Long-term changes in numbers and distribution of wintering waterbirds in the Czech Republic, 1966–2008

Abstract

Capsule Of 26 species of wintering waterbirds, 18 showed an increase in numbers, five showed a decrease and two showed no change.

Aim To assess long-term trends in the numbers and distribution of the 26 most abundant wintering waterbird species in the Czech Republic.

Methods We used International Waterbird Census data from between 48 and 639 wetland sites which had been counted annually in the Czech Republic from 1966 to 2008. From these data long-term changes in numbers and distributions were determined. Log-linear Poisson regression analysis was used to estimate missing data using trim software. The distribution of each species was described as the ratio of the number of sites occupied by that species to the total number of sites investigated.

Results Increasing trends were found for 18 species, five species were found to be declining, one species was stable and two species were found with uncertain trends. Wintering distributions (the ratio of sites occupied by a given species to the total number of sites counted) increased in 16 species and decreased in two species, broadly correlated with the species changes in numbers.

Conclusion In most species changes in numbers as well as changes in distribution followed the Western Palearctic population trends. Those species which increased were mainly piscivores and included geese, ducks and gulls. Scarcer species also exhibited an increase in numbers. The changes in numbers (both positive and negative) were more frequent among species associated with running water, whereas species which showed uncertain trends were more frequently recorded on standing water, which is more affected by variable weather conditions.

Trends in many waterbird species, including species of conservation concern, have changed significantly in recent decades (e.g. Birdlife International 2004, Wetlands International 2006). As in many bird groups, the population dynamics of waterbird species are affected by various factors occurring throughout the year, including the breeding season, spring and autumn migration, and the wintering season. Waterbirds generally breed in low densities over large areas (Scott & Rose 1996, Kear 2005) but aggregate in large numbers in winter, when limited availability of suitable habitats may cause large temporal and spatial variability (Ridgill & Fox 1990). Inter-seasonal variation in numbers and distribution of particular species are considerably affected by weather and habitat changes (Wahl & Sudfeldt 2005, Maclean et al. 2008, Musil et al. 2008a, Musilová et al. 2009). Changes in wintering conditions can be assumed as one of the key factors affecting wintering numbers, over-wintering survival, and consequently the population dynamics of a particular species.

Among wintering waterbirds in Europe, significant population trends (either increases or decreases) have been found in roughly half of all populations (48%; Wetlands International 2006), while the remaining populations are considered as stable, or with uncertain trends. Although counting effort and coverage of particular countries by monitoring programmes should be taken into consideration, it still remains remarkable that, of all the continents, Europe has the highest proportion of populations displaying marked increasing or decreasing trends (e.g. Kear 2005, Wetlands International 2006). Various analyses of wintering waterbird trends at a national (Crowe et al. 2008, Nilsson 2008, Slabeyová et al. 2008, 2009, Fouque et al. 2009, Hustings et al. 2009) or regional (Maclean et al. 2008) level have been carried out. Nevertheless, there are few studies of the status of wintering waterbirds in central Europe.

The Czech Republic is not a major waterbird wintering area because the frost period is somewhat prolonged compared with more western areas of Europe (Hudec 1994, Delany et al. 1999, Musil et al. 2001, Gilissen et al. 2002). Despite this, the climate is relatively mild and there is a high diversity of smaller wetland habitats, which provide some feeding opportunities throughout the winter period for birds which breed in northern Europe, particularly when freezing conditions in the Baltic region may limit the birds' access to feeding areas (e.g. Svazˇas et al. 2001, Nilsson 2008). Conversely, species with a more southerly distribution, which usually leave central Europe to winter in Mediterranean areas (Musil et al. 2001, Cepák et al. 2008), may delay their southbound movement in milder winters when unfrozen wetlands are available. The Czech Republic can therefore provide attractive wintering areas for waterbird species with differing wintering strategies (Hudec 1994, Hudec et al. 1995). To date, there are published analyses of trends in wintering numbers available only for geese (Musil et al. 2008a) and ducks (Musilová et al. 2009).

The long tradition of wintering waterbird monitoring in the Czech Republic started with contributions to the International Waterbird Census in 1966 and now covers almost all sites of national importance. We use these data to try to answer the following questions:

•	Do numbers of Czech wintering waterbirds show long-term changes?
•	Are the changes in numbers and distribution among individual waterbird species correlated?
•	What species-specific variables are responsible for changes in numbers and distribution of individual species? We tested the effect of the following species-specific variables: body size, Western Palearctic population size and trend, mean wintering numbers and distribution in the Czech Republic, biogeographic position, and conservation status at the national and European level. If global trends are more important that local conditions, we would expect that trends in numbers of wintering waterbirds would follow those in the Western Palearctic. We predicted that wintering numbers and distribution would increase particularly in rare and southern species, due to the northward shift in wintering range (see Reif et al. 2008). On the contrary, we expected a decrease in numbers or in distribution of more northern species, i.e. whose main wintering range is located in the Baltic. Moreover, we expected that the protection of individual bird species would have an effect on their population trend (see Vorˇísˇek et al. 2008). We also expected significant increase in numbers in huntable species with larger body size, which were affected by intensive hunting before the 1970s, but which has since declined (see Mooij 2005).
•	Are there any differences in trends in the numbers of birds which exploit running or standing water? We expected more conspicuous changes amongst species on standing water compared to running water, because standing water is more affected by global warming that can increase the extent of non-freezing water bodies suitable for wintering waterbirds.

METHODS

Waterbird data

Long-term trends in the numbers and distribution of waterbird species were analysed using count data recorded in the Czech Republic for the International Waterbird Census (IWC), which is conducted in mid-January each winter. Within the Czech Republic, the IWC counts have been carried out annually at between 48 and 639 wetland sites in January of each year between 1966 and 2008 inclusive. The IWC counts were carried out at between 48 and 200 wetlands sites between January 1966 and 2003. There was a significant increase in the number of sites counted between January 2004 and 2008. In total, between 479 and 639 sites were counted annually in January 2004–2008. Regional coverage of the Czech Republic (Fig. 1) as well as the percentage of sites classified as running and standing waters remained similar in all the sites from 1966 to 2008 (Fiala 1980, Musil et al. 2001, Musilová & Musil 2006). Altogether, 175 wetland sites in the Czech Republic were counted in at least 10 seasons from the 1960s until 2008 (Musil & Musilová 2010).

Figure 1. Distribution of counted sites in the Czech Republic.

Data for the 26 most abundant waterbird species were included in the analysis. These species were included in the analysis if their annual counts exceeded 50 individuals in any year and if they were recorded in more than 20 winter seasons (see Table 1 for list of species, their scientific names, distribution, and recent numbers). The three gull species Larus argentatus, L. cachinnans and L. michahellis were termed ‘large gulls’ and thereafter considered in the text as only one bird species, in accordance with the taxonomic situation current at the beginning of the study period (i.e. in 1966). The proportion of missing counts varied between 52.85% and 79.32% and the proportion of estimated numbers (calculated using trim software: Pannekoek & Van Strien 2005) varied between 27.87% and 85.21% in particular species. Thus, the proportion of missing and imputed counts in any species did not exceed a level of 90%, which can be regarded as an extremely high proportion of imputed counts (Soldaat et al. 2007). Wintering numbers, i.e. numbers of individuals recorded by mid-winter IWC counts, were used to provide the range of individuals counted in 2004–2008.

Table 1. Number of occupied sites, proportion of missing counts, and estimated numbers in the whole study period (between 1966 and 2008), and estimation of recent wintering numbers in January 2004–2008 (see Methods/Waterbird data for explanation of terms). Data for 26 of the most abundant waterbird species are shown.

Common name	Scientific name	Proportion of occupied sites % (n)	Proportion of missing counts %	Proportion of estimated numbers %	Wintering numbers 2004–2008 (individuals)
Little Grebe	Tachybaptus ruficollis	40.2 (337)	68.3	74.6	330–900
Great Crested Grebe	Podiceps cristatus	14.9 (125)	58.2	68.7	40–320
Great Cormorant	Phalacrocorax carbo	44.6 (374)	70.9	38.3	9000–14 200
Great White Egret	Egretta alba	17.7 (148)	70.5	41.6	100–500
Grey Heron	Ardea cinerea	77.5 (649)	74.7	61.7	1900–2900
Mute Swan	Cygnus olor	62.7 (525)	72.1	60.6	2000–3800
Bean Goose	Anser fabalis	9.6 (80)	59.0	85.2	400–6000
White-fronted Goose	Anser albifrons	5.3 (44)	56.3	84.6	1500–13 800
Greylag Goose	Anser anser	9.9 (83)	59.1	77.2	800–2400
Eurasian Wigeon	Anas penelope	10.3 (86)	60.5	58.2	70–170
Gadwall	Anas strepera	9.2 (77)	60.9	39.7	50–300
Common Teal	Anas crecca	28.9 (242)	64.0	66.9	450–1200
Mallard	Anas platyrhynchos	95.0 (796)	77.2	67.6	140 000–180 000
Common Pochard	Aythya ferina	29.8 (250)	61.8	35.9	800–1400
Tufted Duck	Aythya fuligula	32.5 (272)	66.7	27.9	3600–5100
Common Goldeneye	Bucephala clangula	26.1 (219)	60.6	59.3	500–1200
Smew	Mergellus albelus	8.8 (74)	52.9	61.0	40–110
Goosander	Mergus merganser	36.0 (302)	67.6	62.1	1500–3300
White-tailed Eagle	Haliaeetus albicilla	16.2 (136)	65.2	64.9	70–100
Common Moorhen	Gallinula chloropus	32.5 (272)	65.7	61.2	300–700
Common Coot	Fulica atra	53.7 (450)	69.8	60.1	8500–11 000
Black headed Gull	Larus ridibundus	33.1 (277)	64.0	54.6	4000–10 000
Mew (Common) Gull	Larus canus	14.7 (123)	55.4	68.8	1000–4000
Large gulls	Larus spp.	21.2 (178)	59.8	39.1	960–2200
Common Kingfisher	Alcedo atthis	42.1 (353)	72.3	71.6	150–350
White-throated Dipper	Cinclus cinclus	23.6 (198)	79.3	83.1	330–500

Trend analysis

Trend analyses were carried out using IWC data from 838 of the 1078 sites that were counted in at least two winters between 1966 and 2008. Log-linear Poisson regression analysis was used to estimate missing data using trim software (Statistic Netherlands version 3.52, Pannekoek & Van Strien 2005). Missing data resulted from incomplete coverage caused by limited availability of volunteers in some seasons. Serial correlations between annual numbers and over-dispersion in the data were taken into account. The models included change points to allow for changes in the slope parameters at some points in the time series (Pannekoek & Van Strien 2005, Fouque et al. 2007, 2009).

The multiplicative slope (i.e. the change in indices from one year to the next) was the value used to express population trends over the study period. Moreover, the trim classification of the species trends was used in one of six categories, depending on whether the rate of change over the study period was more or less than 5% per year: a strong increase or decrease (> 5% per year), a moderate increase or decrease (< 5% per year), a stable trend (the trend is not significant and the confidence limits were sufficiently small), or an uncertain trend with large values of confidence interval (Pannekoek & van Strien 2005, Fouque et al. 2009).

Additionally, wetlands were classified as standing water (403 sites: fishponds, reservoirs, gravel and sand-pit lakes, and industrial settling ponds) or running water (435 sites: rivers and streams), and separate trend analyses (using the trim software) were undertaken for each type. For running water (rivers and streams), sites were defined as river sections with known boundaries, such as dams, weirs and bridges (for the list of wetland habitats in Czech Republic, see Chytil et al. 1999). Percentages of running and standing waters among counted sites and regional coverage of the Czech Republic did not change during the period covered (Musil & Musilová 2010).

Distribution of species

The distribution of each species was described as the ratio (arcsin transformed) of the number of sites occupied by that species to the total number of sites investigated. Linear regression analysis was then used to identify potentially significant long-term changes in species distribution. Correlation coefficients derived from the linear regression analysis were used to describe the change in species distribution over the study period for each species (Table 2).

Table 2. Changes in distribution (correlation coefficient (r) and significance: * P < 0.05, ** P < 0.01, n.s. = not significant; n = 43; see Methods) and changes in numbers in waterbird species, 1966–2008 (multiplicative rate of change ± se) on all wetlands, running and standing water. The trend categories provided by TRIM software are: SI = strong increase; MI = moderate increase; U = uncertain; MD = moderate decline; S = stable.

Species	Changes in distribution		Changes in numbers
			All wetlands		Running water		Standing water
Little Grebe	–0.367	*	0.973 ± 0.002	MD	0.974 ± 0.002	MD	0.961 ± 0.009	MD
Great Crested Grebe	0.182	n.s.	0.982 ± 0.006	MD	0.949 ± 0.010	MD	1.009 ± 0.012	S
Great Cormorant	0.842	**	1.172 ± 0.009	SI	1.194 ± 0.013	SI	1.109 ± 0.011	SI
Great White Egret	0.778	**	1.262 ± 0.064	SI	1.251 ± 0.048	SI	1.268 ± 0.172	U
Grey Heron	0.916	*	1.045 ± 0.002	MI	1.046 ± 0.003	MI	1.043 ± 0.004	MI
Mute Swan	0.734	*	1.018 ± 0.002	MI	1.019 ± 0.003	MI	1.013 ± 0.004	MI
Bean Goose	0.476	*	1.022 ± 0.015	U	1.063 ± 0.018	MI	1.011 ± 0.034	U
White-fronted Goose	0.691	**	1.114 ± 0.016	SI	1.157 ± 0.043	SI	1.081 ± 0.215	U
Greylag Goose	0.554	**	1.160 ± 0.018	SI	1.357 ± 0.284	U	1.139 ± 0.024	SI
Eurasian Wigeon	0.726	**	1.061 ± 0.011	MI	1.062 ± 0.014	MI	1.047 ±0.023	MI
Gadwall	0.619	**	1.163 ± 0.132	U	1.052 ± 0.017	MI	1.321 ± 1.401	U
Common Teal	–0.319	*	0.974 ± 0.003	MD	0.957 ± 0.004	MD	1.005 ± 0.008	S
Mallard	0.576	**	1.008 ± 0.001	MI	1.004 ± 0.001	MI	1.014 ± 0.003	MI
Common Pochard	0.276	n.s.	1.031 ± 0.006	MI	1.045 ± 0.008	MI	0.965 ± 0.012	MD
Tufted Duck	0.685	**	1.092 ± 0.006	SI	1.100 ± 0.008	SI	1.011 ± 0.008	S
Common Goldeneye	0.181	n.s.	1.016 ± 0.003	MI	1.017 ± 0.003	MI	0.995 ± 0.007	S
Smew	0.236	n.s.	1.057 ± 0.009	MI	1.056 ± 0.011	MI	1.053 ± 0.029	U
Goosander	0.672	*	1.021 ± 0.003	MI	1.019 ± 0.004	MI	1.030 ± 0.007	MI
White-tailed Eagle	0.874	**	1.088 ± 0.007	SI	1.084 ± 0.012	SI	1.088 ± 0.009	SI
Common Moorhen	–0.202	n.s.	0.995 ± 0.003	MD	0.996 ± 0.003	MD	0.953 ± 0.008	S
Common Coot	0.013	n.s.	0.992 ± 0.002	MD	0.993 ± 0.002	MD	0.964 ± 0.006	MD
Black headed Gull	–0.251	n.s.	0.996 ± 0.003	S	0.999 ± 0.003	S	0.973 ± 0.008	MD
Mew (Common) Gull	0.336	*	1.052 ± 0.006	MI	1.061 ± 0.008	MI	1.048 ± 0.011	MI
large gulls	0.826	**	1.197 ± 0.014	SI	1.277 ± 0.169	U	1.187 ± 0.019	SI
Common Kingfisher	0.544	**	1.043 ± 0.004	MI	1.041 ± 0.004	MI	1.057 ± 0.013	MI
White-throated Dipper	0.096	n.s.	1.013 ± 0.003	MI	1.012 ± 0.003	MI	1.052 ± 0.031	U

Species-specific variables

We used the following species-specific variables to find ecological factors responsible for analysed changes in numbers and distribution of individual species.

•	Six eco-taxonomic groups were used: fish-eating birds, geese, dabbling ducks, diving ducks, gulls, and others (see Snow & Perrins 1998).
•	Mean body weight was used as a measure of body size (from Snow & Perrins 1998).
•	Population trends in the Western Palearctic, and midpoints of population range in the Western Palearctic, were obtained from Wetlands International (2006). Moreover, estimation of breeding population size and trends in breeding population (Birdlife International 2004) were used for White-tailed Eagle, Common Kingfisher, and White-throated Dipper, whose data are not included in Waterbird Population Estimates (Wetlands International 2006). These three species are breeding as well as wintering in Europe (Snow & Perrins 1998) and therefore total population size and population trends were used from breeding population data (Birdlife International 2004). We expected similar trends in wintering as well as breeding numbers of these species.
•	The geographical distribution of a species was classified using the latitudinal midpoint (Lemoine et al. 2007), i.e. the mean of the southernmost and northernmost latitudes of the species breeding range (Snow & Perrins 1998).
•	Mean numbers and mean distribution for the Czech Republic were obtained from the Czech IWC data containing values from the period 1966–2008 (this study).
•	The conservation status of a particular species was classified using its listing in Annex 1 of the EU Bird Directive (European level) and using the classification of the species under Czech legislation Act No. 114/92 Coll. and Regulation No. 395/1992 Coll., Annex No. III (list of Specially Protected Animals; Hudec et al. 1999).

Statistical analyses

Effects of species-specific variables on individual specieś trends in numbers, and changes in distribution, were tested by forward selection generalized linear models (GLM) for normal distribution with the identity link function in r software (http://www.r-project.org/). Nine predictors entered the models but the final solution included only the best, confined combination of them. Continuous variables were included as a linear (i.e. mean body mass) as well as squared terms (i.e. mean body mass²) predictors; however, no squared variable was significant. Analysis of independent contrasts were made. Therefore, in the results, there is one ‘missing’ group in more than two-group variables. This group is considered as a ‘baseline’ for comparison with other groups; it means that the significances in tables omit the difference toward the baseline group.

RESULTS

Trend in numbers and distribution across all sites

Population trends were recorded for 26 common wintering waterbird species in the Czech Republic between 1966 and 2008 (Table 1). Among those investigated, 18 species were found to be increasing and only five species were recorded as decreasing. The trend of one species was assessed as ‘stable’ and the trends of two species were assessed as ‘uncertain’ (Table 2). Distribution (i.e. the ratio of occupied sites to total number of investigated sites) increased in 16 species and decreased in only two species. No significant changes in distribution were found in eight species during the study period (Table 2).

A significant increase in numbers and range were found in Great Cormorants, Great White Egrets, Grey Herons, Mute Swans, White-fronted Geese, Greylag Geese, Eurasian Wigeons, Mallards, Tufted Ducks, Goosanders, White-tailed Eagles, Common Gulls, ‘large gulls’ and Common Kingfishers. On the other hand, declines were recorded in Little Grebes and Common Teals. There were significant declines in abundance of Great Crested Grebes, Common Moorhens and Common Coots, although their distribution range remained stable. Moreover, an increase in numbers was recorded in Common Pochards, Common Goldeneyes, Smews and White-throated Dippers, although their distribution range remained stable. On the other hand, no trend (i.e. ‘uncertain trend’) in numbers was found in Bean Geese and Gadwalls, whose wintering range significantly increased. The only ‘stable’ species, without significant change in either numbers or distribution, were Black-headed Gulls (Table 2).

Changes in numbers (indicated by the multiplicative rate of change of values) generally correlated with changes in distribution (correlation coefficient) (Spearman Rank Correlation: r _s = 0.777, n = 26, P < 0.0001; Fig. 2).

Figure 2. Relationship between changes in distribution (correlation coefficient describing trend in the ratio of the number of sites occupied to sites counted) and changes in numbers (multiplicative rate of change).

Changes in number of waterbirds on standing and running waters

Population trends of wintering waterbirds were analysed separately for two main wetland habitats, standing and running waters. Overall, nine increasing, four decreasing, and five stable species were found on standing water, and 18 increasing, four decreasing, and one stable species were recorded on running water. Trends in numbers of individual species in running and standing water were consistent with the overall trends for all species (Table 2, Fig. 3). Exceptions were shown only on standing water and included Great Crested Grebes, Common Teals, Tufted Ducks, Common Goldeneyes and Common Moorhens, whose trends were stable, and Common Pochards and Black-headed Gulls, whose trends on standing water declined (Table 2).

Figure 3. Relationship between changes in numbers (multiplicative rate of change) on standing and running waters. Only 18 species with increasing, decreasing, or stable trends in both habitats are included.

The number of species with an uncertain trend was higher on standing water (6) than on running water (2) (Table 2). The species with uncertain trends include those which use one of the habitats only marginally (e.g. White-throated Dippers on standing water and Greylag Geese on running water). Therefore, we restricted comparison of changes in numbers on standing and running waters to 18 species with increasing, decreasing, or stable trends in both habitats. We found that multiplicative rate of changes in numbers in both habitats were generally correlated among the waterbird species analysed (Spearman Rank Correlation: r _s = 0.716, n = 18, P < 0.001; Fig. 3).

Effect of species-specific variables

Effects of species-specific variables on trends in wintering waterbirds numbers and on changes in their distribution were analysed by forward selection GLMs.

Changes in numbers reflect trends in the Western Palearctic in increasing species but not in decreasing ones. Moreover, changes in numbers were significantly higher in rare species which occupy a fewer wetland sites (Table 3).

Table 3. The forward selection general linear model for changes in numbers (multiplicative rate of change) of bird species in all investigated wetlands (F = 3,718 on 5 and 20 DF, P = 0.015). The model explains 35.22 % of variability in the data set. For all variables entered in the analysis see Methods, species-specific variables.

	Estimate	se	t-value	P
Intercept	1.095	0.025	43.122	<0.000
Western Palearctic population trend
Increasing	0.095	0.033	2.905	0.008
Decreasing	–0.035	0.033	–1.071	0.297
Wintering numbers in Czechia	0.000	0.000	1.930	0.068
Wintering distribution in Czechia	–0.002	0.001	–2.740	0.013
Body size	0.000	0.000	–1.642	0.116

Significant differences in changes in distribution among eco-taxonomic groups were found. An increase took place among fish-eating birds, geese, dabbling and diving ducks, and gulls (Figs. 4a,b). Decreases in distribution were more significant in species which have been decreasing in the whole of the Western Palearctic (Figs. 5a,b). Conservation status in the Czech Republic was also correlated with changes in distribution: increases in distribution were more frequent in non-protected species. Surprisingly, conservation status at the European level (listing in Annex 1 of the EU Bird Directive) was not correlated with changes in the numbers and distribution of particular species. Moreover, the wintering distribution (mean number of occupied sites in the Czech Republic), wintering numbers (mean numbers), body size, and geographical distribution of the species (latitudinal midpoint of the breeding range) was not correlated significantly with a species trend in numbers or changes in distribution (Table 3 & 4).

Table 4. The forward selection general linear model for changes in distribution (correlation coefficient between arcsin-transformed ratio of occupied size and year) of bird species in all investigated wetlands (F = 5.199 on 9 and 16 DF, P = 0.002). The model explained 60.19 % of variability in the data set. For all variables entered the analysis see Methods, species-specific variables.

	Estimate	se	t-value	P
Intercept	–0.078	0.132	–0.595	0.561
Western Palearctic population trend
Increasing	–0.147	0.156	–0.940	0.361
Decreasing	–0.487	0.165	–2.946	0.009
Body size	0.000	0.000	1.266	0.224
Eco-taxonomic groups
Fish-eating birds	1.129	0.187	6.049	<0.001
Geese	0.720	0.194	3.717	0.002
Dabbling ducks	0.647	0.177	3.655	0.002
Diving ducks	0.594	0.175	3.396	0.004
Gulls	0.728	0.235	3.096	0.001
Species protection (Czech legislation)	–0.340	0.136	–2.502	0.024

Figure 4. (a) Changes in numbers (multiplicative rate of change) among waterbird groups. (b) Changes in distribution (correlation coefficient describing trends in the ratio of the number of sites occupied to sites counted) among waterbird groups.

Figure 5. (a) Relationships between changes in numbers (multiplicative rate of change) in the Czech Republic and changes in numbers in the Western Palearctic (Wetlands International 2006). (b) Relationships between changes in distribution (correlation coefficient describing trend in the ratio of the number of sites occupied to sites counted) in the Czech Republic and changes in numbers in the Western Palearctic (Wetlands International 2006).

DISCUSSION

Numbers and distribution of most common wintering waterbird species changed significantly in the Czech Republic between 1966 and 2008. The proportion (88%) of waterbird species whose trends in abundance were significant seems to be higher than in Europe as a whole, where changing (decreasing and increasing) waterbird populations represented 48% of all the species populations recorded (Wetlands International 2006). Although changes in numbers and distribution of particular waterbird species in the Czech Republic often follow the trends in the European populations, increasing trends are more frequent (18 out of 26 species) in this country. Trends in abundance were shown to be correlated to, and possibly affected by, the status of that particular species in the Western Palearctic, and they were also affected by the suitable wintering conditions in the Czech Republic.

In comparison to general European trends (Wetlands International 2006), changes in waterbird wintering numbers in the Czech Republic can probably be explained by a northward shift in the wintering range of many species, due to climatic changes (Ridgill & Fox 1990, Keller 2006, Maclean et al. 2008, Musil et al. 2008a). However, a comparison of population trends between countries is limited by differences in trend period, species included and methods used. Moreover, we can omit species whose distribution is related (especially in the wintering period) to coastal habitats (for example, most waders (shorebirds), seaducks and several goose species), as these occur only in low numbers in inland European countries such as the Czech Republic. Comparable trend analyses covering more than 10 waterbird species are available from, for example: Slovakia (1991–2006, Slabeyová et al. 2008, 2009); Sweden (1967–2006, Nilsson 2008); Ireland (1994–2004, Crowe et al. 2008); the UK (1966–2009, Calbrade et al. 2010); the Netherlands (1976–2008, Hustings et al. 2009); France (1987–2008, Fouque et al. 2009); Bulgaria (1977–2001, Michev & Profirov 2003). Despite methodological differences, we can compare the proportion of species with increasing, decreasing, or no (i.e. stable, uncertain, unknown, fluctuating) trends. In most of these countries (Sweden, Czech Republic, Slovakia, UK, the Netherlands and Bulgaria), the number of increasing species was higher than the number of decreasing species. However, increasing species were in the majority among all investigated species only in Sweden (13 of 17 species, Nilsson 2008), France (12 of 17 species, Fouque et al. 2009), the Netherlands (16 of 28 comparable species, Hustings et al. 2009) and the Czech Republic (18 of 26 species, i.e. this study). Decreasing species were in the majority only in the Irish study (Crowe et al. 2008). Surprisingly, increasing species were more dominant among investigated species than decreasing ones in other western European countries (see above). Some slight north–south differences can be also seen by comparing the Czech data (69% of increasing species) with Slovakia, where only 41% of species were increasing in wintering numbers.

In general, the changes in numbers and distribution of waterbird species in the Czech Republic were consistent with species population trends in the Western Palearctic. However, differences in trends between the Czech Republic and the numbers along recognized flyways were shown for some species. Numbers and distribution of Common Teals have declined significantly in the Czech Republic, whereas this species has increased in other European countries (see for example, Wahl & Sudlfeldt 2005, Fouque et al. 2009, Calbrade et al. 2010). The decrease of Common Teals in the Czech Republic may be related to the negative impact of intensive fishpond management (see, for example, Musil et al. 2001, Musil 2006), which has affected the breeding, migrating and also the wintering numbers of this species. On the other hand, the numbers of three species (Bean Geese, Common Pochards and Mew (Common) Gulls) have decreased in the Western Palearctic flyway (Wetlands International 2006), whereas their numbers and/or distribution have increased in the Czech Republic. Among these species, Common Pochards and Mew Gulls have expanded their breeding and wintering range, not only in the Czech Republic but also in other central European countries (e.g. Slovakia, Slabeyová et al. 2008). The increasing distribution of Bean Geese in the Czech Republic may be related to a shift in wintering sites within Europe (Madsen et al. 1999, Wetlands International 2006, Fox et al. 2010). Nevertheless, the trend in the wintering numbers of this species has been classified as ‘uncertain’, due to its fluctuations in relation to varying mid-winter temperature (Musil et al. 2008a). Mallards are another interesting species whose numbers and distribution have increased in the Czech Republic, whereas there has been a decrease in wintering numbers reported from northwest and northeast Germany (Wahl & Sudlfeldt 2005), Slovakia (Slabeyová et al. 2008, 2009) and also from Ireland (Crowe et al. 2008) and the UK (Calbrade et al. 2010). Numbers of Mallards have been affected by the long-term releasing of this species for hunting purposes in the Czech Republic, especially since 1990 (Musil et al. 2001). Between 2003 and 2007, numbers of released Mallards reached 150 000–220 000 individuals (unpubl. data).

Changes in numbers generally correlated with changes in distribution. Nevertheless, the magnitude of the increase in the range of some species (for example, Grey Herons, Eurasian Wigeons, Mallards, Goosanders and White-tailed Eagles) was high compared with other species, indicating that they had the highest rate of expansion over the study period. These species seem to have become more widespread, as opposed to being concentrated in relatively few sites. On the other hand, Great Crested Grebes, Common Moorhens and Common Coots are decreasing species which do not exhibit significant changes in distribution.

Running water (streams, rivers) and standing water (including fish ponds, reservoirs, gravel and sand-pit lakes, and industrial settling ponds) represent the two main categories of wetland habitats (Chytil et al. 1999) available for wintering birds in the Czech Republic. Although the changes in numbers in these two habitat types were generally consistent among the waterbird species analysed, there were noticeable differences in changes in numbers related to habitat. Species with a highly significant trend (increasing or decreasing) in numbers throughout the entire Czech Republic changed their numbers similarly in both habitat types. In six species, we found differences in trends on standing and running waters which can be related to the habitat preferences of the species (Hudec 1994, Snow & Perrins 1998, Delany et al. 1999, Musil et al. 2001, 2008b, Gilissen et al. 2002). Among these species, numbers of diving ducks (Common Pochards, Tufted Ducks, Common Goldeneyes) increased on rivers and conversely were found to be decreasing or stable on standing water. Those species probably avoid intensively-managed fish ponds affected by the intensive grazing effect of Carp Cyprinus carpio stocks (Musil et al. 2001). Nevertheless, an increase in the importance of non-freezing standing water for other species (e.g. Great White Egrets, Greylag Geese, Gadwalls, Mallards, large gulls) can be expected in the coming years, in accordance with the global climate change forecasts which predict milder winters across Europe, including the Czech Republic (Huntley et al. 2007, IPCC 2007).

The changes in wintering numbers and distribution of particular species were affected by many species-specific variables. Among these, species trends in the Western Palearctic (Birdlife International 2004, Wetlands International 2006) were the most significant. Amongst the various eco-taxonomic groups, the most remarkable increases in numbers and distribution were shown by the fish-eating birds, followed by the geese, dabbling ducks and gulls. This pattern of change is similar to the published data for population changes in Europe as a whole (Wetlands International 2006, Birdlife International 2004).

We found that the rate of change in distribution was lower in species protected under Czech conservation law. These species were listed as specially protected because of their long-term decline in numbers (Musil et al. 2001). However, negative trends shown in these species (e.g. Little Grebes and Common Teals) were not reversed by protection measures imposed under Czech conservation law. Changes in the size of the breeding population of 189 bird species were analysed using data from Atlases of Breeding Distribution in the Czech Republic (Št'astný et al. 2006) between 1985 and 1989 and between 2001 and 2003. Although increasing trends are prevailing among specially protected species, their population trends have not been reversed since the law providing for their conservation at the national level. Increase in numbers continued in species which were increasing before the adoption of conservation laws. Likewise, species which were decreasing before this adoption continued to decrease (Vořísˇek et al. 2008).

Finally, we found that the trends in wintering numbers were negatively correlated with wintering distributions, i.e. number of occupied wetlands. This phenomenon can be explained by the northwards expansion of wintering waterbirds in central Europe, similar to the changes recorded in western or northern Europe (Maclean et al. 2008, Nilsson 2008), or anticipated for breeding populations in most of Europe (Huntley et al. 2007), in response to the expected climatic changes (IPCC 2007). Nevertheless, decrease in wintering distances, as well as shift of range were found to be more significant in terrestrial than wetland bird species (cf. Reif et al. 2008, Visser et al. 2009). Although the effect of climatic changes on numbers and distribution of bird species are often discussed in many papers, relevant studies analysing this phenomenon seem to be scarce.

ACKNOWLEDGEMENTS

We are very grateful to all the volunteers who were involved in waterbird counts and also to the census co-ordinators of the International Waterbird Census (IWC) in the Czech Republic (Bohuslav Urbánek, Vladimír Fiala, Čestmi´r Folk, Josef Křen, Ivana Kozˇená, Jitka Pellantová). IWC in the Czech Republic was organized in cooperation with the Czech Society for Ornithology. We are also grateful to Steve Ridgill and Lucie Fuchsová for language improvement. We thank Tony Fox (NERI, Denmark) for useful comments to earlier versions of the manuscript and Leo Soldaat (Centraal Bureau voor de Statistiek, The Netherlands) for useful advice on trend analysis. We are also grateful to anonymous referees for useful comments which helped to improve our manuscript.

This study was supported by the Ministry of Environment of the Czech Republic, Project VaV MŽP ČR SP/2d3/109/07 entitled ‘The long-term changes in numbers and distribution of waterbirds in the Czech Republic in relation to climatic and environmental changes’.