How should we interpret this spring’s North American breeding-ground survey, the results of which will not be known until late July?

The U.S. Fish and Wildlife Service describes the continental duck survey as its statistical “crown jewel.” It begins in South Dakota and extends northward to the Beaufort Sea. It ranges from eastern Manitoba westward across Alaska. Since 1955 it has monitored the fluctuating populations of our 11 most prized duck species. But today the most important tallies involve mallards and potholes – the two critical elements of Adaptive Harvest Management that determine autumn hunting regulations.

As you read this biologists are flying across the northern prairies counting both ducks and numbers of potholes. Yet, questions remain. How accurate are the breeding-population counts? How should we interpret the results? The answers to both questions are problematical.

The subject of accuracy often arises in hunters’ conversations. A question that I am frequently asked prompts an exchange that goes like this:

“How many ducks are there?”

“You mean, if God counted the ducks, how many would He count?”

“Yes.”

“I have no idea – and neither does the service.”

The surveys are not designed to produce a precise duck census. (Keep in mind the service calls them “estimates.”) But this does not reduce their value or suggest they are fatally flawed. They remain the most valuable measure of our migrant flocks. The confounding problem is that they reveal a relative truth that often is masked by fluctuations in numbers of potholes that cause a change in duck behavior.

We can illustrate the problem by examining the number of potholes and the reaction to pothole fluctuations by three species – gadwalls, pintails and mallards. We begin by looking at potholes over the past decade. This takes us from 1993 (the last year of the 1987-93 drought) to 2002. Low water is defined as less than four million potholes, average water from four million to less than six million, and high water as six million or more.

Figure 1. Pothole numbers during the years 1993-2002 have ranged from a low of 2.8 million to a high of 7.5 million. Drought conditions are indicated in pale yellow, average-water years in green and high water years in blue. Source: USFWS.

As you can see, the number of potholes during the past decade has ranged across the spectrum – three drought years, three average-water years and four years of high water. Potholes numbers peaked in 1996-97. Conditions have gradually deteriorated since that time, notwithstanding a couple of up-ticks

The wide range of conditions allows us to examine the reactions of ducks to varying water conditions. As most of you know, in dry years many species over-fly the prairies to spend the summer in the vast expanse of boreal forest or barren tundra – areas that are lightly surveyed or not surveyed at all. The ducks return to breed on the prairies when water fills the potholes.

We begin our analysis by looking at gadwalls. Ron Reynolds, the highly esteemed U.S. Fish and Wildlife Service waterfowl biologist, notes that gadwalls, unlike mallards and pintails, do not over-fly the prairies in great numbers during dry years. They mostly remain on the prairies where they are less likely to be overlooked by biologists. The species therefore may be considered the best measure of our prairie-nesting ducks.

Figure 2. Breeding flocks of gadwall increased steadily after the 1994 return of water to the prairies. The population peaked at 3.9 million in the spring of 1997 and has declined steadily since that date. Source: USFWS.

As you can see, the gadwall population in 1997 peaked at 3.9 million. Only 2.2 million were found last spring – a 44 percent decline over the past five years. Because gadwalls do not over-fly the prairies, we can conclude this decline represents a real loss. It is not a statistical anomaly caused by ducks being “unsurveyed” in the vast expanse of boreal forest and northern tundra.

Further, if we view gadwalls as the most accurate barometer for all prairie-nesting species, then we can conclude that the breeding-population declines posted for all species last spring also were real and not statistical anomalies.

This will not sit well with some pintail shooters who have complained in recent years that the pintail survey has underestimated the population during drought years. They base their argument on radio-telemetry studies that tracked pintails to breeding sites unsurveyed by biologists. The breeding-population counts for pintails during the past decade can be seen in the following graph.

Figure 3. Over the past decade pintail breeding populations have ranged from a high of 3.6 million in 1997 to a low of 1.8 million in 2002. Source: USFWS.

Because water conditions this spring have improved across the Canadian provinces, many pintails that over-flew the prairies are expected to settle on the intensively surveyed prairies. This can be expected to boost the low pintail count.

This raises the question: How do we determine whether a jump this spring in pintail numbers represents a true population increase or a “false positive” based on a re-distribution of our breeding flocks into intensively surveyed areas?

Assuming water conditions this spring are average, we should expect a breeding population around 2.9 million. This is the average breeding population for the 1998 and 2001 average-water years. If the spring pintail count exceeds 3.3 million (the number counted in 2001, we can reasonably conclude the population has increased.

If the breeding population is less than 2.5 million (the 1998 breeding count), we can conclude that the population continues its decline – a situation that should raise alarms and calls for more severe hunting restrictions, especially in those flyways where the decline of this species has been most severe.

If biologists count between 2.5 million and 3.3 million pintails, the continental population can be described as stable.

Mallards are another story. Their breeding numbers over the past 10 years can be seen in the following graph.

Figure 4. Mallard breeding populations over the past decade have fluctuated from a low of 5.7 million in 1993 to a high of 10.8 million in 1999. The 2002 breeding population totaled 7.5 million. Source: USFWS.

Mallards, like pintails, often have hidden reserves that swell the prairie-nesting population during years of high water. (Note the jump from 5.7 million in 1993 to 7.0 million mallards in 1994, a time when potholes increased from 3.6 million --a drought year -- to nearly six million.)

Although the number of potholes is expected to increase this year, biologists do not expect an increase in the mallard breeding population. They have predicted a 17 percent population decline this spring under average water conditions. This would represent the fourth straight year of decline and would tell us there are no significant numbers of mallard “reserves” in unsurveyed areas to bolster the beleaguered pothole-nesting breeding flocks. This would suggest the decline is not a statistical anomaly. It would tell us the decline is both real and cause for great concern.

Thus, these three species illustrate why breeding population data cannot be taken at face value. Any biological conclusion must be viewed in proper context. It is equally important to note that this does not in any way reduce the value of these surveys. They remain the best measure of our breeding populations.