Spatiotemporal Relationships

Publication:

Nelson, Michael F., Mark Ciochina, and Christopher Bone. “Assessing Spatiotemporal Relationships between Wildfire and Mountain Pine Beetle Disturbances across Multiple Time Lags.” Ecosphere 7, no. 10 (October 1, 2016): n/a-n/a. https://doi.org/10.1002/ecs2.1482.

Background

Native bark beetle outbreaks and wildfire are two of the most prominent disturbances in western North America. The goal of this project was to examine the support of the commonly held view that wildfires are more common in the years following large beetle epidemics. The rationale is that the beetles leave large stands of dead trees, which then dry and become more susceptible to wildfire.

Approach

To examine the relationship between Mountain Pine Beetle (MPB) outbrteaks and subsequent wildfires, we used large-scale spatial datasets of beetle outbreak and wildfire in Washington and Oregon.

We used an overlay approach in which we calculated the spatial overlap between one disturbance and the other disturbance with varying lag times in between. For example, we calculated the area of overlap of wildfire in 2006 with MPB outbreak on 2008 to examine the 2-year lag between fire and subsequent MPB activity. We calculated this overlap area for all possible pairs of years (separated by 2 years) in the datasets. We repeated the process for time lags of up to 26 years. We examined time lags in both directions, that is the lag time between MPB and subsequent fire, as well as fire and subsequent MPB outbreak.

Figure 5: This map shows the spatial distribution of both disturbance types in the study area, as well as details of MPB epidemics before and after a large fire (the Tripod Complex Fires in 2006).

Main Findings

There was much greater than expected spatial overlap between wildfires and subsequent MPB outbreaks at time lags of about 1 - 4 years.

Figure 7: This figure shows the expected and observed spatial overlap of wildfires followed by MPB outbreaks. The panels on the left show the area of overlap, the right panels show the ratio of observed to expected areas.

The pattern of fire following MPB outbreak was more complex, but greater than expected overlap was observed for time lags of up to about 4 years. Severl high-fire years were especially influential in the pattern (1988 and 2006).

Figure 6: This figure shows the expected and observed spatial overlap of MPB outbreaks followed by wildfires. The panels on the left show the area of overlap, the right panels show the ratio of observed to expected areas.

Conclusions

Our results do show that there are higher than expected spatiotemporal overlap between the two disturbances, however the results do not necessarily imply causality. There was no attempt to account for spatial autocorrelation, and other factors, notably drought, are known to affect the likelihood of both disturbances.