Least Bell’s Vireo

The Least Bell’s Vireo (Vireo bellii pusillus; hereinafter vireo) is a small, migratory, songbird that breeds in southern California and northwestern Baja California, Mexico, from April through July (Kus and others, 2020). Historically abundant within lowland riparian ecosystems, vireo populations began declining in the late 1900s as a result of habitat loss and alteration associated with urbanization and conversion of land adjacent to rivers and agriculture (Franzreb, 1989; U.S. Fish and Wildlife Service, 1998; Riparian Habitat Joint Venture, 2004). Additional factors contributing to the vireo's decline have been the expansion in range of the brood-parasitic Brown-headed Cowbird (Molothrus ater; hereinafter cowbird), to include the Pacific Coast (U.S. Fish and Wildlife Service, 1986; Franzreb, 1989; Kus, 1998, 1999; Kus and others, 2020), and the introduction of invasive non-native plant species, such as giant reed (Arundo donax) into riparian systems. By 1986, the vireo population in California numbered just 300 territorial males (U.S. Fish and Wildlife Service, 1986).

In response to the dramatic reduction in numbers of vireos in California, the California Fish and Game Commission listed the species as endangered in 1980, with the U.S. Fish and Wildlife Service (USFWS) following suit in 1986. Since listing, the vireo population in southern California has rebounded, largely in response to cowbird control and habitat restoration and preservation (Kus, 1999; Kus and Whitfield, 2005). As of 2006, the statewide vireo population was estimated to be approximately 2,500–3,000 territories (U.S. Fish and Wildlife Service, 2006a), of which approximately 10 percent was along the San Luis Rey River between Interstate 15 and Interstate 5.

Male vireos arrive on breeding grounds in southern California in mid-March. Male vireos are vocally conspicuous and frequently sing their diagnostic primary song from exposed perches throughout the breeding season. Females arrive approximately 1–2 weeks after males and are more secretive. Often, females are seen early in the season traveling through habitat with the male. The female, with the male's help, builds an open cup nest in dense vegetation approximately 1 meter (m) above the ground. Clutch size for vireos averages three to four eggs. Typically, the female and male incubate the eggs for 14 days, and young fledge from the nest at 11–12 days of age. It is not unusual for vireos to re-nest after a failed attempt provided ample time remains within the breeding season. Vireos rarely fledge more than one brood in a season, although double brooding can be more common during years when breeding conditions are favorable (early initiation or high early fledging success; Ferree and Kus, 2008b; Houston and others, 2017, 2019). Nesting lasts from early April through July, but adults and juvenile birds remain on the breeding grounds into late September through early October before migrating to their wintering grounds in southern Baja California, Mexico (Kus and others, 2020).

Southwestern Willow Flycatcher

The Southwestern Willow Flycatcher (Empidonax traillii extimus; hereinafter flycatcher) is one of four subspecies of Willow Flycatcher in the United States, with a breeding range that includes southern California, Arizona, New Mexico, extreme southern parts of Nevada, Utah, Colorado, and western Texas (Unitt, 1987; U.S. Fish and Wildlife Service, 2002; Sedgwick, 2020). Restricted to riparian habitat for breeding, the flycatcher has declined in recent decades in response to widespread habitat loss throughout its range and, possibly, brood-parasitism by cowbirds (Remsen, 1978; Unitt, 1987; Schlorff, 1990; Whitfield and Sogge, 1999; U.S. Fish and Wildlife Service, 2002; Sedgwick, 2020). By 1993, the species was believed to number approximately 70 pairs in California (U.S. Fish and Wildlife Service, 1993), in small, disjunct populations. The flycatcher was listed as endangered by the State of California in 1992 and by the USFWS in 1995.

Flycatchers in southern California co-occur with vireos; however, unlike the vireo, which has increased tenfold since the mid-1980s in response to management alleviating these threats (U.S. Fish and Wildlife Service, 2006a), flycatcher numbers have remained low. Currently (2023), most flycatchers in California are concentrated at one site: the upper San Luis Rey River at Lake Henshaw in San Diego County (Howell and Kus, 2024). Outside of this site, flycatchers occur as small, isolated populations of one to six pairs.

Male flycatchers typically arrive in southern California in early to mid-May, whereas females arrive approximately 1 week later. While on the breeding grounds, males sing repeatedly from exposed perches. Once the pair bond is established, the female builds an open cup nest that usually is placed in a branch fork of a willow (Salix spp.) or plant with a similar branching structure approximately 1–3 m above the ground. The typical clutch of three to four eggs is laid in May–June. Females incubate for approximately 12 days, and nestlings fledge within 12–15 days, in early July. Adults usually depart from their breeding territory in mid-August through early September to their wintering grounds in central America and northern South America (U.S. Fish and Wildlife Service, 2002).

San Luis Rey Flood Risk Management Project Area

The San Luis Rey Flood Risk Management Project Area (Project Area) spans approximately 233 hectares (ha; 576 acres) of the lower San Luis Rey River in northwestern San Diego County, California (fig. 1; table 1). Authorized in 1970 and constructed during the late 1980s and early 1990s, the flood control Project Area includes single- and double-levee reaches and six off-channel detention ponds, five of which also serve as mitigation sites for alterations to biological resources within the channel. Operation and maintenance of the flood control project within the channel includes periodic vegetation clearing, non-native plant removal, and sediment removal to ensure that sufficient water conveyance capacity is maintained (U.S. Fish and Wildlife Service, 2006b). Management of the off-channel ponds involves adaptive habitat management, such as vegetation clearing, non-native plant removal, and planting native vegetation.

Riparian vegetation communities at the Project Area include willow-dominated riparian, mixed mule fat (Baccharis salicifolia) and sandbar willow (Salix exigua) riparian scrub, freshwater marsh, and areas dominated by non-native giant reed. Dominant plants include red/arroyo willow (Salix laevigata/Salix lasiolepis; we did not distinguish between these species), black willow (Salix gooddingii), Fremont cottonwood (Populus fremontii), sandbar willow, mule fat, and giant reed. Adjacent habitat and land-use types include coastal sage scrub, non-native grassland, and urban housing and commercial developments. Human disturbances, such as transient camps, recreation, illegal dumping, introduction of invasive non-native plants and feral animals, and use by pets from neighboring houses, are pervasive throughout the Project Area (table 2).

The Project Area includes a channelized 10.7-kilometer (km) section of the lower San Luis Rey River from Interstate 5 to College Boulevard and 6 detention ponds outside of the channel in Oceanside, California (table 1). The Project Area is divided into 12 survey sites, 5 of which are primarily within the flood control channel (hereinafter Channel) and 7 that are outside of the channel (hereinafter Off-channel; table 1). The seven Off-channel sites include one historic restored site north of the levee and west of Whelan Lake and six in the detention ponds. There are three areas that have undergone active restoration (hereinafter Restoration). There are two large Restoration sites: one in Reach 1 (Benet Restoration) and one in Whelan Mitigation (Whelan Restoration). One smaller Restoration site is in Reach 4 (College Restoration).

Before the 2006 vireo breeding season, the U.S. Army Corps of Engineers began two flood risk management activities for the San Luis Rey Flood Risk Management Project: (1) non-native species control and (2) flood risk management of river channel vegetation. Non-native species eradication primarily included removal and control of giant reed, but other non-native species such as pepperweed (Lepidium latifolium), tamarisk (Tamarix spp.), and other non-native tree species also were targeted for removal. Non-native species were physically removed or treated with herbicide. The purpose of the flood risk management was to remove vegetation from the San Luis Rey flood control channel to provide a level of flood conveyance of 2,016 cubic meters per second (m³/s; equivalent to a 150-year flood; U.S. Fish and Wildlife Service, 2006b). In addition, a one-time mowing and mulching task was performed in 2005 to reduce the risk of flooding during the 2006 rainy season. On-going flood risk management of river channel vegetation includes rotational and annual vegetation mowing in phases. Phase 1 includes vegetation clearing in Reaches 1–4 to achieve a minimum flow conveyance of 1,500 m³/s (equivalent to a 100-year flood). Phase 2 expands the area mowed beyond that of Phase 1. Phase 3 includes mowing associated with maintaining the Phase 1 mowing area as well as small sections of vegetation mowing in Reach 1 and Reach 3a. Rotational mowing is done every 5 years, alternating between two 18–23-m-wide strips of vegetation (rotations 1 and 2) such that each strip is mowed every 10 years. Before the 2023 vireo breeding season, mowing occurred in the river channel. In fall 2022, annual Phase 1 and 2 mowing in Reaches 1–4 was completed as part of the overall Flood Risk Management Project efforts. For a detailed description of the timeline and vegetation treatments and management, see appendix 1.

There were three Restoration areas within the Project Area: a non-native plant removal program in Reach 1 (Benet Restoration) that began in October 2012 and two other projects that started in March 2014 (app. 2, figs. 2.1–2.4). The largest project, focused on providing flycatcher habitat, was on the north side of the river channel in the eastern section of the Whelan Mitigation site (Whelan Restoration). A rock levee was removed and graded to below ground level, allowing flooding from the main channel to mimic more natural river flows. A second, smaller project was on the south side of Reach 4 near the College Boulevard bridge (College Restoration). Restoration activities included planting, bi-monthly watering, and spraying weeds with herbicide through 2015. Since 2016, there has been no watering or herbicide treatment, but we continued to classify these areas as Restoration sites and evaluate them separately. In addition to the three Restoration sites, in spring 2014, several small areas of the river were planted with pole cuttings, and in 2017, two of the Off-channel ponds were planted as a part of adaptive habitat management. These areas were considered passive restoration and were not included with our Restoration sites or in any of our analyses of Treatment Index.

Surveys

Vireo and flycatcher surveys have been performed at the Project Area every year since 2006. In 2023, four protocol vireo surveys were completed between April 6 and July 20. The surveys followed standard survey techniques developed and recommended by the California Least Bell’s Vireo Working Group (now known as the California Riparian Birds Working Group) and USFWS Least Bell’s Vireo survey guidelines (U.S. Fish and Wildlife Service, unpub. data, 2001). We supplemented these protocol surveys with weekly territory monitoring visits; therefore, most survey sites were visited more than 10 times throughout the breeding season, resulting in complete coverage of the Project Area. During the surveys and monitoring, we used recorded callbacks (as needed) to confirm the absence of vireos in areas in which no birds were detected and to attract males to check leg bands. We completed four protocol flycatcher surveys of the Project Area between May 15 and July 21, 2023, completing one survey during each of four survey periods (U.S. Fish and Wildlife Service, 2000; Sogge and others, 2010). All field activities were covered under 10(a)1(A) Recovery Permit #ESPER0004080_0.2. This study was classified and approved as a field study by the U.S. Geological Survey Western Ecological Research Center Animal Care and Use Committee.

For both species, observers moved slowly (1–2 kilometers per hour) through the riparian habitat while searching and listening for vireos or flycatchers. Observers walked along the north and south levees to survey the flood control channel. In wider stands, observers traversed the habitat to detect all birds throughout its extent. Surveys were completed between dawn and early afternoon, depending on wind and weather conditions. For each bird encountered, investigators recorded age (adult or juvenile), sex, breeding status (paired, single, unknown, or transient), and banding status. Birds were considered transients if they were not detected on two or more consecutive surveys after an initial detection. Vireo locations were mapped using Environmental Systems Research Institute Field Maps (Environmental Systems Research Institute, 2022) on Samsung Galaxy S7 and S8 and LG G5 mobile phones with Android operating systems and built in Global Positioning System (GPS) to determine geographic coordinates (World Geodetic System of 1984 [WGS 84]). Distance to the nearest surface water was recorded for each flycatcher location. Dominant native and non-native plants were recorded within each vireo and flycatcher territory, and percentage cover of native vegetation was estimated using cover categories of less than 5 percent, 5–50 percent, 51–95 percent, and greater than 95 percent. Overall habitat type was specified according to the following categories:

Nest Monitoring

We monitored vireo nests to evaluate the effects of native and non-native vegetation removal on nest success and productivity. Nests were monitored from April 3 until the last nest was discovered to have failed on August 10, 2023. We monitored at three general site types: (1) within the flood channel where non-native and native vegetation removal has occurred regularly (Channel); (2) at three sites adjacent to the flood channel where limited non-native and native vegetation removal has occurred (Upper Pond, Park Pond, and Whelan Mitigation; Off-channel); and (3) at two sites that have been restored (Restoration), one of which is within the Channel (College Restoration in Reach 4) and one which is situated Off-channel (Whelan Restoration within Whelan Mitigation).

Territory boundaries were delineated (for monitored and unmonitored territories) by biologists in the field by circumscribing vireos’ nesting, singing, and foraging locations. In 2017, we began excluding areas in a vireo’s territory that were obviously avoided (for example, the vireo used the edges of the river channel but avoided the center) by creating two unconnected polygons, as opposed to a single large one spanning the area. We classified territories quantitatively by site type in ArcGIS Pro (version 3.1.3; Environmental Systems Research Institute, 2023) using territory boundaries and the boundaries of the three site types (Channel, Off-channel, and Restoration). If a territory straddled more than one site type, we assigned it to the site type that comprised more than 50 percent of the territory.

There were 58 Channel territories (56 pairs and 2 single males), 24 Off-channel territories (23 pairs and 1 single male), and 2 Restoration territories (1 pair and 1 single male) that were monitored during the 2023 breeding season. Pairs were observed for evidence of nesting, and their nests were located. We visited nests as infrequently as possible to minimize the chances of leading predators or cowbirds to nests; typically, there were three to four visits per nest. The first visit was timed to determine the number of eggs laid, the next visits to determine hatching and age of young, and the last to band nestlings (see the “Banding” section). We removed cowbird eggs from nests depending on when they were found. In nests with fewer than three vireo eggs, cowbird eggs were removed no sooner than the seventh day of incubation to minimize the possibility of nest abandonment in response to the removal. We removed cowbird eggs from nests that contained three or more vireo eggs as they were detected. Cowbird nestlings were removed immediately from nests. Fledging was determined through direct observation of fledglings in the territory or, in some rare cases, inferred from an accumulation of feather dust and fecal material in the nest, which is indicative of vireo fledging. We recorded characteristics of nests, including nest height, host plant species, host height, distance to edge of host plant, and distance to edge of host clump after abandonment or fledging of nests.

Analysis of Nesting Data

We used Pearson’s chi-square analysis and Fisher’s exact tests to determine if there were differences in nest success (proportion of nests that fledged at least one young), hatching rates (proportion of eggs and nests that hatched), and fledging rates (proportion of nestlings and nests that fledged) among Channel, Off-channel, and Restoration pairs. We included a Mixed category for instances where (1) a territory was categorized as one site type, but the vireos placed at least one nest in another site type or (2) vireos nested in two different site types within the same territory. For example, we had territories that were categorized as Restoration (50 percent or more of the total territory area fell within the boundaries of the Restoration), but the vireos placed their nest(s) within the Channel, avoiding areas of active restoration. In total, nine territories were considered Mixed in 2023. For analyses involving the nest as the unit of analysis, we used the location of the nest (for example, a territory categorized as Restoration could have a nest in a Channel site). In 2023, we had four nests in three territories that fell outside of the Project Area boundaries. These nests were included in the estimates of nesting success and productivity by pair but were excluded from other analyses. For any analyses of reproductive success or productivity that involved a measure of success per pair or territory, we analyzed Mixed territories separately. Chi-square tests were used when sample sizes were sufficient; Fisher’s exact tests were used when one or more categories contained fewer than five samples. We used two-sample t-tests, Mann-Whitney U-tests (two groups) and Kruskal-Wallis tests (three or more groups) to determine if there were differences in nest site characteristics between successful and unsuccessful nests within and among the Channel, Off-channel, and Restoration sites. When distributions were normal and variances were similar, t-tests were used; Mann-Whitney U-tests or Kruskal-Wallis tests were used when the data violated these assumptions. We used Analysis of Variance and Tukey’s post-hoc pair-wise comparisons to determine if there were differences among Channel, Off-channel, and Restoration sites in average clutch size and average number of young per pair in 2023 and average clutch size and number of young fledged per pair by year (2006–23).

Effects of Treatment

We created an index of treatment (Treatment Index) to evaluate the cumulative effects of vegetation removal over time for 2022 and 2023 vireo territories. We restricted treatment type for this index to the vegetation removal that occurred from 2005, 2007–12, 2014–15, 2017–19, or 2021–22, in which vegetation (native or non-native) was cleared to the ground. We did not include any restoration or herbicide treatments because these treatments were distinct from vegetation removal, and analysis of them was beyond the scope of this project. We limited our analysis of Treatment Index to territories in the Channel because Off-channel adaptive habitat management activities mitigated the effect of the vegetation removal, making the Treatment Index for those territories unsuitable for our analyses. We calculated the percentage of each territory area that was treated in each year (2005, 2007–12, 2014–15, 2017–19, and 2021–22) by overlaying the 2022 and 2023 territory boundaries with the treatment areas in ArcGIS Pro (version 3.1.3; Environmental Systems Research Institute, 2023). Territory boundaries were created to include the bird location points collected using GPS throughout the field season. Treatment boundaries were generated from the location data collected using GPS in the field by a RECON Environmental, Inc. (https://www.recon-us.com/) biologist who walked the treatment boundary during vegetation removal activities. Because treatments have been occurring since 2005, it was possible to have multiple years of treatments as well as overlap among treatments within a single territory. Because early successional vegetation can recover quickly after disturbance, we assumed that the effect of the vegetation removal diminished over time; therefore, when there was overlap in treated areas among years, we used the most recent treatment year to calculate the Treatment Index (fig. 2). To calculate the Treatment Index, we used the percentage of the territory area treated weighted by the squared inverse of the time since treatment (t²) summed across all treatment years. We chose t² rather than t because we wanted to capture the effect of the treatment diminishing quickly over time. We used the following formula:

We performed three sets of analyses using the Treatment Index. For all analyses of the Treatment Index, we excluded territories that overlapped with areas of active restoration:

Nest Survival Analysis

We used RMark (Laake, 2013) to model the effects of vegetation treatment on daily survival rate (DSR) of vireo nests (or the probability that a nest survived from one day to the next; Dinsmore and others, 2002). We used RMark (Laake, 2013) to run program MARK (White and Burnham, 1999) to calculate DSR, which accounts for the variability in exposure days across nests detected at different stages of the nesting cycle and allows for the analysis of the effects of covariates on DSR. In this study, we evaluated the effects of Treatment Index (for Channel territories only) on DSR in 2023. We studied habitat structure at the nest (2006–23) to determine if it was correlated with DSR by analyzing the relationship between total vegetation cover at each height category (see the “Vegetation Study Design” section) and DSR.

We calculated nest survival across a 32-day cycle (2 days to complete the nest and rest before the first egg is laid, 4 days egg-laying, 14 days incubation, and 12 days nestling period). We calculated the age of nests at the time they were discovered by forward- or backward-dating of nests in relation to known dates of nest building, laying, or hatching. We used an information-theoretic approach (Burnham and Anderson, 2002) to evaluate support for statistical models reflecting a priori hypotheses regarding the effect of treatment and habitat variables on DSR. We hypothesized that DSR would be inversely related to the Treatment Index and would increase with increasing foliage cover at nest sites, particularly understory cover within 2 m of the ground, where vireos place their nests. We used logistic regression with a logit link to build models. First, we generated a constant survival model to serve as a reference for the effect of treatment and habitat variables on DSR. We then modeled the treatment and habitat covariates individually and evaluated support for each model in relation to the constant survival model and each other. To evaluate the effect of covariates within our top models, we calculated the odds ratio for each covariate (the odds that the covariate had an effect on DSR where “no effect” equals 1, negative effect is less than 1, and positive effect is greater than 1). We then calculated the 95-percent confidence interval of the odds ratio to determine the likelihood that the effect was significant. Where the confidence interval did not include 1 (was greater than or less than 1), we concluded that we had 95-percent confidence that the covariate had an effect on DSR relative to the reference.

Banding

The primary goals of banding vireos at the San Luis Rey Flood Risk Management Project Area were to (1) assess vireo site fidelity in response to vegetation management, (2) examine natal dispersal within and outside of the Project Area in response to vegetation management, and (3) understand how vegetation removal and alteration affected vireo demography. From 2006 through 2022, we banded nestlings from monitored nests at 6–7 days of age with a single, dark blue anodized numbered federal band. We captured adult vireos in mist nets at monitoring sites and banded them with a unique combination of colored plastic and anodized metal federal bands, including an anodized dark blue band to designate the San Luis Rey River as the bird’s site of origin. Adults previously banded with a single numbered metal federal band as nestlings (natal birds) were target netted to determine their identity, and their original band was supplemented with other bands to generate a unique color combination. These data supplement banding data currently being gathered by the USGS and other investigators regarding nearby vireo populations on the upper San Luis Rey River and the Santa Margarita River on Marine Corps Base Camp Pendleton (MCBCP).

Vegetation Study Design

We sampled vegetation along permanent linear transects within three of the vireo monitoring sites (Channel, Upper Pond, and Whelan Restoration). Sampling points consisted of 2- by 2-m quadrats located at 10-m intervals along each transect; the number of points sampled varied with the length of each transect (fig. 3). Transects were originally established in 2006 using a systematic sampling design. To capture the range of variability of riparian vegetation structure and composition, we positioned transects perpendicular to the river channel. To provide uniform coverage, we placed transects at fixed distances from each other; distances varied with the size of the site. In the Channel, transects were placed at 200- or 400-m intervals depending on the width of the river. In Upper Pond, transects were placed every 100 m. In addition, we sampled two 350-m transects at Whelan Mitigation (Whelan Restoration) that were initially surveyed from 1991 to 1993 to monitor riparian restoration by the U.S. Army Corps of Engineers (Kus, 1998). The Whelan Restoration transects were 75 m apart and oriented approximately parallel (320 degrees) to the flood control channel. Before 2014, this site was sampled and used as an Off-channel site in analyses. However, in 2014, it became a Restoration site and has been analyzed separately from the Channel and the Upper Pond sites since then.

We used several permanent and semi-permanent methods to ensure that transects could be re-sampled in subsequent years. First, a 1.5-m metal rebar was driven into the ground, leaving 75 centimeters (cm) above ground to mark the start of each transect. We spray-painted the rebar pink or orange and placed them at the intersection of the south levee and the riverbed. From the rebar, using a compass and tape measure, two field personnel measured the distances between sampling points. A numbered wooden stake, spray-painted pink or orange, was placed in the ground, and colored plastic flagging was tied nearby to aid in locating the points. Finally, we recorded geographic coordinates for each rebar and wooden stake using a GPS unit (Garmin GPS 12).

Vireo Habitat Use Study Design

In addition to sampling vegetation along transects, we collected vegetation data at 45 randomly selected vireo territories. We sampled one nest and one paired random plot within each territory (hereinafter nest plots and territory plots, respectively) in the Channel and at Upper Pond. We did not sample vegetation at Restoration site nests because the sample was small, and our focus was to determine the response of vireos to treatments (measured by the Treatment Index) rather than their response to restoration activities. Nest and territory plots consisted of four 2- by 2-m quadrats; one quadrat was centered on the nest (or center for random plots), and the remaining three quadrats were located 10 m from the nest/center and oriented at 0, 120, and 240 degrees from it (fig. 4). Territory plot locations were constrained using the Buffer tool in ArcGIS Pro (version 3.1.3; Environmental Systems Research Institute, 2023) to create a 10-m negative buffer on the territory boundaries. This negative buffer ensured that the random sampling point for the quadrats at 0, 120, or 240 degrees would not fall outside of the territory boundaries. A positive 2-m buffer was created around nest locations and vegetation transect points to exclude these locations and prevent duplicate vegetation sampling. We then used the Create Random Points tool on the territory boundary buffer to select one random point between the buffered areas to serve as the center quadrat of the territory plot. Territories were excluded from analysis if they were too narrow to allow a 10-m negative buffer, were in the Mixed category, or overlapped with Restoration.

Least Bell’s Vireo

Population Size and Distribution

We identified a total of 136 vireo territories during surveys and weekly territory monitoring (table 3; app. 2, figs. 2.1–2.4). The number of vireo territories increased from 133 in 2022, a 2-percent increase (table 4; fig. 5). Of the 136 territorial males, 121 (89 percent) were confirmed as paired, and 4 (3 percent) were confirmed as single males. We were unable to confirm breeding status for the remaining 11 territorial males. We detected two transient vireos during surveys in 2023 (table 3).

Of the territories, 70 percent (95/136) were within the Channel, and the remaining 30 percent (41/136) were located Off-channel. Nine Channel territories were in Benet Restoration. The remaining 86 Channel territories were outside of Restoration areas. Of the Off-channel territories, 2 were within Whelan Restoration, 11 were within Whelan Mitigation but outside of the Restoration area, and 28 were in other retention ponds (tables 3, 4; app. 2, figs. 2.1–2.4).

Table 3.    

Number and breeding status of Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Table 3.    Number and breeding status of Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Survey site	Known pairs	Unknown status	Single males	Total territories	Total transients
Channel sites
Reach 1	19	6	0	25	0
Benet Restoration	8	1	0	9	0
Reach 2	12	0	0	12	0
Reach 3a	13	0	0	13	0
Reach 3b	7	0	0	7	0
Reach 4	26	1	2	29	1
College Restoration	0	0	0	0	0
Off-channel sites
Lower Pond	6	2	0	8	0
Tuley Canyon	1	0	0	1	0
Whelan Mitigation	11	0	0	11	0
Whelan Restoration	1	0	1	2	1
Park Pond	5	1	0	6	0
Pilgrim Pond	1	0	0	1	0
Upper Pond	11	0	0	11	0
Riverside Pond	0	0	1	1	0
Total	121	11	4	136	2

Table 4.    

Number and distribution of Least Bell’s Vireo territories at the San Luis Rey Flood Risk Management Project Area, California, in 2006–23.

[—, no data]

Table 4.    Number and distribution of Least Bell’s Vireo territories at the San Luis Rey Flood Risk Management Project Area, California, in 2006–23.

Survey site	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021	2022	2023
Channel sites
Reach 1	13	7	10	23	23	17	10	13	14	22	15	22	27	17	26	21	27	25
Benet Restoration	—	—	—	—	—	—	—	6	6	2	4	4	5	6	6	5	6	9
Reach 2	15	14	16	18	22	16	6	6	6	3	6	7	12	13	12	12	11	12
Reach 3a	14	19	23	28	26	17	14	18	16	12	8	10	15	16	17	10	16	13
Reach 3b	13	14	13	17	16	18	5	4	6	5	4	7	11	10	16	12	9	7
Reach 4	21	21	25	32	24	18	10	14	15	16	12	10	20	22	27	26	27	29
College Restoration	—	—	—	—	—	—	—	—	2	0	0	0	0	0	1	0	2	0
Total	76	75	87	118	111	86	45	61	65	60	49	60	90	84	105	86	98	95
Off-channel sites
Lower Pond	3	2	3	7	7	4	1	7	7	7	6	7	9	8	7	7	8	8
Tuley Canyon	1	1	1	1	1	0	0	0	1	2	2	1	1	1	2	1	1	1
Whelan Mitigation	14	9	12	15	16	12	11	13	6	6	4	6	8	13	15	7	7	11
Whelan Restoration	—	—	—	—	—	—	—	—	4	3	3	3	5	3	5	3	3	2
Park Pond	3	2	4	4	4	6	3	5	5	4	3	4	6	6	6	5	5	6
Pilgrim Pond	1	1	1	1	0	2	0	2	0	0	0	1	1	1	1	1	1	1
Upper Pond	20	17	21	24	17	19	15	14	11	12	11	11	10	12	18	12	10	11
Riverside Pond	1	1	1	1	1	1	1	1	1	1	0	0	0	0	2	0	0	1
Total	43	33	43	53	46	44	31	42	35	35	29	33	40	44	56	36	35	41
Grand total	119	108	130	171	157	130	76	103	100	95	78	93	130	128	161	122	133	136

From 2022 to 2023, the number of vireo territories decreased by 3 percent (from 98 to 95) in the Channel and increased by 17 percent (from 35 to 41) within Off-channel sites (table 4). Vireo territory numbers decreased in Reach 3a and Reach 3b by three and two territories, respectively. Reach 1 (including Benet Restoration) increased by one territory, and Reach 4 (including College Restoration) remained stable. Most Off-channel sites remained stable or increased by one territory, with the exception of Whelan Mitigation (including Whelan Restoration), which increased by three territories for a 30-percent increase (table 4).

Nest Monitoring

We monitored nesting activity in 84 territories within the San Luis Rey River Flood Risk Management Project Area (table 5; app. 2; figs. 2.1–2.4; app. 3). These territories were fully monitored, meaning that all nests within the territory were located and monitored during the breeding season. Of the 84 monitored territories, 4 were occupied by single males and were excluded from nesting analyses. A total of 169 nests were monitored during the breeding season; however, 4 of these were located outside of the Project Area and were only included in estimates of pair success and were not otherwise analyzed.

Table 5.    

Number of Least Bell’s Vireo territories and nests monitored at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

[Mixed territories had a territory classification that differed from one or more nest classifications. Incomplete nests were partially built, but not completed. Abbreviations: —, no data; ±, plus or minus]

Table 5.    Number of Least Bell’s Vireo territories and nests monitored at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Territories/nests	Channel	Off-channel	Restoration	Mixed1	Total
Territories	52	21	2	9	84
Total number of pairs	50	20	1	9	80
Total number of nests monitored2	121	42	32	—	165
Incomplete nests	4	1	0	—	5
Total number of completed nests2	117	41	2	—	160
Completed nests/pair	2.0±1.0	1.8±0.8	2.0±0	2.8±1.3	1.9±0.8
Total number of nest attempts/pair	2.1±1.0	1.8±0.9	2.0±0	2.9±1.5	2.1±1.1

¹

Nine territories had a territory classification that differed from one or more nest classifications (three territories were classified as Channel, but had nests located Off-channel; three territories were classified as Off-channel but nested in the Channel; three territories were classified as Channel but nested outside of the Project Area on the other side of the levy).

²

Four nests by Mixed pairs that fell outside of the Project Area are not included.

³

All nests were in Whelan Restoration.

Nesting Attempts

In 2023, the average number of nesting attempts per pair (including incomplete nests) was 2.1±1.1 (table 5). A total of 73 percent (58/80) of pairs re-nested after their first nest attempt, and 41 percent (24/58) of these pairs initiated at least a third nesting attempt. A total of 24 pairs (39 percent, 24/62) attempted a re-nest after a successful nest; of these, 12 pairs (50 percent, 12/24) successfully fledged 2 broods. Overall, 15 percent of pairs successfully double-brooded in 2023.

Most of the first nesting attempts in 2023 were initiated in mid- to late-April (fig. 6). The peak of nest initiation was during the week of April 23. Between the weeks of April 16 and April 30, 71 percent of vireos began nesting. Nesting began earlier in 2023 (median Julian day 117; April 27) compared to 2022 (median Julian day 120; April 30) and was similar to the 17-year average (mean median Julian day 118.5; April 28–29, 2006–22; fig. 7).

Apparent Nest Success

Overall, 46 percent (74/160) of completed nests were successful (table 6). A total of 46 percent (54/117) of nests in the Channel, 46 percent (19/41) of Off-channel nests, and 50 percent (1/2) of Restoration nests fledged young. Nest success did not differ statistically among the three site types (P=1.0, Fisher’s Exact test).

Causes of nest failure were similar among the Channel, Off-channel, and Restoration site nests. Predation was the primary source of nest failure for all sites (table 6). Predation accounted for 63 percent (40/63), 77 percent (17/22), and 100 percent (1/1) of nest failures in the Channel, Off-channel, and Restoration sites, respectively. Most predators were believed to be birds, mammals, or snakes. Overall, 34 percent (40/117) of Channel nests, 41 percent (17/41) of Off-channel nests, and 50 percent (1/2) of Restoration nests were lost to predation.

Table 6.    

Fate of completed Least Bell’s Vireo nests at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

[Proportion of total completed nests shown in parentheses]

Table 6.    Fate of completed Least Bell’s Vireo nests at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Nest fate	Number of nests
	Channel	Off-channel	Restoration	Total
Successful	54	19	1	74	(0.46)
Failed
Predation	40	17	1	58	(0.36)
Parasitism	1	0	0	1	(0.01)
Other/unknown	22	5	0	27	(0.17)
Failed total	63	22	1	86	(0.54)
Total completed nests	117	41	2	160	(1.00)

Although we could attribute most nest failures to predation in 2023, there were 27 nests that failed for other or unknown reasons. Of these nests, 2 failed with infertile eggs, 1 was found after it had already failed, 2 were found with vireo eggs under the nest, 2 failed when a branch supporting the nest broke, 4 were abandoned with eggs for unknown reasons, 1 was found with a dead nestling in the nest, and 15 nests failed for unknown reasons before eggs were observed.

Brown-headed Cowbird Parasitism

We documented 16 instances of Brown-headed Cowbird parasitism in 2023, a 60-percent decrease from 2022 (40 instances). In total, 16 nests (10 percent of completed nests) in 14 territories (18 percent of territories) were parasitized, slightly higher than the 17-year average of 7 percent of completed nests parasitized (table 7; fig. 8). One nest failed as a result of parasitism (abandoned). We removed 16 cowbird eggs from 16 nests, of which 9 (56 percent) were ultimately successful. Of the remaining parasitized nests, four were depredated, and we found broken eggs under two nests that failed after the removal of a cowbird egg for unknown reasons.

Table 7.    

Number and fate of Least Bell’s Vireo nests parasitized by Brown-headed Cowbirds at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Table 7.    Number and fate of Least Bell’s Vireo nests parasitized by Brown-headed Cowbirds at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Nest fate	Channel	Off-channel	Restoration	Total
Nests parasitized	12	4	0	16
Total cowbird eggs laid	12	4	0	16
Fate of parasitized nests
Abandoned	1	0	0	1
Not abandoned
Successful	7	2	0	9
Unsuccessful	4	2	0	6

Parasitism was first observed during the third week of nesting (April 16), when 11 percent of completed nests were parasitized. The average weekly parasitism rate for completed nests from April 2 to July 9, 2023, was 9 percent (range 0–25 percent; fig. 9).

Reproductive Success and Productivity

Hatching and fledging success were similar in 2023 and 2022. Overall, 69 percent of nests with eggs hatched at least 1 egg, and 75 percent of nests with nestlings fledged at least 1 young (table 8), compared to 60 and 77 percent, respectively, in 2022. Overall, clutch size of non-parasitized nests (3.7±0.5) was higher than in 2022 (3.2±0.5) and higher than the 17-year average (2006–22; 3.4±0.2). In 2023, a greater percentage of hatchings fledged in the Channel (81 percent) compared to Off-channel (69 percent). There were no other statistical differences among Channel, Off-channel, and Restoration sites in clutch size, hatching, or fledging success.

Table 8.    

Reproductive parameters for nesting Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

[Nests were categorized based on their locations within Channel, Off-channel, and Restoration sites rather than based on the pair’s territory site category. Standard deviations presented with means. Channel includes Reaches 1, 2, 3a, 3b, and 4, excluding restoration sites. Off-channel includes Upper Pond and Whelan Mitigation, excluding restoration sites. Restoration includes Benet Restoration, College Restoration, and Whelan Restoration. Asterisks (*) indicate a significant difference. Abbreviations: ±, plus or minus; %, percent]

Table 8.    Reproductive parameters for nesting Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Parameter	Channel	Off-channel	Restoration	Overall
Nests with eggs	102	39	2	143
Eggs laid	360	139	7	506
Average clutch size1	3.7±0.5	3.7±0.3	4.0±0.0	3.7±0.5
Hatching success
Hatchlings	221	87	3	311
Nests with hatchlings	70	28	1	99
Eggs2	61%	63%	43%	61%
Nests3	69%	72%	50%	69%
Fledging success
Fledglings	178	60	3	241
Nests with fledglings	54	19	1	74
Hatchlings4	81%	69%	100%	77%*
Nests5	77%	68%	100%	75%

¹

Based on 77 Channel, 31 Off-channel, and 1 Restoration nests with a full clutch (One-way analysis of variance: F-ratio=0.26, P=0.77).

²

Percentage of all eggs that hatched, among all three sites: P=0.59, Fisher’s exact test.

³

Percentage of all nests with eggs in which at least one egg hatched, among all three sites: P=0.71, Fisher’s exact test.

⁴

Percentage of all hatchlings that fledged, among all three sites: P=0.06, Fisher’s exact test (Channel vs. Off-channel chi-squared=4.13, degrees of freedom=1, P=0.04; other comparisons not significant).

⁵

Percentage of all nests with hatchlings in which at least one young fledged, among all three sites: P=0.58, Fisher’s exact test.

In 2023, Project Area vireos fledged 3.1±2.1 young per pair, higher than productivity in 2022 (2.2±1.7 young per pair) and higher than the 17-year average (2.5±0.7, table 9; fig. 10). There was no difference in productivity (the number of young fledged per pair) or in the proportion of pairs fledgling one or more young among site types. Overall, 79 percent (63/80) of pairs fledged at least 1 young in 2023, higher than in 2022 (72 percent) and the 17-year average of 74 percent. In 2023, 15 percent of pairs (12/80) successfully fledged 2 broods, higher than the 5 percent observed in 2022 and slightly higher than the 17-year average of 12 percent. Without the rescue effect of removing Brown-headed Cowbird eggs from parasitized nests, thus allowing them to potentially fledge young, productivity would have been lower. Had cowbird eggs not been removed from parasitized nests, the fledging rate would have been 2.8±2.2 vireo young per pair, with 70 percent of pairs successfully fledging at least 1 young.

Table 9.    

Productivity of nesting Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

[Standard deviations presented with means. Values in parentheses represent percentages by site type. Abbreviations: ±, plus or minus; %, percent]

Table 9.    Productivity of nesting Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, in 2023.

Parameter	Channel	Off- channel	Restoration	Mixed	Overall
Number of pairs	50	20	1	9	80
Average number of young fledged per pair1	3.2±1.9	2.6±2.5	3.0±0.0	3.3±2.2	3.1±2.1
Pairs fledging at least one young2	41 (82%)	14 (70%)	1 (100%)	7 (78%)	63 (79%)
Pairs fledging two broods3	9 (18%)	3 (15%)	0 (0%)	0 (0%)	12 (15%)

¹

One-way analysis of variance: F-ratio=0.085, P=0.92.

²

P=0.30, Fisher’s exact test.

³

P=0.64, Fisher’s exact test.

Effects of Treatment on Vireo Reproduction

In 2023, the Treatment Index for fully monitored vireo territories within the Channel that did not overlap with active restoration averaged 0.07±0.15 (n=58, range, 0–0.75; fig. 11). There were four fully monitored vireo territories (7 percent, 4/58) within the Channel that had no treatment. There was no effect of Treatment Index on any measures of nesting effort or productivity (number of nests, number of completed nests, or number of fledglings per pair; table 10; fig. 12).

Table 10.    

Generalized linear model results, parameter estimates, standard errors (SE), Z statistics, and P-values for Least Bell’s Vireo per-pair reproductive parameters as a function of Treatment Index within the Channel at the San Luis Rey Flood Risk Management Project Area, California, 2023.

[<, less than]

Table 10.    Generalized linear model results, parameter estimates, standard errors (SE), Z statistics, and P-values for Least Bell’s Vireo per-pair reproductive parameters as a function of Treatment Index within the Channel at the San Luis Rey Flood Risk Management Project Area, California, 2023.

Variable	Estimate	SE	Z	P
Number of nests
Intercept	0.8	0.1	7.6	<0.01
Treatment Index	0.3	0.5	0.6	0.54
Number of completed nests
Intercept	0.7	0.1	6.9	<0.01
Treatment Index	0.2	0.6	0.3	0.73
Number of fledglings
Intercept	1.1	0.1	13.3	<0.01
Treatment Index	0.3	0.4	0.8	0.44

We analyzed 2023 nest survival within the Channel to determine if there was an effect of Treatment Index on DSR. The best model for DSR was the constant model (table 11). There was some support for the model that included Treatment Index (∆AIC_c less than 2); however, an analysis of the odds ratio indicated that it was not a significant contributor to the model (β=0.19, SE=0.8, odds ratio=1.21, 95-percent confidence interval=0.26–5.71).

Table 11.    

Logistic regression models for the effects of Treatment Index (within the Channel only) on daily nest survival of Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, 2023.

[Models are ranked from best to worst based on Akaike’s Information Criterion for small samples (AIC_c), change in AIC_c (ΔAIC_c), and AIC_c weights. Akaike’s Information Criterion for small samples is based on −2×log_e likelihood, the number of parameters in the model, and sample size]

Table 11.    Logistic regression models for the effects of Treatment Index (within the Channel only) on daily nest survival of Least Bell’s Vireos at the San Luis Rey Flood Risk Management Project Area, California, 2023.

Model	Deviance	Number of parameters	AICc	ΔAICc	AICc weight
Constant	335.4	1	337.4	0.0	0.73
Treatment Index	335.4	2	339.4	1.9	0.27

Nest Survival and Habitat Use

We investigated the effect of habitat structure on vireo nest survival by constructing a series of models relating the vegetation cover variables at nest sites to the daily nest survival rate (table 12). The best model was the constant DSR model; however, there was support for several of the other models (ΔAIC_c less than or equal to 2). We looked at the beta estimates for the first two models: the model that included percent cover 0–1 m+1–2 m and the model that included percent cover 1–2 m. For both of these models, the beta values for the percentage cover estimates were low (0–0.01), and the confidence interval for the odds ratio spanned 1, so we concluded that percent vegetation cover 0–1 m and 1–2 m were not significant contributors to the models (table 13).

Table 12.    

Logistic regression models for the effects of habitat structure on the daily survival rate of Least Bell’s Vireo nests at the San Luis Rey Flood Risk Project Area, California, 2006–23.

[Models are ranked from best to worst based on Akaike’s Information Criterion for small samples (AIC_c), change in AIC_c (ΔAIC_c), and AIC_c weights. Akaike’s Information Criterion for small samples is based on −2×log_e likelihood, the number of parameters in the model, and sample size. Abbreviations: DSR, daily survival rate; m, meter; +, plus]

Table 12.    Logistic regression models for the effects of habitat structure on the daily survival rate of Least Bell’s Vireo nests at the San Luis Rey Flood Risk Project Area, California, 2006–23.

Model	Deviance	Number of parameters	AICc	ΔAICc	AICc weight
Constant DSR	2,373.0	1	2,375.0	0.0	0.2
Percent cover 0–1 m+1–2 m	2,369.8	3	2,375.8	0.9	0.1
Percent cover 1–2 m	2,372.1	2	2,376.1	1.1	0.1
Percent cover 5–6 m	2,372.7	2	2,376.7	1.7	0.1
Percent cover 6–7 m	2,372.7	2	2,376.7	1.7	0.1
Percent cover 3–4 m	2,372.7	2	2,376.7	1.8	0.1
Percent cover 4–5 m	2,372.8	2	2,376.8	1.9	0.1
Canopy height	2,372.9	2	2,376.9	1.9	0.1
Percent cover 0–1 m	2,372.9	2	2,376.9	1.9	0.1
Percent cover 7–8 m	2,372.9	2	2,376.9	2.0	0.1
Percent cover 2–3 m	2,373.0	2	2,377.0	2.0	0.1
Percent cover 2–3 m+3–4 m+ 4–5m+5–6 m+6–7 m+7–8 m	2,371.7	7	2,385.7	10.8	0.0

Table 13.    

Parameter estimate (β), standard error (SE), odds ratios, and 95-percent confidence intervals (CI) for models explaining the effect of habitat structure on the daily survival rate of Least Bell’s Vireo nests at the San Luis Rey Flood Risk Management Project Area, California, 2006–23.

[Models are in order of best-supported to least-supported. Abbreviations: m, meters; +, plus]

Table 13.    Parameter estimate (β), standard error (SE), odds ratios, and 95-percent confidence intervals (CI) for models explaining the effect of habitat structure on the daily survival rate of Least Bell’s Vireo nests at the San Luis Rey Flood Risk Management Project Area, California, 2006–23.

Effect	β	SE	Odds ratio	95-percent CI
Percent cover 0–1 m+1–2 m
Intercept	3.80	0.15	44.73	33.51–59.71
0–1 m	−0.01	0.01	0.99	0.98–1.00
1–2 m	0.01	0.01	1.01	1.00–1.02
Percent cover 1–2 m
Intercept	3.72	0.14	41.37	31.56–54.22
1–2 m	0.00	0.00	1.00	1.00–1.01

Southwestern Willow Flycatcher

Vegetation Structure and Composition

A total of 46 transects (516 sampling points) were sampled in the Project Area in 2023 (table 28; app. 9, figs. 9.1–9.4). Of these points, 71 percent (368/516) were in the Channel, and 22 percent (113/516) were located Off-channel in Upper Pond. The remaining 7 percent (35/516) of points were in the Whelan Restoration site. The number of points per transect varied between 4 and 18. Ten points were in transient camps and were not sampled in 2023. Global Positioning System coordinates for the start and end point of each transect are provided in appendix 10.

Vegetation Structure and Composition

At all sites, there was generally more vegetation cover below 1 m with cover decreasing with increasing height (fig. 13). However, foliage cover from 1 to 3 m was similar across sites. Overall, foliage cover was low above 3 m but was highest in the Channel. Average live canopy height and maximum canopy height (including live and dead vegetation) were similar to each other at all sites and were highest in the Channel (5.6±3.8 m and 5.6±3.8 m, respectively) compared to Upper Pond (4.7±2.8 m and 4.7±2.7 m) and Whelan Restoration (3.9±2.0 m and 4.0±2.0 m).

Vegetation composition in the Channel differed from that in Upper Pond and Whelan Restoration (fig. 14). Tree cover, dominated by red/arroyo willow, made up a higher proportion of total vegetation cover in the Channel (46 percent) relative to Upper Pond (20 percent) and Whelan Restoration (28 percent). Conversely, shrub cover made up a higher proportion of the total cover at Upper Pond (41 percent) and Whelan Restoration (39 percent) compared with the Channel (11 percent). Mule fat was the dominant shrub species at all three sites. Herbaceous cover was at least 2 to 3 times greater in both the Channel (24 percent) and Upper Pond (17 percent) than in Whelan Restoration (8 percent). Marsh species, such as bulrush (Scirpus sp.), sedge (Carex sp.), and cattail (Typha latifolia), comprised 58 percent of the total herbaceous cover in the Channel. Non-native species cover also was higher in both the Channel (17 percent) and Whelan Restoration (21 percent) compared to Upper Pond (4 percent). In Whelan Restoration, the non-native cover was entirely herbaceous species. Non-native cover in the Channel was almost half (47 percent) giant reed. Dead woody cover was highest in Upper Pond (13 percent) compared to Whelan Restoration (4 percent) and the Channel (2 percent). Dead vegetation in Upper Pond was primarily red/arroyo willow, whereas most dead vegetation at Whelan Restoration and the Channel could not be identified.

Vegetation Changes 2006–23

When we examined the cumulative effects of management activities on vegetation structure over time at the Project Area, we detected significant changes from 2006 to 2023 in the Channel (figs. 15A–15C). Overall, for all height classes over 3 m, foliage cover has decreased and remains below historic levels. We detected significant negative trends in average percentage of foliage cover for the 2–3 m height class and all height classes greater than 6 m (2–3 m: R²=0.16, P=0.05 [fig. 15A]; 6–7 m: R²=0.18, P=0.04; 7–8 m: R²=0.25, P=0.02; greater than 8 m: R²=0.53, P<0.001 [fig. 15C]). Decreases were steeper during 2008–13 and 2014–16 than during 2006–08. Average percentage of foliage cover was lowest at all height categories in 2016. Since 2016, percentage of foliage cover increased at all height classes. Herbaceous cover more than doubled between 2016 (13 percent) and 2021 (29 percent) and has remained high through 2023 (24 percent) and accounts for most of the foliage cover observed in the 0–1 m height class. Generally, vegetation cover below 2 m has remained high. In contrast, despite increases since 2016, percentage of foliage cover over 2 m has remained lower than levels detected before 2009.

At Upper Pond, especially in the lower height classes, we detected very few trends or patterns with regard to vegetation cover from 2006 to 2023 (figs. 16A–16C). Vegetation cover under 4 m has fluctuated annually with no discernable pattern, and in 2023, was comparable to what was observed in 2006 (fig. 16A). For the higher strata, which represented only a small fraction of the overall cover at Upper Pond, we detected a significant negative trend from 2006 to 2023 at the 4–5 m height class (4–5 m: R²=0.25, P=0.02; fig. 16B).

At Whelan Restoration, we detected a significant positive trend for the 5–6 m height class (R²=0.14, P=0.07 [fig. 17B]) and a significant negative trend above 8 m (R²=0.32, P=0.01); however, overall foliage cover in these strata make up only a small percentage of the total cover at this site (fig. 17C). The general pattern observed at Whelan Restoration has been one of decline and recovery, with foliage cover declining at all height classes from 2006 to 2010, after which cover has steadily increased at most height classes. Current (2023) cover is similar to 2006 levels for most height classes (figs. 17A–17C).

Vireo Habitat Use

We measured vegetation characteristics at 36 nest plots and 36 territory plots (288 sampling points) within the Channel and 9 nest and 9 territory plots (72 sampling points) at Upper Pond and compared these with the available habitat (transect data). In the Channel, we determined that vireos were selective with regard to territory, establishing territories in areas with greater foliage cover between 3 and 7 m and less cover below 1 m compared to available habitat. With regard to nest selection, we determined that vireos selected nest sites within their territories largely at random, with the exception that nest sites had less foliage cover between 4 and 5 m than in the rest of the territory (fig. 18). Vireos in Upper Pond were also selective with regard to territory establishment, although below 1 m, selectivity was opposite that of vireos in the Channel, with birds selecting territories with more foliage cover relative to the available habitat. In Upper Pond, vireos established territories in areas with greater foliage cover below 4 m and between 5 and 6 m. However, vireos selected nest sites generally at random, with the exception that nest sites had less foliage cover below 1 m compared to available habitat within the territory (fig. 19).

Southwestern Willow Flycatcher

There were no breeding pairs of flycatchers detected in the Project Area in 2023, which was the 17th year that the Project Area has not been occupied by Willow Flycatchers since it was colonized and monitoring began in 2000 (Kus and Rourke, 2005). It is unknown why flycatchers did not breed in the Project Area, although conditions in the historical flycatcher territories have been gradually changing since 2006, when three pairs occupied the Project Area. At Whelan Restoration in particular, large canopy trees, such as red/arroyo willows and black willows, have died. This historic breeding habitat is likely still too dry, sparse, and open for flycatchers.

Vegetation clearing does not seem to have played a role in the decline of the flycatcher in the San Luis Rey Flood Risk Management Project Area population. In addition to habitat change, more likely explanations include other life history factors, such as first-year or adult mortality experienced during migration or on the wintering grounds. At nearby Marine Corps Base Camp Pendleton, the resident population of flycatchers also declined for 10 years, from 26 individuals in 2007 to 0 individuals in 2017 (S. Howell, U.S. Geological Survey, unpub. data, 2017). In 2018, a social attraction study using speakers to broadcast flycatcher vocalizations commenced at Camp Pendleton. In 2018–21, there were between one and three resident breeders documented; however, there were no breeders documented in 2022 and one single female in 2023 (S. Howell, U.S. Geological Survey, unpub. data, 2018, 2019, 2020, 2021, 2022, 2023). Collection of more information is warranted for this declining species.