email: kmckelvey[at]fs.fed.us Wildlife Genetics Laboratory Website Google Scholar profile In recent years I have concentrated on developing methods to evaluate status and trends of organisms across broad spatial and temporal domains. For broad-scale evaluation of population description, status, and trends, genetic monitoring techniques hold great promise. Spatially referenced occurrence or abundance can be related to the landscape features; new statistical methods allow variance to be partitioned across multiple scales greatly improving our ability to evaluate habitat quality. If these occurrence data were derived from genetic samples, many additional analyses are possible. By relating pair-wise genetic differences across space we have developed methods to measure population connectivity across complex landscapes. The combination of occurrence data and genetic patterns together provide a rich and cost effective approach to determine population status. We are currently prototyping broad-scale genetic monitoring using native fish samples from across Montana and Idaho. Additionally, we are testing environmental DNA based detection of brook, bull, lake, and westslope cutthroat trout.
My research interests are focused on providing better methods to assess resources across large spatial and temporal domains. For example, the Forest Service manages over 190 million acres. Even spending a single dollar per acre to assess all of the resources --wood, plant and animal species, soild conditions, recreation use--would be too expensive. So, we need to develop methods that produce reliable information for many resources for only pennies per acre. As current methods will not deliver this, we need to use new technologies and statistical methods to develop methods that do. This is my primary interest
importance of research
My research interests have been varied, but have in common a focus on large scale issues. I began work for the Forest Service building a spatially explicit simulator for the northern spotted owl, a model that allowed detailed modeling of various reserve options and which was eventually used by the state of California, the BLM, and in the President's Forest Plan. I then was leader of the fire modeling sub-group in the Sierra Nevada Ecosystem Project, and developed the first comprehensive fire history for the Sierra Nevada. This history showed that fire had been stable in its patterns throughout the 20th Century, and was therefore predictable. I then became part of the Lynx Science Team and was the Science lead for the National Lynx Survey, the most extensive DNA-based carnivore survey ever attempted. This survey provided reliable presence absence data Canada lynx for over 50 national forests. 5 national parks, and numerous other areas managed by the BLM and several Tribal Nations. I then was a member of the wolverine science team, and worked with Jeffrey Copeland to define the bioclimatic factors that controlled wolverine range and habitat use across it's northern hemispheric range.
Simply put, you cannot run a store if you have no idea what is on the shelves. And the store, its shelves, and the customers are relatively stable and don't change much from day to day. For example, if you don't sell a jar of peanut butter, it will be there to be sold tomorrow. Natural resources are, however, in a continual state of change. They grow, die, burn, fly or run away, interact, and adapt. It is therefore much more difficult and if we are to manage our resources intelligently, it is much more important to be able to assess current resource status. Without this, our management will consist of guesswork and our guesses will too frequently wrong. As professional management is fundamentally dependent on resource assessment and monitoring, there are few tasks more important to do well.