|
Since the avian species richness of a physiographic unit is directly related to the atlassing effort, these two variables were compared using a nonlinear regression model. The value of the residual calculated for each physiographic unit can be used to identify the units in which there is a difference between the observed richness and predicted richness of breeding bird species. Some units had richness values much greater than that predicted by the model, which indicates that, for a given atlassing effort, a greater number of breeding species will be detected than in units where the value of the residuals is negative. There was no relationship between the residuals and observed richness in breeding birds in each unit (r² = 0.083), which indicates that calculating the residuals allows the bias associated with atlassing effort to be eliminated. In addition, the value of the residuals was not related to field hours (r² = 0.011), the total area of units (r² = 0.009), or the area surveyed (r² = 0.0004; number of squares from the Breeding Bird Atlas surveyed in the units).
Physiographic units with positive (n = 42) and negative (n = 26) residuals occurred in all the regions of the St. Lawrence. Units with the highest positive residuals were located in the Carleton, Montreal and Bergeronnes regions, while units with species richness much lower than predicted by the model (negative residuals) were located in the Sainte-Marie Islands, Trois-Rivières and Brion Island sectors.
|
Table 1. Physiographic units with the highest and lowest residuals |
|
Physiographic unit |
Natural province |
Residual |
|
Carleton |
Appalachians |
34,7 |
|
Montréal |
St. Lawrence Lowlands |
30,9 |
|
Bergeronnes |
Laurentians |
24,4 |
|
Pointe-du-Lac |
St. Lawrence Lowlands |
21,8 |
|
Grondines |
St. Lawrence Lowlands |
21,5 |
|
Cap Tourmente |
Laurentians |
21,1 |
|
New Richmond |
Appalachians |
19,7 |
|
Ragueneau |
Laurentians |
18,7 |
|
Contrecoeur |
St. Lawrence Lowlands |
18,0 |
|
Les Escoumins |
Laurentians |
17,4 |
 |
|
Mingan Archipelago |
Lower North Shore Plateau |
-13,7 |
|
Anticosti-East |
Anticosti Island |
-15,4 |
|
Matamec River |
Lower North Shore Plateau |
-15,5 |
|
Anticosti-North |
Anticosti Island |
-16,8 |
|
Mont-Joli-South |
Appalachians |
-17,3 |
|
Mont-Sainte-Anne |
Laurentians |
-18,5 |
|
Sept-Îles |
Laurentians |
-26,9 |
|
Brion Island |
Magdalen Islands |
-31,8 |
|
Trois-Rivières |
St. Lawrence Lowlands |
-37,5 |
|
Sainte-Marie Islands |
Lower North Shore Plateau |
-54,2 |
Units favourable to breeding birds
The clusters of physiographic units that were the most favourable to breeding birds were identified for each of the ten typologies developed, by disregarding the atlassing effort used to confirm breeding. Points were awarded to units to calculate an index identifying the units most favourable to breeding birds (Table 2). For example, the mean residuals for units with a "hot" or "cool" climatic context were quite high in comparison with those with a "cold" climatic context, with a significant difference between the means (P < 0.05). Therefore, the units belonging to these clusters received 2, 2 and -2 points respectively for the typology.
|
Table 2. |
Mean residuals by cluster of physiographic units for each typology. The number following the name of the cluster represents the number of units in the cluster |
Climatic
contexts |
Cluster |
Mean |
Points |
|
Bioclimatic
domains |
Cluster |
Mean |
Points |
| Hot (17) |
5,9 |
2 |
Maple (25) |
5,4 |
1 |
| Cool (31) |
4,2 |
2 |
Balsam fir (33) |
2,4 |
1 |
| Cold (20) |
-5,2 |
-2 |
Black-spruce-tundra (10) |
-9,0 |
-1 |
| |
p = 0,0353 |
|
|
p = 0,0770 |
|
|
Coastal
landscapes |
Cluster |
Mean |
Points |
|
Land use (NOAA) |
Cluster |
Mean |
Points |
| Meandering-flat (13) |
11,6 |
2 |
Deciduous (22) |
8,1 |
1 |
| Straight-rugged (19) |
3,8 |
0 |
Agricultural-
heterogeneous (13) |
3,4 |
0 |
| Jagged-rugged-bay (8) |
-5,0 |
-2 |
Field crops (12) |
2,7 |
0 |
| Straight-cliff (15) |
-5,2 |
-2 |
Conifer (18) |
-4,7 |
-1 |
| |
p = 0,0044 |
|
|
p = 0,0918 |
|
|
Aquatic
typology |
Cluster |
Mean |
Points |
|
Land use (LANDSAT) |
Cluster |
Mean |
Points |
| Estuary (15) |
8,5 |
0 |
Mixed (14) |
8,4 |
0 |
| Freshwater (13) |
4,6 |
0 |
Agricultural-
urban (18) |
4,0 |
0 |
| Gulf-strong-tides (17) |
0,9 |
0 |
Deciduous (10) |
2,9 |
0 |
| Gulf-weak-tides (13) |
-5,1 |
0 |
Softwood (23) |
-2,1 |
0 |
| |
p = 0,1678 |
|
|
p = 0,2520 |
|
|
| Drainage |
Cluster |
Mean |
Points |
|
Fragmentation (NOAA) |
Cluster |
Mean |
Points |
| Imperfect (12) |
3,9 |
0 |
Slightly
fragmented (19) |
3,0 |
0 |
| Rapid (36) |
3,1 |
0 |
Moderately
fragmented (38) |
2,1 |
0 |
| Good (7) |
-7,8 |
0 |
Highly
fragmented (11) |
-1,2 |
0 |
| |
p = 0,6945 |
|
|
p = 0,7747 |
|
|
| Terrain type |
Cluster |
Mean |
Points |
|
Fragmentation (LANDSAT) |
Cluster |
Mean |
Points |
| Flat (29) |
5,5 |
0 |
Highly
fragmented (7) |
12,6 |
0 |
| Rolling (13) |
0,0 |
0 |
Slightly
fragmented (27) |
2,9 |
0 |
| Rugged-hilly (11) |
-2,2 |
0 |
Moderately
fragmented (33) |
-0,3 |
0 |
| |
p = 0,3210 |
|
|
p = 0,1171 |
|
|
* Number in parentheses represents the number of units in the cluster |
Differences in the means for the clusters of units were observed in only four of the ten typologies analysed; three of these four variables (climate, bioclimatic domain, and NOAA land use) showed a gradient of values at the scale of the river. The North Shore region was found to be characterized by a cold climatic context and softwood-dominated vegetation, characteristics that are not very favourable to breeding bird richness, unlike those of southern regions with a more favourable climate. Physiographic units dominated by hardwoods, identified by analysing NOAA images, are favourable to breeding birds, unlike units where field crops, agricultural areas and conifer forests dominate the landscape. Lastly, units with a sinuous coastline, low relief, and a mainly muddy-sandy foreshore are favourable to breeding birds, since this type of coastline is conducive to the formation of coastal marshes that are used by many species of breeding birds. Conversely, units characterized by steep relief along the upper shore and where muddy foreshores were more or less absent do not favour breeding birds.
The index calculated from the points awarded to the clusters of units for each of the ten typologies allowed us to identify units with landscape contexts that are the most favourable for breeding birds. Therefore, units with a high index value were favourable to breeding, compared with units associated with clusters less favourable to breeding birds (low mean residuals), which had a low index value.
This index allows the relative importance of physiographic units within the same region to be compared, taking account of the disparities between units in their respective landscape contexts. Therefore, a unit located on the north shore of the Gaspé Peninsula is not very favourable to breeding due to the harsher climate and presence of cliffs, compared with units in Chaleur Bay, where the climate is less harsh, the terrain flatter, and the shoreline more sinuous. Indeed, the four physiographic units in Chaleur Bay had the highest indices (Table 3). Similarly, units in the St. Lawrence Lowlands east of Trois-Rivières were less favourable to breeding birds, perhaps because of the tides, which limit the establishment of beds of aquatic vegetation used by many species and because of the steep relief on the river's south shore (in the Lotbinière region, for example). High indices calculated for units in the Upper North Shore reflect the flat relief and cool climate in that region, the presence of salt marshes no doubt contributing to the establishment of a diverse breeding avifauna. The regions with landscapes the least favourable to breeding birds were in the Middle and Lower North Shore and Anticosti Island, due to the cold climate and coniferous forest cover.
|
Table 3. |
Physiographic units with the highest indices
|
|
Physiographic unit |
Natural province |
Index |
|
New Richmond |
Appalachians |
6 |
|
Carleton |
Appalachians |
6 |
|
La Grande Rivière |
Appalachians |
6 |
|
Paspébiac |
Appalachians |
6 |
|
Pointe-du-Lac |
St. Lawrence Lowlands |
5 |
|
Bécancour |
St. Lawrence Lowlands |
5 |
|
Montréal |
St. Lawrence Lowlands |
5 |
|
Lake Saint-François |
St. Lawrence Lowlands |
5 |
|
Lake Saint-Pierre |
St. Lawrence Lowlands |
5 |
|
Trois-Rivières |
St. Lawrence Lowlands |
5 |
|
Grondines |
St. Lawrence Lowlands |
5 |
|
Beauport |
St. Lawrence Lowlands |
5 |
|
Contrecoeur |
St. Lawrence Lowlands |
5 |
|
Saint-Siméon |
Laurentians |
4 |
|
Bergeronnes |
Laurentians |
4 |
|
Les Escoumins |
Laurentians |
4 |
|
Ragueneau |
Laurentians |
4 |
|
Pointe-des-Monts |
Laurentians |
4 |
Although the Lower North Shore region does not provide very favourable habitat for a number of breeding birds, the index values show that the physiographic unit near Blanc-Sablon is more favourable than the other units in the region, even in the absence of detailed information on the breeding avifauna of this area. The less jagged shoreline and presence of muddy-sandy foreshores probably help to increase the attractiveness of this unit to breeding birds.
A relationship (r² = 0.385) was found between the breeding bird richness observed in each physiographic unit and the value of the index calculated to identify units favourable to breeding birds. This shows that the units identified as favourable to breeding birds tend to be the same as those units with high observed species richness, despite the fact that the atlassing effort to confirm breeding was eliminated in the residuals analysis process. The high breeding bird richness observed in the units in the southern part of the St Lawrence can probably be explained in part not only by the atlassing effort in this area, but also by the landscapes favourable to the species that occur in these units. We can therefore assume that the results of the analyses presented in the preceding sections (priority species, population trends, rarity, restricted species) were less biased by atlassing effort than previously thought.
Benoît Jobin and Jean-Luc DesGranges
|
|
