Effects of Habitat Change on the Wintering Waterbird Community in China’s Largest Freshwater Lake

Abstract

Poyang Lake wetland in the middle and lower Yangtze River floodplain provides important wintering sites for migratory waterbirds. Extreme climatic events and human activities have resulted in the degradation and redistribution of habitat over the last few decades. However, the effects of habitat changes on the abundance of waterbirds remain unclear. We used long-term waterbird monitoring data and Landsat remote-sensing data to characterize changes in abundance and the relationship between habitat variation and abundance. A total of 113 waterbird species were recorded in the wintering period between 1999 and 2021, including 23 globally threatened species. Greater White-fronted Goose (Anser albifrons), Swan Goose (Anser cygnoides), Bean Goose (Anser fabalis), and Tundra Swan (Cygnus columbianus) were the dominant species. A total of 46 species with more than 6 annual surveys and average abundance >100 were recorded between 1999 and 2021. For most species, the mean abundance across all sub-lakes was higher in the first year than in the last year, and no obvious changes were observed over the last 20 years. The mean abundances of the vulnerable species Hooded Crane (Grus monacha) and White-naped Crane (Grus vipio) significantly declined between 1999 and 2021. No significant changes in the mean abundance of all foraging groups were observed. The area of water bodies increased, and the area of mudflats decreased. For most species with significant changes in abundance, habitat change did not greatly contribute to variation in mean abundance. The reduction in the area of mudflats greatly contributed to declines in the mean abundance of the gray heron (Ardea cinerea) and gadwall (Anas strepera).

1. Introduction

Wetlands are the transitional zones between aquatic and terrestrial systems, and they provide several important ecosystem functions, including flood reduction, water purification, carbon storage, erosion control, and biodiversity conservation [1]. Wetlands are also ecologically fragile areas because of intensive external stress associated with climate change and human disturbance. For example, climate change has resulted in the loss of 47% of the area of inland Ramsar wetlands between 1980 and 2014, and the area of these wetlands is expected to be 6000 km² by 2100 [2]. Land reclamation supports rapid urbanization and economic development and has resulted in reductions of 50% of the area of coastal wetlands in China; the reclamation rate has increased from 24,000 ha year⁻¹ (1950–2000) to 40,000 ha year⁻¹ (2006–2010) [3]. The decrease in wetland areas and the loss of their associated functions pose major threats to wetland function and biodiversity [4,5].

The middle and lower Yangtze River floodplain is a wetland area of global significance that comprises rivers, shallow water, and freshwater lakes [6]. These wetlands provide important wintering sites for migratory waterbirds along the East Asian–Australasian Flyway (EAAF) [7]. However, human activities, land reclamation, the invasion of alien species, water contamination, and overfishing have caused the functional degradation of 278 of 423 freshwater lakes (65.72%), according to the Second National Survey of Wetland Resources [8]. Waterbird populations have been concentrated in a few freshwater lakes in recent decades, such as Poyang Lake [9].

Poyang Lake wetland is an important component of the middle and lower Yangtze River floodplain [10], and complex hydrological fluctuations have resulted in the formation of different types of habitats, which has supported high waterbird biodiversity [11]. Wetland vegetation is the main food source for herbivores; mudflats provide foraging habitats for invertebrate eaters, fish eaters, and omnivores; and water bodies provide foraging habitats for tuber feeders, fish eaters, invertebrate eaters, and omnivores [12]. Over 480,000 waterbirds belonging to 69 species use these habitats as their wintering home [13]. Geese, swans, and shorebirds are considered the most common groups of waterbirds. Twelve globally threatened species defined by IUCN have been documented in this region, including Siberian Crane (Leucogeranus leucogeranus), Baer’s Pochard (Aythya baeri), Oriental White Stork (Ciconia boyciana), Dalmatian Pelican (Pelecanus crispus), Lesser White-fronted Goose (Answer erythropus), Swan Goose (Anser cygnoides), Hooded Crane (Grus monacha), and White-naped Crane (Grus vipio) [13]. Large proportions of the global populations of several threatened species occur in this region, including nearly 100%, 90%, 50%, and 20% of the global populations of Siberian Cranes (critically endangered), Oriental White Storks (endangered), Swan Geese (vulnerable), and White-naped Cranes (vulnerable), respectively [14,15,16].

Extreme climate events and human activities have resulted in dramatic changes in the Poyang Lake wetland over the past few decades [17]. Recent dam construction in the Yangtze River has greatly affected the hydrological regimes of Poyang Lake. This has reduced the connectivity between the Yangtze River and Poyang Lake wetland, which has led to a decrease in the water level in the dry season between October and March and an increase in the area of water in surrounding rivers [18,19]; this will eventually result in the redistribution of habitat resources. Although previous studies have explored changes in land use caused by human activities and their ecological effects on waterbird habitats in the Poyang Lake wetland [20], the effects of habitat variation on waterbird community structure remain unclear.

To evaluate the effect of habitat change on the wintering waterbird community in the Poyang Lake wetland, we characterized variation in waterbird diversity and habitat area over the last 20 years, as well as the relationship between habitat change and variation in waterbird diversity. We then examined the effects of variation in habitat (water bodies, mudflats, and wetland vegetation) on waterbird diversity in the Poyang Lake wetland.

2. Methods

2.1. Study Area

Poyang Lake (28°11′–29°51′N, 115°49′–116°46′E) is the largest freshwater lake in China, and it is connected to the Yangtze River. The water level of Poyang Lake varies seasonally. In the dry season, mudflats and wet meadows are present in the main wetland areas [6]. From late November to early March, a low water period creates multiple habitats for migratory waterbirds, including sparse vegetation, mudflats, water bodies with a depth of 0–30 cm, water bodies with a depth of 30–60 cm, and water bodies with a depth of >60 cm. These natural wetlands support most wintering waterbird populations in Poyang Lake [13]. However, the Poyang Lake wetland experiences rapid change due to human activity and extreme events. The spatial distribution and area of the main habitats, including sparse vegetation, mudflats, and water bodies, have been altered (Figure 1), and this has affected the waterbird community structure.

2.2. Waterbird Survey

Wintering waterbird surveys at the sub-lake scale in Poyang Lake have been conducted since 1999. Waterbird survey data between 1999 and 2021 (9 January 1999; 9 January 2001; 9 January 2002; 9 January 2003; 9 January 2004; 9 January 2005; 29 December 2006; 3 January 2008; 13 February 2009; 27 February 2010; 12 January 2011; 8 January 2012; 18 January 2013; 10 January 2014; 1 January 2017; 1 January 2018; 1 January 2019; and 1 January 2021) were acquired from the Wildlife Protection Administration of Jiangxi Province. All surveys were conducted in winter because waterbird populations peak during this period. Two to three investigators counted the number of waterbirds using single-tube telescopes via the direct counting method if large waterbirds dominated the survey site. Surveys were conducted by counting ecological groups (geese, cranes, storks, ducks, shorebirds, and gulls) of small waterbirds that were abundant at the survey site. Each ecological group comprised 50, 100, and 200 individuals [21].

2.3. Abundance Trends

We analyzed changes in waterbird abundance from 1999 to 2021 at the species and foraging group level. We analyzed variations in the abundances of species that had been detected in more than 6 annual surveys and with average abundances greater than 100. We also examined changes in the mean abundances across all sub-lakes of five foraging guilds, herbivores, invertebrate eaters, fish eaters, omnivores, and tuber feeders, which were classified following the criteria of [12].

We used interannual variation and long-term abundance changes to identify abundance trends for each species and foraging group. We calculated the average abundance size for all sub-lakes in each year for each species and foraging group; we then compared the abundances (population counts) of each species and foraging group in the first survey year and last survey year (1).

$$p=\frac{C_{last}-C_{first}}{C_{first}}\times 100\%$$

(1)

p is the percentage change in the population count between the first survey year and the last survey year; C_last and C_first indicate the population count in the first survey year and last survey year, respectively.

Long-term abundance trends were analyzed using linear trend-fitting models at the 5% significance level in IBM (International Business Machines Corporation, New York, NY, USA) SPSS (Statistical Product and Service Solutions, Chicago, IL, USA) Statistics 22.0 software. The goodness of fit of the models was assessed on the basis of the p-value, and p < 0.05 indicated a good model fit with significant trends. We considered a species and foraging group to be declining if the percentage change was negative and the linear trend-fitting model revealed a significant declining trend.

2.4. Habitat Condition

According to the habitat requirements of the 5 different foraging guilds, the habitat types were divided into sparse vegetation, mudflats, water bodies with a depth of 0–30 cm, water bodies with a depth of 30–60 cm, and water bodies with a depth of >60 cm [11,12]. Detailed habitat divisions for different foraging guilds are provided in Supplementary Table S1.

Remote-sensing images used for extracting waterbird habitat data were obtained from the United States Geological Survey (http://glovis.usgs.gov/ (accessed on 16 July 2023)). Five Landsat TM (18 December 1999; 6 January 2007; and 14 January 2010, (accessed on 1 January 2012)) and Landsat OLI (13 February 2015; 23 January 2019, (accessed on 1 July 2020)) remote-sensing images were used. The spatial resolution of these data was 30 m, and the selected images were cloud free. The ISODATA unsupervised classification and the hierarchical decision tree classification were used to classify the landscape types [22]. According to the habitat characteristics of waterbirds [23], we first classified the wetland landscape into three main landscape types: water, land–water transition zone, and vegetation. The Automated Water Extraction Index (AWEI), which comprises the AWEIsh and AWEInsh indices, was used to extract water surface data [24]. The Normalized Difference Vegetation Index (NDVI) was used to extract vegetation area data [25].

According to the habitat requirements of the five different foraging guilds, we further classified the water into deep water (>60 cm water depth), moderate-depth water (30–60 cm water depth), and shallow water (<30 cm water depth). The procedure for calculating the water depth was as follows. First, the “Raster To Polygon” and ‘Polygon To Lines” tools were used to extract the boundary of the water area; second, the “Feature Vertices To Points” and “Extract Values to Points” tools were used to generate points at the water boundary and obtain the elevation of points, respectively; the “Natural Neighbor” interpolation tool was used to obtain the elevation of the water surface; and the “Minus” tool was used to subtract the elevation of the water body and surface to generate water depth. Mudflats and grassland in the low-elevation zone also provide habitat for some waterbirds; we thus classified the land–water transition zone and vegetation into mudflats, sand, sparse grassland (<13.5 m elevation), dense grassland, and high grassland (>13.5 m elevation). The digital elevation model (DEM) used to subdivide the water and vegetation areas was provided by the Jiangxi Provincial Bureau of Surveying and Mapping. Field survey data and ground truth data were compared to ensure that the overall classification accuracy was greater than 80%, which was sufficient for the objectives of this study [26]. All remote-sensing images were processed in ArcGIS 10.7 software (ESRI, Redlands, CA, USA).

2.5. Effects of Habitat Change on Population Size

According to the results of the population trend analysis at the species and foraging group level, we analyzed species and foraging groups that showed significant variation over the survey period. We conducted Spearman’s correlation analyses at the 5% significance level in IBM SPSS Statistics 22.0 to evaluate the relationships between habitat area and population size between 1999 and 2019.

3. Results

3.1. Waterbird Diversity

A total of 113 waterbird species were recorded during the wintering period between 1999 and 2021 (Supplementary Table S2). The mean number of species was 56 (±16 SD), and the mean population size was 363,640 (±146,549 SD). A total of 23 globally threatened species were observed, including the critically endangered species Siberian Crane, Baer’s Pochard, and Scaly-sided Merganser (Mergus squamatus); the endangered species Far Eastern Curlew (Numenius madagascariensis), Oriental White Stork, Black-faced Spoonbill (Platalea minor), and Nordmann’s Greenshank (Tringa guttifer); the vulnerable species Hooded Crane (Grus monacha), White-naped Crane, Saunders’s Gull (Larus saundersi), Common Pochard (Aythya ferina), Red-breasted Goose (Branta ruficollis), Swan Goose, Swinhoes Yellow Rail (Coturnicops exquisitus), Chinese Egret (Egretta eulophotes), and Lesser White-fronted Goose (Anser erythropus); and the near-threatened species Eurasian Curlew (Numenius arquata), Bar-tailed Godwit (Limosa lapponica), Northern Lapwing (Vanellus vanellus), Black-tailed Godwit (Limosa limosa), Dalmatian Pelican (Pelecanus crispus), Eurasian Oystercatcher (Haematopus ostralegus), and Falcated Duck (Anas falcata). Species were considered dominant when their relative abundance was greater than 10% of the total number of waterbirds. In most survey years, Tundra Swan, Swan Goose, White-fronted Goose, and Bean Goose were the dominant species.

3.2. Abundance Trends

A total of 46 species that were detected in more than 6 annual surveys and with average abundances >100 were recorded between 1999 and 2021 (

Table 1. Percentage change analysis and linear trend-fitting model analysis. % change. ^a: The percentage change in the population count between the first survey year and last survey year. N ^b: The sample size of the linear trend-fitting model analysis. F ^c: The degree of freedom of the linear trend-fitting model analysis. p ^d: The 5% significance level of the linear trend-fitting model analysis. ^e: Change trends of population size.

	% Change a	N b	F c	p d	Population Trends e
Species
Greater White-fronted Goose	−50.5%	18	1.682	0.213	No significant trend
Common Coot	−1.6%	17	1.682	0.021	Declining
Siberian Crane	−53.0%	16	0.712	0.413	No significant trend
Eurasian Spoonbill	84.7%	18	9.866	0.006	Increasing
Hooded Crane	−71.3%	15	9.973	0.008	Declining
Eurasian Curlew	−33.3%	8	0.89	0.382	No significant trend
White-naped Crane	−88.6%	16	12.443	0.003	Declining
Spot-billed Duck	14.0%	17	0.712	0.412	No significant trend
Grey Heron	−66.5%	18	10.370	0.005	Declining
Gadwall	−87.7%	13	9.225	0.011	Declining
Eurasian Wigeon	107.9%	16	1.636	0.222	No significant trend
Ruddy Shelduck	32.9%	18	1.442	0.247	No significant trend
Eastern Great Egret	−79.0%	12	0.291	0.601	No significant trend
Oriental White Stork	−60.6%	18	0.76	0.396	No significant trend
Bean Goose	1534.1%	17	22.646	0.000	Increasing
Pied Avocet	272.8%	15	0.131	0.724	No significant trend
Northern Lapwing	−9.3%	18	0.104	0.752	No significant trend
Great-crested Grebe	746.8%	13	7.108	0.022	Increasing
Tufted Duck	−42.9%	12	0.01	0.922	No significant trend
Spotted Redshank	−86.6%	14	4.886	0.047	Declining
Black-winged Stilt	−93.3%	7	1.780	0.240	No significant trend
Dunlin	−12.3%	13	0.385	0.548	No significant trend
Common Moorhen	928.6%	13	2.838	0.12	No significant trend
Black-tailed Godwit	−76.7%	12	3.107	0.108	No significant trend
Common Redshank	−95.1%	16	9.986	0.007	Declining
Common Pochard	1267.9%	11	4.139	0.072	No significant trend
Common Black-headed Gull	42369.4%	18	2.743	0.117	No significant trend
Swan Goose	199.6%	18	0.626	0.44	No significant trend
Baikal Teal	−73.9%	12	0.046	0.835	No significant trend
Common Crane	619.5%	18	2.779	0.115	No significant trend
Grey-lag Goose	150.6%	17	9.854	0.007	Increasing
Falcated Duck	−89.6%	15	0.0045	0.835	No significant trend
Common Teal	−75.5%	17	0.511	0.486	No significant trend
Mallard	−82.8%	17	1.614	0.223	No significant trend
Northern Shoveler	−95.3%	12	2.169	0.172	No significant trend
Great Cormorant	−92.2%	16	6.831	0.02	Declining
Common Greenshank	−87.7%	16	1.102	0.312	No significant trend
Baer’ s Pochard	−74.5%	11	1.211	0.300	No significant trend
Lesser White-fronted Goose	−64.0%	16	0.01	0.921	No significant trend
Little Egret	200.6%	18	0.999	0.332	No significant trend
Little Grebe	−15.6%	18	0.638	0.436	No significant trend
Tundra Swan	9.3%	18	0.322	0.578	No significant trend
Whiskered Tern	−91.8	6	1.022	0.369	No significant trend
Black-crowned Night Heron	−98.6	7	3.364	0.126	No significant trend
Herring Gull	−65.5	18	0.004	0.95	No significant trend
Northern Pintail	−81.6	17	0.116	0.738	No significant trend
Foraging groups
Herbivores	2.7%	18	2.1	0.167	No significant trend
Invertebrate eaters	−25.9%	18	0.317	0.581	No significant trend
Fish eaters	−24.0%	18	2.97	0.104	No significant trend
Omnivores	−1.6%	18	0.862	0.367	No significant trend
Tuber feeders	57.3%	18	0.148	0.705	No significant trend

). For most species, the mean abundance across all sub-lakes was higher in the first year than in the last year (Table 1), and no significant changes were observed over the last 20 (Figure 2). For a few species, the mean abundance significantly decreased, including Common Coot (Fulica atra), Hooded Crane, White-naped Crane, Grey Heron (Ardea cinerea), Gadwall (Anas strepera), Spotted Redshank (Tringa erythropus), and Great Cormorant (Phalacrocorax carbo). For a few species, the mean abundance significantly increased, including Eurasian Spoonbill (Platalea leucorodia), Bean Goose, Great-crested Grebe (Podiceps cristatus), and Grey-lag Goose (Anser anser) (Table 1; Figure 2). No significant changes were observed in the mean abundances of foraging groups (Table 1; Figure 2).

3.3. Habitat Change between 1999 and 2019

From 1999 to 2019, the area of water bodies with a depth of >60 cm, 30–60 cm, and <30 cm and sparse vegetation increased in Poyang Lake, and the area of mudflats decreased (Figure 3). The area of habitat with water bodies with a depth of >60 cm first increased by 69.06%, decreased, and then increased. The area of habitat with water bodies with a depth of 30–60 cm increased by 26.27%, and little variation was observed throughout the study period. The area of habitat with water bodies with a depth of <30 cm increased by 44.93%, and this change was consistent with observed variation in the area of habitat with water bodies with a depth of >60 cm. The area of habitat with mudflats decreased by 50.46% over the study period, but it first decreased, increased, and then decreased. The area of sparse vegetation increased by 23.85% over the study period, but it first increased and then decreased (Figure 3).

3.4. Relationship between Change in Habitat Area and Change in Waterbird Abundances

We evaluated the relationship between habitat area and the mean abundance of species that showed significant changes over the study period. For most species, no significant relationships were observed between changes in the habitat area and changes in mean abundance (

Table 2. The relationship between habitat change and mean abundance change for species showing significant changes in abundance.

	Water Bodies with a Depth of >60 cm		Water Bodies with a Depth of 30–60 cm		Water Bodies with a Depth of <30 cm		Mudflat		Sparse Vegetation
Common Coot	r	−0.618	r	−0.740	r	−0.643	r	0.691	-	-
	p	0.102	p	0.036	p	0.086	p	0.058	-	-
Eurasian Spoonbill	-	-	r	−0.061	r	0.085	r	−0.036	-	-
	-	-	p	0.887	p	0.842	p	0.932	-	-
Hooded Crane	-	-	r	0.073	r	0.218	-	-	-	-
	-	-	p	0.864	p	0.604	-	-	-	-
White-naped Crane	-	-	r	−0.073	r	−0.073	-	-	-	-
	-	-	p	0.864	p	0.864	-	-	-	-
Grey Heron	-	-	r	−0.764	r	−0.812	r	0.764	-	-
	-	-	p	0.027	p	0.014	p	0.027	-	-
Gadwall	r	−0.837	r	−0.473	r	−0.667	r	0.909	-	-
	p	0.010	p	0.237	p	0.071	p	0.002	-	-
Bean Goose	-	-	-	-	-	-	-	-	r	0.255
	-	-	-	-	-	-	-	-	p	0.543
Great-crested Grebe	r	0.147	r	0.147	r	0.147	r	0.265	-	-
	p	0.781	p	0.781	p	0.781	p	0.612	-	-
Spotted Redshank	-	-	-	-	r	−0.267	r	0.218	-	-
	-	-	-	-	p	0.523	p	0.604	-	-
Common Redshank	-	-	-	-	r	−0.182	r	0.182	-	-
	-	-	-	-	p	0.666	p	0.666	-	-
Grey-lag Goose	-	-	-	-	-	-	-	-	r	0.194
	-	-	-	-	-	-	-	-	p	0.645
Great Cormorant	r	−0.061	r	0.158	r	0.158	r	0.036	-	-
	p	0.887	p	0.709	p	0.709	p	0.932	-	-

). For a few species, habitat change greatly contributed to variation in mean abundance, including the Common Coot, Grey Heron, and Gadwall.

4. Discussion

Environmental change and its effects on biodiversity have been widely explored at the global scale [27,28]. For example, the effects of land use change on global terrestrial biodiversity [27], cropland expansion on post-2020 global biodiversity [29], and natural wetland loss at stopover sites on migratory shorebird populations along the EAAF migratory route [30] have been examined in previous studies. Many studies have explored the relationship between habitat change and biodiversity variation. Changes in waterbird distributions, waterbird biodiversity, and habitat at Poyang Lake have been examined in previous studies. Most of these studies have focused on specific species, especially cranes [15,31,32]. However, our study is the first to evaluate the effects of habitat change on waterbird populations at Poyang Lake using long-term survey data and remote-sensing images.

Consistent with the results of previous studies, Poyang Lake provides an important wintering site for multiple globally threatened waterbird species, especially the Siberian Crane, Oriental White Stork, Hooded Crane, White-naped Crane, and Swan Goose [7]. No significant changes were observed in the mean abundances of most species and foraging groups over the last 20 years, and the mean abundance was higher in the first year (1999) than in the last year (2021); these findings were not consistent with the observed increases in population size for most species in [13]. This might be explained by the lack of consistency in the survey periods and differences in data sources. Waterbird population data between 1997 and 2014 were collected using books, reports, and papers in the work of [13], whereas systematic surveys between 1999 and 2021 were conducted in this study. The mean abundances of the dominant species Hooded Crane and White-naped Crane significantly decreased. The mean abundances of the Bean Goose and Grey-lag Goose significantly increased; these findings were consistent with the results of [13].

The area of water bodies increased, and the area of mudflats decreased. Over the last few decades, the distribution of sand vessels has continually expanded southward, and these mining activities now occur in the center of the lake [33]. Sand mining activities have altered the lakebed, and this might have led to an increase in the area of deep water bodies and a decrease in the area of mudflats. In addition, a previous study has shown that increases in precipitation in the winter in the Poyang Lake basin over the last few decades have contributed to these habitat changes [34]. In our study, no significant relationship was observed between habitat change and the mean abundance of most species and foraging groups. Potential explanations for this finding are manifold. First, although the unsupervised classification method can easily extract image information, the classification accuracy greatly depends on the initial segmentation parameters, and errors can lead to decreases in the classification accuracy [35,36]; this might introduce errors in estimates of the relationship between habitat change and population variation. Second, the small sample size caused by the short time interval over which habitat change was examined might also explain the weak relationship. However, we still managed to detect significant effects of the decline in the area of mudflats on the mean abundance of the Grey heron and Gadwall. The conservation of mudflats thus requires special attention.

In future studies, habitat information should be extracted over multiple periods using long-term remote image data in Poyang Lake to reveal the relationship between habitat change and population variation. In addition, multiple sites along the middle and lower Yangtze River area, EAAF, and the globe provide important habitats for migratory waterbirds [7,37]. Whether habitat change poses a serious threat to these regions remains unclear. Long-term surveys at multiple sites are needed, and this will require the participation of experienced birdwatchers, non-governmental organizations, and researchers.

5. Conclusions

We explored the effects of habitat change on wintering waterbird abundance in Poyang Lake. The Poyang Lake wetland provides an important wintering site for multiple rare and endangered species. For most species and foraging groups, mean abundances were higher in the first year than in the last year, and no significant changes were observed over the last 20 years. The mean abundances of two vulnerable species, the Hooded Crane and the White-naped Crane, significantly declined. The area of water bodies increased, and the area of mudflats decreased. For most species showing significant changes in mean abundance, habitat change did not greatly contribute to mean abundance variation. Nevertheless, we found that a reduction in the area of mudflats greatly contributed to declines in Gay heron and Gadwall mean abundances. Multiple sites along the EAAF and in other regions provide key habitats for migratory waterbirds, which have been negatively affected by environmental and land use change. The methodology used in our study could be used to clarify the effects of habitat change on waterbird biodiversity at the global scale.