USGS ScienceSearch

USGS · 70214031

Habitat suitability and ecological associations of two non-native ungulate species on the Hawaiian island of Lanai

Abstract

The ability to effectively manage game species for specific conservation objectives is often limited by the scientific understanding of their distribution and abundance. This is especially true in Hawai‘i where introduced game mammals are poorly studied and have low value relative to native species in other states. We modeled the habitat suitability and ecological associations of European mouflon sheep (“mouflon”; Ovis musimon ) and axis deer ( Axis axis ) on the island of Lāna‘i using intensive aerial survey and environmental data that included climate, vegetation, and topographic variables. We conducted diagnostic tests on a suite of primarily categorical predictors and determined most were highly correlated. We therefore developed a suite of other spatial predictor layers with continuous variables. We tested several modeling approaches but settled on generalized linear models (GLM) and random GLMs because they could account for group size of animals and were based on curvilinear responses of each species to environmental variability. Both mammal species were habitat generalists showing little affinity to particular plant species or communities. Continuous predictors associated with plant productivity such as mean annual precipitation, normalized difference vegetation index (NDVI), and cloud cover were important explanatory factors in a GLM of axis deer and a random GLM of mouflon habitat suitability. The presence of axis deer was also an important explanatory predictor for mouflon distribution, but deer were not influenced by mouflon distribution, indicating asymmetrical competition. Consequently, mouflon were restricted to lower elevation arid and very dry slopes, whereas axis deer were more broadly distributed throughout other upland environments of the island, but avoided steep terrain. Findings indicate that removal of a substantial portion of the more abundant axis deer population may lead to an increase in abundance and distribution of mouflon without containment. Resulting spatial models of game mammal habitat suitability will be employed to inform land use prioritization analyses and to help resolve long-standing conflicts between native species conservation and sustained-yield hunting.

Explore related subjects

90° N90° S · 180° W ← longitude → 180° E
Source-reported bounding extent: 20.72400644605942° to 20.931941310423674° latitude; -157.06329345703125° to -156.79962158203125° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Steve C. Hess, Lucas Berio Fortini, Christina Leopold, Jacob Muise, Jonathan Sprague. 2020. Habitat suitability and ecological associations of two non-native ungulate species on the Hawaiian island of Lanai. https://pubs.usgs.gov/publication/70214031

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

Effectiveness of rapid 'ōhi'a death management strategies at a focal disease outbreak on Hawai'i Island

The ongoing spread of rapid ‘ōhi‘a death (ROD) in the Hawaiian Islands threatens the long-term sustainability of ‘ōhi‘a lehua (Metrosideros polymorpha) forests throughout the state. First identified in the Puna district of Hawai‘i Island in 2014, the disease caused by the novel fungi Ceratocystis lukuohia and Ceratocystis huliohia has now spread island-wide and was recently detected on Kaua‘i, O‘ahu, and Maui. The leading hypothesis for the spread of ROD is through airborne ambrosia beetle frass particles that contain viable Ceratocystis propagules, thus management efforts focus on containing this frass. At the time of this study (2017–2018), the Waipunalei site was the northernmost outbreak of ROD on Hawai‘i Island. The focal nature of the outbreak and accessibility of the location provided the opportunity to monitor the effectiveness of two types of proposed management methods to reduce the airborne spread of potentially infective ambrosia beetle frass: tree felling and insecticide treatments. We placed 23 passive environmental samplers (PES), which monitored for airborne frass and wood particles containing C. lukuohia and C. huliohia in a grid that spanned the outbreak area over 22 weeks. Cross-vane panel traps with 1:1 methanol:ethanol lures were attached to nine of the PES to document wood-boring ambrosia and cerambycid beetle populations during the latter three months of the study. Monitoring with PES began three weeks before management and continued for one month after the last infected trees were felled. Glass microscope slides from the 23 PES were examined for airborne ambrosia beetle frass and wood particles by microscopy. DNA was extracted from the slides and tested by qPCR (quantitative polymerase chain reaction) for C. lukuohia and C. huliohia. We also investigated the correlation of beetle gallery counts with tree height and tested the efficacy of Bifen I/T insecticide (active ingredient: bifenthrin 7.9%) for preventing beetle attacks on the cut surface of ‘ōhi‘a bolts (tree stem sections). Beetle trapping data revealed that the area supports a diverse community of wood-boring beetles, some of which likely attack ‘ōhi‘a and may facilitate the spread of ROD. The number of beetle galleries on felled ‘ōhi‘a trees decreased linearly as tree height increased. We also observed significantly fewer beetle attacks on Bifen I/T treated ‘ōhi‘a bolts than non-treated bolts, but gallery formation nearly ceased in both treated and control bolts by week three. Ceratocystis lukuohia DNA was detected twenty-six times and C. huliohia was detected five times in the PES throughout this study. DNA detections were correlated to frass and wood counts, and the number of felled trees were correlated to wood particle counts but not frass counts. Both the timing and distribution of detections across the sampling grid indicate that tree felling may have reduced airborne detections of Ceratocystis DNA soon after tree felling was completed. A subsequent increase in detections after tree felling ceased may indicate that incomplete removal of infected trees and the appearance of new infections in previously asymptomatic trees could have allowed airborne detections of potentially infectious fungal propagules to once again increase.

Hawaii

Field trials to test new trap technologies for monitoring Culex populations and the efficacy of the biopesticide formulation VectoMax® FG for control of larval Culex quinquefasciatus in the Alaka'i Plateau, Kaua'i, Hawaii

Mosquito-borne avian malaria Plasmodium relictum is a key limiting factor for endemic Hawaiian forest birds. In the past decade, populations of Kaua‘i’s endemic forest birds have been in a steep decline due to an increase in malaria transmission. To evaluate the use of available biopesticides for short-term mosquito control we tested the efficacy of the biopesticide VectoMax® FG against Culex quinquefasciatus larvae in naturally occurring perched stream pools, seeps, and ground pools in forest bird habitat in Kaua‘i’s remote Alaka‘i Plateau. We also tested the efficacy of conventional and newer traps and attractants for the capture of adult Culex quinquefasciatus in Hawaiian rain forests and monitored adult mosquito populations at the Kaua‘i field site. During field trials conducted on Hawai‘i Island we captured more Culex quinquefasciatus in gravid traps than in host-seeking traps. Among the host-seeking traps, Biogents BG-Sentinel 2 traps baited with CO2 and BG-Lure caught more Culex quinquefasciatus and Aedes japonicus japonicus than CDC (Centers for Disease Control and Prevention) traps baited with compressed CO2, CDC traps baited with dry ice, or Biogents BG-Sentinel 2 traps baited with BG-Lure and octenol but not CO2. Both Biogents BG-Sentinel 2 and CDC miniature traps baited with compressed CO2 or dry ice captured significantly more Culex quinquefasciatus than Biogents BG-Sentinel 2 traps baited with octenol and BG-Lure but without CO2. We also found that gravid traps baited with timothy hay infusions caught significantly more Culex quinquefasciatus than traps baited with either a commercial gravid mosquito attractant or an infusion made with pelleted rabbit feed. Traps baited with an infusion of timothy hay and donkey dung were the most effective for Culex quinquefasciatus. On Kaua‘i, we operated Biogents BG-Sentinel 2 traps baited with CO2 and gravid traps and captured 29 mosquitoes in 182 trap-nights from October–November 2016 and 126 mosquitoes in 254 trap-nights from September–October 2017. Contrary to our findings on Hawai‘i Island, most mosquitoes (96%) were captured in Biogents BG-Sentinel 2 traps indicating considerable site-to-site variability in trap efficacy. Weekly adult trapping on Kaua‘i indicates Culex quinquefasciatus populations peaked in October but provided no reliable evidence that larval control had any significant effect on adult populations. Overall, VectoMax® FG was very effective at larval control reducing larval abundance by 95% at 48 hours and out to 1-week post-treatment. Treatment was most effective (100% at 1-week post-treatment) in perched pools when early instar larvae were present and least effective in seeps when pupae and fourth instar larvae were most common. Although post-treatment counts fluctuated dramatically, we observed no evidence of population level impacts to the two most common non-target invertebrates: the water strider Microvelia vagans and endemic damselfly naiads (Megalagrion sp.). VectoMax® FG appears to be an effective and safe biopesticide for the local control of Culex quinquefasciatus larvae in forest bird habitat in the Alaka‘i Plateau. Further studies will be necessary to determine if local larval control significantly reduces adult mosquito abundance and, ultimately, avian malaria transmission, and if there are long term, non-target effects associated with repeated use of VectoMax® FG in natural Hawaiian waterways.

Hawaii

Hawaiian hoary bat (Lasiurus cinereus semotus) acoustic monitoring at Hawaii Army National Guard (HIARNG) installations statewide

Acoustic sampling for occurrence of the endangered Hawaiian hoary bat (Lasiurus cinereus semotus) was conducted at 18 “long-term” acoustic monitoring stations on 12 Hawai‘i Army National Guard (HIARNG) installations across the islands of Hawai‘i, Maui, Moloka‘i, O‘ahu, and Kaua‘i between 2012 and 2018. Bats were confirmed as present at 10 of these installations: Kealakekua Armory, Keaukaha Military Reservation (KMR), Hanapēpe Armory, Kekaha Firing Range (KFR), Pu‘unēnē Training Facility, Ukumehame Firing Range (UFR), Kaunakakai Armory, Bellows Regional Training Institute, Kalaeloa, Barber’s Point, and the 487th Military Parking Facility, Wahiawā. Seasonal frequency of bat detection was similar to previous acoustic studies for the islands of Hawai‘i, Maui, and O‘ahu. Hawaiian hoary bats were recorded at HIARNG installations during periods of pregnancy, lactation, and pup fledging. Our acoustic sampling did not record bat vocalizations at Fort Ruger and Waiawa Armory. Foraging activity was observed at nine acoustic monitoring stations on the islands of Hawai‘i, Kaua‘i, O‘ahu, and Maui. No foraging activity was observed on Moloka‘i and a single station on O‘ahu recorded one feeding buzz in September 2017 at Kalaeloa, Barber’s Point. Within-night detections showed bat activity was mostly confined to the first six hours of the night but was also variable among stations. In addition to long-term bat acoustic monitoring at HIARNG installations, Hawaiian hoary bat insect prey sampling with paired acoustic monitoring was conducted at KMR on Hawai‘i Island from May through August 2018. Insect abundance and bat activity were sampled within areas where goats and sheep were used to control weeds to determine if grazing by these ungulates attract and support potential prey for the insectivorous Hawaiian hoary bat. The assessment focused on types of flies that are often associated with livestock (muscoid flies, including house flies [Muscidae], blow flies [Calliphoridae], flesh flies [Sarcophagidae], and biting midges [Ceratopogonidae]), and moths that may be impacted by changes in the availability of grass. Insect abundance was found to vary in both space and time across the study area, with numbers of muscoid flies and biting midges increasing in the presence of livestock at some stations. Although these insects appeared to respond to livestock grazing in some instances, we did not find statistically significant responses in bat foraging as measured by echolocation activity between grazed stations and the ungrazed reference station. Thus, we found no evidence that suggested bats are drawn to foraging resources in grazed areas. This result may be influenced by several factors, including the size of Hawaiian hoary bat foraging ranges compared to the scale of study area, the type of ungulate and their dung, and the timing of insect activity. Hawaiian hoary bats use KMR and forage seasonally as evidenced by long-term acoustic studies and their presence in the ungulate grazing areas; however, the relatively small size of the ungulate herd and the area that they graze may not be able to support enough prey to have a significant influence on bat foraging rates.

Hawaii