< Back

A simple model based on averages (deterministic) was used to predict changes in bison populations and/or seroprevalence rates should a given alternative be implemented. Because a single severe winter, such as the 1996–97 winter, could alter estimates of bison numbers significantly, the analysis also includes a section on the effects of “stochastic” events on the population size.

The deterministic model predicts the continued implementation of alternative 1 would result in a growing bison population. From 1997 to 2006, the bison population would increase at 4% per year to approximately 3,100. Management actions in this alternative would not measurably affect the age/sex distribution or reproductive rates of bison in this or any alternative except for alternative 5. Bison distribution outside the park is indicated in table 4. In this, and all other alternatives except alternative 5, 100–200 bison would freely range on public lands in the Eagle Creek/Bear Creek area.

Alternative 2 would result in the largest and fastest growth of the bison population of all alternatives. From 1997 to 2006, the population is expected to increase to 3,500, moderately more bison (14%) than in alternative 1.

Alternative 3 would result in growth of the bison population, with numbers controlled primarily through hunting. From 1997 to 2006, the bison population would be expected to increase from about 2,200 to 3,500 (average increase 6%/year). Limited capture operations, agency shooting, hunting, and periodic severe environmental conditions would likely maintain the bison population near the upper management range of 1,700 to 3,500. It is estimated that alternative 3 would result in moderately more bison in the population (14% increase) compared to alternative 1.

In alternative 4, bison population numbers would be controlled through capture, shipment of seropositive bison to slaughter, and hunting. This alternative would result in a slowly increasing bison population with lower population numbers than alternatives 1, 2, 3, or 6. From 1997 to 2006, the bison population would be expected to increase from about 2,200 bison to 2,800 (average increase 3%/year). This would be a minor decrease (8% lower) in bison population size relative to alternative 1.

For alternative 5, the bison population would be expected to decline from 2,200 bison to approximately 1,250 bison by 1999. The bison population would be expected to number approximately 2,000 by 2006, and approximately 2,900 bison by 2011, 10 years after capture, test, and slaughter operations have ceased. No bison would be expected in Reese Creek, Eagle Creek/Bear Creek, or West Yellowstone in this alternative. The bison population would experience a major decrease in this alternative, representing a nearly 47% reduction, compared to alternative 1, over a period of only three years.

No bison would be allowed anywhere outside Yellowstone National Park boundaries under alternative 5. Management actions in alternative 5 could affect the age/sex distribution or reproductive rate of the bison population. Bison distribution within the park would likely be affected, and several areas would likely have few or no bison for as long as 10 years.

In alternative 6, all bison would be vaccinated for approximately 10 years (beginning in the year 2000) to reduce seroprevalence in the population. After whole herd vaccination, bison would be captured, tested, and seropositives slaughtered, similar to alternative 5. Two different estimates of population size were calculated based on the effectiveness of the vaccine. Assuming a 70% effectiveness, the bison population would be expected to increase during the vaccination phase from 2,200 bison to approximately 3,500 bison in 2010, a negligible to minor increase compared to alternative 1. After 10 years of vaccination (2010), capture and slaughter would begin, and the population would drop from 3,500 to about 2,900 in a single year, a moderate (17%) decrease compared to alternative 1. If the vaccine was only 25% effective, the population would drop from 3,500 animals in 2010 to 2,500 the following year, when parkwide capture and slaughter began. This would represent a major short-term adverse impact (28% reduction) on the population. The herd would begin to increase following completion of the test and slaughter program; from 2,900 to 3,400 bison by 2014 (assuming 70% effectiveness), or from 2,500 to about 3,000 animals (assuming 25% effectiveness) by 2014.

Unlike other alternatives, in alternative 7 the agencies would attempt to manage the bison population within the more narrow range of 1,700 to 2,500 animals. Given the mix of management tools described above in “Alternatives,” the model predicts the bison population would be expected to increase from about 2,200 bison to 2,700 (average increase 2.6%/year) in 2004, and level off at or about 2,700 throughout the remainder of the 15-year plan. This alternative would result in a bison population 12% lower than alternative 1 in 2006 and 23% lower in 2011. However, because of limitations with the deterministic model, the differences between alternatives 1 and 7 might be less. Slaughter, quarantine, agency shooting, and hunting are predicted to remove an average of 132 to 137 bison per year. If bison exited the park in larger numbers during severe winters, more would be killed if the bison population was near or above 2,500 animals. During mild winters, fewer bison would exit the park and thus fewer bison would be killed.

The modified preferred alternative provides for an increasing bison population and would maintain a population of around 3,000. The use of management tools described in volume 1, “The Alternatives” would likely maintain the population near 3,000; modeling indicated the mean population would be similar to alternative 1 in the long term and was consistently about 20% higher than alternative 7 (identified as the preferred alternative in the Draft Environmental Impact Statement). This is considered a moderate to major benefit of the modified preferred alternative.

Stochastic Influence on Bison Population. In the period following the release of the Draft Environmental Impact Statement and the publication of the final environmental impact statement, the National Park Service funded development of a stochastic model to examine the influence of random events, such as severe winters, on bison management. Table 5 shows the model predictions of impacts on the bison population for all eight alternatives.

Seroprevalence Rate. Modeling efforts using the deterministic model to predict impacts of management scenarios on seroprevalence in the Draft Environmental Impact Statement assumed 50% seroprevalence in the bison population. The more refined stochastic model described above was also used to check predictions of impact on seroprevalence; however, research after the release of the Draft Environmental Impact Statement indicated seroprevalence in 246 bison tested in the winter of 1996–97 was 39% (NPS, unpubl. data). Both models assumed either a 70% rate of effectiveness of the bison vaccine (based on current success with cattle) or 25% rate of effectiveness (based on effectiveness in bison calves). Bison calves were assumed to be vaccinated with a safe and effective vaccine beginning in 2000 in the deterministic model; however, additional research has indicated a safe and effective vaccine for calves would probably not be available until later (2002/2003), so vaccination was assumed to begin in 2002 in the stochastic model.

Using the deterministic model, and assuming a vaccine that was 70% effective and calfhood vaccinations began in 2000, the population seroprevalence rate under alternative 1 would be expected to decline from a starting point of 50% seropositive in 1997 to at least 33% seropositive in 2006 (see table 6). If the vaccine was 25% effective, seroprevalence was predicted to drop from 50% to 40% by 2006. Continued management efforts and calfhood vaccination (assuming 70% efficacy) would reduce seroprevalence to 24% in 2011. The stochastic model predicted mean seroprevalence would fall to about 11% in 2013 (assuming 70% efficacy). This is a 69% reduction in the first 11 years of vaccination compared with a 49% reduction in 11 years of vaccination predicted by the deterministic model.

In alternative 2, the population seroprevalence rate would be expected to decline to at least 34% seropositive in 2006 (assuming 70% efficacy) or to 42% by 2006 (assuming 25% efficacy). Continued management efforts and calfhood vaccination (70% efficacy) would reduce seroprevalence to 26% in 2011. This would represent a minor adverse impact (3% to 8% less reduction) compared to alternative 1. The stochastic model predicted the seropositive rate would drop to about 13% by 2013, or a 62% reduction in 11 years, compared with a 42% reduction in 11 years of vaccination estimated by the deterministic model under this alternative.

In alternative 3, the population seroprevalence rate would be expected to decline to at least 36% seropositive in 2006, assuming a 70% vaccine efficacy. With calfhood vaccination and a vaccine efficacy of 25%, seroprevalence was predicted to drop to 45% by 2006. Continued management efforts and calfhood vaccination (70% efficacy) would reduce seroprevalence to 28% in 2011.

This would be a minor to moderately higher seroprevalence (9%–17% higher) than that predicted for alternative 1. The stochastic model predicted a 60% drop in seroprevalence from 11 years of vaccination to 15% seropositive, compared to a 40% reduction predicted by the deterministic model.

In alternative 4, capture and removal of seropositive bison, and calfhood vaccination (70% efficacy) was predicted to decrease seroprevalence to at least 34% in 2006 and 26% in 2011. Assuming a 25% vaccine efficacy, seroprevalence would drop to 42% by 2006. This would be a minor adverse impact (3%–5% higher seroprevalence) compared to alternative 1. The stochastic model predicted seroprevalence would fall to 13% in 11 years of vaccination. This is a 65% reduction compared to a 42% reduction predicted by the deterministic model.

In alternative 5, the seroprevalence rate in bison would be expected to drop from 50% in 1997 to 0% in 2001, assuming 70% vaccine efficacy, capture, test, slaughter operations, and whole-herd vaccination. In the 25% vaccine efficacy model the seroprevalence rate dropped to 0% by 2001. This would be a significant decrease in the seroprevalence rate and a major beneficial impact compared to alternative 1. Results using the stochastic model were comparable.

In alternative 6, the seroprevalence rate would remain similar to alternative 1 during the vaccination phase (2000–2010), and then drop to 0% by 2013. This would be a major reduction in seroprevalence compared to alternative 1. The stochastic model predicted that stabilization of seroprevalence (e.g., the end of phase 1) would take longer than the 15-year life of the plan. Phase 2 would drop seroprevalence to near zero by 2020.

In alternative 7, the population seroprevalence rate would be expected to decline from a starting point of 50% seropositive in 1997 to at least 32% seropositive in 2006 due to removal of seropositive bison leaving Yellowstone National Park in the West Yellowstone and Reese Creek area, and calfhood vaccination (70% efficacy) beginning in 2000. Continued management efforts and calfhood vaccination (70% efficacy) would reduce seroprevalence to 23% in 2011. With calfhood vaccination and a vaccine efficacy of 25%, seroprevalence was predicted to drop from 50% to 40% by 2006. This would be a negligible to minor beneficial impact (0–4% lower seroprevalence rate) compared to alternative 1. The stochastic model predicted a 61% decline to 14% in 2013 compared with a 49% decline in seroprevalence predicted in the same period of time by the deterministic model.

The deterministic model predicts that seroprevalence under the modified preferred alternative would decline to about 33% in 2006 due to removal of seropositive bison and remote calfhood vaccination. Continued management efforts and vaccination would reduce seroprevalence to 25% in 2011, similar to that predicted under alternatives 1 (24%) and 7 (23%). The stochastic model predicted a decline to about 15% in 2012 and 13% by 2013 after 11 years of vaccination. This is a reduction of 63% and is a greater reduction than the 46% drop predicted by the deterministic model in the same period.