Introduction
Birds are an exceptionally diverse class, occupying almost all habitats on Earth (Lees et al. Reference Lees, Haskell, Allinson, Bezeng, Burfield and Renjifo2022). However, many species are threatened due to anthropogenic activities, associated habitat loss and resulting competition between humans and birds for the same resources (Amulike et al. Reference Amulike, Fuller, Houlihan and Griffin2020; Fakarayi et al. Reference Fakarayi, Mashapa, Gandiwa and Kativu2016; Harris and Mirande Reference Harris and Mirande2013; Tilman et al. Reference Tilman, Clark, Williams, Kimmel, Polasky and Packer2017). Birds display distinct behavioural changes in response to environmental changes, making them valuable ecological indicators (Fraixedas et al. Reference Fraixedas, Lindén, Piha, Cabeza, Gregory and Lehikoinen2020; Lees et al. Reference Lees, Haskell, Allinson, Bezeng, Burfield and Renjifo2022; Marcelino et al. Reference Marcelino, Silva, Gameiro, Silva, Rego and Moreira2020). For example, some environmental changes result in a mismatch of important resources or optimal conditions during key life stages, forcing migratory species to adjust their migration patterns (Sillett et al. Reference Sillett, Holmes and Sherry2000; Visser et al. Reference Visser, Gienapp, Husby, Morrisey, De La Hera and Pulido2015; Zurell et al. Reference Zurell, Graham, Gallien, Thuiller and Zimmermann2018), whilst non-migratory bird species attempt to shift their home ranges (Santisteban et al. Reference Santisteban, Benkman, Fetz and Smith2012). Furthermore, delayed insect emergence and reduced food availability associated with climatic changes disrupt hatchling growth and extend the time it takes for adults to replenish the energetic costs incurred during breeding (Charmantier and Gienapp Reference Charmantier and Gienapp2014; Møller et al. Reference Møller, Rubolini and Lehikoinen2008; Pearce-Higgins et al. Reference Pearce-Higgins, Dennis, Whittingham and Yalden2010).
Larger birds with longer generation times and delayed reproduction, as well as those that rely on specialised habitats, face a higher extinction risk, especially if they cannot adapt to changing environments (Bird et al. Reference Bird, Martin, Akçakaya, Gilroy, Burfield and Garnett2020; Brown and Orians Reference Brown and Orians1970; Owens and Bennett Reference Owens and Bennett2000; Toussaint et al. Reference Toussaint, Brosse, Bueno, Pärtel, Tamme and Carmona2021). Cranes (Gruidae) are among the most long-lived bird species, with generational lengths of approximately 12 years (vs the average of approximately three years for all bird species) (Bird et al. Reference Bird, Martin, Akçakaya, Gilroy, Burfield and Garnett2020; Zelelew et al. Reference Zelelew, Bekele and Archibald2019). This, together with their habitat specialisations, magnifies their extinction risk (Edwards et al. Reference Edwards, Converse, Swan and Moehrenschlager2022; Owens and Bennett Reference Owens and Bennett2000), making cranes among the most threatened bird families globally (Beilfuss et al. Reference Beilfuss, Dodman and Urban2007; Harris and Mirande Reference Harris and Mirande2013). The Grey Crowned Crane Balearica regulorum is sub-Saharan Africa’s most vulnerable crane species and was uplisted to “Endangered” on the International Union for Conservation of Nature (IUCN) Red List in 2012 (Harris and Mirande Reference Harris and Mirande2013). Grey Crowned Cranes have a strong reliance on wetlands for breeding and grasslands for foraging (Beilfuss et al. Reference Beilfuss, Dodman and Urban2007); however, commercial afforestation and agricultural intensification now expose this species to the risk of habitat loss (Fakarayi et al. Reference Fakarayi, Mashapa, Gandiwa and Kativu2016; Morrison and Bothma Reference Morrison and Bothma1998; Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022; Weyer et al. Reference Weyer, Granger, Hill and O’Connor2015). The resulting close association of cranes and agriculture (Olupot Reference Olupot2016) is linked to increasing reports of breeding failures following damming and drainage of wetlands, powerline collisions, disease, and predation; while poisoning, hunting, and illegal trading of these birds are also increasing (Amulike et al. Reference Amulike, Fuller, Houlihan and Griffin2020; Galloway-Griesel et al. Reference Galloway-Griesel, Roxburgh, Smith, McCann, Coverdale and Craigie2022; Harris and Mirande Reference Harris and Mirande2013).
The long-term persistence of a population, particularly in response to climate change, is directly related to an individual’s reproductive output per breeding event or lifetime, as well as the probability that their offspring will reach sexual maturity (Assersohn et al. Reference Assersohn, Marshall, Morland, Brekke and Hemmings2021). Grey Crowned Cranes breed during the summer rainy season when seeds and insects are abundant, a strategy which improves parent condition and chick survival (Austin et al. Reference Austin, Morrison and Harris2018; Gichuki Reference Gichuki2000; Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022). Grey Crowned Cranes lay large clutches compared to other cranes (Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022), and their strategy of synchronised parental incubation assists with protecting eggs from aerial predators while nesting in tall, inaccessible wetlands offers concealment from terrestrial predators (Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Ndung’u2020; Wen et al. Reference Wen, Su, Hu, Xiong, Liu and Long2021). These reproductive behaviours and nest-site preferences have made traditional on-foot data collection of breeding metrics challenging, resulting in data paucity and ineffective conservation planning for this species (Morrison Reference Morrison1998). More detailed metrics are needed to better understand the factors that directly or indirectly impact adults, eggs or chicks (e.g. through fluctuating incubation temperatures, food quality, and availability or protection from predators; Wu et al. Reference Wu, Zou and Ma2014), ultimately impacting breeding success (Soriano-Redondo et al. Reference Soriano-Redondo, Inger, Sherley, Rees, Abadi and McElwaine2023; Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022; Wu et al. Reference Wu, Zha, Zhang and Yang2009).
To our knowledge, no study has investigated how surrounding macro-environmental characteristics influence the reproductive outcomes of Grey Crowned Cranes, and the absence of extensive breeding data hinders our ability to determine the breeding outcomes of this species. Therefore, our study aimed to (1) determine the general breeding outcomes of this species in a key agricultural area in South Africa, (2) identify the stage where reproductive failure is most likely by assessing the transition probabilities between early life stages, and (3) understand how specific macro-environmental factors (both human-related and natural) affect reproductive outcomes.
Methods
Study area
Nest-site monitoring took place during the 2022/3 Grey Crowned Crane breeding season in the Underberg, Himeville, Franklin, and Kokstad regions of KwaZulu-Natal, South Africa (Figure 1). This summer-rainfall region (650–1,000 mm per annum) has large areas of open grasslands and wetlands while agriculture (intensive cropping and dairying, extensive beef and sheep) and commercial forestry (Pinus spp. and Eucalyptus spp.) are the primary land-use types occurring in these regions. Rainfall records supplied by farmers in each study region showed that rainfall during the 2022/3 breeding season exceeded the yearly average captured in previous years (Supplementary material Table S1), likely resulting from a La Niña event (Jones Reference Jones2022).
Figure 1. Distribution of the four main study sites within the southern Drakensberg, highlighting land-cover within surroundings. Specific nest-site locations have been omitted because this is potentially sensitive information and we do not have the landowner’s permission to share this information publicly. Land-cover is based on a provincial land-cover map produced by the provincial conservation authority, Ezemvelo KwaZulu-Natal Wildlife (Ezemvelo KZN Wildlife Reference Ezemvelo2008).
Data collection
General breeding outcomes and transition probabilities
We identified potential Grey Crowned Crane breeding pairs using behavioural cues following Wamiti et al. (Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022) – a lone crane foraging close to a water-body usually implied that its partner was nesting nearby. As Grey Crowned Cranes nest in areas that are difficult to access, we used a standard Mavic Air 2S drone (DJI Technology Co., Shenzhen, China) (595 g, 1-inch 20 MP sensor, 8 × zoom, 65 dB low noise propeller) to monitor all nests following the nest-site approaching methodology detailed in Demmer et al. (Reference Demmer, Demmer and McIntyre2024) (Figure S1). Drone flights were kept to a minimum to decrease disturbance and lower the risk of nest or chick abandonment by parents. When monitoring families with chicks, the small size and cryptic colouring of Grey Crowned Crane hatchlings required closer and lower flights for accurate counts. During these flights, the pilot cautiously manoeuvred the drone, approaching the grouping at a diagonal angle, without hovering, to capture breeding metrics (spending approximately 20 seconds at the nest). Video recording was enabled throughout all approaches to facilitate post hoc recording of breeding metrics.
We considered a nest-site “active” if eggs were present or a crane was incubating on an identified nest (i.e. incubating; Figure 2B and D). Locations where courting pairs were observed without nests were revisited every two to three weeks. During each breeding site observation, we counted the number of eggs, hatchlings, juveniles, and fledglings and differentiated between these stages based on their size and plumage. Distinguishing juveniles from fledglings depended on whether chicks could fly (commonly takinge place 55–100 days after hatching; Gichuki Reference Gichuki2000). Each nest-site was revisited at 2–3-week intervals from incubation until the offspring could fly or until there was no evidence of the family after two subsequent monitoring events. If a nest-site was not monitored from incubation, it was excluded from our analyses as missing the initial nesting stage results in biases towards identifying nest-sites that successfully hatch chicks (Jehle et al. Reference Jehle, Yackel Adams, Savidge and Skagen2004). Monitoring families after fledging was sometimes challenging due to the family’s increased mobility. For this reason, we considered a breeding attempt successful when the offspring reached the juvenile stage. This usually included fledged chicks, but in other cases, families had moved off to join nearby flocks soon after chicks became more mobile, making it nearly impossible to identify specific families.
Figure 2. Drone photographs depicting: (A) the view of a nest-site from afar; (B) a close-up of a Grey Crowned Crane Balearica regulorum nest-site; (C) a flooded nest-site; (D) incubating parent; (E) parents with two hatchlings and one egg at their nest-site; (F) parent foraging with juveniles.
Breeding outcomes were measured in three ways: (1) total counts of offspring at each stage; (2) probability of offspring transitioning from one stage to another (total count in subsequent stage divided by total count in the previous stage); and (3) nest-site success when at least one juvenile or fledgling was observed.
Macro-environmental impacts
Daily rainfall data for the 2022/3 breeding season were sourced from local farmers known to keep accurate records by measuring millimetres of rainfall with standard 10 cm2 circular rain gauges. These data were collected for each of the five regions where nest-sites were situated. We also collected macro-environmental data describing the broader patch and landscape-scale habitat type around each nest-site, further details of which are provided in Table 1.
Table 1. Macro-environmental descriptors measured for each nest-site
We used GIS software to measure direct distances between each nest-site and important environmental descriptors (i.e. nearest building, road, agricultural land, and water-body) and to determine the dam and wetland surface areas (measured in m2) (Table 1). The only categorical descriptor, vegetation height, was categorised through post hoc assessments using drone photographs taken within 1 m of each nest-site. The five vegetation height categories included “short” for grass, “medium” for reeds, and “tall” for bulrushes. Nest-sites with a combination of these vegetation types were labelled as “short to medium” or “medium to tall”.
Statistical analyses
All statistical analyses were conducted in R version 4.3.1 (R Core Team 2022). To assess whether there was bias in our observations towards a particular study region, we compared the number of observations at each nest-site across regions with a Poisson generalised linear model (GLM) and a log link function.
General breeding outcomes
The date at which each developmental stage was most likely to be detected (an integer with 1 representing the first observation in the breeding season) was determined using a log-linear mixed effects model using the “lmer” function from the lme4 package (Bates et al. Reference Bates, Machler, Bolker and Walker2015). Developmental stage (eggs, hatchlings, juveniles, and fledglings), breeding outcome (success or failure), and the interaction of these two factors as independent variables were included as fixed effects.
The number of offspring observed for a breeding attempt was assessed using a Poisson generalised linear mixed effects model (GLMM), implemented using the “glmer” function from the lme4 package, with developmental stage as the independent variable. Changes in the number of offspring over time (used to describe how and when stage transitions took place) were assessed with a generalised additive mixed model with a Poisson error distribution and a log link function, using the “gamm” function from the mgcv package (Wood Reference Wood2011). Here, the number of offspring in each stage was allowed to vary as a function of days since nest-site detection. Since repeated observations took place at each nest-site throughout the breeding season, we controlled for this in all the above-mentioned mixed effects models by including a random intercept of the nest-site itself.
Transition probabilities
We used a binomial GLM with a logit link function to assess the probability of transitioning from one stage to the next. The transition types that we assessed were the three sequential transition types (eggs–hatchlings, hatchlings–juveniles, juveniles–fledglings), the egg-to-juvenile transition as a measure of nest-site success (the chance that an egg will survive to a juvenile), and the chance of obtaining 100% success (whether all eggs would transition to juveniles). In a second model we assessed the change in transition probability of two transition types (eggs–hatchlings and hatchlings–juveniles) across a gradient of rainfall intensity using a binomial GLMM with a logit link function. We calculated rainfall intensity (mm/day) as the sum of the daily rainfall during each transition divided by the number of days the transition took to complete where higher values indicate greater rainfall over a shorter time period. Here, we included a random intercept of “region” to control for differences in total rainfall between the regions (Table S1). There was also a difference in the mean rainfall intensity experienced during these two transition types (χ2 = 17.366, df = 1, P <0.001) and across months (F7,1236 = 340.34, P <0.001, Table S1), with greater rainfall intensity observed earlier in the breeding season. We controlled for this by incorporating a random slope effect of rainfall intensity within months to capture the variable rainfall intensity observed between the months. Juvenile fledgling transitions did not occur in all the regions and so this transition was excluded from this analysis. Unless otherwise stated, all results from these regression models are presented as means (± 95% confidence interval [CI]).
Macro-environmental impacts
We used a Principal Components Analysis (PCA) to explore trends in Grey Crowned Crane breeding outcomes across biophysical descriptors collected for each nest-site. Since the environmental descriptors differed by orders of magnitude, these measurements were log(x + 1) transformed and then standardised by dividing each measurement by the maximum transformed distance recorded for that measurement type, scaling each measurement between zero and one. We conducted the PCA using the “rda” function from the vegan package (Oksanen et al. Reference Oksanen, Simpson, Blanchet, Kindt, Legendre and Minchin2022). We tested for the separation of centroids between successful and unsuccessful hatching (hatching success) and the production of juveniles (overall success) by implementing two analyses of variance, using Euclidean distance matrices, with the “adonis2” function in the vegan package. For both the multivariate and the multi-model inference analyses, a nest-site was considered to have successfully hatched offspring if at least one egg hatched. Similarly, a nest-site was considered successful overall if at least one juvenile was observed at the nest or with the monitored family.
We identified the most important environmental descriptors using multi-model inference to complement the multivariate analyses. A binomial GLM including all main effects was fitted initially. Models with all possible combinations of predictor variables were then assessed using the “dredge” function from the MuMIn package (Bartoń Reference Bartoń2023). From these we selected all models within two Akaike information criterion (AIC) units of the best-performing model (∆AIC ≤2). To reduce the influence of potentially non-significant parameters included in these selected models, we applied a model averaging procedure, using the “model.avg” function from the MuMIn package to produce the final model (Arnold Reference Arnold2010).
There is a strong correlation between likelihood ratio tests and ∆AIC. A ∆AIC <2 (indicating that there is little evidence that models are significantly different) translates to P <0.157, approximating the 85% CI (Arnold Reference Arnold2010). As a result of this, we report both the 95% and 85% CIs in the final averaged model’s presentation for these analyses (Arnold Reference Arnold2010; Sutherland et al. Reference Sutherland, Hare, Johnson, Linden, Montgomery and Droge2023). Effects significant at the 85% but not at the 95% level were referred to as “moderately significant”.
Results
A total of 95 nests were identified and monitored across the five regions during the 2022/3 breeding season (Table S1). On average, each nest-site was observed 3.72 (3.44; 4.00) times, totalling 351 observations across all 95 nest-sites. The number of observations per nest-site did not differ between regions (χ2 = 3.238, df = 4, P = 0.519; Table S1), indicating no monitoring bias between regions. Rainfall intensity did vary across regions, with Himeville recording the highest and Kokstad the lowest rainfall intensity respectively (F4,1239 = 2.523, P = 0.039; Table S1).
The average dates where observations of the different stages were made are illustrated in Figure 3. These dates were significantly different between developmental stages (F1,618.43 = 106.14, P <0.001), but did not differ between successful and unsuccessful nests (F1,110.00 = 0.01, P = 0.913). The interaction between these two variables was also non-significant (F1,946.65 = 0.41, P = 0.666).
Figure 3. Mean (± 95% confidence interval) dates at which Grey Crowned Crane Balearica regulorum developmental stages were observed between successful and unsuccessful sites.
General breeding outcomes
The number of offspring per nest differed between developmental stages, with significantly more eggs recorded than juveniles or fledglings (χ2 = 27.08, df = 3, P < 0.001; Figure 4A). The average number of eggs detected was 2.53 (2.30; 2.78). This average includes interesting observations of four nest-sites, each with four eggs. The number of eggs continuously declined and were no longer detected after approximately 30 days of monitoring (estimated-df = 3.82, F = 70.69, P <0.001). Hatchlings (estimated-df = 3.95, F = 29.19, P <0.001), juveniles (estimated-df = 4.49, F = 23.26, P <0.001), and fledglings (estimated-df = 3.13, F = 35.32, P <0.001) all showed unimodal-type responses with peak counts occurring at 24, 48, and 146 days, respectively (Figure 4B).
Figure 4. Counts of Grey Crowned Crane Balearica regulorum offspring at different developmental stages. (A) Error bars indicate 95% confidence interval, and shared letters indicate no evidence of significant differences. (B) Change in counts throughout the season. Dashed vertical lines indicate when the most offspring were detected for a developmental stage.
Transition probabilities
Grey Crowned Crane transition probabilities varied significantly among stage types (χ2 = 30.73, df = 2, P <0.001; Figure 5A). No significant difference was found in the probability of transitioning from eggs to hatchlings (0.384 (0.293; 0.484)) in comparison to either transitioning from eggs to juveniles (0.311 (0.266; 0.361)) or transitioning from juveniles to fledglings. However, transitioning from eggs to hatchlings was less likely to occur than the transition from hatchlings to juveniles (0.660 (0.521; 0.776)), but the probability of transitioning from hatchlings to juveniles was not different from the probability of transitioning from juveniles to fledglings (0.590 (0.432; 0.731)). Transitioning from eggs to juveniles was less probable than transitioning from hatchlings to juveniles, and from juveniles to fledglings. Furthermore, 10.6% (7.8; 14.2) of nests showed a 100% success rate, where all eggs transitioned to juveniles – less than all other transition types considered. The probability of developing from eggs into hatchlings was negatively affected by increased rainfall intensity (χ2 = 4.071, df = 1, P = 0.044; Figure 5B), but transitioning from hatchlings to fledglings was not (χ2 = 0.054, df = 1, P = 0.816; Figure 5B).
Figure 5. Transition probabilities (A) (mean ± 95% confidence interval) from one development stage to the next where (i) shared letters denote no difference, (ii) eggs–juveniles represents the probability that an egg was successfully reared, and (iii) 100% success indicates the probability that all eggs in a nest were raised successfully. (B) Across a rainfall intensity gradient (points are empty and jittered to show overlap).
Macro-environmental impacts
The first principal component (PC1) explained 22.8% of the variation in nest locations (eigenvalue: 0.16). It included wetland areas (21.2%) and distances to buildings (16.0%), water edges (15.6%), tar roads (14.8%), and farm roads (13.3%) as the most closely related descriptors (Figure 6). The second principal component (PC2) explained 16.1% of the variation (eigenvalue: 0.11) and represented distances to natural grasslands (20.1%), buildings (15.7%), tar roads (14.8%), rivers (11.9%), and water edges (10.7%) (Figure 6). There were significant differences in the composition of nest-site metrics when considering hatching success (PERMANOVA, 999 permutations, pseudo-F1,95 = 3.04, P = 0.006; Figure 6 – empty circles), but not in their overall success (PERMANOVA, 999 permutations, pseudo-F1,95 = 1.86, P = 0.073; Figure 6 – filled circles). These separations occurred along PC1.
Figure 6. Grey Crowned Crane Balearica regulorum nest-sites and environmental descriptors along PC1 (eigenvalue = 0.16) and PC2 (eigenvalue = 0.11) of a Principal Components Analysis. Black points represent nest-sites, solid black lines are environmental descriptors. Envelopes represent the 95% confidence interval for the centroids (larger coloured points) of successful (blue) or failed (red) hatching (empty circle, dashed envelope) and overall breeding (filled circle, solid envelope) success. All descriptors are distances except for island (nest on island), vegetation height (height class around the nest-site), and areas (size of wetland or dam).
Multi-model inference identified 11 candidate models to explain hatching success and eight models to define overall breeding success. Model averaging procedures (Figure 7 and Table S2) revealed that hatching success decreased when nest-sites were further away from natural grasslands and buildings (at the 95% CI level) and when located at greater distances from dirt roads (at the 85% CI level). Considering overall success, nest-sites further away from buildings were less successful (at the 95% CI level), while reduced distance to natural grasslands and larger nearby dam surface areas were positively associated with increased breeding success (at the 85% CI level).
Figure 7. Model averaged effect sizes of covariates on Grey Crowned Crane Balearica regulorum breeding success. Positive values indicate greater success with greater distance or area. Negative values indicate greater success with reduced distance (measured in metres) or area (measured in m2/1,000). Error bars indicate 95% (red) and 85% (blue) confidence intervals. Dashed vertical lines indicate no effect. Shaded values are non-significant at P <0.157 (representative of ∆AIC = 2).
Discussion
General breeding outcomes and transition probabilities
Grey Crowned Cranes lay some of the largest clutches in the crane family, with findings in Kenya reporting average clutch sizes of approximately 2.5 eggs (Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022), which aligns with our observations. Pomeroy (Reference Pomeroy1980) observed larger crane clutch sizes at higher altitudes, possibly explaining our rare finding of four nests with more than three eggs (n eggs = 4) (Ivey and Dugger Reference Ivey and Dugger2008; Sundar and Choudhury Reference Sundar and Choudhury2005). Other research reported that the reproductive success of cranes increases with stronger pair bonds, while frequent socialisation and having previously undertaken multiple breeding attempts also increases their reproductive success (Barwisch et al. Reference Barwisch, Mewes and Schmitz Ornés2022; Hammers et al. Reference Hammers, Richardson, Burke and Komdeur2012; Ivey and Dugger Reference Ivey and Dugger2008; Teitelbaum et al. Reference Teitelbaum, Converse and Mueller2017).
Unlike other crane species which frequently only invest in one hatchling (e.g. Wattled Cranes Grus carunculata), Grey Crowned Cranes attempt to raise all hatchlings (Morrison Reference Morrison2015; Morrison and Bothma Reference Morrison and Bothma1998; Zelelew et al. Reference Zelelew, Bekele and Archibald2019). Gichuki (Reference Gichuki1993) estimated an 83.6% hatching rate in Grey Crowned Cranes in Kenya, while Zelelew et al. (Reference Zelelew, Bekele and Archibald2019) reported a higher hatching rate (91.3%) in Black Crowned Cranes Balearica pavonina but inferred a low overall breeding success since less than half of the clutches in their study fledged. We observed substantially lower hatching rates than these studies, with only 38.4% of eggs hatching, 66.0% of hatchlings transitioning to juveniles, and 59.0% of juveniles surviving to the fledging stage. Multiplying these transition probabilities suggests a 14.95% chance that an egg will reach the fledging stage. These results support the earlier mentioned concerns of Assersohn et al. (Reference Assersohn, Marshall, Morland, Brekke and Hemmings2021) and suggests that Grey Crowned Crane breeding is most vulnerable during their nesting phase. After hatching, crane chicks remain vulnerable for another two to three months until they fledge. However, their survival probability usually increases thereafter, as they become older and more adept at identifying danger and escaping predators themselves (Fox et al. Reference Fox, Semel and Ward2019).
In our study, 10.6% of Grey Crowned Crane pairs managed to ensure that all eggs in their clutch transitioned to juveniles. The hatchling to juvenile/fledgling transition probability in Grey Crowned Cranes (one to two fledglings per pair) is higher than that of some other crane species. For example, Wattled Cranes are known to display low reproductive outputs, with the highest transition rate of hatchling to juvenile being recorded as 0.42 chicks per pair (in 1999, vs 0.66 for Grey Crowned Cranes) and approximately 0.25 chicks per pair in subsequent wetter years (Bento et al. Reference Bento, Beilfuss and Hockey2007). Another comparison can be made with the Endangered Whooping Crane Grus americana, whose recovery is believed to be highly dependent on their ability to reproduce successfully (Butler et al. Reference Butler, Metzger and Harris2017). Yet, over eight years, Spalding et al. (Reference Spalding, Folk, Nesbitt, Folk and Kiltie2009) estimated a 30% hatching rate, with 47% of these Whooping Crane chicks surviving until the fledging stage – this translates to an overall breeding success of only 14%. Grey Crowned Cranes’ general reproductive output throughout the duration of our study is comparable to that of Whooping Cranes (Spalding et al. Reference Spalding, Folk, Nesbitt, Folk and Kiltie2009), which supports similar concerns about declining Grey Crowned Crane population trends.
Several studies have identified extreme weather events and human disturbances as major drivers of breeding failures among some Critically Endangered coastal and wetland ground-nesting birds (Ferreira et al. Reference Ferreira, Hansen, Parrish, Pierce, Pulham and Taylor2005; Powell et al. Reference Powell, Hodgman, Glanz, Osenton and Fisher2010; Smart et al. Reference Smart, Gill, Sutherland and Watkinson2006; Spalding et al. Reference Spalding, Folk, Nesbitt, Folk and Kiltie2009; Van De Pol et al. Reference Van De Pol, Ens, Heg, Brouwer, Krol and Maier2010). Therefore, in the sections that follow, we discuss how these climatic and landscape-level factors impact the reproductive outputs of Grey Crowned Cranes.
Macro-environmental impacts
Breeding birds incur high energetic costs when defending territories, incubating, and raising young (Gichuki Reference Gichuki2000), but these costs are escalated under suboptimal climatic conditions. It is noteworthy that our observations were made during a La Niña event, which resulted in unusually high rainfall; thus, the reported observations should be considered in this context. Whilst rainfall generally increases avian breeding success by optimising habitat requirements (Woolley et al. Reference Woolley, Hereford and Howard2022), prolonged and heavy rain (especially over a short period) can lead to flooding or cooler weather and wet nest conditions, which may affect egg temperatures (Ivey and Dugger Reference Ivey and Dugger2008). Maintaining egg temperature is crucial for species that share incubation duties between adults (including most crane species), as eggs are regularly exposed to the elements and predators during rotation (Du Rant et al. Reference Du Rant, Hopkins, Hepp and Walters2013; Ivey and Dugger Reference Ivey and Dugger2008). Our findings showed that the egg phase was more vulnerable to intense rainfall than the hatchling and fledgling stages. Indeed, we observed nest abandonment by Grey Crowned Cranes shortly after heavy rainfall and hail events (Demmer, personal observation), as has been reported for Sandhill Cranes Grus canadensis following rainfall events of >50 mm of rain within 24 hours (Fox et al. Reference Fox, Semel and Ward2019; Nesbitt Reference Nesbitt1988).
Future climate predictions indicate more extreme events such as droughts and high rainfall in alternating years (Connell et al. Reference Connell, Hall, Nimmo, Watson and Clarke2021; Marcelino et al. Reference Marcelino, Silva, Gameiro, Silva, Rego and Moreira2020) and so the ability to adapt to a changing environment may influence the survival of a species (Charmantier and Gienapp Reference Charmantier and Gienapp2014). Endangered bird species often exhibit reduced genetic variability, which can limit their ability to adapt to changing environmental conditions (Canteri et al. Reference Canteri, Fordham, Li, Hosner, Rahbek and Nogués-Bravo2021; Evans and Sheldon Reference Evans and Sheldon2008; Frankham Reference Frankham2005). This reduced adaptability may be particularly pronounced in non-migratory species like the Grey Crowned Crane, as they cannot easily relocate to more suitable habitats (Donnelly et al. Reference Donnelly, Moore, Casazza and Coons2022; Ramírez et al. Reference Ramírez, Rodríguez, Seoane, Figuerola and Bustamante2018). Despite higher rainfall over the past season, we found no difference in breeding outcomes when breeding began earlier or later in the 2022/3 breeding season. However, since these observations are based on only a single breeding season, we encourage future studies to examine whether Grey Crowned Cranes show any changes in their breeding strategy between years of extreme and non-extreme climatic conditions.
Several observations of crane species attempting to renest within the same breeding season in an attempt to replenish lost broods have been reported (Nesbitt Reference Nesbitt1988; Thompson et al. Reference Thompson, Gordon, Bolt, Lee and Szyszkoski2022; Woolley et al. Reference Woolley, Hereford and Howard2022). Renesting decisions depend on factors such as the time available to renest, environmental and parental condition, and the species’ overall breeding strategy (Saalfeld et al. Reference Saalfeld, Hill, Hunter, Frost and Lanctot2021). Whilst renesting can improve breeding success in some bird families (Hayes Reference Hayes2022; Ledwoń et al. Reference Ledwoń, Flis, Banach, Kusal, Łożyńska and Atamas2023), it may not always be beneficial, as the long-term fitness of adults can decrease when energetic costs outweigh reproductive benefits (Jones et al. Reference Jones, Heck, Lewis, Mungroo, Slade and Cade1995; Swift et al. Reference Swift, Anteau, Ring, Toy and Sherfy2020). Some renesting attempts by Grey Crowned Cranes in our study (n = 4; excluded from analyses in this study) were successful, but others renested towards the end of the breeding season which may have resulted in a resource mismatch during the early stages of offspring development (Martay et al. Reference Martay, Pearce‐Higgins, Harris and Gillings2023; Zelelew et al. Reference Zelelew, Bekele and Archibald2019). Our small renesting sample size precluded analyses of renesting impact on overall reproductive output, but this could be explored in future studies.
Nest-sites were evidently vulnerable to flooding during high rainfall periods, which is likely exacerbated in agricultural landscapes due to wetland degradation and destruction. To effectively address the threat of breeding failures caused by habitat loss and degradation, future conservation strategies and management interventions should focus on protecting habitats that are important to the entire life cycle of Grey Crowned Cranes. Restoring degraded wetlands for example, will not only increase biodiversity and food availability, but could also mitigate severe water level changes during periods of intense rainfall. Responsible land-use practices and wetland buffer zones are vital tools in wetland restoration, but they must be implemented in a way that supports and informs landowners to ensure their cooperation (Jepthas and Swanepoel Reference Jepthas and Swanepoel2019; Wood Reference Wood1999). Further protection from flooding may also take the form of artificial nesting platforms in suitably vegetated areas, since Grey Crowned Cranes readily nest in man-made water-bodies (Chuyu et al. Reference Chuyu, Fengqin, Xuesong, Qijiang and Yumin2022).
Human disturbances and anthropogenic landscapes
Cranes are generally sensitive to human disturbances (Coverdale Reference Coverdale2006; Ivey and Dugger Reference Ivey and Dugger2008; Végvári et al. Reference Végvári, Barta, Mustakallio and Székely2011; Wang et al. Reference Wang, Li, Beauchamp and Jiang2011), but in South Africa, cranes occur primarily outside of protected areas. While this increases their vulnerability to human pressures and unmanaged changes in their environment (Olupot et al. Reference Olupot, Mugabe and Plumptre2010), Hemminger et al. (Reference Hemminger, König, Månsson, Bellingrath‐Kimura and Nilsson2022) reported that several crane species, including the Grey Crowned Crane, benefit from readily available food supplies resulting from crop residues. An unfortunate drawback of this is that cranes may cause crop damage, with Grey Crowned Cranes observed affecting up to 4.1% of maize plants (van Niekerk Reference van Niekerk, Austin, Morrison and Harris2018). Losses caused by insects are comparably higher (10–16%) (Bradshaw et al. Reference Bradshaw, Leroy, Bellard, Roiz, Albert and Fournier2016; Oerke Reference Oerke2006; Raphela and Pillay Reference Raphela and Pillay2021). However, both crane and insect-related crop damage can be substantially reduced by planting seeds treated with insecticide, which reduces crane-related crop damage to just 0.23% (van Niekerk Reference van Niekerk, Austin, Morrison and Harris2018). The residual effects of consuming pesticide-treated seed in small amounts is largely unknown, and this, together with other studies on the impact of Grey Crowned Cranes on agricultural activities, should be explored further.
Several alternative, but highly effective, approaches exist to deter cranes from agricultural lands (Austin and Sundar Reference Austin, Sundar, Austin, Morrison and Harris2018). These methods include changing the timing of crop planting to avoid overlap with crane presence, selecting less preferred crops when cranes are present in high numbers, planting marginal lands as “diversionary fields” to distract cranes away from the principal croplands, providing supplementary feeding, and applying responsible seed treatments to deter birds (Austin and Sundar Reference Austin, Sundar, Austin, Morrison and Harris2018). Each of these methods can be applied regardless of the scale of agricultural activity being undertaken. Despite the benefits of increased foraging opportunities, intensive agricultural practices (e.g. grazing, cropping, and damming; Dalu et al. Reference Dalu, Wasserman and Dalu2017) drastically transform natural ecosystems, modifying sensitive breeding habitats (Bartzen et al. Reference Bartzen, Dufour, Clark and Caswell2010; Sica et al. Reference Sica, Gavier‐Pizarro, Pidgeon, Travaini, Bustamante and Radeloff2018). Nesting closer to human disturbances during vulnerable and energetically taxing breeding stages (e.g. incubation) may also lower breeding success (Seress et al. Reference Seress, Sándor, Evans and Liker2020). This is supported by observations of the migratory White-naped Cranes Grus vipio, which have shifted their nest-sites further away from roads and settlements (Wu et al. Reference Wu, Zou and Ma2014).
The physical nest-site location directly or indirectly influences breeding success or failure (Nalwanga et al. Reference Nalwanga, Lloyd, Du Plessis and Martin2004; Wu et al. Reference Wu, Zou and Ma2014). For example, nesting on islands is common in several waterbird species (e.g. Burgess and Hirons Reference Burgess and Hirons1992; Scarton and Valle Reference Scarton and Valle2020) and was also observed to promote positive breeding outcomes in our study – likely through a decreased risk of predation by terrestrial mesopredators. Grey Crowned Cranes commonly locate their nest-sites in moderate (50 cm) water depth amongst taller (60–90 cm) vegetation that is further away (100 m) from shore (Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Ndung’u2020). Most nest-sites in our study were surrounded by tall vegetation, likely to conceal nest-sites from terrestrial predators (Olupot Reference Olupot2016). Yet, our results indicated that vegetation height did not necessarily promote reproductive success. This may be explained by nest-sites still being visible to aerial predators despite tall vegetation. Indeed, we observed three instances when birds of prey were feeding on exposed eggs at nests (Figure S2). Considering the simplicity of earlier recommendations, we suggest that new dam sites consider incorporating these features into their planning to offset potential losses to pre-existing breeding habitats due to wetland flooding or degradation. Furthermore, our observations revealed that nearly all nests were situated in dense bulrushes (Typha capensis). Therefore, both new and existing dam sites should aim to encourage the establishment of this species to promote the suitability of nesting sites around the dam.
Although our study did not assess predator presence around nest-sites, predation via natural predators is an on-going challenge for ground-nesting birds that inhabit farmland landscapes (Bravo et al. Reference Bravo, Sarasa, Bretagnolle and Pays2023). We observed and received reports from landowners of some predatory incidences on Grey Crowned Cranes during this study (n = 6; Demmer, personal observation) by generalist mesopredators (i.e. serval Leptailurus serval and jackal Lupulella mesomelas). These types of predators are commonly associated with intensively managed pastures during calving and lambing seasons (Humphries et al. Reference Humphries, Ramesh, Hill and Downs2016; Skead Reference Skead1979) and are known predators of ground-nesting birds (Beja et al. Reference Beja, Schindler, Santana, Porto, Morgado and Moreira2014), and we presume that mesopredator density would be lower in natural grasslands and near buildings during this time. We therefore speculate that this could account for some of the observed positive associations between Grey Crowned Crane reproductive success and the closer proximity of their nesting to natural grasslands and buildings. Interestingly, our findings suggest that nesting closer to secondary roads and buildings (an indicator of human presence) is associated with increased hatching success in Grey Crowned Cranes, suggesting a tolerance to (if not a preference for) some human disturbances (Archibald et al. Reference Archibald, Meine, Garcia, del Hoyo, Elliott, Sargatal, Christie and de Juana2020; Wamiti et al. Reference Wamiti, Gichuki, Sigana, Njoroge, Morrison and Schröder2022). One potential explanation for this is that mesopredators typically avoid regions where humans are present (Reilly et al. Reference Reilly, Suraci, Smith, Wang and Wilmers2022) which, in turn, benefits nesting Grey Crowned Cranes. Alternatively, Fox et al. (Reference Fox, Semel and Ward2019) suggest that such relationships between birds and nesting location may be in response to decreased suitable breeding habitats, forcing cranes to choose nest-sites in suboptimal locations (e.g. Miller and Barzen Reference Miller and Barzen2016).
Finally, nest-sites situated in water-bodies with larger surface areas improved overall breeding success – perhaps because large, healthy water-bodies act as biodiversity hotspots, meeting a greater number of habitat requirements for significant life events of inhabiting species (Creed et al. Reference Creed, Lane, Serran, Alexander, Basu and Calhoun2017; Donnelly et al. Reference Donnelly, Moore, Casazza and Coons2022; Toussaint et al. Reference Toussaint, Brosse, Bueno, Pärtel, Tamme and Carmona2021). These larger water-bodies may also be able to buffer flooding instances better than smaller water-bodies, especially given the greater flood attenuation ability of highland wetlands (like those in our study) compared with lower altitude wetlands (Åhlén et al. Reference Åhlén, Thorslund, Hambäck, Destouni and Jarsjö2022; Wu et al. Reference Wu, Sun, Hu, Zhang and Rousseau2023).
Conclusions
Our study provided key insights into the breeding outcomes of Grey Crowned Cranes in unprotected areas and identifies macro-environmental factors that influence their reproductive success or failure. We observed significant attrition from egg to hatchling stages, which increased with higher rainfall intensity. Successful breeding outcomes were higher when Grey Crowned Cranes nested near buildings and natural grassland vegetation, suggesting some adaptability to human disturbances. However, the benefits of nesting near agricultural areas may be short-lived if wetland vegetation becomes degraded, reducing future nesting sites. We recommend refining monitoring methods and implementing more frequent, detailed, and long-term monitoring of Grey Crowned Crane breeding to gain a better understanding of this species’ long-term breeding success. Insights from long-term monitoring will undoubtedly assist with maintaining wetland ecosystem health and ensure continued breeding opportunities for this Endangered species.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0959270924000376.
Acknowledgements
Two anonymous reviewers provided constructive feedback on an earlier version of this manuscript. We thank the farmers and landowners who allowed this research to be conducted on their property; this study would not have been possible without your participation. Julie Braby, we appreciate the substantial contributions, time, and effort you gave to the initiation of this project as well as creating awareness of this species and the aims of this project in our surrounding communities. Last but certainly not least, we acknowledge the KZN Crane Foundation for their continued support and financial contributions to aspects of our fieldwork. This study was performed under the ethics clearance obtained from the University of South Africa’s Animal Research Ethics Committee (2022/CAES_AREC/173), and experiments were permitted by Ezemvelo KZN Wildlife (permit no. OP 3311/2022). All drone flights were conducted following the regulations of the South African Civil Aviation Authority.
Open access
