Using GIS to Model the Risk of Mountain Pine Beetle Infestation in the Black Hills National Forest of South Dakota Melissa L. Mathis Department of Resource Analysis, Saint Mary’s University of Minnesota, Winona, MN 55987 Keywords: Mountain Pine Beetle, Dendroctonus ponderosae, Bark Beetles, GIS, Black Hills National Forest, South Dakota, Susceptibility Model, Ponderosa Pine, Pinus ponderosa Abstract The mountain pine beetle (Dendroctonus ponderosae) is causing large problems in South Dakota’s Black Hills National Forest. From 1996–2001, these destructive beetles killed more than half a million ponderosa pine trees in the northern and central parts of the forest. The dead trees are not only safety hazards but also increase the risk of wildland forest fires particularly near populated areas. A GIS was used in this project to examine the relationships between ponderosa pine (Pinus ponderosa) stand characteristics and the occurrences of mountain pine beetle infestations. A study area was selected in a severely infested site to identify three forest factors mountain pine beetles prefer. These forest factors were used in a susceptibility model to identify other tree stands that may be vulnerable to mountain pine beetle outbreak in the Black Hills National Forest. According to the model, the tree stands in the Northern Hills Ranger District were found to be the most susceptible to mountain pine beetle infestation. Introduction The mountain pine beetle (Dendroctonus ponderosae), a tiny insect about the size of a grain of rice, is killing thousands of acres of trees in the Black Hills National Forest (BHNF) of western South Dakota. Within the last six years, these destructive bark beetles have killed more than a half million ponderosa pine trees (Pinus ponderosa) in the northern and central part of the forest. These dead trees are not only safety hazards but also increase the risk of wildland forest fires, particularly near populated areas. Currently, the populations of the native mountain pine beetles are at epidemic levels and increasing annually (Allen and Long 2001). This epidemic is occurring largely because of favorable climatic conditions and favorable stands of ponderosa pine.

Ninety-seven percent of the 1.2 million acre Black Hills National Forest consists of ponderosa pine trees. Therefore, the US Forest Service and local logging companies are making large investments to control the mountain pine beetle (MPB) infestations. Management alternatives include: removing infested trees, thinning tree stands around infested areas, application of preventive insecticides, and identifying and protecting valuable tree stands (Allen and McMillin 2000). This study explores the spatial relationship between beetle-infested sites and ponderosa pine tree stands. A geographic information system (GIS) was used to examine the relationship between ponderosa pine stand characteristics and the occurrences of mountain pine beetle infestations. A susceptibility model was

Forest offers many recreational opportunities, some of which include, camping, fishing, gold panning, hiking, mountain biking, mountain climbing, and rock hunting, making the area a year round tourist destination for many people.

created to identify tree stands vulnerable to mountain pine beetle outbreak in the BHNF. Both ESRI ArcView 3.3 and ArcGIS 8.2 were used to produce a hazard map, which identifies high-hazard risk areas (Figure 11). These tools may be used to identify those areas where substantial losses of pine to MPB could occur; thus resources may be targeted to particular areas. Using this analysis, forest managers can also determine where suppression strategies are needed to meet management objectives and minimize ponderosa pine tree mortalities.

Mountain Pine Beetle The mountain pine beetle (MPB) is the number one insect killer of coniferous trees throughout western North America (McMillin and Allen 1999). The beetle is a native insect and attacks ponderosa pine trees in the Black Hills National Forest of South Dakota. The Black Hills National Forest is divided into three forest management districts. They are the Northern Hills, Mystic and Hell Canyon forest ranger districts. For this study, a project area spanning 36 miles wide by 60 miles long was chosen (Figure 1). In the BHNF, the life cycle of the

Black Hills The Black Hills range in elevation from 2900 ft-7050 ft and contain many national well-known attractions which include, Badlands National Park, Wind Cave National Park, Jewel Cave National Monument, Mount Rushmore National

Study Area

Detail of roadless area within the study area.

South Dakota Black Hills

Figure 1. Black Hills National Forest boundary map, showing the three forest management districts, populated areas, and study area.

mountain pine beetle lasts one year (Allen and Long 2001). In late summer, adults

Memorial, Crazy Horse and the Black Hills National Forest. The Black Hills National

2

forests, infestations are successful in tree stands that are dense, large and over 80 years old (Schmid and Mata 1992). As a result, these tree stands are often more susceptible to stress and beetle infestation. Parameters for tree stand density and size are provided in the analysis section of this paper. Once the beetle population reaches epidemic levels, natural predators like woodpeckers cannot reproduce quickly enough to maintain insect populations. The most effective natural control of beetle infestations is cold winter temperatures (Bellows et al. 1998). In order to kill off the beetles, temperatures of –24 degrees Fahrenheit need to be reached under the tree bark for a period of 24 hours (Schmid et al. 1993). The western portion of South Dakota has not had a severely cold winter for many years. Mild winters have resulted in high beetle survival rates and therefore population increases. The first recorded outbreak in the BHNF occurred in the 1890s in which an estimated 10 million trees were killed (Allen and Long, 2001). Approximately five outbreaks, each lasting 8-13 years, have occurred since that time though none has reached the same magnitude (Allen and Long, 2001). The outbreak in the early 1970s resulted in the loss of more than 440,000 trees. The last outbreak occurred from 1988 to 1992 and resulted in the death of approximately 50,000 trees (Pasek and Schaupp 1992). The most recent outbreak has been developing since 1999. Currently, mountain pine beetle populations are at epidemic levels in certain areas of the BHNF (Figure 2). Beetle outbreaks have caused considerable changes in the BHNF’s forest stands, including a reduction in average stand diameter and stand density (McCambridge et al. 1982). The outbreaks are also conflicting with land management objectives: they reduce timber stocking levels, affect wildlife habitat, increase short term fire risks, and can negatively alter

leave infested trees and attack uninfested green trees. An adult beetle chews through the tree bark and spends most of its short life there. The female produces approximately 75 eggs in an egg gallery (Samman and Logan 2000). Once the eggs hatch, the remaining larvae tunnel horizontally away from the vertical egg gallery. The larvae spend the winter under the bark of the infested tree. They eventually cut off the tree’s water supply and flow of nutrients. The adult beetle introduces a blue-stain fungus, which clogs the sapwood and blocks water from going to the roots and crown. The combination of the feeding beetle larvae and the blue fungus cause the tree to die (Samman and Logan 2000). Within a year the tree dies and the next generation of adults emerge. Mountain pine beetle populations are usually found at endemic (normal) levels. At normal population levels, beetles play a pivotal role in the development and maintenance of pine stands in the BHNF. The beetle will often attack trees under stress from injury, poor site conditions, fire damage, overcrowding, root disease, and old age (Samman and Logan 2000). Thus, as old trees die, openings in the forest promote biodiversity and overall healthier tree stands. Healthy trees can often resist beetle attacks by secreting pinesap or “pitch” that expels invading beetles (Samman and Logan 2000). In order for an outbreak to occur, four conditions must be present in the forest. The conditions include: stressed ponderosa pine trees, increasing number of beetles, favorable environmental conditions, and a triggering mechanism. An environmental disturbance may trigger a beetle outbreak. Some of the natural factors that may trigger an outbreak are, wind blown tree stands, dry conditions over several years (drought), mild winters, lighting strikes and fire (Samman and Logan 2000). All of these natural occurrences may lead to weakened, beetleprone trees. Tree stand conditions also contribute to outbreaks. In ponderosa pine

3

analysis was a vegetation coverage created by the USDA Black Hills National Forest Service office in Custer, SD. This dataset contains information pertinent to the forest from the RMRIS Oracle database as well as the spatial location of existing vegetation stands. The RMRIS database is very large and contains a wealth of detailed forest stand information. The data is current as of 05/10/2002. Other coverages used in this project that were obtained from the Forest Service: administration boundaries, digital elevation model, digital raster graph, forest coverage and a forest service boundary coverage. Six years of beetle infestation coverages were obtained from the Forest Service’s Rocky Mountain Research Center in Denver, CO. The coverages were created from aerial surveys flown to identify dead trees killed by the mountain pine beetle. A sketch mapping method, performed in a fixed wing airplane, was used to collect the data. Forest Service personnel sketched polygons on hardcopy 1:100,000 scale topographic maps and later digitized maps into an ArcInfo coverage. The resulting polygon coverage represents the estimated number of dead ponderosa pine trees. Coverage attributes include the number of dead trees, size of infestation, and severity of attack.

visual and recreation values (Samman and Logan 2000).

Photo from the Rapid City Journal

Figure 2. MPB outbreak near Beaver Park area.

In the BHNF, a number of factors have contributed to epidemic mountain pine beetle infestations. Some of these factors include drought, mild winters, dense, mature ponderosa pine stands, and past forest management practices. Decades of fire suppression have produced a landscape dominated by mature stands. As a result, larger, more contiguous landscapes have become susceptible to bark beetle outbreaks in the BHNF (Figure 3).

Methods A geographic information system (GIS) was used to examine the relationship between ponderosa pine stand characteristics and the occurrences of mountain pine beetle infestations. The extent and severity of the mountain pine beetle in the BHNF was estimated by comparing six years (19962001) of beetle infestation GIS coverages. These coverages were created from aerial survey flights flown every September for six years. The purpose of the survey was to estimate damage levels caused by the mountain pine beetle on ponderosa pine.

Photo from the Rapid City Journal

Figure 3. Dead trees killed by the MPB.

Data There were several datasets used in this project. The primary data set used in the 4

coverage showed a few dead trees scattered throughout the Black Hill’s three forest management districts. In 1997, the eastern part of the Northern Hills Ranger District experienced a noticeable increase in pine tree mortality. In 1998, another sharp increase was detected in the eastern part of the Northern Hills Ranger District. In 1999, the beetles spread through all three-forest districts. As listed in Table 1, a total of 27,700 trees were killed throughout the Black Hills National Forest in 1999. This represents a significant increase from 9,688

After acquiring ArcInfo polygon coverages from the Rocky Mountain Research center, this study proceeded by creating a project in ArcView 3.3. The Xtools extension produced for ArcView was used to create six beetle centroid point coverages. The centroid points represent the location of MPB infestation pockets. These were used in a point analysis to reveal patterns defining intensity and spatial distributions of the beetle infestations. According to the point analysis (Figure 4), the 1996 beetle infestation

1996

1999

1997

2000

1998

2001

Figure 4. Black Hills National Forest Districts showing location of Mountain Pine Beetle Infestation from 1996 – 2001 (one point represents 10 dead trees).

5

Table 1. Number of trees killed by the Mountain Pine Beetle in the BHNF.

Year 1996 1997 1998 1999 2000 2001 Total

occurred in the Beaver Park area. According to Table 1, a total of 352,170 trees were killed over a six-year time period in the Black Hills. The Beaver Park area contained over 100,000 of these dead trees. This amounts to an estimate of over 8 million cubic feet of volume lost (Allen and Long 2001). The second part of the point analysis entailed calculating a new field and adding it to the 1996-2001 beetle incident coverages. The new field, number of dead trees per acre (TPA), was calculated by using the map calculator in ArcView. The tree per acre index indicates where the greatest numbers of trees were killed, also known as a measure of MPB severity. In ArcView, the centroid points were displayed by using the graduated legend color type. The TPA attribute was used as the classification field. Three classifications were chosen to show low, moderate, and severe damage caused by MPB infestations (Figure 7). According to Figure 7, the Northern Hills Ranger District was impacted most severely, particularly the Beaver Park Area.

Dead Trees 1,105 4,795 9,688 27,700 38,000 275,030 352,170

trees killed in 1998 (Figure 5). In 2000, the tree mortality increased to 38,000 trees. Much of the damage was concentrated in groups of 25 trees or more. More than 60% of the total trees killed in the 1999 and 2000 beetle coverages occurred in the Beaver Park area. In 2001, a tremendous increase occurred in the number of trees killed (Figure 6). A total of 275,030 trees were killed compared to 38,000 trees in 2000. Based on the 2001 beetle coverage, most of the tree mortality appeared scattered in small groups or occurred as single trees. In 2001, large areas of concentrated mortality

1996-2000 Black Hills Ponderosa Pine Mortality due to the Mountain Pine Beetle

Dead Trees

40000 30000 20000 10000 0 1996

1997

1998 Year

Figure 5. 1996-2000 Pine mortality chart.

6

1999

2000

1996-2001 Black Hills Ponderosa Pine Mortality due to the Mountain Pine Beetle

300000 Dead Trees

250000 200000 150000 100000 50000 0 1996

1997

1998

1999

2000

2001

Year Figure 6. 1996-2001 Pine mortality chart.

2002). In order to determine forest characteristics beetles prefer, a study area of approximately 28,000 acres was selected around Beaver Park. The study area lies within Lawrence and Meade counties and is located within the eastern edge of the Northern Hills Ranger District (Figure 1). The site is located six miles southeast of the city of Sturgis and can be characterized by thick, dense ponderosa pine stands. The Forest Service owns 97% of the land in the Beaver Park Study area, including a 5,000acre federally protected roadless wilderness area. Forest Stand Factors

Figure 7. Black Hills National Forest districts showing the severity of Mountain Pine Beetle outbreaks, 1996-2001.

For the analysis, a forest polygon coverage was obtained from the Black Hills Forest Service Office. This dataset included all of the standard forest cover attributes from the Forest Service’s RMRIS Oracle database. The forest coverage feature attribute table contained approximately one hundred fields and 32,500 records. Some of the forest

Analysis Mountain pine beetles are at epidemic levels in the Beaver Park area (Allen and Schaupp

7

According to Figure 8, the densest forest stands are located in the eastern part of the study area. Figure 8 shows tree stands attributed by density in the Beaver Park Study Area. The density or basal area was classified into three categories according to low, medium and high density values which range from 0–280.

coverage attributes used in this project included: acres, administration district, basal area, elevation, forest stand age, ownership, slope position and tree size. The large number of attributes in the coverage made it difficult to determine which factors contributed to MPB infestations. ArcGIS Spatial Analyst extension provided a useful tool for ranking and summarizing multiple forest stand factors. The forest coverage was projected into UTM coordinates and clipped to the study area boundary by using ArcGIS Toolbox commands. The six polygon beetle coverages were also clipped to the study area boundary. The geoprocessing wizard in ArcView was used to perform an intersect overlay operation between the beetle and vegetation coverages. This resulted in six new intersected coverages containing vegetation data similar to the known MPB infestation sites. A series of summary tables and statistics were then created to determine the predominant forest characteristics within the 1996–2001 beetle intersect coverages. The determination of forest factors was based on a MPB hazard model developed by Stevens, McCambridge, and Edminster in 1980. The three predominant tree stand factors in the Beaver Park area are basal area (tree stand density), average diameter at breast height-DBH (tree size) and average tree stand age. The average basal area, or mean density, in the study area is relatively high (refer to Table 2).

Figure 8. Dense tree stands in Beaver Park Area

The second predominant forest factor is tree size, based on DBH. The average tree stand sizes in the Beaver Park area are listed below in Table 3. As shown in Figure 9, a majority of the Beaver Park area contain large trees, 9 inches and above. The third factor, tree stand age, is relatively constant in the study area. Most of the tree stands are over 80 years old. Therefore, three classifications were created. The younger tree stands ranged from 10-60 years old, medium trees were 60-80 years old and any tree stands over 80 years were classified as old.

Table 2. Mean Density of tree stands

Year 1996 1997 1998 1999 2000 2001

Mean Density ft2/acre 123 124 113 114 103 104

Susceptibility Modeling These three predominant factors that characterize the forests in the Beaver Park area were used to create a susceptibility

8

Table 3. 1996-2001 Acres impacted by the MPB and average tree stand size attacked Tree Size, DBH

< 5 in Small 5 - 8.9 in Medium 9-15.9 in Large Very Large >16 in

1996 acres 0 6 146 0

1997 acres 7 177 2190 0

1998 acres 28 634 2350 62

1999 acres 71 820 4120 72

2000 acres 49 418 2034 3

2001 acres 41 2124 5323 51

given to forest stands that are considered more susceptible to the mountain pine beetle. Therefore the grids were weighted accordingly: Reclass of Density Grid

.50 (50%)

Reclass of Age Grid

.25 (25%)

Reclass of Tree Size Grid

.25 (25%)

Each of the grid data sets were weighted according to their importance in making a forest stand susceptible to mountain pine beetle outbreaks. The tree density grid was given a higher weight than the age and size grids. Based on the Beaver Park outbreak, tree stand density was the most important factor in mountain pine beetle outbreaks. The higher the percentage, the more influence the grid dataset has in the hazard model. The raster calculator in Spatial Analyst was used to combine the three grid datasets. The equation used in the raster calculator is listed below.

Figure 9. Average tree stand size in the Beaver Park Area

model in the BHNF. The purpose of this model is to identify tree stands that may be susceptible to a mountain pine beetle outbreak. ArcGIS Spatial Analyst was used to create the susceptibility model. Three 30meter grid datasets were derived from the BHNF forest coverage. The grids were attributed by tree density, tree size and tree age. Each grid dataset was reclassified to a common rating scale (1-3). The reclassify menu under the Spatial Analysis toolbar was used to input the three number classifications. A lower value (1) indicates that a particular forest stand is less susceptible to mountain pine beetle infestation. The highest number, 3, was

[Density grid] * 0.5 + [Tree Size grid] * 0.25 + [Age grid] * 0.25 The output grid was classified into three data ranges and saved as a permanent grid. The raster calculator produced an output grid displaying tree stands throughout the BHNF matching the predominant characteristics of the study site. 9

(DOQs) in both districts, two areas were outlined as potential outbreak sites (Figure 10). Based on the image, dense ponderosa pine trees dominate the mountainous terrain. The second outlined area lies in the Mystic Ranger District. The risk model projected both of these areas to be highly susceptible to MPB infestations. In Figure 11, the MPB infestation severity data (Figure 7) was overlaid on the high-risk forest stands. According to Figure 11, some MPB infestations have already occurred in the neighboring areas especially in the Northern Hills and Mystic Ranger Districts. Due to the close proximity of MPB pockets, these two areas are prime candidates for an outbreak of epidemic proportions. Figure 12 shows an example of the terrain and tree density. If appropriate management strategies are applied soon, this outbreak could possibly be lessened. Possible management strategies may consist of thinning tree stands, removing infested trees and creating buffer strips around healthy forest stands.

Results According to Figure 10, tree stands in the Northern Hills and Mystic Ranger Districts appear to be the most susceptible to mountain pine beetle infestations.

Figure 10. Forest stands ranked as highly susceptible in the analysis grid.

According to the ArcView susceptibility model, the Northern Hills Ranger District contains 272,185 acres of ponderosa pine trees. In this district, 55,976 acres or 21% of the ponderosa pine trees were found to be highly susceptible to the MPB. The Mystic District contains 270,000 acres of ponderosa pines and the Hell Canyon District contains 303,950 acres. The Mystic Ranger District contained 29,694 highly susceptible acres (11%) and the Hell Canyon Ranger District contained 17,717 highly susceptible acres (.05%). Both the Northern Hills and Mystic Ranger Districts contain contiguous forest stands that meet the three forest factors beetles prefer. After looking at digital orthophotography quadrangle images

Figure 11. Overlay of MPB infestation severity centroid points and forest stands ranked as highly susceptible in the analysis grid.

10

Figure 12. Detail of DOQ within the outlined area in the Northern Hills Ranger District, displayed at a 1:6,000 scale to show an example of tree density.

Discussion forest inventory databases. Combined with GIS analysis tools and display techniques, this raw data has been extracted and transformed into useable information. The susceptibility model that has been generated will place valuable data in a format accessible to resource managers.

Given the epidemic proportions of the MPB infestation in the Beaver Park area, it will be crucial to assess the condition of the forest at the end of 2003. Comparing the susceptibility model produced in this study with the new data will add an important layer of information. Since the model proved effective in mapping previously susceptible stands, it can be used to determine if this trend continues. Also, areas where mitigation practices are put into place can be examined in view of this model. There might be relationships between the susceptibility of tree stands and the effectiveness of various levels of protection. This project makes use of the wealth of information stored in national USDA

Acknowledgements I would like to thank the members on my graduate committee Martha Roldan and John Ebert for their guidance and suggestions. I would also like to thank Dr. David McConville for his enthusiastic teaching style and patience. Also thanks to Beth Collins for her technical support and forestry expertise.

11

Forest Service, Rocky Mountain Region, Biological Evaluation R2-9204. Samman, S. and J. Logan, tech. eds. 2000. Assessment and response to bark beetle outbreaks in the Rocky Mountain area. Report to Congress from Forest Health Protection, Washington Office, Forest Service, USDA Forest Service General Technical Report RMRS-GTR-62. 46 p. Schmid, J.M. and Mata, S.A. 1992. Stand density and mountain pine beetlecaused tree mortality in ponderosa pine stands. USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Research Note RM-515. Schmid, J.M., Mata, S.A., Olsen, S.K. and Vigil, D.D. 1993. Phloem temperatures in mountain pine beetleinfested ponderosa pine. SDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Research Note RM-521.

References Allen, K.K. and Long, D.F. 2001. Evaluation of mountain pine beetle activity on the Black Hills National Forest. USDA Forest Service, Rocky Mountain Region, Biological Evaluation R2-02-02. Allen, K. K. and McMillin, J.D. 2000. Evaluation of mountain pine beetle activity on the Black Hills National Forest. USDA Forest Service, Rocky Mountain Region, Biological Evaluation R2-01-01. Allen, K.K.and Schaupp, W.C. 2002. Evaluation of mountain pine beetle activity in the Beaver Park Area of the Black Hills National. USDA Forest Service, Rocky Mountain Region, Biological Evaluation R2-03-01. Bellows, T.S., Meisenbacher, C. and Reardon, R.C. 1998. Biological control of arthropod forest pests of the western United States: A Review and Recommendations. USDA Forest Service, Forest Health Technology Enterprise Team-Morgantown, WV. FHTET-96-21. McCambridge, W.F., Hawksworth, F.G., Edminster, C.B. and Laut, J.G. 1982. Ponderosa pine mortality resulting from a mountain pine beetle outbreak. USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Research Paper RM-235. McMillin, J.D. and Allen, K.K. 1999. Evaluation of mountain pine beetle activity in the Black Hills National Forest. USDA Forest Service, Rocky Mountain Region, Biological Evaluation R2-00-03 Pasek, J.E. and Schaupp, W.C. 1992. Status and trends of mountain pine beetle populations in the Bear Mountain and White House Gulch areas of the Harney Ranger District, Black Hills National Forest, South Dakota. USDA

12