Oecologia (1995) 101:478-486 ORIGINAL
9 Springer-Verlag 1995
PAPER
Mark L. Taper 9Katrin Brhning-Gaese - James H. Brown
individualistic responses of bird species to environmental change
Received: 30 December 1993 ! Accepted: 26 October 1994
Abstract We investigated how the population dynamics of the same bird species varied in different environments, and how the population dynamics of different species varied in the same environment, by calculating long-term population trends for 59 insectivorous songbird species in 22 regions or strata of eastern and central North America using data from the North American Breeding Bird Survey. Of the 47 species that occurred in more than one region 77% increased in some regions and declined in others. Of the 22 regions 91% had some species that increased and others that decreased. There were only slightly more significant correlations between strata in species trends and between species for stratum trends than would be expected by chance. Because of nonlinearities in the data, the actual patterns of population fluctuations of the same species in different regions and of different species in the same region were even more heterogeneous than suggested by our analyses of linear trends. We conclude that these bird species respond to spatial and temporal variation in their environment in a very individualistic fashion. These individualistic responses show that the extrapolation of population trends gained from a few local studies to a larger spatial scale, and the use of a few indicator species to monitor the status of a broader community, are suspect. Key words Community dynamics. Insectivorous birds Large-scale environmental variation 9Long-term population trends 9North American Breeding Bird Survey M.L. Taper1 9J.H. Brown Department of Biology,Universityof New Mexico, Albuquerque, NM 87131, USA K. Brhning-Gaese(~) Abteilung for Verhaltensphysiologie, Beim Kupferhammer8, D-72070 Ttibingen, Germany Present address:
I Dept. of Biology,MontanaState University, Bozeman, MT 59717, USA
Introduction One of the main questions in community ecology has been how assemblages of species respond to environmental variation in space and time. How does the same species respond to a variety of environments? How do different species respond to the same environment? These questions raise fundamental conceptual issues about the nature of ecological communities and the kinds of interrelationships among the component species. Alternative viewpoints date back to at least the classic debate between Clements (1916) and Gleason (1917, 1926). To what extent are communities integrated units comprised of closely interacting, tightly coevolved species, or to what extent are they facultative assemblages of species that respond individualistically to abiotic conditions and other organisms, each according to its unique requirements and life history characteristics? Insight into the nature and dynamics of communities can be obtained from comparing changes in the abundance of ecologically similar species over time at different locations. With regard to changes in the abundance of a single species in different environments three different patterns might be expected. Firstly, if a single pervasive factor affects the abundance of a species, we might expect all populations to exhibit correlated dynamics. For example, tropical deforestation on the wintering grounds might be a factor that causes correlated declines in populations of a migratory bird species in all regions within its breeding range (Terborgh 1989; Robbins et al. 1989). Secondly, the abundance of a species might be determined by factors such as climate that have different effects in different parts of the species' range (Heggberget 1987; Root 1988). For example, we might expect global warming to cause declines in the southern areas of a species' range and increases in the northern regions. This would result in a spatially explicit pattern of both positively and negatively correlated changes in the abundance of a species over its entire range. Thirdly, we might find spatially independent uncorrelated changes in the abundance of a species. This would indicate that dif-
479
ferent populations respond individualistically to changes in multiple independent environmental factors at relatively small spatial scales. For example, a species might increase in New England because of reforestation, decrease in Illinois in response to increasing predation pressure (B6hning-Gaese et al. 1993), and have stable populations in Ohio because of the two processes balancing each other out. With regard to changes in ecologically similar species within the same region, again, three different patterns can be expected. Firstly, a single strong factor such as urbanization could cause similar declines in an entire guild of ecologically similar species, e.g., farmland species. In this case we would expect correlated responses in the population dynamics across multiple species. Secondly, the population dynamics of a group of species could be influenced by factors that have negatively correlated effects on different species. For example, if temperatures are increasing with global climate change, within a region we would expect increases in species that are near the northern margin of their breeding ranges and declines in species near the southern margin of their ranges. Thirdly, if different species respond individualistically to different environmental factors influencing their abundances, we expect no consistent patterns of correlations among their population dynamics. In order to be able to distinguish among these possible patterns, changes in the abundance of different species have to be monitored over large geographical areas. But few studies have been conducted at sufficiently large spatial and temporal scales (Kareiva and Andersen 1988; Tilman 1989). There have been several long-term investigations of the abundance of a group of organisms within a single locality or region (e.g., Jfirvinen and V~iis~inen 1977a,b; Enemar et al. 1984; Holmes et al. 1986; Brown and Heske 1990; Marchant et al. 1990). These studies offer important insights into the dynamics of communities. However, additional insights might be obtained by comparing the patterns among different regions on a continental scale. One of the few data sets available on a continental scale is the North American Breeding Bird Survey (BBS), conducted by the United States Fish and Wildlife Service and the Canadian Wildlife Service since 1965 (Robbins et al. 1986). The BBS is attractive because of its broad spatial scale and its long duration. It is a massive compilation; each year during the breeding season the populations of all breeding birds are censused at approximately 2000 routes throughout North America. Many sites have been censused for more than 20 years. Despite limitations of its sampling design and data (Robbins et al. 1986; Droege 1990; B6hning-Gaese et al. 1993), the BBS represents one of the few data sets available for investigating long-term, large-scale changes in populations and communities of any organism in North America. In this paper we investigated how the population dynamics of the same bird species varied in different environments, and how the population dynamics of different
species varied in the same environment. We focused mainly on the long-term population trends.
Methods Each year, the BBS counts birds along approximately 2000 roadside routes distributed across the North American continent north of the United States-Mexico border (Robbins et al. 1986). Each route is censused by an experienced volunteer one morning per year during the breeding season in June. A route consists of 50 sampling locations 0.8 km apart. At each location, the observer counts by species all birds heard or seen during a 3-rain period. Our statistical analysis began by confronting the problem that single BBS routes do not give reliable counts of the number of individuals of each species present in single years. For example, errors are caused by true sampling error, observer bias, and day-today differences in avian activity. We therefore did not calculate population trends for individual BBS routes. Fortunately, the importance of the error in counts for individual routes can be reduced by changing the spatial scale and calculating averages over multiple routes. When this is done there are clear regional trends that cannot be attributed to errors in counts at individual routes. The 95 physiographic areas or strata (Robbins et al. 1986) of North America were selected as spatial units for these averages. The strata are based largely on a map of biome types of North America (Aldrich 1963) and are ecologically more meaningful than political provinces (e.g., states). In fact, the geographic range boundaries of many species correspond closely to many of the stratum boundaries (Robbins et al. 1986). Our goal was to calculate for each stratum and each species the mean number of individuals counted per year, averaging over the routes within the stratum. However, obtaining unbiased averages over multiple routes is not straightforward (Robbins et al. 1980; Geissler and Noon 1981). Most routes have not had high-quality surveys continuously from 1965 to 1987. To overcome this problem, we considered only the 20 years from 1968 to 1987, because few routes were sampled before 1968. Additionally, we limited our analysis to routes of high quality: of the routes that were run from 1968 to 1987, we used only those that had no more than two missing counts, no two missing counts in a row, and no missing counts in 1968 and 1987. For analysis the data were transformed to log abundances [In(count+l)]. This formulation, like most other ways of quantifying abundance, creates problems when estimating variance in population size, but these problems did not affect our analyses, which were based on temporal trends in the means of multiple counts within a stratmaa (see below). More details of methods can be found in B6hning-Gaese et al. (1993). We restricted our analyses to an ecologically similar group of passerine species: wood warblers (subfamily Parulinae of the Emberizidae), vireos (Vireonidae), gnatcatchers and kinglets (subfamily Sylviinae of the Muscicapidae), titmice and chickadees (Paridae), nuthatches (Sittidae), the brown creeper (Certhia americana, Certhiidae), wrens (Troglodytidae) and bluebirds (Sialia, Muscicapidae). These species comprise taxonomically defined groups whose members are small, insectivorous, mostly foliage-gleaning songbirds (Ehrtich et al. 1988). For each stratum the mean number of individuals per species was calculated for each year, averaging over all the routes within the stratum at which the particular species was observed at least once during the 20 years. Stratum averages were calculated only for species that had been recorded on at least eight routes per stratum. We ended with time series for 59 species (Appendix 1) in up to 22 strata, all in eastern and central North America (Appendix 2). Long-term population trends were calculated in each stratum for each species by regressing log abundance as a function of time (proc GLM, SAS/STAT 1988). Since in all cases investigated in this paper the slope of this regression was small, the slopes were essentially equal to fractional changes in population size per year. We multiplied by 100 and reported our slopes as per cent growth rate per year.
Table 1 Population trends (% growth rate/year) and their significance levels for 59 species of insectivorous bird species in 22 different strata of North American. Right margin: number of strata per species, number and % of significant increases and declines Species
Stratum 04
Viol Viph Vigi Vifl Viso Vigr Vibe Mnva Prr Lisw Heve Vepi Vech Veru Vepe Paam Deti Depe Decr Deeo Dema Dece Depn Deca Dest Defu Dedo Devi Depi Deds Seau Seno Semo Opfo Opph Getr Icvi Wici Wipu Wica Seru Thlu Thbe Trae Trtr Cipl Cipa Ceam Sica Sicn Sipu Pabi Paat Paca Pahu Resa Reca Poca Sisi n Sig.pos Sig.neg
per species. Bottom margin: number of species per stratum, number and % of significant increases and declines per stratum. Species and stratum abbreviations: see Appendix
1,06"
05 -2.06***
-0.24 -0.23 1.99"**-1.87"**
10
12
5.40***
0.47
0.43 0.54
4.23*** 2.59*** 0.85
4.02***
0.18 1.07"** 1.23"** 0.43
0.26
13
14
-0.02
-1.t6
0.09
1.04"** 0.65
-1.33"
-1.1l*
-0.78
-0.60* -0.06
-0.57 1.83" -0.96 -1,45 *** -0.57
0.98***
0.40 -0.34
15
16
17
18
-1.27
1.40
3.55***
1.69"
0.67 1.18"
4.37*** -1.95"** 3.24*** 1,63"** :
-1.15
-0.20
3.41"**
5.08*** -1.58"** -0.10
1.24 0.04 -0.18 -0.46 -1.57" 2.33*** -0.30
1.85"**
20 0.50
24 0.28
-2.83*** 1.42" 1.71"** -0.27 2.26***
0.94
-0.32
1.27" -2.81"** -0.48 -0.06
1.28"*
0.43
19
0.10 1,83
2.92***
0.37
5.64***
0.69
1.51"* 0.23 3.05"** 1.57"**
0.83
0.33
1.52"** -1.43"* -2.78***
1.38"*
0.41 0.24
-0.04 -1.84"**
0.36
0.50 3.46*** 0.94 3.02*** 1.76"** -1.09 -2,28*** -1.45"** 0.87 0.86 0.78 0.39 0.04 1,73"** 0,87 -0.25 -0.80 1.22 -1.64"** -0.93
3.70*** -0.81" 2.48*** 0.42* -0.01
-1.75"** -4.43*** 1.53" -0.22 -5.76*** -0.87 0.91"
-0.71" 0.21 -0.68 -1.81"** -0.44
-1.34" 1.74 0.70
0.23
-l.93"** 2.46** -0.24
0.46 1.11 *
0.42
0.29
-0.24 -1.50 -6,06 -3.87*** 1.12"** 6.61"**
0.05 5.02*** -0.39 0.29
3.29***
0.62
0.05
0.87 1.48"**
0.83
3.39***
9.71"** 2.18"**
1.65"*
2.98***
1.84"**
1.15
1.27
1.74"** -0.76 2.99*** 0.70 1.59"
1,66"** 3,42***
0.28
2.41"** 4.08*** -0.83
0.38 0.05 3.09*** -1.25
29 14 (48%) 4(14%)
2.75**
1.72"* 0.57 0.76
1.76" -2.04*** -0.66
2.36*** -1.16 0.11
2.35***
1.33
2.64*** 4.68***
0.17
-2.13"
-3.37*** 0.76 -4.85* -0.48
1.30
-0.19
-0.93* 0.69 -0.89
-0.29
0.31
2.94* 2.19"** 1.62"
1.74 *
1.85" 6.66***
0.27
8 19 28 25 29 15 19 1 (13%) 9 (47%) 10 (36%) 4(16%) 6(21%) 4 (27%) 11 (58%) 4(50%) 1(5%) 3(11%) 7(28%) 6(21%) 4(27%) 1(5%)
2.53*** -0.15 1.04 ***
2.62*** 0.04 2.66*** -0.14 -1.28" 1.87"
-3.65 -2.41" -0.99 * -0.45
0.84***
1.62"*
0.05 0.35
1.46
-0.73 0.72 1.80" -4.21"** -4.26*** -4,78*** -2.03*** -0.04
0.53 0.35
2.94***
2.10" -1.08 -1.75
4.9***
0.12 0.06 0.78* 0.64 3.73***
3.48***
-0.91 9 8 0(0%) 5 (63%) 1(11%)0(0%)
3.97*** -0.03 -2.04
-0.11 -0.15
24 17 33 7 (29%) 8 (47%) 12 (36%) 3(13%) 2(12%) 2(6%)
481 Table 1 (continued) Species
Stratum 27
28
Viol Viph Vigi Vifl Viso Vigr Vibe Mnva Prci Lisw Heve Vepi Vech Veru Vepe Paam Deti Depe Decr Deco Dema Dece Depn Deca Dest Defu Dedo Devi Depi Deds Seau Seno Semo Opfo Opph Getr Icvi Wici Wipu Wica Seru Thlu Yhbe Trae Trtr Cipl Cipa Ceam Sica Sicn Sipu Pabi Paat Paca Pahu Resa Reca Poca Sisi
-0.30
1.06 -0.43 0.19 0.18 1.97"**
n Sig.pos Sig.neg
30 38 4(13%) 11 (29%) 9(30%) 4(11%)
1.60" -0.18 1.16
30
31
32
33 -1.13"
1.99"
0.71
-0.10
1.05
1.44 0.89*
1.84"** -0.30 -0.62
-0.03 1.12 -2.07***
34
39
0.35
2.69 *
-0.40 -5.19"** 4.47***
1.90"**
1.75"** 0.89 -0.04 0.87* -1.02"* 1.15 1.67"* -3.08*** -1.63" 0.50 2.48*** 3.32*** -0.60 0.30
-0.58
-2.47** -1.46" -2.63*** 1.93"** -0.38 0.05 -0.68 0.03
2.33 ***-0.12
-0.97*
-0.34
-0.27 -0.86 -0.48 1.13" 0.03 0.34 1.61" 0.92 0.37
2.19"** -2.04*** -0.27
1.65
3.87*** -2.44***
0.93* -3.21"** -0.09 -3.72*** 0.39 -4.5*** 1.32 -0.42 -0.33 -0.37 0.25 3.92*** 0.83
-1.02"** -0.93 -0.95
0.45
1.25"
2.57*** -1.56"
3.66***
1.84"** -1.06 5.14"** -4.07***
0.36 -1.29"** -3.75*** -0.59 0.32 -1.11" 0.33
0.53
1.05
-0.91"
0.29 0.67 0.78 2.99***
1.74"* -1.89"** 2.28**
1.97"** 1.30 0.59 3.49***
-2.47*** 0.95 -3.00** -1.09"
-0.55
*P<0.05; **P<0.01; ***P<0.001
1.70" -1.45
-0.02
2.19"* -0.32
40
0.53
6 14 14 15 6 2(33%) 6(43%) 4(29%) 3 (20%) 3 (50%) 1(17%) 3(21%) 3(21%) 6 ( 4 0 % ) 0 ( 0 % )
3 0(0%) 0(0%)
n
20 1 1.76 19 14 4 9 4 9 3 1 3 6 3 4 1 5 1 -0.43 19 3 4 3 1 6 1 1 4 4 4 7 7 12 4 8 7 4 -0.33 22 11 5 1 4 12 11 4 3.23*** 20 2 5 2 4 15 4 1 14 2.31"** 13 8 1 1 1 11 -1.05 18 7 2 (29%) 0(0%)
Sig.pos
Sig.neg
5 (25%) 0 (0%) 8 (42%) 5 (36%) 2 (50%) 2 (22%) 1 (25%) 2 (22%) 1 (33%) 1 (100%) 0 (0%) 3 (50%) 0 (0%) 1 (25%) 0 (0%) 3 (60%) 0 (0%) 7 (37%) 0 (0%) 4 (100%) 1 (33%) 0 (0%) 1 (17%) 0 (0%) 0 (0%) 1 (25%) 0 (0%) 1 (25%) 5 (71%) 1 (14%) 5 (42%) 0 (0%) 3 (38%) 0 (0%) 2 (50%) 9 (41%) 0 (0%) 2 (40%) 1 (100%) 0 (0%) 1 (8%) 0 (0%) 0 (0%) 5 (25%) 0 (0%) 1 (20%) 1 (50%) 1 (25%) 9 (60%) 1 (25%) 0 (0%) 8 (57%) 10 (77%) 4 (50%) 0 (0%) 0 (0%) 0 (0%) 5 (45%) 3 (17%)
2 (10%) 0 (0%) 2 (11%) 0 (0%) 0 (0%) 3 (33%) 2 (50%) 2 (22%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (67%) 0 (0%) 0 (0%) 0 (0%) 1 (100%) 3 (16%) 1 (33%) 0 (0%) 0 (0%) 1 (100%) 2 (33%) 0 (0%) 0 (0%) 1 (25%) 1 (25%) 1 (25%) 0 (0%) 3 (43%) 0 (0%) 0 (0%) 0 (0%) 1 (14%) 0 (0%) 3 (14%) 8 (73%) 1 (20%) 0 (0%) 3 (75%) 3 (25%) 3 (27%) 3 (75%) 2 (10%) 0 (0%) 2 (40%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 (21%) 0 (0%) 0 (0%) 1 (100%) 0 (0%) 1 (100%) 1 (9%) 2 (11%)
482
Results Table 1 summarizes the basic results of this analysis. The growth rates for each species in each stratum are given, along with significance levels. In the right margin we list for each species the number of strata with significant positive and negative trends. The bottom margin indicates for each stratum the number of species with significant positive and negative trends. The most striking thing about Table 1 is that, whether one looks across species or down strata, there was wide variation in growth rates. Of the 47 species with data from more than one stratum 77% showed increases in at least one stratum and decreases in at least one other (Table 2). Further, 91% of the 22 strata had some species that increased and others that decreased (Table 3). More detailed analysis confirmed the impression that there is little pattern in Table 1. We compared strata for similarity in the trends of their species; that is, we calculated the correlation between strata using the trends of different species as observations. Only 9.4% of the 221 possible correlations were significantly positive and 0.9% significantly negative at the 5% level (compared to 2.5% expected on the basis of chance alone). Neighboring strata showed a slight tendency to be more correlated than distant strata. We tested this by comparing the similarity matrix of the strata with a physical proximity matrix of the strata. The similarity matrix of the strata was calculated as above. The correlation coefficient can range from -1 (complete dissimilarity of the strata) to +1 (complete similarity of the strata). The physical proximity matrix of the strata was obtained by defining the distance of strata with a common boundary as 1 and the distance of strata without a common boundary as 0. The matrix correlation between these two similarity matrices was tested using Mantel tests (Smouse et al. 1986). The resulting matrix correlation coefficient was r = 0.33 with a significance of P < 0.001 (2000 bootstrap replications). We also compared species for similarities in their trends in different strata, i.e., we compared correlations among species using stratum trends as observations. This analysis was restricted to the 15 species that occurred in more than ten strata. Of the 105 possible correlations only 7.6% were significantly positive and 2.9% were significantly negative at the 5% level. To summarize, both strata and species comparisons showed a larger, but only slightly larger, proportion of significant correlations than expected by chance. When comparing the population trends among species and strata, considering only the long-term trends is a simplification of the actual population dynamics. However, when curvilinear regressions or the actual data were used to compare the year-to-year variation in the abundance of species across regions, the individualistic responses of species to their environments were even more apparent. Two figures are shown as typical examples: the population dynamics of one species, here the red-eyed vireo (Vireo olivaceus), varied greatly in different strata
Table 2 Patterns of population change for one species in different strata Number of species Species that occur in two or more strata
47
Species with declines in at least one stratum, and increases in at least one other stratum Species with significant declines in at least one stratum and significant increases in at least one other stratum
36 (77%) 18 (38%)
Species with increases in all strata Species with significant increases in all strata
7 (15%) 1 (2%)
Species with declines in all strata Species with significant declines in all strata
4 (9%) 0 (0%)
Table 3 Patterns of population change for different species within one stratum Number of strata 22 Strata with declines in at least one species, and increases in at least one other species Strata with significant declines in at least one species and significant increases in at least one other species
20 (91%) 17 (77%)
Strata with increases in all species Strata with significant increases in all species
2 (9%) 0 (0%)
Strata with declines in all species
0 (0%)
(Fig. 1). Comparing the population dynamics of different species in the same stratum, here Southern New England, a similarly large amount of variation can be seen (Fig. 2).
Discussion To summarize, we have four important results: (1) 77% of the species increased in some regions and declined in others; (2) 91% of the strata had some species that increased and others that decreased; (3) there were only slightly more significant correlations between strata in species trends and between species for stratum trends than would be expected by chance; and (4) the actual patterns of population fluctuations of the same species in different strata and of different species in the same stratum were even more heterogeneous than suggested by our analyses of linear trends. We conclude that these bird species respond to spatial and temporal variation in their environments in a very individualistic fashion. We find both individualistic responses of a single species to different environments and individualistic responses of different species to the same environment. It is easy in retrospect to imagine scenarios
483 Fig. 1 Population dynamics of the red-eyed vireo (Vireo olivaceus) in 12 different strata fitted by linear and quadratic regressions. Strata abbreviations see Appendix 2
O4 2.2
05
0.8
m'
2.1
,
,
0.7
. "
=
i
2.0
,
10
9
2.5
9
,
,
k
,
o
~ 9
0.6
lm
9
9
2.0
1.9 0.5
1.8
9
9
0.4
1.7 %6
0.3
1.8
0.2
1.5
I
13 1.7
I
I I
a
i
I
i
1,0
15
~=
"
i
I
'
'
"
6
9
""
i
1.4
r/."
"
1
'"
"
0.9
Z
12
,.,
9
'
'
u
L
t
0
1
-
,,
28
3.6
t
30 2.4
3.3
, 9
9
0.6,
9
3.2 3.1
OA
2.0
3.0
0.3 1.8
2.9
0.2
2.8
1.6
0.1
2.7 8
2'65
33
0.5
2.2
9
~!
381
2.0
"!t 3.4
0.8
l
!
I
!
70
75
80
85
90
14 '-65
I
I
1
I
70
75
80
85
9O
0.0
65
I
I
1
i
70
75
80
85
9r
YEAR
that would cause such a lack of associated dynamics. For example, a certain species might have declined in one region because of nest predation and in another because of harsh winters, while it increased in a different region because of winter food supply and in another because of changes in habitat. Within the same region, one species might have increased because it benefited from successional changes in its breeding habitat while another species might have declined because it was negatively affected by the same habitat changes or by hunting pressure on its wintering grounds. Thus, each species has its own unique niche, and its abundance and distribution are controlled by a different combination of abiotic and biotic factors. These specific niche requirements interact with local and regional environmental variation to produce spatially heterogeneous dynamics within species and equally variable dynamics for the multiple species that coexist within a given region. These results are con-
sistent with an individualistic or Gleasonian view of ecological communities (Gleason 1917, 1926). This result is strengthened by similar results obtained in another analysis of bird population dynamics in eastern North America. B6hning-Gaese et al. (1994) asked whether it is possible to extrapolate the results on population dynamics obtained in small-scale studies to larger spatial and temporal scales. They showed that these extrapolations are not reliable because different factors seem to determine long- and short-term population dynamics and because in different regions different variables seem to be important. They concluded that the population dynamics of each species are controlled by an individualistic combination of factors. Similar results have also been found for other groups of organisms. Individualistic responses to environmental change were reported, e.g., by Davis (1986) and Cole (1982) for trees, by Graham (1986) for small mammals,
484 Fig. 2 Population dynamics of 12 different bird species in the Southern New England stratum fitted by linear and quadratic regressions. Species abbreviations see Appendix 1
Viol
Vigi
3.3
1.4
V i s o
9
,
,
0,5
,
3.2
0.4 k
1.2 3.1
0.3 F
9
1.0
3.0 2.9
0.8
2.8 0.6
2.7 2.6
0.4
-O.1
Mnva
Vepi
1.6
9
1.5
9
9
Vech
1.4 =
1.4
0.4
,
1.3 1.2
0.3
.
1.1
0.2
9
1.0 1.3
0.9
1.2 9 Z
0.1
0.8
1.1
=9
0.0
0.7
n
n
n
0.6
nO~1
0.5
0.5
0.4
0.4
Deco 0.5
.
.
.
0.4
.
9
,
,
y
..y
0.3 0.2
0.3 0.2 0.2
0.1
~
0.0
I
-0.1
9
=m
0.1
= 9149 m
n
0.0 i
,
-0.1
I
Deds
I
=
0,0
=
m
m
I
Pabi
0.4
1.2
9
0.3
2
1.0
0.2
0.8
0.1
0.6
0.0
65
m
Seno
1.4
o.41
0.1
70
75
80
85
90
-0.1
..~.~.. 9 9 !
65
70
1 mm
/
I
I
75
80
85
90
0
65
70
75
80
85
90
YEAR
by Coope (1986) and Elias (1991) for insects and by Valentine and Jablonski (1993) for mollusks (see Foster et al. 1990 for a review). All these studies worked over paleontological time scales, e.g., comparing communities from the late Pleistocene with contemporary assemblages. In contrast to these paleoecological studies, in our study we were able to detect individualistic responses of species within time periods of just 20 years. Birds potentially respond very rapidly to environmental change in both space and time because they are highly mobile. Further, the species of small insectivorous passerines that we worked with have short generation times. Additionally, in contrast to the paleoecological studies cited above, we focused on a group of ecologically similar species that was selected a priori, before any analysis. Because the species were members of the same guild, one might expect more similarity in their response to environmental
variation than in the more heterogeneous groups of species used in the paleoecological studies. Nevertheless, our decades-long analysis of an avian guild revealed a degree of individualism comparable to that observed over tens of thousands of years in assemblages of ecologically diverse species. The individualistic responses of species to environmental change have important implications for environmental management. First, we cannot conclude from stable or increasing populations of a species in a particular locality, that the species is performing similarly in other regions. We might, however, be able to make predictions about population trends in nearby localities, because population trends tend to be spatially autocorrelated (Pollard 1991; Hanski and Woiwod 1993; B6hningGaese et al. 1994). But there is no justification for extrapolating from census data gathered in one region of the species' range to make inferences about other areas
485 or even more global trends. Conversely, when specieswide population trends have been identified, they cannot be assumed to be reflected in all local populations. For example, several large-scale studies have identified longterm changes in the overall abundances of insectivorous passerines (Terborgh 1989; Robbins et al. 1989; B0hning-Gaese 1992; BOhning-Gaese et al. 1993). Our analysis would suggest, however, that local and regional population trends are often extremely variable and sometimes even opposite to the global trends. Secondly, it is hazardous to use one or a few species as "indicator species" for general environmental changes (either degradation or improvement) that can be expected to affect other organisms similarly (see also Landres et al. 1988). Although we worked with a relatively homogeneous group of bird species, including some that were both very closely related and extremely similar in ecology, there was little consistency in population trends across species or regions. There are few shortcuts for environmental managers. Both the extrapolation of the results of a few local studies to a larger spatial scale, and the use of a few indicator species to monitor the status of a broader community must be regarded as suspect. The fact that species respond individualistically to environmental variation does not imply that species do not respond deterministically and predictably to abiotic conditions and to other species. On the contrary, working with the same bird species in the same strata, B6hningGaese et al. (1993) showed that factors such as increased predation pressure on their breeding grounds and perhaps deforestation on their wintering grounds can influence the long-term population trends of species over their entire geographical ranges. Between 1978 and 1987 a group of 14 long-distance migrants with traits indicating high vulnerability to nest predation (low, open nests and high cowbird parasitism) showed on average significant declines, whereas the majority of species without such traits increased or showed no significant change. By choosing sensible analysis techniques and by explicitly taking into account regional differences in population trends, pervasive factors influencing the more global population dynamics of a single species over its entire geographic range or a suite of species with similar ecologies and susceptibilities can be identified. Acknowledgements We especially thank the BBS volunteer ob-
servers, recorders, and coordinators for their dedication in organizing, gathering, and compiling the data that made this study possible. We are grateful to the U.S. Fish and Wildlife Service and especially to S. Droege for making the BBS data set available for us and J. Long for providing the Mantel test program. The research was supported by NSF Grant BSR88-07792 to J.H.B. and by a fellowship of the Studienstiftung des deutschen Volkes to K.B.G. Appendix 1 Species used in the analysis and their abbreviations (mostly the first two letters of the genus and species names): Viol Viph Vigi Vifl
Red-eyed Vireo (Vireo olivaceus) Philadelphia Vireo (Vireo philadelphicus) Warbling Vireo (Vireo gilvus) Yellow-throated Vireo (Vireoflavifrons)
Appendix 1 (continued) Viso Vigr Vibe Mnva Prci Lisw Heve Vepi Vech Veru Vepe Paam Deti Depe Decr Deco Dema Dece Depn Deca Dest Defu Dedo Devi Depi Deds Seau Seno Semo Opfo Opph Getr Icvi Wici Wipu Wica Seru Thlu Thbe Trae Trtr Cipl Cipa Ceam Sica Sicn Sipu Pabi Paat Paca Pahu Resa Reca Poca Sisi
Solitary Vireo (Vireo solimrius) White-eyed Vireo (Vireo griseus) Bell's Vireo (Vireo bellii) Black-&-white Warbler (Mniotilta varia) Prothonotary Warbler (Protonotaria citrea) Swainson's Warbler (Limnothlypis swainsonii) Worm-eating Warbler (Helmitheros vermivorus) Blue-winged Warbler (Vermivora pinus) Golden-winged Warbler (Vermivora chrysoptera) Nashville Warbler (Vermivora ruficapilla) Tennessee Warbler (Vermivora peregrina) Northern Parula (Parula americana) Cape May Warbler (Dendroica tigrina) Yellow Warbler (Dendroica petechia) Black-thr. Blue Warbler (Dendroica caerulescens) Yellow-rnmped Warbler (Dendroica coronata) Magnolia Warbler (Dendroica magnolia) Cerulean Warbler (Dendroica cerulea) Chestnut-sided Warbler (Dendroica pensylvanica) Bay-breasted Warbler (Dendroica castanea) Blackpoll Warbler (Dendroica striata) Blackburnian Warbler (Dendroica fusca) Yellow-throated Warbler (Dendroica dominica) Black-thr. Green Warbler (Dendroica virens) Pine Warbler (Dendroica pinus) Prairie Warbler (Dendroica discolor) Ovenbird (Seiurus aurocapillus) Northern Waterthrnsh (Seiurus noveboracensis) Lousiana Waterthrush (Seiurus motacilla) Kentucky Warbler (Oporornisformosus) Mourning Warbler ( Oporornis philadelphia) Common Yellowthroat (Geothlypis trichas) Yellow-breasted Chat (Icteria virens) Hooded Warbler (Wilsonia citrina) Wilson's Warbler (Wilsonia pusilla) Canada Warbler (Wilsonia canadensis) American Redstart (Setophaga ruticilla) Carolina Wren (Thryothorus ludovicianus) Bewick's Wren (Thryomanes bewickii) House Wren (Troglodytes aedon) Winter Wren (Troglodytes troglodytes) Sedge Wren (Cistothorus platensis) Marsh Wren ( Cistothorus palustris) Brown Creeper (Certhia americana) White-breasted Nuthatch (Sitta carolinensis) Red-breasted Nuthatch (Sitta canadensis) Brown-headed Nuthatch (Sitta pusilla) Tufted Titmouse (Parus bicolor) Black-capped Chickadee (Parus atricapillus) Carolina Chickadee (Parus carolinensis) Boreal Chickadee (Parus hudsonicus) Golden-crowned Kinglet (Regulus satrapa) Ruby-crowned Kinglet (Regulus calendula) Blue-gray Gnatcatcher (Polioptila caerulea) Eastern Bluebird (Sialia sialis).
Appendix 2 Biogeographic regions or strata used in the analyses and their coding according to Robbins et al. (1986): Atlantic Coastal Plain
04 05
Upper Coastal Plain Mississippi Alluvial Plain
Eastern Piedmont Plateau
l0 12 13 14 15
Northern Piedmont Southern New England Ridge and Valley Highland Rim Lexington Plain
486 Appendix 2 (continued) Eastern Piedmont Plateau (continued) 16 Great Lakes Plain 17 Wisconsin Driftless Area 18 St. Lawrence Plain 19 Ozark-Ouachita Plateau 20 Great Lakes Transition Appalachian Mountains and Boreal Forest 24 Allegheny Plateau 27 Central New England 28 Spruce-Hardwood Forest 30 Aspen Parklands Great Plains 31 Till Plains 32 Dissected Till Plains 33 Osage Plains 34 High Plains border 39 Unglaciated Missouri Plateau 40 Black Prairie
References Aldrich JW (1963) Life areas of North America. J Wildl Manage 27:530-531 Bthning-Gaese K (1992) Causes for the decline of European songbirds: an analysis of the migratory bird trapping data of the Mettnau-Reit-Illmitzprogram. J Ornithol 133:413-425 Bthning-Gaese K, Taper ML, Brown JH (1993) Are declines in North American insectivorous songbirds due to causes on the breeding range? Conserv Biol 7:76-86 Bthning-Gaese K, Taper ML, Brown JH (1994) Avian community dynamics are discordant in space and time. Oikos 70:121-126 Brown JH, Heske EJ (1990) Temporal changes in a Chihuahuan desert rodent community. Oikos 59:290-302 Clements FE (1916) Plant succession: an analysis of the development of vegetation (Publication 242). Carnegie Institute, Washington D.C. Cole KL (1982) Late Quaternary zonation of vegetation in the eastern Grand Canyon. Science 217:1142-1145 Coope GR (1986) The response of late Quaternary insect communities to sudden climatic changes. In: Gee FHR, Giller PS (eds) Organization of communities, past and present. Blackwell, Oxford, pp 421-438 Davis MB (1986) Climatic instability, time lags, and community disequilibrium. In: Diamond J, Case TJ (eds) Community ecology. Harper and Row, New York, pp 269-284 Droege S (1990) The North American Breeding Bird Survey. In: Saner JR, Droege S (eds) Survey designs and statistical methods for the estimation of avian population trends (Biological Report 90). U.S. Fish and Wildlife Service, Washington D.C., pp 1-4 Ehrlich PR, Dobkin DS, Wheye D (1988) The birder's handbook: a field guide to the natural history of North American Birds. Simon and Schuster/Fireside, New York Elias SA (1991) Insects and climate change. Bioscience 41: 552-559 Enemar A, Nilsson L, Sjostrand B (1984) The composition and dynamics of the passerine bird community in a subalpine birch forest, Swedish Lapland. A 20-year study. Ann Zool Fenn 21: 321-338
Foster DR, Schoonmaker PK, Pickett STA (1990) Insights from paleoecology to community ecology. Trends Ecol Evol 5: 119-122 Geissler PH, Noon BR (1981) Estimates of avian population trends from the North American Breeding Bird Survey. Stud Avian Biol 6:42-51 Gleason HA (1917) The structure and development of the plant association. Bull Torrey Bot Club 44:463481 Gleason HA (1926) The individualistic concept of plant association. Bull Torrey Bot Club 53:7-26 Graham RW (1986) Response of mammalian communities to environmental changes during the late Quaternary. In: Diamond J, Case TJ (eds) Community ecology. Harper and Row, New York, pp 300-313 Hanski J, Woiwod JP (1993) Spatial synchrony in the dynamics of moth and aphid populations. J Anita Ecol 62:656-668 Heggberget TM (1987) Number and proportion of southern bird species in Norway in relation to latitude, spring temperature and respiration equivalent. Holarct Ecol 10:81-89 Holmes RT, Sherry TW, Sturges FW (1986) Bird community dynamics in a temperate deciduous forest: long-term trends at Hubbard Brook. Ecol Monogr 56:201-220 J~irvinen O, V~iis~inenRA (1977a) Recent quantitative changes in the populations of Finnish land birds. Pol Ecol Stud 3: 177-188 J~irvinen O, V~is~inenRA (1977b) Long-term changes of the North European land bird fauna. Oikos 29:225-228 Kareiva P, Andersen M (1988) Spatial aspects of species interactions: the wedding of models and experiments. In: Hastings A (ed) Community ecology. Springer, Berlin Heidelberg New York, pp 38-54 Landres PB, Verner J, Thomas JW (1988) Ecological uses of vertebrate indicator species: a critique. Conserv Biol 2:316-328 Marchant JH, Hudson R, Carter SP, Whittington PA (1990) Population trends in British breeding birds. British Trust for Ornithology, Tring Pollard E (1991) Synchrony of population fluctuations: the dominant influence of widespread factors on local butterfly populations. Oikos 60:7-10 Robbins CS, Bystrak D, Geissler PH (1980) Techniques and results of nongame bird monitoring in North America. In: Oelke H (ed) Bird census work and nature conservation. Dachverband deutscher Avifaunisten, Gtttingen, pp 82-91 Robbins CS, Bystrak D, Geissler PH (1986) The Breeding Bird Survey: its first fifteen years, 1965-1979 (Resource publication 157). U.S. Fish and Wildlife Service, Washington D.C. Robbins CS, Saner JR, Greenber RS, Droege S (1989) Population declines in North American birds that migrate to the neotropics. Proc Natl Acad Sci USA 86:7658-7662 Root T (1988) Energy constraints on avian distributions and abundances. Ecology 69:330-339 SAS/STAT (1988) SAS/STAT user's guide, release 6.03 edn. SAS Institute, Cary Smouse PE, Long JC, Sokal RR (1986) Multiple regression and correlation extensions of the Mantel test of matrix correspondence. Syst Zool 35:627-632 Terborgh J (1989) Where have all the birds gone? Princeton University Press, Princeton Tilman D (1989) Ecological experimentation: strengths and conceptual problems. In: Likens GE (ed) Long-term studies in ecology. Approaches and alternatives. Springer, Berlin Heidelberg New York, pp 136-157 Valentine JW, Jablonski D (1993) Fossil communities: compositional variation at many time scales. In: Ricklefs RE, Schluter D (eds) Species diversity in ecological communities. University of Chicago Press, Chicago, pp 341-349