Introduction
As part of the JMMB program, trends of 34 waterbird species for the international Wadden Sea and the four regions - The Netherlands, the Federal States of Germany, Niedersachsen/ Hamburg and Schleswig-Holstein, and Denmark will be presented.

Data and methods
Data used in the analyses are a mixture of total counts (two internationally, up to five nationally) and counts of a selection of sites which are counted more frequently (12-25 times a season). At present a total of 594 counting units are defined in the Wadden Sea, which are included in the analyses. For this report, the original counting data, available at the smallest level have been used.
Trends are calculated and presented for 34 waterbird species. These are species which use the Wadden Sea during stop-over on migration or as a wintering area with large parts of their flyway population. Species which only occur in low numbers or species which cannot be counted with sufficient representativeness have been excluded from the analyses (for a more detailed explanation see Rösner et al., 1994).

Despite a large dataset with lots of real count data available also missing counts are present. A complete dataset involves counts for all counting units in all months of the year.To analyse the waterbird count data, UINDEX (Bell, 1995) was used to account for missing counts in the dataset, and then TrendSpotter is applied to calculate trends (Visser, 2004, Soldaat et al 2007). The program UINDEX is estimating bird numbers for missing counts (imputing) taking into account site- ,year- and month-factors (Underhill & Prys-Jones 1994). Sites are grouped in four regional strata representing the four different Wadden Sea “countries”. The counted and imputed values for each month are added to yearly averages for the respective “bird-years”, covering the period from July to June of the following year (Figure 1). After that with the program TrendSpotter so-called “flexible trends” are calculated. These are particularly suitable for time series data with different periods of decreasing, stable or increasing trends (Visser 2004, Soldaat et al., 2007). A trend line calculated by TrendSpotter hardly deviates from a moving average or a smoothed trend line as calculated by a Generalized Additive Model (GAM) (e.g. Atkinson et al., 2006). TrendSpotter calculates also confidence intervals and differences between the trend level of the last year and each of the preceding years can be assessed (Soldaat et al. 2007). This way trend estimates can be given for any period, as for example the last 10 years and the whole time period, as in the current analyses.


Figure 1. Example of the treatment of data for the trend analyses. First the seasonal pattern is reconstructed by using counted numbers and imputed numbers for each month for a certain species (left graph of the figure, dark blue is counted, light blue is imputed). Than the average over all months is taken and this is the ‘yearly estimate’ to be used in the trend analyses (right graph). The trend line and confidence limits are calculated over all year estimates.

Acknowledgements
In Denmark the counts were carried out by the National Environmental Research Institute (NERI, University of Aarhus). Aerial counts were carried out by NERI up to 1992, and during the years after they were organized through a collaboration between NERI and Ribe Environmental Center, Ministry of the Environment.
In Schleswig-Holstein the monitoring was initiated by the Ornithological Society Schleswig-Holstein (OAG SH) in the 1960s; regular monitoring was jointly organized by the OAG SH and the World Wide Fund for Nature (WWF) in 1987 and during the first period until 1994 funded by the federal state Schleswig-Holstein and the Federal Ministry of Environment (Federal Environment Agency) as part of an ecosystem research project. Since then it was funded by the National Park Administration Schleswig-Holstein Wadden Sea. The coordination of the project moved from WWF to the Schutzstation Wattenmeer e.V. in 2004. The aerial surveys of Common Eider and Shelduck were separately financed by the National Park Administration Schleswig-Holstein Wadden Sea.
In Niedersachsen and the Hamburg regions the counts were organized by the Bird Conservation Station in the Lower Saxony Water Management, Coastal Defence and Nature Conservation Agency (NLWKN), formerly Lower Saxony Agency for Ecology (NLÖ). The aerial surveys of Common Eider were financed by the Lower Saxony Wadden Sea National Park Authority.
The waterbird counts in the Dutch Wadden Sea are part of the national monitoring program of waterbirds in The Netherlands, which is a cooperation between the Ministry of Agriculture, Nature and Food Quality, the Ministry of Water Management and Public Works, Statistics The Netherlands (CBS), Vogelbescherming Nederland and SOVON Dutch Centre for Field Ornithology. The aerial surveys of Common Eider were carried out under the responsibility of the Ministry of Water Management and Public Works.