The impact of earlier flood recession on metacommunity diversity of wintering waterbirds at shallow lakes in the middle and lower Yangtze River floodplain

1.   Introduction

Shorebird populations are declining worldwide, mainly due to human disturbances and degradation of coastal habitats and loss of coastal wetlands (Stroud et al., 2006; Birdlife International, 2022). However, supratidal habitats (natural or artificial) could play a role in buffering human impact, and some studies have proven the importance of anthropogenic supratidal habitats in maintaining shorebird populations (Masero et al., 2000; Sanchez-Guzman et al., 2007). These habitats include coastal saltpans (saltworks, salt ponds or salinas), anthropogenic supratidal areas where salt is extracted from seawater through solar evaporation in shallow ponds separated by dykes. Saltpans have shown to be important as feeding or breeding sites of some shorebird species, including the Kentish Plover (Charadrius alexandrinus), the Little Tern (Sternula albifrons), or the Avocet (Recurvirostra avosetta) (Masero, 2003; Chokri and Selmi, 2011; El Malki et al., 2018). Although most of the world's saltpans were abandoned in the middle of the last century, restoration of saltpans may be crucial for the maintenance of some species, especially considering the potential decline or and even disappearance of many saltmarshes and beaches due to the effects of the global change scenario (Galbraith et al., 2002; Horton et al., 2018). Nonetheless, the restoration of abandoned saltpans for bird sustainability requires a thorough understanding of the species-specific habitat conditions shorebirds need to feed and breed to make salt extractions compatible with bird conservation. This would also be true for saltpans that are still operational. This understanding would help us to better manage the saltpans for enhancing bird populations.

A potential conservation strategy to increase shorebird populations in saltpans is to manipulate the cues that birds use to select optimal breeding habitat. Here it is hypothesized that shorebirds are attracted to bivalve shells due to the advantages they offer improving hatching success by providing increased nest protection against severe weather conditions like flooding, enhance camouflage and/or create improved thermal conditions for egg development (Hansell, 2000; Rounds et al., 2004; Colwell et al., 2011).

In this study, we identified an area in a restored saltpan and supplemented it with marine shells, expecting an increase in the number of breeding bird nests, and hypothesising that if breeding birds selected the area with shell supplementation, the expected result would be an increase in hatching success.

Moreover, it is likely that coastal ground-nesting birds also face the risk of having their nests flooded by water during the high tide (Winton et al., 2000; Anteau et al., 2012; Bailey et al., 2017). This also occurs in saltpans where some of the ponds must be flooded for salt extraction purposes. In this case, birds must choose between building their nests close to food (and consequently in close proximity to water), despite the risk of nest flooding, or opting for the central region of the wall, which offers a higher elevation. Therefore, in line with the study of shell influence on patch selection, we assessed whether the location of nests on the saltpan walls (central/side) had any effect on this selection.

If these hypotheses apply, the result may constitute a valuable management tool, either to attract breeding birds after restoring a saltpan or to increase numbers of nesting birds in active ones.

2.   Material and methods

2.1   Study area

Our study was conducted in southern Spain, in a 35-ha restored saltpan (Salinas de La Esperanza) within the Bay of Cádiz Bay Natural Park (36°30ʹ53.4ʺ N, 6°09ʹ23.3ʺ W). The study was carried out between 2019 and 2021 during the breeding season (early March to end of July). The conditions experienced during these years were comparable to those observed in previous periods. Three bird species regularly breed in this saltpan: the Kentish Plover (KP hereafter), the Little Tern (LT hereafter) and the Avocet, with a total of 100–150 KP nests, 70–100 LT nests and 70–90 Avocet nests per season (data drawn from 15 years of monitoring; SCI-SE, 2021). Most nests are found in the crystallisation part of the saltpan, where low bare walls (200 m long, 2–5 m wide and < 10–20 cm high) are surrounded by water (Fig. 1).

Figure  1.  (A) Study area of the restored saltpan ‘Salina de la Esperanza’, located in the inner area of the Bay of Cádiz, SW Spain. (B) Experimental zone: shell patches in a quincunx pattern. White spots indicate patches supplemented with shells positioned along the edges of the wall (edge_ext on the external side; edge_int on the internal side). Orange spots represent patches supplemented with shells positioned in the central part of the wall. (C) Detail of a shell patch. (D) Detail of a nest in a patch supplemented with shells. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

DownLoad: Full-Size Img PowerPoint

2.2   Experimental design

The experiment was carried out in an area of the saltpan where no more than eight KP nests and 10 LT nests have been found in the last 15 years (Castro et al., 2019). Two similar parallel walls measuring 139 and 144 m long and 13 and 15 m wide were chosen (Fig. 1A and B). Both walls were supplemented with marine shells (mainly Tapes spp. and Donax spp.) taken from a beach close to the saltpan. The walls were divided into 63 parallel strips 2-m wide and each strip was supplemented with 1-m² patches of marine shells (10 kg approx.) following a quincunx pattern. For example, if strip one was supplemented with a patch in the central area, strip two was supplemented with patches on each side but not in the centre (Fig. 1B). This design allowed patches with and without shells, which helped to evaluate whether the location of patches (centrally or laterally) has an effect on hatching success.

Two close walls (~350 m away from the experimental zone) of the active saltpan were selected as the control zone (Fig. 1A).

2.3   Nest fate

All nests found on the walls were identified marked with stones placed close to the nest. Each stone bore an alphanumeric code and was checked every four days. Nest visits lasted less than 5 min. Extremely windy and very hot days were avoided. A nest was considered successful if at least one egg hatched. Successful hatching was determined either when recently hatched chicks were seen in or close to the nest, or when the nest was found empty near the date of hatching with small pieces of eggshell left by adults when removing the eggs after hatching. If the laying date of a nest was unknown, we estimated the number of days that the nest had been active by employing an equation that considered the daily rate of mass loss in eggs during incubation relative to egg volume. For detailed information, refer to Fraga and Amat (1996) and Amat et al. (1999). Nests were considered predated when they were found empty far from the date of hatching or when evidence of predation was found (broken eggs, marks in the ground). When fate could not be determined, nests were considered as ‘fate unknown’ (<15%).

2.4   Statistical analysis

In order to ascertain whether the rate of selection by KP and LT of areas with or without provided supplemented nesting material (shells) differed significantly from expected random values, we used chi-square or Fisher's exact tests (the latter for samples with few records) (Sokal and Rohlf, 1970).

Subsequently, in order to determine the diagnostic significance of wall location in nesting selection, a Generalized Linear Model (GLM) with a binomial distribution and log-link function was employed (Zuur et al., 2010). In this analysis, the dependent variable was the presence or absence (1/0) of nests within the patch, evaluating a total of 409 patches. In addition to considering the presence or absence of shells and the location of nests (central or lateral) within the wall, we took into account the relative position of the nest within the walls-colony system. Specifically, we distinguished whether the nest was oriented towards the outside of the colony (edge_ext) or towards the interior of the colony (edge_int), while categorising those located in the central wall as oriented towards the central colony (edge_central). A forward stepwise generalized linear model was used to identify possible predictors of the binomial outcome ‘nest’ out of the following candidate variables: ‘shell’ with two levels (0/1); ‘location’, with two levels (central/lateral); ‘edge’, with three levels (edge_external/edge_central/edge_internal). At each step, variables were added based on p-values, and the AIC (Akaike, 1973) was used to set a limit on the total number of variables included in the final model. Once the final model was selected, we used it to predict the probability of nesting in the different areas of the wall applying the boot.predict function of the boot package (Canty and Ripley, 2021). The effect of these variables for habitat selection were tested for the KP and LT species, both individually and jointly, to demonstrate their role as a functional group of waders adapted to living in anthropized habitats such as saltpans. All analyses were performed with R software (R Core Team, 2022).

3.   Results

A notable increase in the number of nests of both species, LT and KP, was observed in the experimental area: 148 in total, with 42 and 26 LT nest and 57 and 23 KP nests in 2019 and 2021, respectively. In 2019, there were four times more nests in the experimental zones compared to the control zone (Table 1).

Table  1.  Percentage of total and hatched Kentish Plover and Little Tern nests in patches with shells and in patches without shells (bare).

		Experimental zone								Control zone
		2019				2021				2019			2021
		N	%	% success		N	%	% success		N	% success		N	% success
Kentish Plover	Shells	45	78.95	51.11		18	78.26	5.56		10	60		13	36.5
	Bare	12	21.05	50.00		5	21.74	20.00
Little Tern	Shells	34	80.95	47.06		21	80.77	4.76		5	100		14	36.13
	Bare	8	19.05	50.00		5	19.23	0.00
The number of nests and the percentage of hatched nests in the control zone (not supplemented with shells) are presented. % = percentage of total nest found in each substrate; % success = percentage of nests where at least one egg had successfully hatched.

 | Show Table

DownLoad: CSV

After supplementing the study area with marine shells, more that 75% of KP and LT nests were found in patches with shells, with the number of nests found in shells for both years being significantly different (Table 1, χ² = 11.308, df = 1, p-value < 0.001 and χ² = 24.362, df = 1, p-value < 0.001 for KP and LT, respectively).

Hatching success (% of hatched nest) was similar in nests found in shell patches and in nests found on bare ground (Table 1, Fisher test = 0.861, p-value = 1 and Fisher test = 0.808, p-value = 0.755 for KP and LT, respectively).

After testing the effect of the individual variables and the combinations of those significant in the presence or absence of nests (Appendix Tables S1–S3), the best model for KP or LT indicates that the contribution of shells is the factor that correlates most strongly with the location of nests on the wall (Table 2). The probability of using a place with shells to make their nests versus a bare place more than doubles for KP and triples for LT (Table 3).

Table  2.  Top explanatory Generalized Liner Models (GLM) results of all models tested for nest location selection of Kentish Plover and Little Tern on shell-supplemented walls in a saltpan.

Species		Model		Variables		Estimate		Std. error		z value		Pr (>|z|)
Kentish Plover		Nest–shell		Shell		1.268		0.52		2.441		0.015
Little Tern		Nest–shell		Shell		1.248		0.477		2.616		0.009
Kentish Plover + Little Tern		Nest–shell + location		Shell		1.315		0.363		3.627		0
				Location_central		0.621		0.32		1.937		0.053
Categorical variable ‘location’ uses location_lateral as reference. All models are depicted in Appendix Tables S1–S3.

 | Show Table

DownLoad: CSV

Table  3.  Predicted probability of habitat selection for Kentish Plover and Little Tern based on best explanatory models.

Species		Shell		Nest probability
Kentish Plover (KP)		1		0.09 (0.05–0.13)
		0		0.03 (0.01–0.05)
Little Tern (LT)		1		0.10 (0.06–0.15)
		0		0.03 (0.01–0.06)
Both species (KP + LT)
Shell		Location		Nest probability
1		Central		0.24 (0.16–0.33)
1		Lateral		0.15 (0.09–0.20)
0		Central		0.08 (0.03–0.14)
0		Lateral		0.04 (0.02–0.08)

 | Show Table

DownLoad: CSV

When modelling both species together as a saltpan wader functional group, we observe that in addition to the shells, the central location of the nest within the wall is also positively correlated (Table 3); the areas with shells that are also located in the centre of the wall have a six times greater probability than those without shells located on the side of the wall (Table 3).

4.   Discussion

4.1   Nesting habitat selection

Our research indicates that the Little Tern and the Kentish Plover showed a significant nesting preference for shell-supplemented patches over bare patches. More than 75% of Kentish Plover nests and 80% Little Tern nests were found in patches supplemented with marine shells.

Beaches are a natural breeding place for both species and most nests are usually built on marine shells (Gochfeld et al., 2020; del Hoyo et al 2021). While shells are common elements on beaches, they are scarce in marshes and saltpans. In saltpans, common nest materials also include pebbles and plants collected around the nest (Gómez et al., 2018). Nonetheless, our study suggests that the Kentish Plover and the Little Tern choose shells for nesting when they are available.

There are several potential explanations for the observed strong preference for shells. Firstly, it is possible that marine shells contribute as a calcium-rich supplement, aiding in the formation of eggs (Reynolds and Perrins, 2010). Additionally, the shells could enhance camouflage and/or create more favourable thermal conditions for optimal egg development (Hansell, 2000; Colwell et al., 2011), playing a crucial role in maintaining closely packed eggs and thereby preventing rolling. Furthermore, these shells could offer protection against potential hazards such as flooding or mud (Bailey et al., 2017).

While our study does not provide definitive answers to the aforementioned hypotheses, it is worth noting that neither of the two species studied directly consume shells, and consequently, the hypotheses concerning calcium intake may be discounted.

An experimental study carried out in the same saltpan showed that camouflage (and no thermoregulation) seems to be the main factor in choosing nest material in this site (Gómez et al., 2018), since shells could help to hide nests when birds are unable to attend to them. Nonetheless, hatching success was similar both in nests placed on bare ground and in shelled patches, indicating that, shell patches seems not to improve camouflage in this case.

The prevention of egg rolling does not appear to be the primary cause in this area, as fewer than 10% of nests are found with rolled-out egg (SCI-SE, 2021).

While conducting a comprehensive study would be necessary to definitively ascertain the most accurate hypotheses, our observations tend to support that the selection of shells is likely a strategy to prevent flooding or muddy conditions. In instances of rain or flooding, there is a risk that eggs could become embedded in the mud and potentially fracture when the female turns them during the incubation process (personal observation).

An alternative explanation to consider could be the shells’ role as a key factor in identifying suitable breeding sites, similar to their presence on beaches (Anteau et al., 2012). Beaches have been their natural and ancestral breeding habitats for both species, and the prevalence of plentiful shell patches might signify to the birds that the location is indeed a favourable breeding ground.

Regarding the best location within the saltpan walls, the best models for habitat selection have shown that first shell availability and then central location on the wall are the main factors explaining nest location for both the Kentish Plover and the Little Tern. While our experiment did not explicitly reveal the reason for the selection of the central area, we hypothesise that its central location could provide protection from potential rising tide water levels in the saltpan and subsequent nest flooding (Norte and Ramos, 2004). Conservation measures should prioritize flood prevention to improve the extent and the quality of suitable breeding habitats with adequate nesting substrates (Wu et al., 2020).

4.2   Hatch success rate

We expected to find a higher hatch success rate in nests located on shells. Nonetheless, hatch rate was similar in both types of patches. Nest failure or success probably depends on many variables, such as weather conditions, usual predators in the habitat, human activities and food availability (Norte and Ramos, 2004). In our study, predation by gulls is the most frequent cause for nest failure, which is responsible for more than 60% of predated nests in the saltpan (Liñán-Cembrano et al., 2021; SCI-SE, 2021). Given these results, our study suggests that the presence of shells does not prevent attacks from aerial predators.

We found a very different hatch success rate between nests on shelled patches in 2019 and 2021 (~80% in 2019 vs. ~5% in 2021). Avocets bred in the experimental zone next to Kentish Plovers and Little Terns in 2019, but not in 2021. This suggests that joint breeding of the above species acts as a real defence from aerial predators, as other studies have shown (Powell, 2001; Rocha et al., 2016). Indeed, the observed higher nesting success within the control zone in 2021 may be attributed to the fact that Avocets nested in the control zone in proximity to Kentish Plovers and Little Terns in both years. However, this result also prevents us from recommending a generalized use of marine shells to attract nesting birds to the area. It is necessary to carry out a prior study on predation and anti-predator behaviours to increase hatch success rate. Otherwise we would effectively be attracting nesting birds into a trap.

4.3   Conservation implications

Saltpans were abandoned in the last century due to their low profitability. Today, the Mediterranean coast holds only 34 artisanal saltpans out of the hundreds that have actively produced salt in the last century (Medartsal-Sustainable Management Model for Mediterranean Artisanal Salinas MEDARTSAL, 2021). However, a new demand for organic products and ecotourism experiences has seen a resurgence of many saltpans with restoration projects worldwide (Crisman et al., 2009; Martins et al., 2020; Ramos et al., 2022). Furthermore, due to the importance of saltpans in preserving certain ecosystem services (de Melo Soares et al., 2018), some are being restored only to maintain ecosystem functions, with bird conservation as a top priority (Stralberg et al., 2006; Athearn et al., 2012; Pruaño et al., 2012).

The results of this study can be used as a valuable tool for attracting birds to suitable restored saltpans, improving the conditions of crystalliser pond walls and supplementing them with shells, mainly in central areas, as our model suggests. This supplementation could lead to greater success with expensive restoration projects, where time is usually a common constraint.

Ethics statement

This study did not require any ethical approval.

CRediT authorship contribution statement

Macarena Castro: Writing – review & editing, Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Andrés De la Cruz: Writing – review & editing, Methodology, Formal analysis, Data curation, Conceptualization, E. Nuria Martin-Sanjuan: Writing – review & editing, Methodology, Investigation. Alejandro Pérez-Hurtado: Methodology, Investigation, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We thank to Servicio de Gestión del Medio Natural–Delegación de Cádiz from Consejería de Sostenibilidad, Medioambiente y Economía Azul (regional government), as well as the director of the Parque Natural Bahía de Cádiz for the support on conduct the study. Universidad de Cádiz provided facilities at Salinas La Esperanza (SCI-SE) during field work. Saltpan Initiative Project (MAVA Foundation) and MEDARTSALT project (EU-ENICBC) provided the funding for the study. We extend our gratitude to SALARTE ONG for generously supplying the shells for the experiments. Andrés De la Cruz was funded by the Margarita Salas Grant (2021-067/PN/MS-RECUAL/CD) from the Ministry of Universities of the Government of Spain and the European Union. We would especially like to thank the volunteers who participated in the shell supplementation.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.avrs.2024.100158.