Are life history events of a northern breeding population of Cooper's Hawks influenced by changing climate?

Abstract Numerous studies have demonstrated earlier timing of spring migration and egg‐laying in small passerines, but documentation of such responses to recent climate change in the life histories of higher trophic feeding birds such as raptors is relatively scarce. Raptors may be particularly susceptible to possible adverse effects of climate change due to their longer generation turnover times and lower reproductive capacity, which could lead to population declines because of an inability to match reproductive timing with optimal brood rearing conditions. Conversely adaptively favorable outcomes due to the influence of changing climate may occur. In general, birds that seasonally nest earlier typically have higher reproductive output compared to conspecifics that nest later in the season. Given the strong seasonal decline in reproductive output, and the heritability of nesting phenology, it is possible that nesting seasons would (adaptively) advance over time. Recent climate warming may release prior ecological constraints on birds that depend on food availability at the time of egg production, as do various raptors including Cooper's Hawks (Accipiter cooperii). Under this scenario, productivity, especially clutch size, might increase because it is likely that this reproductive demographic may be the most immediate response to the earlier seasonal presence of food resources. We demonstrated a statistically significant shift of about 4–5 days to an earlier timing of egg‐hatching in spring across 36 years during 1980–2015 for a partially migratory population of Cooper's Hawks in Wisconsin, United States, which is consistent with a recent study that showed that Cooper's Hawks had advanced their timing of spring migration during 1979–2012. Both studies occurred in the Great Lakes region, an area that compared to global averages is experiencing earlier and increased warming particularly in the spring in Wisconsin. The nesting period did not lengthen. We suggest that the gradual shift of six consecutive generations of hawks was likely in response to recent climate change or warming. We did not detect any long‐term temporal change in average clutch or brood sizes. However, such indices of reproduction are among the highest known for the species and thus may be at their physio‐ecological maximum for this population. Our study population appears to show resilience to and does not appear to be adversely influenced by the recent rate of changing climate at this time.

tors including Cooper's Hawks (Accipiter cooperii). Under this scenario, productivity, especially clutch size, might increase because it is likely that this reproductive demographic may be the most immediate response to the earlier seasonal presence of food resources. We demonstrated a statistically significant shift of about 4-5 days to an earlier timing of egg-hatching in spring across 36 years during 1980-2015 for a partially migratory population of Cooper's Hawks in Wisconsin, United States, which is consistent with a recent study that showed that Cooper's Hawks had advanced their timing of spring migration during 1979-2012. Both studies occurred in the Great Lakes region, an area that compared to global averages is experiencing earlier and increased warming particularly in the spring in Wisconsin. The nesting period did not lengthen. We suggest that the gradual shift of six consecutive generations of hawks was likely in response to recent climate change or warming. We did not detect any long-term temporal change in average clutch or brood sizes. However, such indices of reproduction are among the highest known for the species and thus may be at their physio-ecological maximum for this population. Our study population appears to show resilience to and does not appear to be adversely influenced by the recent rate of changing climate at this time.

| INTRODUCTION
Much evidence suggests that recent climate change has altered life history events of numerous species of birds (e.g., Brommer, 2004;Jenni & Kery, 2003;Wood & Kellermann, 2015). Studies particularly demonstrate earlier timing of spring migration and egg-laying in birds throughout the Northern Hemisphere, although these findings are strongly biased toward research on small passerines (Dunn, 2004;Lehikoinen, Saurola, Byholm, Lindén, & Valkama, 2010;Nielsen & Møller, 2006). There are relatively few studies that have investigated the possible effects of climate change on higher-level avian consumers, including predatory hawks (Clutton-Brock & Sheldon, 2010;Lehikoinen et al., 2010Lehikoinen et al., , 2013. However, some recent studies on raptors have shown earlier spring migration (Jaffré et al., 2013;Lehikoinen et al., 2010;Sullivan, Flaspohler, Froese, & Ford, 2016), earlier hatching dates (Lehikoinen et al., 2013), and shifts to earlier (Filippi-Codaccioni, Moussus, Urcun, & Jiguet, 2010) and later timing of autumnal migrants (Rosenfield, Lamers, Evans, Evans, & Cava, 2011;Van Buskirk, 2012), with such seasonal shifts being consistent with compensatory response to global warming (Sullivan et al., 2016). Sullivan et al. (2016) recently demonstrated phenological shifts to earlier spring migration of several raptor species, including Cooper's Hawks during 1979-2012 in the Great Lakes region of North America, and they indicated that such shifts were consistent with decadal climatic oscillations and global climate change. Such shifts could likely lead to earlier arrival times at breeding territories, which in turn could lead to earlier breeding schedules, a lengthened nesting period, and possibly increased productivity (perhaps an adaptive response sensu Lehikoinen et al., 2010;Fontaine, Stutzman, & Gannes, 2015).
Alternatively, it is possible that changes in timing of migration could produce maladapted individuals given a possible decoupling between migration (and breeding) schedule of hawks and the temporal variation in available prey (Lehikoinen et al., 2009;Rosenfield et al., 2011).
However, the possible consequences on the aforementioned life history events are unknown for the raptors in the Great Lakes region studied by Sullivan et al. (2016). Indeed, Sullivan et al. (2016) indicated concern for the viability of these species' regional populations given in part the lack of data to indicate whether behavioral responses are sufficient to maintain adequate productivity, compounded with the fact that raptors tend to exhibit relatively low reproductive capacities.
This concern is significantly aggravated by the fact that Sullivan et al. (2016) do not know the specific destinations and/or breeding locales of the individuals of the various species counted and thus where population data for each species could be gathered to potentially address their concern. Moreover, Sullivan et al. (2016) noted that certain life history attributes (e.g., longer generation turnover times) appear to constrain responsiveness to changing climate, which could lead to asynchrony in phenological events across trophic levels and cause population declines. Mismatches between reproductive timing and food availability are expected to become more severe with increasing trophic level (Both, Van Asch, Bijlsma, Van Den Burg, & Visser, 2009;Visser, Both, & Lambrechts, 2004). Lehikoinen et al. (2010) demonstrated earlier timing of spring arrival dates and hatching dates in the migratory Eurasian Sparrowhawk (Accipiter nisus) in Finland and attributed these results to climate change (increased warming in the spring) in their study areas across 29 years . This small raptor is a specialized predator of passerines and perhaps responded to the earlier arrival of their migratory prey (Lehikoinen et al., 2010). They suggested that this temporal dynamic provided improved hunting conditions and allowed sparrowhawks to gain resources for breeding more easily during the preincubation period and, in turn, could advance timing of breeding and increase productivity. Notably, they documented increased mean brood sizes in this raptor but claimed that this result was likely due to reduced exposure to organochlorine pollutants. However, they indicated that they could not exclude the possibility that an earlier onset of breeding had caused increased production.
Recent climate change may release prior ecological constraints on birds that depend on resource availability at the time of egg production (e.g., Cooper's Hawks [Accipiter cooperii]; Drent, Fox, & Stahl, 2006;Gienapp & Visser, 2006;Rosenfield, Bielefeldt, & Cary, 1991;Rosenfield, Sonsthagen, Stout, & Talbot, 2015). Cooper's Hawks in our Wisconsin population and other birds that nest earlier in the season typically have higher reproductive output of eggs and/or offspring compared to conspecifics that nest later in the season (Perrins, 1970;Newton, 1979;Rosenfield & Bielefeldt, 1999;R.N. Rosenfield, unpubl. data). It is possible that nesting season would advance over time as a consequence of an adaptive response by breeding birds especially given the strong seasonal decline in reproductive output, and the heritability of nesting phenology (van Noordwijk, van Balen, & Scharloo, 1981;Van der Jeugd & McCleery, 2002). Thus, it is possible that an advanced timing of nesting could result in increased productivity, as was shown for Ficedula flycatchers in Europe in which earlier advancement of laying prompted by climate change (i.e., increased spring warming) resulted in increased clutch sizes .
Clutch size is perhaps the most central index to avian reproduction as it exhibits a strong correlation with age at maturity, and offspring and adult survival in birds (Jetz, Sekercioglu, & Bohning-Gaese, 2008;Ricklefs, 2000). Seasonality of resources is globally the predominant driver of clutch variation in birds across geographic gradients (Jetz et al., 2008). We contend it would have been particularly revealing if Lehikoinen et al. (2010) had found that clutch sizes had increased over the study period (clutch counts were not presented) when hatching dates advanced because it is likely that this reproductive demographic may be

K E Y W O R D S
Accipiter cooperii, adaptive response, climate change, clutch size, Cooper's Hawk, egg-hatching dates, shift in reproductive timing the most temporally proximate or immediate response to the earlier arrival of prey by sparrowhawks. We note that higher mean clutch counts, but not mean brood counts, occur in years with earlier hatch dates in our Wisconsin populations of Cooper's Hawks (Rosenfield & Bielefeldt, 1999; and see Section "2"). Clutch size also indexes reproductive potential, life history trade-offs pertinent to migratory behavior in birds including possibly Cooper's Hawks (Meiri & Yom-Tov, 2004;. To our knowledge, there are no studies of higher trophic predatory birds in which variation in clutch size was investigated as a potential demographic response to climate change. We conducted long-term nesting studies of a partially migratory population of Cooper's Hawks in Wisconsin, which population is near the northern extent of this species' continental range in the Great Lakes region (Rosenfield & Bielefeldt, 1993;Bielefeldt, Rosenfield, Stout, & Vos, 1998;R.N. Rosenfield, unpubl. data (Rosenfield, Bielefeldt, Booms, Cava, & Bozek, 2013;Rosenfield, Hardin, Bielefeldt, & Anderson, 2016).
We note that Sullivan et al. (2016) reported Cooper's Hawk to have advanced its spring migration in our study region during the study years included herein and that changes to higher average temperatures from 1950 to 2009 in the Great Lakes region exceed global averages (Hayhoe, Vandorn, Croley, Schlegal, & Wuebbles, 2010).
Wisconsin is among the 10 fastest warming states in the United States (Tebaldi, Adams-Smith, & Heller, 2012), particularly regarding increased spring temperatures documented in our study areas (Kucharik, Serbin, Vavrus, Hopkins, & Motew, 2010;Zhao & Schwartz, 2003). Further, there is strong evidence of a general and gradual trend of advanced spring plant phenology and delays in the onset of fall throughout the Great Lakes region including Wisconsin throughout our study years (Bradley, Leopold, Ross, & Huffaker, 1999;Ewert, Hall, Smith, & Rodewald, 2015;Kucharik et al., 2010). Several species of migratory birds preyed upon by Cooper's Hawks in our study sites (e.g., Bielefeldt, Rosenfield, & Papp, 1992;R.N. Rosenfield, un-publ. data) exhibited a significantly gradual and earlier spring arrival to Wisconsin during our study years (Bradley et al., 1999). The advanced timing of spring migration of these birds was linked to earlier and higher spring temperatures due to climate change (Bradley et al., 1999). Notably, one of these species, the American Robin, is the most commonly detected prey item in spring during the preincubation period at Cooper's Hawk nest sites on our study areas throughout all study years (Bielefeldt et al., 1992;R.N. Rosenfield, unpubl. data), and thus, likely an important resource for accumulation of body reserves requisite for egg production (Rosenfield & Bielefeldt, 1999).

American Robin [Turdus migratorius], red-winged blackbird [Agelaius
We use a 36-year data set involving six consecutive generations (estimated at ~6 years per generation (Rosenfield, Bielefeldt, Affledt, & Beckmann, 1995)) of Cooper's Hawks during 1980-2015 to investigate whether our study population has gradually advanced its egg-laying period during spring in accord with documented gradual spring warming in Wisconsin during our study years. We also assessed whether average clutch and/or brood sizes have increased (sensu Lehikoinen et al., 2010) such that potential changes in all three life history events may be consistent (or adaptive) with recent climate warming in Wisconsin (Kucharik et al., 2010) for this primarily birdcatching hawk whose diet includes migratory birds (Bielefeldt et al., 1992(Bielefeldt et al., , 1998Rosenfield et al., 2010, R.N. Rosenfield unpubl. data).

| Study areas
We studied breeding Cooper's Hawks during 1980-2015 at two principal areas in central and southeastern parts of Wisconsin as described by Rosenfield et al. (1995Rosenfield et al. ( , 2010 and Rosenfield and Bielefeldt (1996). Forest, South Unit. These study sites were chosen without preconceptions about their suitability for nesting Cooper's Hawks (Bielefeldt et al., 1998). We note that we have been unable to show in our multidecadal studies that habitat (i.e., rural or urban, conifer plantation or nonplantation, presumptive site quality as indexed by consistency of nesting area use, and high breeding density) is related to size of clutch or brood counts, nesting phenology, annual adult survival, and production of recruits, or fitness in Cooper's Hawks in Wisconsin (Rosenfield & Bielefeldt, 1999;Rosenfield, Bielefeldt, Sonsthagen, & Booms, 2000;Rosenfield, Bielefeldt, Rosenfield, Booms, & Bozek, 2009;Rosenfield, Hardin et al., 2016, R.N. Rosenfield, unpubl. data).

| Field procedures
Each year, we found most nests (>90%) before egg-laying by listening for dawn vocalizations  or by searching for partially constructed nests during the preincubation stage, about mid-March through late April (Bielefeldt et al., 1998). We are thus able to examine productivity without adjusting for the biases that might have resulted from excluding breeding attempts that failed prior to discovery (Rosenfield et al., 2000). Further descriptions of our study design, study areas, and nest finding techniques can be found in Rosenfield andBielefeldt (1991, 1996).
We made at least two visits to nests to assess reproduction. One visit included climbing to nests during the mid-incubation period (about mid-May) to obtain completed clutch counts and another climb about mid-June when nestlings were 16-19 days of age, or about 70% of fledgling age, to ascertain brood size, age, and band young at successful nests (Rosenfield & Bielefeldt, 2006;Rosenfield, Grier, & Fyfe, 2007). This schedule avoided the criterion of 80% of fledgling age suggested for other raptor species, an age at which visits could result in premature fledging of some nestlings and/or inaccurate brood counts of young (Rosenfield, Grier et al., 2007;Rosenfield, Grier, & Fyfe, 2007).
Brood counts herein are from successful nests in which at least one young reached 16-19 days of age (Rosenfield et al., 2013). Hatching dates per nest were determined by backdating from estimated nestling ages of the oldest chick based on plumage development of known-age birds (Meng & Rosenfield, 1988;Bielefeldt et al., 1998, R.N. Rosenfield unpubl. data).
Estimated hatching dates were based only on young aged by RNR, who aged >95% of all bandable young each year across all study years.

| Data analyses
We used simple linear regression to assess how Julian date for the 50th percentile (i.e., median) of the of the total number of nests with estimated hatching dates in each of 36 study years changed across time (i.e., to ascertain whether a statistically significant shift in hatching date per year had occurred following procedures in Rosenfield et al. (2011)).
We used linear regression because it is often used to detect possible changes of timing of events in migratory birds due to climate change (e.g., Wood & Kellermann, 2015); linear regression was also used to investigate the possible effects of climate warming on timing of abiotic events and phenologies of 55 different species of plants and birds (including songbirds preyed upon by Cooper's Hawks) in Wisconsin during our study years (Bradley et al., 1999;Ewert et al., 2015;Serbin & Kucharik, 2009). Invocation of similar analytical techniques by different researchers studying similar phenomenon facilitates a tenable discussion of results across studies (Whitlock & Schluter, 2009). We used the median because it was less sensitive to the effects of data outliers that could skew results (see Rosenfield et al., 2011).  (Table 1) was same as the average of median Julian hatch dates for the first (1980)(1981)(1982)(1983)(1984)(1985) of six study generations of Cooper's Hawks.
We also report the extent of the shift in hatching dates for all com-

| RESULTS
We  nestlings/year and an overall average of 3.7 young/year (SE = 0.04; median = 4) across all study years (Figure 2). We did not detect a statistically significant trend in mean clutch (slope = −0.001, SE = 0.004, p = .89) and mean brood sizes/year (slope = 0.002, SE = 0.004, p = .61) across 36 years in Wisconsin (Figure 2). We conclude that the gradual change to an earlier nesting phenology did not appear to influence average clutch or brood counts in our study population of Wisconsin Cooper's Hawks.

| DISCUSSION
Although numerous studies have demonstrated earlier timing of spring migration and egg-laying in small passerines (Wood & Kellermann, 2015), documentation of such responses to recent climate change in the life histories of secondary consumer birds such as raptors is scarce (Lehikoinen et al., 2009(Lehikoinen et al., , 2010 . We are unaware of any study that investigated potential variation or increase in clutch counts in raptors as a possible adaptive response to changing climate. Sullivan et al. (2016) (Bradley et al., 1999;Ewert et al., 2015). For example, the gradual shift of 0.31 days/year in spring to earlier egg-laying in Wisconsin Cooper's Hawks is similar to a gradual and average of 0.21 days/year of earlier timing of collective spring phenologies for 55 species of Wisconsin plants and birds during our study years (these birds were predominately migratory songbirds whose earlier arrival may trigger egg-laying in Cooper's Hawks, see below; Bradley et al., 1999). We note that a gradual pace of response to climate warming is common for many species (including birds), both globally and in Wisconsin (Bradley et al., 1999;Kucharik et al., 2010;Stanley, 1988;Wood & Kellermann, 2015). We also highlight that the shift to earlier hatching of Cooper's Hawks occurred seasonally during spring, which, along with winter, are the seasons exhibiting the greatest warming in Wisconsin on our study areas during our study years (Kucharik et al., 2010;Zhao & Schwartz, 2003). Further, the overall average shift of about 4-5 days to earlier egg-laying in Cooper's Hawks is similar to an average advancement of about 6-7 days for spring in Wisconsin during our study years (Kucharik et al., 2010;Serbin & Kucharik, 2009 is inversely related to nesting phenology, such that larger birds tend to nest seasonally earlier and, and as aforementioned, exhibit greater productivity (Newton, 1979;Rosenfield & Bielefeldt, 1999;Warkentin, Espie, Lieski, & James, 2016). Given that size is heritable in the Cooper's Hawk, it is possible that the nesting season would advance over time as a consequence of an evolution toward larger body size in breeding birds that breed earlier (as an adaptive response) given the strong seasonal decline in reproductive output and the heritability via body size of nesting phenology (van Noordwijk et al., 1981;Rosenfield & Bielefeldt, 1999;Van der Jeugd & McCleery, 2002). However, we have shown that body size for both male and female breeding Cooper's Hawks has not changed during our study years (Rosenfield et al., 2013;. We previously linked body size in north-central, North American breeding populations of Cooper's Hawks, including Wisconsin birds, to a phenotypically plastic response to prey size (i.e., size of hawks track size of prey; Rosenfield et al., 2010;Sonsthagen et al., 2012). We note however that the diet of breeding Cooper's Hawks in Wisconsin has not changed over our study years (Bielefeldt et al., 1992;R.N. Rosenfield, unpubl. data).
Further, long-term high survivorship and relatively high nesting densities (Rosenfield et al., 2009(Rosenfield et al., , 2013 R.N. Rosenfield, unpubl. data), along with the long-term and relatively high average productivity indices presented herein, suggest that possible density-dependent factors such as varying food availability and disease (investigated by Rosenfield et al., 2002;Stout et al., 2005) in Wisconsin Cooper's Hawks have unlikely influenced the reproductive ecology and productivity we document here (Bielefeldt et al., 1998;. We reiterate that we have been unable to show in our multidecadal studies that habitat (i.e., rural or urban, conifer plantation or nonplantation, presumptive site quality as indexed by consistency of nesting area use, and high breeding density) is related to size of clutch or brood counts, nesting phenology, annual adult survival, and production of recruits, or fitness in Cooper's Hawks in Wisconsin (Rosenfield & Bielefeldt, 1999;Rosenfield et al., 2000Rosenfield et al., , 2009Rosenfield, Bielefeldt et al., 2016, R.N. Rosenfield, unpubl. data Lehikoinen et al. (2010) who documented a significant temporal increase in mean brood sizes (they did not report clutch counts), we found a stable trend in mean clutch and brood sizes across our study years despite an earlier nesting schedule. While they suggested that increased average brood size was a possible response to climate change, we detected no change in average clutch and brood sizes in our study population of Cooper's Hawks during the years when climate warming was occurring in Wisconsin. We suggest that timing of food availability by perhaps earlier arriving migratory passerines in Wisconsin (especially the American Robin (Bradley et al., 1999), which is a common prey item of Cooper's Hawks in spring) is favorable for breeding when females are acquiring resources across weeks before laying eggs. We note that Snyder and Wiley (1976) found that egg-laying by Cooper's Hawks in Arizona was triggered by the arrival of migrant songbirds, and Millsap, Breen, and Phillips (2013) (Peck & James, 1983). However, these larger clutch counts appear to be anomalies (Rosenfield & Bielefeldt, 1993). We are unaware of any studies documenting more than six nestlings at one Cooper's Hawk nest. The largest clutch and brood sizes on our study areas were six eggs and six young, which maxima occurred throughout our study years but rarely (3% and 0.9% of 613 and 740 clutches and broods, respectively) and without any temporal trend (R.N. Rosenfield, unpubl. data Further, it is tenuous to compare reproductive indices from our work with those from studies that that did not principally locate their nests (as did we) before eggs were laid; as such, latter studies likely have calculated reproductive indices that were biased toward successful nests, whereas our productivity findings were not (Rosenfield et al., 2000).
Irrespective of these possible methodological shortcomings, we note that the overall average of 4.3 eggs per clutch for our study population is exceeded by only one study (4.4 for a 5-year study in British Columbia ), while our overall average of 3.7 young per nest is higher than any other average brood count reported for any other Cooper's Hawk population in North America (Rosenfield & Bielefeldt, 1993;Rosenfield, Bielefeldt, et al., 2007). It is possible that adult hawks in our study years on average exhibited their maximum eco-physiological output of eggs and young per nest given the interannual variation in environmental attributes on our apparently high breeding quality study areas (Bielefeldt et al., 1998;, the intrinsic behavioral, physiological, and genetic qualities of breeding Cooper's Hawks on our study sites, and noting that average productivity indices were stable across six consecutive generations when egg-hatching dates became earlier. Adaptive responses to recent climate change in many migratory birds, which generally are believed to occur mechanistically through phenotypic plasticity in long-lived species such as raptors ( (Ewert et al., 2015). We were unable to address these hypotheses or alternative explanations with our study design. We urge raptor researchers with long-term reproductive data sets, especially those with cross-generational data sets, to analyze their data within the context of how climate change potentially has (or has not) stressed various life histories in their study subjects and thus allow for elucidation of the ability (or lack thereof) of secondary avian consumers to respond to climate change and maintain or possibly enhance individual fitness and/or population viability (Ewert et al., 2015).
We highlight that our long-term, cross-generational stable high nesting densities, relatively high productivity indices, marked flexibility in habitat use, and high annual adult survivorship for both breeding male and female Cooper's Hawks on our study areas indicate a healthy breeding population of Cooper's Hawks (e.g., Rosenfield et al., 1995Rosenfield et al., , 2009R.N. Rosenfield, unpubl. data). Our study population appears to show resilience to and does not appear to be adversely influenced by the recent rate of changing climate. This demographic synopsis seems to favorably counter the concern regarding the aforementioned possible ill effects of an earlier migration of Wisconsin Cooper's Hawks due to recent climate warming in the Great Lakes region expressed by Sullivan et al. (2016). However, the rate of climate change is expected to increase in the Great Lakes region, and thus, further monitoring of our study population is requisite for further insights into the potential influence of a changing climate on the life histories of breeding Cooper's Hawks in Wisconsin (Ewert et al., 2015). We emphasize that such monitoring cannot be predicated on indefinite extrapolation of historical data and that reliable insights into future population status and viability during a changing climate will require periodic updates on the same ilks of demographic information assessed in earlier baselines (sensu York, Dowsley, Cornwell, Kuc, & Taylor, 2016).