575 items found

CPW website species profile: LynxLynx populations in Colorado plummeted in the late 1800s and early 1900s for various reasons, including general predator poisoning and unregulated trapping. The last known lynx was illegally trapped near Vail in 1974, a year after the state listed the lynx as endangered. In 1997, Colorado Parks and Wildlife undertook what was to become one of North America’s most high-profile carnivore reintroductions to date. Four years after the last lynx was released into the state in 2006, CPW deemed the initial lynx introduction effort a success. Research has now focused towards determining and maintaining the long-term success of the reintroduction. Learn more about lynx and Colorado’s successful lynx reintroduction in our Lynx Fact Sheet. Or, learn how to identify a lynx and report a lynx sighting.

Lynx Research Projects:

• Evaluating Options to Monitor Canada Lynx Reintroduction in Colorado
• Mammal and Breeding Bird Response to Bark Beetle Outbreaks

1. Motion-activated cameras (“camera traps”) are increasingly used in ecological and management studies for remotely observing wildlife and are amongst the most powerful tools for wildlife research. However, studies involving camera traps result in millions of images that need to be analysed, typically by visually observing each image, in order to extract data that can be used in ecological analyses.
2. We trained machine learning models using convolutional neural networks with the ResNet-18 architecture and 3,367,383 images to automatically classify wildlife species from camera trap images obtained from five states across the United States. We tested our model on an independent subset of images not seen during training from the United States and on an out-of-sample (or “out-of-distribution” in the machine learning literature) dataset of ungulate images from Canada. We also tested the ability of our model to distinguish empty images from those with animals in another out-of-sample dataset from Tanzania, containing a faunal community that was novel to the model.
3. The trained model classified approximately 2,000 images per minute on a laptop computer with 16 gigabytes of RAM. The trained model achieved 98% accuracy at identifying species in the United States, the highest accuracy of such a model to date. Out-of-sample validation from Canada achieved 82% accuracy and correctly identified 94% of images containing an animal in the dataset from Tanzania. We provide an r package (Machine Learning for Wildlife Image Classification) that allows the users to (a) use the trained model presented here and (b) train their own model using classified images of wildlife from their studies.
4. The use of machine learning to rapidly and accurately classify wildlife in camera trap images can facilitate non-invasive sampling designs in ecological studies by reducing the burden of manually analysing images. Our r package makes these methods accessible to ecologists.

Led By

Dr. Eric R. Fetherman​

Study Area

Gunnison River and Harrison Creek

Project Status

Ongoing​

Research Objectives

• To maintain wild brood stocks of whirling disease resistant rainbow trout to supplement hatchery stocks, as necessary.
• To evaluate wild stocks for continued disease resistance.

Project Description

Fishery managers stock whirling disease resistant rainbow trout (known as the Hofer strain) in waters across the state to supplement and recover populations previously lost to whirling disease.

In addition to maintaining hatchery brood stocks (fish used for spawning) of whirling disease resistant rainbow trout, two wild brood stocks have been established to supplement and replace hatchery brood stocks, as necessary. Researchers periodically evaluate that these brood stocks retain resistance to whirling disease. These evaluations allow Colorado Parks and Wildlife biologists and researchers to determine if the resistance characteristics of these populations are changing or remaining static, and ensure that eggs collected from these populations and used to supplement hatchery brood stocks will continue to produce rainbow trout that are resistant to whirling disease.

One of these wild brood stocks is located in Harrison Creek, a tributary of Lake Catamount in Steamboat Springs, Colorado. This wild brood stock is being used to rear crosses of the Hofer and Harrison Lake rainbow trout strains (known as the HxH). Known to be partially resistant to whirling disease, the Harrison Lake strain of rainbow trout originates from Harrison Lake, Montana. Recent research has shown that fish stocked in Harrison Creek return to the creek to spawn, facilitating future wild egg collections. Additionally, resistance to whirling disease is increasing in this population as more HxHs become established.

The other wild brood stock is located in the East Portal of the Gunnison River in the Black Canyon of the Gunnison National Park. Once managed for crosses of the Hofer and Colorado River Rainbow trout strains (known as the HxC), recent research shows that the HxC constituted only a small proportion of the total adult spawning rainbow trout population. Despite this, exposure experiments conducted using eggs from the East Portal showed that these fish had started to develop a resistance to whirling disease, likely a result of low infection levels and continued natural reproduction. Eggs are taken from this brood stock on an annual basis to stock other locations within the Gunnis​​on River, and to maintain the Gunnison River Rainbow trout brood stock in Colorado hatcheries. ​

Associated Publications

• Fetherman, E. R., and G. J. Schisler. 2016. Sport Fish Research Studies. Federal Aid Project F-394-R15. Federal Aid in Fish and Wildlife Restoration, Job Progress Report. Colorado Parks and Wildlife, Aquatic Wildlife Research Section. Fort Collins, CO. 
• Fetherman, E. R., and G. J. Schisler. 2015. Sport Fish Research Studies. Federal Aid Project F-394-R14. Federal Aid in Fish and Wildlife Restoration, Job Progress Report. Colorado Parks and Wildlife, Aquatic Wildlife Research Section. Fort Collins, CO. 
• Fetherman, E. R., and G. J. Schisler. 2014. Sport Fish Research Studies. Federal Aid Project F-394-R13. Federal Aid in Fish and Wildlife Restoration, Job Progress Report. Colorado Parks and Wildlife, Aquatic Wildlife Research Section. Fort Collins, CO.
• Fetherman, E. R., and G. J. Schisler. 2013. Sport Fish Research Studies. Federal Aid Project F-394-R12. Federal Aid in Fish Restoration Job Progress Report​. Colorado Parks and Wildlife. Aquatic Research Section. Fort Collins, CO.

Led ByJake Ivan

Study AreaStatewide

Project StatusCompleted

Research Objective: To determine how bark beetle outbreaks have impacted breeding birds and mammals in Colorado.

Wildlife Cameras

As part of a project to learn about the impacts of bark beetle outbreaks on wildlife, Colorado Parks and Wildlife deployed 300 remote wildlife cameras at randomly selected locations throughout spruce/fir and lodgepole pine forests in Colorado.  To the right are samples of the almost 400,000 photos collected from the project.

Project Description

Since 1996, millions of acres of Colorado forests have fallen victim to bark beetle infestations. This unusual epidemic has left researchers wondering how local wildlife living in these forests will respond.

This study will focus on tracking how mammals and breeding birds alter their use of subalpine forests during the course of an outbreak. 

Researchers expect wildlife response to the beetle epidemic to vary by species and forest system. Some species may stay in the area during the full course of the outbreak, while others may disappear once the area is infected. For example, researchers expect red squirrel use of lodgepole pine forests to decrease as the outbreak affects cone-producing trees, which serve as the squirrel's primary food source. 

The focal species for this study are those listed as a priority species under the Colorado Wildlife Action Plan and/or listed as sensitive species by the United State Forest Service Region 2 and/or listed as vulnerable at the sub-national scale by NatureServ. 

These species include:

• American marten
• American Three-toed Woodpecker
• Broad-tailed Hummingbird
• Cassin's Finch
• Cordilleran Flycatcher
• Dusky Flycatcher
• Dusky Grouse
• Evening Grosbeak
• Northern Goshawk
• Olive-sided Flycatcher
• Red Crossbill
• Red-naped Sapsucker
• Williamson's Sapsucker

Two additional species, the snowshoe hare and red squirrel, are focal species as well because together they compromise nearly 100 percent of the diet of the Canada lynx, which is a state and federally listed species.  

Researchers plan to survey the selected mammal species using motion-sensor cameras. Bird species will be surveyed during early morning counts throughout the breeding season (late May to early July). Because the survey methods are general, researchers should be able to gather information about variety of additional game and non-game species beyond those species of concern listed above. Researchers will use this data to determine which species decrease their use of these forests as bark beetle infestations run their course, which species increase their use, and which species are unaffected.

Associated Publications

Ivan, J. S. 2013. Small mammal and breeding bird response to bark beetle outbreaks in Colorado. Pages 1-13 in Wildlife Research Report, Mammals Research Program, Colorado Division of Parks and Wildlife, Fort Collins, CO. 310 pp.