This is interesting. In many grasslands forest islands are actually the result of fires. I have worked long in Kenya, where roundish forest islands mark often the previous extension of forests that were affected by savanna fires, or where small forest islands are a result of germination of tree species in termite mounds that subsequently protect them from fires. How much anthropogenic influence did the Chaco grasslands receive previously? In Peru we see that even dense "pristine" cloudforest areas or subalpine grasslands almost always show clear signs pf pre-columbian management (e.g. terrasses). Was there anthropogenic fire influence in the Chaco in pre-columbian times (I would assume so). From this perspective, could the forest islands not mark outposts of previos forest extension? Or could the forest islands actually be remnants of human management activities (which should reflect in the species composition)?

I do not know of the use of forests specifically as fire breaks, but forest-savanna boundaries are often thought to be maintained by fires that burn up to, but not into adjacent savannas.  See:

http://www4.ncsu.edu/~wahoffma/Savanna_Forest_Boundaries/Savanna-Forest.html

And maybe start with:

Hoffmann, W.A., Jaconis,S.Y, Geiger, E.L., Gotsch, S.G., Franco, A.C. (2012)  Fuels or microclimate? Understanding the drivers of fire feedbacks at savanna-forest boundaries. Austral Ecology, 37:634-643.

Hoffmann, W.A., Geiger, E.L., Gotsch, S., Rossatto, D.R., Silva, L.C.R., Lau, O.L., Haridasan, M., & Franco, A.C. (2012) Ecological thresholds at the savanna-forest boundary: How plant traits, resources and fire govern the distribution of tropical biomes. Ecology Letters, 15:759-768.

Hello Rainer, thank you for your comments. Your examples from the savanna are quite unheard for me also and interesting. The humid chaco was quite an inhospitable place during pre-columbian times, with only a few scattered hunting tribes that limited any kind of  agriculture or other land use changes. Different to the dry Chaco which is dominated by forests/shrublands and now has one of the highest deforestation rates in the world, the humid chaco is a rather wetland/swamp dominated landscape with very few and small forest (remnants?), with seasonal flooding and poor drainage soil which (we only think) limits the development of forests. The strongest impacts came with cattle grazing in the past two centuries or so (as we can tell from some old drawn maps). Some estancias still conserve the original landscape, where we presume only cattle was induced to native grasslands with no actual deforestation taking place, except for the past decades where forest islands are being cut to plant new more productive exotic grasses. Thanks again and wishfully this can all be put to a new research topic in a near future. Here's an old map of an estancia an a modern map of current vegetation, we have little knowledge of what happened before, but history tells us it was not impacted by any native/indigenous civilization but mostly by hunting gatherers. I'll be glad to let you and anybody interested about the place and our work here.  

In South Africa, Coert Geldenhuys proposed that forest shape and location was heavily influenced by the prevailing bergwinds (hot, dry winds) which carry fires, and the topography. He also discussed the shape of the forest boundary (natural forests that have a rounded profile, rather than an abrupt edge). I saw him present these ideas at a talk many years ago, but here is a paper on the topic (paywall).

http://www.jstor.org/stable/2845603?seq=1#page_scan_tab_contents

In essence, if I remember correctly, the shape and location of the forest patches prevented fire penetration. If the forest-grassland transition were damaged, then the fire would be more likely to be able to penetrate into the forest itself.

I haven't ever seen it formally documented, but under certain conditions in the Rocky Mountains of the US, aspen stands have been used as fuelbreaks when burning grassy meadows.  If memory serves, usually these burns were conducted in the fall when the grass was cured but the aspen stands were relatively moist in the understory.  Don't recall if the leaves were still on the trees or had fallen.  However, aspen leaves generally don't carry fire well when down.  The fuelbed is very compressed and the leaves lack volatile compounds that would aid in burning.  Knowing the structure of the fuelbed and dead fuel moisture content in the forest patches can go a long way in identifying which forest patches are more likely to serve as fuelbreaks and which ones won't.

One thing to consider when trying to understand what role previous societies may have playing in shaping historical landscapes is that hunter-gather cultures do not extinguish their campfires.  They simply abandon them.  Also that fire was a primary tool used by these cultures to alter the landscapes and provide them with various goods and services, even in wet environments such as marshes.  They would have known the conditions under which such sites could burn.  So between the accidental ignitions that could have happened during droughty periods and any deliberate ignitions, the pre-Columbian landscape could have been at least partly shaped by hunter-gather fires.

Hi all

very interesting discussion. One thing I can add is that certain trees can be planted to serve as firebreaks. For example in Colombian tropical dry forest Guazuma ulmifolia has this reputation...

cheers Evert 

Evert, Cool! Do you have a reference for this?

Thanks Truman for these very pertinent and interesting articles. As Hoffman and collaborators put It, it is a possibility that specific edaphic factor could explain forests from advancing towards savannas (in not most of the cases), in our case it would be flooding and bad drainage soils (Or we think that could be, but we are unsure). It would be interesting to determine if this is the main driver or fire for this particular site. Our idea is that native forests as fire-breaks could add to the list of environmental services of forests in savannas of the humid chaco, and could represent another good reason (as if the ones already acknowledged weren't enough!) to conserve this forest islands as natural protection from fires, which, depending on the year can be quite troublesome, putting livestock and human lives in danger. I'd like to thank all other interesting contributions by Kenneth, Louisa and Alan, I am very glad to see this question has been interesting for you. 

Hello Kenneth,  maybe a totally different subject but it was brought to my mind when you talked about human induced fires. On a little experience I had when working in Mexico's fire prevention program in my first years after college. We noticed that fire outbreaks were not only created for production purposes (clearance) or as accidents or due to natural reasons, on the other hand we noticed very poor rural communities started wild fires on purpose, to obtain attention from government agencies (such as ours), which translated into temporary benefits for the community: with wild fires came emergency health services, equipment and tools, occasional paid work for local men to help in fire-breaks and lots of hungry firefighters who needed cooks in town and local produce. We could never be certain about this situation,  but it seems fire and the response obtained from government created an economic opportunity for this very marginalized areas (some totally inaccessible by land). We did work later on convincing this communities to join local voluntary fire-fighter groups mostly focused on prevention. This is just anecdotal, to perhaps think on the opposite effect that public policies can sometimes cause. Thank you for your very interesting comments and inputs.

Hello Evert, 

Coincidentally a dominant species of the forest edge in the humid chaco is Guazuma ulmifolia. Is it the forest or the tree? This is quite interesting! Thanks. As Truman points out, it would be good to have some literature if it is in your possibilities to share.

Truman: don't have a literature reference for this; but this is what Colombian restoration practitioners say... (three rows of Guazuma will stop almost any fire)

Karim,

No studies I can cite, but in the northern Peten region of Guatemala, when local swidden agricultors burn slash prior to planting, they usually do not make any sort of fire break, and unless it is unusually dry, the fire generally does not spread into surrounding forest, at most scorching the forest edges. However, one practice that aids this is doing a preliminary "black-lining" or burning the downwind edge of the slash first. Then, if there is substantial breeze, when the fire arrives at the downwind edge, after a mighty surge across the slash, the tall flames don't reach the forest edge. This, by the way, is not a truly wet region, but rather a sub-humid lowland region with a mix of evergreen and deciduous tree species.

Of course, in unusually dry years, such fires do burn a lot of forest in this area, so when very dry this fire-resistance of the forest diminishes considerably. And studies in the Amazon suggest that, once burned, forests there become more burnable in a vicious cycle.

Very interesting what Evert Thomas says, suggesting that Guazuma ulmifolia is fire-resistant. I have regarded it as simply a common species in human-altered sub-humid landscapes in Mexico and Guatemala, and not considered that it may be particularly fire-resistant. 

Dave Whitacre

Dear Karim

Accidental fires in vegetation and grasslands destroy the forage available for herbivores, often in seasons when the pasture is not abundant. Firebreaks protect pasture, as well as forests and plantations. Firebreaks contribute to protection of vegetation from bushfires in the dry season, isolating separate areas by wide strips designed to reduce or halt the progression of fire in bush, pasture or forest. They constitute an important technique to control or prevent fire, especially when used together with other fire prevention and control measures. This technique contributes to the protection of the vegetation and the fauna affected by fire, the protection of the vegetal biomass useful to wild and domestic herbivores and the regeneration of vegetation. In pastoral areas, firebreaks made more secure the controlled fires.

Techniques:

Firebreak consists in strips of land, about 40 m wide, able to stop the progress of a fire front, or at least reducing the violence. The direction is across the dominant direction of wind.  We distinguish different types of firebreaks, selected according to type of soil, the type of vegetation, the climate and the available equipment:

bare soil firebreaks, (expensive and demanding in terms of labour and equipment).

Planted firebreaks, Firebreaks with native vegetation, (only in dense forest, without lowerlayer - understorey).

Cropped firebreaks, the least expensive option. Land that is continuously cultivated and actively managed. This must be linked to intensification of agricultural practices to avoid inherrent risks of fertility losses and soil erosion.

Best Regards

I do not know the humid Chaco in Paraguay, but in other regions with wet forest there is a recognition that intact humid forests generally will not burn ... while neighbouring open vegetation often will (if frequent fires occur this is normally a grassland).  Fire is often what maintains a sharp forest-grassland boundary.  Forests thus block fire.

The difference in flammability allows the more open areas to burn repeatedly while forests burn only rarely (but can switch as a result of prolonged drought or canopy damage or both).  There is a big literature on this (especially the positive feedback that once a forest burns once it is much more likely to burn again).

See, e.g., Hoffmannet al. 2012 Ecological thresholds at the savanna‐forest boundary: how plant traits, resources and fire govern the distribution of tropical biomes. Ecology Letters, 15(7), 759-768.

Cochrane et al, 1999 Positive feedbacks in the fire dynamic of closed canopy tropical forests. Science, 284(5421), 1832-1835.

Mayer & Khalyani 2011. Grass trumps trees with fire. Science, 334(6053), 188-189.

If you look at these in google and see the citing articles you will get a good overview of the topic.

I see in East Africa that the grass-forest boundary may shift in as fire enters the forest margin or shift out as evergreen understorey vegetation (herbs like Acanthus spp)  spread out ... this allows forest blocks to change in size.  Much depends on the timing of the fires versus fuel availability at the forest edge ... and the  behaviour of animals such as elephants and other browsers at the forest margin.  This is not so well documented but is widely recognised.

Douglas

Hi Karim,

In the Northern Bolivian Amazon forests are often left as a barrier between grasslands and agricultural fields and agroforestry systems. Practices are similar to those described in previous answers, and smallfarmers generally burn their fields about two days after a heavy rain, knowing that the sun will have dried up sufficiently the plant material in their previously slashed field to guarantee a good burn, while the shade of the surrounding vegetation secures enough humidity to prevent any significant spreading of the fire into the forests. I have seen some smallfarmers experimenting with fire in the edge of their plots to make sure humidity is on the desired level. Knowing about this capacity of forests, if possible local make sure there is an area of forest between fireprone vegetations (e.g. grasslands and young secondary vegetation) and their agricultural fields. Although I do not know of any specific literature on this subject, I believe Arienne Henkemans describes these practices in more detail in her publication "Between tranquility and hardship in the Bolivian Amazon". In my latest book "agroforestry systems in the bolivian amazon" (see my RG site) I also include several comments on this subject. And finally I'd like to suggest you check the text: "Como combatimos los incendios" (available from my ISSUU-site: http://issuu.com/vincentvos/docs/comit___interinstitucional_2015_c__) describing actions by an interinstitucional commitee for the control of fires... including the implementation of fire breaks with mango trees.