Population Studies. Steve Davis Department of Family Medicine, Box G Brown University Providence, RI

Transcription

1 Population Studies The Hooded Merganser A Preliminary Look at Growth in Numbers in the United States as Demonstrated in the Christmas Bird Count Database Steve Davis Department of Family Medicine, Box G Brown University Providence, RI abunoah@aol.com Peter Capobianco Brown University Providence, RI peter_capobianco@brown.edu The online Christmas Bird Count database can be a useful tool to guide further, more in-depth research on bird distributions across North America. A prior study in the journal North American Birds (Davis, 2001) reported the growth in the New England winter population of Hooded Mergansers (Lophodytes cucullatus). The Audubon Christmas Bird Count database was used as the basis of the data, with data downloaded self-serve to the public through the historical results tool set. Although that study showed a marked increase in the numbers of Hooded Mergansers, it examined a limited region. What was not investigated was whether the growth in numbers in New England was mirrored in other regions of North America, or whether the New England increase was at the expense of other regions. This larger-scope initial study looks at the reports of Hooded Mergansers throughout the 49 states (not including Hawaii, which is not in the American Birding Association area) and the District of Columbia for the 86th through 105th Christmas Bird Counts (winters of through ). This study supports the hypothesis that the New England increase is consistent with the growth of the Hooded Merganser population in most of the United States. Additionally, the data analysis in this study highlights that the online CBC database should offer optional methods of standardization techniques beyond what is currently available on the Christmas Bird Count website. Overview Christmas Bird Count results, as reported in the historical results section of the Audubon online database ( were the source of this study s data. Tables of the CBC results for specific species, areas, and time periods are easily created online. For visual learners, maps and line graphs for each combination of categories can also be created by pressing a few keys. (For those of us who have been analyzing CBC data through the years by extracting species numbers from the Christmas Bird Count reports, this is a fantastic breakthrough and simplification; now time and effort can be spent in analysis rather than in data extraction! Thank you, National Audubon and Cornell.) Although there are some analytic complications in the interpretation of Steve Davis is a family doctor and clinical associate professor of family medicine at Brown University; two of his hobbies are birding and math. Peter Capobianco is a senior biology/geology major at Brown University; he has extensively photographed birds and butterflies. Both have been mentored in their birding careers by mutual friend Hugh Willoughby, the dean of Rhode Island birders. THE 106TH CHRISTMAS BIRD COUNT AMERICAN BIRDS 27

2 individual species numbers that make comparisons difficult (see Table 1), the raw data on the numbers of Hooded Mergansers reported on CBCs suggest that there has been a significant increase in the population throughout most of the 49 states. When each state is tabulated for the numbers of Hooded Mergansers reported for each of the past 20 years, 39 of the 50 regions show apparent significant positive trends. Methods and Results We downloaded Hooded Merganser data from the Christmas Bird Count database directly through the historical results link of the CBC website (Butcher, ). Given the fact that this was a preliminary look at what is happening with Hooded Mergansers continentally, and also that we disagree with the pre-programmed effort adjustment as available through the CBC historical results tool set, we elected to use only raw numbers of Hooded Mergansers without effort adjustment, and we made only minor standardization of the data based on the number of CBCs used and the expected numbers of birds per count over time. We then downloaded data from several different regions across the United States and compared the Hooded Merganser results from region to region. Using these methods, 39 states appear to have significant increases in Hooded Merganser numbers for the 20-year period (the District of Columbia does not). Rough and preliminary linear regression analysis can be used to produce approximate trend lines for each state. In this analysis, we used the raw data for the number of Hooded Mergansers counted in the CBCs for each state for each year. The value of the trend in Table 1 indicates the annual rate of growth in the number of Hooded Mergansers. The values for each state are grouped below geographically for comparison purposes. The column Significant? tells whether the trend appears to be increasing significantly in that state. The columns 86 and 105 below give the estimated Table 1. State-by-state Rough Trends of Hooded Merganser Numbers on CBCs Trend Significant? Predicted # in 86 Predicted # in 105 (Northeast) Maine 13.1 yes New Hampshire 13.4 yes Vermont 8.5 yes Massachusetts 94.7 yes Rhode Island 17.3 yes Connecticut 70.2 yes New York yes New Jersey yes Pennsylvania 37.6 yes Delaware 7.5 yes (Southeast) Maryland 49.2 yes District of Columbia 1.3 no 34 Virginia 97.1 yes North Carolina 80.8 yes South Carolina 77.9 yes Georgia 58.5 yes Tennessee 32.8 yes Alabama no 1500 Florida yes Mississippi no 2113 (Midwest) Ohio 25.6 yes Indiana 36.8 yes West Virginia 8.0 yes Kentucky 5.4 no 362 Illinois 73.1 yes Michigan 36.8 yes Wisconsin 6.7 yes Minnesota 5.2 no 0 Iowa 7.2 yes Missouri 21.6 yes (Central/Great Plains) Louisiana 8.0 yes Arkansas 10.6 no 423 Texas 1.2 no 760 Oklahoma 29.0 yes Kansas 30.1 yes Nebraska 1.3 yes 0 24 South Dakota 0.2 no 1 North Dakota 0.1 no 0 Montana 8.5 yes (West) Idaho 30.7 yes Utah 27.7 yes Colorado 25.0 yes New Mexico 1.9 no 35 Arizona 1.2 yes 9 33 Nevada 3.9 yes California 50.7 yes Oregon 27.2 yes Washington 57.2 yes Alaska 0.04 no AMERICAN BIRDS

3 numbers of Hooded Mergansers for those counts if the trend line had been followed exactly. For states with a trend that is not statistically significant, the baseline predicted number is given; negative numbers are rounded up to zero. Let us consider some individual states. First we look at Mississippi, which for the past 20 years has had the largest number of Hooded Mergansers on Christmas Bird Counts. Even though its rate of increase is not significant, at per year it has the highest yearly growth of Hooded Mergansers, though this fluctuates widely (see Figure 1). Of concern in Mississippi are the last few years from the 101st count to the 103rd there was a drastic decline in Hooded Mergansers. In the 101st CBC, a count high of 10,882 Hooded Mergansers was tallied; for the 102nd CBC the number dropped to 3660, then to 1103 in the 103rd CBC. Fortunately, the number rebounded somewhat for the 104th count at 2263, and the 105th count at Partly because of these decreases in Mississippi, the total number of Hooded Mergansers for the nation also decreased from 51,877 in the 100th count and 51,810 in the 101st count to 47,875 in the 102nd count and 38,371 in the 103rd count. A similar rebound was noted for the 104th count at 41,793, as well as for the 105th at 42,401. It can be noted in Figure 2, the graph showing the sum of the Hooded Mergansers for the United States, that the total for the 103rd count reflects what is happening in Mississippi. It is interesting to compare the Mississippi and U.S. graphs. Contiguous with Mississippi, Alabama (Figure 3) has the only negative trend for all the 50 regions. Actually, if the one outlier value for the 100th count of 3919 is eliminated, the negative trend becomes even more pronounced. This observation, combined with an evaluation of the Mississippi graph for the last 10 years, suggests that perhaps the status of Hooded Mergansers in the Deep South is not as robust as it might appear. Figure 1. Mississippi Figure 2. United States Figure 3. Alabama THE 106TH CHRISTMAS BIRD COUNT AMERICAN BIRDS 29

4 Figure 4. Florida Figure 5. Louisiana Figure 6. Arkansas Numbers from Florida (Figure 4) and other areas of the Southeast, however, appear to be going strong. Florida and Georgia show strongly increasing numbers, and other southeastern states (South Carolina, North Carolina, and Virginia) are all showing positive trends. Even in Tennessee, more inland, Hooded Merganser numbers are increasing on CBCs. In Louisiana (Figure 5), just west of Mississippi, numbers are more lackluster. In recent decades there have not been large numbers of Hooded Mergansers in Louisiana, although reports continue to grow slowly. The graph has an interesting biphasic pattern, with small numbers and a general downward trend in the first decade, and then a relatively large increase in numbers and another downward trend for the past eight to 10 years. Arkansas (Figure 6), Louisiana s neighbor to the north, has also had a recent downward trend. As reported in our previous articles, Hooded Mergansers have increased in numbers on Christmas Bird Counts in the Northeast. New York, New Jersey, Massachusetts, Connecticut, Maryland, and Pennsylvania all have significant and increasing trends. The graph for Connecticut (Figure 7) is a good example of results from these states. In the heartland, Hooded Mergansers have also done well, with numbers increasing strongly in Illinois, Indiana, Michigan, Tennessee, and Missouri. Closer to the Mississippi River, however, Iowa, Wisconsin, and Minnesota only show slight increases. Indiana (Figure 8) illustrates this pattern. Farther west, most of the Great Plains states have few Hooded Mergansers and small growth trends. Nebraska (Figure 9) has small but strongly increasing numbers, partly a statistical artifact because of the very few Hooded Mergansers counted in the first six years of this period. And Texas is, as usual, in a class by itself (Figure 10), with good annual numbers of Hooded Mergansers that are neither increasing nor decreasing. The central mountain west states show reasonably positive numbers. 30 AMERICAN BIRDS

5 Colorado, Utah, and Idaho have strong increasing trends, but both farther south (New Mexico, Arizona, and Nevada) and farther north (Wyoming and Montana) trends are more neutral. Perhaps this is a demonstration that to tally reasonable numbers of ducks in winter there must be water available! Hooded Mergansers are doing well in the West, with strongly increasing numbers in California, Oregon, and Washington. However, the graph for Oregon (Figure 11) shows a decrease for the last six years. Hooded Merganser numbers in California show a less striking leveling off. Far to the north, Alaska has few Hooded Mergansers and no particular growth or trend in their numbers (Figure 12). Discussion The analyses above were derived from the raw numbers of Hooded Mergansers reported and have not been standardized by any adjustment for observer effort, such as by numbers of counts, party hours, or number of observers. Clearly, before we can make definitive statements about the population trends of Hooded Mergansers, analyses must be conducted that properly adjust for changes in observer effort, because analysis of the raw numbers of birds reported can be very misleading. Observer effort has been generally increasing in the Christmas Bird Count each year. Due to this increased effort, one might expect to find annual increases in the raw numbers of Hooded Mergansers reported as reported here, even if there was no actual increase in the Hooded Merganser population. Nonetheless, by using the online CBC tools to take an initial look at Hooded Merganser numbers over the past two decades, the steep increase in raw numbers reported suggests that the number of Hooded Mergansers have in fact increased robustly. The distribution maps that can be generated online are also a good representation of the increases. The following three maps (Maps 1, 2, and 3) are for the 85th count, the Figure 7. Connecticut Figure 8. Indiana Figure 9. Nebraska THE 106TH CHRISTMAS BIRD COUNT AMERICAN BIRDS 31

6 Figure 10. Texas Figure 11. Oregon Figure 12. Alaska 94th count, and the 103rd count, which is the most recent one available. Given the strong indications from this study that Hooded Mergansers are increasing in frequency in most areas of the United States on Christmas Bird Counts, it seems appropriate that further, more rigorous study is warranted. Major advances have been published in recent years regarding the methods appropriate for in-depth, statistically backed investigations into large citizen science generated databases such as the Christmas Bird Count. In fact, papers have been published in the pages of American Birds over the past few seasons as well as in other more technical journals describing these methods (Sauer and Link, ; and Sauer, Niven, and Link, ). Any such in-depth studies will certainly require that researchers request complete sets of raw data from Audubon, which they will then use to pursue their own methods of inquiry and statistical analyses. When viewed through the Christmas Bird Count website, CBC data can be standardized ( effort corrected ) between circles and years by considering the number of party hours and party miles of effort for each count circle. It is interesting to note the differences in graphs when looking at raw numbers of birds reported over time and that of effortcorrected data. However, a few cautionary comments are in order regarding the current data display through Historical Results. The online database does standardize the tabulated state results by party hour for each year; however, it considers only the party hours of the count circles that reported Hooded Mergansers (or any species of interest) in each state or region selected. Thus effort data from counts not reporting Hooded Mergansers are not included, and not all the effort from a state is used as the divisor. Consequently, the results can be misleading. For example, consider that the CBC historical results tool reports that there was one Hooded Merganser reported in Iowa in the 83rd count year, from one count circle that had 56 party 32 AMERICAN BIRDS

7 Map 1: 85th CBC Map 2: 94th CBC Map 3: 103rd CBC hours, which calculates to 0.02 Hooded Mergansers per party hour for Iowa. However, there were actually 19 other count circles in Iowa that season that did not report Hooded Mergansers. Those count circles amassed additional party hours, which equates to Hooded Mergansers per party hour for Iowa. This difference in birds per party hour is a factor of 20 for some investigations, it can be very important to include the total effort in the state, not just that of the counts reporting the species of interest. Also, although the number of party hours for each individual count circle tends to remain approximately the same from year to year (a very rough personal observation), the number of count circles in each state or region has generally increased through the years; consequently, the number of party hours per state has not remained stable through the passing decades. We strongly recommend the development of additional data summary options, all thoroughly documented and properly labeled, for display of data to the public through the historical results tool set on the CBC website. Graphs and Hooded Merganser drakes and hen, Weekapaug, Rhode Island. Photo/Geoff LeBaron charts should be generated showing raw numbers of birds, birds per party hour for counts in the region reporting the species of inquiry, and birds per party hour for the species of inquiry for all circles in the region. Each graph and chart should be clearly labeled, and an explanation of the different measures presented. Additionally, it would be desirable to have more versatility in both species and region options for inquiry through the historical results tool set. Furthermore, and perhaps most obviously, the mapping data for species from seasons after the 103rd CBC should be available. These more inclusive tools would allow even more complete initial inquiries into the Christmas Bird Count database online through the CBC website, helping define the direction of future research projects. Summary This study demonstrates that, based on a preliminary analysis of count data that did not include any attempt at effort adjustment, there appears to be an increase in the numbers of Hooded Mergansers in the United States over the past two decades. Thirty-nine of the 49 states showed increases with statistically significant trend lines, based on raw counts of birds (unadjusted for the amount of observer effort). Recent decreases in the number of reported Hooded Mergansers from the Deep South are of concern. Further, more rigorous study should bear out these predictions. Moreover, the value of the online CBC historical results tools would be greatly increased if changes are made to allow for more than one type of effort adjustment. References Butcher, Greg. Enjoying Christmas Bird Count Data on the Internet. American Birds, Vol. 57, The 103rd Christmas Bird Count, , pp Davis, Stephen. The Change in the Numbers of Wintering Hooded Mergansers (Lophodytes cucullatus) in New England. Bird Observer, Vol. 38, No. 3 (2000), pp Davis, Stephen. The Increase in Wintering Hooded Mergansers in New England. North American Birds, Vol. 55, No. 4 (2001), pp Root, Terry. Atlas of Wintering North American Birds: An Analysis of Christmas Bird Count Data, University of Chicago Press, Sauer, John, and William A. Link. Using Christmas Bird Count Data in Analysis of Population Change. American Birds, Vol. 56, The 102nd Christmas Bird Count, , pp Sauer, John, Daniel K. Niven, and William A. Link. Statistical Analyses Make the Christmas Bird Count Relevant for Conservation. American Birds, Vol. 58, The 104th Christmas Bird Count, , pp The Christmas Bird Count database online: Hooded Merganser drakes, Weekapaug, Rhode Island. Photo/Geoff LeBaron THE 106TH CHRISTMAS BIRD COUNT AMERICAN BIRDS 33