The study was conducted at Yongnianwa Wetland in Yongnian County, Hebei Province, China (36°40°60''-36°41°06"N, 114°41°15"-114°45°00"E, elevation 40.3 m). As a natural depression, the Yongnianwa Wetland belongs to the alluvial plain of the Fuyang River, which is a tributary of the Haihe River. It is located at the confluence of the Fuyang River Basin and the Zhang River Basin, where the water system is well developed and consists of many tributaries. The region has a typical warm-temperate, semi-humid, continental monsoon climate with four distinct seasons with both rain and heat during the same period. The main vegetation type of the wetland is reeds (Phragmites australis) mixed with cattail (Typha latifolia) and other herbaceous plants (for more details, see Ma et al. 2018a, 2018b).

Both ORW and blunt-winged warbler (Acrocephalus concinens) (BW) belong to the genus Acrocephalus, family Acrocephalidae, and order Passeriformes (Zheng et al., 2017). The ORW is a marsh-nesting species which depends on reedbed-dominated wetland habitats, with many published studies on its reproductive ecology (Dyrcz and Flinks, 2000; Dyrcz and Nagata, 2002; Wang et al., 2013; Yang et al., 2014; Zhao et al., 2014; Li et al., 2016). Wang et al. (2013) reported the breeding ecology of ORW in a Chinese population (Zhao et al., 2014), and Yang et al. (2014) discussed the differences between two sympatric cuckoo hosts breeding in China (Li et al., 2016). In comparison, the BW is widely distributed, but has a low abundance, and there are limited existing records on its reproductive ecology (del Hoyo et al., 2006). Both the reed parrotbill (Paradoxornis heudei) (RP) and vinous-throated parrotbill (Sinosuthora webbiana) (VP) belong to the family Sylviidae, order Passeriformes (Zheng, 2017). The former bird species is highly dependent on wetland reeds for various stages of its life history, such as breeding and overwintering (Ma, 1988; Xiong et al., 2007; Xiong and Lu, 2008; Boulord et al., 2011, 2012; Li et al., 2015, 2016; Ma et al., 2019). The latter species mainly nests and breeds in shrubs, and there are many studies on various aspects of its breeding behavior, such as parental provisioning behavior and egg-color dimorphism in South Korea, as well as nest-site choice and nestling growth in China (Kim et al., 1995; Guo et al., 2006; Lee et al., 2010).

The fieldwork was conducted during the bird breeding seasons from May to July in 2016 and 2017. We searched systematically for nests of all four species in the study area, and the locations of nests were marked using a GPS followed by taking measurements of a range of nest-related landscape, micro-habitat, and nest structure features. We revisited the nests regularly (every 2-3 days) to determine the ultimate fate of each nest. The fate was not defined as “success” unless at least one of the eggs laid resulted in fledging success (not a cuckoo chick). Moreover, the factors causing breeding failure, such as predation, parasitism, and poor weather conditions such as heavy rainstorms and wind, were recorded (Yang et al., 2011). The actual fates of the nests successfully parasitized by cuckoos, but whose nest eggs did not successfully hatch, were recorded as breeding success (Antonov et al., 2007).

The following nest-site parameters were measured and recorded after the birds laid their eggs, following Li et al. (2016): 1) Distance to road (m): the distance between the nest-site and the nearest road with human activity; 2) Distance to perch (m): distance from the nest-site to the nearest perch that a predator or parasitic cuckoo could use to observe the nest; 3) Distance to reed edge (m): the distance from a nest to the edge of the nearest reed; 4) Distance to water edge: the distance from the nest to the nearest boundary of the water body around the nest-site (m); 5) Vegetation cover: the degree of canopy cover 10 cm above the nest; 6) Nest height (m): the distance between the nest and the ground/water surface; 7) Reed height (m): with a nest as the center, the height of the nesting reeds in a 1m × 1 m square was measured and recorded to quantify the mean natural height of the reeds above the water surface; 8) Number of reed stems: the number of reeds within a 1 m × 1 m square; 9) Water depth (cm): the average water depth directly below the nest.

Two experiments were performed using an infrared camera (LTL Acorn 6210) to monitor the different types of predators. In the first experiment, 28 natural nests were monitored using video surveillance. In the second experiment, 29 empty nests with either successful breeding, predation or desertion were randomly selected and two quail eggs were placed inside each of the chosen nests. Due to an unexpected flood (Ma et al., 2019), this experiment aimed to investigate the presence and types of local predators; thus, predation rate was not determined.

The four species of breeding birds were similar in terms of nest-site choice, which showed a certain degree of differentiation. Based on the logistic regression results, only three variables—nest height, reed height, and number of reeds—were included in the final prediction model, with a significant difference in nest height existing among the four species (Chi-square = 46.795, df = 3, P < 0.01). The breeding nests of the ORW were positioned significantly higher than those of the BW, RP and VP (P₁ = 0.006, P₂ < 0.01, P₃ < 0.01). Reed height of nest-site was significantly different among the four breeding birds (Chi-square = 13.635, df = 3, P = 0.003), with the reed height of the ORW being significantly greater than those of the BW, RP and VP (P₁ = 0.012, P₂ = 0.001, P₃ < 0.001). The reed height of the BW was significantly greater than those of the RP and VP (P₁ = 0.012, P₂ = 0.003). Number of reeds (defined as the number of reeds in a 1 m × 1 m square centered around the nest) was also significantly different among the four species (Chi-square = 8.045, df = 3, P = 0.045). Number of reeds of the ORW was significantly lower than those of the BW and RP (P₁ = 0.003, P₂ = 0.004) (Table 1).

In addition, the timing of breeding of the four species differed. Notably, the two species of parrotbill are local residents and started breeding earlier in the season (the first clutches were laid in 2017 by RP: April 28 and VP: May 3), while the two species of warbler are summer visitors and started breeding later in the season (ORW: May 15 and BW: May 29). The starting time of reproduction for ORW was from May 20 to June 20, during which 84% (115/137) of the birds had laid the first clutch; and for RP it was from April 28 to May 17 and June 6 to June 18, during which 86% (24/28) of the birds had laid the first clutch. In addition, 43% (12/28) of the RP had laid the first egg before the starting time of reproduction of ORW.

For all four species, the breeding success rate was about 35%-41%, and the rates for each species were 34.5% (86/249) for ORW, 35.3% (6/17) for BW, 38.5% (15/39) for RP, and 40.9% (9/22) for VP in the two years. For all four species, the main cause of breeding failure was predation, corresponding to 36.9% (92/249) for ORW, 41.2% (7/17) for BW, 35.9% (14/39) for RP and 54.5% (12/22) for VP, and the secondary cause of breeding failure was weather. For the ORW, nest parasitism by the common cuckoo was a key factor causing breeding failure, with a 2-year parasitism rate of 11.6% (n = 249) (Table 2).

Regardless of whether the first or second year was assessed, there was no significant difference in the overall breeding success rate among the four species (Fisher’s exact test: P₁ = 0.432, P₂ = 0.426, P₁₂ = 0.893); however, there was a significant difference in the overall breeding success rate of the four species between the two years (Fisher’s exact test: P = 0.038). There was no significant difference in predation rate between the two years (Fisher’s exact test: P = 0.139), but the weather-caused breeding failure rate in the first year was significantly higher than that in the second year (Fisher’s exact test: P < 0.01).

The main nest predators documented through videos from the ground were snakes, such as the red-banded snake (Dinodon rufozonatum) and twin-spotted ratsnake (Elaphe bimaculata), and mammals, like the yellow weasel (Mustela sibirica) (Fig. 3). In addition, some bird species, such as the magpie (Pica pica) and rufous-backed shrike (Lanius schach), were observed to prey on chicks more than once.

This study demonstrated that the four sympatric bird species had similar breeding site and nest structures in the wetland, but that certain characteristics (nest height, reed height, and reed density) differed along with the timing of the onset of breeding. All four species bred in the reed habitat portion of the wetland, mostly nesting on the middle and lower parts of the reeds. Nests were well-concealed from above, while the water surface below the nests potentially limited access to predators, such as rodents. Moreover, the nest-sites of the four species of birds were far from roads and located deep in the reeds where anthropogenic disturbance was minimal.

The ORW and RP mainly relied on wetland reeds for breeding, which is consistent with previous studies (Dyrcz and Flinks, 2000; Dyrcz and Nagata, 2002; Boulord et al., 2011, 2012). Birds that coexist in the same environment inevitably have different needs for food resources, nest-site choices, and breeding phenology. Thus, they might successfully coexist by partitioning the limited resources and reducing interspecific competition in the same region through niche differentiation (Macarthur, 1958; Martin and Thibault, 1996; Tokeshi, 1999; Sommer and Worm, 2002; Atiénzar et al., 2013). The results of the present study demonstrated that the nest sites of the four species exhibited a mixed distribution; however, a certain degree of differentiation was detected in nest-site choices in terms of microhabitat and reproductive phenology. First, the four species used the reeds differently. The population of the ORW was dominant in the region. This species used higher reeds for nesting than the other three species, which occupied comparatively low reeds. Second, there was a clear vertical stratification in nest height. The ORW placed nests higher than the other three species, with nests of the VP being positioned the lowest in terms of both nest height and reed height. Previous studies on great reed warbler, an allied species of the ORW, showed that it also nested on the highest reeds (3.29 ± 0.551 m), while the sympatric reed warbler (Acrocephalus scirpaceus) used shorter reeds (2.53 ± 0.385 m) for nesting (Honza et al., 1999; Prokešová and Kocian, 2004). Furthermore, in their own territory, the former species expelled the latter species, and even destroyed their nests (Honza et al., 1999; Prokešová and Kocian, 2004). A similar situation was observed in our study area, in which one species expelled another invader out of the breeding territory due to competition. In addition, the four species of birds in this study initiated reproduction at different times. For instance, the two parrotbill species started breeding earlier than the two warbler species, supporting the findings of Li et al. (2016), which showed that the RP started breeding earlier than the ORW to avoid cuckoo parasitism. The observed differences in the timing of breeding in the current study were more likely due to the fact that the four species of birds exhibit different types of residency in the region instead of being due to encountering cuckoos. For instance, the two parrotbill species are year-round residents, while the two warbler species are summer visitors (Zheng, 2017). Earlier breeding for the two parrotbill species might avoid the breeding peak and reduce the overlap of the breeding period with the remaining two species, thus alleviating the interspecific competition for resources. A similar dynamic has been demonstrated in previous studies on sympatric breeding birds, which mitigate interspecific competition and coexist through differences in reproductive phenology (Macarthur, 1958; Prokešová and Kocian L, 2004; Atiénzar et al., 2013). In addition, the shape of beak of the two parrotbill species allowed them to obtain food through preying on insect larvae hidden in the dry reeds (Xiong et al., 2007).

The relationship between nest-site choice and breeding success is important for understanding whether the observed patterns influence individual fitness (Clark and Shutler, 1999). In the present study, the ORW had the largest breeding population and the BW had the smallest population; however, the four species had similar breeding success rates. The main factor causing breeding failure of the four species was nest predation, followed by poor weather conditions such as flooding. For the ORW, parasitism of nests by the common cuckoo was one of the important factors causing breeding failure. Nest predation is a major factor influencing the breeding failure of birds in general (Martin, 1995). In particular, nest height could be directly related to breeding success for birds, as higher nests are more vulnerable to threats from aerial predators, while lower nests are more likely to be attacked by ground predators (Piper and Catterall, 2004; Mainwaring et al., 2014; Tabib et al., 2016). The ground predators in our study area were mainly snakes and yellow weasels, and as both of these groups have a certain ability to swim and climb, they present a greater threat to lower nests than to higher nests. Moreover, there were large populations of magpies and rufous-backed shrikes in the study area, and both species were observed holding chicks in their mouths (L.M., personal observations). There was a significant difference in nest height among the four species of birds; however, there was no significant difference in predation rate, which might be attributed to the high level of concealment of the nests, on average, by each of these species. The nests of the four species were constructed on the middle and lower parts of the reeds, making nests well concealed from above. The nests placed in the higher reed by the ORW might reduce the threat from ground predators. However, a higher nest is more vulnerable to aerial predators and potential parasites such as cuckoo parasitism. In comparison, the other three species with lower nests used more reeds to increase the concealment of their nests, which reduced the probability of their being discovered by ground predators. Furthermore, the water environment below the nests reduced the threat from ground predators. However, the intensity levels of predation by ground predators and aerial predators in the area require further investigation.

There was a significant difference in the rate of breeding success of the four bird species between the two study years. In the first year (2016), the study area was subjected to a heavy rainfall event, which resulted in the breeding failure of all nests and consequently reduced the overall breeding success when the event occurred (Ma et al., 2018). Extreme weather is rare and is difficult to predict, but it exerts a devastating impact on the survival of birds, especially reproduction, as documented in previous studies (Dugger et al., 2002; Altwegg et al., 2006; Jenouvrier, 2013). Geographical variation exists in nest-site choice among populations of the same species of birds which inhabit different geographical areas. For example, Dyrcz and Nagata (2002) studied the Japanese population of the Oriental reed warbler, which also nests in the reeds of wetlands, and found that most individuals in the population nested at higher positions in the reeds (1.25 m). This difference was attributed to two main factors. First, the predators in the breeding area for the population of the Oriental reed warbler in Japan are mainly snakes and yellow weasels; thus, predation pressure is mainly from the ground, so nesting at higher positions represents an adaptive strategy for increasing survival from ground predators (Vanderwerf, 2012). Second, summer rainfall in the breeding area for the population in Japan is high, causing large water-level fluctuations. The region studied here was located in an inland area with low rainfall and relatively small water-level fluctuations often accompanied by windy weather, leading to a low risk of nest submergence; thus, lower nests are more beneficial for mitigating the risk of strong winds that can cause nest tipping.