example, many different government agencies and others collect data that is
relevant to understanding the health of forests: The Forest Service does an
inventory of plots around the country every five to ten years that assesses the
trees, the ground vegetation, the soils, and the air quality; the U.S. Environmental
Protection Agency assesses water quality around the country; the Census
assesses population levels and housing densities; private groups monitor at-risk
species; the list goes on and on. Understanding the true state of our forests and
the threats to them requires integrating this data coherently.
“Forests on the Edge” is a project that is doing just that, combining allthose data sources into a single map and analyzing the results. But the data
doesn’t line up neatly. The plots the Forest Service analyzes, for example, are
different from the plots the Geological Survey analyzes. The data are of varying
quality and are gathered in different ways. Scales vary. The project has
developed techniques to use the combined data to produce the clearest picture
of the state of our forests and the threat to it, but new techniques are needed to
quantify the uncertainty of the combined data they produce.
In other situations, the hard data scientists need simply don’t exist. It’sdifficult and expensive too, for example, count all the caribou in a ten-thousand-
square-mile area. In some such situations, however, knowledgeable,
experienced folks have some good ideas about what is going on – they just can’t
back their opinions up scientifically. Inuit in northern Canada, for example, may
have a strong sense of whether the caribou population is rising or falling, based
on their long experience traveling across the land and sharing information with
one another. Mathematical scientists are working on developing unbiased ways
based on mathematical and biological principles to combine this “soft” data with
the limited available hard data. For caribou, as an example, scientists could
survey a few, limited areas carefully and then test out how accurate the expert
knowledge is on those particular areas in order to determine how much weight to
give it in an overall assessment. This approach has only begun to be explored.
Even when the funding exists to gather the data needed, mathematicalquestions arise about how to do so most efficiently. For example, the National
Ecological Observatory Network (NEON) is collecting data at twenty sites across
the U.S. to get a continent-wide picture of the impacts of climate change, land
use change and invasive species on natural resources, and biodiversity. Those