When I worked at the Biological Diversity Lab at Utah State University my boss was advising a PhD student named Josh Lawler, who was studying the habitat relationships of cavity nesting birds in the Uintah mountains of northern Utah (I actually got to assist Josh with his field work for a couple of weeks, but that's another story).
Josh interviewed last week for the Ecologist position on the faculty of the College of Forest Resources at the University of Washington, and I attended his seminar. The talk was absolutely mind-blowing. He is doing cutting-edge research and really breaking new ground methodologically. He talked about two of my all-time favorite topics, global climate change and land use change, all within the context of how these forces of ecological change are impacting biological diversity.
The first half of his talk was devoted to the work heÂ’s done modeling changes to species ranges over the next ~100 years as a result of climate change. He looked at nearly 3,000 species ranges from three taxonomic groups (amphibians, birds, mammals) in the western hemisphere and used the model outputs from 30 General Circulation Models (GCMs) to predict how climate change will alter species ranges. The resolution is 50 km. He ran each of the 3,000 species through each of the 30 GCMs, for a combination of 90,000 *final* model runs. In order to get to those final runs, he reserved 20% of the range of each species, then took the remaining 80% and used random "“forest"” classifiers (nothing to do with real forests) that use model averaging, classification tree models, and machine learning (basically neural networks) to find which models most effectively predicted the reserved 20% based on current climate conditions. The upshot of this work is BIG changes over the next 100 years for all of the species, but particularly for the amphibians. He mapped hotspots of species loss (where species will leave cell to move elsewhere or simply die) and conversely areas of climate "refugia"” where changes are predicted to be less severe. Very interesting work, and I'm looking forward to learning more about it--he only had 20 minutes to present it.
The second half of his talk was on modeling future land use change in the Umpqua River watershed, in the southern Coast Range of Oregon. He's doing this work as part of CLAMS with Norm Johnson and Tom Spees, and they are simulating the impact of different management schemes on forests. Josh used the PATCH model and presented the effects of landscape modification on the Olive-side flycatcher and the Willow flycatcher. This model was also pretty amazing, it uses a hexagonal grid, where the size of the hex is determined by the home range size of the animal. Each species can move, reproduce, or die in each cell. The model incorporates population dynamics, individual animal behavior (i.e., nest site fidelity), sociodemographic changes, vegetation structure, and pattern.
He is now working on linking the CLAMS model with his climate change work in the Northwest, and will soon be producing new species range maps for the entire NW at 1 km resolution.
Pretty exciting and inspiring stuff, and hopefully weÂ’ll have him over here again to talk in much more detail about his work.
No comments:
Post a Comment