Intraspecific variation in microhabitat selection in reintroduced Chinese giant salamanders

Abstract

Reintroduction of captive-bred Chinese giant salamanders is a primary approach for restoring wild populations. Despite previous studies have investigated the habitat preferences of reintroduced Chinese giant salamanders, the intraspecific variation in their habitat selection has been neglected. In the present study, 30 captive-bred Chinese giant salamanders belonging to 3 groups (i.e., 10 males, 10 females, and 10 juveniles) were released into a montane stream to explore whether intraspecific variation in habitat selection occurred in this species using radiotelemetry. Our results indicated that linear home range and daily movement of males were significantly higher than those of females and juveniles. Male sedentariness was significantly lower than that of females and juveniles. No significant differences were detected between females and juveniles in these measures. Importantly, we found that males preferred microhabitats with low water conductivity and deep water depth. Females preferred microhabitats with high water conductivity, low dissolved oxygen and ammonium-nitrogen, and slow current velocity, while juveniles occupied microhabitats with low ammonium-nitrogen. In addition, males and juveniles exhibited higher niche breadth than females. Niche overlap was high between adults and juveniles but low between males and females. Our study revealed the presence of spatial segregation in reintroduced Chinese giant salamanders. Males, females, and juveniles exhibited variation in microhabitat selection. These results provide important information for use when planning strategies for conservation of Chinese giant salamanders.

Habitat loss and degradation are considered to be important drivers of biological diversity loss (Brook et al. 2008). Therefore, understanding habitat use and requirements can aid in the planning of conservation strategies (Johnson and Derocher 2020). Indeed, spatial segregation and differences in habitat selection occur widely in wild populations (e.g., between sexes or age groups; Salvidio 2002; Ficetola et al. 2013). This phenomenon may be due to the distinct requirements of different groups within species (Kearney et al. 2007; Ficetola et al. 2013). For instance, adult and juvenile European cave salamanders Hydromantes (Speleomantes) strinatii have been found to exhibit different habitat preferences that can be attributed to distinct risk-taking strategies. Specifically, the juveniles tended to live near the cave entrance where invertebrates were abundant. This is because juveniles need sufficient food for growth and development. The adults preferred to live deep in the cave, as their survival is the most important contribution to the population growth rate (Ficetola et al. 2013). Similar spatial segregation patterns can also be detected between sexes, such as the tendency for male Australian freshwater crocodiles Crocodylus johnstoni that have been reintroduced to move further than females to claim territory and gain breeding access to females. The decreased movement of females has been found to be associated with the need to attain sufficient energetic stores for reproduction (Tucker et al. 1997). Although habitat selection has been recognized to vary among demographic groups in many species, little attention has been devoted to Chinese giant salamander Andrias davidianus.

The Chinese giant salamander is the largest extant amphibian species in the world (Zhao 1998) and has lived on Earth for millions of years (Fei et al. 2006). This species is a top predator in freshwater ecosystems (mainly feeding on fish, frogs, small mammals, water birds, crabs, and insects; Song 1994) and thus exerts strong top-down effects on ecosystem functioning (Zhang et al. 2017). Historically, Chinese giant salamanders were broadly distributed in the tributaries of 3 major river systems (i.e., the Yellow River, Yangtze River, and Pearl River) in central, eastern, and southern China (Wang et al. 2004; Fei et al. 2006; Zhao et al. 2020), covering 18 provinces (Fei et al. 2006). However, wild populations of Chinese giant salamander have declined sharply since the 1950s (Wang et al. 2004), mainly due to the over exploitation, water pollution, and habitat degradation (Turvey et al. 2018; Jiang et al. 2021). As a consequence, Chinese giant salamanders have rarely been observed in nature in recent decades (Turvey et al. 2018). Therefore, this species is listed as critically endangered in China (Jiang et al. 2016) and by the IUCN Red List (IUCN 2021). This species has also been regarded as a protected species by the Chinese government (grade II) since 1988.

The Chinese government and international organizations have expended much effort on the conservation of Chinese giant salamanders over the last 50 years (Zhao et al. 2017). Specifically, 47 natural reserves have been created to protect this species in China since 1980, covering approximately 2.5% of the total area of China (Zhao et al. 2017). Another primary protective measure for the conservation of Chinese giant salamanders is the introduction of captive-bred individuals into natural environments (Aquatic Wildlife Conservation Association 2015). By the end of 2019, over 287,840 captive-bred individuals were released back into the wild across 16 provinces and 98 counties in China (Shu et al. 2021). Studies were then conducted to quantify the spatial distribution and seasonal movement patterns of reintroduced Chinese giant salamanders (Zhang et al. 2019), as well as their habitat selection (Zhang et al. 2017). These studies found that captive-bred Chinese giant salamanders exhibited relatively small home ranges, short daily movement distances, and high sedentariness at their release sites when compared with observations from studies on hellbenders (e.g., eastern hellbender, Burgmeier et al. 2011 and Ozark hellbender, Bodinof et al. 2012). Moreover, large boulders were more important for the settlement of reintroduced Chinese giant salamanders, followed by water depth and canopy cover. Although these studies provided the first evidence of habitat preference of reintroduced individuals, the differences in habitat selection between sexes and between juveniles and adults of this species have been neglected.

The main objective of the present study was to assess whether intraspecific variation in habitat selection occurs in this species using radiotelemetry approaches. Specifically, we 1) tested the spatial segregation of the 3 groups (i.e., juveniles, males, and females); 2) identified the microhabitat preferences of the 3 groups; and 3) calculated the habitat niche widths of the 3 groups, as well as the habitat niche overlap between them. Based on previous studies (e.g., Tucker et al. 1997; Ficetola et al. 2013; Lunghi et al. 2015), we predicted that spatial segregation would occur between groups, and we also predicted that intraspecific variation in habitat selection would exist in reintroduced Chinese giant salamanders, with different age and sex classes exhibiting specific microhabitat preferences.

Material and Methods

Study area

Field work was conducted in a montane stream located in Dujiangyan, Sichuan Province, China (103°38ʹE, 31°02ʹN, Figure 1). Following the recommendations of Lunghi et al. (2019), the exact coordinates are not provided because the endangered Chinese giant salamander is still collected as food by local people. This area is characterized by a subtropical climate, with an average annual temperature of 15.2°C and annual precipitation of 1244 mm (Yu et al. 2004). The elevation of the study area was approximately 950 m and the main vegetation cover along the streambank consisted of deciduous broadleaf forest and bamboo forest. The mean width and the mean depth of the stream were 5 m and 40 cm, respectively. Deep pools and large quantities of food resources (e.g., fish, crabs, tadpoles, and shrimp) can be observed in this stream, providing potential suitable habitats for reintroduced Chinese giant salamanders.

Figure 1.

(A) Geographical location of the study area. (B) Photos of the study stream. The source of the map was from https://www.webmap.cn.

Study animals

A total of 30 captive-bred Chinese giant salamanders belonging to different life stages (juveniles/adults) and sexes were obtained from a hatchery located in Hongya, Sichuan Province, China (103°10ʹ05ʹʹE, 29°52ʹ36ʹʹN). Specifically, these individuals included 10 males (4 years old), 10 females (4 years old), and 10 juveniles (2 years old). Since the 4-year-old individuals could not be identified to sex based on external morphology during the nonbreeding season, we followed Su (2018) and determined their sex by using Doppler B-ultrasound equipment (Accuvix V10, Samsung, South Korea) in West China Hospital of Sichuan University. The total length of each individual was measured to the nearest 0.1 cm by a tape measure and the body weight was measured to the nearest 0.01 kg with a scale before release. After measurements were completed, a radio transmitter (PIP41, Lotek UK Ltd., UK) was attached to the back of each individual by a piece of cotton thread (Supplementary Figure S1). The transmitter weighed approximately 0.62 g, accounting for 0.03−0.62% of the body weight of the salamanders used in the present study. More importantly, this approach is noninvasive when compared with the previous surgical implantation approach (e.g., Marcec et al. 2016; Zhang et al. 2016). Therefore, we believe that this noninvasive approach has less effect on Chinese giant salamanders than does traditional surgical implantation (Bodinof et al. 2012; Borzée et al. 2018). Finally, these individuals were released at 6 sites (i.e., 2 sites each at upstream, midstream, and downstream locations; Supplementary Figure S2) in the stream on 22 May 2020. These sites were selected based on accessibility and diverse microhabitat features (e.g., deep pools, shallow water, and fast-flowing segments with large boulders). At each site, we released 5 individuals, including several males, females, and juveniles (Supplementary Figure S1). Individuals could freely move from one site to the others.

Data collection

After being released, each individual was separately monitored and located daily (between 13:00 and 18:00) using a radio receiver (SIKARX8, Lotek UK Ltd., UK). The rationale for this timing of observations is that the Chinese giant salamander is a nocturnal predator and we assume that the daytime GPS locations primarily reflected individuals’ preferred locations (Lunghi et al. 2015, 2018). We continued tracking them for 32 days (until 23 June 2020), during their period of peak activity, and until the transmitter batteries were almost exhausted. We chose this period to perform the experiment, as Chinese giant salamanders are more active during this time (Liu 2020). Importantly, the study was conducted at the end of the dry season, thus avoiding the effects of the flooding. We measured 11 potential microhabitat factors that may affect the distribution of Chinese giant salamanders at their final settlement sites by the end of the experiment (i.e., 23 June 2020). Specifically, water temperature, water conductivity, and dissolved oxygen (DO) were measured using a portable fluorescence photometer (Orion, Thermo Fisher Scientific, USA). Water width and water depth were measured using a tape measure. Current velocity was measured using a portable current meter (LS1206B, China). Sediments were classified into 3 categories based on particle size (i.e., silt: <1 cm; gravel: 1–20 cm; and boulder: >20 cm). The chlorophyll a concentration was measured using a portable fluorescence photometer (Manta + 35, USA). Water surface area was estimated by the same person. For ammonium-nitrogen (NH₄-N), water samples were collected using new bottles in the field (bottles were washed using filtered local water before the sampling), which were brought back to the laboratory immediately for analyses. In addition, we also recorded the potential food resources (e.g., fish, crabs, and shrimp; Song 1994) that were caught in traps every day and food richness was represented using the Margalef index (Death 2008).

Statistical analyses

Ovitalmap (www.ovital.cn) was used to calculate linear home range, which is the distance between the initial and the final position of each individual (Burgmeier et al. 2011). This index reflects the length of the stream utilized by salamanders and has been widely used for species living in linear and fairly homogenous environments (e.g., small streams and rivers; Burgmeier et al. 2011). Moreover, we defined consecutive locations ≤4 m in distance as stationary movements. Sedentariness was calculated as the ratio of the number of days with stationary movements to the total number of monitoring days (Zhang et al. 2019). Individual daily movement distance was represented by the mean length of the straight line between the starting point and the ending point per day (Ayala et al. 2019). The movement distance was not recorded when it is less than 4 m, as this could reflect a regular movement in the same pool. One female died 3 days after release and 1 male was captured by native people 16 days after release. Therefore, 29 individuals were used to calculate LHR, sedentariness, and daily movements (10 males, 9 females, and 10 juveniles) and 28 individuals were used to investigate habitat selection, niche width, and niche overlap (9 males, 9 females, and 10 juveniles).

Since the values of LHR, sedentariness, and daily movements were not normally distributed based on the Shapiro–Wilk tests, differences between groups were tested using Kruskal–Wallis rank sum tests (Kruskal and Wallis 1952; Conover and Iman 1979), and the Conover test was applied for post hoc pairwise tests (Conover 1999). We then performed Pearson correlation analyses to detect the pairwise correlations between 11 microhabitat factors (Pearson 1920). Based on the results, water width and water surface area were highly correlated (|r| ≥ 0.7; Quinn and Keough 2002) and we retained water surface area for further analyses, as this measure is more important for the distribution of aquatic animals (Wang et al. 2007). Then, generalized linear models (GLMs) were constructed to determine the potential relationships between the distributions of different age and sex classes of released Chinese giant salamanders and the recorded microhabitat variables. In the models, presence/absence was the dependent variable and all microhabitat variables were the independent variables. All the candidate models for each age or sex class were ranked and the best model was selected based on the minimum corrected AIC value because of the small sample size (AICc; Burnham and Anderson 2002; Burnham et al. 2011). In addition, we used hierarchical partitioning to calculate the relative contributions of different microhabitat variables to the microhabitat selection of each group (Chevan and Sutherland 1991; Nally 2002).

Finally, the microhabitat niche width of each group was quantified using the Levins index (B) according to the formula:

$${\displaystyle \begin{array}{l}{\displaystyle B=\frac{1}{{{\sum }_{j=1}^R\left(P_j\right)}^2},}\end{array}}$$

where $P_j$ denotes the proportion of individuals found at sampling point j and R is the total number of sampling points (Simpson 1949). The microhabitat niche overlap between groups was quantified using the Levins index (O_ik) based on the formula:

$${\displaystyle \begin{array}{l}{\displaystyle O_{ik}=\frac{{\sum }_{j=1}^R{P}_{ij}{P}_{kj}}{{{\sum }_{j=1}^R\left(P_{ij}\right)}^2},}\end{array}}$$

where P_ij and P_kj are the abundance of group i and group k at sampling point j, respectively. R is the total number of sampling points (Levins 1968).

All statistical analyses were conducted in R v4.0.3 (R Development Core Team 2020). The Shapiro–Wilk tests were performed using the stats package (R Development Core Team 2020). Pearson correlation analyses were undertaken using the psych package (Revelle 2020). Kruskal–Wallis rank sum tests with post hoc Conver tests were performed using the PMCMR package (Pohlert 2014). GLMs were constructed using the MuMIn package (Burnham and Anderson 2002). Hierarchical partitioning was performed using the vegan (Oksanen et al. 2020) and hier.part packages (Nally and Walsh 2004). Niche width and niche overlap were calculated based on the spaa package (Zhang 2016).

Results

Before release, the mean total lengths of males, females, and juveniles were 63.5 cm ± 1.170 (SE), 59.9 cm ± 1.123 (SE), and 29.5 cm ± 0.630 (SE), respectively. The mean body weights of males, females, and juveniles were 1.63 kg ± 0.052 (SE), 1.30 kg ± 0.061 (SE), and 0.14 kg ± 0.008 (SE), respectively.

LHR, sedentariness, and daily movement

In total, 7 males, 2 females, and 3 juveniles moved further than 4 m during the monitoring and the remaining individuals only exhibited regular movement (i.e., < 4 m) at the location where they were released. During the experiment, each individual spent a mean of 1.28 ± 0.46 (SE) days (the dates between the initial movement and the final movement) selecting the preferred habitat and settling down. The mean LHR values were 166.75 m ± 63.57 (SE), 6.33 m ± 4.35 (SE), and 9.15 m ± 4.53 (SE) for males, females, and juveniles, respectively (Supplementary Figure S3). Kruskal–Wallis rank sum tests demonstrated that LHR sizes differed significantly among groups (P = 0.03). Specifically, the LHR of males was significantly higher than those of females and juveniles. However, there was no significant difference between female LHR and juvenile LHR (Table 1 and Figure 2A).

Table 1.

The results of Conover tests between Chinese giant salamander groups

Spatial movement indices	Groups	T value	P value
Daily movement	Juvenile-female	0.327	0.373
	Juvenile-male	0−2.282	0.016*
	Female-male	0−2.543	0.009**
LHR	Juvenile-female	0.344	0.367
	Juvenile-male	0−2.408	0.012*
	Female-male	0−2.687	0.006**
Sedentariness	Juvenile-female	0−0.167	0.435
	Juvenile -male	2.269	0.016*
	Female-male	2.374	0.013*

^* P < 0.05,

^** P < 0.01.

Figure 2.

The linear home range (A), daily movement (B), and sedentariness (C) of 3 groups of released Chinese giant salamanders. Different letters on top of the error bars indicate a significant difference between groups (P < 0.05).

The mean sedentariness was 0.94 ± 0.022 (SE), 0.99 ± 0.004 (SE), and 0.99 ± 0.003 (SE) for males, females, and juveniles, respectively. Kruskal–Wallis rank sum tests demonstrated that sedentariness differed significantly among groups (P < 0.05). Specifically, female and juvenile sedentariness were significantly higher than that of the males. However, there was no significant difference between juvenile sedentariness and female sedentariness (Table 1 and Figure 2B).

The mean daily movements were 3.44 m ± 1.30 (SE), 0.12 m ± 0.08 (SE), and 0.17 m ± 0.09 (SE) for males, females, and juveniles, respectively. Kruskal–Wallis rank sum tests demonstrated that daily movement differed significantly among groups (P = 0.04). Specifically, males’ daily movements were significantly further than those of females and juveniles. However, there was no significant difference between female daily movement and juvenile daily movement (Table 1 and Figure 2C).

Preference for microhabitat features

In the optimal GLM, microhabitat variables significantly impacted the spatial distribution of reintroduced Chinese giant salamanders of different life stages and sexes. Specifically, male distribution was negatively correlated with water conductivity but positively correlated with water depth and ammonium-nitrogen. Female distribution was negatively correlated with DO, current velocity, and ammonium-nitrogen and positively correlated with water conductivity. Moreover, juvenile distribution was negatively correlated with ammonium-nitrogen. Regarding the independent contribution of each selected microhabitat factor to the distributions of different groups, hierarchical partitioning analyses indicated that water depth and water conductivity contributed the most to the distribution of males (71.1% and 23.2%, respectively, Figure 3A). Water conductivity was the most important contributor to the habitat preference of females (40.9%), followed by DO (24.7%), current velocity (20%), and ammonium-nitrogen (14.4%, Figure 3B). The contribution of ammonium-nitrogen to juvenile distribution was 77.9% (Figure 3C).

Figure 3.

Relative contribution of different microhabitat factors to the microhabitat selection of 3 released Chinese giant salamander groups (A: Males, B: Females, and C: Juveniles). WC, water conductivity; DO: Dissolved oxygen; DE, water depth; CV, current velocity; SE, sediment; NH4_N, Ammonia nitrogen; CH, chlorophyll α; FR, food richness.

Niche breadth and niche overlap

Habitat niche width (i.e., Levins index B) for males, females, and juveniles was 7.14, 3.86, and 7.36, respectively. In addition, habitat niche overlap (i.e., Levins index O_ik) was 8.6% between males and females, 57.3% between males and juveniles, and 81.8% between females and juveniles.

Discussion

This study examined the intraspecific variation in microhabitat selection of reintroduced Chinese giant salamanders in a montane stream. Our results indicated that spatial segregation occurred across sex and ontogenetic stage. Specifically, the LHR and daily movement of males were significantly greater than those of females and juveniles. Males were significantly less sedentary than females and juveniles. All of these results demonstrated that the reintroduced males exhibited more movement than females and juveniles. This observation contrasts with findings for some other reintroduced species of Caudata, such as Italian crested newts Triturus carnifex in Ameisensee Lake, Austria, and marbled newts Triturus marmoratus in southern France, in which females migrate farther than males (Schabetsberger et al. 2004; Trochet et al. 2017). Therefore, our results suggest that male Chinese giant salamanders follow specific strategies to benefit their living, foraging, and mating in streams. More importantly, the contrast results between Chinese giant salamander and other tailed species suggested that the conservation of Chinese giant salamander could not simply follow the experiences of other species, vice versa. Interestingly, movement distances of reintroduced Chinese giant salamanders were generally short overall during our monitoring activities (LHR for males, females, and juveniles was 166.75 m ± 63.57 [SE], 6.33 m ± 4.35 [SE], and 9.15 m ± 4.53 [SE], respectively), with sedentariness exceeding 90% for all individuals. Our results contrast with those of a previous study from May 2013 to August 2014 showing that the LHR of reintroduced Chinese giant salamanders was as long as 1730 m in the Donghe River, central China, and sedentariness was only approximately 26% (Zhang et al. 2019). This is because summer and early autumn form the rainy season in central China and the reintroduced Chinese giant salamander individuals could be transported far from the release sites by floods (Zhang et al. 2019). Indeed, salamanders are poor dispersers displaying strong site fidelity (Mayasich et al. 2003; Schulte et al. 2007). Our results thus support this claim, showing that the reintroduced Chinese giant salamanders can only select preferred microhabitats within limited areas near release sites. Since individuals may exhibit distinct movement performance in habitats with different features (Ayala et al. 2019), further studies will be focused on the spatial segregation of individuals within a group of males, females, or juveniles.

Similar to patterns in some other salamanders (e.g., Hydromantes italicus, Lunghi et al. 2015), the preferred microhabitat was determined by distinct factors for different groups of reintroduced Chinese giant salamanders. Specifically, males preferred the environments with deep water. This may be attributed to the large numbers of caves in deep water areas (Tao et al. 2004), which are important breeding sites for males to attract females (Fei et al. 2006; Wang 2015). Females were usually found in microhabitats with high water conductivity, but low current velocity. Previous studies have demonstrated that this kind of water environment can benefit the growth of aquatic plants (Bogotá-Gregory et al. 2020) and can thus maintain a high abundance of aquatic animals (e.g., tadpoles, shrimp, and fish; Sun et al. 2021). Therefore, females can acquire more food resources to reproduce and spawn in such a location. In addition, juveniles preferred microhabitats with low ammonium-nitrogen (<0.4 mg/L in the present study), probably because their potential foods (e.g., shrimp) were more vulnerable in water bodies with high ammonium-nitrogen (>0.4 mg/L; Jia et al. 2017). These results indicated that microhabitat assessment should be conducted before the release of Chinese giant salamanders, and distinct microhabitats preferred by salamanders of different sexes and developmental stages need to be included in the same stream to allow their survival.

We also found that niche breadth values were high for males and juveniles, indicating that individuals from these two groups occupied more diverse habitats in this stream. Since habitat diversity is usually positively related to the diversity of available resources, individuals from these two groups may potentially utilize more diverse food items (Araújo et al. 2011). Future studies could verify our inference using dietary analyses or stable isotope analyses. Females showed lower niche breadth values, demonstrating that they were concentrated within several pools. These results suggest that female salamanders could be considered as microhabitat specialists, displaying stricter habitat requirements in the study stream. This is not surprising, as a female salamander usually needs a specific environment in which to forage, nest, and breed (Wu 2009). More importantly, habitat niche overlap was low between males and females, which could reduce the competition (for habitat and food) between them. In contrast, habitat niche overlap was high between juveniles and males/females, indicating that juveniles and adults can coexist in the wild. This could be attributed to the ontogenetic trophic niche shift, which is widely observed in salamanders (e.g., Ambystoma laterale, Schriever and Williams 2013). Overall, our study revealed the presence of intraspecific variation in the spatial movement patterns of reintroduced Chinese giant salamanders, which was determined by differing preferences for microhabitat features by males, females, and juveniles. This phenomenon have also been observed in fish (e.g., Carcharhinus melanopterus, Schlaff et al. 2020) and reptiles (e.g., Platysternon megacephalum, Shen et al. 2010), so intraspecific variation in microhabitat utilization may exist in all the cold-blood vertebrates.

Supplementary Material

Supplementary material can be found at https://academic.oup.com/cz.

Funding

This work was supported by the National Key Programme of Research and Development, Ministry of Science and Technology (2016YFC0503200); the Biodiversity Survey and Assessment Project of the Ministry of Ecology and Environment, China (2019HJ2096001006); the Construction of Basic Conditions Platform of Sichuan Science and Technology Department (2019JDPT0020); and China Biodiversity Observation Networks (Sino BON).

Authors’ Contributions

T.Z., J.J., and B.Y. conceived the original idea and procured funding. C.Z., J.F., Z.S., and W.F. conducted the field work and collected the data. C.Z. and T.Z. analyzed the data and wrote the manuscript. W.Z. and J.C. commented the manuscript. All authors read and approved the final manuscript.

Conflict of Interest Statement

The authors declare that there is no conflict of interest.

Ethical Approval Statement

Animal procedures were approved by the Animal Care and Use Committee of Chengdu Institute of Biology.