A Word from Alpine Tundra: Watch Out, Forests Are Invading! – Spatial
Detection of Alpine Treeline Ecotones in the Western United States
Abstract
Human-mediated climate change over the past century has made significant
impacts on global ecosystems and biodiversity including accelerating
redistribution of the geographic ranges of species. In mountainous
regions, the transition zone from continuous closed-canopy subalpine
forests to treeless alpine tundra areas at higher elevations is commonly
referred to as ‘Alpine Treeline Ecotone’ (ATE). Globally, warming
climate is expected to drive the ATE upslope, which could lead to
negative impacts on local biodiversity and modify ecosystem function.
However, existing studies rely primarily on field-based data which are
difficult and time consuming to collect. In this research, we define
three critical characteristics of the ATE including 1) an abrupt spatial
shift in vegetative activity as elevation varies, 2) reduction in
vegetative activity as elevation increases, and 3) vegetative activity
is at an intermediate level. Using the geospatial tools provided by
Google Earth Engine, we construct an index (ATEI) to identify areas with
the three ATE features based on the image gradients of vegetative
activity and elevation datasets. Based on the ATEI and Google Earth
imagery in 115 Landsat pixels, we establish a Logistic regression model
to estimate the probability of whether or not a sampled pixel is located
within the ATE. The prediction accuracy is approximately 80%.
Furthermore, the ATEI-estimated ATE elevation is strongly correlated (r
= 0.96) with a set of field-based data at 20 sampling sites from across
the region. Based on the average annual ATEIs from 2009 to 2011, we
estimate the average ATE elevation for each mountain range in the
western U.S. The result varies from 1,183 m to 3,584 m. The detection
metric developed in this study facilitates monitoring the geographic
location and potential shifts of ATEs as well as the general impact of
climate change in mountainous regions during recent decades. We also
expect this approach to be useful in monitoring other ecosystem
boundaries.