PART 0
Data Preparation
For FRAGSTATS
White spruce raster data was first clipped to the study area, then reclassified into 4 classes: 0-4.9%, 5-9.9%, 10-39.9%, 40-100%.
​
For Maxent
ArcMap model builder was used to convert data to formats readable by Maxent (Figure 1).
After calculating the XY coordinates for each grid, cells with a value of 40% or higher were reclassified as white spruce stands. A total of 408 stands were found within the study area. Only the stand data were used in Maxent analysis.
​
​
​
​
​
​
​
​
Figure 1. ArcMap model for processing tree data
Bioclimate data were clipped to the study area then converted into ACSII format (Figure 2).
​
​
​
​
​
​
​
​
Figure 2. ArcMap model for processing environmental variables
​
Overview
Warming temperature is expected to fragment boreal forests and undermine their ability to provide ecosystem services (Murray et al., 2017). However, limited statistical research has been done on the extent of fragmentation within the context of the Boreal Shield East of Canada. Extreme weather events coupled with large-scale forest fire have caused tremendous forest disturbances in Ontario (Natural Resources and Forestry, 2019). Mean annual temperature change during 2001-2011 fluctuated but was seen as part of the rising historical climate trends (Ontario Centre for Climate Impacts and Adaptation Resources, n.d.).
​
​
​
To address this knowledge gap, FRAGSTATS was used in the first part of this study to calculate the different indexes indicating the level and extent of fragmentation in the study area.
FRAGSTATS is a set of spatial statistics for landscape ecology that facilitates the evaluation of landscape processes. Class metrics were utilized in this analysis. The short definitions for each metric used are provided in Table 1.
​
​
​
FRAGSTATS analysis
​
Overview
​
MaxEnt is one of the most popular tools for species niches and distributions modelling. It is a type of Species Distribution Model that is commonly used to predict the geographic range of a species given presence-only occurrence data and environmental variables assumed to influence its distribution (Merow et al., 2013). Applying a machine-learning approach, MaxEnt predicts the distribution of a species based on the theory of maximum entropy (Merow et al., 2013). The model generates a probability distribution from sets of environmental grids and occurrence records.
​
In this stage of the research, MaxEnt was employed to (1) examine which bioclimate variables contribute the most to the distribution of white spruce; (2) investigate how well does the model explain the actual distribution of white spruce; and (3) produce pictures of predictions that can be used as a baseline reference.
​
The analysis was replicated 10 times for cross-validation. Two kinds of analysis were done with MaxEnt.
​
Pictures of predictions
Regions with suitable conditions for white spruce were identified under the current climate conditions. The result will be used as a reference compared with future projection. The output is in the cloglog format.
​
Variable importance
Environmental variables were narrowed from 19 to 8 to refine the results. However, it is noted that annual variables have a bigger influence on long-term climate change-induced species distribution.
Maxent
Current Distribution
Overview
How does climate change influence the distribution of white spruce? Are the impacts detrimental or beneficial? Will this result in a northward shift in tree range in the study area?
​
To answer these questions, the final section of this study focuses on the modelling of white spruce distributions under changing climate conditions. Future climate conditions under 4 climate regimes estimated using general circulation models were examined. Special attention was paid to the migration of suitable habitat range along the longitude, as well as the change in suitable habitat size and distribution. The results of this section were compared to the anticipated response of boreal forest in other research.
Finally, results were exported to ArcMap to visualize changes across different climate scenarios against the current presence of white spruce.
Maxent
Future projection
Table 1: Short description of class level metrics used in this analysis.