An Analysis of Gypsy
Moth Damage Potential in Minnesota

 By Dr. Thomas Eiber, MN DNR-Forestry


With the specter of gypsy moth knocking at our door, we need to take a deep breath, lean back, evaluate the situation, and earnestly begin to prepare for a forest ecosystem that includes an increased defoliation stress component. Over the next few issues, we will engage in several important discussions relating to this pending, inevitable "invasion". In this issue, we will discuss the damage potential to Minnesota's forests caused by gypsy moth. In subsequent issues, we will discuss the specifics of predicted damage and some approaches to modified silvicultural practices that can minimize the impacts, both economic and ecological, to our forested systems.

A Primer of Gypsy Moth Damage Potential In Minnesota

Defoliation is easy to predict while damage potential, ie. impact, is much more difficult. It is safe to assume that effectively all forested systems in Minnesota will feel the bite (no pun intended) of gypsy moth defoliation. Oaks and aspen, along with birch and maple, are all devoured with varying degrees of vigor. With gypsy moth established in the Sudbury area in Ontario and central Maine, it is unwise, perhaps downright foolish for us to assume that Minnesota's "legendary" winter will provide anything like immunity from the pest. There is little evidence that indicates that we will NOT be completely infested within the next ten to fifteen years.

Defoliation is not damage, at least not in any serious sense other than visual and aesthetic. Real? Yes, but limited. True damage will occur when trees loose vigor and begin to die. As we have progressed through the maxi-drought in 1988 and localized mini-droughts in the 1990's, we have come to understand that the ultimate impact of stress is tree death, a.k.a. mortality. We have also learned that tree mortality is not as easy to understand as it is to say. Trees die from stress, more specifically, from too much stress. We also know that stress is cumulative and that it comes from many sources.

As a result, we clearly must understand that damage from gypsy moth will occur NOT just in areas that are defoliated, but will occur in areas where trees are stressed from other causes and are defoliated. One of these "other causes" is the competition caused by normal stand structure and overcrowding. Silviculturists have long called small, understory trees in a forest "suppressed" for good reason. Gypsy moth will add another stress, but that is a story for another day.

Environmental Stress Plays an Important Role

Minnesota lies astride one of the most dramatic ecological boundaries in North America. The divide between the grassy plains in the central continent and the blanket of forest to the east. We clearly understand the issues of the edge. Climate, particularly rainfall patterns, plays a major role in the health of forests over a great proportion of the state, particularly the hardwood band swiping from northwest to southeast called the hardwood forest. This is a zone of ecological tension. Wet years and dry years. Rainfall, largely expressed by shortfalls in evapotranspiration, drives forest ecology in this zone back and forth, prairie to forest to prairie to forest and so forth.

Soils play an important role in this tug-of-war by effectively buffering short and medium term shortfall events. We have long since thought of sandy soils as "droughty". Actually, there is no evidence that it rains less on sandy soils. Simply, sandy soils have a poor ability to hold rainfall and to effectively store moisture. Loamy soils do a really good job. Clay soils do an even better job of holding soil moisture, but are totally miserable at releasing it. Banded soils, where sands and silts (loam) form layers, act as water storage devices and tend to be our most productive soils.

Structure of the Damage Potential Model

A four layer, weighted model was created to assess the potential damage to forested stands in Minnesota due to the introduction and outbreak of the exotic gypsy moth, Lymantria dispar. In this analysis, the damage potential of the gypsy moth was identified relative to where defoliation is (1) likely to be heavy and sustained and (2) have the greatest chance of "contacting" additional environmental stress.

The damage potential model was created by combining four Minnesota data sets. These sets can be classified into two groups, biological and environmental, based on their general nature and dynamic state. Biological data sets consisted of land cover type and forest cover density. These describe the general vegetative cover of the land and are considered to be dynamic variables in the analysis, ie. they can change with time. The Environmental variables of Soil Type and Environmental Stress, measured as Evapotranspirational Shortfall (EVTS), are generally considered to be fixed save major catastrophic events of geology or climate.

The analysis model was created by combining the two biological factors into a sub-model and the two environmental factors into a second sub-model each with separate weighting factors. As a result of this weighting, each sub-model contained scale values of relative damage potential. Operating at the landscape level, these models evaluate the relative damage potential for locations based on the individual factors. In a second phase, the sub-models were added together to derive a damage potential value for the entire state. While no specific research has been applied in the overall model, effectively all components are known to affect damage potential. The structure and weighting of the model was derived with this considerable body of research in mind.

Biological Component: Forest Cover Type

Forest cover type for all lands within the state is surprisingly elusive. The only available statewide cover type information was derived from a 1990 USGS study which used AVHRR (weather) satellite data to generate a national cover type map. This data was minorly adjusted based on our current knowledge of ECS, FIA, and land cover into the modified AVHRR Land Cover layer found in the MGC100 data set in EPIC (see attached map).

This layer was given the heaviest weighting of all model elements at 60%. A maximum of 60 rating units (points) were assigned to areas that had a high proportion of aspen or oak. Nonpreferred hosts were assigned 5 or fewer units.

Biological Component: Forest Cover Type Density

Forest cover type density is derived from another AVHRR study. Forest density was considered to be useful in appraising damage potential since areas with fewer trees have fewer trees to damage. With a weighting of only 5%, a maximum of 5 rating units could be assigned to a given area. The areas with the highest forest density are almost exclusively located in the Laurentian Forest Province in the northeastern part of the state.

Environmental Component: Soil Type

Soil data from the Minnesota Soil Atlas has been available in automated, fully registered format for many years. Currently part of EPIC's MGC100 data set, it is easily the best statewide soil information. Scoring soils was based on a general understanding of soil to site quality relationships. Soil quality was assigned a 20% weighting in the model with a maximum score of 20 being assigned to sandy soils. In effect this component estimates a soil unit's LACK of productive capacity. Better quality soil units received lower scores.

Environmental Component: Climatic Stress

Climatic stress, as measured by the evapotranspiration shortfall (summer rainfall minus water needed by the vegetation) has been understood for several years in Minnesota. It has been used in various work ranging from windbreak management to Ecological Classification Systems work. It is effectively a measure of the stress created by our "normal" shortfall of growing season precipitation. This "shortfall" fails to provide an optimum photosynthetic environment even in normal years. It is particularly important in Minnesota in understanding the west to east, prairie to forest transition that occurs in Minnesota.

The Climatic Stress component was given a weighting of 15%. While important in predicting general stand vigor over time, it is probably somewhat less important than soil, which tends to "buffer" its effect. A maximum unit score of 15 was assigned to areas with a EVTS shortfall greater than 8". These areas show a clear tendency to occur in the western Prairie Province. An area of low EVTS shortfall (1-2") occurring near Lake Superior was given a minimal score of 1 effectively showing minimal stress caused by a lack of rainfall.

A Technical Note

All spatial modeling was accomplished with the EPIC GIS using the MGC100 data set. Two 2-way reclass models were applied to the primary thematic layers to generate the two sub-models of Biological and Environmental Components. These were additively combined using EPPL's EVLAUATE command. All component and sub-models are available in EPIC as MGC100 data layers. The final results were a run of damage potential values from a low of 7 to a high of 96. These values were simplified to the three basic class illustrated on the attached map.


The attached color map is a first in two ways. First, it is the first full color map page in this newsletter. Perhaps the sign of things to come. Second, it graphically shows the final result of the Damage Potential model simplified into the three classes we know so well, low, medium, and high. Several things are immediately apparent. First, most areas of the state can expect to see damage caused by gypsy moth to some degree. Two, there are significant areas of high damage potential in the northern parts of the state. This is largely due to the large expanses of highly-favored aspen in these areas.

How do the forests in your area fare? The list below is the average danger class by county. The numeric score for each county is based on Minimal = 1, Moderate = 2, and High = 3. The counties are ordered from greatest to least damage potential. For example, Winona County has the highest average score of 2.09, indicating that on average we expect somewhat more than moderate damage. Hint: This is not a nice list on which to score high. We will be using this damage prediction model for various purposes over the next few years as we prepare for establishment of the gypsy moth. By using it, we can focus our attention and efforts into those areas of the state most likely to be damaged by gypsy moth defoliation. We expect gypsy moth to be everywhere, but damage to be concentrated in more specific areas.


As the result of this analysis, we are able to predict that the introduction of gypsy moth into Minnesota's forest ecosystems is likely to have a major impact on the nature of the forest cover. If the pattern of damage holds from other states, this change will occur over a period of years, likely decades, as gypsy moth outbreaks wax and wane. Damage will likely occur in individual stands as mortality of preferred host species, oak and aspen, leads to stands with a higher component of nonpreferred species. Mortality will be highest in intermediate and suppressed trees and be worse on sandy or droughty soils. It is likely to accelerate dramatically in areas where drought and defoliation due to forest tent caterpillar occurs.

Damage will concentrate in the southeast ( DNR Areas 531 and 533). Regions 2 and 3 will also feel a significant sting in several areas (Areas 221 and 222 in Region 2 and Areas 323 and 351 in Region 3). Regions 1 and 4 will be relatively little damaged. Region 6 will suffer some damage but will likely bear the brunt of the "nuisance damage factor" due to high public contact in the urban forest.

1 WINONA 2.09
3 ITASCA 1.84
4 KANABEC 1.72
5 WABASHA 1.69
7 CASS 1.63
8 CROW WING 1.61
10 PINE 1.58
11 TODD 1.58
13 HUBBARD 1.49
14 ANOKA 1.45
16 STEARNS 1.36
18 CHISAGO 1.31
19 BECKER 1.28
20 MILLE LACS 1.25
21 LAKE 1.21
22 OLMSTED 1.18
23 CARVER 1.17
24 AITKIN 1.16
25 COOK 1.14
26 BENTON 1.13
27 ST. LOUIS 1.08
28 WADENA 1.06
29 ISANTI 1.03
30 SCOTT 1.01
32 BELTRAMI 0.96
33 WRIGHT 0.94
35 GOODHUE 0.91
37 HENNEPIN 0.90
38 DOUGLAS 0.83
39 RICE 0.79
40 MAHNOMEN 0.74
42 POPE 0.63
43 RAMSEY 0.62
44 LE SUEUR 0.62
46 LINCOLN 0.49
47 DAKOTA 0.48
48 MEEKER0.41
49 ROSEAU 0.36
50 MCLEOD 0.35
51 STEELE 0.28
52 SIBLEY 0.27
53 NICOLLET 0.24
54 WASECA 0.23
55 DODGE 0.20
57 CHIPPEWA 0.17
58 SWIFT 0.17
59 FREEBORN 0.14
60 BROWN 0.13
63 MOWER 0.12
64 LYON 0.12
65 KITTSON 0.12
66 BLUE EARTH 0.11
67 MARSHALL 0.09
68 RENVILLE 0.08
69 REDWOOD 0.07
70 BIG STONE 0.07
72 CLAY 0.05
73 STEVENS 0.04
74 POLK 0.04
75 MURRAY 0.04
76 JACKSON 0.03
78 NOBLES 0.02
79 NORMAN 0.02
80 ROCK 0.02
81 MARTIN 0.02
82 GRANT 0.02
83 WATONWAN 0.02
84 WILKIN 0.02
85 RED LAKE 0.01
87 TRAVERSE 0.00