Lynx in the Šumava Mountains
183
national parks, the Šumava (680 km2) and the Bavarian Forest National Park (230 km2) form the
center of a large protected area. They are surrounded by two landscape protected areas (Šumava
Biosphere Reserve 790 km2, Bavarian Nature Park 2680 km2). The summit of the Arber reaches an
elevation of 1457 m above sea level. Forest cover is more than 90%, and human population density is
low (20 inhabitants/km2). For our analyses we included the foothills and adjacent mountain ranges
namely, the Šumava Foothills, the Blanský Les, the Novohradske hory, the Outer Bavarian Forest, the
Èeský Les, the Fichtelgebirge, and the Slavkovský Les. In these areas, forest cover varies from 60 to
80%, and human density reaches 100 inhabitants/km2. The climate is continental with some maritime
influence from the west. Mean temperature in the center of the study area is 3.5 to 6.5°C, with an
annual precipitation varying from 2500 mm (west) to 400 mm (east). Permanent snow cover lasts up to
7 months. Other large predators such as wolf Canis lupus and brown bear Ursus arctos are not present
in the area. Main prey species for the lynx are roe deer Capreolus capreolus. Other ungulate species
such as red deer Cervus elaphus, wild boar Sus scrofa, and mufflon Ovis ammon also occur in the area.
Methods
Data sources and quality
We collected data for the Czech Republic by reviewing available literature, unpublished records of
questionnaires distributed to various regional hunting grounds, results of snow tracking and
incidental sightings, tracks, killed prey or lynx found dead (for detailed description see Èervený and
Bufka 1996, Bufka 1997, Èervený et al. 1997). In Austria only incidental findings were recorded (W.
Proksch, pers. comm.). In Germany incidental findings were gathered (Poost 1996, Wölfl 1996, 1997,
Habel 1997). Since 1994 data have been collected using snow tracking and questionnaires as well (for
details refer to Poost 1996, Heurich 1997, Kiener 1997).
When approaching our data set, we have to take into consideration several aspects linked with the
process of data gathering and public relation work in the region. First, only few of the data were
verified. We estimate verification of data to be about 30% in Czech Republic, 30% in Germany and
about 20% in Austria. Especially data gathered by questionnaires generally could not be verified.
Second, in the beginning of the nineties only minimal efforts were made to gather data. Only few
people knew of lynx presence, and even fewer could recognize field signs of lynx. Coupled with more
intense public information, data were collected on a larger scale during the next few years. For
example, the questionnairies sent to hunters in Czech Republic since 1994 covered more districts each
year. Therefore, our data likely overestimate the rate of increase of the lynx population. Moreover,
data flow is dependent on other variables. The main problem with our data set is that it is not possible
to measure the searching effort. Therefore, it is difficult to relate data quality to any population trends
because we lack any index linking the amount of data to searching effort (for example, numbers of
tracks found per km trackline searched, numbers of lynx observed per hour observation time). That is
why a change in the yearly amount of data is not necessarily linked with a change in population
numbers but may be related to other factors as well. For example, weather conditions, especially snow
cover and snow conditions, strongly influence our findings of lynx presence. Additionally, the
motivation level of people living in the area to cooperate with researchers is changing. This depends on
factors such as mutual trust or hunting practice on roe deer. Lastly, the effort researchers make to
procure data varies over space and time. In Austria, for example, a network for systematically
gathering lynx information has only recently been installed. In spite of the shortcomings stated above
our data are the best available. Because most of the data were not classified into different reliability
categories, a splitting into different quality groups was not feasible.
Data plotting and processing
For plotting data in the Czech Republic, we used the RFME grid system with a square size of
approximately 11.2 by 12 km (Èervený et al. 1997). For Germany and Austria, data were transcribed
into 10 ´ 10 km squares according to the Gauß-Krüger-grid system. For estimating population trends
the period was divided into yearly intervals from 1990 to 1998. For each interval and square, lynx