Methods
Data Sources
Two separate tree ring databases were used: a version of the International Tree Ring Database (ITRDB) revised by Manvailer and Hamann in 2022, and another database for Spruce tree rings in Alberta received from the Canadian Forest Service and used by Hynes and Hamman in a 2020 study. The databases were filtered to time series with values between 1901 – 2020 and within Canada. The Alberta Spruce TRDB had 443 trees at 113 sites, and 578 trees at 526 sites were used from the ITRDB.
Ecozones
Figure 1. A map showing the ecozones within Canada.
Each site was projected in ArcGIS Pro 3.0.0 (Esri Inc, 2022) and the locations were spatially joined to a shapefile outlining Canadian ecozones (Agriculture and Agri-Food Canada, 2021). This process associated each site location with an ecozone. Figure 1 shows the relative position of each of these ecozones within Canada. Trees in ecozones with less than 10 trees were removed, as well as species with less than 10 trees within an ecozone.
Climate Moisture Index
Figure 2. A map showing the distribution of the normal annual Climate Moisture Index (CMI) from 1961-1990 throughout Canada.
The monthly Climate Moisture Index (CMI) values for each site from 1901-2020 were extracted from ClimateNA (Wang, Hamann, Spittlehouse, & Carroll, 2016), (ClimateNA, 2022) by inputting the site locations. CMI is calculated as the difference between precipitation and potential evapotranspiration derived from Thornthwaite’s equation (Pereira & Paes De Camargo, 1989) and can be used as a proxy for relative moisture levels (Hogg, Barr, & Black, 2013). These monthly values were converted to annual values by summer the CMI for September to August for each year. The average geographic distribution of annual CMI values between 1961 and 1990 can be seen in figure 2, which demonstrates the relative moisture levels of each part of Canada.
Detrending
Figure 3. Example of detrending using the ring widths from series FHJB15--BAN-XS3-2-T4-C. The ring widths are shown in the top graph with the spline drawn throughout. This spline can be seen again in the bottom graph, reprojected as a flat line throughout the time series. The ring widths can be seen adjusted in relation to that line.
Tree ring widths were detrended using the Friedman method for each time series. Cross-validation was incorporated to choose the best method and a bass value of 0 was used to preserve annual variation. Figure 3 displays an example of detrending and how the values are adjusted.
Drought
Drought years were determined using CMI values. Values that fell below 10% of the difference between the mean and the minimum were considered as severe drought. This was calculated individually for each time series. The goal was to select CMI values that were likely to have a large effect relative to other environmental factors while accounting for series with higher variation and skewed distributions.
CMI Threshold = (meanCMI – minCMI)*0.1 + minCMI
The threshold and period for pre- and post-drought growth was chosen by iteratively calculating the mean difference between pre-drought ring widths and ring widths during drought using a pairwise t-test Table x.
Resilience
Figure 4. A graph displaying resilience indices and what they each represent relative to growth and drought (Lloret, Keeling, & Sala, 2011) .
Resistance(Rs), recovery(Rc), and resilience(Rs) were calculated for each instance of drought within each times series that also coincided with the presence of 5 years of ring widths centered on the drought year. Pre- and post-drought growth (PreDr, PostDr) were calculated as the average detrended ring width over 2 years either immediately before or after the drought year. Drought growth (Dr) was the detrended ring width during the year where the drought occurred.
Rt = Dr/PreDr
Rc = PostDr/Dr
Rs = PostDr/PreDr
Resilience indices were averaged for each timeseries and then for each species within each ecozone. A graphical representation of each index and how they relate to drought timing and growth can be seen in figure 4.