Our modeled species of this project, white spruce (Picea glauca) is widely distributed in the North American boreal forest and extensively favored for its high commercially and ecologically value. However, climate change may directly influence the health and productivity of white spruce. Building the species distribution model for white spruce can help us identify the major climatic factors influencing its distribution, and then estimate the habitat suitability for the future climate. As one dominant tree species in the boreal forest, its distribution pattern also plays essential roles in the local species composition and diversity. Modeling the past can also provide valuable evolutionary insight of the species.
In this study, the species distribution model(SDM) of white spruce is built on over 60,000 observations of the species occurrence across North America. We collected the climate condition for each point and selected ten the most commonly used climatic factors as predictor variables. Random forest, one ensemble machine learning method, is chosen for its high accuracy comparing with other approaches.
With the Community Climate Model(CCMl) and Geophysical Fluid Dynamics Laboratory model(GFDL) we reconstructed the white spruce distribution in the periods of 21,000, 14,000, 11,000, 6,000 years before present(BP). The prediction approximately matches the fossil pollen records from Neotoma database. The distribution patterns are indicating the white spruce populations may used Beringia as refugium during the last glacial maximum.
Our modeled species of this project, white spruce (Picea glauca) is widely distributed in the North American boreal forest and extensively favored for its high commercially and ecologically value. However, climate change may directly influence the health and productivity of white spruce. Building the species distribution model for white spruce can help us identify the major climatic factors influencing its distribution, and then estimate the habitat suitability for the future climate. As one dominant tree species in the boreal forest, its distribution pattern also plays essential roles in the local species composition and diversity. Modeling the past can also provide valuable evolutionary insight of the species.
In this study, the species distribution model(SDM) of white spruce is built on over 60,000 observations of the species occurrence across North America. We collected the climate condition for each point and selected ten the most commonly used climatic factors as predictor variables. Random forest, one ensemble machine learning method, is chosen for its high accuracy comparing with other approaches.
With the Community Climate Model(CCMl) and Geophysical Fluid Dynamics Laboratory model(GFDL) we reconstructed the white spruce distribution in the periods of 21,000, 14,000, 11,000, 6,000 years before present(BP). The prediction approximately matches the fossil pollen records from Neotoma database. The distribution patterns are indicating the white spruce populations may used Beringia as refugium during the last glacial maximum.
DISCLAIMER: This is a class exercise based on modified or randomly generated datasets, and this website is developed only for the purposes of RENR 690 course.