Discussion
Does retention forestry (specifically the creation of island remnants) during a harvest successfully mimic island remnants created during a wildfire with respect to tree mortality?
In terms of harvest islands and fire islands having similar mortalities, we actually see that there is a significant difference in these disturbance types but that harvest islands have lower mortality than fire islands. This is different from what I expected, although may be explained by the how edges are created differently between natural and anthropogenic disturbances. Wildfires may burn more along the edge of an island remnant, creating less obvious edges and initially leaving behind trees that were partially burned but did not die during the disturbance. However, these trees may then be more susceptible to factors that cause mortality, whereas a cutblock has a very distinct edge and trees left behind intentionally may not be as damaged following the disturbance.
Interestingly, edge effects don't seem to influence tree mortality amounts when all explanatory variables are combined, this could be because the boreal forest commonly has natural edges (wetlands, lakes, etc.) and so this forest type is better adapted to edges. Additionally, our interior plots were located only 30 m interior from the disturbance edges, so it's possible that the interior plots were still getting that edge influence and we didn't have a stark enough contrast between the two site types. However, fire island edges did have higher overall mortality than other plot types, but overall edge and interior plots of all treatment types combined were not significantly different in mortality amounts. Further research could look at studying if the influence of fire created higher mortality rates in island edges--potentially the effects of high heat or burned boles weakened trees in island edge plots, and this contributed to higher mortality rates seen in fire island edges. A potential hypothesis to test would be that fire island edge plots have higher mortality because of partially burned trees that eventually die due to high heat and flame exposure, this could be tested by recording burn marks and evidence of charring on standing stems in plots.
For the most common boreal tree species, there doesn't seem to be a clear relationship between explanatory variables and mortality response, although a handful of our plots had higher mortality rates of certain species (Betula spp. and Populus spp.), but this was due to the fact that these species weren't present in all our sampled plots. Future studies or data analysis could explore further the impact of fire and harvest on specific boreal species.
Island size does not show a relationship to mortality, however our island sizes were fairly tight in distribution. Because we wanted to compare fire and harvest islands of similar size, we had to narrow down the pool of island candidates to get a decent comparison. The distribution of fire islands in general (including islands we did not select for field sites) shows a much larger distribution in size with more islands of larger size, however harvest islands were generally on the smaller size. Future work could track larger fire islands and corresponding mortality rates to see if there is evidence for leaving larger island sizes during a harvest to mitigate tree death.
Overall, we can say that the creation of island remnants during a harvest doesn't result in higher mortality than naturally created island remnants from fire, which gives evidence to support continued implementation of Ecosystem Based Management for forest recovery and resilience following disturbances.
In terms of harvest islands and fire islands having similar mortalities, we actually see that there is a significant difference in these disturbance types but that harvest islands have lower mortality than fire islands. This is different from what I expected, although may be explained by the how edges are created differently between natural and anthropogenic disturbances. Wildfires may burn more along the edge of an island remnant, creating less obvious edges and initially leaving behind trees that were partially burned but did not die during the disturbance. However, these trees may then be more susceptible to factors that cause mortality, whereas a cutblock has a very distinct edge and trees left behind intentionally may not be as damaged following the disturbance.
Interestingly, edge effects don't seem to influence tree mortality amounts when all explanatory variables are combined, this could be because the boreal forest commonly has natural edges (wetlands, lakes, etc.) and so this forest type is better adapted to edges. Additionally, our interior plots were located only 30 m interior from the disturbance edges, so it's possible that the interior plots were still getting that edge influence and we didn't have a stark enough contrast between the two site types. However, fire island edges did have higher overall mortality than other plot types, but overall edge and interior plots of all treatment types combined were not significantly different in mortality amounts. Further research could look at studying if the influence of fire created higher mortality rates in island edges--potentially the effects of high heat or burned boles weakened trees in island edge plots, and this contributed to higher mortality rates seen in fire island edges. A potential hypothesis to test would be that fire island edge plots have higher mortality because of partially burned trees that eventually die due to high heat and flame exposure, this could be tested by recording burn marks and evidence of charring on standing stems in plots.
For the most common boreal tree species, there doesn't seem to be a clear relationship between explanatory variables and mortality response, although a handful of our plots had higher mortality rates of certain species (Betula spp. and Populus spp.), but this was due to the fact that these species weren't present in all our sampled plots. Future studies or data analysis could explore further the impact of fire and harvest on specific boreal species.
Island size does not show a relationship to mortality, however our island sizes were fairly tight in distribution. Because we wanted to compare fire and harvest islands of similar size, we had to narrow down the pool of island candidates to get a decent comparison. The distribution of fire islands in general (including islands we did not select for field sites) shows a much larger distribution in size with more islands of larger size, however harvest islands were generally on the smaller size. Future work could track larger fire islands and corresponding mortality rates to see if there is evidence for leaving larger island sizes during a harvest to mitigate tree death.
Overall, we can say that the creation of island remnants during a harvest doesn't result in higher mortality than naturally created island remnants from fire, which gives evidence to support continued implementation of Ecosystem Based Management for forest recovery and resilience following disturbances.