• Monitoring of hazel grouse (Bonasa bonasia) distribution range and population trends in the swiss Jura Mts.
• Build GIS-based habitat suitability maps.
• Examine population dynamics with TetrasPool.
• Provide management support for hazel grouse.

• Hazel grouse group created in 1998.
• No official structure, registration is free.
• Over 18 members collect data on hazel grouse in the swiss Jura Mts. Members are wildlife rangers, foresters, ornithologists, hunters and biologists.
• Everyone interested in collecting/sending data on hazel grouse in Switzerland could register.

• Intensive field survey will be ended at the end of this spring. Up to now, 600 hazel grouse signs have been collected. Most signs consists in droppings or tracks in the snow. Direct observations remain rare events.

Grouse data

• Field work was conducted during three winters (December-March), 1998/99, 1999/2000, 2000/01. We performed snow trailing investigations 10 days after fresh snowfall to get sufficient detections.
• Birds presence was assessed by searching for winter droppings lying under feeding trees or shrubs. Winter Hazel grouse droppings have 6mm radius for both sexes (Leclercq, 1981).

• Between 1998-2001, we prospected 100 kilometric squares but recorded habitat parameters only on plots were both grouse species were present. Finally, 140  hazel grouse feeding sites over 70 squares kilometers were sampled.
• For each square kilometer, sampling design rely on two hazel grouse droppings and two random plots. This design increases the probability of data independance while hazel grouse has small area requirements with winter territories down to 0.2km2. Every square kilometer plot was explored radially from its centre towards the borders until a grouse dropping was discovered.
• To standardize research effort, grouse signs were search during maximum one hour in each kilometric plot. Hazel grouse is a forest species, areas outside forests were excluded from randoms points. At these points, we checked for the absence of the species by searching for droppings following the same protocol as in occupied plots. A presence sign was never discovered in the 140 random points.

Habitat data

• For each grouse sign and random points, we recorded 25 habitat variables including topographical features, structural and compositional stands structure. 
• We used 10m diameter plots centered on each dropping or at random points locations for variables estimation. Grouses plots locations were recorded on a GPS at +-10m and position of random points were located in the same way. 
• Topographical features measured were altitude, slope and exposure. 
• Habitat structure was assessed by vegetation cover for each of the three lignous layers. Cover estimation were made visually and estimations are at the nearest 5%. Species with low cover were assigned a value of 1%.
• Presence/absence data were randomly partitionned into two equal-sized calibration and validation datasets.

Results

• GLM models should assess habitat suitability at the local scale. As data collection is an ongoing process unitl this spring, further analysis will be conducted by a diplome work at the University of Lausanne (summer-winter 2001).

Summary

• Hazel grouse is an endangered species in the mountaneous forests of central Europe. In Switzerland, capercaillie abundance and distribution strongly decrease during the last decade. In this study, we use the Ecological Niche Factor Analysis (Hirzel et al. 2000) for predicting hazel grouse habitat suitability from indirect  presence data issued from the Jura Montains, western Switzerland, in 1995-2000.
• Habitat was described with land-use databases including topography, vegetation and human infrastructures for each one hectare plot with hazel grouse presence and plots with unknown hazel grouse presence in study site. Human disturbance was investigated indirectly by integration of roads, railways, houses and farms networks into ENFA.
• Hazel grouse observation points were expanded according to hazel grouse home range size and randomly assigned to calibration and validation maps.
• Habitat suitability is high for most forest between 1100-1550m a.s.l. while areas between 500-1100m are irregularly suitable (Figure 1). These low elevated areas are majoritarly unsuitable.
• Model fiability was strong as predicted suitability was elevated by comparison to the validation dataset. Highly significant bootstrap resampling confirmed strong accuracy of the predicted suitability map.
• Overall, it appears that hazel grouse occurrence was a result of several interacting factors. At the landscape scale, hazel grouse presence probability increased with forests availability, mainly forests with elevated canopy cover. The bird avoided meadows and similar habitats where trees were rare or absent while human infrastructure do not affect hazel grouse occurrence probability.
• Model results suggest that effective management plans should include actions on habitat structure to increase carrying capacity. By contrast with capercaillie, human disturbance reduction should not be set as a priority in hazel grouse habitats.