Using an Automatic Trapping Grid to Reduce Rodents on Kaua’i: First Five Years
Uma Nagendra, Chiemi Nagle, Kawika Winter; Limahuli Garden and Preserve
Invasive rodents (Rattus sp.) remain a major threat to native plants and animals across the Pacific Islands. Although landscape-scale control of black rat populations is necessary to preserve endemic and critically endangered species, tools appropriate for remote areas are limited in Hawai’i. In addition, conservationists are challenged to find effective rodent control that is safe for endemic animals and has support of the local community. Automatic self-resetting kill traps for rodents (Goodnature A24) have been used to eradicate rodents from small islands and areas isolated by predator-proof fencing. Its effectiveness in unfenced systems on Kaua’i is still unknown.
In this rodent management trial, we assess changes to rodent activity, kills per trap, and trap success over the course of four years.
Trap success (measured by kills / activity) indicates the ongoing effectiveness of the rodent management activities.
Since our rodent management area is not bounded by a predator-proof fence, we expect "edge effects" to be visible. We assessed the magnitude of these edge effects by using trap density as a proxy.
In 2012, the National Tropical Botanical Garden implemented a landscape-scale rodent management trial using GoodNature A-24 kill traps in the Upper Limahuli Valley of Northwestern Kaua’i. A grid of 63 GoodNature A24 kill traps was installed along intersecting trails in a 30-acre drainage without a predator-proof fence. Number of kills per trap was documented with the help of automatic counters attached to the traps. Field staff verified kills and conducted trap maintenance each month.
|
|
Rodent presence/absence was monitored quarterly with peanut butter-baited ink cards placed in monitoring tunnels.
Yearly rodent activity was calculated for each location by the percent of times/year a monitoring tunnel indicated rodent presence. |
We then used ArcGIS to extrapolate rodent presence from tunnel locations to the entire rodent management area, including the trap locations. For each trap location, yearly trap success was then calculated by (# kills in a year / extrapolated rodent presence ). Edge effects were assessed using trap density as a proxy, also calculated in ArcGIS.
Change in rodent activity and trap success as a function of rodent kills and trap density over time were tested using repeated measures in R.
Methods
In 2012, the National Tropical Botanical Garden implemented a landscape-scale rodent management trial using GoodNature A-24 kill traps in the Upper Limahuli Valley of Northwestern Kaua’i. A grid of 63 GoodNature A24 kill traps was installed along intersecting trails in a 30-acre drainage without a predator-proof fence. Number of kills per trap was documented with the help of automatic counters attached to the traps. Field staff verified kills and conducted trap maintenance each month.
|
|
Rodent presence/absence was monitored quarterly with peanut butter-baited ink cards placed in monitoring tunnels.
Yearly rodent activity was calculated for each location by the percent of times/year a monitoring tunnel indicated rodent presence. |
We then used ArcGIS to extrapolate rodent presence from tunnel locations to the entire rodent management area, including the trap locations. For each trap location, yearly trap success was then calculated by (# kills in a year / extrapolated rodent presence ). Edge effects were assessed using trap density as a proxy, also calculated in ArcGIS.
Change in rodent activity and trap success as a function of rodent kills and trap density over time were tested using repeated measures in R.
Traps in interior, high density areas experienced a drastic reduction in rodent activity in 2016. Despite initially higher baseline activity levels in the interior zone, the whole rodent management area is now maintained at a low rodent population level. Starting in 2016, low-density exterior traps also have higher kills per trap, indicating a possible influx of rodents from the surrounding landscape.
Trap success fluctuated slightly from year to year. In each year, however, traps in exterior, low-density areas had greater trap success than those in interior, high-density areas.
Traps in interior, high density areas experienced a drastic reduction in rodent activity in 2016. Despite initially higher baseline activity levels in the interior zone, the whole rodent management area is now maintained at a low rodent population level. Starting in 2016, low-density exterior traps also have higher kills per trap, indicating a possible influx of rodents from the surrounding landscape.
Trap success fluctuated slightly from year to year. In each year, however, traps in exterior, low-density areas had greater trap success than those in interior, high-density areas.
Overall rodent presence fell from 48% in 2014 to 19% in 2017 year. As expected, numbers of rodents killed declined along with rodent activity levels. Most strikingly, trap success remained roughly the same each year. The degree of reduction in rodent presence and kills depended on trap density and edge effects. Traps in interior, high density areas experienced a greater reduction in rodent activity over trial period.
Results
Overall rodent presence fell from 48% in 2014 to 19% in 2017 year. As expected, numbers of rodents killed declined along with rodent activity levels. Most strikingly, trap success remained roughly the same each year. The degree of reduction in rodent presence and kills depended on trap density and edge effects. Traps in interior, high density areas experienced a greater reduction in rodent activity over trial period.
Contrary to expectations, trap success did not decline rapidly within five years of implementing the rodent management plan. Even as the rodent population dwindled, traps overall had roughly the same success rate of catching remaining rodents. Edge effects were important: while interior trap zones exhibited the greatest drop in rat activity, exterior traps consistently had the greatest trapping success.
Both of these points indicate that there is a continued influx of rodents from elsewhere in the landscape. While the current rodent management plan is effectively reducing the rodent population in this localized area, connection to the wider landscape prevents outright elimination.
Even if Goodnature traps alone cannot completely eliminate rodent populations, this level of population control still provides substantial benefit to endangered seabird colonies by reducing rates of rodent predation on seabird eggs.
