Ely, Minnesota - Gateway to the BWCA Canoe Country

NATURAL DISTURBANCE AND VARIABILITY OF FORESTED ECOSYSTEMS

NATURAL DISTURBANCE AND VARIABILITY OF FORESTED ECOSYSTEMS IN
NORTHERN MINNESOTA

Lee E. Frelich
University of Minnesota
Department of Forest Resources
June 17, 1998

INTRODUCTION

Northern Minnesota Forests at a Glance
          The forests of northern Minnesota have been shaped by three primary factors: climate, unique soil and landform combinations, and disturbances. A given combination of disturbance regime (simply a description of the types, severities, and frequencies of disturbances that occur in a given region), soil type and climate generally leads to forest vegetation with a characteristic range of composition, age structure and physical structure. The current structure and composition of forests is a result of disturbances that have occurred during the last 2-3 centuries (Frelich and Reich 1995a,b). Therefore, we may safely ignore the large-scale climate changes that have occurred over the last 12,000 years in Minnesota, associated with the retreat of the last major glaciation, and take the 18th and 19th centuries to be representative of presettlement condition (i.e. condition prior to settlement by Europeans, generally during the late 1800s). Climate change for the last 1,000-2,000 years has been relatively slight, compared to the previous several thousand years (Davis 1981, Webb 1987). Therefore the forests have changed little during the last 1000 years, so that makes the 18th-19th Century condition a good, stable, benchmark with which to compare the modern forest (Swain 1973, Davis 1981, Webb et al. 1983). Currently, the forest is still in the process of adjusting to the switch from wind and fire to timber harvesting as the most common disturbance. How the forest will respond to this change will not be completely known for another century. However, we can at least compare current forest inventory data with presettlement forest data from the General Land Office surveys done during the 19th Century, and from historical reconstruction of remnant forests that were never logged.
          Looking at northern Minnesota with a broad-brush approach, five major forest types can be recognized, which are introduced here and then discussed in separate sections below. Three of these five are what ecologists call 'near-boreal forest', because they are literally near the true boreal forest that extends across Ontario about 50-100 miles north of the Minnesota-Canada border. Minnesota near-boreal forests have characteristics intermediate between the true boreal forests of Canada and the hardwood forests to the south. I recognize three types of near-boreal forest:
          (1) Jack Pine or Jack Pine Barrens-. Forests on poor-quality sites with extremely sandy or rocky soil, that support jack pine forests that may also have some black spruce, aspen, white cedar, white pine, and balsam fir. About 7% of all presettlement forests in Minnesota were of this type, which was concentrated in the Border Lakes, with a few scattered stands elsewhere in northern Minnesota. Both the acreage and percentage of forest in jack pine today has declined at least 50% compared to presettlement times (Table 1).
          (2) Spruce-fir-aspen-birch-. Forests on good quality sites with deep fine-textured soil that support a mixture of aspen, paper birch, white spruce, balsam fir and red maple. These forests were widespread in northern Minnesota, and occupied about 24% of all forested land in the state at the time of European settlement. Today, both the acreage and percentage of the forest occupied by this forest type has increased (Table 1). Nearly half of all forest in the state fits into this category, which includes second-growth aspen that originated after cutting/burning of forests that were other types in presettlement times.
          (3) Swamp conifers-. Forests occupied by the few species able to tolerate the saturated soils of peatlands--including black spruce, tamarack, and white cedar. Swamp conifers actually occupied more acreage in pre-European settlement times than any other forest type in the state--a huge 6.7 million acres and 25% of all forest. Thus, swamp conifers edged out the spruce-fir-aspen-birch type, which was at 6.4 million acres and 24% of all forest land (see above and Table 1). Today, the area occupied by swamp conifer forest is slightly more than half of what it was 100 years ago, although the percentage of the forest in this category is still similar. The vast majority of swamp conifer forest occupied the Red Lake peatlands, rather than areas now managed by the U.S. Forest Service. However, large patches of swamp conifers occur throughout northern Minnesota. A very important characteristic of swamp conifers is that 20-30% of all stands were (and are) unproductive--meaning that the growth rate of the trees is considered too slow for timber harvesting operations to be economically sustainable.
          Forests types 4 and 5 are more closely related to the Great Lakes forests that stretch across the northeastern U.S. and southern Ontario and Quebec, than to boreal forests:
          (4) Northern hardwood or maple basswood-. Forests dominated by sugar maple, along with yellow birch, basswood, green ash, red maple and red oak. These forests grew on sites that were, for northern Minnesota, relatively warm and with very good soils, including major stands in the Chippewa National Forest-Grand Rapids area, and along the North Shore. These stands of northern hardwoods on the Chippewa and Superior, are really the northern-most outposts of a forest type that mainly lies in east-Central Minnesota, northern Wisconsin and Michigan. About 20% of Minnesota's presettlement forests were northern hardwoods, compared with 9% today. The acreage also decreased from about 5.3 to 1.5 million acres (Table 1).
          (5) White and red pine-. Forests dominated by the two species that attain the largest size among all species in northern Minnesota: white and red pine. Major presettlement concentrations occurred on the Chippewa National Forest, on the North Shore, and in the Ely area. About 3.5 million acres, or 13% of Minnesota's original forests were dominated by red and white pine, compared with a half million acres and 3.0% of the forest today (Table 1). White pine has an especially broad range of tolerance for different site conditions, and was frequently mixed with all of the previously mentioned tree species. Red pine also has a broad range, although not quite to the extent of white pine. It is important to remember that most white pine trees do not occur within 'white pine' stands, but as scattered individuals within other forest types.

Disturbances and How They Shape the Forest
          The following background on the two major types of disturbance--wind and fire--will set the stage for detailed consideration of individual forest types below. Please note while reading these discussions that ecologist's estimates of average intervals between natural disturbances are necessarily rough. Fluctuations in disturbance frequency lasting a few decades at a time, related to climatic cycles, and a small number of observations of large disturbance are among the reasons. Often, estimates of frequency for low-severity disturbances, such as surface fires and partial windthrow, are quite accurate (plus or minus a decade), because a large number of observations exist. However, large severe disturbances cause the most dramatic transformation of the landscape (i.e. the top 5% of disturbances, in terms of severity of wind or fire, and size of the blowdown or burn), contribute 90% or more of the total forest damage. In the boreal forest, for example, the largest 3% of all fires burn 97% of the landscape (Johnson 1992). Not many observations exist of these large landscape-transforming events. Therefore confidence in the disturbance interval estimates is low for extreme events (plus or minus 50%).
          Wind as a disturbance. All forest types in northern Minnesota have blowdowns ranging from 10-year average intervals between storms that cause a few scattered treefalls, to 300-year intervals for storms that remove 30-40% of the canopy trees, to 2000-year average intervals for storms that totally flatten the canopy of a forest stand (Simiu et al. 1979, Changery 1982, Canham and Loucks 1984, Frelich and Lorimer 1991a). These intervals get shorter as one moves southwest, so that intervals between total canopy blowdowns could be as short as 1000 years at the prairie-forest border.
          Tree size, type of anchorage provided by the roots, species, and degree of rot in the trunk determine the chance of treefall. Large trees and old growth are more susceptible to blowdown than young trees and even-aged forest stands (Frelich and Lorimer 1991b). Windstorms are caused by thunderstorm downbursts--large parcels of air that descend from the bottom of the thunderhead, going almost straight down until they hit the ground, where they splatter in all directions like a glass of spilled water. Straight line winds are another common type of thunderstorm wind, which acts like a breaking wave at the gustfront at the front of the storm. Wind speeds reach as high as 150 mph in either of these two types of winds (Fujita 1978), and winds of that speed will completely flatten the forest canopy, although many small seedlings and saplings on the forest floor and understory will survive (Glitzenstein et al. 1986, Webb 1988, Frelich and Lorimer 1991a). Tornadoes also cause forest damage, but these are very rare in northern Minnesota and the tornadoes have seldom reached the F3-F4 level with 200 mph winds, such as the one that devastated St. Peter in southern Minnesota (Thom 1963). Individual blowdowns severe enough to remove the previous canopy and create a young even-aged stand range from an acre or less to 9000 acres, with a mean of 230 acres (Canham and Loucks 1984). Larger areas of blown down forest rarely occur, but may result when a thunderstorm produces multiple downbursts, episodes of straight line wind and/or tornadoes as it moves across the state (Fujita 1978).
          Please note that a large majority of forest-damaging severe thunderstorms and tornadoes have windspeeds in the F1 category, 74-112 mph (Fujita 1978). These storms cause only partial damage to the forest canopy, principally removing large, old and weak trees. These storms occur frequently (on the order of once every 10-20 years at a given location), and create many canopy gaps of the size occupied by 1-3 canopy trees. It is best to think of these storms as causing minor damage over a large area (i.e. a large disturbance with low intensity), since one storm may affect thousands of square miles. Typically, these minor treefalls affect 5-7% of the canopy in older forests each decade (Frelich and Lorimer 1991a).
          Accumulations of heavy wet snow accompanied by winds of 40-60 mph may also occur in November and April in northern Minnesota, and these may cause loss of a small percentage of canopy trees over a large region, such as northeastern Minnesota. There may be a few areas 1-10 acres in size where most of the canopy is taken down in such a storm, especially in mature, tall-thin jack pine or black spruce (Frelich and Reich 1995a). Crown damage, rather than trunk breakage or uprooting, is common in white pine from these storms. Ice storms do not cause much damage in northern Minnesota, since the main path of these storms takes them through extreme southeastern Minnesota, and regions to the south of that.
          Fire as a disturbance. Fires in Minnesota were historically very large in size--those fires that burned most of the landscape were often hundreds of thousands of acres in size (Heinselman 1973). However, there was large variability in the severity of fires and their frequency across the forested region of the state. Near the prairie-forest border, fires were very frequent, having occurred once or twice a decade. These areas were covered with savanna-like vegetation with scattered groves of trees embedded within grasslands. The general trend as one moved towards the northeast, was for fire frequency to decrease, but for fire severity to increase. Thus, the forest mixture of conifers and birch and aspen that covered much of northern Minnesota, had very severe canopy-killing fires with average fire intervals ranging from 50-200 years among forest types (Heinselman 1963, 1970, 1981). Fire was generally rare in northern hardwood stands with sugar maple, and the fires did occur there were generally smaller in size with much smaller flame lengths than in coniferous forests (Stearns 1951, Frelich and Lorimer 1991a). The different structure of the sugar maple and coniferous forests explains the vast difference in frequency and intensity between them. Spectacular crown fires require relatively high density of foliage and low water content in the foliage--both characteristic of coniferous forests--in order to get started and spread across the landscape (Johnson 1992, Van Wager 1977). Sugar maple forests also occur on soils with good water-holding capacity that minimize the effect of droughts, and on sites that have major natural fire breaks such as lakes and rivers on the south and west sides.
          Two interesting facts about fires in Minnesota's forests are worth noting before proceeding to detailed description of natural disturbance variability in specific forest types. First, within a given forest type, fires occurred at random with respect to stand age. That means that forest stands of all ages were equally likely to burn, so that some stands were burned twice within a short period of time, while other stands were skipped, purely by chance for centuries (Johnson 1992). Thus, there were a lot of very young even-aged post-fire stands in northern Minnesota, as well as old-growth stands (Heinselman 1973). This random-fire regime lead to an age distribution of stands across the landscape with the highest number of stands in the young age classes, and a steady decline in the number of stands as age increased. This smooth decline with age is known as a negative exponential distribution, shown in Figure 1. The reason that stands of all ages were equally likely to burn, is that fire weather--the degree of drought in preceding 1-2 months before the fire, and the temperature, humidity and windspeed on the day of the fire, were far more important in determining the likelihood of ignition and intensity of fire within closed canopy northern coniferous forests, than the total amount of fuel in the forest (Johnson 1992).
          The second interesting fact about fire is that the size, severity and frequency of fires has changed dramatically since settlement by Europeans (Heinselman 1981). Conversion of much of the landscape from conifers to aspen and birch (which do not carry severe crown fires very well), to cities, farms, and resorts, has disrupted the flow of fires across the landscape. These changes in land use alone could account for most of the reduction in the role of fire--especially large fires that burned most of the landscape in presettlement times--over the last century. In addition, direct suppression of fires and climate change have probably also contributed to the reduction in number of large fires. Large fires are still possible, and will eventually occur in northern Minnesota. However, the annual probability of a catastrophic fire--similar in size and severity to the 1988 Yellowstone fires-- is somewhat lower at this point than in the past.

MAJOR FOREST ECOSYSTEMS OF NORTHERN MINNESOTA AND THEIR DISTURBANCE REGIMES

Near-Boreal Jack Pine Forests on Rocky Lands in the Border Lakes
          Drought-prone, shallow rocky soils and droughts once or twice each decade favored a very severe fire regime. Until the early 20th century, the fire regime was very similar to that of Yellowstone National Park, and crown fires with flame lengths of 50-100 feet swept across large tracts of forest ranging from a few thousand to several hundred thousand acres in size. The one major difference between the historical fire regime in Yellowstone and that of Minnesota's Border Lakes was in the frequency of fires. Such fires occurred 4-5 times as often in Minnesota (50-year average interval between fires) as compared to Yellowstone (200-year average interval) (Van Wagner 1978, Romme 1982). In Minnesota, only a few tree species are able to tolerate such poor quality soil and severe disturbances, and these include jack pine and black spruce. Both species have major adaptations to fire in that a storehouse of seeds exists in cones that remain closed for many years, until they are opened by the heat of a fire; these are known as serotinous cones. After a crown fire, jack pine and black spruce seeds rain down on the forest floor by the millions, and a new even-aged forest begins to rise almost immediately. Forests dominated by jack pine and black spruce are very flammable, because they have relatively low moisture levels in the foliage, the density of foliage is high, and they grow on excessively well drained droughty sites. Most of these stands were young and even-aged, but they also had a huge number and volume of standing dead snags, because most of the trees from the pre-fire stand were killed, but the trunks were mostly not consumed by fires (Heinselamn 1973).
          In most cases there was little succession (replacement of one group of tree species by another) after fires, because the same species regenerated after fires as were previously present (Heinselman 1973, Frelich and Reich 1995a,b). However, two circumstances led to episodes of succession (Figure 2A). Some stands, purely by chance, were missed by fires for periods of up to 250 years. In these stands windstorms, insect infestation, disease, and general old age slowly chip away at the jack pine canopy, forming gaps that are invaded by balsam fir, white cedar and paper birch between 150 and 200 years after fire (Frelich and Reich 1995a). Black spruce maintains a significant presence in these older stands. Because it is shade tolerant and has the ability to survive on the forest floor for several decades while waiting for the tree above to die, it reproduces in the absence of fire through rooting of the lower branches when they contact the soil, and although many cones remain closed until a fire occurs, a few cones open each year and shed seeds on the forest floor (Heinselman 1973). Many people think of paper birch as a 'pioneer species'. However, it invades these stands as they enter the old-growth stage. The reason is that rotten logs and moss of the forest floor provide an ideal seedbed for germination and survival of paper birch seedlings, and gaps in the canopy allow significant amounts of light to the forest floor, at least in a few places at any given time. The overall structure of these old-growth forests is a mixture of small groves (0.1 acre or less) of fir, cedar, black spruce and paper birch, of all ages. A thick moss layer is usually present along with tons of rotten logs on every acre (Frelich and Reich 1995a).
          Spruce budworm is common in these older stands, and infests mostly balsam fir. Most fir trees do not live beyond 40 years for this reason, even though the species is capable of living 200 years. There are virtually always young fir in the forest understory that are not killed, and when they mature they are infested themselves (Heinselman 1981). Fires in these old-growth forests probably favored a mixture of paper birch, aspen and black spruce after the fire. The reasons are that there would be no jack pine seed source, the cedar and fir (including seeds) would be killed by the fire, the serotinous cones of black spruce would shed seeds after the fire, paper birch could resprout from the stump, and aspen could invade from afar due to its long-distance seed-dispersal ability.
          The second circumstance leading to successional replacement of jack pine and black spruce, is the occurrence of two major fires within a 10-15 year period (Figure 2A). If this happens, then most of the conifers will not be old enough to bear seed. The burned stand is therefore open to invasion by other species from the outside (Heinselman 1973). Quaking aspen is the most successful species under these conditions, since its seeds are very light and can enter the burn from a mile or more away. This is the explanation for the existence of large areas of so-called 'virgin' aspen stands that existed at the time of settlement, including an aspen forest >100,000 acres in size in the eastern Boundary Waters Canoe Area Wilderness (BWCAW), that burned in 1864 and 1875.

          One may ask the question, how often do the two circumstances allowing succession from jack pine and black spruce to some other forest type occur? This can be answered by observing that the chance of fire in conifer-dominated boreal forest is approximately equal regardless of stand age. A 10-year-old stand is just as likely to burn as a 100 or 200-year old stand. Note that with a 50-year average interval between fires, a substantial proportion--approximately 10%--of the landscape would burn before age 10-15 and be replaced by aspen. One property of the negative exponential age distribution (Figure 1), is that approximately 36% of all stands survive for one full average fire interval, and 36% of the remaining stands survive one additional mean fire interval, and so on. Thus, about 36% of all stands would survive more than 50 years, 13% as long as 100 years, 3% for 150 years, and 1% for 200 years. The random timing of fires was/is a double edged sword for jack pine. Most of the time, fires occurred when stands were between ages of 20-120, where jack pine can replace itself via massive post-fire seeding. However, some fires occurred too soon before jack pine was old enough to have seeds, and some fires occurred after too long, when only a few old jack pine remained in the stand (Frelich and Reich 1995a).
          To summarize the near-boreal forest on rocky lands, we can say that there is a complex successional system that includes jack pine, aspen, black spruce, balsam fir, paper birch, and white cedar (Figure 2A). These species replace each other in response to timing and severity of fires. In presettlement times, most of the forest stands were young and even-aged with heavy dominance by jack pine and black spruce or aspen. A few stands attained old-growth status with dominance by black spruce, balsam fir, paper birch and white cedar. Because of the distribution of stand ages, in the 1800s there were both more young stands and more old stands than there are on a modern commercial forest landscape, such as that just outside the BWCAW.
          Within the modern BWCAW, however, there are probably more old stands now than in the last few centuries, due to the decrease in fire frequency since the early 1900s (Heinselman 1973). Fires have been less frequent due to climate change, fire exclusion caused by changing land use around the BWCAW, and fire suppression. Many people underestimate the effect that fire exclusion has had at the southern margin of the boreal forest. Most of the large historical fires in the BWCAW started outside the area, and burned into it coming from the southwest. Today, these fires rarely get going. Thus, many fires that would have occurred during the past century have been prevented, or excluded from occurring at all.

Near-Boreal Spruce-Fir-Aspen-Birch Forests
          Areas in northern Minnesota with soils that are deeper, or more fine textured than soils in the jack pine forest, allowed development of dense forests of mixed aspen, birch, balsam fir, white spruce, and red maple. Good soils favor paper birch and aspen immediately after fire, rather than the conifers. Therefore, these forests were undergoing constant succession (Figure 2B). Fires would wipe out fir and spruce, replacing them with birch and aspen. In some areas where the landscape is dotted with small wetlands--especially parts of Chippewa National Forest--tamarack also was an important component of post-fire forests. The small wetlands shielded small groves of tamarack from damage by fires, so that tamarack seeds would be widely available after a fire. Since tamarack is intolerant of shade, it was capable of growing on burnt lands along with paper birch and aspen. Between fires, fir, spruce and red maple would invade from intact forest surrounding the burnt areas, attaining dominance by 80-100 years after fire, and remaining until removed by the next fire. The fir and spruce were often high in density, and able to propagate crown fire as well as jack pine forest. But because they were on less droughty soils, they did not burn as often. Average intervals between stand-killing fires were about 100 years (Heinselman 1981).
          Beaver in this and the other forest systems, can cause local pockets of tree mortality (0.1-several acres). The areas are flooded along streams, when the beaver have consumed all of the aspen in the vicinity, they move on, the dam eventually fails, and the water level returns to the normal stream level, and aspen reinvades the formerly flooded forest area. Beaver may return when the aspen is mature and repeat the cycle. The time period for this cycle depends on the population density of beaver, which fluctuates in response to hunting by wolves, trapping, severe winters, and diseases.
          These forests formed a successional system of aspen-paper birch, spruce, fir, and red maple. Fires were just as large as in the jack pine forest, but not as frequent (Heinselman 1981). About 2/3 of all stands would be less than 100 years old, and a few stands (10-15%) would have survived 200 years or more before burning. Therefore, the landscape was a mosaic with young even-aged stands dominated by aspen and birch, middle-aged stands of mature birch and aspen with conifers in the understory, and some older stands almost completely dominated by conifers but with some red maple. Fire acted to set back succession. In the modern commercial-forest landscape, relatively few stands go for 100 years or more before harvesting. At the same time, harvests are not random with respect to stand age, so that young stands are not harvested. Like the jack pine forest, the presettlement spruce-fir-aspen-birch forest had both more young stands and more old stands than the current commercial-forest landscape, although within reserved forests, there are more old stands today, due to fire exclusion, than in the past.
          Infestations of spruce budworm in stands greater than 80 years old worked in a similar manner to that described above for old jack pine forests that succeed to balsam fir. In presettlement times, fire systematically removed fir from large pieces of the landscape (Frelich and Reich 1995a). Fir was constantly reinvading stands after fire, but was unable to form contiguous stands over large areas. This process was especially important in the far northeastern part of the state. The removal of fir no longer occurs, because logging is the major disturbance, and small fir seedlings are not usually killed by logging (especially winter logging), as they would have been during presettlement fires. Therefore, fir is now more abundant and contiguous in distribution across the landscape. Spruce budworm finds it easier to spread from one fir grove to the next, leading to prolonged and serious infestations (Heinselman 1973).

Swamp Conifer Forests
          There are three major types of swamp conifer forest in northern Minnesota that intergrade with one another: black spruce, which has a tendency to occupy the sites with sphagnum peat, very acidic and very low nutrient supply. Second, are tamarack forests on peats of intermediate pH and nutrient supply, and thirdly, white cedar, on sites of relatively high pH and nutrient levels. Some of the lowland forests (especially cedar forests) are on muck over a mineral soil layer, rather than deep peat (Heinselman 1963, 1970). The locations of each of these forests types are determined by the flow of water across the landscape, which is influenced by small variations in the topography. Areas that are cut off from ground water, and get most of their water from rainfall, are the most acidic. These are the raised peatlands that hold onto rain water like a sponge, occupying most of the 'Big bog' near Red Lake. Other areas with slowly moving ground water have minerals which neutralize the acid produced by sphagnum mosses, leading to dominance by tamarack or cedar.
          The lowland conifer forests did support canopy-killing fires, but only half as often as uplands, so that the average interval between fires was about 150-200 years (Heinselman 1981). Burning usually did not cause much change in tree composition (Figure 2C). The main change that occurs in upland areas after fire is invasion of paper birch and aspen, but these two species do not grow well on peatlands. Succession from one conifer to another on peatlands is determined more by long-term changes (over centuries and millennia) in the thickness of peat, changes in drainage patterns, and climate change than by fires (Heinselman 1963).
          The one disturbance acting at short time intervals that can cause major change, is flooding by beaver. Beaver can raise the water table 1-2 feet, and the resulting flood can kill large acreages of forest on the northwestern peatlands, because the landscape is almost flat. This is especially true for areas with muck soils or thin layers of peat that are not buoyant enough to float the forest at the prevailing water level. The trees die due to lack of oxygen for the roots, which causes the root system to shut down, leading to drying out of the foliage in the upper part of the tree. Black spruce and tamarack are the most tolerant of saturated soil among the conifers. Stands killed by flooding are replaced by lowland shrubs or marshes until the water level goes back down. If the flood is short in duration (a few weeks or months), then forest may be reestablished in the first few years after the flood. The frequency with which beaver killed lowland conifer forest is not known precisely, but it is likely that certain tracts of forest situated at the edge of the peatlands where the beaver had access to aspen, were killed repeatedly, while the vast majority of the lowland forests were rarely or never affected.
          Because fires were relatively infrequent in the peatlands, many stands reached old ages. Generally, even aged structure of stands held together for 150-200 years, and about 2/3 of all stands were even-aged (Heinselman 1963, Groot and Horton 1994). Stands over 200 years old become multi-aged, and occupied the remaining 1/3 of the landscape, with a few stands surviving as long as 400-600 years before burning.

Red and White Pine Forests on Rocky or Sandy Lands
          Near the borders of lakes, on peninsulas, and on islands within the Border Lakes, the severity of fires was often lower than on the contiguous uplands dominated by jack pine and aspen or black spruce. The lake edges are often so rocky that fuel for fires was not contiguous, the humidity was higher, the vegetation in some cases could tap into the water table and remain wet even during droughts, and finally, in some areas, a concentration of large lakes served as fire breaks that interfered with the free movement of crown fires across the landscape. On moraines and sandy land types in the Chippewa National Forest, there were similar situations, where lakes or extensive lowlands provided a fire break. In such areas, crown fires occurred much less often (150-200 years average interval) than in the jack pine forest, and many fires dropped to the ground to become surface fires. Such a disturbance regime, with infrequent crown fires and frequent surface fires, favors the development of red and white pine forests (Heinselman 1973, Frelich and Reich 1995b). The area around Ely was a good example. Basswood Lake and other large lakes prevented the spread of large crown fires, and this was one of the large stands of white and red pine that was removed early in Minnesota's logging era.
          There are certain similarities and dissimilarities between these forests and the previously described jack pine forest. Like the jack pine forest, fires were equally likely in stands of all ages. Although the mean age at time of stand-killing fire is 150-200 years, some stands (5-15%) go for 400-600 years without stand-killing fires (Heinselman 1973). White and red pine can be maintained for 600 years or more, as long as surface fires occur regularly (Figure 2D). Without the surface fires, dominance by white and red pine will be lost through succession to shade-tolerant species (Heinselman 1973, 1981, Frelich and Reich 1995a). In the Border Lakes, the replacement species were mainly black spruce, white cedar and balsam fir, whereas farther to the southwest in Chippewa National Forest, red maple, red oak, with some balsam fir were the most common trees to replace the pines. Windstorms hasten succession to spruce and fir, or red maple and red oak because the much taller pines are susceptible to blowdown and lightning strikes. With frequent crown fire--less than 80-year interval between fires--dominance by red and white pine will also be lost, because these species take at least 80 years to establish a canopy that bears high numbers of seeds to replace the trees after fires.
          The dissimilarities with the jack pine forest include a completely different life-history strategy for red and white pine to perpetuate themselves, and that these forests underwent constant succession. White and red pine do not have serotinous cones, and thus heavy seed rain cannot be assured after the major crown fires, and they cannot resprout like aspen and birch do after having the above-ground part of the tree killed by fire. Red and white pine can only survive fire as a mature adult, meaning that they can only survive surface fire, during which the peak temperature inside the bark does not get high enough to kills living cells, and the foliage is not killed. Saplings are usually killed by fire, since their bark is too thin to provide that much insulation, and the tree can be girdled by fire. Individual mature trees that survive the fire are the source of seed.
          If there is a major crown fire, most of the large pines are killed, and paper birch invades rapidly and grows much faster than any pine seedlings present. The few large pines that survived after crown fire, as well as pines outside the burned area provide the seed source for a gradual reinvasion of the stand underneath the young birch canopy. This wave of reinvasion by white pine typically takes 1-4 decades (Frelich 1992). As the stand matures, white pine break through the birch and begin to dominate the canopy by age 100. Minor surface fires remove shrubs, invading spruce, fir, maple and oak, and thick duff, allowing for establishment of new cohorts of young pines. If surface fires continue to occur, the stands become multi-aged, usually with 2-4 main age groups (Frissell 1973, Heinselman 1973, Frelich 1992).
          In the rocky lands and sandy lands, white and red pine form stands in areas where fire has less than average presence than the surrounding landscape, due to protection from fire by the landscape and topographical setting. These pines cannot perpetuate themselves without surface fire, and they cannot get along with frequent crown fire either. The fire regime and topographic setting has to be just right, which is why red and white pine occupied a relatively small proportion of the Minnesota's forest landscape (about 13%) even in presettlement times (Frelich 1995). The timing of crown fires and surface fires determined whether these forests underwent succession from paper birch to white and red pine, and then on to more shade-tolerant species, or whether they remained multi-aged pine stands for several centuries, or whether they reverted to the jack pine or aspen forest type.
          Today, most red and white pine stands have dense understories of fir, spruce, red maple, or shrubs. If fire exclusion continues, these species will eventually replace pine. The suppression of surface fires has been very effective, and the result is to set the red and white pine forests onto new successional pathways that may not include much naturally regenerated red and white pine in the future. The buildup of fuels directly beneath pine crowns--especially balsam fir--may also lead to fires of increasing severity in the future. These severe fires will be hard to control and could kill the remaining large red and white pine.
          Another major difference between presettlement red and white pine forests and modern commercial forest, is in age strcuture. More than half of all stands would have attained old-growth status (using the Minnesota DNR definition, also proposed to be used by the U.S. Forest Service), of >120 years old, under the natural disturbance regime, whereas under the 20th-century harvesting regime, only 2-3% of all stands have attained an age of >120 years (Frelich 1995).

Northern Hardwoods Forest
          Forests dominated by sugar maple, often with some yellow birch, red maple, red oak, and basswood on sites with deep loamy soil, are called northern hardwood. Unlike the other forest types in northern Minnesota, these forests had a relatively low incidence of fire, and grew in areas with (for northern Minnesota) relatively warm climates. Thus, only small scatterred tracts of this forest type occurred on the Superior National Forest, while much larger tracts occurred on the Chippewa National Forest. The disturbance regime consisted mainly of windstorms that occurred every few decades, which removed senescent, weak and/or hollow trees from the canopy (Frelich and Lorimer 1991a). The resulting gaps were colonized by either pre-existing seedlings of shade-tolerant species like sugar maple and basswoods, or by new seedlings of mid-tolerant species like yellow birch, red maple or red oak. A typical acre of forest had several hundred trees of all sizes and ages, with many gaps in different stages of recovery. An individual acre of forest could exist in this state, termed old-multi-aged state, for many centuries (Frelich and Lorimer 1991b). Disturbances that leveled the forest, resulting in establishment of a young even-aged stand, were rare. The rotation period for extreme windstorms that leveled the upper canopy of the forest was 1000-2000 years (Canham and Loucks 1984, Frelich and Lorimer 1991a). Typically, a seedling layer would be left nearly intact after such storms to regenerate the forest to a very similar composition as before the storm (Dunn et al. 1983). Thus, little succession occurred due to these disturbances caused by wind (Figure 2E). Eighty to 90% of the landscape would covered by old growth forest, with the remainder recovering from canopy-killing disturbance (Frelich and Lorimer 1991b).
          A major change in forest composition only occurred when the slash from an extreme windstorm burned (Figure 2E), leading to severe fire that killed the understory trees and seedbank. In this situation, paper birch was the main invader (Frelich and Reich 1995b). Succession could then lead to replacement of the paper birch by white pine, red maple or red oak (or a mixture of these). Thus, white pine maintained itself in areas that had a higher presence of fire than average across the landscape, just the opposite of the boreal forest, where white pine was confined to areas with less severe fire than the rest of the landscape. There would then be three alternative successional pathways for these mid-successional white pine-oak forests. If surface fires occurred, then the white pine and oak could be maintained for a few centuries. If a crown fire occurred, then paper birch would dominate the stand. If no additional fires occurred for several decades, then sugar maple becomes reestablished, and the chance of fire then goes way down, since maple fuels are not very flammable (Frelich 1992). Windstorms were (and are) the most common cause of loss of white pine within mixed pine-hardwood stands (Tester et al. 1997). In some stands, sugar maple may invade a paper birch forest directly, so that the mid-successional white pine-oak stage is skipped. The first two pathways were more likely to occur on somewhat dry sandy loams, while the latter two pathways were more likely to occur on moister loams and silt loams.
          Although these forests were heavily dominated by sugar maple most of the time, there was a significant presence of red oak, yellow birch, green ash, red maple and other mid-tolerant species. This mixed nature of the forest can be accounted for in two ways. First, canopy gaps created by windstorms were large, because trees in these old-growth forests were large. Some of the gaps, formed when a group of 2-4 canopy trees were toppled by wind, were up to 0.1 acre in size. This is large enough to support growth of mid-tolerant species. Secondly, the yellow birch and oak are favored by occasional surface fires, of very low severity, which merely burn the duff in spring or fall, scarring or killing a few of the overstory trees (Frelich and Lorimer 1991a). In areas towards the prairie-forest border, as well as on sandier soils, these surface fires are more common, and the abundance of oaks in the forest may be higher than at the northeastern edge of sugar maple's range.
          Lowland hardwood forests were also important, particularly on the Chippewa National Forest, where there are a lot of sites that are close to the water table, but not actually flooded. These forests are a bit too wet for sugar maple, but support mixed stands of black ash, green ash red maple, and yellow birch. White cedar is an occasional component in lowland hardwoods as well. These are very productive sites with large trees--not the stunted black ash on flooded sites that many associate with the term 'lowland hardwoods'. The disturbance dynamics of these lowland hardwood forests is similar to that of nearby sugar maple forests, except that susceptibility to windthrow is slightly greater, due to shallower rooting.
          The northern hardwood forests had the lowest presence of fire, the highest proportion of old growth, and the greatest stability over long time periods of any forest type in northern Minnesota. Although rare, surface fires were important for maintaining diversity of tree species in the hardwood forests, even if they only occur once every few centuries. For example, if surface fires occur with 500-600 year rotation, and a grove of red oak is established after most of the fires and survive for 300 years, then half of the landscape would have red oak mixed in with the sugar maple, leading to a much more diverse landscape than one almost completely dominated by sugar maple alone. Human settlement and management of the northern hardwood forests has caused some species composition change, through suppression of surface fires which formerly favored red oak and yellow birch, and has changed the age structure dramatically, since most stands now are even-aged and young, rather than multi-aged old growth (Frelich 1995).

Literature Cited

Canham, C.D. and O.L. Loucks. 1984. Catastrophic windthrow in the presettlement forests of
          Wisconsin. Ecology 65: 803-809.
Changery, M.J. 1982. Historical extreme winds for the United States: Great Lakes and adjacent
          regions. NUREG/CR-2890. National Climatic Center, NOAA, Washington, D.C. USA.
Davis, M.B. 1981. Quaternary history and the stability of forest communities. Pages 132-153 In: D.C.
          West, H.H. Shugart, and D.B. Botkin, Editors. Forest succession, concepts, and application.
          Springer-Verlag, New York, New York, USA.
Dunn, C.P., G.R. Guntenspergen, and J.R. Dorney. 1983. Catastrophic wind disturbance in an
          old-growth hemlock-hardwood forest, Wisconsin. Canadian Journal of Botany 61: 211-217.
Foster, D.R. 1983. The history and pattern of fire in the boreal forest of southeastern Labrador.
          Canadian Journal of Botany 61: 2459-2471.
Foster, D.R. 1988. Species and stand response to catastrophic wind in central New England, USA.
          Journal of Ecology 76: 135-151.
Frelich, L.E. 1992. The relationship of natural disturbances to white pine stand development. Pages
          27-37 In: R.A. Stine and M.J. Baughman, editors. White pine symposium proceedings: history,
          ecology, policy and management. Minnesota Extension Service, University of Minnesota, St.
          Paul, MN USA.
Frelich, L.E. 1995. Old forest in the lake states today and before European settlement. Natural Areas
          Journal 15: 157-167.
Frelich, L.E. and C.G. Lorimer. 1991a. Natural disturbance regimes in hemlock-hardwood forests of the
          Upper Great Lakes Region. Ecological Monographs 61: 145-164.
Frelich, L.E. and C.G. Lorimer. 1991b. A simulation of landscape-level stand dynamics in the northern
          hardwood region. Journal of Ecology 79: 223-233.
Frelich, L.E. and P.B. Reich. 1995a. Spatial patterns and succession in a Minnesota southern boreal
          forest. Ecological Monographs 65: 325-346.
Frelich, L.E. and P.B. Reich. 1995b. Neighborhood effects, disturbance and succession in forests of
          the western Great Lakes Region. Ecoscience 2: 148-158.
Frissell, S.S., Jr. 1973. The importance of fire as a natural ecological factor in Itasca State Park,
          Minnesota. Quaternary Research 3:397-407.
Fujita, T.T. 1978. Manual of downburst identification for project NIMROD. SMRP research paper 156.
          University of Chicago Press, Chicago, Illinois, USA.
Glitzenstein, J.S., P.A. Harcombe, and D.R. Streng. 1986. Disturbance, succession, and maintenance
          of species diversity in an East Texas forest. Ecological Monographs 56: 243-258.
Groot, A., and B.J. Horton. 1994. Age and size structure of natural and second-growth peatland Picea
          mariana stands. Canadian Journal of Forest Research 24:225-233.
Heinselman, M.L. 1961. Black spruce on the peatlands of former glacial Lake Agassiz and adjacent
          areas in Minnesota: a study of forest sites, bog processes, and bog types. Ph.D. Thesis,
          University of Minnesota.
Heinselman, M.L. 1963. Forest sites, bog processes, and peatland types in the glacial Lake Agassiz
          region, Minnesota. Ecological Monographs 33:327-374.
Heinselman, M.L. 1970. Landscape evolution, peatland types, and the invironment in the glacial Lake
          Agissiz Peatlands Natural Area, Minnesota. Ecological Monographs 40:235-261.
Heinselman, M.L. 1973. Fire in the virgin forests of the Boundary Waters Canoe Area, Minnesota.
          Quaternary Research 3: 329-382.
Heinselman, M.L. 1981. Fire intensity and frequency as factors in the distribution and structure of
          northern ecosystems. Pages 7-57 In: Fire regimes and ecosystem properties. USDA Forest
          Service, General Technical report WO-26.
Heinselman, M.L. 1996. The boundary waters wilderness ecosystem. University of Minnesota Press,
          Minneapolis, MN, USA.
Johnson, E.A. 1992. Fire and vegetation dynamics: studies from the North American boreal forest.
          Cambridge University Press, Cambridge, England.
Lorimer, C.G. 1977. The presettlement forest and natural disturbance cycle of northeastern Maine.
          Ecology 58:139-148.
Romme, W.L. 1982. Fire and landscape diversity in subalpine forests of Yellowstone National Park.
          Ecological Monographs 52: 199-221.
Simiu, E.M., J.Changery, and J.J. Filliben. 1979. Extreme windspeeds at 126 locations in the
          contiguous United States. United States Department of Commerce, National Bureau of
          Standards, Building Science Series 118.
Stearns, F.W. 1951. The composition of the sugar maple-hemlock-yellow birch association in northern
          Wisconsin. Ecology 32: 245-265.
Swain, A.M. 1973. A history of fire and vegetation in northeastern Minnesota as recorded in lake
          sediments. Quaternary Research 3: 383-396.
Tester, J.R., A.M. Starfield, and L.E. Frelich. 1997. Modeling for ecosystem management in
          Minnesota pine forests. Biological Conservation 80: 313-324.
Thom, H.C.S. 1963. Tornado probabilities. Monthly Weather Review 91:730-736.
Van Wagner, C.E. 1977. Conditions for the start and spread of crown fire. Canadian Journal of Forest
          Research 7: 23-34.
Van Wagner, C.E. 1978. Age-class distribution and the forest fire cycle. Canadian Journal of Forest
          Research 8: 220-227.
Webb, S.L. 1988. Windstorm damage and microsite colonization in two Minnesota forests. Canadian
          Journal of Forest Research. 18: 1186-1195.
Webb, T. III, E.J. Cushing, and H.E. Wright, Jr. 1983. Holocenc changes in the vegetation of the
          Midwest. Pages 142-165 In: Late-quaternary environments of the United States. Volume 2. The
          Holocene. University of Minnesota Press, Minneapolis, MN, USA.
Webb, T. III. 1987. The appearance and disappearance of major vegetational assemblages: long-term
          vegetational dynamics in eastern North America. Vegetatio 69: 177-187.
Whitney, G.G. 1986. Relation of Michigan's presettlement pine forests to substrate and disturbance
          history. Ecology 67: 1548-1559.

Table 1. Comparison of Minnesota forest acreage and percent, for presettlement (ca. 1850-1900) and today (ca. 1990). Adapted from Frelich (1995). 'Other' forest refers to oak-hickory and riverbottom forest, which mainly occurred in southern Minnesota. Acreages in millions, rounded to the nearest 0.1 million.

Forest Type 18th Century Today

Acreage % Acreage %

Jack Pine 1.9 7.0 0.6 3.5

Spruce-fir-birch 6.4 23.7 8.1 48.5

Swamp conifer 6.7 24.6 3.5 21.2

Red-white pine 3.5 12.9 0.5 3.0

Northern hardwoods 5.4 19.9 1.5 8.9

Other 3.2 11.9 2.5 14.9

Total 27.1 100.0 16.7 100.0

Figure 1. The negative exponential distribution of stand ages on landscapes where conifer forests burn at random with respect to stand age. The mean stand age is equal to the mean interval between fires (also sometimes called the fire cycle or natural rotation period). 63% of all stands are younger than the mean age, while 36% survive more than one fire cycle, 13% survive two fire cycles and 3% survive 3 fire cycles.

Figure 2. Idealized timelines showing episodes of disturbance and succession for 1000 years in a typical stand for each of the five major forest types.