This model is a heavily simplified approach at understanding the spatial-temporal distribution of the MPB. Although simple and reductionist in nature, this a strength of the model. Large, complex holistic models are limited by increasing amounts of error due to uncertainties about the interaction of variables as well as their relative importance. Pragmatically, holistic models also require more understanding and more computing power, which although broader and more "complete" in scope, can affect the timlieness and accuracy of analysis. With this in mind, this model provides the basis to be enhanced to better reflect the dyanmic properties of nature. Many of these properties could be be enhanced by a better understanding of population dynamics and host-prey interactions.
1. Rate of spread
In this study, we assumed a uniform semi-circular radial expansion. However, we know this not to be true for various reasons such as that MPB populations flucuate spatially-temproally reflecting non-uniform outbreak events. As well, dispersal can cause rapid and random appearance far from establish populations (e.g. the MPB can disperse over 100 km over the Rocky Mountains).
2. Host density
In this study, we used an estimate of host density (percent pine x total tree volume) while assuming all species and individuals are equally suitable hosts. This is problematic as ecological interactions are species-specific, while variation exists in interaction exists at an individual level. Understanding different interactions between the MPB and different hosts (e.g. lodgepole vs. jack pine) and different individuals (e.g. saplings vs. mature stands) may lead to a more accurate prediction. As well, we assumed that the resistance of a cell was inversely related to the pine tree volume (as similar to Koch 2008). However, the resitance equation was only scaled to reflect the maximum pine tree volume and may not reflect the true slope of the line. Understanding how the MPB interacts with hosts at a community and population level will also aid in more accurate predictions. Further, understanding how tree composition will change in the future will help the researchers and the decision makers understand the potential spread of the MPB. With increasing temperatures, an increasing expansion of trees northward should be expected, allowing for new regions for potenital MPB establishment.
3. Climate suitability
In this study, we were only concered with -40oC minimum temperatures as it corresponds with 100% population mortality rates. However, understanding how MPB populations change overwinter could aid in this analysis. With decreasing temperatures, MPB mortality increases. Understanding how a 50% population motality event versus a 90% population mortality event affects rate of spread could provide better projections.
For all these variables, they are subject to change over long temporal scales due to evolutionary reponses. Overall, this model have many different areas that could be investigated to better understand its ecological framework.
Acquired data ranged from 10 m - 1 km resolution. This creates issues in terms of scalability. This analysis was performed at a 250 m resolution (same as MODIS data), with MPB data being scaled up. Climate data was at 1 km resolution, althougth it did not affect the analysis, finer data may result in more areas of limiting winter conditions.
Acquired data for the MPB came from diferent authorities which had different standards for collecting and managing data. This creates issues when trying to join datasets. BC data was at a coarser resolution than Alberta data, while, BC distributions were published anually, while Alberta data was more or less, every two years.