What determined in whether a spot was prairie or woodland?
Another way to ask, annoyingly to some (see Endnote 1): What determined "the most advanced stages of succession” for our prairies and forests for any given place.
Southern Wisconsin, northern Illinois, and much of adjacent states, when first entering recorded EuroAmerican history, consisted of patches of prairie and timber (a word often used then for savanna and woodland). In some areas it was said that every 80-acre farm had some prairie and some timber.
Henry Allan Gleason and other smart observers puzzled over the seemingly random nature of these patches in some areas. Gleason noticed prairie sites where the evidence suggested they once were woods - and visa versa. While it’s true (and reassuring to the orderly mind) that prairie occurred predominantly wherever the land was level (or on southwest-facing slopes) and that timber was predominant wherever the land was hilly (especially on north-facing slopes), striking exceptions were common.
Fire tends to burn hotter and spread faster uphill than downhill. North slopes tend to be wetter and more likely wooded. Southwest slopes are the driest as they’re baked by the afternoon sun. Rivers may stop fires, so the east sides of rivers tended to be wooded. Wetlands sometimes stop fires; then again, with increased fuel loads, when they do burn they can burn very hot. Oaks (especially bur oak) tend to resist fire damage. But maple, the “climax” tree of long-unburned areas, tends to be wiped out when chance brings a hot fire through it.
And what other principles might apply? Might knowledge of them help conservation decisions be rescued from the “playing God” charge - by appealing to history?
One day it came to me, having watched brush patches and prairie patches move in response to controlled burns, that in this region the “climaxes” moved. Indeed there was order, but it was a different kind of order.
I had noticed as I managed little nature preserve areas that patches of trees and shrubs sometimes shaded out and killed off nearby prairie areas. But also, with fires predominantly from the west, vigorous prairie fires sometimes killed trees and shrubs to the east of the grassland.
What could have prevented this sort of thing from operating on the larger scale … especially where so many patches made such a complex mosaic? Also, what could have prevented one community type from just taking over?
My hypothesis came to be that the “climax” vegetation in some areas was the northeast edge of both prairie and timber patches. There was a “rock, paper, scissors” dynamic going on here. On average, given the prevailing westerly winds, a prairie would tend to burn into and replace timber on the east edge of the prairie (the west edge of the woods). Also, because here in the temperate zone the sun at midday shines not from straight above but from the south, patches of timber would tend to shade out and replace prairie on the north edges.
Oak Groves Move Northeast - Passing Two Hills
In this graphic, up is north, and the nested circles represent topo lines of hills.
Green represents oak woodland. White is open prairie. (See Endnote 2.)
Some hilly, north-facing slopes were prairie. And some flat, rich areas were timber. Most any piece of land in complex mosaic areas might alternate between prairie or timber. But perhaps the patches would move (at varying speeds, depending on topographics) generally northeast, irregularly, like giant amoebas. The southwest-facing slopes would last longest as prairie, but (except in extreme situations) they’d eventually be shaded by woody plants and become savanna or woodland for a while. Northeast-facing slopes (except in ravines and other extremes) would last longest in the woodland phase.
It's tempting to think that savannas might have been more stable, as their rich turfs had components of most all prairie and woodland species. But it would also be tempting to think that during periods of mild burning a savanna might evolve into woodland, and during periods of severe burns, pure prairie might win out.
Such processes may have made for the grassland and woodland biodiversity that we find today and which is vanishing so rapidly and completely in many areas from lack of good stewardship … and understanding.
Endnotes
1. Many wise people over the years have argued strenuously that “succession” is a word that is not useful today; it has been made scientifically confusing and misleading by misuse. The problem with “succession” (a fact that happens in nature, as keenly noticed by Henry Chandler Cowles) is that the concept was ossified wrongly by Frederick Clements. People began to believe that what happens without fire or other “disturbance” is true nature. It's “good” and leads to nature's most noble state, according to which most land in the Midwest would be maple forest.
Even as late as Curtis (The Vegetation of Wisconsin 1959), the perspective was that oak woodlands would naturally become maple. But our understanding of “nature” now includes “biodiversity.” Curtis himself described how a diverse mesic oak woodland (with hundreds of species of plants and thousands of species of animals) loses most of its species when shady maples take over. Most conservationists today would see that as an ecologically tragic loss of a rare remnant of nature – resulting from neglect or mismanagement.
On the other hand, if we’re not to use the word succession, we need other words to describe what happens in situations when, after some form of degradation followed by good stewardship, conservative species gradually displace most “weedy” species. Some people refer to prairies and savannas as examples of "fire climax."
A related question: how much attention should we pay to such “climax” states when making conservation decisions? Some people look back on the past for our model. And whether or not we seek to restore past states, knowing a site's former ecological state can certainly inform conservation decisions.
2. The graphic above is considerably less than half-baked and way simplified. How this dynamic would work in various areas would also depend on soils and many other features. Also, the panel representing "Future, Sometime" would only hold if we had fires as ferocious as past fires sometimes were.
3. It may be a minor question what the “original vegetation” of a site was. For millions of years, our grassland and woodland communities evolved under the influence of lightning-caused fires. Following the retreat of the most recent glacier, lightning fire is widely believed to have been largely replaced by human-set fires. But the species, relationships, and communities were largely those same millions-of-years-old ones.
Consider this thought experiment: A patch of prairie, in the absence of fire for a century or two, is invaded by the plants and animals of adjacent savanna and oak woodland. Many of the conservative prairie plants and animals die out, and many conservative savanna animals and plants become established in a mosaic of rich savanna and oak woodland communities. Then the adjacent former prairie and original savanna/woodland get destroyed. But the "patch of former prairie" becomes a nature preserve. Would it be best to cut down the trees to return this preserve to its original prairie state - if it loses its savanna/woodland biodiversity in the process and gains back no prairie biodiversity, because it no longer exists nearby? To me the answer in clearly no; save the inter-relationships among the oaks and the other savanna species; this is especially true as savanna biodiversity is even rarer than prairie biodiversity.
4. Timber patches vs Prairie patches. It would be interesting to use existing topographies and ask a computer to model how vegetation patterns might have changed over time. The graphic above shows a "prairie grove" moving through surrounding grasslands. But an isolated patch of prairie might move through timber somewhat similarly. In an area that had equal amounts of both, the amoeba-like movements would be endlessly blending into each other or swallowing each other. The principles followed would include:
- Prairie patches on flat ground would tend to expand into timber to the north and east, where fire would be hottest. They'd need to shrink on their south edges because of shade and from their west edges because fire would have no momentum there.
- Timber patches on flat ground would head (and last longest) to the north because of shade and, to some degree, to the east because fire in frequently burned woods is moderated by lack of fuel. And a grassland on the east edge of timber might get shaded out by young trees because fire would have little opportunity to build momentum and heat while burning through the woods.
- Prairie would spread fastest and last longest on south and southwest-facing slopes, in proportion to how steep the slopes were.
- Timber would spread fastest and last longest on north and east-facing slopes.
- Timber would also spread fastest and last longest when protected by bodies of water to the west and to a lesser extent to the south, because fires would tend to have to back into those places (giving the prevailing west or southwest winds during fire season); a backfire is much less destructive of timber than a head fire.
5. How fast did these patches travel? Once again, the answer would depend on soils, slope, aspect (the direction the slope faces), nearby water bodies, etc. But two additional elements are fun to think about. Bur oak is the principal tree in this dynamic because they are so much more resistant to fire than any other. The other tree species grow behind the protection of the bur oaks.
Bur acorns are big and heavy (It's hard to get established in a dense turf of grasses.) and thus not dispersed by birds or wind. They're planted by squirrels. Gray squirrels tend not to go more than 50 yards (?) from trees to cache their acorns. (Might fox squirrels have been different?) It takes many years for those oaks to grow big enough to make new habitat that the squirrels feel confident to venture out from.
Bur oaks live for 300 or 400 years - said to be this region's longest-lived tree. It might be hard for fire to dislodge mature oaks from an area. If so, would that suggest that an area colonized by bur oaks would comfortably remain as timber for three or four hundred years? Other tree species can in time grow where bur oak reduces prairie fire intensity, but bur oak cannot reproduce in the shade of most other tree species. Perhaps intense fire has a better chance to eat away at the edge of a grove once the other species have replaced the bur oaks.
6. The map below by Marlin Bowles and Jenny McBride shows the 1880s vegetation of six townships along the north edge of Cook County. Yellow is prairie. Green is timber. Blue is wet.
Deer Grove is the large woods to the northwest. Busse Woods is the large woods in the south middle township, northeast of a wetland.
Timber is on the east side of the DesPlaines river, which spans this map from north to south on the east.
The Somme preserves are in the upper right (northwest) corner. Somme Prairie is west of the West Fork of the North Branch of the Chicago River. Notice that the prairie crosses the river there. The savanna and woodland are on the poorer soils of the moraine.
And what about the many smaller, isolated prairie groves? Isn't it fun to imagine what was going on with them?!
7. The maps of McHenry and Lake Counties, Illinois, shown below were suggested by Don Osmond.
The striking difference between McHenry (left) and Lake (right) probably reflects differences in what the map-makers wanted to bother with more than differences on the ground. That is, the Lake County map seems more detailed.
Don also supplied a letter from 1835 in which a settler (C. Fletcher) described the prairie southeast of Wauconda with these words: “Traveling on them you are out of sight of land, as it is called here, that is timber. Then a grove of timber that runs along by some stream a mile or two wide, and then prairie as far as you can see except small groves of one to five acres.”
Map Key
beige = prairie
pale green = savanna
dark green = woodland or forest
blue = water
On the above map in Lake County (right), there are a dozen small prairie openings in the oak savanna and about that many isolated denser groves surrounded by prairie or savanna. But these maps were extrapolated from the original Public Land Survey which mostly mapped features that crossed survey lines, so there may have been many more smaller prairies or groves. McHenry County (left) looks very different, but that may just mean a less detailed mapping.
For the letter by C. Fletcher, see here.
8. John Curtis, author of the Vegetation of Wisconsin (1959), is one of the people who has wondered about these questions. In a somewhat different context, on page 304 he wrote:
Thus it appears probable that the mesic prairies moved eastward by a series of jumps, each following a catastrophic destruction of the pre-existing forest. If that forest had developed sufficiently close to a climax condition so that it would have eliminated the oaks, then a true prairie resulted, whereas if the forest contained any proportion of oak (other than red oak), then a brush prairie was likely.
9. When this "moving ecosystem edges" concept first filled my mind for a while, decades ago, I was thrilled with it. I imagined there might be here – if not a “Great Discovery” – at least something useful. Then I forgot about it. There are so many competing challenges in conservation, and life can get busy. But Will Overbeck recently remembered a conversation about all this and asked if I had a graphic for his Wild Things talk. So I resurrected it. Thanks, Will. Perhaps some people will find it interesting.
