Population Size and Distribution

Resident flycatchers were detected at 6 of the 33 locations (18 percent) surveyed between 2015 and 2019 (table 1). Most of the resident flycatchers occurred on the San Luis Rey River; resident flycatchers were found on only one other drainage in San Diego County, at San Dieguito River. Based on the initial protocol survey at each location, we detected a minimum of 80 resident flycatchers from 2015 to 2019 (table 1). Of the resident flycatchers, 99 percent (79/80) were found on the San Luis Rey River; most of these were on the upper San Luis Rey River upstream and downstream from Lake Henshaw. Resident flycatchers were documented for the first time at Lake Henshaw, the only new location surveyed that supported flycatchers.

Transient flycatchers were observed at 14 of the 33 (42 percent) locations surveyed; 30 transient flycatchers were detected at 10 locations along the San Luis Rey River, and 8 along 4 other drainages/locations (table 1). Most of the transients were detected between May 15 and June 12.

Comparison to Historical Occupancy

Over the course of this study, 11 locations historically occupied by resident flycatchers were resurveyed; only 5 (45 percent) were found to support resident flycatchers: (1) Bonsall, (2) Cleveland National Forest, (3) Rey River Ranch, (4) San Dieguito, and (5) Vista Irrigation District (table 1; fig. 2). No resident flycatchers were found at El Capitan, Couser Canyon, Middle San Luis Rey, Pala, Santa Ysabel, or Sweetwater Authority (table 1). At Bonsall, the number of territories declined from a high of four (2014; B. Kus, U.S. Geological Survey, unpub. data, 2014) to two (table 1). The number of territories at San Dieguito declined from the previous confirmed high of three (1997; Kus and Beck, 1998) to one (table 1). The distribution at San Dieguito had also contracted; in 1997, there were two separate areas along the river with resident flycatchers, but only one was occupied during this study.

The remaining three historically occupied locations that were resurveyed were (1) Cleveland National Forest, (2) Rey River Ranch, and (3) Vista Irrigation District. These areas were previously surveyed for Willow Flycatchers in 1999 (table 2; Kus and others, 1999; Varanus Biological Services, 2000, 2001). Overall, the number of territories on the upper San Luis Rey River in the historically occupied area declined 71 percent between 1999 and 2019; the population dropped by 42 percent between 1999 and 2016, and by 50 percent during the 3 years after that. From 2016 to 2019, the largest declines in the number of territories occurred at Vista Irrigation District between 2016 (8 territories detected) and 2017 (3 territories detected), at Rey River Ranch between 2017 (9 territories) and 2018 (5 territories), and at Cleveland National Forest between 2018 (10 territories) and 2019 (6 territories).

The distribution of flycatcher territories at the upper San Luis Rey River in 2016 changed relative to the distribution in 1999 (table 2); the proportion of all territories that were at the Cleveland National Forest and Rey River Ranch decreased to 36 percent each, whereas the proportion at Vista Irrigation District increased to 29 percent, creating a more equal distribution of territories across the historically occupied area. By 2019, the distribution changed relative to 2016, with most of the territories spread equally between Cleveland National Forest and Rey River Ranch (43 percent each), while the proportion of territories at Vista Irrigation District declined to 14 percent.

Banded Birds and Dispersal Observations

We encountered 15 previously banded birds at 4 locations while performing surveys across San Diego County from 2015 to 2019. Of the 15 birds, 3 were known birds that had been previously observed at Bonsall in 2014 (1 male, 1 female; B. Kus, U.S. Geological Survey, unpub. data, 2014) and Cleveland National Forest in 2010 (1 male; B. Kus, U.S. Geological Survey, unpub. data, 2010). Two of the birds were females originally banded as nestlings at MCBCP and were observed for the first time as breeding adults; one female was last seen in 2013 (B. Kus, U.S. Geological Survey, unpub. data, 2013) and dispersed a distance of 55 kilometers (km) to breed at Cleveland National Forest (2015), while the second was last seen in 2010 (B. Kus, U.S. Geological Survey, unpub. data, 2010) and dispersed 41 km to San Dieguito (2017). We also documented an adult male who was originally observed at the beginning of the 2017 breeding season at Bonsall before moving within the same season to breed at San Dieguito (dispersal distance, 31 km; B. Kus, U.S. Geological Survey, unpub. data, 2017). At Lake Henshaw, we encountered nine banded birds including seven birds banded as nestlings and two adults; these birds are discussed further in chapter B.

Overview of Nest Monitoring Area

Based on the results of positive surveys in 2015 (detailed in chapter A), we designated a reach of the upper San Luis Rey River encompassing three locations (fig. 2; Cleveland National Forest, Rey River Ranch, and Vista Irrigation District) as a nest monitoring study area (hereafter, “monitoring area”). The monitoring area consisted of an approximately 6.0-km (3.7-mile [mi]) reach of the San Luis Rey River, including Vista Irrigation District property downstream from Lake Henshaw, Cleveland National Forest land downstream from the Vista Irrigation District property, and private property downstream from the Forest Service property. The entire length of the monitoring area was previously monitored for Southwestern Willow Flycatchers in 1999 (Kus and others, 1999; Varanus Biological Services, 2000, 2001).

The upper San Luis Rey River monitoring area was monitored annually from 2016 to 2019. Standardized territory names and boundaries were created by plotting all GPS points collected for resident flycatchers (including all nest locations within an individual year) and outlining the maximum extent of each concentration of points.

Breeding Productivity—Data Collection

All resident flycatchers detected within the monitoring area were visited at least weekly from mid-May until mid- to late August to determine their breeding status (paired or unpaired) and if paired, to determine pairing type (monogamous or polygynous). Monogamous pairings consisted of one male paired with one female, while polygynous pairings consisted of one male paired with more than one female. Behaviors used to establish polygyny included males interacting with and attending nests of more than one female simultaneously or sequentially. Pairs were visited weekly from their date of detection until the completion of their final nesting attempt to obtain a complete record of breeding activity within the territory. Pairs were observed for evidence of nesting, and their nests were located and monitored following standard protocol (Rourke and others, 1999). To minimize the chances of leading predators or Brown-headed Cowbirds to nest sites, nests were visited only as frequently as needed to collect sufficient data, targeting transition dates (for example, hatch date). Typically, there were three to four visits to a nest, corresponding to approximately one visit per week. The first visits were timed to determine the number of eggs laid and to remove cowbird eggs from parasitized nests; subsequent visits were to determine whether eggs had hatched and age of young, and the last to band the nestlings.

Nests were assigned six possible fates based on the following parameters. Nests that fledged at least one young were considered successful (SUC). Fledging was confirmed by audial or visual detection of young outside the nest, or on rare occasions, by the observation of an empty nest on or after the estimated fledge date in combination with parents carrying food within the territory. Unsuccessful nests were placed into one of four nest fate categories. Nests found empty or destroyed before the estimated fledge date and where the adult flycatchers were not found tending fledgling(s) were considered depredated (PRE). Previously active nests that were subsequently abandoned by adult flycatchers after one or more Brown-headed Cowbird eggs were laid in the nest were considered to have failed because of nest parasitism (PAR). Any nests that fledged cowbird young without fledging flycatcher young also were considered to have failed because of nest parasitism (PAR). Nests failing for reasons such as poor nest construction or the collapse of a host plant that caused the nest contents to be dumped onto the ground, or the presence of a clutch of infertile eggs (eggs that were dark with a visible air pocket, or that did not hatch after 20 days of incubation), were classified as failing because of other causes that were known (OTH). Nests that appeared intact and undisturbed but were abandoned with flycatcher eggs before the earliest hatch date, or nests that were completed but failed before flycatcher eggs could be confirmed, were classified as having failed because of unknown causes (UNK). Finally, nests that were seen during the construction phase but never completed or showed evidence of being dismantled were classified as incomplete (INC).

Whenever possible, we followed our standard protocol for manipulating nest contents when cowbird eggs or nestlings were detected in flycatcher nests. For nests with two or more flycatcher eggs, cowbird eggs were removed from nests as they were found. For nests with fewer than two flycatcher eggs, we waited until the next visit to remove the cowbird egg to minimize the possibility of nest abandonment in response to the removal. Cowbird nestlings were removed immediately from nests. Performed in this way, nest manipulation allows many parasitized nests to remain active and potentially fledge young where they would otherwise fail to fledge flycatcher young.

Breeding Productivity—Data Analysis

We used information collected during territory and nest visits to determine the number of nesting attempts per pair, nest initiation dates (date first egg laid), egg clutch size, hatching success for both eggs (percent of eggs that hatched) and nests (percent of nests with eggs in which at least one hatched), fledging success for both hatchlings (percent of hatchlings that fledged) and nests (percent of nests with hatchlings from which at least one fledged), apparent nest success (the proportion of completed nests that fledged young), productivity (the number of young fledged per pair), and fledge dates (the date on which at least one fledgling was detected outside the nest). We used the presence of cowbird eggs in flycatcher nests that contained at least one flycatcher egg to calculate parasitism rate (percent of nests parasitized), based on nests in which the contents were observed.

We used Pearson’s chi-square analysis and Fisher’s Exact tests to determine if there were differences across years in the proportion of pairs re-nesting after failed first nests, egg hatch rate, hatchling fledge rate, parasitism rate, apparent nest success, and productivity. 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 Analysis of Variance (ANOVA) and Tukey’s post-hoc pair-wise comparisons to determine if there were between-year differences in the number of completed nests per pair, average clutch size, and average number of young fledged per pair. Results were considered significant if P≤0.10.

We modeled daily survival rate (DSR) of flycatcher nests, or the probability that a nest would survive from one day to the next (Dinsmore and others, 2002) in Program MARK (White and Burnham, 1999) using the ‘RMark’ package (Laake, 2013) in R (R Core Team, 2020). Nest survival was calculated across a 35-day cycle length beginning with nest completion and encompassing egg-laying, incubation, and nestling periods. Age of nests at the time they were discovered was calculated by forward- or backward-dating of nests in relation to known dates of nest-building, egg-laying, or hatching. We used an information-theoretic approach (Akaike’s Information Criterion for small sample size, or AIC_c; Burnham and Anderson, 2002) to evaluate support for nest survival models reflecting a priori hypotheses regarding the effects of year and precipitation on survival. Monthly precipitation data were gathered from the Vista Irrigation District Lake Henshaw Dam station (R. Larsen, Vista Irrigation District, written commun., 2020). Because most of the precipitation in our study area accumulates during the fall and winter before the breeding season, we focused our analyses on ‘bio-year’ precipitation, which we defined as starting on October 1 of the calendar year before breeding and ending on March 31 of the breeding season year. Although this definition of bio-year begins before the flycatcher breeding season, fluctuations in food resources that may affect productivity and survival are thought to be driven by precipitation, and therefore, the defined bio-year is biologically relevant to the period when flycatchers are present. We further sub-grouped bio-year precipitation into two periods (fig. 3): October to December (early winter), and January to March (late winter), to examine whether timing of rainfall, in addition to amount of rainfall, influenced DSR. We created a set of three separate models using bio-year precipitation (hereafter, “precipitation”; PrecipBY), early winter precipitation (PrecipEW), and late winter precipitation (PrecipLW), as well as an additive model with early and late winter precipitation together to estimate coefficients for one variable while controlling for the other. 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 effects of year and precipitation on DSR. We then modeled the covariates and evaluated support for the models in relation to the constant survival model. Criteria for well-supported models included: ∆AIC_c less than 2, and AIC_c weight greater than 0.05. We evaluated the significance of covariates within our top models by examining whether the 95-percent confidence intervals of the beta estimates included zero; we proceeded with inference when the 95-percent confidence interval of an estimate did not cross zero. If more than one model in the model set had ΔAIC_c value less than 2, we averaged the parameters in the model set using the ‘RMark’ package (Laake, 2013) in R (R Core Team, 2020).

Survivorship, Fidelity, and Movement

To facilitate analyses of survival and dispersal, we attempted to capture and color band any unbanded adult flycatchers in our monitoring area. We used mist nets and song playback to capture adults. When captured, birds were banded with colored leg bands to create a unique color-band combination so that individuals could be identified in the future without recapture. Flycatcher nestlings from accessible nests were banded at 7–9 days of age with a single blue anodized federal numbered band on the left or right leg alternating by year (that is, 2016 and 2018 nestlings were banded on the left leg and 2017 nestlings were banded on the right leg). Birds banded as nestlings that returned as adults (“natals”) were captured and given a unique combination of color bands to supplement their single federal band. Adult flycatchers were sexed based on breeding characteristics observed during the banding process (the presence of a brood patch for females or a cloacal protuberance for males) or by behavioral characteristics observed during subsequent encounters (that is, only females incubate nests; males broadcasting song or other territory defense behaviors). During surveys and nest monitoring activities, we attempted to resight all flycatchers to determine whether they were banded, and if so, to confirm their identity by reading their unique color band combination.

Habitat and Nest Site Characteristics—Data Collection and Analysis

Dominant native and exotic plants were recorded at each monitored flycatcher territory, and percent 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:

In 2018, we began noticing dead and dying oaks in the monitoring area. To aid in evaluating habitat changes in flycatcher territories related to dead and dying oaks, we examined visible/near infrared aerial imagery from the National Agricultural Inventory Program (NAIP; U.S. Geological Survey, 2021). The NAIP image sets were captured between April 23, 2016, and April 24, 2016, and between April 15, 2020, and May 25, 2020, and consisted of 1-m ground sample distance resolution, and four-band visible and near-infrared georeferenced orthoimages. A normalized difference vegetation index (NDVI) was computed from the NAIP image using Image Analysis in ArcMAP 10.4.1 (Environmental Systems Research Institute, 2021). Zonal statistics were calculated from the NDVI values for each buffered territory boundary. Percent change in mean NDVI from 2016 to 2020 was calculated for each territory in the study (((mean NDVI 2020-mean NDVI 2016)/mean NDVI 2016)*100).

We examined the vegetation characteristics of flycatcher nest locations. Following the abandonment or fledging of flycatcher nests, we recorded nest height, host plant species, host height, distance to edge of host plant, and distance to edge of host clump. Clump boundaries were defined where leaves or branches of neighboring plants no longer overlapped. We tested for differences in flycatcher nest site characteristics across years and between successful and unsuccessful nests within years using Mann-Whitney U-tests (two groups), and Kruskal-Wallis tests (three or more groups) because the data violated the normality assumptions of t-tests and ANOVA tests. Mann-Whitney U-tests and Kruskal-Wallis tests were performed using base tools in R (R Core Team, 2020).

Composition of Monitored Population

From 2016 to 2019, we monitored 82 territories (14–27 per year; table 3; appendix 1, figs. 1.1–1.4) at the upper San Luis Rey River monitoring area. Most of the territories supported pairs, although single or undetermined breeding status birds were present every year (six in 2016, three in 2017, one in 2018, and three in 2019). Additionally, in 2018 we detected a non-territorial floater that was present late in the breeding season (table 1; appendix 1, fig. 1.4, V10). Among the paired birds, we observed polygynous pairings in all years, with the lowest rate of polygyny (number of polygynous pairs/total number of pairs) observed in 2016 (10 percent) and the highest in 2017 (70 percent; table 4). The proportion of paired males that were polygynous ranged from 5 to 54 percent between 2016 and 2019 and varied with the ratio of females to males in the adult population. The total number of males in the monitoring area declined 41 percent from 2016 to 2017, which increased the sex ratio of females to males and encouraged polygynous pairings (table 4).

Breeding Productivity

We monitored the nesting activity of 11–21 pairs annually, all of which were fully monitored, meaning that the territory was visited at least weekly and all nests within the territory were found and documented during the breeding season (table 3). A total of 126 nests were monitored between 2016 and 2019 (18–41 nests per year; table 3). Eight of these nests were not completed and were subsequently excluded from calculations of nest success, productivity, and daily nest survival.

Nest Initiation

The Southwestern Willow Flycatcher breeding season began in May or June and ended in July or August. Most of the first nesting attempts were initiated during late May and early June (fig. 4). Nest initiation (date first egg laid) occurred earlier in 2016 and 2017 (median Julian day 157; June 5 and June 6) than in 2018 (median Julian day 159; June 8) and 2019 (median Julian day 160; June 9; fig. 5). The earliest confirmed nest initiation (first egg of the season) was May 24 (Julian day 144; 2017) and the latest initiation was July 11 (Julian day 192; 2017).

The earliest and latest fledge dates occurred in 2016 (June 28; Julian day 180 and August 18; Julian day 231, respectively; fig. 6). The median fledge date occurred earlier in 2016 (July 13; Julian day 195) and 2017 (July 16; Julian day 197) than in 2018 (median Julian day 207; July 26) and 2019 (July 19; Julian day 200).

Nesting Attempts

Monitored flycatcher pairs completed between zero and four nests per year; one pair in 2017 did not nest. The average number of completed nests per pair ranged from 1.5±0.9 to 2.1±0.8 across years (table 4), but did not differ significantly by year (F_3,65=1.53, P=0.22).

In most years, most of the flycatcher pairs (47–76 percent) attempted more than one nest, all following an unsuccessful first attempt (fig. 7). A few pairs attempted third and fourth nests after failures. Significantly more flycatcher pairs were successful on their first nest attempt in 2017 (37 percent; 7/19; Fisher’s Exact test, P=0.09), compared to 2016 (29 percent; 6/21), 2018 (6 percent; 1/17), and 2019 (9 percent; 1/11). None of the successful pairs in any year re-nested after a successful nest.

Productivity

We documented 262 Southwestern Willow Flycatcher eggs in 96 monitored nests from 2016 to 2019 (table 7). Clutch size estimated from full clutches in non-parasitized nests ranged from two to four and averaged 2.8±0.8 to 3.1±0.8 annually (fig. 8). Clutch size did not vary significantly between years (F_3,55=0.22, P=0.88).

Table 7.    

Reproductive success and productivity of nesting Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

[Calculations of average clutch size include only non-parasitized nests that had a known full clutch. Abbreviation: ±, plus or minus]

Table 7.    Reproductive success and productivity of nesting Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

Parameter	2016	2017	2018	2019
Nests with eggs	31	24	30	11
Eggs laid	90	66	80	26
Average clutch size	3.1±0.7	3.1± 0.7	3.1±0.8	2.8±0.8
Hatchlings	49	43	30	5
Nests with hatchlings	18	16	11	2
Hatching success
Percentage of eggs that hatched	54	65	38	19
Percentage of nests with eggs in which at least one hatched	58	67	37	18
Fledglings	34	26	12	5
Nests with fledglings	13	11	5	2
Fledging success
Percentage of hatchlings that fledged	69	60	40	100
Percentage of nests with hatchlings from which at least one fledged	72	69	45	100
Average number of young fledged per pair	1.6±1.5	1.3±1.3	0.7±1.2	0.5±1.0
Pairs fledging at least one young (percentage)	13 (62)	11 (55)	5 (29)	2 (18)

Hatching success of eggs differed significantly between years (P<0.001, Fisher’s Exact test; table 7) and was higher in 2017 (65 percent) than in 2016 (54 percent), 2018 (38 percent), and 2019 (19 percent). Significant differences occurred in all paired year comparisons except 2016 and 2017 (2016 versus 2017: χ2=1.39, P=0.24, df=1; 2016 versus 2018: χ2=4.23, P=0.04, df=1; 2016 versus 2019: P<0.001, Fisher’s Exact test; 2017 versus 2018: χ2=9.98, P<0.001, df=1; 2017 versus 2019: P<0.001, Fisher’s Exact test; 2018 versus 2019: P=0.10, Fisher’s Exact test).

Similarities within years between the percent of eggs that hatched and the percent of nests with eggs that hatched indicate that the factors affecting hatch rate were operating at the nest level, instead of the egg level (for example, infertile eggs, partial predation, and so forth).

Fledging success differed significantly between years (P=0.02, Fisher’s Exact test; table 7) and was significantly lower in 2018 than in 2016 (χ2=5.46, P=0.02, df=1) and 2019 (P=0.02, Fisher’s Exact test). As with hatching success, factors influencing fledging success appeared to be operating at the nest level.

The number of fledglings per pair ranged from zero to four and averaged 0.5±1.0 to 1.6±1.5 annually (table 7). There was a significant difference in the number of young fledged per pair between years, with pairs in 2016 producing more than three times the number of fledglings as in 2019 (F_3,65=2.76, P=0.05; table 7; fig. 9). The percent of pairs fledging at least one young appeared to decline over time but did not differ significantly by year (P=0.37, Fisher’s Exact test; table 7).

Several of the nests that successfully fledged young had been parasitized by cowbirds (2 nests in 2016, 1 nest in 2017, and 1 nest in 2019 [table 6]) and were subsequently “rescued” by removing the cowbird eggs. These rescued nests were responsible for the production of 2–3 additional young fledged per year, increasing annual productivity in 2016 by 7 percent (from 1.5±1.5 to 1.6±1.5 [table 7] young fledged per pair), in 2017 by 8 percent (from 1.2±1.3 to 1.3±1.3 [table 7]), in 2018 by 0 percent, and in 2019 by 150 percent (from 0.2±0.6 to 0.5±1.0 [table 7]) over what would be expected had all parasitized nests been allowed to fail.

Daily Nest Survival

Analysis of DSR showed that from 2016 to 2019, the best supported model for predicting flycatcher nest survival was the model that included both early winter precipitation and late winter precipitation (table 8). The year model was also within two AIC_c units of the top model and carried an AIC_c weight of 0.24. Early winter precipitation and late winter precipitation acted in opposition to one another, with increases in early winter precipitation promoting DSR while increases in late winter precipitation reduced it (table 9). This pattern was largely driven by changes in DSR in 2018 and 2019, which were the only years considered significant predictors to the year model (table 9). The DSR declined between 2017 and 2018 coinciding with a steep decline in early winter precipitation (99 percent) and a lesser decline in late winter precipitation (56 percent; fig. 10). The DSR failed to increase between 2018 and 2019 when early winter precipitation increased a bit but late winter precipitation more than doubled (fig. 10), which was probably the source of the negative relationship between late winter precipitation and DSR. Overall, DSR varied little across years (table 10).

Table 8.    

Logistic regression models for the effects of precipitation and year on daily nest survival (S) of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

[Models are ranked from best to worst based on Akaike’s Information Criterion for small samples (AICc), ΔAICc, and Akaike weights (w). AICc is based on −2 x log_e likelihood and the number of parameters (K) in the model. Abbreviations: +, plus; PrecipBY, pre-breeding season precipitation from October of the prior calendar year to March of the breeding year; PrecipEW, early winter precipitation from October to December of the calendar year prior to the breeding season; PrecipLW, late winter precipitation from January to March of the breeding year]

Table 8.    Logistic regression models for the effects of precipitation and year on daily nest survival (S) of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

Model	Deviance	K	AICc	ΔAICc	w
PrecipEW + PrecipLW	413.8	3	419.8	0.00	0.59
Year	413.6	4	421.6	1.83	0.24
PrecipEW	419.1	2	423.1	3.34	0.11
Constant	423.8	1	425.8	5.99	0.03
PrecipBY	422.7	2	426.8	6.97	0.02
PrecipLW	423.7	2	427.7	7.94	0.01

Table 9.    

Parameter estimate (β), standard error (SE), and 95-percent confidence intervals (95% CI) for models with ΔAIC<2 explaining daily nest survival of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

[+, plus; PrecipEW, early winter precipitation from October to December of the calendar year prior to the breeding season; PrecipLW, late winter precipitation from January to March of the breeding year]

Table 9.    Parameter estimate (β), standard error (SE), and 95-percent confidence intervals (95% CI) for models with ΔAIC<2 explaining daily nest survival of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

Effect	β	SE	95% CI
PrecipEW+PrecipLW
Intercept	3.21	0.31	2.60	–	3.83
PrecipEW1	0.15	0.05	0.06	–	0.25
PrecipLW1	−0.07	0.03	−0.12	–	−0.01
Year
Intercept2	3.29	0.22	2.86	–	3.71
Year2017	0.01	0.32	−0.61	–	0.64
Year20181	−0.70	0.29	−1.27	–	−0.13
Year20191	−0.74	0.35	−1.42	–	−0.06

¹

Indicates a significant effect and denotes a 95-percent CI that does not cross zero.

²

Intercept here represents Year2016.

Table 10.    

Model averaged real parameter estimate, standard error (SE), and 95-percent confidence intervals (95% CI) for daily nest survival of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

Table 10.    Model averaged real parameter estimate, standard error (SE), and 95-percent confidence intervals (95% CI) for daily nest survival of Southwestern Willow Flycatchers at the upper San Luis Rey River monitoring area, San Diego County, California, 2016–19.

Year	Estimate	SE	95% CI
2016	0.955	0.008	0.936	–	0.968
2017	0.955	0.008	0.936	–	0.968
2018	0.947	0.012	0.918	–	0.965
2019	0.946	0.014	0.910	–	0.968

Survivorship, Fidelity, and Movement

Habitat and Nest Site Characteristics

Habitat Characteristics

Flycatcher territories in the monitoring area between 2015 and 2019 occurred in five habitat types (table 24). Of the 5 habitat types, 47 percent (20/43) of territories were in willow-oak, 19 percent (8/43) in willow-ash, 16 percent (7/43) in oak-sycamore, 12 percent (5/43) in mixed willow riparian, and 7 percent (3/43) in willow-sycamore. All flycatcher territories occupied during the study period contained greater than 95-percent native plant cover. The most commonly recorded dominant species at flycatcher territories included red or arroyo willow, coast live oak, California sycamore, and velvet ash.

Table 24.    

Habitat characteristics of Southwestern Willow Flycatcher territories at the upper San Luis Rey River monitoring area, San Diego County, California, 2015–19.

[Territory: See appendix 1, figs. 1.1–1.4 for territory locations. Habitat type: Mixed willow riparian: Habitat dominated by one or more willow species, including Goodding’s black willow, and red or arroyo willow (SAL), with mule fat as frequent co-dominant. Oak-sycamore: Woodlands in which coast live oak (QUE) and California sycamore (PLT) exist a co-dominants, Willow-ash: Willow riparian habitat in which velvet ash (FRX) is a co-dominant. Willow-oak: Willow riparian habitat in which coast live oak is a co-dominant. Willow-sycamore: Willow riparian habitat in which California sycamore is a co-dominant. Dominant species: ALN, White Alder (Alnus rhombifolia). Years occupied: X, occupied, —, not occupied. NDVI change: The positive or negative percent change in values from 2016 to 2020. Other abbreviations: NDVI, normalized difference vegetation index; −, negative]

Table 24.    Habitat characteristics of Southwestern Willow Flycatcher territories at the upper San Luis Rey River monitoring area, San Diego County, California, 2015–19.

Territory	Habitat type	Dominant species	Years occupied					NDVI change
			2015	2016	2017	2018	2019
C1	Mixed willow	SAL	X	—	X	—	—	−0.08
C2	Willow-oak	QUE, SAL	X	X	—	—	—	−0.03
C3	Willow-oak	QUE, SAL	—	—	—	X	X	0.01
C4	Willow-oak	QUE, SAL, FRX	—	—	X	X	—	−0.07
C5	Willow-oak	QUE, SAL, FRX	X	—	X	X	—	−0.10
C6	Willow-ash	FRX, SAL, QUE	X	X	X	—	X	−0.04
C7	Willow-ash	FRX, SAL, QUE	X	X	X	X	X	−0.08
C8	Willow-ash	FRX, SAL	X	—	—	—	X	0.04
C9	Willow-ash	FRX, SAL, QUE	X	X	X	X	X	−0.05
C10	Willow-oak	QUE, SAL, ALN	—	X	X	—	—	−0.07
C11	Willow-ash	FRX, SAL, QUE	—	X	X	X	—	−0.02
C12	Willow-ash	FRX, SAL, ALN	X	X	X	X	X	−0.03
C13	Willow-oak	QUE, SAL	—	X	X	X	—	−0.08
C14	Willow-oak	QUE, SAL	—	—	—	X	—	0.01
C20	Willow-oak	QUE, SAL	X	—	—	—	—	0.00
C22	Oak-sycamore	QUE, PLT	X	X	X	X	—	0.03
C23	Mixed willow	SAL	—	X	—	—	—	0.04
C24	Willow-oak	QUE, SAL	X	—	—	—	—	0.02
R1	Willow-oak	QUE, SAL	—	X	—	—	—	0.07
R2	Willow-oak	QUE, SAL	—	X	—	X	X	0.06
R3	Willow-sycamore	PLT, SAL	X	—	—	—	—	0.04
R4	Willow-sycamore	PLT, SAL	—	—	X	—	—	0.02
R6	Oak-sycamore	QUE, PLT	X	X	—	—	—	0.08
R9	Oak-sycamore	QUE, PLT	X	X	X	—	—	0.05
R10	Oak-sycamore	QUE, PLT	X	—	X	—	X	0.06
R11	Willow-oak	QUE, PLT, SAL	—	X	X	X	X	0.08
R12	Oak-sycamore	QUE, PLT	X		X	X	X	0.05
R13	Willow-oak	QUE, PLT, SAL	X	X	X	X	X	0.05
R14	Willow-oak	QUE, SAL	X	X	X	X	X	0.06
R15	Willow-oak	QUE, SAL	—	—	X	—	—	0.05
R16	Oak-sycamore	QUE, PLT	X	X	X	—	—	0.04
R17	Oak-sycamore	QUE, PLT	X	X	—	—	—	0.08
V1	Willow-sycamore	SAL	—	X	X	X	—	−0.04
V2	Mixed willow	SAL	X	X	—	X	X	−0.02
V3	Willow-oak	SAL	—	X	—	—	—	−0.06
V4	Willow-oak	QUE, SAL	X	X	—	—	—	−0.11
V5	Mixed willow	SAL	X	—	—	—	—	−0.05
V7	Willow-ash	FRX, SAL	X	X	X	X	X	−0.05
V8	Mixed willow	SAL	—	—	X	—	—	−0.03
V9	Willow-oak	QUE, SAL	—	X	—	—	—	−0.02
V10	Willow-oak	QUE, SAL, ALN	X	X	—	X	—	0.03
V11	Willow-ash	FRX, SAL, ALN	X	—	—	—	—	−0.03
V12	Willow-oak	SAL	X	X	—	—	—	−0.01

In 2018, we began noticing dead and dying oaks in the monitoring area. We investigated the change in mean NDVI values between 2016 and 2020 in all territories occupied by resident flycatchers between 2016 and 2019 and found that NDVI values decreased in 49 percent of flycatcher territories (21/43; table 24). The decrease in NDVI values occurred along a gradient of severity from upstream to downstream, with the territories furthest downstream (C20–24; R1–17; appendix 1, figs. 1.1–1.2) showing no negative change and the territories upstream (C1–14; V1–12; appendix 1, figs. 1.2–1.4) experiencing the largest decrease in NDVI values. The change is visually apparent with the spectrum of NDVI values overlaid on aerial imagery in one example territory (C13; fig. 11A, 11B; appendix 1, fig. 1.3) that showed a decline in overall NDVI values: live vegetation appears green on the image, and dead vegetation appears red (the San Luis Rey River also appears as “dead” vegetation within the image). This particular territory was occupied for 3 years of the study (2016–18) but was vacant in 2019. We did not see a clear relationship between territory occupancy and a change in NDVI values, as 43 percent (6/14) of territories occupied in 2019 were territories that showed a decrease in NDVI values; however, there may be a lag time associated with NDVI change and territory abandonment.

Composition

Annually, most of the territories at the upper San Luis Rey monitoring area were occupied by pairs, with a variable number of single males present each year. Among the territories containing pairs, we documented monogamous and polygynous pairings. While previous surveys and limited nest monitoring of this population allowed for speculation about the degree of polygyny (Kus and others, 1999; B. Kus, U.S. Geological Survey, unpub. data, 2010), our banding studies allowed us to characterize this with more certainty. The degree of polygyny we documented at the upper San Luis Rey River monitoring area fluctuated annually but fell within the range reported in other studies (10–50 percent of paired males; Sedgwick, 2000; Davidson and Allison, 2003; Kus and others, 2017). Polygyny in Willow Flycatchers is thought to be facultative and has been linked to differences in adult sex ratios (Davidson and Allison, 2003; Kus and others, 2017). Kus and others (2017) documented rates of male polygyny up to 100 percent in a declining population at MCBCP, where the rate increased as the ratio of adult females to males increased. We observed a similar relationship at the upper San Luis Rey River monitoring area when the highest rate of male polygyny (54 percent) occurred following a decline in the number of males between 2016 and 2017, which created an adult sex ratio that favored females.

Breeding Productivity

The measures of breeding productivity we documented at the upper San Luis Rey River monitoring area were highly variable, and within the timeframe of this study, showed somewhat biennial rather than annual variation. Flycatchers experienced greater success in the first two years of the study, with multiple measures of productivity including hatching success, fledging success, nest success and seasonal productivity significantly higher in 2016 and 2017 than in 2018 and 2019. However, higher productivity in 2016 and 2017 did not result in a population increase in 2018 and 2019; instead, the population continued to decline. Although some of this decline could be attributed to emigration to Lake Henshaw, it appears to be largely a response to declining productivity over time. The highest productivity we documented was 1.6 young per pair in 2016 (comparable to 1.5 reported by Kus and others [1999]), and by the end of the study, productivity had dropped to 0.5 young per pair. Apparent nest success was highly variable, ranging from a high of 37 percent in 2016 and 2017 to a low of 11 percent in 2019, with the high comparable to that collected by Kus and others (1999) of 38 percent. The decline in nest success in the latter 2 years of the study appeared to coincide with higher rates of predation; 66 percent of all nests were depredated in 2018, compared to the low of 49 percent in 2016. Cowbird parasitism also increased in the latter years and averaged 25 percent of all nests in 2018–19 compared to 7 percent in 2016–17. Kus and others (1999) observed no cowbird parasitism in 1999, although cowbird trapping was performed in that year, and likely decreased the cowbird population (J.M. Wells, U.S. Forest Service, written commun., 1999). In contrast, over an 18-year period, the flycatcher population at MCBCP averaged higher average nest success (54±21 percent), higher overall seasonal productivity (1.9±0.7 young per pair), lower predation (32±17 percent of all nests depredated), and near zero cowbird parasitism.

Our analysis of daily nest survival produced a pattern across years mirroring that of apparent nest success, with DSR declining between the first two and the last two years of the study. The analysis identified precipitation as possibly influencing nest survival, although the contribution of precipitation relative to other factors discussed above and how those vary across years is unclear. In particular, it is unclear whether our results indicate a causal relationship between precipitation and DSR, or whether they are a statistical artifact driven by extreme differences between 2018 and 2019.

Although we were able to “rescue” most parasitized nests by removing cowbird eggs and thereby increasing annual nest success by up to 45 percent, parasitism reduced productivity through its effect on flycatcher clutch size, which sets the limit on the potential number of young that can be produced. Clutch size reduction is a typical feature of parasitism and occurs when cowbird females remove a host egg and replace it with one of their own (Rothstein and others, 2003). This reduces flycatcher clutch size, which averaged approximately three eggs annually, by about 30 percent, a substantial limit on reproductive potential. Moreover, parasitism may be contributing to higher predation rates observed in this population; Stumpf and others (2011) found predation rates were higher in parasitized nests compared to unparasitized nests. Parasitized nests may be at greater risk from predation because of increased activity at nest sites from adult flycatchers attempting to thwart parasitism attempts and increased noise from cowbird nestlings that may increase auditory cues for predators. Moreover, cowbirds themselves may depredate flycatcher nestlings. While the overall rate of parasitism during our study was low (16 percent), higher rates in 2018 (23 percent) and 2019 (27 percent) are cause for concern. The threshold set by the Southwestern Willow Flycatcher recovery plan (U.S. Fish and Wildlife Service, 2002) to trigger management intervention if parasitism rates exceed 20–30 percent for more than 2 years was exceeded in the latter years of this study.

One other possible explanation for the greater success observed in the first two years of the study is the timing of nest initiation: in 2016 the median initiation date was June 5, compared to June 9 in 2019. It is possible that there is greater food availability earlier in the season, less predation and parasitism pressure, or a combination of other unknown variables contributing to greater success. Earlier nest initiation also translates into earlier fledge dates, which can influence subsequent survival of young. Paxton and others (2007) demonstrated that at their study site in Arizona, earlier fledge dates were associated with higher first-year survivorship. For each day later into the season a nestling fledged, its odds of survival decreased by 2 percent. Earlier fledge dates in 2017 may have translated to higher first-year survival in our study, as most of the first-year flycatchers we encountered originated in 2017.

Survivorship and Movement

Habitat and Nest Characteristics

As previous studies have shown, coast live oak is one of the primary components of the habitat at the upper San Luis Rey River monitoring area (Kus and others, 1999; B. Kus, U.S. Geological Survey, unpub. data, 2010). The high proportion of oak in the upper San Luis Rey River monitoring area is unusual in that most of the occupied flycatcher habitats are primarily dominated by more traditional riparian vegetation such as willow, cottonwood, alder, and so forth. During our study, coast live oak was a co-dominant species in 63 percent of flycatcher territories. Coast live oak was also the most commonly used nest host, with 70 percent of nests in our study placed in this species. In 2018, we began to observe dead and dying oaks in our study area, which we believe to be the result of goldspotted oak borer (Agrilus auroguttatus) infestation. At the conclusion of this study, we investigated the overall change in NDVI in flycatcher territories within the monitoring area and found that the mean NDVI values decreased in 49 percent of monitored territories between 2016 and 2020. In the areas with dead oaks, the canopy appeared to be more open and there were more gaps in the habitat; this could create additional opportunities for predation and parasitism.

We found no differences between years in nest placement and no indication that flycatchers have adjusted their nest placement to avoid oak trees. Coast live oak remained the most commonly used nest host even in later years after we observed dead and dying oaks.