Times are displayed in (UTC-05:00) Central Time (US & Canada) Change

Ancient Trees on The Move: Mapping the Historical Distribution of Blue Oaks in California

Session: 37th Annual Undergraduate Research Exhibition Sponsored by Sigma Gamma Epsilon (Posters)

Presenting Author:

Lily Weissman

Authors:

Weissman, Lily¹, Rose, Brooke², Viteri, Maria³, Mychajliw, Alexis⁴

(1) Earth and Climate Science and Biology Departments, Middlebury College, Middlebury, VT, USA, (2) Geography Department, San Diego State University, San Diego, CA, USA, (3) California Chapter (Frank and Joan Randall Preserve), The Nature Conservancy, Tehachapi, CA, USA, (4) Earth and Climate Science and Biology Departments, Middlebury College, Middlebury, VT, USA,

Abstract:

Blue oaks (Quercus douglasii) are a keystone species in California woodlands, and their future distribution is expected to shift due to climate change, but it remains uncertain which areas will serve as viable refugia. As the climate warms and droughts become more frequent in California, it is predicted that the southern extent of their habitat will become less suitable. Consequently, it is expected that the trees will migrate northward, seeking higher altitudes, toward the foothills of the Sierra Nevada Mountains. We modeled the historical habitat range of the blue oak to both better predict future habitat suitability and to inform modern conservation efforts.

We developed an ensemble of species distribution models (SDMs), using over 13,000 unique occurrence records compiled from public databases, including iNaturalist, Calflora, and iDigBio from 1900 to present. We assessed variable importance to determine key environmental drivers shaping blue oak habitat suitability, using eight predictors informed by ecological knowledge and prior work: minimum and maximum temperature, precipitation, percent clay, soil depth, available water capacity, soil pH, and percent rock fragments. The ensemble model incorporated multiple statistical and machine learning algorithms, including random forest, MaxEnt, boosted regression trees, artificial neural networks, generalized additive models, generalized linear models, and Gaussian process models. Model outputs were integrated based on individual model performance, enabling more robust habitat suitability predictions.

In addition to mapping current tree distributions, we hindcasted these models using Holocene paleoclimate data (1km resolution CHELSA data) to reconstruct habitat distributions from the Last Glacial Maximum to present. This temporal perspective allows us to identify past refugia areas and assess how long-term environmental drivers have shaped the species’ range. Ongoing comparisons with modern recruitment data and future climate projections (using 270m resolution Basin Characterization Model outputs) will help evaluate trends between historical refugia, future suitability, and modern recruitment rates.

To compare modeled habitat suitability with current recruitment patterns, we conducted fieldwork at the Frank and Joan Randall Preserve, a Nature Conservancy preserve near Tehachapi, California. This fieldwork assessed whether projected habitat losses impact current blue oak recruitment rates. Our research supports the development of tailored conservation strategies utilizing comprehensive analyses of paleoclimate trends and historical habitat ranges to help enhance blue oak resilience in the face of climate change today.