Statistics
|
Importance is important: Using more reliable but less common methods for determining the importance of ecological variables “The results of a regression are like a messy storehouse: your task is to decide what you pull out and use!” Nick, our statistical specialist, suggests some tools for evaluating variable importance. |
||
|
Oases along the flyway: Identifying stopover sites for migratory birds in the southeastern U.S. How do you find stopover habitat of migrants moving under the cover of night? David La Puma uses weather radar to see in the dark and identify sites across the southeastern U.S. |
||
|
It’s all a game – Land Use and Conservation in everyone’s hands What if someone told you that you could be of great help to science in general and conservation in particular and have fun at the same time? Forward Trails is a Massive Multiplayer Online Game that allows you to do just that! |
||
|
Why and where did farmers abandon their fields in the Ukraine? The breakdown of the Soviet Union in 1991 triggered rapid and widespread farmland abandonment across Eastern Europe. Using remote sensing and statistical models, Matthias Baumann mapped where abandonment occurred in Ukraine and explained why it happened where it did. |
||
|
Predicting Broad-scale Patterns of Avian Biodiversity with Landsat Image Texture There is not a map that predicts bird species biodiversity for the whole United States at scales that are relevant for a forester or a county planner. However, such a map is utterly needed to make realistic conservation plans. |
||
|
New multi-scale landscape indices for spatial pattern analysis Understanding how spatial patterns of vegetation explain the distribution of organisms is a central theme within Landscape Ecology. Avi Bar Massada developed a novel method to quantify these patterns, which may be more effective than existing methods. He illustrates its effectiveness with bird data from Wisconsin’s Fort McCoy. |
||
|
Highway to the danger zone? Effects of sample size, number of parameters, and collinearity on error estimates from cross-validation. Ecologists often approach their research as a “measure everything, predict everything” endeavor. While this is advantageous when resources are unlimited, this is rarely the case, and often there are only a few observations. Nick Keuler, resident statistician in the SILVIS lab, offers some guidelines to developing good predictive models. |
||
|
Things aren’t as they used to be: forest disturbance, birds, and the shifting baseline syndrome Can a bumper sticker inspire innovative research? In the case of Chad Rittenhouse PhD, a chance sighting motivated an innovative line of research that questions how changes occur in the natural world and how we perceive and measure these changes. |
