernaea sp., petechial hemorrhaging, and lamprey wounds (Markle and others, 2014; Burdick and others, 2015a). Afflictions were then quantified and compared to observed afflictions relative to previous years.

Upper Klamath Lake Year-Class Strength and Apparent Survival

During the 2020 juvenile monitoring sampling, 55 suckers were captured in Upper Klamath Lake, and most (60 percent) were age-0 (table 3; fig. 3). Of the 33 age-0 suckers captured in Upper Klamath Lake, 20 were Lost River sucker, 11 were shortnose sucker, 1 was a Klamath largescale sucker, and 1 was indeterminant (table 3). Of the 22 suckers age-1 or greater captured in Upper Klamath Lake, 3 were Lost River sucker, 15 were shortnose sucker, and 4 were Klamath largescale sucker (table 3). Age-0 suckers from Upper Klamath Lake ranged from 42 mm to 99 mm SL (fig. 4). Most suckers captured in Upper Klamath Lake were less than 150 mm SL, with the exception of one individual caught near Mid-North at 159 mm and another individual caught at Rattlesnake Point at 236 mm (fig. 5). The 236 mm individual was a recently released shortnose sucker from the USFWS SARP.

The 2020 August CPUE for all sucker species combined in Upper Klamath Lake was lower than in most other sampling years and most similar to 2017 and 2018. Most of the 2020 year-class was composed of Lost River suckers (table 4). Age-0 Lost River Sucker CPUE in 2020 was slightly larger than in 2017 and 2018, somewhat smaller than in 2015 and 2019, and only about 10 percent of CPUE in 2016 (table 4). The 2020 shortnose sucker CPUE was between the 2017 and 2018 year-classes CPUE, and about a quarter of the 2015, 2016, and 2019 year-classes CPUE (table 4).

Cohort tracking among years indicated that within year and overwinter apparent survival of suckers was low in Upper Klamath Lake (tables 5, 6, 7, 8). Shortnose suckers had greater apparent survival than Lost River suckers within most years and all years for overwinter apparent survival (tables 7 and 8). A total of 18 age-1 suckers from the 2019 cohort, two age-2 suckers from the 2018 cohort, and two age-3 suckers from the 2017 cohort were captured throughout the sampling season in 2020 (fig. 3). CPUE of the 2019 cohort decreased to near zero starting in July 2020 (table 5). The 2020 cohort was caught during the July sampling period but did not fully recruit to the gear until August (table 6). Catch rates for the 2020 cohort declined between August and September (table 6). August to September survival indices for the 2020 cohort were moderate, with apparent survival greater than the 2015, 2016, 2017, and 2019 cohorts but lower than the 2018 cohort (table 7). August to September survival indices for the 2020 cohort were greater for shortnose suckers than for previous cohorts except 2018 (table 7).

Clear Lake Reservoir Year-Class Strength and Apparent Survival

The majority of suckers captured in Clear Lake Reservoir in 2020 were classified as the combination of Klamath largescale suckers and shortnose suckers from the Lost River Basin from 3 to 5 years old, and ranged from 136 to 325 mm SL (figs. 3 and 4). Of the suckers captured during the 2020 Clear Lake Reservoir sampling, seven were under age-3 and ten were older than age-5. None of the 129 suckers captured in Clear Lake Reservoir during the 2020 juvenile monitoring sampling were age-0 (fig. 3; table 9). Only 6 of the 129 suckers were not classified as the combination of Klamath largescale and shortnose suckers from the Lost River Basin: 4 were Lost River suckers, and 2 were shortnose suckers based on Klamath Basin genetics (figs. 3 and 5).

The age distribution of suckers in Clear Lake indicates better annual survival of juvenile suckers than in Upper Klamath Lake, where very few age-1 or older suckers are captured. The oldest sucker we captured in Clear Lake was a 410 mm standard length age-9 sucker from the 2011 cohort (fig. 3). We captured two each age-7 and age-8 suckers from the 2012 and 2013 cohorts, five age-6 suckers from the 2014 cohort, and 15 age-5 suckers from the 2015 cohort. The most abundant cohorts captured were from 2016 (age-4) and 2017 (age-3). Age-1 and age-2 suckers from the 2018 and 2019 cohorts were also represented in our catches but in low numbers. No age-0 suckers were captured in Clear Lake in 2020.

There were several year-classes of suckers that were not captured as age-0 suckers; however, these year-classes were captured for the first time as age-1 to age-3 (table 10). The peak of the 2015 cohort catches was during the September 2017 sampling effort, or when the cohort was age-2 (table 10). The 2016 cohort catches peaked during September 2016 when the cohort was age-0. The 2017 cohort catch peaked during the July 2019 sampling effort or when the cohort was age-2. The 2018 cohort catch peaked during June 2019 sampling effort when the cohort was age-1. Although the 2019 cohort was not detected during the 2019 sampling, it also followed the same trend as other cohorts being caught at greater rates once it reached age-1 (table 10).

Afflictions

Unlike other years, in 2020, suckers in Clear Lake generally had more macro parasites, deformities, and skin afflictions than suckers in Upper Klamath Lake. The primary affliction observed during the 2020 monitoring season in both lakes was attached Lernaea sp. Other afflictions observed were lamprey wounds on twelve Clear Lake Reservoir suckers and two Upper Klamath Lake suckers. Petechial hemorrhaging of the skin was observed on seven Clear Lake suckers and one Upper Klamath Lake sucker. Six suckers from Clear Lake had missing or blind eyes whereas no suckers from Upper Klamath Lake were observed with this affliction. We did not observe any fish with black spot (metacercariae of Bolbophorus sp.) from either lake in 2020 (table 11).

Missing or deformed opercula were observed only in Clear Lake juvenile suckers in 2020, whereas in previous sampling years this affliction was only observed in age-0 Upper Klamath Lake suckers (table 12). Except for 2015, opercular deformities were more common in Lost River suckers than shortnose suckers. Opercular deformities occurred in Lost River suckers from 0 to 62 percent of the time within each year, whereas this deformity occurred less than 10 percent of the time in shortnose suckers within each year. Of the three suckers with opercula afflictions in 2020, two had one affected operculum, and one had both missing or deformed. All suckers with missing or deformed opercula in 2020 were age-4 suckers classified as the combination of Klamath largescale suckers and shortnose suckers from the Lost River Basin.

Lernaea sp. were the most common parasite seen on juvenile suckers in 2020. All Lernaea sp. observations in age-0 suckers occurred from Upper Klamath Lake, and there was only one Lernaea sp. per individual (table 13). Age-1 and older suckers in Upper Klamath Lake had a greater proportion of fish with Lernaea sp. attached than those in Clear Lake (table 14). The most Lernaea sp. attached to an individual juvenile sucker was six, but most often only one Lernaea sp. was attached. Although this parasite was a relatively common occurrence in Upper Klamath Lake and Clear Lake, we could not detect any obvious signs that Lernaea sp. caused mortality of juvenile suckers in 2020.

Petechial hemorrhaging of the skin on age-0 fish was only observed in one Upper Klamath Lake sucker (table 15). The proportion of age-0 suckers in Upper Klamath Lake with petechial hemorrhaging in 2020 was low relative to previous years (table 15). Petechial hemorrhaging was observed on a lower proportion of age-1 and older suckers relative to 2019 in both lakes (table 16). Although petechial hemorrhaging was observed on Clear Lake suckers age-1 and older in 2020, it was still a relatively rare affliction not commonly observed in Clear Lake.

Upper Klamath Lake

The lack of substantial recruitment to the spawning population continues to be the bottleneck for the recovery of shortnose and Lost River suckers in Upper Klamath Lake. Since the early 2000s, the abundance of both species has decreased by more than 40 percent (Hewitt and others, 2018). Nearly all adult suckers in Upper Klamath Lake are older than the average life span expected for each species and shortnose suckers are approaching the maximum known age for their species (Hewitt and others, 2018). As the adult sucker populations diminish, we continue to catch small numbers of juvenile suckers during our monitoring efforts. Without the balance of recruitment by new individuals to the spawning population, Lost River and shortnose suckers will continue their downward trend until extirpated from Upper Klamath Lake.

The scarcity of age-1 and older suckers in Upper Klamath Lake is likely attributable to juvenile mortality. Several observations support the presumption that mortality, rather than reduced selectivity or emigration from sampled areas, explains the reduction in catch by age. Most of our catch in Clear Lake were age-1 and older suckers, indicating older, larger fish are vulnerable to our trap nets. A substantial lack of recruitment to the adult populations indicates that juvenile suckers have unsustainably low survival rates (Hewitt and others, 2018). A lack of directed movement toward the lake’s outlet suggests that emigration is not the primary reason for a lack of older juvenile suckers in Upper Klamath Lake (Burdick and others, 2009b).

Although mortality is substantial for Lost River and shortnose suckers, it appears to be greater for juvenile Lost River than shortnose suckers as seen by the overwinter survival indices. Although earlier years did not separate out shortnose suckers from Klamath largescale suckers, the 2020 age-1 catch was composed of a greater number of shortnose suckers than Lost River or Klamath largescale suckers. There is always the possibility that the index for past years reflects a greater portion of Klamath largescale suckers surviving to age-1; however, the lack of age-1 Klamath largescale suckers captured during 2020 and the identification of more shortnose suckers than Klamath largescale suckers using vertebrae counts in past studies (Burdick and others, 2008, 2009a; Burdick and Brown, 2010) suggest that the past 5 years may also have a lack of age-1 Klamath largescale suckers. Burdick and Martin (2017) reported the same survival trend; however, they were unable to determine why there were differences in species mortality rates. One hypothesis is that the Lost River suckers, which hatch earlier than shortnose suckers, may encounter unfavorable water quality at a critical juncture in their development.

There are several possible explanations for why we only detected one of the thousands of PIT-tagged suckers that were released into Upper Klamath Lake from the SARP program through 2020. Given the large size of Upper Klamath Lake, these are relatively small numbers of fish to detect, even when ignoring post-release mortality. PIT-tag antennas operating in the Link River at the outlet of Upper Klamath Lake and in the Williamson and Sprague Rivers detected 3 SARP fish in 2018, 19 SARP fish in 2019, and 13 SARP fish in 2020, indicating that directed emigration was an unlikely explanation for the disappearance of these fish. As of the writing of this report, USGS crews have detected 181 PIT tags from SARP-released suckers on bird colonies cumulatively from 2018, 2019, and 2020 scanning efforts, indicating bird predation may be a factor in the survival of some SARP fish (Evans and others, 2016). However, Evans and others (2021) estimated that predation rates on SARP fish were only 4.4–8.8 percent from 2018 through 2020, indicating that colonial bird predation accounts for a small percentage of the mortality of these suckers.

Although it is typical for survival to be low in the early life stages of fish (Houde, 1989), near-complete disappearance of entire cohorts within the first 2 years is of concern. High fecundity may be a life-history strategy to overcome high mortality for juvenile suckers in the Klamath Basin, but near complete mortality is unsustainable (Rasmussen and Childress, 2018). Although age-0 survival was intermediate from August–September for the 2019 cohort compared to age-0 survival of other cohorts, the overwinter survival of the 2019 cohort was low, resulting in a substantial loss of this cohort. Given that the adult populations of Lost River and shortnose suckers have decreased by more than 50 percent since the early 2000s (Hewitt and others, 2018), there would have to be a substantial recruitment event soon for both species to recover naturally. As of the writing of this report, we have no indication that such an event is imminent.

Clear Lake Reservoir

With greater juvenile sucker survival than in Upper Klamath Lake, intermittent recruitment of new spawners has been documented for Clear Lake populations (Hewitt and Hayes, 2013). The mechanisms behind intermittent cohort success are not completely understood. Hypotheses include (1) limited access to spawning habitat in dry years and (2) differential juvenile sucker mortality among years. It has been shown that varying rates of juvenile sucker mortality is associated with differential rates of avian predation among years that are mediated by water levels and fish size, especially for small fish (Evans and others, 2021). Other hypothesis of why survival is better in Clear Lake than Upper Klamath Lake include the differences in water quality between the two lakes, and the suspicion that the Clear Lake population is composed mainly of Klamath largescale suckers, which rear in the river before entering the lake.

A lack of age-0 suckers captured in the low water years of 2014 and 2015 in Clear Lake Reservoir led Burdick and others (2016) to conclude cohorts were not formed in those years. Lake-surface elevations less than 1,378.9 m prevented adult suckers from migrating up the spawning tributaries in the spring of both years (Hewitt and others, 2021). Therefore, Burdick and others (2016) concluded that spawning did not occur without access from the lake to the river. After high flows in the Willow Creek drainage increased lake-surface elevations in 2016 (fig. 6), the 2014 and 2015 cohorts were detected in Clear Lake Reservoir, and these cohorts continued to be collected through 2020. The 2015 cohort continued to be detected in high numbers through 2019 and was detected again in 2020. The presence of the 2015 juvenile cohort in Clear Lake challenged Burdick and others (2016) presumption that high springtime lake elevations are required to form year-classes in Clear Lake Reservoir (Bart and others, 2020a). We suspect that the majority of the 2015 cohort is offspring of stream resident Klamath largescale suckers that only recruited to Clear Lake when high water flushed them from the tributaries. Although the current genetic techniques can separate Klamath largescale suckers from shortnose suckers in the Klamath Basin, there is still uncertainty in separating these species in the Lost River Basin (Smith and others, 2020). Most of the suckers from the 2015 cohort were genetically classified as the combination of Klamath largescale suckers and shortnose suckers from the Lost River Basin. However, a stream resident life history is consistent with these fish being Klamath largescale suckers.

There are several possible explanations for annual variation in age-0 sucker catches. Juvenile and adult suckers have been documented in disconnected pools and reservoirs throughout the Willow and Boles Creek drainages during summer 2018 (Martin and others, 2021). These fish appear to be stranded in these pools and unable to reenter the lake because of low water levels. Furthermore, several cohorts of suckers in Clear Lake were not captured during our sampling until age-1 (Bart and others, 2020a, b; Bart and others 2021). One possible explanation is that juvenile suckers rear for a year or more in tributaries before recruiting to Clear Lake. The lack of access between the tributaries and Clear Lake does not explain why age-0 sucker catches were low or absent in 2017–20. Lake levels were relatively high and adult suckers were detected on remote PIT-tag arrays migrating into Willow Creek during springs of 2017–20, indicating that spawning likely occurred (Hewitt and others, 2021). With higher water levels and access to Willow Creek for spawning habitat, we expected to see large numbers of age-0 suckers in Clear Lake from 2017 to 2020, but this was not the case. In 2018 and 2020, creek flows were relatively low as indicated by a lack of rapid spring lake level increase (fig. 6). Therefore, in these years a long instream residency time, if it occurred, may not be entirely voluntary. Suckers may have made spawning runs during high flows, but by the time larvae hatch water may be insufficient to allow for outmigration to the lake thus trapping suckers in disconnected pools.

Clear Lake surface elevation and Willow Creek flows may affect the annual rate of bird predation on suckers (Hewitt and others, 2021). Double-crested cormorants (Nannopterum auritum) and American white pelicans (Pelecanus erythrorhynchos) prey upon vulnerable suckers as they enter Willow Creek to spawn. Hewitt and others (2021) hypothesized that adult sucker mortality is greatest when lake-surface elevation ranges from 1,378.9 to 1,379.2 m above mean sea level during the spawning migration. Several cohorts that cease to be captured in trammel net catches at the age of first spawning may indicate that young or small adult suckers are especially vulnerable to bird predation. When the lake-surface elevation is less than 1,378.9 m or if instream flows are very low, suckers cannot access spawning habitat in Willow Creek and are therefore less susceptible to predation. The formation of nesting islands for American white pelican and double-crested cormorant at lake-surface elevations greater than 1,378.9 m provides protection for the birds’ eggs from predators, resulting in greater numbers of birds to be present for a longer period (Evans and others, 2016). As lake-surface elevation increases above 1,378.9 m, bird islands shrink in size, thus reducing bird nesting habitat and the number of nesting birds available to prey on suckers. At lake-surface elevation above 1,379.2 m, deep water also provides cover for migrating suckers (National Marine Fisheries Service and U.S. Fish and Wildlife Service, 2013). Poor survival of suckers that first attempt to spawn in low-water years may explain the absence of some cohorts from adult sucker sampling.

Another possibility is that differences in water depth among years affects our ability to capture suckers in Clear Lake. In Upper Klamath Lake, age-0 Lost River and shortnose sucker habitat use is optimized at depths less than 1 m and decreases with deeper depths (Burdick and Hewitt, 2012). Water depth in Clear Lake changes substantially among years (fig. 6). Although we attempted to adjust our sampling sites to account for yearly variation in depths, our sites were deeper in 2017–20 than in 2011–16. Although we aimed to sample depths of 3 m or less when Clear Lake water levels were high, we were still sampling deeper depths than depths sampled in 2011–16. Therefore, it is possible that our sampling did not coincide with high densities of age-0 suckers from 2017 to 2020 if they were in shallower water than what we were sampling. Whatever the reason for the lack of age-0 sucker captures in 2020, it is still possible that a cohort was formed and not detected. Given our continued pattern of catching cohorts once they are past age-0, sampling Clear Lake in 2021 is needed to confirm the existence of the 2020 cohort.

Afflictions

Opercular deformities that are more common for Lost River than shortnose suckers make the gills more vulnerable to parasites and poor water quality, ultimately increasing the chances of mortality. Because there were no age-1 or older suckers in Upper Klamath Lake with deformed opercula, there is the potential that it is serious enough of an affliction that it increases the probability of mortality before fish can reach older ages. However, we did see opercular deformities in several of the age-4 suckers classified as the combination of Klamath largescale suckers and shortnose suckers from the Lost River Basin in Clear Lake in 2020, indicating there may be additional confounding factors. Although this affliction has been observed on other sucker species (Barkstedt and others, 2018), the exact cause of deformed opercula is difficult to determine, but potential explanations could be inbreeding, hybridization (Winemiller and Taylor, 1982; Tringali and others, 2001), nutrient deficiency (Chávez de Martínez, 1990; Lall, 2002), heavy metals, pesticides, high egg incubation temperature (Boglione and others, 2013) or a combination of these factors.

Lernaea sp. parasitism is one of the most common afflictions on juvenile suckers captured in both lakes, occurring to some degree in all sampling years. Wounds that form at Lernaea sp. attachment sites may provide a pathway for bacterial infection (Berry and others, 1991). Inflammation associated with Lernaea sp. attached to juvenile suckers from Upper Klamath Lake is most often limited to a focal area in the skin and skeletal muscle directly surrounding the attachment site, indicating this parasite is unlikely to cause systemic infections that result in mortality (Burdick and others, 2015a).

The causes of petechial hemorrhaging, which generally appears more frequently on suckers from Upper Klamath Lake, are unknown. Petechial hemorrhages of the skin are a common observation in Upper Klamath Lake and have been documented since monitoring for them began in 2014 (Burdick and others, 2015a). Petechial hemorrhages of the skin have been found to be caused by irritants including abrasion, bacteria, or toxins (Ferguson and others, 2011). The low prevalence of observed hemorrhages in Clear Lake relative to Upper Klamath Lake indicates that abrasions caused by our method of capture is unlikely to be the primary reason for the hemorrhaging. Burdick and others (2018) examined the hemorrhages microscopically and did not observe associated bacterial disease or other parasites. Janik and others (2018) observed petechial hemorrhaging on collected fish from Upper Klamath Lake canals; however, they could not observe it through histology, which indicated that the infection was likely confined to the skin.

Lamprey wounds were seen in both lakes but are likely not a large source of mortality. All lamprey species in the Upper Klamath Basin are native (Kostow, 2002), some of which are endemic. Given the low prevalence of lamprey wounds and that lamprey have coevolved with suckers in the Klamath Basin, it is unlikely that they are the primary cause of annual juvenile sucker year-class failure.

Incidence of black spot was hypothesized to be associated with high mortality of juvenile suckers (Markle and others, 2014). In previous years monitoring, black spot was only recorded on a small proportion of fish, and during the 2018 through 2020 monitoring seasons, there were no suckers with black spot (Bart and others 2020a, b; Bart and others, 2021). In years when black spot was observed, it was more prevalent in Upper Klamath Lake than in Clear Lake (Burdick and others, 2016, 2018; Bart and others, 2020b). There is the potential that we are missing cases of black spot in suckers when it is not visible externally. Markle and others (2020) found that out of 55 fish observed without external black spot, 10 had internal muscle or gill infections of black spot. Although this would indicate that black spot is underrepresented in our data, there is no indication from our data that it is a substantial source of mortality for juvenile suckers.