Dataset of mountain pine beetle outbreak dynamics and direct control in Cypress Hills, SK

The data presented in this article are related to the research article entitled “Mountain pine beetle outbreak duration and pine mortality depend on direct control effort” [1]. This article provides presence of mountain pine beetle infested trees detected by the Saskatchewan Forest Service on a grid covering the spatial extent of the Saskatchewan portion of the Cypress Hills interprovincial park between 2006 and 2018. For each grid cell, associated ecological and environmental covariates, such as topography, weather and vegetation, are also provided. These data cover the spatio-temporal extent of an almost entire mountain pine beetle outbreak and contribute to the understanding of mountain pine beetle outbreak dynamics.


Data description
The data presented in this article are related to the research article entitled "Mountain pine beetle outbreak duration and pine mortality depend on direct control effort" [1]. Data are provided for each year between 2006 and 2018 and for each cell of a grid covering the spatial extent of the study area. Two datasets are provided for two grid spatial resolution: a grid of 18,317 100 Â 100 m cells (dataset 1) and a grid of 722 500 Â 500 m cells (dataset 2). For each cell of a grid, the dataset contains the presence or number of mountain pine beetle infested trees as well as topography, weather, vegetation, and management factors ( Table 1). The topography factors included are latitude, longitude, elevation, slope, aspect, northerness, easterness. The weather factors included are maximum temperature in summer, minimum temperature in summer, minimum temperature over the year, degree-days, cold tolerance, relative humidity in spring, soil moisture index, wind speed in summer, and an estimate of mountain pine beetle emergence peak from degree-days. The vegetation factors include pine cover, pine height, pine age, and number of pines. The management factors included are presence of managed and unmanaged mountain pine beetle infested trees the previous year in the same cell and in neighbouring cells, and the distance between the cell and a limited patch of unmanaged infestations outside the park's southern limit.

Mountain pine beetle infested trees
Saskatchewan Ministry of Environment Forest Service is responsible for surveying mountain pine beetles in this portion of the park. Every year, a complete aerial survey of the park extent is performed Specifications Table   Subject  Ecology  Specific subject area  Insect outbreak dynamics  Type of data  Tables  How data were acquired  Aerial and ground surveys, collection of data across literature and  Value of the Data These data are useful for understanding the fine-scale dynamics of an almost complete mountain pine beetle outbreak limited in its spatial range. Researchers and managers can benefit from these data to understand outbreak dynamics or to inform decision-making processes and management strategies. These data can be used to further insight spatial dynamics of dispersal and can be used for modelling.
in order to detect red-top trees [3]. These are later ground-truthed for beetle attacks and trees fading from other causes than mountain pine beetle are removed from the survey. Ground surveys with a radius of 50 m are conducted around each red-top tree containing beetles to find currently infested pines with live brood. However, in some areas, red-top trees are clustered together. To thoroughly  Distance from the centre of the cell to the park's southern limit close to external infestations m survey these areas, polygons are delineated by hand around each red-top trees cluster. The extent of each polygon is then entirely checked for infested trees using line surveys. Polygon locations and shapes typically change from one year to the other. However, they are consistently located for the most part in the same highly infested and therefore highly surveyed areas. Infested pines located outside of the circular plots or polygons would be most likely missed by managers. All detected infested trees are treated, primarily using a fall and burn tactic to ensure that the broods are killed. The Forest Service has been following this procedure since the mountain pine beetle infestation was detected in 2006 up to the current state of the outbreak in 2018. In summary, we obtained the following data for each year: locations of red-top trees, locations and shapes of polygons, number of infested trees for each circular survey, and locations of infested trees within the polygons. From this information, we estimated the locations of all infested trees.
To get an estimate of all of the infested trees location for every year, we used the following methods.
For the controlled infested trees found in circular survey plot (called type C1), the exact coordinates of the infested trees were not recorded so we used the location of the plot's centroid. For the controlled infested trees found in line surveys (i.e. within polygons, called type C2), we used the exact location of the infested trees.
For the uncontrolled infested trees that were missed at year t, became red-top trees at year tþ1 and were not part of a polygon at year tþ1 (called type U1), we used the exact location of the red-top trees at year tþ1. For the uncontrolled infested trees that were missed at year t, became red-top trees at year tþ1 and were part of a polygon at year tþ1 (called type U2), the exact coordinates of red-top trees were not recorded. We first assumed that, in Cypress Hills, the proportion of uncontrolled infested trees within these highly surveyed areas is 0.11 [4]. Therefore, the proportion of controlled infested trees within ground survey areas is 1e0.11 ¼ 0.89. Note that this number is different from the proportion of controlled infested trees in the entire park as ground surveys do not cover the entire park extent. Using these proportions, the number of controlled trees in polygons is equal to 0.89 times the number of infested trees. So, the number of infested trees in polygons is the number of controlled trees in polygons divided by 0.89. Therefore, the number of uncontrolled infested trees in polygons is # uncontrolled infested trees ¼ 0.11 Â # infested trees ¼ 0.11 Â # controlled infested trees /0.89 Second, using the equation above, we modelled, for each tþ1 polygon, the number of infested trees that would have been missed and fell within the polygon areas at year tþ1 as a Poisson random variable with mean equal to the number of uncontrolled infested trees at t that fall into a tþ1 polygon.
Third, we randomly distribute in space the U2 infested trees in each polygon area. In total, from 2006 to 2018, we had 2672 trees of type C1, 740 trees of type C2, 1819 trees of type U1, and 93 trees of type U2 (Fig. 1). These numbers are consistent with the previously estimated control efficiency over the entire park extent: 60.6% and 73.7% of infested trees were controlled in 2011 and 2012, respectively [4]. We were able to estimate the location of all infested trees every year with a precision of a few meters for 48% of the infested trees (types C2 and U1), 50 m for 50% of the infested trees (type C1), and several hundred meters depending on the polygons size for the remaining 2% (type U2). Finally, a grid with cell size 100 Â 100 m or 500 Â 500 m is superimposed over the park extent and the presence or number of infested trees in each cell is recorded.

Ecological and environmental covariates
We obtained topography variables from the Canadian Digital Elevation Map downloaded from the Geogratis website (geogratis.cgdi.gc.ca). We extracted the elevation, slope, and aspect. The aspect was converted into northerness and easterness using, respectively, a cosine and sine transformation. We estimated weather variables at the centroid of each grid cell for each year using the software BioSIM and the associated weather station data [5]. We obtained an estimate of the mountain pine beetle emergence peak each year by using the model and parameters provided in Ref. [6] (model 1) and cumulative degree-days from BioSIM [5].
From Ref. [2], we estimated, for each cell in 2001 and 2011, the leading species height, the pine cover and the tree volume. Then, we obtained pine height by using the leading species height when the pine cover is greater than 50%. We spatially interpolated the pine height values at the location of each cell using bicubic spline interpolation provided by the function interp of the R package akima [7]. We obtained the pine volume by multiplying the pine cover by the tree volume. To obtain vegetation variable values for every year, we linearly interpolated the values over the time period for each cell. We estimated the number of pine trees in each cell from the pine volume using the process and equation described in Ref. [8]. The expected number of pines greater than 10 cm at breast height E(S) depends on the pine volume per hectare V following equation E(S) ¼ A V exp(-d V) where A and d are free parameters. Pines with a diameter at breast height smaller than 10 cm are seldom the target of mountain pine beetle attacks and therefore were not included in the pine count [9]. Based on [10], Cypress Hills has a site index of 15e18 m at 50 year breast height age for lodgepole pine which corresponds to a medium to good site in Ref. [11]. Therefore, we fit A and d to data from the yield tables for medium and good sites provided in Ref. [11] using a nonlinear regression and obtained the values A ¼ 18.18 (±0.23 SE) and d ¼ 5.4 Â 10 À3 (±0.4 Â 10 À3 SE).