Our goal is to improve conservation by providing powerful wildlife monitoring tools to identify status, distribution, and population trends

We provide rigorous data for impact assessments, cost-effective tools for measuring ecological changes after restoration and management actions or development projects, and statistically sound approaches for estimating population trends over time.

We look forward to hearing from you about your monitoring challenges.

Data analysis

Acoustic sensors, camera traps, and aerial imagery are powerful monitoring tools; capable of generating rich datasets for a variety of applications. These technologies enable users to collect data over long periods of time, often in remote and hard-to-access locations. CMI specializes in analyzing these large monitoring datasets. We help our clients design and implement surveys that are tailored to the specific goals of their project. We’ve collaborated on more than 200 biological monitoring and resource management projects around the globe, ranging from 20 gigabytes to >10 terabytes of data per project.

Example projects include:

data analysis for wildlife monitoring and conservation
  • Detecting the presence of rare/elusive species across the landscape (California Red-legged Frog, Critically Endangered bat species in Africa (IUCN Red List), Night Parrot, Northern Spotted Owl, Ashy Storm-petrel, feral cats, etc.)
  • Before/After Control/Impact studies of conservation outcomes (Rat Island, Alaska; Isla Choros, Chile).
  • Long term monitoring of trends – Quantify breeding population trends (Wedge-tailed Shearwaters in Queensland), breeding phenology, and local attendance patterns (Sandhill Cranes, Newell’s Shearwater, Ashy Storm-petrel, Brown Booby, etc.).
  • Wildlife population surveys using triggered or time-lapse images from game cameras (kittiwakes, feral goats, etc.), and aerial imagery (Olive ridley sea turtle, Adelie penguin, etc.)
  • Soundscape Ecology -calculating indices correlated with species richness and as covariates to improve estimation of detection probabilities.
  • Impact assessment and prevention – monitoring for infrequent acoustic events related to wildlife impacts (wildlife collisions with powerlines, gunshots, engines, chainsaws, etc.)

Machine learning

Our team develops machine learning algorithms (Deep Neural Networks, Convolutional Neural Networks, etc.) and can deploy these in efficient production workflows to detect events of interest in acoustic and visual data. We have extensive experience developing robust training datasets, quantifying model performance, and tuning model parameters to meet specific project goals. Additionally, we have also developed a number of methods to accelerate the initial labeling process and quickly create models for new projects without any prior labeled data.  We can deploy early models in production workflows to rapidly iterate and improve performance.

applying machine learning for wildlife monitoring

Scalable workflows

We create scalable workflows to transfer, manipulate, and analyze datasets ranging from a few gigabytes to hundreds of terabytes. Our team employs custom software solutions, automated data checks, programmatic file manipulation, cloud computing, and other tools to build repeatable workflows that can grow with our clients’ needs.

applying machine learning for wildlife monitoring

Survey design

We collaborate with clients to design statistically robust surveys. This includes combining acoustic or image data collection with more traditional survey methods such as point counts, burrow searches, gps tracking data, etc. Well-designed surveys can enable statistically rigorous measures of abundance, change detection before-after management actions, and long-term activity trends.

extensive experience with survey design in challenging habitats

Statistical expertise

We apply cutting edge statistical techniques to our analysis of data streams. These include survey power analyses, Bayesian modeling, inter-annual trend analyses, and others. Our analysts are skilled at distilling complex datasets and creating intuitive figures and maps. These techniques strengthen our clients’ understanding of animal activity patterns and relative abundance through space and time.

sophisticated data analysis and statistics

Technical advice

We have a working knowledge of many types of acoustic and visual sensor hardware and software packages currently available – and have developed some of our own proprietary tools as well.

Examples of the tools we work with include:

  • Song Meters (Wildlife Acoustics, Inc.)
  • Bioacoustic Recorders (Frontier Labs)
  • AudioMoths
  • Swifts and Automated Recording Units (ARUs) (Cornell Lab of Ornithology)
  • Game cameras from Reconyx, Bushnell, and others
  • XBAT, Raven, seewave, WildSpectra, SoX, Ishmael, Kaleidoscope, SonoBat, and other bioacoustic analysis packages.
expertise with wildlife monitoring hardware

Pro-bono projects

Each year we collaborate on 3-5 pro-bono conservation projects; helping projects test novel approaches to wildlife monitoring for species of conservation concern.