- Climate Change Bird Atlas

Current Distribution Maps
Climate Change Scenario Maps
Predictor Maps

Current Distribution Maps

The legend shows the incidence values (IV) of the species as calculated from the North America Breeding Bird Survey data (BBS). The BBS Incidence (on a scale of 0 – 1) represents the proportion of times a species was observed on a route over a decade. The value is the bases for modeling and projecting bird incidence across the eastern US.

Climate Change Scenario Maps

The abundance change maps provide a pairwise view of a suite of combinations of current (1961-1990 baseline) conditions vs. future (2071-2100) General Circulation Models (three different models and their "average") and two emission scenarios (B1, a 'Lo'- relatively high conservation emission scenario and A1fi, a 'Hi' - keep increase fossil fuel emissions). The pull-down menus allow any combination, and the animation button allows a flashing (on/off) view between selected scenarios.

The legend shows the incidence values (IV) of the species as calculated from the North America Breeding Bird Survey data (BBS). The BBS Incidence (on a scale of 0 – 1) represents the proportion of times a species was observed on a route over a decade. The value is the bases for modeling and projecting bird incidence across the eastern US.

Potential Changes in Abundance and Range (Future)

Percent Area Occupied: The percent of 20x20 km cells within the total area of the eastern United States that have been modeled to be suitable for the species. There are a total of 9767 cells used in the model.
Ave. IV: The average importance value (IV) across all 20x20 km cells that have been modeled to be suitable for the species.
Sum IV: The sum of importance values (IV) across all cells. Thus it is a metric that considers both the abundance and the range of the species, perhaps the best metric of overall species importance.
Future/Current IV: [Only applies to Projected Future Habitat] The ratio of change between the modeled current Sum IV and any of the GCM scenarios Sum IV.

Predictor Maps

These maps are tied to the tree diagram for the 38 model variables, which drive the distribution of a species. The 38 variable are divided into five categories including Climate (7), Soil Properties (12), Soil Types (9), Elevation (5), and Landscape (5). Due to the nature of the non-climatic variables, the seven climate variables were only mapped for each scenario to compare the differences in the models. Within the Climate tab, two maps have been provided so that a comparison can be made by selecting two models. Maps are based on a 20 x 20km grid of cells that make up the extent of the United States east of the 100th Meridian. The scale is 1:33,000,000 and is projected in Clarke 1866 Albers.

Climate

The climate will be affected by the emission levels of each scenario, and the results of nine models have been included for each variable. The legend contains the temperature in degrees Celsius; the scale for the current climate differs slightly from the other scenarios which all have the same temperature scale.

Mean Annual Temperature (°C): average annual temperature for the counties occupied by the species. Data from USEPA (1993).
Mean January Temperature (°C): average January temperature for the counties occupied by the species. Data from USEPA (1993).
Mean July Temperature (°C): average July temperature for the counties occupied by the species. Data from USEPA (1993).
Mean May-September Temperature (°C): Annual Precipitation (mm) - average annual precipitation for the counties occupied by the species. Data from USEPA (1993).
Mean May-September Precipitation (mm): Total monthly precipitation from May to September
Mean Difference between July and January Temperature (°C): measure of seasonality and here the differences between July and January monthly mean temperature.

Elevation

Maximum Elevation (m): average maximum elevation for the counties occupied by the species. Data from USGS 1:250,000 Digital Elevation Model (DEM) files.
Minimum Elevation (m): average minimum elevation for the counties occupied by the species. Data from USGS 1:250,000 Digital Elevation Model (DEM) files.
Mean Elevation (m): For each cell the average elevation from the maximum and minimum elevation
Range of Elevation (m): For each cell the difference in Maximum and minimum Elevation

This tool allows you to compare two predictor variables side-by-side. Select a predictor variable to display as the "Left Image" using the menu, and a predictor variable to display as the "Right Image."

If you select a Climate predictor variable, you can also select the climate scenario (e.g. HadleyCM3 – A1FI (High, "Harsh") or Current) as well.

Click and drag the orange handle over the maps to compare.

Left Image
Select a Predictor and scenario

Right Image
Select a Predictor and scenario

You are viewing the same predictors and scenario. Select a different predictor or climate scenario to compare.

Left Image:
Annual precipitation (mm) Current Modelled

Right Image:
Annual precipitation (mm) HadleyCM3 – A1FI (High, "Harsh")

Annual precipitation (mm) Current Modelled

Legend

Predictor Importance for () Relative to All Species

Boxplots provide a quick visual of the distribution of the variable importance from the random forest models from all 147 species (black boxplot) and how each species fits into the overall distribution (cyan line). The box encompasses from the 25th to the 75th percentile of the data. The thick line within the box identifies the median (50th percentile). The whiskers extend 1.5 times the interquartile range (75th-25th). This shows the spread of the data. Data points that lay beyond this range are shown with circles. The first figure shows all 38 predictor variables together (numbers on the x-axis). By clicking on each group (climate, elevation, vegetation set1, vegetation set2, and vegetation set3) a plot of each group will be opened.

Predictor Statistics

Predictor	Predictor Code	Predictor Name
1	PPT	Annual precipitation (mm)
2	PPTMAYSEP	Mean May-September precipitation (mm)
3	TJAN	Mean January temperature (°C)
4	TJUL	Mean July temperature (°C)
5	JULJANDIFF	Mean difference between July and January Temperature (°C)
6	TMAYSEP	Mean May-September temperature (°C)
7	TAVG	Mean annual temperature (°C)
8	ELV_MIN	Minimum elevation (m)
9	ELV_MAX	Maximum elevation (m)
10	ELV_RANG	Range of elevation (m)
11	ELV_MEAN	Average elevation (m)
12	12	Distribution of balsam fir
13	97	Distribution of red spruce
14	110	Distribution of shortleaf pine
15	111	Distribution of slash pine
16	126	Distribution of pitch pine
17	129	Distribution of eastern white pine
18	131	Distribution of loblolly pine
19	261	Distribution of eastern hemlock
20	316	Distribution of red maple
21	318	Distribution of sugar maple
22	371	Distribution of yellow birch
23	372	Distribution of sweet birch
24	375	Distribution of paper birch
25	462	Distribution of hackberry
26	491	Distribution of flowering dogwood
27	531	Distribution of American beech
28	541	Distribution of white ash
29	611	Distribution of sweetgum
30	621	Distribution of yellow-poplar
31	693	Distribution of blackgum
32	746	Distribution of quaking aspen
33	762	Distribution of black cherry
34	802	Distribution of white oak
35	832	Distribution of chestnut oak
36	833	Distribution of northern red oak
37	835	Distribution of post oak
38	972	Distribution of American elm

Variable importance scores for this species is plotted in cyan and can be compared to all species importance scores which are plotted as boxplots. Thus for example, if the cyan line extends above the boxplot (or the median of the boxplot,shown as the bold black line), it indicates that the variable for that particular species is a more important predictor than for most of the other species.

If you click on one grouping of the plot (e.g., climate, elevation, etc), you will get a higher resolution image of that section of the plot.

Close

Predictor Importance for (Cyan Line) Relative to All Species (Boxplot)

Predictor Statistics

Predictor	Predictor Name
1	Annual precipitation (mm)
2	Mean May-September precipitation (mm)
3	Mean January temperature (°C)
4	Mean July temperature (°C)
5	Mean difference between July and January Temperature (°C)
6	Mean May-September temperature (°C)
7	Mean annual temperature (°C)
8	Minimum elevation (m)
9	Maximum elevation (m)
10	Range of elevation (m)
11	Average elevation (m)
12	Distribution of balsam fir
13	Distribution of red spruce
14	Distribution of shortleaf pine
15	Distribution of slash pine
16	Distribution of pitch pine
17	Distribution of eastern white pine
18	Distribution of loblolly pine
19	Distribution of eastern hemlock
20	Distribution of red maple
21	Distribution of sugar maple
22	Distribution of yellow birch
23	Distribution of sweet birch
24	Distribution of paper birch
25	Distribution of hackberry
26	Distribution of flowering dogwood
27	Distribution of American beech
28	Distribution of white ash
29	Distribution of sweetgum
30	Distribution of yellow-poplar
31	Distribution of blackgum
32	Distribution of quaking aspen
33	Distribution of black cherry
34	Distribution of white oak
35	Distribution of chestnut oak
36	Distribution of northern red oak
37	Distribution of post oak
38	Distribution of American elm

Predictor Statistics

Predictor	Importance Value
PPT
PPTMAYSEP
TJAN
TJUL
JULJANDIFF
TMAYSEP
TAVG

Predictor Statistics

Predictor	Importance Value
ELV_MIN
ELV_MAX
ELV_RANG
ELV_MEAN

Predictor Statistics

Predictor	Importance Value
12
97
110
111
126
129
131
261

Predictor Statistics

Predictor	Importance Value
316
318
371
372
375
462
491
531

Predictor Statistics

Predictor	Importance Value
621
693
746
762
802
832
833
835
972

Potential Changes in Abundance and Range (Future)
GCM SCENARIO	% Area Occ	Ave IV	Sum IV	Future/Current IV

This tool allows you to compare two climate models side-by-side. Select a scenario to display as the "Left Image" using the menu, and a scenario to display as the "Right Image."

Click and drag the orange handle over the maps to compare.

Potential Changes in Abundance and Range (Future)

Percent Area Occupied simply is the percent of 20x20 km cells within the total area of the eastern United States that have been modeled to be suitable for the species. There are a total of 9767 cells used in the model.

Ave. IV is the average importance value across all 20x20 km cells that have been modeled to be suitable for the species.

Sum IV is the sum of importance values across all cells. Thus it is a metric that considers both the abundance and the range of the species, perhaps the best metric of overall species importance.

Future/Current IV is the ratio of change between the modeled current Sum IV and any of the GCM scenarios Sum IV.

Left Image:
Current Modelled

Right Image:
HadleyCM3 – A1FI (High, "Harsh")

Legend

You are viewing the same species. Select two different species.

(in decreasing order of importance)

Predictor Name	Importance Score	Rank

Class Map

Close

Predictor Map Help

The Geographic Predictors Maps are derived from Regression Tree Analysis, where each terminal node of the 'tree' corresponds to a legend and map color that is represented on the map. As such, one can determine what variables may be driving the species' distribution and abundance at a particular part of its range. For example, a precipitation variable may determine the western boundary and a temperature variable may determine the northern boundary for a species. Also note that variables important high in the tree diagram relate to larger portions of the species' range, while those variables lower (closer to the terminal nodes) relate to more localized variables driving the distribution.

IMPORTANT: It is important to note that if the model reliability is NOT high for the species, then the confidence that the predictors in a single regression tree are driving the distribution is LOW. It is therefore worthwhile to compare the predictor importance according to the RandomForest model. In the species page, click the "Statistics, Tables & Interpretations" button in the "Current Distriubtion" panel for viewing the predictor importance table under RandomForest model.

NOTE: Example: If tavg > < 6.5 means that if tavg> 6.5 deg C, traverse the left branch and if tavg < 6.5 deg C traverse the right branch.
If tavg <> 6.5 means that if tavg is less than 6.5, traverse the left branch. and if tavg > 6.5 traverse the right branch.

Close

Predictor Listings

Abbreviation	Predictor Name	Importance Score	Rank

Close

Class Map Help Predictor Acronyms

The regression-tree diagram along with the corresponding class-map shows what predictors are driving the distribution for this species.

Bird Species Not Found