A climate analog approach to evaluate seed transfer and vegetation transitions

• Published in: Frontiers in Forests and Global Change
• Main Authors: Richardson, Bryce A., Rehfeldt, Gerald E., Sáenz-Romero, Cuauhtémoc, Milano, Elizabeth R.
• Format: Article in Journal/Newspaper
• Language: unknown
• Published: Frontiers Media SA 2024
• Subjects: Arctic
• Online Access: http://dx.doi.org/10.3389/ffgc.2024.1325264; https://www.frontiersin.org/articles/10.3389/ffgc.2024.1325264/full

Introduction There is growing recognition that restoring species diversity is crucial to maintaining ecological functions and services. Increasing the diversity of species used in restoration programs has placed greater emphasis on determining the seed transfer needs for a wider array of plants. However, many plants, outside of commercial forestry, lack information that would provide guidance on seed transfer for current or future climates. Generalized seed transfer approaches use climate partitioning to approximate adaptive differentiation among populations and provide an estimation of seed transfer distance for such species. Methods Herein, we describe a generalized seed transfer approach that uses Euclidean distance of 19 climate variables within North America (from northern Honduras to the Arctic). Euclidean distances are used to identify climate analogs from vegetation databases of about 685,000 plots, an average density of 1 plot per 32 km 2 . Analogs are classified into three thresholds (strong, moderate, and weak) that correspond to altitudinal climate gradients and are guided by the scientific literature of observed adaptive variation of natural tree populations and seed transfer limits. Results For strong threshold observations, about 97% of the analogs had climate distances equivalent to ≤300 m elevation, whereas for the weak threshold observations, 53% had an elevation equivalence of ≤300 m. On average 120, 267, and 293 m elevation separated two points under strong, moderate, and weak thresholds, respectively. In total, threshold classification errors were low at 13.9%. Discussion We use examples of plot data identified from a reference period (1961–1990) and mid-century (2056–2065) analogs across North American biomes to compare and illustrate the outcomes of projected vegetation change and seed transfer. These examples showcase that mid-century analogs may be located in any cardinal direction and vary greatly in spatial distance and abundance from no analog to hundreds depending on the site. The ...