What percentage of the fall mallard flight do we kill? It is a critical issue, striking at the heart of whether hunting is responsible for declining numbers of greenheads winging southward each autumn and subpar breeding populations in the spring. In five of the past six years, the Midcontinent mallard breeding population has fallen below the 8.2 million goal set by the North American Waterfowl Management Plan.    

Adaptive Harvest, the U.S. Fish and Wildlife Service biological population model that purports to define the relationship between mallard deaths and mallard survival to set our hunting regulations, relies on band data to determine the kill. It keys on adult males. During the most recent five-year period (2003-07) for which all data is available, its predicted harvest rates for adult male mallards in the three western flyways ranged from 11.91 percent to 13 percent, as shown below.

The harvest rates for other mallard age-sex classes is fixed and based on the historic ratio of each to the adult male kill. The ratios are 1.5407 for juvenile males, .7191 for adult females and 1.1175 for juvenile females. The average harvest rate for each group during the past five years can be calculated as follows:

This tells us that on average each autumn and winter from 2003-07 we shot 13 percent of the adult male mallards, 19 percent of the juvenile males, 9 percent of the adult females and 14 percent of the juvenile females.

The question we must ask: Are these harvest estimates reasonable?

To find our answer, we must first assemble two additional sets of data. We begin (as does Adaptive Harvest) on the breeding grounds. The first column shows the year, the second the breeding population (BPoP) on the traditionally surveyed North American breeding grounds. The third column shows the breeding population the following spring in Year Y+1.

We also know that each year our breeding flocks produce young, which is measured by age-ratios – the number of juveniles per adult in the bags of hunters. The average annual mallard age-ratio for the three western flyways is as follows:

We now can begin our calculation. To simplify the math we will use five-year averages.

STEP 1. We first must determine the number of adult mallards that winged southward. We know that males outnumber females on the breeding ground. We assume 52 percent males and 48 percent females in the breeding population. We further know that females suffer higher mortality than males on the breeding grounds. We will assume a one percent mortality rate for males and 19 percent for females. We can calculate the fall flight as follows:

STEP 2. Band data tells us that hunters killed on average 13 percent of the adult males and 9 percent of the adult females. This allows us to calculate the adult kill.

STEP 3. We know from age-ratio data that hunters killed 1.31 juveniles per adult. To calculate the juvenile kill, we multiply the adult kill by the age-ratio (1,009,230 x 1.31). This tells us the juvenile kill totaled 1,322,196.

STEP 4. We know from band data that the juvenile kill averaged 16.5% of the juvenile fall flight.  Therefore, the juvenile mallard fall flight totaled 8,013,309.

STEP 5. We now can calculate both the total mallard fall flight and the kill.

STEP 6. We now can divide the kill by the fall flight to determine the average annual harvest rate. This tells us we killed 15.7 percent of the mallard fall flight.

Here is the problem. The data tells us that during the years 2003-07 a total of
14. 9 million mallards on average flew south each autumn. We shot an average of 2.3 million. We know that 7.5 million mallards on average returned to the breeding grounds. These are all the ducks we can account for. The numbers do not add up We face a significant deficit, as you can see below.

We find 5.1 million mallards missing in action, a number equal to 68 percent of the surviving breeding population. Put another way, for each duck we shot, 2.2 mallards died of unspecified, presumably natural causes. Natural mortality exceeded gunning mortality by 120 percent.

The September-to-March non-shooting mortality rate amounted to 34 percent of the fall flight, more than twice as high as the gunning mortality rate and three times greater than the 10 percent mortality rate on the breeding grounds, the time when hens are believed most vulnerable to predation. It suggests that from September-to-March a total of 21,250 mallards died each day of natural causes

The extraordinary number of missing ducks raises two questions: (1) What causes the death of 5.1 million missing mallards? Adaptive Harvest is silent on this issue, admitting it has no idea. (2) Where are the 5.1 million carcasses? If two ducks die of natural causes for each one we shoot, shouldn’t we encounter mallard carcasses more often when afield?  A documented natural die-off of 50,000 ducks of any species makes newspaper headlines, but these are exceedingly rare. When is the last time you can recall a mallard die-off of that magnitude? For that matter, when is the last time you read of a 10,000-duck die-off?

We could fine tune our analysis, adjusting for such things as crippling losses, juvenile gunning vulnerability and so on. To keep our analysis simplified I will spare you the calculations. The adjustments do not make a material difference.

It is important to note that Adaptive Harvest avoids its MIA problem by “adjusting for bias,” a polite way of admitting to a now-you-see-it, now-you-don’t statistical sleight-of-hand. It consists of inserting a calculated mathematical “adjustment” that effectively removes the MIA problem from its population model. This raises doubts about Adaptive Harvest’s “revealed truth” and its reliance on spurious harvest rates to determine the hunter kill.

 If Adaptive Harvest does not portray a reasonable “ducks in, ducks out” population assessment, how can waterfowl management rely on it to adjust the kill upward or downward to meet breeding-population goals?   The answer is that it cannot. 

The time is long over-due to recognize Adaptive Harvest for what it is – a poster child for junk science -- and toss it on the trash heap. We should demand from the U.S. Fish and Wildlife Service a rigorous, predictive population model on which to base our hunting regulations. 

__________________

¹ Young ducks are believed to come off the nest equally divided between males and females.  The both-sex juvenile kill rate therefore can be calculated as follows: (.19 + .14)/2 = .165.

² Harvest data from HIP surveys suggest hunters during the years 2003-07 shot and retrieved an average of 4.3 million mallards annually.  If we add a 20 percent crippling rate (defined as ducks downed but not retrieved) and add another 20 percent for ducks struck by shot that fly away, and further assume these ducks succumb to their wounds.  We find the hunter kill rises to 6 million mallards.  Thus, 6 million dead mallards, plus 7.5 million surviving breeders, accounts for 13.5 million mallards out of a fall flight of 14.9 million, leaving only 1.4 million (9% of the fall flight) to die of natural mortality.  Some studies have suggested only 8 percent of all mallards die of natural mortality over the autumn-winter period.