Prehistoric Fires Research Articles

Fire History and Vegetation Pattern in Mesa Verde National Park, Colorado, USA

piñon–juniper woodlands (Pinus edulis, Juniperus osteosperma, and J. scopulorum) and petran chaparral communities (Quercus gambelii, Amelanchier utahensis, Cercocarpus montanus, and other tall shrub species) cover much of the Colorado Plateau in the southwestern United States. Long-term fire history and successional dynamics are poorly understood in these vegetation types. Therefore, we lack a suitable historical context for interpreting the ecological significance of large fires and dramatic vegetative changes that have occurred recently in these ecosystems. For example, in Mesa Verde National Park, located in southwestern Colorado, four large intense fires in the last 50 years have threatened significant cultural and natural resources and have caused debate over whether Mesa Verde's fire regime has been significantly altered by human activities in the last century. In this study, we dated prehistoric fires in shrublands dominated by Gambel oak (Quercus gambelii) by aging stems that resprouted after fire. We mapped the spatial extent of all fires >10 ha that occurred during the last 150 years within a 6600-ha, shrub-dominated portion of Mesa Verde National Park. The turnover time (years required to burn an area equal to the entire shrubland zone) was ∼100 years under the “natural” fire regime of the mid- to late 19th century. Fire occurrence was reduced substantially during the first half of the 20th century, but the current fire regime (since about 1950) appears to be similar to that of the 19th century—despite a continuing policy of total fire suppression. The “natural” fire turnover time in piñon–juniper woodlands of Mesa Verde is about 400 years. A sharp boundary exists between piñon–juniper woodlands at slightly lower elevations in the southern portion of the park and petran chaparral at slightly higher elevations in the north. This pattern is explained, in part, by more extensive fires in the northern area, which favor resprouting shrubs and eliminate the fire-sensitive piñon and juniper. The less frequent occurrence of large fires, and resulting persistence of woodland in the southern portion of the park, may be due in part to natural barriers to fire spread (cliffs and sparsely vegetated slopes) to the south and west of the piñon–juniper woodlands. Our findings demonstrate that fire frequency and extent in Mesa Verde during the last 50 years have not been greatly different from the “natural” fire regime of the late 1800s. Therefore, the recent large fires in the park, and the vegetative responses to those fires, appear to be within the historic range of variation for this ecosystem.

FIRE HISTORY AND VEGETATION PATTERN IN MESA VERDE NATIONAL PARK, COLORADO, USA

Pifion-juniper woodlands (Pinus edulis, Juniperus osteosperma, and J. sco- pulorum) and petran chaparral communities (Quercus gambelii, Amelanchier utahensis, Cercocarpus montanus, and other tall shrub species) cover much of the Colorado Plateau in the southwestern United States. Long-term fire history and successional dynamics are poorly understood in these vegetation types. Therefore, we lack a suitable historical context for interpreting the ecological significance of large fires and dramatic vegetative changes that have occurred recently in these ecosystems. For example, in Mesa Verde National Park, located in southwestern Colorado, four large intense fires in the last 50 years have threatened significant cultural and resources and have caused debate over whether Mesa Verde's fire regime has been significantly altered by human activities in the last century. In this study, we dated prehistoric fires in shrublands dominated by Gambel oak (Quercus gambelii) by aging stems that resprouted after fire. We mapped the spatial extent of all fires >10 ha that occurred during the last 150 years within a 6600-ha, shrub-dominated portion of Mesa Verde National Park. The turnover time (years required to burn an area equal to the entire shrubland zone) was -100 years under the natural fire regime of the mid- to late 19th century. Fire occurrence was reduced substantially during the first half of the 20th century, but the current fire regime (since about 1950) appears to be similar to that of the 19th century-despite a continuing policy of total fire suppression. The natural fire turnover time in pifion-juniper woodlands of Mesa Verde is about 400 years. A sharp boundary exists between pinon-juniper woodlands at slightly lower elevations in the southern portion of the park and petran chaparral at slightly higher ele- vations in the north. This pattern is explained, in part, by more extensive fires in the northern area, which favor resprouting shrubs and eliminate the fire-sensitive pifion and juniper. The less frequent occurrence of large fires, and resulting persistence of woodland in the southern portion of the park, may be due in part to barriers to fire spread (cliffs and sparsely vegetated slopes) to the south and west of the pifnon-juniper woodlands. Our findings demonstrate that fire frequency and extent in Mesa Verde during the last 50 years have not been greatly different from the natural fire regime of the late 1800s. Therefore, the recent large fires in the park, and the vegetative responses to those fires, appear to be within the historic range of variation for this ecosystem.

Using fire scars and growth release in subfossil Scots pine to reconstruct prehistoric fires

Fire scars indicating low- to moderate-intensity fires on peat deposits have been sampled from subfossil Scots pine ( Pinus sylvestris L.) at sites in England, Wales and south-west Ireland. Analysis of ring-width responses to one fire event in 2800 BC illustrates its rejuvenating effect on Pinus sylvestris woodland, supporting a growing body of modern and palaeoecological data that illustrates the regenerative role played by fire in persistence of Pinus sylvestris woodland. Both the scale and timing of these fires suggest that infrequent low- to moderate-intensity fires are sufficient to stimulate Pinus sylvestris growth. This effect is shown by average increases in ring-width following the fire of between 0.62 and 1.16 mm in non-scarred trees and between 0.92 and 2.74 mm in fire-scarred individuals. Growth release in non-scarred trees may prove to be a more reliable method of detecting fire than using the relatively rare fire scars alone. Radii at time of scarring varied between 1.85 and 11.2 cm, much smaller than is predicted to survive from modern studies.

Spectral investigation of the thermoluminescence of heated flint (silex)

Test measurements to determine the sufficient heating of silex (flint) artifacts from prehistoric sites for TL dating revealed a large variation of TL-characteristics. To determine TL-emission wavelengths suitable for dating, the spectra of several samples from different geological origin were measured. A dominant emission in the orange band was found, consisting of three peaks centred around 90–130°C, 190 and 270°C or only a broad single peak at around 250°C. Two UV peaks appear at 100–190°C and around 360°C. At the latter temperature a blue emission occurs as well. Simulation experiments of prehistoric fires for various temperatures were performed. In contrast to the blue emission the UV emission around 360°C shows slight fading and/or bleaching. The blue band also requires slightly lower temperatures for complete resetting than the UV–TL. Furthermore, this blue-TL emission is less light sensitive, especially for red laboratory light, making sample preparation much more convenient. Sufficiently heated silex samples can easily be detected by the position of the high temperature TL-peak at about 360°C and a Gaussian-like shape of the blue peak, and of course by the plateau test. The observed TL-variation of heated archaeological silex might be due to the different genesis of the material, but it can also be attributed to different degrees of ancient heating.

Prehistoric fires and land-cover change in western Kenya: evidences from pollen, charcoal, grass cuticles and grass phytoliths

High rates of sediment accumulation in Lake Simbi, western Kenya, has enabled fine-resolution fossil assemblages to reveal short-term changes in vegetation cover between 4320 ± 250 yr BP and 2880 ± 90 yr BP. Pollen analysis of a 4.2 sediment core covering this period reveals a vegetation change around 3300 yr BP. Analysis of charcoal shows that wildfires were prominent prior to the pollen-inferred vegetation change. Remains of grass cuticles and grass phytoliths show an expansion of grasslands and an increase in the diversity of grass types following the period of extensive fires. Also following the period of wildfires and the vegetation change, there were shifts within the grass flora to more Chloridoid grass types and less Panicoid grass types. This paper addresses one of the major problems of environmental reconstruction in lowland savannah: namely differentiating palaeoecological changes within the Gramineae.

Does the rare Banksia goodii have inferior vegetative, reproductive or ecological attributes compared with its widespread co‐occurring relative B. gardneri?

Banksia goodii (rare) and B. gardneri var. gardneri (widespread) are closely related rhizomatous evergreen subshrubs of south western Australian scrub‐heath and woodland. B. goodii is restricted to sixteen populations, most of which are remnants on roadsides due to landscape fragmentation from agricultural activities. Bioclimatic profiles indicate that B. gardneri spans a wider range of climates than B. goodii. However, their post‐fire survival and resprouting vigour, rates of seedling growth and soil penetration and susceptibility to seedling predators are similar. Both have the potential for geographic range extension, especially B. goodi. Seedlings of both species survived for at least 3 years at sites with 2 (but not 3) months shorter and 1 month longer growing seasons than experienced by natural B. goodii populations. Seedling establishment did not occur in the absence of fire and was low (1.3% of released intact seeds for B. goodii) even following en masse release of seeds after fire. Sowing of intact seeds, especially in cleared areas and with protection from herbivores, increased this to 35%. Fire survival among adults of both species was 100% following two fires. B. goodii, which roots along its rhizomes, produces ramets, while B. gardneri, with a lignotuber instead, does not. Seed production, pre‐dispersal granivory and canopy seed storage are low but similar for both species, although B. gardneri has a superior level of germination. The biology of B. gardneri thus provides few insights as to why B. goodii is rare. There were no differences in topo‐edaphic attributes, community structure, shading or litter cover between localized sites where either species occurred alone or together. Differences in present or prehistoric fire regimes, clearing (habitat fragmentation), susceptibility to generalist fungal pathogens, or impacts of alien invasives, cannot explain the distribution of B. goodii relative to B. gardneri. The rarity of B. goodii may be a result of its recent origin, habitat specialization (dense woodland), and the impacts of habitat fragmentation within its current range, rather than inferior vegetative, reproductive and/or ecological attributes.

Dating prehistoric bog‐fires in northern England to calendar years by long‐distance cross‐matching of pine chronologies

The ages of prehistoric fires can be approximated by radiocarbon dating of charcoal or associated material, but such dating is often inaccurate and at best imprecise. Pine trunks preserved in British and Irish peats occasionally show firescars, which might be dated through dendrochronology to yield calendar-year dates. However, unlike oak, there is no master pine chronology to provide absolute dates, so dating is dependent on interspecies cross-matching; for sites in the British Isles with no dated oaks, calendar-year dating of prehistoric pines has hitherto proved impossible. We present a first success in dating, accurately and precisely, prehistoric fire events recorded in subfossil bog-pine trunks, using long-distance cross-matching of pine chronologies between White Moss, Cheshire, and the Humberhead Levels, England. Results demonstrate a bog-fire in Cheshire in spring 2800 BC, and again in 2710 BC, between spring and summer. Further successful long-distance cross-matching of pine would permit international climatological comparisons. © 1997 John Wiley & Sons, Ltd.

Dating prehistoric bog-fires in northern England to calendar years by long-distance cross-matching of pine chronologies

The ages of prehistoric fires can be approximated by radiocarbon dating of charcoal or associated material, but such dating is often inaccurate and at best imprecise. Pine trunks preserved in British and Irish peats occasionally show firescars, which might be dated through dendrochronology to yield calendar-year dates. However, unlike oak, there is no master pine chronology to provide absolute dates, so dating is dependent on interspecies cross-matching; for sites in the British Isles with no dated oaks, calendar-year dating of prehistoric pines has hitherto proved impossible. We present a first success in dating, accurately and precisely, prehistoric fire events recorded in subfossil bog-pine trunks, using long-distance cross-matching of pine chronologies between White Moss, Cheshire, and the Humberhead Levels, England. Results demonstrate a bog-fire in Cheshire in spring 2800 BC, and again in 2710 BC, between spring and summer. Further successful long-distance cross-matching of pine would permit international climatological comparisons. © 1997 John Wiley & Sons, Ltd.

Fire in the Forests of Maine and New Hampshire

FAHEY, TIMOTHY J. and WILLIAM A. REINERS (Dept. Biol., Dartmouth Coll., Hanover, New Hampshire 03755), Fire in the forests of Maine and New Hampshire. Bull. Torrey Bot. Club 108: 362-373. 1981.-Modern fire records for Maine and New Hampshire and other historical evidence were investigated to infer the occurrence and distribution of fire in pre-settlement time. Between the decades of 1910-1920 and 1960-1970 fire incidence increased and average fire size decreased sharply, with the net effect being a several-fold diminution of land burned per year. Fire was unequally distributed among forest types. Pine-dominated forests in the south-coastal part of the States were burned most frequently, with northern hardwood forests in central portions intermediate and spruce-fir forests least influenced by fire. The spruce-fir type in northern New England appears to be an exception to the generalization that coniferous forests burn more readily than hardwoods. Historical evidence suggests that in preColumbian time a significant potential for fire ignition existed in association with lightning and the incendiary activities of Native Americans. Because human control of fire size was lacking, the extent of prehistoric fires may have equalled that in the early Twentieth century.