Breeding Bird Atlas 2 Results
Blocks Covered and Effort
Atlas 2 volunteers surveyed 1,037 of the 1,055 blocks that were defined for the state. Of these blocks, 896 received at least 20 hours of survey effort, and 141 received fewer than 20 hours, or did not have the total number of hours recorded.
- Atlas 1 surveyed 969 blocks, and hours were not recorded for those surveys.
Species Found
From January 20, 2007, to October 16, 2011, volunteers submitted 149,470 records from 222 species. Some of the records were multiple reports of the same species in a block, and the final data that only lists the highest code for each species in a block uses 76,200 records.
- Atlas 1 collected evidence on 199 species.
- The final Atlas 1 data set uses 55,200 records.
- During Atlas 2 there were 191 species with Confirmed breeding as the highest evidence, 18 had a Probable code as their highest evidence, and 13 species had only Possible as their highest code. Many of the species with only Probable or Possible as their highest code were very thinly distributed, often occurring in only 1 or 2 blocks.
- During Atlas 1 there were 194 species with Confirmed breeding as their highest evidence, 6 that had a Probable code as their highest evidence, and 0 species that had only Possible as their highest code.
Species Confirmed During One Atlas, But Not The Other
Twelve species were Confirmed during Atlas 2 but not in Atlas 1. They are:
- Yellow-crowned Night-Heron
- Bald Eagle
- Sharp-shinned Hawk
- Merlin
- Peregrine Falcon
- Sandhill Crane
- Common Raven
- Cerulean Warbler
- Prothonotary Warbler
- Hooded Warbler
- Clay-colored Sparrow
- White-winged Crossbill
Fifteen species were Confirmed during Atlas 1 and not Atlas 2. They are:
- Northern Shoveler
- Northern Pintail
- Ring-necked Duck
- Red-breasted Merganser
- Ruddy Duck
- American Coot
- Least Sandpiper
- Wilson's Phalarope
- Long-eared Owl
- Common Nighthawk
- Red-headed Woodpecker
- Olive-sided Flycatcher
- Golden-winged Warbler
- Yellow-breasted Chat
- Rusty Blackbird
There were no species recorded in Atlas 1 that were not recorded in Atlas 2, although there were several species found during Atlas 2 that were not recorded during Atlas 1. Many of these are very rare species in the state during the summer, and only a few were Confirmed during Atlas 2. They are:
- Forster's Tern
- American Wigeon
- Great Cormorant
- Black Vulture
- Bald Eagle
- Merlin
- Peregrine Falcon
- Sandhill Crane
- Ring-billed Gull
- Monk Parakeet
- Bicknell's Thrush
- Prothonotary Warbler
- Kentucky Warbler
- Summer Tanager
- Blue Grosbeak
- Dickcissel
- Clay-colored Sparrow
- Lincoln's Sparrow
- White-winged Crossbill
Most Widespread Species
Widespread species, those found in the most blocks, are one of the metrics derived from Atlas data. These widespread species, and the number of blocks in which they were recorded, are listed below.
Graph of the species found in the most blocks, and the number of blocks where they were recorded during Atlas 2. Note: There were 1,037 blocks surveyed during Atlas 2.
Species per Block
- On average, Atlas 2 volunteers found 74 species per block.
- This is notably higher than the 57 species per block average that were found during Atlas 1.
Minimizing Overestimation of Species Gain
Atlas 2 found, on average, many more species per block than Atlas 1 - and the disparity is greater than differences found in atlases from neighboring states. For instance, the New York Atlas 2 found, on average, a gain of 4 species per block between the Atlas 1 and Atlas 2 data sets (McGowan & Corwin 2008). This differential is similar to the same metric from the Vermont Atlas 2 (Renfrew pers. com.), and both are notably smaller than the 17 species differential between Massachusetts Atlas 1 and 2.
Regional Patterns of High Variability Between Atlas 1 and Atlas 2 Samples
While the average species differential was high, it was not evenly distributed around the state, and in most cases where there was a notable disparity between the species totals in a block the Atlas 2 species totals were higher than Atlas 1. Additionally, the low species totals during Atlas 1 were clustered into some regions of the state. For example, of the 956 blocks that were surveyed in both Atlases, 127 had a species differential of more than 40 species. Fifty-five (55) of those blocks were located in Worcester County, and 16 were located in Franklin County (with many on the Worcester County border). The remaining 72 blocks with a 40+ species differential were scattered throughout the state. In all but one of the Worcester and Franklin County blocks with a 40+ species differential, Atlas 2 surveys found more species than Atlas 1.
It is not likely that the species differential Atlas 2 revealed was entirely driven by new species colonizing a block, rather, it is more likely that Atlas 1 surveys suffered lower species totals for a variety reasons, mostly related to field methods and tools available in the field. Atlas 1 was the first Breeding Bird Atlas in North America, and as such observers were "blazing a new path." Their efforts appear to have focused on achieving the highest code for each species in a block, rather than the highest species count in each block. This can take a significant amount of time and effort, and keep observers from maximizing species totals. Atlas 2 volunteers were directed to first maximize species count during Safe Dates, then to upgrade species to higher breeding codes.
If Atlas 1 is compared to Atlas 2 without any adjustment, there is a real risk of overestimation of "new" species in a block. This would mask species declines when data were pooled across the state.
To most accurately compare species totals and infer range changes, we needed to eliminate the most poorly covered Atlas 1 and Atlas 2 blocks from discussions of species range change. Controlling species counts to reflect effort is not possible, since Atlas 1 did not record field effort. Instead, we used the species totals as a surrogate measure of effort.
Change Data Set
In this document we report the regional distribution of a species during Atlas 1 using the full Atlas 1 data set (969 blocks). When we discuss the current regional distribution during Atlas 2, we use the full Atlas 2 data set (1,037 blocks). However, in the sections that discuss the change in a species breeding footprint in the state, we use a subset of the data that is referred to as the "best surveyed blocks" or the "effort-controlled" blocks (829 blocks).
We used the species total as a surrogate for effort, and we selected blocks where there was a differential between the two Atlas species of fewer than 40 species. This reduced the sample size of the blocks in the "change" data set, but reduced the likelihood that we were overestimating species growth.
- There were 829 blocks in the change data set. These blocks were surveyed in both Atlas 1 and 2, and the difference in the species totals were within 40 species.
- When we reduced the sample to only include blocks with fewer than a 40-species differential, the average species per block for Atlas 1 climbed from 57 to 61, and the average species per block for Atlas 2 went from 75 to 74. The average difference in the total species found per block was reduced to 14 species.
- This correction likely still overestimates "gain" for species. This contributed to our estimation that any drop below 100% block occupancy may represent an actual decline for a species. This metric is used in the species accounts in the discussion regarding the species trend and conservation status.
"Change" data is reported as the percentage of the Atlas 1 occupied blocks detected during the Atlas 2 surveys. If the same number of blocks is occupied in Atlas 2 as were occupied during Atlas 1, the percent is 100. If the number of blocks doubles, the percent remaining is 200, and if the number of blocks occupied is halved, the percent of blocks occupied is 50.
A Note on Other Uses of the Change Data
The "change" data is also used in discussions of gain, loss, and stability for species. Atlas 2 surveys reveal the number, or percent, of blocks that remained occupied (stable), had new detections (immigration or gain), or lost a species both statewide, and for each region in the state. These numbers are occasionally discussed in the species accounts, often with regard to a species' persistence within a region. Those tables, however, are not included in this volume.
These metrics of the patterns of gain-loss-stability are often instructive. Occasionally, a species will remain stable (be reported in a similar number of blocks in both Atlases) in an ecoregion, or statewide, while shifting its range and showing similar numbers of blocks gained and lost. Ruffed Grouse show this pattern well.
Breeding Footprint Changes by Breeding Habitat and Life History Characteristics
There are several ways to view the change in breeding footprint for species in the state. Some species increased or decreased their footprint tremendously - the absolute number of blocks they occupied during Atlas 2 was much larger or smaller than during Atlas 1. Those with the greatest increase or decrease in the number of blocks occupied are charted below. Note that these are gross figures - a bird species which disappeared from a large number of blocks but also appeared in a large number of other blocks could appear on both charts.
Species with the greatest increase in the absolute number of blocks occupied since Atlas 1, as well as the total number of new blocks they occupied.
Species with greatest decrease in the absolute number of blocks occupied since Atlas 1, as well as the total number of blocks from which they disappeared.
Understanding the change in a species' status in the state requires that we also understand how each species' present distribution has changed relative to its 1979 distribution. More blocks were surveyed during Atlas 2, so all species could have increased in the absolute number of blocks occupied, and while it is instructional to look at absolute growth, it is also important to understand that the percent of growth is adjusted for effort. Looking at the absolute number of blocks can also obscure changes for species that occupy specialized habitats, such as Willet. These species may occupy a habitat that only occurs in a few blocks, so their growth is limited compared to a species with a more common habitat, like a forest dwelling species.
To get a clearer picture we can look at the percent change in distribution for each species, setting each bird's 1979 distribution as 100%. Evaluating the data through that lens, and looking at the species with the largest increase and decrease relative to 1979 distribution shows that the following species have had dramatic swings in the number of blocks they occupy:
Change in percent of blocks occupied by the species with the largest gains. Percent change noted at end of bar. A species whose range did not grow or shrink at all would have 100%, a species whose range doubled would have 200%, etc.
* The Red-bellied Woodpecker's real percent of blocks change from BBA1 to BBA2 was 15,825%. It was shown as less in the graph above for ease of display.
Similarly, we can look at the species that have the greatest percent loss in distribution relative to Atlas 1.
Change in percent of blocks occupied by the species with the largest loss of footprint. Percent change noted at end of bar. A species whose range did not grow or shrink at all would have 100%, a species whose range halved would have 50%, etc.
Each breeding species was assigned to one or more breeding habitats, with the assignments relating to both the foraging and nesting habitats of the species during the breeding season. Using the "change" data set, we calculated the percent of the species the Atlas 1 breeding footprint still occupied, and used that metric as a measurement of the species' trend since Atlas 1.
The breeding habitats were:
- Grassland and Agricultural Landscapes - including human structures, like barns
- Shrublands - also called early successional or young forest
- Forest
- Coastal - including dunes, cliffs, offshore islands, beaches
- Tidal Marsh
- Open Freshwater - including rivers, lakes, and ponds Freshwater Marsh - only grass and shrub-height emergent vegetation
- Freshwater Swamp - dominated by trees
- Urban
- Suburban
Breeding species were also assigned other behavioral characteristics and protection status and tested to see if those species had increasing or decreasing breeding footprints. The behavior characters and protection status examined were:
- Listed Species - State listed as Special Concern, Threatened, or Endangered
- SWAP Listed Species - State Wildlife Action Plan listed
- Ground and near-ground breeders
- Cavity Nesting
- Colonial Nesting
- Nesting on Human Structures - including boxes, platforms, buildings
- Aerial Insectivores - species that pick insects from the air while flying
- Common Feeder Birds
- Introduced Species
- Game Species
- Generalists - three or more breeding habitats used
- Birds at the Northern/Southern Edge of their Breeding Range
- Migration Status - residents, short-range migrants, long-range migrants
Birds assigned to the same habitat were compared to all other species in the state to measure if the species in those habitats were increasing or decreasing their footprint when relative to other species. The results (using log transformed percent data and SAS ANOVA) indicated that species in the following categories are definitely (p<.05) or likely (p<.1) increasing more than other species:
- Forest Breeders
- Open Water Breeders
- Freshwater Swamp Breeders
- Birds Nesting on Human Structures
- Birds at the Northern Edge of their Range in Massachusetts
- Resident Species (compared to long-distance migrants)
Likewise, species showing declining footprints when compared to all other species were:
- SWAP Listed Species
- Grassland/Agricultural Land Breeders
- Shrubland Breeders
- Ground and Near-Ground Breeders
- Aerial Insectivores
- Birds at the Southern Edge of their Breeding Range in Massachusetts
- Long-Distance Migrants (compared to residents)