| OTHER FEATURE STORIES |
|
SCWA Chapter Local Volunteer Conservation |
for South Carolina
By Walt Rhodes and Tommy Strange South Carolina Department of Natural Resources,
July 21, 2003
South Carolina (SC) is one of 17 states and 6 Canadian provinces that comprise the Atlantic flyway (AF). Flyways are a concept of migration corridors developed as a means of expressing the direction of passage and the geographic distribution of waterfowl between breeding and wintering areas. The United States (US) is delineated into four administrative regions following political boundaries of states with common borders. Even though flyways are arranged in a north-south orientation, Bellrose (1976) acknowledged that the passage of migrating waterfowl was seldom in a north-south fashion.
Beginning in the early 1990s, continental waterfowl biologists and managers began establishing a process known as Adaptive Harvest Management (AHM) to determine continental US waterfowl hunting seasons. Williams and Johnson (1995) stated that the adaptive approach explicitly recognizes that the consequences of hunting regulations cannot be predicted with certainty, and provides a framework for making objective decisions in the face of that uncertainty (USFWS 2001). Appropriate season alternatives are based on combinations of mallard breeding population size, environmental conditions (pond numbers), and competing model weights (Johnson et al. 1997).
Mallards breed in significant numbers across the northern US through Canada and into Alaska. Geographic differences in life-history traits suggest that optimal levels of harvest differ. Because no flyway exclusively receives mallards from one breeding stock, the challenge of AHM is to suggest flyway-specific harvest strategies that account for multiple breeding stocks that are exposed to a common harvest (USFWS 2001).
When AHM was implemented in 1995, it was based on mallard breeding stocks and pond conditions in the mid-continent region (federal survey strata 1-18, 20-50, and 75-77), and in Minnesota, Wisconsin, and Michigan. Recognizing an increase in eastern breeding mallard populations, and thus, increased harvest potential, several AF states called for the establishment of an eastern mallard model. Based on mallards breeding in southern Ontario and Quebec and in the northeastern US, the eastern mallard model has been used to establish AF duck harvest strategies since 2000.
Waterfowl biologists in SC have contended for decades that the majority of duck species wintering in the state originate from breeding areas associated with the Great Lakes and prairie-pothole region of the north-central US and Canada, the area encompassed by the mid-continent mallard model. Relative to other AF states and provinces, mid-winter survey (MWS) data reveals that SC has historically wintered a large proportion of species commonly associated with the mid-continent breeding areas (Northern shovelers, gadwalls, blue-winged teal, Northern pintails, American wigeon, American coots, and mallards). Canvasbacks, which are rare in SC, and green-winged teal are exceptions; the latter, while very abundant in SC and a breeder in the mid-continent region, primarily originates from the Canadian Maritimes.
Simulations conducted by Johnson (pers. comm.) determined that the AF waterfowl seasons based on the eastern mallard model would be in the liberal package (60 days/ 6-bird bag) nearly 100% of the time whereas waterfowl seasons based on the mid-continent mallard model would not be liberal as often. The dynamic wetland conditions of the mid-continent region and resulting variable mallard breeding populations are the cause for the difference. More recently, Johnson (2003) evaluated multiple-stock AHM rather than solely mallards under the mid-continent model. He found that the current AHM protocol should not prevent several other sympatric duck species, except scaup, American wigeon, and Northern pintails, from maintaining average population sizes above their respective North American Waterfowl Management Plan goals.
Realizing the potential that SC could be offered a liberal package under the eastern mallard model when its mid-continent-derived wintering waterfowl population should be subjected to more restrictive seasons recommended by the mid-continent mallard model, the state’s waterfowl biologists have become increasingly concerned about the negative consequences of such a scenario. A potential outcome is over-harvest of South Carolina’s wintering waterfowl population.
The objective of this report is not to evaluate AHM or determine the decline of mallards in SC. Rather, our intent is to review past analyses of mallard distribution in regards to SC, reexamine various datasets, and lay the foundation that the state (and possibly the Southeast AF) belongs as a separate management unit under the mid-continent mallard model. We appreciate the review of this document by Dr. Dave Otis.
Background
Historically, large numbers of mallards have wintered in SC. Over 100,000 mallards wintered there in the 1960s, and the majority of mallards in the AF wintered in SC until the mid-1980s. Less than 5,000 mallards have been observed on the SC MWS in recent years.
Several studies of mallard recovery distributions that include data about SC have been conducted over the last three decades. Most of this information is contained in the “Mallard Volumes” but a few researchers have analyzed data at a finer scale.
One of the earliest works is by Anderson and Henny (1972). They divided North America north of approximately 36 degrees latitude into 16 major breeding reference areas (RA) and 44 minor reference areas. RAs were grouped based on similar recovery distribution patterns. The authors tabulated the distribution of band recoveries from each reference area from 1950-69. The proportion of bands recovered in the AF from the Missouri River Basin and Great Lakes RAs was often small (<10%) but the largest proportion (>20%) of recoveries in the AF was in SC. Recovery proportions in SC became smaller in an easterly direction through E. Ontario-W. Quebec, Mid-Atlantic, and Northeastern U.S. reference areas
As part of the analysis of survival and recovery rates of mallards banded postseason in SC, 1960-78, Rakestraw (1981) also provided a distribution of band recoveries. He found that 30-41% of recoveries (direct and indirect), depending on age and sex, occurred back in SC and another 30-38% were in the Mississippi flyway (MF). Fourteen to 29% of recoveries were in Canada but a province was not specified. Only 7-11% of recoveries occurred in the AF outside of SC.
Munro and Kimball (1982) analyzed the distribution and derivation of the mallard harvest in North America based on preseason bandings (1961-75). They stated SC is representative of the southeastern AF in terms of magnitude and derivation of the mallard harvest. Total SC mallard harvest was derived mainly from the Great Lakes (29%), E. Ontario-W. Quebec (24%), the Missouri River Basin (13%), and N. Saskatchewan-N. Manitoba-W. Ontario (11%) RAs, which was similar to most of the MF and portion of the AF, New York and New England excluded. The area currently covered by the state plot survey accounted for <10% of the SC harvest.
Trost (1985) made an assessment of the distribution and derivation of the mallard harvest in the US based on recoveries of breeding-ground bandings from 1975-84. Canada was excluded as a harvest area from his results. He found that SC accounted for <1% of the mallard harvest that occurred in the US. Proportion of the mallard harvest for SC from major breeding RAs was E. Ontario-W. Quebec (42%), N. Saskatchewan-N. Manitoba-W. Ontario (20%), Great Lakes (13%), the Missouri River Basin (7%), and S. Saskatchewan (5%).
Nichols and Hines (1987) provide recovery distributions of winter-banded mallards. Their Southern Atlantic Flyway RA included SC, GA, and FL. The RA only wintered 1-2% of North American mallards, but they stated SC was the most important state in this RA. Recoveries of birds banded, 1950-77, in this RA occurred primarily in the AF and MF. For mallards winter banded in the minor reference area of SC and GA, nearly an equal number of total recoveries (all ages and sexes combined) occurred in the AF and MF (40% and 38%, respectively). In the AF, 74% of the recoveries were in SC. In order of abundance, MI, WI, MN, IL, and OH accounted for two-thirds of the recoveries in the MF. Eighteen percent of total recoveries occurred in Canada, with southeastern Ontario (68%) and southern Manitoba and Saskatchewan (26%) the major recovery areas.
Sheaffer and Malecki (1996) estimated the distribution and derivation of the 1990-94 harvests of mallards in the Atlantic Flyway. They found that SC accounted for <3% of the distribution of the harvest of mallards banded in the northeastern US and eastern Canada. SC was grouped with FL, GA, and NC to estimate harvest derivations, which could bias estimates because NC has been shown to be different than SC (Otis, pers. comm.). The derivation they calculated revealed a north-south gradient, with lesser proportions of northeastern US- and eastern Canada-banded mallards in a southerly direction. Percent derivations for the Southern region were 92%, 36%, and 49% for adult males, adult females, and immatures, respectively.
Dr. Dave Otis of the IA Cooperative Fish and Wildlife Research Unit at IA State University spent considerable time while at the Co-op Unit at Clemson University analyzing information pertaining to mallards in North and South Carolina. Like previous studies, his initial investigations based on 1961-98 bandings showed that SC derived mallards from the N. Saskatchewan-N. Manitoba-W. Ontario, E. Ontario-W. Quebec, Missouri River Basin, Great Lakes, Mid-Atlantic and Northeastern U.S. RAs. Otis’ analysis of mallards banded postseason in SC, 1961-90, was similar to Rakestraw’s findings. Thirty-seven to 42% of recoveries, depending on age and sex, were in SC. The MF, most notably the region defined by the states of MN, WI, IA, IL, IN, OH, MI, MO, and KY, accounted for 26-29%. Approximately 15-21% occurred in Canada, primarily Ontario and Quebec. Only 6-9% of recoveries occurred in the AF outside of SC.
Otis attempted to determine the 1990-98-harvest derivation for SC but he cautioned there were data constraints. The most important RAs were N. Saskatchewan-N. Manitoba-W. Ontario, E. Ontario-W. Quebec, Great Lakes, Mid-Atlantic, and Northeastern U.S. He stated his results placed more importance on the latter two RAs than earlier analyses. However, because his harvest-distribution analysis did not reveal any obvious changes in the proportion of breeding mallards from the Mid-Atlantic and Northeastern U.S. RAs recovered in SC, he warned it was difficult to know if this result was due to the increased breeding population estimates, which, in turn, increased the weight assigned in the derivation calculations, or a real increase in the importance of eastern breeding stocks to the SC harvest.
Initial work would indicate that there is currently enough justification to place SC under the mid-continent mallard model. No one to date has proven any temporal changes in mallard harvest distribution and derivation. It appears that the greatest proportion of mallards harvested in the state originate from areas not represented by the eastern mallard model.
Additional Analysis
All previous analyses have been conducted with political boundaries defining regions. While this may be suitable for management, it can cause errors in data interpretation due to the juxtaposition and size of states and provinces. We propose a longitudinal approach that eliminates state/province boundaries. Distributions of pre-season bandings recovered in SC and recovery distribution of mallards winter-banded in SC are the two main datasets. Harvest survey data was not examined because of potential contamination by hand-reared mallard releases in the Mid-Atlantic region and, more recently, in SC. Population weighting was not needed because we are not comparing RAs (Munro and Kimball 1982) nor are we attempting to determine harvest derivations. After discussing an approach with Dr. Otis, it was decided to keep any analysis as simple as possible.
We plotted the distribution of all mallards banded pre-season north of 36-degrees latitude and recovered in SC by decade, 1960-2001. This approximates the area encompassed by the two mallard models. We used Mardia’s test as described by Nichols and Haramis (1980) and Munro and Kimball (1982) to test the null hypothesis of no difference between the distribution of pre-season bandings recovered in SC by decade, 1960-2001. The distributions between the 4 decades appear similar. There was no difference (P > 0.09) between 1970-79 and 1980-89 and 1980-89 and 1990-2001. However, 1960-69 differed (P < 0.001) from the 3 other periods and 1970-79 differed slightly (P = 0.03) from 1990-2001. Mardia’s test is known to be sensitive to small changes in distributions. Most of the differences we recorded are probably a result of the high number of recoveries from eastern Wisconsin and Michigan, especially during the 1960s. Biologically, it seems that the overall distribution of pre-season band recoveries in SC has been constant since 1960.
A longitudinal approach to the distribution of pre-season banded mallards recovered in SC is important in regards to which model SC should fit. Between 60- to 110-degrees longitude, which encompasses >99% of SC’s pre-season banded recoveries, at least 74% of the recoveries were west of 79 degrees. Moreover, the swath between 79 and 81 degrees has become an increasingly important area, which may result from a reduced vulnerability to hunting relative to other mallards migrating to SC from areas farther west. For example, the banded cohort of locally-produced mallards from around Lake Erie would be exposed to hunting for a shorter period of time before they made an overnight flight to SC. A similar cohort of prairie-pothole-produced mallards headed for SC would be exposed to hunting beginning in mid-September, and would have had to cross waterfowl seasons in 4 states before any reached a similar point as the Lake Erie cohort to make an overnight migration to SC. Thus, the number of banded mallards available to SC hunters would be less. The point is the importance of the Lake Erie vicinity to SC could be exaggerated while the influence of the western areas could be understated.
Banding effort, reporting rate, and hunter effort are known to influence band-recovery data. None of these factors affect the interpretation of the results in regards to SC. Although pre-season banding has increased since 1960, the proportion across longitudinal gradients has been similar. There is no reason to believe reporting rate has declined in SC. On the contrary, it has probably increased with the inception of the 1-800 number and hunters are more likely to report bands as their distance increases from banding stations. While duck-hunter numbers have been declining in some locales, the number of duck stamps sold, active adult duck hunters, and days hunted have been increasing in SC since 1986, and currently exceed estimates from the early 1970s (Serie and Raftovich 2002). All of this evidence only strengthens our results that the majority of mallards harvested in SC originate from a region encompassed by the mid-continent mallard model.
Distribution patterns of winter-banded mallards can provide information on migration corridors, and allow researchers to make hypotheses about potential breeding RAs. More than 37,000 mallards have been banded postseason in SC between 1961-90, with greater than 4,000 recoveries. Very little postseason banding took place during the 1990s but there has been a renewed effort in recent winters. As expected, most of the recoveries occur back in SC. A large number were recovered at the western end of Lake Erie in MI, OH, PA, and Ontario, southeast WI, and along the upper Mississippi River and northeastern ND. A smaller number were recovered around Lake Ontario but very few came from the Chesapeake region (except SE VA) or east of eastern NY and NJ. When winter bands were adjusted for regional reporting rates (Nichols et al. 1995), 89% were from west of 79 degrees. The distribution of all winter-banded recoveries appears to be strongly mid-continent.
Conclusions for SC can be drastically affected since so many SC pre-season band recoveries are clustered in an area around central and western Lake Erie. A north-south line placed east or west of the cluster could have two different meanings.
The 86-degree longitude line was selected when the AF September blue-winged teal season was evaluated (Fuller 2002). This line correlates with the farthest western edge of the eastern mallard model, however, we feel selection of this line for mallards is problematic in relation to the mid-continent versus eastern model. If we choose this boundary, more of SC’s recoveries would be east of the line, thereby implying the eastern mallard model would be appropriate. But, the pre-season bandings that tip the scales in favor of the eastern model for SC originate from MI and OH. MI’s mallard breeding population is a component of the mid-continent mallard model and Ducks Unlimited is currently studying OH’s breeding population as a portion of the Great Lakes Mallard project. We feel is it more prudent to select a breakpoint line that is farther east and places MI appropriately to the west. We do not deny the importance of the London and Sudbury portions of Ontario and northwest PA to SC, but this region occupies a very small percentage of the total area used for the eastern mallard model.
The justification to form a separate management unit under the mid-continent mallard model for SC (and possibly the Southeast AF) is warranted. Failure to do so will not affect the continental mallard population but it could have dire effects on the mallard population that has wintering fidelity to SC. Additional analysis may be possible with current band recoveries, but previous researchers have noted problems. A renewed emphasis on winter banding will help but this takes time before results can be interpreted, possibly too long. A well-designed and funded satellite-telemetry project for mallards might provide more immediate results, if added information is even needed.
Indications are the Southeast AF receives a major segment of its wintering waterfowl from mid-continent breeding RAs. Because FL and GA have historically wintered very few mallards, SC is probably in a position unlike any other state in the US. We propose that the Mallard Committee determine if enough information currently exists to recommend SC being placed under the mid-continent mallard model.
Literature Cited
Anderson, D.R. and C.J. Henny. 1972. Population ecology of the mallard: I. A review ofprevious studies and the distribution and migration from breeding areas. U.S. Fish andWildl. Serv. Resour. Publ. 105. 166pp.
Bellrose, F.C. 1976. Ducks, Geese and Swans of North America. Stackpole, Harrisburg, PA.540pp.
Fuller, J. 2002. Final report on experimental September teal seasons in the Atlantic Flyway, 1998-2000. Report by the Wood duck and other Dabblers Committee to the Atlantic Flyway Technical Section. 19pp.
Heusman, H.W. and J. R. Sauer. 2000. The northeastern state’s waterfowl breeding population survey. Wildl. Soc. Bull. 28:355-364.
Johnson, F.A., C.T. Moore, W.L. Kendall, J.A. Dubovsky, D.F. Caithamer, J.R. Kelly, Jr., and B.K. Williams. 1997. Uncertainty and the management of mallard harvests. J. Wildl.Manage. 61:202-216.
_____. 2003. Population dynamics of ducks other than mallards in mid-continent North America. U.S. Fish and Wildl. Serv., unpubl. rep. 15pp.
Munro, R.E. and C.F. Kimball. 1982. Population ecology of the mallard: VII. Distribution and derivation of the harvest. U.S. Fish and Wildl. Serv. Resour. Publ. 147. 127 pp.
Nichols, J.D. and G. M. Haramis. 1980. Sex-specific differences in winter distribution patterns of canvasbacks. Condor 82:406-416.
_____ and J.E. Hines. 1987. Population ecology of the mallard: Winter distribution patterns and survival rates of winter-banded mallards. U.S. Fish and Wildl. Serv. Resour. Publ. 162. 154pp.
_____, R.E. Reynolds, R.J. Blohm, R.E. Trost, J.E. Hines, and J. P. Bladen. 1995.
Geographic variation in band reporting rates for mallards based on reward banding. J. Wildl. Manage. 59:697-708.
Rakestraw, J.L. 1981. Survival and recovery rates of mallards banded postseason in South Carolina. J. Wildl. Manage. 45:1032-1036.
Serie, J. and B. Raftovich. 2002. Atlantic flyway waterfowl harvest and population survey data. U.S. Fish and Wildl. Serv. 107pp.
Sheaffer, S.E. and R.A. Malecki. 1996. Distribution and derivation of the 1990-94 mallard harvest in eastern North America. Northeast Wildlife 53:45-54.
Trost, R.E. 1985. A preliminary assessment of the recent distribution and derivation of the mallard harvest in the United States based on recoveries from breeding ground bandings,1975-1984. U.S. Fish and Wildl. Serv., unpubl. rep. 69pp.and Wildlife Service. 2001. Adaptive harvest management: 2001 Hunting season.U.S. Dept. Interior, Washington, D.C. 47pp.
Williams, B.K. and F.A. Johnson. 1995. Adaptive management and the regulation of waterfowl harvests. Wildl. Soc. Bull. 23:430-436.
 |
|
|
|
