2.2 | Field Methods
2.2.1 | Longspur Settlement
We deployed 24 autonomous acoustic recorders (Wildlife Acoustics model SM4, Maynard, MA; hereafter “song meters”) to assess settlement patterns of territorial male longspurs on the breeding grounds. We consulted local biologists and used observations from the USFWS Breeding Bird Survey (BBS) and eBird (Sauer et al., 2020; Sullivan et al., 2020) to identify locations previously used by longspurs. We deployed 8 and 16 song meters in 2020 and 2021, respectively, with half (4 in 2020, 8 in 2021) in crop fields and half in native grassland sites. We selected sites that were no more than 25 km apart to minimize regional variation in weather patterns between site types. We deployed song meters on 7 April and retrieved them on 30 April after territory establishment (With, 2021). We affixed song meters to 1.8-m t-posts at a height of 1.2 m and covered each microphone with an extra layer of foam to reduce recorded wind noise.
We programmed song meters to collect a 3-minute recording every half hour starting 15 minutes before sunrise and ending by 09:00 hours to coincide with morning breeding choruses of longspurs (With, 2021), resulting in six 3-minute recordings collected each morning. Upon removal from the field, a trained technician manually reviewed each 3-minute recording and documented longspur presence. We excluded any recordings ≥ 25% obscured by wind or other noise.
2.2.2 | Longspur Abundance
Initial Occurrence Surveys. – To identify areas used by thick-billed longspurs, we randomly generated 100, 64-ha sampling plots in both crop and native sites. We only included survey plots that had rangeland productivity ≤1,100 kg ha-1 (Lipsey and Naugle, 2017) and that were located within land parcels where we secured permission to conduct fieldwork. Plots were separated by ≥200 meters to ensure independence. Random plots in crop fields that contained >1 crop type were discarded because different crop types grow at different rates, potentially confounding results. In native sites, we discarded plots if they contained badlands or water bodies covering ≥1/4 of the plot because longspurs do not nest in such areas (DuBois, 1937; Felske, 1971; With and Webb, 1993).
We conducted initial surveys within the 64-ha plots during 30 Apr–11 May, 2020–21. We surveyed ≥25 plots in each site type each year. Observers walked a U-shaped line transect within each plot, starting 200 meters inward from a randomly selected plot corner (Figure 3). We identified species and recorded perpendicular distance and direction from the transect line for each bird or group of birds seen or heard to maintain consistency with distance sampling methods. Estimated distances were recorded in bins: 0–25, 26–50, 51–75, 76–100, and 101–200 meters. We walked at a pace of 2–3 km hr-1 and completed each transect within 40 minutes. Observers were trained to avoid double-counting birds. Surveys began one half hour before sunrise and were completed by 10:00 hours. We did not conduct surveys if wind speed was >25 km hr-1 or it was raining. We recorded survey covariates including observer, cloud cover, temperature, wind speed, date, GPS starting point, and transect start/end times.
Abundance surveys. – Within initial plots occupied by longspurs, we generated smaller, 16-ha survey plots within which we monitored longspur populations for the remainder of the nesting season. We identified occupied crop fields, randomly selected ≥20 of these fields, and generated a single 16-ha plot within the center of each selected field. This ensured crop plots were ≥200 m from field edges, roads, and other plots. In native sites, we delineated large patches of occupied habitat by tracing the extent of occupied areas on foot with a GPS unit and later transferred this information to ArcMap 10.7.1 (ESRI, 2019). Patches were discovered after determination of longspur occupancy during initial surveys, and patch edges were defined by presence/absence of singing longspurs and were typically coupled with apparent changes in vegetation composition. We then overlaid a grid of 16-ha cells over occupied patches and used ArcMap to randomly select ≥20 cells from these patches. Only non-adjacent cells were used to ensure plots were ≥200 m apart. Each site selection process allowed us to select only occupied sites and guaranteed independence.
We conducted six rounds of line transect surveys within each 16-ha survey plot during 10 May–15 July 2020–21. Survey rounds were separated by ≥5 d. Observers walked a U-shaped line transect within each plot, starting 100 meters inward from a randomly selected plot corner (Figure 4). We collected data as described above but limited observations to thick-billed longspurs and distance bins included 0–25, 26–50, 51–75, and 76–100 m. We completed each transect within 30 minutes and surveys began 15 minutes before sunrise and ended at 9:00 hours.
2.2.3 | Nest Phenology, Survival, and Reproductive Output
Nest Searching. – We searched for nests during 9 May–22 July, 2020 and 5 May–8 July, 2021 to assess reproductive effort throughout the entire nesting season (With, 2021). Nest searching began at sunrise and ended at 11:00 hours on days without precipitation and observers were randomly assigned a group of plots to search each morning. Observers alternated between crop and native sites during subsequent days and used behavioral observations to find nests (Martin and Geupel, 1993; Winter et al., 2003). We observed longspurs from a distance of ≥30 m and moved to a new plot after 60 minutes if no female longspurs were detected. In addition, we supplemented behavioral nest searching with standard rope dragging methods (Klett et al., 1986; Koford, 1999).
Nest Monitoring. – Upon finding a nest, we recorded the geographical coordinates and marked the nest location with 15-cm bamboo stakes placed 2 m north and east of the nest to aid in relocation. Nests were checked every 2–4 days until fledging or failure (Martin and Geupel, 1993; Ralph, 1993). We recorded adult behavior, number of eggs and young, number of brown-headed cowbird (Molothrus ater ) eggs or nestlings, date, time, observer, time spent at nest, and any relevant notes. We aged nestlings according to developmental cues described in Jongsomjit et al. (2007) so the nest could be checked on predicted date of fledging. We considered a nest failed if eggs were gone before expected hatch date, if nestlings disappeared before nearing expected fledge date, or if dead nestlings or depredated eggs were found in or near the nest. A nest was considered successful if ≥1 chick fledged. We deemed nests successful if nearby adults were observed feeding fledglings, ≥1 fledgling was observed, territorial adults were present with food or directed aggressive behaviors toward observers, or fecal material was present and nestlings had reached the appropriate age to fledge (Ralph, 1993; Jones et al., 2010).
2.2.4 | Habitat Conditions
We collected vegetation measurements at two spatial scales, the nest site and the survey plot (16-ha). Measures were collected at each nest within 3 days of fledge or expected fledge for failed nests. In addition, we randomly selected 3 and 10 habitat sampling points within the 16 ha survey plots in crop fields and native prairie sites, respectively. Vegetation conditions in crop fields were fairly homogenous and required fewer sampling points (Swicegood, 2022). We measured vegetation conditions three times throughout the longspur breeding season, once in May, June, and July. At each sampling point, we recorded visual obstruction readings (VOR) in each cardinal direction from a distance of 4 m and a height of 1 m (Robel et al., 1970). We measured overlapping percent cover of grass, forb, shrub, litter, and bare ground within a 20 × 50 cm sampling frame at the sampling point and at 4 locations 0.5 m from the point in each cardinal direction (Daubenmire, 1959). Cover was recorded within six percentage classes (0%, 1–5%, 6–25%, 26–50%, 51–75%, 76–95%, and 96–100%). We listed all plant species, lichen, and spikemoss in order of decreasing abundance, found within a 2-m radius of the point center. We measured litter depth (mm) in the northwest corner of the Daubenmire frame and recorded the species, distance (m), and height (cm) of the nearest shrub within 25 m as shrubs influence nest density or detection of nests for many grassland passerines (Davis, 2005; Pulliam et al., 2021).

2.3 | Analyses

2.3.1 | Longspur Settlement
We used multi-season occupancy models to evaluate whether longspur settlement patterns differed between crop and native sites (MacKenzie et al., 2003). Multi-season occupancy models use detection/non-detection data collected with a robust design (Pollock, 1982) to estimate initial occupancy and subsequent rates of local colonization (e.g., settlement) and extinction (e.g., site abandonment) while accounting for spatially variable detection probability (MacKenzie et al., 2003; Mackenzie, 2006). The design used k secondary survey periods within Tprimary periods; each day represented a primary period and each 3-minute recording a closed secondary period (i.e., 6 secondary periods occurred over 24 days in each year).
We fitted multi-season occupancy models using the ‘colext’ function in R package ‘unmarked’ (Fiske and Chandler, 2011; Kéry and Chandler, 2016) and used information theory to evaluate support for competing models representing hypotheses about detection probability, initial occupancy, and settlement patterns (MacKenzie et al., 2003). We evaluated support for our a priori models in a phased approach. First, we evaluated how well a fully parameterized model fit the data and estimated a variance inflation factor (ĉ ) using the mb.gof.test in the R package ‘AICcmodavg’. Bootstrapping was based on 500 simulations to generate a chi-squared statistic and to calculate average ĉ , where a ĉ value >1 indicates overdispersion in the data, but much higher values (>4) may indicate lack-of-fit (Mazerolle, 2020). We found evidence of moderate overdispersion (ĉ = 1.9) and inflated estimated standard errors by\(\sqrt{}\)ĉ and based subsequent model evaluation and inference on the quasi-Akaike’s Information Criterion adjusted for finite samples (QAICc; Burnham and Anderson, 2002).
We developed models that evaluated the effects of survey conditions on detection probability. Variables hypothesized to influence detection included daily precipitation, minimum daily temperature, minutes past sunrise, and Julian day (Table 1). We evaluated a quadratic effect of minutes past sunrise because bird detections were previously found to be highest mid-morning (With, 2021). Because all detection covariates we measured are known to affect the detectability of songbirds, we used a backward selection approach based on QAICc to eliminate uninformative parameters (Pagano and Arnold, 2009; Arnold, 2010; Montgomery et al., 2021). Models with large relative weights (wi ) and QAICc values ≤2 from the best-fit model were considered equally parsimonious (Devries et al., 2008; Arnold, 2010; Burnham et al., 2011). After we identified a parsimonious sub-model for detection, it was retained in subsequent evaluations of occupancy and settlement.
Because some birds had already arrived at the study area prior to song meter deployment, we evaluated whether initial occupancy differed by habitat type (crop vs. native) before evaluating the effects of habitat type and Julian day on settlement probability (Table 1). In addition to these main effects, our candidate set for settlement probability included a model with an interaction between habitat type and Julian day because we hypothesized that settlement rates would change over the season differentially by habitat type. We hypothesized that abandonment rates would be extremely low; once territorial longspurs arrive at a breeding site post-migration, they are unlikely to abandon the site (With, 2021). Therefore, we did not include any models with covariates on abandonment rates. Model selection was again based on QAICc (Burnham and Anderson, 2002). We used empirical Bayes methods to derive estimates of latent occupancy from the most parsimonious model for each primary period from predicted posterior distributions using the ‘ranef’ function in R package ‘unmarked’ (Fiske and Chandler, 2011). All analyses were performed using R Statistical Software (v 4.1.2; R Core Team 2021).
2.3.2 | Longspur Abundance
We used open-population distance sampling models to estimate longspur abundance in crop and native sites and assess whether abundance changed differentially throughout the breeding season (Royle et al., 2004; Sollmann et al., 2015). Distance sampling is a common method for estimating abundance or density of wildlife populations and allows simultaneous estimation of detection probability without requiring repeat site visits (Buckland et al., 2001). Open-population distance sampling models allow explicit modeling of population dynamics over space and time, where data from repeat distance sampling surveys are used and populations are assumed open between survey periods (Sollmann et al., 2015).
We fitted open-population distance sampling models using the ‘distsampOpen’ function in R package ‘unmarked’ (Fiske and Chandler, 2011) and used information theory to evaluate support for competing models representing hypotheses about detection, initial abundance, and trends in abundance over the breeding season (Sollmann et al., 2015). We evaluated support for our a priori models in a phased approach. First, we used null models with the ‘trend’ dynamics parameterization to estimate the best-fitting detection function and mixture type based on our data. We then evaluated how well a fully parameterized model fit the data and estimated a variance inflation factor (ĉ ) using theNmix.gof.test in the R package ‘AICcmodavg’ from 500 bootstrapped simulations. Because the negative binomial model can overestimate population abundance (Ver Hoef and Boveng, 2007; Kery and Royle, 2015), we used the Poisson distribution for all subsequent models, inflating estimated standard errors by \(\sqrt{}\)ĉ and basing model evaluation and inference on the quasi-Akaike’s Information Criterion adjusted for finite samples (QAICc; Burnham and Anderson, 2002). We found evidence of moderate overdispersion using the Poisson distribution (ĉ = 1.9 for 2020 data, ĉ = 1.7 for 2021 data).
We developed models to evaluate the effects of survey conditions on detection probability. Variables hypothesized to influence detection probability included observer, wind speed, temperature, and start time (minutes past sunrise; Table 2). We evaluated a quadratic effect of start time because bird detections are usually highest 1–2 hours after sunrise (With, 2021). Initial screening indicated that detection was variable across observers, so we separated observers into 2 groups for each year (‘high’ and ‘low’ detection rates) based on relative coefficient estimates from a full model to reduce the number of parameters in candidate models while retaining large observer effects on detection. We used the backward selection approach described previously to eliminate uninformative parameters and identify a parsimonious sub-model for detection probability, which was retained in subsequent evaluations of abundance and seasonal trend.
We evaluated if initial abundance and seasonal trends differed by habitat type (crop vs. native; Table 2). We developed models that included the effect of habitat type on both initial abundance and trend, as well as all submodels. Model selection was again based on QAICc (Burnham and Anderson, 2002). We used Bayesian methods to derive true abundance estimates from the most parsimonious model for each survey round from predicted posterior distributions using the ‘ranef’ function in R package ‘unmarked’ (Fiske and Chandler, 2011). We analyzed data separately for the two years because differences in weather and drought conditions were likely to produce different population responses.
2.3.3 | Nest Phenology, Survival, and Reproductive Output
Nest Phenology. – For each nest, we calculated initiation date as the day the last egg was laid, although actual initiation of incubation is variable for passerines (Hébert, 2002; Badyaev et al., 2003). Initiation date was estimated based on clutch size, hatch date, or chick age and assuming an incubation period of 12 d (With, 2021). For nests found after clutch completion but destroyed before hatch, we assumed initiation to be 6 d prior to the midpoint of the active period. We plotted nest initiation dates to visualize patterns of nest initiation between crop and native sites and to assess differences between years.
Nest Survival. – We used the nest survival model in program MARK to model daily nest survival rate (DSR) and we fitted models in the R package ‘RMark’ (White and Burnham, 1999; Rotella et al., 2004; Laake, 2013). We built and evaluated a set of competing models representinga priori hypothesized relationships between DSR and habitat type (crop or native), nest initiation date, and year (2020, 21). We evaluated 15 models with all combinations of habitat type, initiation date, and year (Table 3). We also included a model with a quadratic effect of initiation date because other studies have shown DSR to be higher or lower mid-season (Weintraub et al., 2016; Skagen et al., 2018). We predicted DSR may exhibit a pseudo-threshold response in crop sites only, being low for nests initiated early and leveling off after fields were planted. Therefore, we included a model with a pseudo-threshold effect of initiation date and one including an interaction term with habitat type. We evaluated the relative support of models using Akaike’s Information Criterion corrected for finite sample size (AICc). Supported models with large model weights (AICc wi ) and AICc values ≤2 from the best fit model were considered parsimonious; when supported models differed by one parameter, we considered this parameter uninformative (Arnold, 2010; Burnham et al., 2011). To estimate nest survival probability, we used a 26-day nesting cycle beginning with the start of the laying period and multiplied DSR for each daily interval over a 25-day period from nest initiation to fledging (e.g., DSR25 for constant model). We calculated standard error for nest survival estimates using the Delta method (Powell, 2007).
Reproductive Output. – We calculated an index of nest density for each plot by dividing the number of nests located in each plot by the total search effort (hours) for that plot. We report the mean and standard deviation of relative nest density for each habitat type (crop vs. native). Incidental nests located outside of survey plots and nests found via rope dragging methods were excluded from this calculation. Importantly, we were unable to account for detectability of nests with behavioral search methods and it is possible detectability differed in crop and native sites. Detectability almost certainly differed by observer (Diefenbach et al., 2003; Giovanni et al., 2011); therefore, observers were rotated through different plots each day.
We tabulated maximum clutch size for all nests with known fates and the number of young fledged per successful nest. The number of young fledged was recorded as the number of chicks present 8–10 d after hatching, the typical fledging time for longspurs (With, 2021), unless some dead and some live fledglings were found during the final visit. We developed a set of generalized linear models to analyze the effects of habitat type and initiation date on the number of young fledged per successful nest using a Poisson distribution with a log link. We included an interaction term to test if the number of young fledged differed by both habitat type and initiation date (Table 4). Nests were removed from analysis if the number of young fledged was unknown. We evaluated relative model support AICc and used the best-fitting model to estimate the number of young fledged per successful nest.
2.3.4 | Habitat Conditions
We used generalized linear models to test hypotheses that specific vegetation attributes differed significantly between crop and native sites, longspur habitat changed structurally over the summer as plants grew, and such changes were more extreme in crop sites than in native sites. Variables included VOR, bare ground cover, grass and forb cover, litter cover and litter depth. For proportional response data (e.g., percent coverages), we used the binomial distribution and logit link function to fit GLMs (Chen et al., 2017). For all other vegetation measures, including VOR and litter depth, we used the identity link and log transformed the response variables to meet the assumptions of linear regression (Dunn and Smyth, 2018). For each habitat variable, we built and evaluated the same set of competing models representing a priori hypothesized relationships between habitat type and survey round (Appendix I).
We evaluated relative model support using AICc. Supported models with large model weights (AICcwi ) and AICc values ≤2 from the best fit model were considered equally parsimonious (Burnham et al., 2011). When a supported model differed from a top model by a single parameter, the additional parameter was considered uninformative (Arnold, 2010). We based inferences on effect sizes from a single top model and calculated model averaged estimates when models shared support (ΔAICc ≤2; Burnham et al., 2011).

3 | RESULTS

3.1 | Longspur Settlement
We deployed song meters at 8 and 16 sites in 2020 and 2021, respectively, half in crop fields and half in native sites. Due to equipment malfunction and failure of longspurs to establish territories at some sites, we were able to obtain data from 2 song meters in native sites and 4 song meters in crop sites in 2020, and 7 song meters in native sites and 7 song meters in crop sites in 2021. Overall, we collected >37 hr of useable recordings in 2020 and >100 hr in 2021.
Detection probability. – The top model for detection probability contained an effect of Julian day, minimum temperature, and a quadratic effect of minutes past sunrise (QAICcwi = 0.97; Table 1). Detection probability increased with Julian day (ꞵ = 0.99 ± 0.13 SE) and increased in response to minimum temperature (ꞵ = 0.08 ± 0.02). Detection probability was highest at ~90 – 100 minutes past sunrise, or 1.5 hours after sunrise (Figure 5).
Initial occupancy and settlement probability. – We found no evidence for an effect of habitat type on initial occupancy with the null model carrying virtually all support (QAICcwi = 0.98; Table 1). We found no evidence that settlement probability differed by habitat type with the model containing an effect of Julian day carrying virtually all support (QAICc wi = 0.98). Settlement probability increased for both habitat types with Julian day (ꞵ = 2.24 ± 0.68). Derived estimates of true occupancy for both crop and native sites increased from 0.52 (± 0.17 SE) on 7 April to 0.99 (± 0.01) on 30 April (Figure 6).
3.2 | Longspur Abundance
In 2020, we conducted initial occurrence surveys in 80 plots (36 crop and 44 native); 67% of crop and 20% of native plots were occupied. In 2021, we conducted initial surveys in 62 plots (35 crop and 27 native); 91% of crop and 33% of native plots were occupied. In 2020, we conducted 287 longspur abundance surveys at 24 crop sites and 22 native sites during 14 May – 19 July. We observed 5.4 ± 4.4 (mean ± SD) male longspurs in crop sites and 4.2 ± 3.3 in native sites. In 2021, we conducted 325 surveys at 25 crop sites and 25 native sites during 10 May – 14 July. We observed an average of 3.8 ± 3.2 and 3.2 ± 2.3 males per plot in crop and native sites, respectively. Most crop plots contained spring wheat (28 plots); we surveyed 4 summer fallow plots in 2020 and 8 in 2021 (Table 5).
Detection probability. – For both years, the top model contained an effect of observer group (Table 2). Detection probability was lower for observer group 2 and effect sizes were -1.67 ± 0.54 SE in 2020 and -0.95 ± 0.39 in 2021 (Figure 7). Confidence intervals for the effect sizes for other covariates on detection overlapped 0; therefore, only observer group was retained in subsequent abundance modeling (Arnold, 2010).
Initial abundance and seasonal trends. – We found support for an effect of habitat type on both initial abundance and seasonal trend for data collected in 2020 (QAICc wi= 0.91; Table 2). Expected initial abundance in crop sites was 17.4 ± 4.1SE birds per plot and the estimated seasonal trend was λ = 0.84 ± 0.04, indicating that abundance decreased by 16% over the season. Empirical estimates of true abundance for crop sites decreased from 16.8 (95% CI = 15.7–18.0) during the first survey round to 6.5 (5.6–7.8) during the sixth round. Expected initial abundance in native sites was 8.6 ± 2.0 birds per plot and increased slightly during the season (λ = 1.02 ± 0.05). Derived empirical abundance for native was 8.7 (95% CI = 7.8–9.7) during the first survey round and 9.4 (8.4–10.7) during the sixth round (Figure 8).
In 2021, we found no support for an effect of habitat type on either initial abundance or seasonal trend, with the null model carrying the most support (QAICc wi = 0.54; Table 2). Because of model uncertainty, we averaged results across all four supported candidate models. Expected initial abundance was similar in crop and native sites (12.5 ± 3.3 SE) and seasonal population sizes did not change much during the season (λ = 1.03 ± 0.04 SE in crop sites; 1.01 ± 0.04 in native sites). Derived estimates of true abundance for crop sites increased slightly from 12.3 (95% CI = 11.1–13.3) during the first survey round to 15.1 (13.2–17.0) during the sixth round. Derived estimates of true abundance for native sites were fairly stable across the season (12.7 (95% CI = 11.5–14.1) during the first survey round; 12.1 (10.8–13.4) during the sixth round; Figure 8).
3.3 | Nest Phenology, Survival, and Reproductive Output
We located 240 longspur nests, 111 in crop sites and 129 in native sites. Of these, 174 were located using behavioral cues of adults, 14 using rope-dragging methods, and 52 were incidental finds while observers were conducting other fieldwork. We spent 515 hours behavioral searching in crop fields and 421 hours behavioral searching on native sites, for a total of 936 hours nest searching using behavioral cues. Using 2–3 observers, we spent 76.5 person-hours rope dragging in crop fields and 22.5 person-hours rope dragging in native sites, for a total of 99 rope dragging person-hours.
Of the 240 nests, 222 had known fates (96 crop, 126 native). For the 18 remaining nests, we were unable to determine nest fate due to either conflicting clues at the nest site or weather events/farming operations preventing timely nest checks near expected fledge date. We were able to estimate the number of young fledged for 87 successful nests, 41 crop and 46 native. Apparent nest success was 44% in crop sites and 37% in native sites. Predation was the main cause of nest failure in both crop fields and native plots (Table 6). Other causes included weather, farming operations (crop only), and abandonment. Brown-headed cowbird (Molothrus ater ) parasitism rates were 1.8% and 7.8% in crop and native sites, respectively.
Nest Phenology. – Patterns of nest initiation were similar within crop and native sites each year, but median initiation dates in native sites were 6–11 d later than median dates in crop sites (Figure 9). In addition, the first and third quartiles were 6–10 d later in native sites. In 2020, median initiation date was 29 May (IQR = 25 d, n = 68) and 9 June (IQR = 26 d, n = 71) in crop fields and native sites, respectively. Longspurs nested through mid-July and there were two prominent peaks in nest initiation. In 2021, median date of initiation was 28 May (IQR = 17 d, n = 28) in crop sites and 3 June (IQR = 13 d, n = 55) in native sites. Nesting slowed significantly in late-June – early-July and there was only one main peak in nest initiation. Notably, the interquartile distance for initiation dates was 32% shorter in crop sites and 50% shorter in native sites during the 2021 drought year than during 2020.
Nest Survival. – The null model of constant daily nest survival was the best supported in the candidate set (AICcwi = 0.18; Table 3). Models including effects of habitat type, year, and initiation date, including models with different functional forms of initiation date, had approximately equal support as the null model, indicating that these parameters were uninformative. Average daily nest survival estimated from the null model was 0.944 ± 0.005SE and estimated nest survival over the 26-day exposure period (DSR25) was 0.236 ± 0.028.
Reproductive Output. – Relative nest density (±SD) was 0.153 ± 0.215 nests/hour/plot in crop sites and 0.233 ± 0.317 nests/hour/plot in native sites. Mean clutch sizes ± SD were 3.5 ± 0.8 and 3.3 ± 0.8 for nests occurring in crop fields and native sites, respectively. The mean number of young fledged per successful nest was 3.0 ± 1.1 SD in crop sites and 2.8 ± 0.9 in native sites. The null model was the best supported model in our candidate set of generalized linear models for number of young fledged per successful nest (AICcwi = 0.49; Table 4), indicating that neither nest initiation date nor habitat type was related to the number of young fledged. Models including the effects of habitat type and initiation date had approximately equal support as the null model, indicating that these parameters were uninformative. Estimated from the null model, the average number of young fledged per successful nest in both crop and native sites was 2.90 ± 0.18 SE.
3.4 | Habitat Conditions
We observed significant differences in vegetation conditions between crop and native sites that varied across survey rounds (Figure 10). Visual obstruction reading (VOR) estimates (cm ± SE) in 2020 changed from 0.81 ±1.42 in May to 17.81 ±1.43 in July in crop sites and from 1.95 ±1.51 to 2.61 ±1.46 in native sites. In 2021, VOR estimates (cm ± SE) changed from 0.72 ±1.35 to 1.48 ±1.35 in crop sites and from 0.68 ±1.34 to 0.28 ±1.34 in native sites. Bare ground coverage was significantly lower on native sites than crop sites during both years. Estimated bare ground (% ± SE) in 2020 was 45 ±6 in crop fields and 10 ±4 in native sites. In 2021, estimated bare ground was 42 ±6 in crop fields and 14 ±4 in native sites.
Estimated litter coverage in 2020 (% ± SE) was 25 ±5 in crop fields and 8 ±3 in native sites. In 2021, estimated litter coverage was 26 ±5 in crop fields and 11 ±3 in native sites. Estimated litter depth in 2020 (mm ± SE) changed from 4.66 ±1.30 in May to 0.58 ±1.31 in July in crop sites and from 1.57 ±1.34 to 1.05 ±1.32 in native sites. In 2021, estimates changed from 2.75 ±1.12 in May to 1.35 ±1.12 in July in crop sites and from 1.01 ±1.12 to 0.91 ±1.12 in native sites. Models of residual, forb, and grass cover indicated that these vegetation conditions were similar across habitat types and survey rounds. The same results were true for grass cover when we only compared native sites to wheat crop types (e.g., all crop plots classified as forb were removed).

4 | DISCUSSION

Collectively, our results did not support the hypothesis that crop fields are ecological traps for breeding thick-billed longspurs because, compared with longspur use of native grassland sites, there was no evidence of preference for crop habitat or of suppressed reproduction in crop fields. Specifically, settlement patterns of singing males were similar between crop and native sites and relative nest density was slightly lower in crop sites, providing no evidence for preferential selection of crop habitat. Nest survival, average clutch size, and the number of young fledged were similar between crop and native sites, providing no evidence for suppressed reproduction in crop fields. Additionally, precipitation and associated vegetation growth appeared to mediate longspur abundance and use of crop fields. Longspur abundance decreased throughout the breeding season in crop fields during a normal year (2020) as plant biomass increased whereas abundance did not decrease during a drought year (2021). Annual variation in timing of seeding coupled with drought effects on vegetation may increase the unpredictability of crop habitat among years.
We found that median nest initiation dates occurred 6–11 days earlier in crop sites despite similar settlement patterns for the two habitat types. Longspurs appeared to shift timing of nesting in crop sites, and perhaps this phenological shift is beneficial in habitat that changes structurally to become unsuitable late in the breeding season. Based on our results, crop sites may provide thick-billed longspur populations with viable alternative nesting opportunities in an area where native habitat has been reduced.
4.1 | Crop Fields as Potential Traps
We observed similar settlement patterns within breeding territories in crop fields and native prairie, indicating that selection cues and preference of longspurs were similar between habitat types. Although 50% of our study plots were occupied prior to song meter deployment in early April, increases in daily longspur occupancy were similar across habitat types and all sites were occupied by 27 April. Also, our nest density index was 29% lower in crop sites, though estimated precision was low and confidence intervals overlapped. Together, these findings suggest a similar preference of longspurs for crop fields and native rangeland habitats.
All measures of reproductive output (nest survival, clutch size, number of young fledged per successful nest) were similar between habitat types. Although we observed higher early nest failures in crop sites as a result of farming activities (e.g., seeding, discing, and plowing), thick-billed longspurs are quick to renest (<10 d) (Mickey, 1943; Felske, 1971; With, 2021), and we often found new nests close to failed nest locations. While the most common cause of nest failure in both habitat types was predation, higher predation rates on native sites resulted in overall similar nest survival rates (~24%) in crop fields and native prairie habitat.
In contrast to expectations, some farming activities, including rolling (field leveling) and spraying, did not result in nest damage or abandonment, and harvest of crops occurred too late in the season to affect nesting longspurs. Longspur nest bowls were constructed below the soil surface; farming activities such as rolling that did not disturb the soil did not negatively affect nests (n=9) regardless of nest stage. Most (>95%) of our crop fields were sprayed with herbicides (glyphosate, 2-4 dichlorophenoxyacetic acid [2-4D]) twice per season and ≥5 fields were also sprayed once with organic fertilizers. Herbicide application typically consisted of a pre-spray to eliminate weeds around the time of seeding and a second application in June when plants were 12–15 cm tall (M. Sather, USFWS, pers. comm.). Although nests were active when fields were sprayed, spraying did not directly result in losses of eggs or nestlings. However, we did not assess potential indirect effects of herbicide and fertilizer spraying on nestling growth rates or subsequent fledgling survival.
Flooding and hail destroyed nests in both crop (n=8) and native sites (n=4). Nest abandonment was often due to partial predation, weather, brown-headed cowbird parasitism, and possibly frequent disturbance by predators or perceived predators. On a few occasions in native sites (n=3 nests), we found all nestlings apparently uninjured but laying outside the nest. These nestlings never survived and were never returned to the nest by adult longspurs. We suspect this to be the activity of brown-headed cowbirds or other passerine nest predators (Pietz and Granfors, 2000; Pietz et al., 2012).
4.2 | Longspur Abundance and Use of Crop Fields
Precipitation and vegetation structure appeared to mediate longspur abundance in crop fields but not native sites. Longspur abundance was relatively stable throughout the season within native sites in both years and averaged 8–12 birds per plot (0.63 birds per ha). In a season of normal precipitation (2020), longspur abundance was higher in crop sites than native sites early in the season (April–May) when crop biomass was low but declined with the growth of crops. In contrast, longspur abundance increased slightly throughout the breeding season in crop fields during a drought year (2021) when crop growth was minimal. Because rates of nest abandonment were ubiquitously low, declining abundances of longspurs across time imply reduced nesting attempts in crop sites during a year of normal precipitation, though we could not confirm this with unmarked birds.
Longspurs used all types of crop fields in our study area, including lentil, pea, flax, wheat, canola, mixed cover crop, and summer fallow. Although we didn’t have sufficient sample plots in summer fallow treatments (n = 12) to include this as a separate category in our analyses, we consistently observed fewer longspurs in summer fallow fields compared to other crop types. We found very few nests in summer fallow fields over both years (n = 10). Summer fallow fields were planted in strips, with fallow sections intermixed with planted sections. Fields planted in the narrowest strips, and hence having more abrupt edges, were rarely used by longspurs (on 4–5 out of 6 surveys we detected 0 birds). Lower abundance of longspurs in these areas is consistent with avoidance of habitat edges in grassland birds (Johnson and Igl, 2001; Renfrew et al., 2005; Sliwinski and Koper, 2012; Thompson et al., 2015). In addition, fallow portions are tilled multiple times during the breeding season to control weeds. Therefore, nests in unplanted portions of summer fallow fields have a higher risk of being destroyed later in the season, unlike nests in annual crop sites. It is also possible that frequent tilling of fallow fields results in different soil conditions, generating different invertebrate resource availability than that found in minimum-tilled fields (Stinner and House, 1990; Kladivko, 2001).
4.3 | Timing of Nesting
In native prairie habitats, longspurs select territories on south-facing slopes during the early breeding season where snow melts and the ground warms faster (Felske, 1971; Greer, 1988; Shaffer et al., 2019). Bare ground cover was higher in crop fields than native sites throughout the breeding season, and exposed soils warm faster than vegetated soils (Song et al., 2013). Although territory settlement phenology was similar between habitat types, median nest initiation dates during both years occurred 6–11 days earlier in crop fields than in native sites. Thus, earlier warming of crop fields may allow earlier nest initiation and egg laying resulting from favorable microclimatic conditions or an earlier invertebrate food supply (Felske, 1971; Greer and Anderson, 1989). However, we did not assess thermal or other microclimatic conditions at nests. In addition, the range of nest initiation dates and therefore length of nesting period was significantly shorter during the drought year in 2021. Longspurs are known to forego nesting or experience lower reproductive success during periods of extreme drought (Felske, 1971; Shaffer et al., 2019). Our results indicate that longspurs may initiate nests earlier in crop than native sites but experience a shorter breeding period in both site types during drought.