Sandhill Crane Colt Survival in Minnesota

Sandhill crane Antigone canadensis populations rebounded beginning in the late 20th century; thus, conservationists often tout this rebound as a success story. Dubovsky (2016) estimated that the Eastern Population (EP) increased 3.9%/y from 1982 to 2012 and that the Midcontinent Population (MCP) increased 0.8% per y from 1979 to 2009. More recently, there has been an overlap in the breeding distributions of these two sandhill crane populations in central Minnesota, largely due to recolonization by EP sandhill cranes of formerly unoccupied areas (Lacy et al. 2015; Wolfson et al. 2017). At the population level, survival rates of sandhill crane juveniles in their first year of life (hereafter colts) can affect recolonization and population persistence. For example, fledged subadult sandhill cranes in Minnesota were more exploratory in their movements than adults (Wolfson et al. 2020) and are therefore more likely to influence recolonization and range expansion. Sandhill cranes are also less fecund than most other migratory game birds, with a typical brood size of one or two (Drewien et al. 1995); relatively low recruitment of colts into the breeding population also likely affects sandhill crane population ecology (Gerber and Kendall 2016). Although population growth rate is typically most sensitive to adult survival in long-lived vertebrates (Gaillard et al. 1998; Eberhardt 2002), variation in reproductive parameters can drive population dynamics when adult survival varies little (Gerber and Kendall 2016; Wilson et al. 2016), and aspects of sandhill crane reproductive success, such as nest success and brood survival, influence population recruitment (Ivey and Dugger 2008). Predator densities, access to forage and roosting habitats, and proximity to metropolitan areas show associations with reproductive success in sandhill cranes (Fox et al. 2019). For example, a predator control program in Oregon reduced American mink Neovison vison predation on colts from 27–35% to 9% (Ivey and Scheuering 1997).

Population models using age composition or sightings of marked birds as indicators of nest success and first-year survival address the lack of estimates of colt survival rates (Johnson and Kendall 1997; Gerber and Kendall 2016). However, Arnold et al. (2016) concluded that there is considerable uncertainty when modeling populations of sandhill cranes because of low precision in juvenile survival rate estimates. Colts typically fledge at 70–75 d posthatch, but they remain with their parents throughout the first year, and estimates of survival rates from hatch to fledging represent a critical information need since the mid-1990s (Tacha et al. 1994). Despite this need, there remains a lack of published information regarding prefledging colt survival, and estimates of survival rates to fledging vary widely between different studies, from 3% (95% CI = 0.1–19%) in Wisconsin (Berzins 2020) to a pooled estimate of 31% (±2.8% SE) in central Wisconsin and northeastern Illinois (Fox et al. 2019). Furthermore, Fox et al. (2019) reported a survival rate in Illinois (27%) that was about half of the observed survival rate in Wisconsin (54%). Currently, there are no estimates of colt survival rates during the prefledging period in the region where the EP and the MCP have recently been found to overlap (Wolfson et al. 2017). Our objective was to estimate a colt survival rate during their first summer in this region of breeding range overlap by using data derived from sandhill cranes marked with radio transmitters as part of a larger movement ecology study (Wolfson et al. 2017, 2020; Wolfson 2018).

We captured and marked colts during June–October 2014 and 2015 in seven Minnesota counties (Figure 1; 46 to 48°N, −96 to −93°W) in a zone between historical breeding ranges of MCP and EP sandhill cranes (Wolfson et al. 2017, 2020). These counties in Minnesota encompass the junction of sections of several ecoregions, including the Northern Minnesota Drift and Lake Plains and Minnesota and Northeast Iowa Morainal sections (MNDNR 2015). The study area contains an ecocline between eastern tallgrass prairie and northern hardwood forest. The landscape is a mosaic of wetlands and lakes in the lowlands and pasture, agricultural crops, second-growth forest patches, and residential development in the uplands. Wetland areas consist primarily of emergent vegetation communities dominated by cattails Typha spp. and sedge Carex spp. meadows. Known potential predators of colts (Ivey and Scheuering 1997) in the study area include gray wolves Canis lupus, coyotes Canis latrans, bobcats Lynx rufus, raccoons Procyon lotor, and American mink and bald eagles Haliaeetus leucocephalus, great horned owls Bubo virginianus, and other raptors.

Colts leave the nest within 48 h of hatching (Fox et al. 2019); we located colts by identifying and monitoring adult sandhill cranes known or thought to be associated with a nest (Wolfson et al. 2017). We captured colts (aged 19–77 d) by hand during daylight hours and marked them with 6.7-g glue-on very-high-frequency (VHF) radio transmitters (Advanced Telemetry Systems, Isanti, MN; Figure 2). The VHF transmitters were designed with a mortality sensor that resulted in a change in signal following a period of inactivity. We sewed each transmitter into the center of a fabric patch (colored to approximate the plumage of a colt) and attached the fabric square to the colt's middorsal region by using waterproof, nontoxic, quick-drying eyelash adhesive (Andrea Lashgrip Eyelash Adhesive; Ardell, Los Angeles, CA; Spalding et al. 2001; Fox et al. 2019). We collected a blood sample from the metatarsal vein just below the tibio-tarsus joint of each captured sandhill crane and determined sex via subsequent genetic analyses (Avian Biotech, Tallahassee, FL). We weighed colts and extrapolated from an age-to-mass curve to assign an age at capture (Figure S1, Supplemental Material; Fox et al. 2019).

We monitored marked colts by using VHF receivers (Advanced Telemetry Systems) and a handheld three-element antenna. Once colts gained sufficient mass to carry a larger transmitter (just before fledging, ∼50–60 d old), we recaptured and fit them with a global positioning system–global system for mobile communications (GPS-GSM) transmitter (models CTT-1060a-LB and CTT-1060-LM-BT3; Cellular Tracking Technologies, Rio Grande, NJ). We attached GPS-GSM transmitters (transmitters weighed ∼60 g, ∼3% of body mass) above the left tibio-tarsus joint by using a two-piece leg band that recorded locations every 15 min during daylight. Upon capture (or recapture), we inspected each colt to assess overall health and body condition. No colts had any physical impairments, and we released them in good condition. For a full description of capture and handling methodology, see Wolfson et al. (2017). The University of Minnesota Institutional Animal Care and Use Committee (protocol 1403-31362A) approved all capture and handling methods.

We monitored colts equipped with VHF transmitters approximately every 10 d. For colts equipped with GPS-GSM transmitters, we remotely verified survival each day based on movement. If a VHF transmitter indicated mortality or a GPS-GSM transmitter indicated a tight clustering of locations, field crews investigated the location. We assigned the proximate cause of mortality based on the preponderance of evidence when possible (e.g., colt remains, carcass location, predator sign), but generally the approximately 10 d between monitoring colts equipped with VHF transmitters precluded definitive determinations of cause of death.

We right censored observations at 120 d if the colt was still alive and its transmitter was still functioning. We wrote a function in R (R Core Team 2021) to calculate the likelihood of the parameter values given the capture dates, last known date alive, and date of death or censoring (Text S1, Supplemental Material). We used the optim function in R (R Core Team 2021) to find the maximum likelihood estimates of β₀ and β₁. We used a cluster-level bootstrap in which we resampled all colts from the same brood with replacement to relax the assumption that brood mates have independent fates when calculating confidence intervals (CIs; Fieberg et al. 2020). The cluster-level bootstrap allows for positive or negative correlation between fates of brood mates. We provide all raw data and code in Table 1 and in the Supplemental Material (Table S1; Text S1). The data and code used to estimate the parameters and plots are available in the Data Repository of the University of Minnesota (Table S1; Text S1; https://hdl.handle.net/11299/225562). We used hatch success and survival estimates until 3 wk of age (our youngest colt captured was 19 d old) from the literature (Fox et al. 2019) to also estimate survival rate from nest initiation to the end of summer.

We captured and marked 36 colts (12 female, 20 male, 4 unknown sex) from 29 broods (seven sets of brood mates, 22 colts that were the only member of their brood marked) during June–August 2014 and 2015 (Table 1). Mean age at capture was 42 d (SD = 13; range, 19–77 d). We never observed one colt (2015-6) again after capture (suspected collar failure) and did not include it in further calculations. We excluded a second colt (2015-31) because we did not weigh it at capture and therefore could not estimate its age. Estimates of daily mortality probabilities rapidly decreased as colts aged (Figure 3A), resulting in the cumulative survival curve that approached an asymptote at approximately 80–120 d (Figure 3B). We estimated period survival rate from 19 to 120 d to be 0.52 (90% CI, 0.36–0.71; Figure 3A). Nine colts died during our monitoring period; eight of the nine deaths occurred prefledging. We could not determine the cause of eight of these deaths. One of these eight colt carcasses was in a bald eagle nest, but could have been scavenged. The cause of the remaining mortality was a collision with a vehicle. Of the nine colts with confirmed mortalities, six had been first fit with a VHF transmitter followed by a GPS-GSM transmitter, two had been fit with VHF and GPS-GSM transmitters simultaneously, and one had been fit with only a GPS-GSM transmitter. The remaining 26 colts survived beyond the duration of this study (Table 1).

Fox et al. (2019) estimated colt survival rate to fledging (11 wk) at 0.54 in central Wisconsin. Using their data on weekly survival for weeks 1, 2, and 3 (∼0.925, 0.875, and 0.90, respectively), we estimated an apparent survival rate from age 0 to 21 d to be 0.73 (Fox et al. 2019; Figure 2). Next, using that apparent survival rate as a proxy for survival rate from 0 to 19 d in our study, we estimated survival rate from 0 to 120 d to be 0.38 (0.73 × 0.52).

Survival during the first summer of life is an important parameter for understanding population dynamics of sandhill cranes. Our survival rate estimate is conditional on colts surviving until we handled them (earliest age at handling was 19 d). Our survival rate estimate from 0 to 120 d was 0.38, calculated using the best available 0–3-wk survival estimate (Fox et al. 2019). Wheeler et al. (2019) estimated colt survival from earliest age at banding (35 d) to 1 y at 0.818 (95% CI, 0.755–0.882) in Wisconsin. This rate is much higher than our observed rate, but does not account for losses of colts less than 35 d old. Fox et al. (2019) reported 31% of colts that hatched survived to fledge in Illinois and Wisconsin, but demonstrated that colts from nests nearer urban areas exhibited higher survival rates. Two additional studies in central Wisconsin reported extremely low colt survival rates to fledging of 1–2% in Necedah National Wildlife Refuge (McLean 2019) and 3.4% in Horicon Marsh (Berzins 2020). Our estimate provides additional context to the high degree of spatial variation in regional colt survival rates, with implications for population modeling and understanding sandhill crane population dynamics.

Based on our analyses, colt survival rates increased with age until remaining relatively constant postfledging, likely because colts became less vulnerable to predation as they grew large enough to evade mammalian and avian predators (Jones et al. 2017). Wheeler et al. (2019) observed that 1- and 2-yr-old sandhill cranes displayed much higher survival rates than colts. The age at fledging in our study site was approximately 70–75 d (D.W.W., unpublished data), which roughly corresponded to just before the period when daily survival rates approached a constant value. Other studies of sandhill crane colt survival cited predation as the leading source of mortality (Ivey and Scheuering 1997; Nesbitt et al. 2008); however, only Fox et al. (2019) reported daily survival rates, finding that daily survival rate increased with colt age. Increasing survival rates with age is also a common pattern in young animals of other taxa (Cox et al. 2014; Daly et al. 2015; Severud et al. 2019).

Our estimates of colt survival rate in a region where the breeding distribution of two sandhill crane populations has recently overlapped were much higher than colt survival rate estimates from other midwestern U.S. studies where only one of these two populations breeds. Both of these sandhill crane populations are subject to harvest, likely including the population segments that breed in Minnesota. In Minnesota, a harvest season for MCP sandhill cranes was established in 2010 (Kruse et al. 2012), although there is not currently a harvest season for EP sandhill cranes. However, there are some locations in Minnesota where both EP and MCP sandhill cranes co-occur during the MCP sandhill crane hunting season (Wolfson et al. 2017), and sandhill cranes from both populations that breed in Minnesota are subject to harvest elsewhere. The MCP sandhill cranes are harvested in several states in the central United States and adjacent provinces in Canada. The EP sandhill cranes are harvested under a quota system in Alabama, Kentucky, and Tennessee; allocation of quotas is based on counts from annual surveys conducted at areas where sandhill cranes concentrate during fall and winter (Seamans 2021). Our estimates of colt survival rates can inform population models of both MCP and EP sandhill cranes, especially at the scale where the breeding distributions of these two populations overlap.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content of functionality of any supplemental material. Queries should be directed to the corresponding author.

Text S1. R code to conduct survival analyses and create plot (Figure 3) of sandhill crane Antigone canadensis colts in Minnesota, 2014–2015.

Available: https://doi.org/10.3996/JFWM-21-097.S1 (763 KB HTML) and https://doi.org/10.13020/qy1k-8269

Table S1. Raw survival data of sandhill crane Antigone canadensis colts (n = 36) in Minnesota, 2014–2105. Column headings are colt_ID (unique identifier for each colt), nest_ID (unique identifier for each nest), sex, Cap.age.days (colt age in days at capture), capture.date (date of capture in month/day/year format), Surv (binary indicating if colt survived [1] or died [0]), cause_death (cause of death, if known), lastdateknownalive (last date colt was observed alive in month/day/year format), firstdateknowndead (first date colt observed dead in month/day/year format), and bd (birthdate of colt in year/day/year format).

Available: https://doi.org/10.3996/JFWM-21-097.S2 (2 KB CSV) and https://doi.org/10.13020/qy1k-8269

Figure S1. Observed relationship between mass (grams) and age (days) of sandhill crane Antigone canadensis colts (n = 105 from 64 colts) during 2008–2014 in northern Illinois and southern Wisconsin (Fox et al. 2019). Colts were of known age (monitored from hatch to measurement). Black dots depict observed measurements, blue line is a loess smooth, and the shaded area is a 95% confidence interval.

Available: https://doi.org/10.3996/JFWM-21-097.S3 (94 KB TIFF)

We thank N. Cross, J. Dachenhaus, J. Fox, D. Fronczak, K. Kovach, G. Kramer, E. Ulrey, E. Wells, and S. Zudrow for assistance with fieldwork and many private landowners for granting access during our study. We thank W. Brininger, T. Buker, L. Domine, W. Ford, A. Hewitt, G. Knutsen, M. North, and H. Saloka for logistical support. Thanks to K. Barrett, G. Henderson, K. Larson, B. Liddell, M. Loss, E. North, J. Provost, T. Stursa, E. Thorson, and many more for information on sandhill crane locations. Thanks to B. Maas for piloting the helicopter during surveys. Funding for this project was provided by the U.S. Fish and Wildlife Service (USFWS) and U.S. Geological Survey (Science Support Program) through Research Work Order No. 101 at the U.S. Geological Survey, Minnesota Cooperative Fish and Wildlife Research Unit; by the Minnesota Environmental and Natural Resources Trust Fund as recommended by the Legislative-Citizen Commission on Minnesota Resources; by the USFWS Webless Migratory Game Bird Program; and by the Minnesota Department of Natural Resources. J.F. received partial salary support from the Minnesota Agricultural Experimental Station. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the University of Minnesota or the State of Minnesota. We thank D. Fronczak and D. Johnson for reviewing earlier versions of this manuscript. We also thank three anonymous reviewers and the Associate Editor for providing comments that improved the manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.