Hawk Migration on the Kittatiny Ridge and Climate Change
By Glenn Nelson
ABSTRACT: Much has been recorded about the phonological changes in the migration cycles of birds, as related to climate change. These studies typically focus on modeling for future changes in migration dates for a given range of species. However, there appears to be little focus on recording the change in numbers of birds seen in migration, as they relate to temperature and weather phenomena. This study hopes to find first, whether climate change is observable over time in a given location, and second, whether those changes appear to have a correlation with the total number of migrating hawks. Anecdotally, it has been noted that certain species will stay in their breeding ground as long as there is food source availability. Thus, we hypothesize that as temperature rises, fewer birds will get the impulse to migrate during the typical migration season, (or in warmer years at all), lowering the number of birds observed during the migration period. Initial results from a statistical analysis of NOAA weather data and migration data from Hawk Mountain suggest that there is a likely correlation between warming regional climate and lower than average hawk flights, particularly since the 1980s, which has been a period of increasingly rapid warming. There are a few possible explanations for this change. While this study hypothesizes fewer birds are migrating, it is also possible they are using thermals instead of the updrafts found along the mountain ridges. Whichever explanation accounts for the observed changes, it still represents a change in migration behavior, which could have ecological consequences.
- INTRODUCTION:
Hawk migration has long been seen as a harbinger of seasonal changes, associated with the coming winter. However, less attention has been paid to how the phenology of hawk migration may change over time, on the various hawk watches in North America. While still politically debated, the scientific community has come to a consensus that the world is warming and that this effect is likely anthropogenic.[1] However, the warming effect is not universal, and climates are changing in a variety of ways. Thus it is important for this study to establish a trend in temperature within the given study years, as a primary goal, before trying to understand the relationship between temperature and hawk migration.
The usual volumes associated with hawk watching in the northeast, such as Pete Dunne’s Hawks In Flight and Jerry Ligouri’s Hawks From Every Angle, generally chart when to expect high flights of certain species.[2] (An example of such migration time tables can be seen in Figure 1.1).[3] Looking at these charts, the question arises as to whether such projections change and whether any changes are associated with climate change.
Fig: 1.1
There is a good deal of available research which has produced projections about the changing timetable for raptor migration, but the available research focuses more on the changing dates of a specie’s migration, as opposed to whether the total numbers changed as well. For instance, a 2012 study by Josh Van Buskirk, asserts that:
“The migratory period has become more extended, especially for short-distance migrants. Opposite responses during the two seasons had the effect of extending time spent to the north of the study area, by up to 30 days in some species since the early 1970s. These phenological shifts—potentially related to climate change— are causing dramatic changes in the annual cycle of North American raptors.”[4]
Another study, conducted in Europe by Mikael Jaffre, (et al), resulted in similar findings.
“We found that when the temperatures increased, birds delayed their mean passage date of autumn migration. Such delay, in addition to an earlier spring migration, suggests that a significant warming may induce an extension of the breeding-area residence time of migratory raptors, which may eventually lead to residency.”[5]
Yet, a change in total numbers should also be concerning for two reasons. First, hawk migration data is often used to estimate the health of a species population. This fact is highlighted in a forest service study by Kyle McCarty and Keith Bildstien. Their study points out that, “One particularly cost-effective method for monitoring populations of these birds is to sample regional and even continental populations at traditional migratory bottlenecks and concentration points.”[6] Second, even if the population remains healthy but the migration numbers change, it could signal a change in habits that may have detrimental effects to a species, or other ecological repercussions within the bird’s traditional range. For instance, if a species lingers too long in their breeding range, they may stress available resources in the breeding area, or be surprised by a rapid change in weather such as an unusually cold year. Furthermore, prey species may suffer from exponential population growth in areas that were in the hawks’ previous migration and southern range. Similar relationships have been noted in lower trophic levels, when keystone predators are removed from ecosystems.[7]
Inevitably, there are a number of factors that complicate the results. First, the period of acceptable data has been shortened to control for DDT. As was evidenced by Rachel Carson, the spraying of DDT had an effect on migratory bird populations, including raptors.[8] From 1934 to 1972, DDT spraying was allowed in the United States, (as well as in the migration range countries, where DDT still sometimes persists). DDT has since been illegalized in the United States, (1972).[9] It is clear that numbers jump drastically between the 1960s and the 1970s, as the DDT ban was enacted and conservation efforts for migratory raptors were put into place, (fig 1.2). Since there is no effective way to control for this effect, data prior to the illegalization of DDT was discarded.
Fig: 1.2
Decade 1=1950
There are also issues in the standardization of the data, since the period of the hawk watch varies from year to year, and from weather and other factors. From 1943-to1945 no hawk watch was conducted, due to World War 2. More telling results could be gleaned from comparing population numbers to the percent of migrating birds, but systematic data on that topic does not currently exist, and collecting it would present logistical challenges.
Other issues must also be raised as possible factors. I had been informed, and since found anecdotal accounts in both Jerry Liguori’s Hawks From Every Angle, as well as Don Heintzelman’s Guide to Hawk Watching in North America, as well as research from Tarra E. Gettig, that hawks have a tendency to fly in advance of and behind certain frontal systems.[10] Since the number of frontal systems affecting this area per year is difficult to accurately calculate, it is not a factor included in this study. However, it is something that should be looked at in the future, in order to glean the full picture of what weather processes are affecting migration.
There is also research to suggest that hawks may be using thermals if the temperature is high, or if wind is unavailable. When hawks use thermals, they soar much higher, and are thus harder to observe. This is what is posited in Michael J. Lanzone’s study “Flight responses by a migratory soaring raptor to changing meteorological conditions This offers an alternate explanation for changes in migration numbers.”[11]
Overall, this study hypothesizes that, first, climate change will be observable in the temperature data, and may be observable in the rainfall data. I expect that this will affect migration numbers negatively and residency numbers positively, though the effect is likely gradual, requiring a large data set, or decade means, to see the trend.
- METHODOLOGY:
Fig: 2.1
This study is designed firstly to establish whether climate change is occurring in the Lehigh Valley area of the Kittatiny Ridge. To this end, climate summary data for the study period was requested of and provided by NOAA, for the Allentown weather station, ______ miles East of Hawk Mountain. All data were collected for the period of the fall hawk watching season on Hawk Mountain, between August and December. This data was used to establish decade means and charted as the dependent variable, while the decade functioned as the independent variable. (fig 2.1) Temperatures were measured in degrees Fahrenheit, in order to be more relatable to a domestic audience. As can be seen in figure 2.2, the decade mean for temperature generally rose, excepting one anomaly. The temperature for the period of the 2000s was 53.6 degrees (f), about a degree above the period average.
Fig 2.2:
Yearly hawk totals from Hawk Mountain Sanctuary in Kempton, Pennsylvania were then plotted as the dependent variable, (fig. 2.3). This showed us both the average flight, as well as the trend over the study period.
Fig 2.3
Precipitation data was collected similarly, and used as a stand in for weather variability.
Fig 2.4
The data were then compared using standard deviation to show variability, and pearson’s correlation coefficient, using the following equations:
Standard Deviation:
Pearson’s Correlation Coefficient:
Finally, data were collected from the Winter Raptor Survey results from Pennsylvania Birds Magazine.[12] The data are only available during the 2000s, and are thus of limited applicability, but can be used to suggest a possible explanation for the observed phenomena.
- Results:
The following data tables represent the results from the statistical analysis at hawk mountain, (Tables 3.1 – 3.3) and (Figures 3.1-3.2)
Table 3.1
Decade Averages | Temp | Rainfall | Number of Hawks | BW | SS | RT | |
50 | 50.8 | 19.4 | 15535 | 8600.63 | 2858.55 | 2539.09 | |
60 | 51.3 | 16.9 | 16162 | 9729.82 | 2187.18 | 2788.82 | |
70 | 52.2 | 20.4 | 21513 | 11011.82 | 5535.55 | 3393 | |
80 | 52 | 17.6 | 22042 | 8220.82 | 7190.09 | 4035.18 | |
90 | 52.5 | 18.7 | 18631 | 6426.27 | 5316.36 | 3676.64 | |
00 | 53.6 | 20.5 | 17971 | 7052.55 | 4397 | 3090.36 | |
correlation | 0.3788 | 0.1126 | -0.4651 | 0.4459 | 0.4260 | ||
correlation from 1980s | -0.8336 | -0.8725 | -0.4645 | -0.9179 | -0.9972 | ||
correlation from 1970s | -0.852923146 | -0.399197261 | |||||
Fig. 3.1: X= Temperature, Y=Hawk Migration Numbers
Fig. 3.2: X=Rainfall, Y=Hawk Migration Numbers
Table: 3.2
Year | 1950-1960 | 1960-1970 | 1970-1980 | 1980-1990 | 1990-2000 | 2000-2010 |
Average Temp. (F) | 50.8 | 51.31 | 52.15 | 52 | 52.46 | 53.63 |
Average Rainfall (In) | 19.44 | 16.9 | 20.41 | 17.62 | 18.71 | 20.53 |
Average Number of Hawks | 15535 | 16162 | 21513 | 22042 | 18632 | 17971 |
Sharp-shinned | 2859 | 2187 | 5536 | 7190 | 5316 | 4397 |
Broad-winged | 8601 | 9730 | 11012 | 8221 | 6426 | 7053 |
Red-tailed | 2539 | 2789 | 3393 | 4035 | 3677 | 3090 |
Standard Deviations | ||||||
Temp. | 1.14 | 1.36 | 1.69 | 1.44 | 1.49 | 1.53 |
Rainfall | 3.75 | 2.87 | 4.28 | 4.81 | 4.37 | 4.92 |
Number of Hawks | 3008 | 2170 | 8679 | 4202 | 4105 | 3606 |
Sharp-shinned | 800.3763319 | 564.5103751 | 3155.206788 | 1997.933055 | 1146.100543 | 1057.040586 |
Broad-winged | 2443.611642 | 1709.996188 | 6480.008084 | 3327.400451 | 3009.409214 | 2702.533158 |
Red-tailed | 714.1127998 | 633.1852522 | 1010.681651 | 839.4813659 | 889.580606 | 861.2376295 |
Table 3.3
Year | Winter hawks | RT | Temp | Rainfall |
2010 | 5915 | 2665 | 53.1 | 19.96 |
2009 | 5789 | 2275 | 53.3 | 20.13 |
2008 | 5634 | 2390 | 52.3 | 21.68 |
2007 | 4948 | 2218 | 55.2 | 19.68 |
2006 | 5359 | 2184 | 54.9 | 18.5 |
2005 | 4512 | 2610 | 55 | 23.22 |
2004 | 3959 | 2052 | 53.6 | 26.01 |
2003 | 2374 | 1182 | 54.4 | 29.64 |
2002 | 2539 | 1399 | 52.8 | 19.75 |
2001 | 2192 | 1141 | 55.1 | 10.55 |
Statistical data do not always tell their own story (or they tell multiple stories depending on how they are interpreted). As has been shown over the years by the climate change debate, data can be less or more convincing, depending on how it is analyzed. This study looks at total hawks per year, averaged over a decade. It can be said that this was done to improve R2 values, in order to strengthen certain conclusions. Certainly, when looked at on a closer scale, like taking yearly averages as individual anomalies and running a correlation, the R value is reduced. However, this was not our justification for looking at means.
The justification for using decade averages has to do with the high degree of variability in individual anomalies, concerning hawk migration data. As has arisen in the climate debate, you may have hot years and you may have cold years, but individual anomalies do not establish a larger trend. In the case of looking at long term trends, relationships can be obscured by scaling too close, and what may be only a small change from year to year presents itself as quite pronounced over time.
When looking at hawk data (fig 3.3) this is particularly pronounced. You have good years and bad years, and there don’t appear to be any particular reasons for the individual anomalies. However, the means help to smooth out the often pronounced outliers, (such as 1979, which had more than 40,000 hawks, in what is the single biggest year in Hawk Mountain’s history).
If we go back as far as 1930, and we look at the lowest outliers, there is certainly a linear improvement. However, the number of days on the lookout has also improved, as has the reliability of the count’s protocol. This is a problem with the Winter Raptor Survey data as well, since in recent years count hours have gone up, alongside the number of resident hawks.
Fig: 3.3
There are a number of problems with looking at the data in this kind of resolution, since there are changes in variables, through a number of periods during the study. The charted data showed lower than average flights through most of the DDT period. In the period immediately following, there appears to be a bump, followed by a gradual decline. The polynomial trend line does a much better job of expressing the shifting periods, than does the linear trend line. This helped to separate the data by three distinct periods. The DDT era, the post-DDT recovery, and a notable decline.
The first period was defined as 1946-1974, since this period was dominated by the effects of DDT, as a limiting factor to hawk populations. In this period we see far fewer than average hawk flights, at 15,458 birds per year. We also see below average precipitation and temperatures. Despite that fact, there is a warming trend present, even in these years. Between 1950 and 1970, the average decade temperature increased by nearly 2 degrees (f). In this period, the correlation coefficient for temperature was -0.25. This aligns with the hypothesis, that as temperature increases, fewer birds would migrate, though, during this period the effect is slight. Furthermore, the correlation coefficient for precipitation is .011, which is too weak to suggest any relationship.
The next period, 1974 to 1992, represents what would appear to be a recovery from DDT and a relative temperature stagnation. These two phenomena transpired to produce higher than average flights for the period at 22,933 per year, on average. The correlation between temperature and hawks for the period was
-0.068, which is too weak to suggest any relationship. What is more, the -0.14 coefficient for precipitation suggests that there is a weak inverse relationship between rainfall and hawk numbers.
The final period looked at, was between 1980 and 2014. While there is an overlap with the previous period, there seemed to be certain trends in the charts, which needed to be followed back to 1980. In this period, mean birds per year fell to 19,540, which is 225 below the average for the whole study period of 19,765. Furthermore, both average temperature and average rainfall were higher during this period. Both temperature and precipitation correlation coefficients increased their previous trends to -0.097 and -0.227 respectively. Both of these trends remain slight, but suggest a very gradual, possible relationship, where both temperature and rainfall adversely effect hawk flights.
Finally, the data was broken up into decade averages between 1970 and 2010. By zooming out on the data, certain trends did emerge as more apparent. First, through the 1980’s bird numbers are on an increasing path, which suggests that there has been a certain level of success in conservation efforts. However, the 1990s and 2000s show a marked decline in those averages. The averages remain higher than they had been in 1950, but suggest an alarming trend. Furthermore, by averaging, we can see a steady rise in temperatures, though there appears to be three segments. Temperatures rise into the 1970s, stagnation for a decade, and then rising rapidly (by 1.1 degree (f)) between the 1990s and 2000s. In that same period, average hawks per decade drop off by 340. While, if we look at the correlation coefficient for the averages of each decade since 1950 we get the slightly positive .0379, if we look at it just spanning the high point in the 1970s, up to 2010, we get a more alarming coefficient of -0.85. The precipitation averages also rise in this period, and amount to a coefficient with hawk numbers of -0.4. These numbers are suggestive of a strong relationship between high temperatures, high rainfall and fewer hawk flights. While the temperature relationship relates closely with the hypothesized results, the precipitation data seems somewhat less suggestive of a relationship between variable weather and hawk flights.
While over the study period the story appears to be a slight linear positive, a closer examination of recent trends is more alarming. By putting the data into three periods we can see the gradual changes, but by averaging the decades and examining the relationship more closely, we can see a more startling trend, which appears to establish a strong relationship between rising temperature and lower average hawks per decade. Overall, 72.25% of the variation in average fall hawk numbers by decade can be described by a negative linear relationship with temperature. Only about 16% of the variation can be described by a linear relationship with precipitation in inches.
Finally, in looking at the decade of data from the Winter Raptor Survey, we can see a that hawks observed per year is rising, though with negligible correlation to fall temperatures. While the data is incomplete, and can’t be fully analyzed into decade means, there are a few emergent trends worth noting. First, 2010 saw 5915 hawks observed, which is 1,593 birds more than the period average of 4,322. There is a P = -0.26, which is a bit below statistically significant and suggests that only about 6-7% of the variation in hawks can be described by the fall temperature anomaly. This result may be more significant if averaged out.
Fig: 3.4
- DISCUSSION:
In the process of conducting this study, a number of questions have arisen. First there has been the problem of accounting for the DDT period and the period immediately afterward, (which appears to exhibit a recovery). While this is a very important finding, it is not helpful to the study, because the exceptionally low number of hawks in the DDT period creates an artificially low floor, while the recovery due to conservation efforts presents an artificially high ceiling. In many ways, it shows that human effort has helped to solve a human problem, which is optimistic for future challenges. The idea that we can overcome our mistakes once they are identified is encouraging, but the problem of identifying the problem remains.
There were certain trends, which appeared from the data to be correlative, that emerged when looking at the data from the height of the recovery period (1970) into the present. First, temperatures, which had stagnated during the recovery period rose after 1980; at first gradually and then rapidly after the year 2000. Furthermore, there appeared to be greater disparity in the scatter in rainfall data in the current period. These trends are magnified when examined by decade.
The increase in climate change related measurements, and a significant drop in birds per year, appears to have a significant, if gradual relationship. This relationship may have significant ramifications ecologically. First, as the paper by ____ notes, migration data is one of the cheapest ways to monitor hawk populations. If migration is occurring less frequently, or differently, it will be more difficult to monitor populations. Birds that stay in their breeding range to the north longer would likely effect both their breeding habitat and their wintering habitat, by stressing prey populations in one and not limiting them enough in the other. Even if that is not the correct assumption, and hawks are just travelling at higher altitudes on thermals, as ___ suggests, this is still significant, as it impedes our ability to monitory the populations of migratory raptors. This study also has ramifications for phenology, as hawk migration is seen as a harbinger of the changing seasons.
- CONCLUSIONS:
While I can see that there may be concerns that the decade averages could be construed as an alarmist manipulation of statistics, I believe that it was necessary to view the data on such a scale, in order to establish long-term trends. While the overall trends show a growing number of migrating hawks since the 1930s, which do not correlate with climate change, (when plotted in a linear regression model)—the polynomial projection suggests a different relationship; depending on what period you are focused on. Averaging decades and discarding data that presented too many variables provided an amplified and simplified window for comparison. I was able to observe how higher and lower numbers, in each category, were skewing the mean in different periods. Furthermore, the patterns that appeared gradual, in a year over year analysis, emerged in a more pronounced manner. The years where DDT likely effected hawk populations were particularly apparent, as were the recovery years. The recovery years appear to be aided by a stagnation in temperature, as well. However, as the world economy began growing again in the 1980s, and greenhouse gas emissions rose precipitously though the 1990s (citation needed), many of the gains made by conservation efforts in the 1970s have been lost. The relationship between higher temperatures and lower decade averages of hawk numbers, since 1970, is thus an alarming one.
When we look at the data we see this process, but it occurs gradually, as indicated by the polynomial line on figure ______. There are waves of both progress and setbacks, that cannot be shown by a simple linear projection, even in the decade means. It is important to isolate the current period, one which is nearly a degree Fahrenheit warmer than the study average, and which has risen at a faster rate, from periods where other factors were likely dominant. Even when we look at the yearly averages, (instead of the decade averages) the correlation exists, but it is not as drastic, because we see it spread out over time. The decade averages isolate each period, and make a vague picture clear.
I think it is important to acknowledge how gradual this process has been, and that the low rate of change certainly effects the correlation coefficient. That said, the data set does tell a story. For instance, there have been 21 years in which high temperatures correlated with low flights, 11 of them occurring in the period of rapid warming after 1980. Higher than average precipitation correlates with lower hawk flights for 23 of the study years, but only 8 since 1980. Meanwhile, there has been an increasing frequency of higher than average temperatures since 1980, with 24 higher than average years, including every year since 2001. This compares with 13 before 1980. In that time, there has only been one year significantly below average. In the same period, there have been 14 years with higher rainfall than average. The frequency appears to have accelerated since the year 2000, with half those years occurring in that period. The number of lower than average rainfall years since 1980 have also increased from 5 in the previous 46 year period to 5 in just 34 years, with 2 since the year 2000. This is suggestive of more extreme and variable weather. There have been 15 years with lower than average hawk flights in this period, 10 of them since the year 2000. These numbers are why I think it is important to focus heavily on this period.
When we look at the R value between temperature and hawk migration numbers between the decades of 1970-2000, the value is a statistically significant -0.8529. When that is couple with a 0.73 R2 in a linear regression model, with a slope of -2434.419, it is clear that there is a strong relationship with temperature, that cannot be ignored. While there can be a number of explanations for this trend, all of the explanations have to acknowledge this relationship. The probability exists that 72.25% of the variation in hawk migration numbers can be described by the hypothesis of y= 2434.4x + 148029. In other words, there is a high probability that temperature is affecting hawk migration. This cannot be said with more certainty, because there are too many variables, and hawk migration itself is extremely variable.
We still have a net positive picture, as even the decade averages show. The decade of the 2000s has better hawk flights than that of the 1950s, when DDT was an active problem. If we plot the data in a linear manner, we have an upward trend. However, since the 1980s that trend has reversed, and that reversal has accelerated in the 2000s, in correlation with temperature and to a lesser degree rainfall.
I believe that this evidence is suggestive, then, of a change in migration that is likely due to climate change. The process is gradual, and it would be expected to be, but it appears to be occurring. While there are certain ways that this study could be strengthened, I think it shows that this understudied area would benefit from a greater intensity of study.
Appendix 1: Hawk Migration Data from Hawk Mountain Sanctuary and Weather Data from NOAA.
Year | Temp | Percipitation | BW | SS | RT | # of Hawks |
1934 | 50.9 | 22.83 | 3 | 1703 | 5426 | 7874 |
1935 | 50.3 | 20.01 | 3873 | 4168 | 3214 | 14681 |
1936 | 50.7 | 15.45 | 6990 | 4406 | 3162 | 16083 |
1937 | 50.8 | 14.85 | 4343 | 4791 | 4932 | 15446 |
1938 | 52.1 | 18.16 | 10754 | 3105 | 2228 | 17007 |
1939 | 51.4 | 10.84 | 5736 | 8620 | 6208 | 22488 |
1940 | 49.5 | 17.45 | 3159 | 2406 | 4725 | 11228 |
1941 | 53.3 | 11.68 | 5170 | 3908 | 4698 | 15424 |
1942 | 50.5 | 25.71 | 4362 | 3200 | 2378 | 11014 |
1943 | 49.8 | 17.06 | ||||
1944 | 50.6 | 14.41 | ||||
1945 | 50.2 | 20.88 | ||||
1946 | 52.8 | 14.62 | 2886 | 2382 | 2306 | 8729 |
1947 | 52.3 | 13.12 | 6664 | 1726 | 1680 | 11366 |
1948 | 52.3 | 16.07 | 15026 | 1650 | 2343 | 20483 |
1949 | 52.5 | 17.02 | 9579 | 2963 | 2749 | 17092 |
1950 | 50.8 | 16.19 | 5305 | 2667 | 3674 | 13366 |
1951 | 50.8 | 22.5 | 10997 | 3008 | 2307 | 17890 |
1952 | 51.5 | 26.45 | 12603 | 3566 | 2754 | 20737 |
1953 | 53 | 16.87 | 7247 | 2791 | 2051 | 13542 |
1954 | 51.6 | 20.44 | 5956 | 3183 | 2070 | 12606 |
1955 | 51.1 | 24.19 | 9542 | 4709 | 3764 | 19867 |
1956 | 51.8 | 16.7 | 8734 | 2048 | 1525 | 13469 |
1957 | 52.4 | 15.11 | 8935 | 2662 | 2730 | 15858 |
1958 | 49.9 | 15.58 | 8880 | 1752 | 2951 | 15128 |
1959 | 53.8 | 19.99 | 5301 | 2825 | 1904 | 11585 |
1960 | 50.6 | 19.83 | 11107 | 2233 | 2200 | 16832 |
1961 | 53.5 | 12.5 | 8642 | 1723 | 2566 | 14716 |
1962 | 49.1 | 20.59 | 8254 | 2181 | 2772 | 14651 |
1963 | 50.9 | 18.4 | 9791 | 1518 | 3402 | 15900 |
1964 | 51.7 | 11.62 | 10180 | 1259 | 2626 | 15202 |
1965 | 51.7 | 15.12 | 9235 | 3103 | 3297 | 17371 |
1966 | 51.4 | 18.73 | 10110 | 2883 | 2126 | 16582 |
1967 | 49.8 | 16.37 | 8000 | 2330 | 1854 | 13604 |
1968 | 52.5 | 16.38 | 14041 | 2253 | 3765 | 21789 |
1969 | 50.2 | 17.85 | 8515 | 2670 | 3566 | 16176 |
1970 | 53 | 18.6 | 9153 | 1906 | 2503 | 14960 |
1971 | 53.8 | 24.17 | 5603 | 2135 | 1781 | 10536 |
1972 | 50.8 | 22.24 | 8131 | 2233 | 3463 | 15285 |
1973 | 54 | 20.17 | 6404 | 3347 | 3098 | 14448 |
1974 | 51.3 | 24.03 | 9146 | 4477 | 3658 | 18519 |
1975 | 53.1 | 22.75 | 10390 | 5354 | 2880 | 20121 |
1976 | 48.1 | 18.48 | 8461 | 5376 | 3694 | 18941 |
1977 | 51.7 | 25.78 | 13009 | 10612 | 3504 | 29123 |
1978 | 51.8 | 19.18 | 29519 | 6826 | 2852 | 40576 |
1979 | 53.1 | 19.1 | 11173 | 10306 | 4175 | 27639 |
1980 | 53 | 10.01 | 10141 | 8319 | 5715 | 26495 |
1981 | 50.3 | 11.95 | 8660 | 9464 | 3939 | 24890 |
1982 | 52.8 | 16.54 | 7163 | 4541 | 5025 | 18742 |
1983 | 52.4 | 22.53 | 6922 | 6517 | 3954 | 19681 |
1984 | 54.3 | 14.27 | 13619 | 3796 | 3157 | 22343 |
1985 | 52.3 | 23.6 | 3415 | 5766 | 2895 | 13931 |
1986 | 50.9 | 19.46 | 13996 | 9239 | 3305 | 29200 |
1987 | 51.2 | 24.07 | 8409 | 6776 | 4215 | 22366 |
1988 | 51.2 | 14.1 | 5944 | 6714 | 4687 | 20034 |
1989 | 49.8 | 16.5 | 7504 | 9832 | 3710 | 24700 |
1990 | 53.8 | 20.78 | 4656 | 8127 | 3785 | 20084 |
1991 | 53.3 | 14.41 | 5858 | 5678 | 2970 | 17219 |
1992 | 50.6 | 17.68 | 10661 | 4629 | 3288 | 21125 |
1993 | 52 | 26.26 | 3592 | 5449 | 3744 | 15829 |
1994 | 53.7 | 17.34 | 3513 | 4934 | 4433 | 15713 |
1995 | 51.3 | 16.68 | 10077 | 6217 | 4854 | 24363 |
1996 | 52.2 | 24.85 | 1809 | 4468 | 2734 | 11589 |
1997 | 51.4 | 15.81 | 5519 | 4218 | 2402 | 15533 |
1998 | 54.6 | 12.6 | 9935 | 5835 | 4331 | 24238 |
1999 | 54 | 22.75 | 8634 | 4416 | 4999 | 22491 |
2000 | 50.2 | 16.68 | 6435 | 4509 | 2903 | 16767 |
2001 | 55.1 | 10.55 | 3843 | 4817 | 3741 | 16137 |
2002 | 52.8 | 19.75 | 12228 | 3211 | 3499 | 22212 |
2003 | 54.4 | 29.64 | 6134 | 3651 | 3385 | 16474 |
2004 | 53.6 | 26.01 | 6387 | 2958 | 2847 | 15027 |
2005 | 55 | 23.22 | 5273 | 4545 | 4551 | 18346 |
2006 | 54.9 | 18.5 | 11804 | 5480 | 3898 | 24940 |
2007 | 55.2 | 19.68 | 7836 | 5099 | 2426 | 19495 |
2008 | 52.3 | 21.68 | 4289 | 3358 | 1807 | 12205 |
2009 | 53.3 | 20.13 | 6640 | 4299 | 1762 | 15590 |
2010 | 53.1 | 19.96 | 6709 | 6440 | 3175 | 20492 |
2011 | 55.7 | 40.07 | 13323 | 4447 | 1697 | 22902 |
2012 | 54.2 | 18.97 | 8394 | 5222 | 2876 | 20078 |
2013 | 52.4 | 21.41 | 6430 | 3772 | 2030 | 15271 |
2014 | 53.8 | 12.7 | 6369 | 4772 | 2266 | 17415 |
[1] UN climate report.
[2] Ligouri: pg 11, Heintzleman: pg 79-80.
[3] Hawk Mountain
[4] Buskirk: pg 1. http://www.zora.uzh.ch/70159/1/auk%252E2012%252E12061.pdf
[5] Jaffre, et al: pg 1.
[6] McCarty and Biddlestein: pg 718. http://www.fs.fed.us/psw/publications/documents/psw_gtr191/psw_gtr191_0718-0725_mccarty.pdf
[7] environmental sustainability
[8] (citation from Silent Spring).
[9] Citation needed
[10] Liguori: pg 10, Heintzleman: pg 77-84, Gettig (http://www.gammathetaupsilon.org/the-geographical-bulletin/2010s/volume53-2/article2.pdf).
[11] http://rsbl.royalsocietypublishing.org/content/8/5/710
[12] PA Ornithological Society