Measuring the "naturalness" of landscapes: a look at the U.S.

St. Mary Lake in Glacier National Park. Credit, Ken Thomas.A new study in the journal Landscape Ecology presents an approach for measuring the level of naturalness across a region.

Colorado State University researcher David Theobald tests this approach on the lower 48 states of the U.S. to determine the naturalness of the nation and how it is projected to change over the next 20 years.

The method develops a composite score of naturalness based on the proportion of area that can be considered "natural" as opposed to "human dominated."

Specifically, Theobald used existing land used data to apply scores at a scale of 30 meters. Urban/built-up areas, roads and cropland were assigned a score of "0". Natural areas (i.e. forests, grassland, wetlands, etc.) were assigned a score of "1". Roads and rural development negatively impacted the scores of adjacent areas.

Based on the data, the study calculated a natural landscape score of .6621 for the conterminous United States in 2001.

This means that roughly 1/3 of the nation or (2.6 million out of 8 million km2) had been converted to human dominated uses and was no longer natural.

Using projections for future development, Theobald calculated that by 2030, an additional 92,200 km2 of natural lands would be converted to human dominated uses.

One advantage of the approach is that it is scalable to regional or sub-regional levels. For example, the study found that central Idaho, Yellowstone National Park (WY), southeastern Utah, northwestern Arizona, southwestern New Mexico, southeastern Oregon, central Nevada, and Owens Valley (CA) had the highest naturalness scores.

The metric could be particularly important for understanding landscape changes around protected areas and within ecologically important regions. It could also be useful for doing regional-scale prioritization of conservation efforts.

However, the metric could probably benefit from some refinement. For example, the approach scores both cropland and urban areas as "0" even though farmland may provide much better habitat and ecosystem services.

Similarly, the metric gives a score of "1" to all natural areas even though some forests and grasslands may be subject to high levels of disturbance from logging, grazing, and other human uses.

Refinement of the scoring to better reflect the variability in the conservation values of land cover types will do better justice to the concept of "naturalness" and make the metric a more useful conservation tool.

--by Rob Goldstein   

Theobald, D. (2010). Estimating natural landscape changes from 1992 to 2030 in the conterminous US Landscape Ecology DOI: 10.1007/s10980-010-9484-z