Return to ecosystems summary page

14. Cool temperate forest (closed forest, usually mainly deciduous. Includes mixed conifer-broadleaved forest)

(Corresponds to subdivision of Olson temperate broadleaved and mixed forest).

References directly cited in these pages (does not at present include secondary citations)


This type of closed, mainly broadleaved forest vegetation exists in various parts of the world under a mean annual biotemperature of below about 12 C according the the Holdridge (1967) classification (above 12 C is classified here as warm temperate forest). The cool temperate broadleaved forests have supported agriculture and forestry for thousands of years, and have been very extensively cleared or cut over wherever they occur in the world (Tallis 1990). Thus, the forest we see at present is nearly all at various stages of recovery from agricultural clearance, cutting or replanting. There can be no doubt that the 'youthfulness' of many of these forests means that soil, vegetation and litter carbon storage are below what they would have been in their primaeval state several thousand years before. This is particularly evident when one sees the large numbers of big senescent trees and fallen branches in forest reserves that have not undergone significant forest management for thousands of years (e.g. the ancient forest reserves of Poland and Hungary; Walter 1971). Natural grazing and other animal disturbance (e.g. by beavers) may have been removed from certain old forest areas (although in the case of the central European forests wild animal grazing is still very important), but it has often been replaced by the grazing of domesticated animals up until the relatively recent past.

Lightning-induced fire in temperate forests generally seems to be very rare, although at its climatic margins, and when high proportions of conifers are also mixed in, it may occur much more frequently. For instance, in a 1 km study area in NW Minnesota, Clark (1990) found evidence of fire hitting the same area of forest at intervals as frequent as 7-15 years during the last 750 years, although for the most part this was only ground fire (consuming mainly litter) and not the more destructive crown fire. Wein & MacLean (1982) suggest that mixed temperate conifer-hardwood forests are actually destroyed by fire much less frequently; about every 100 - >200 years, which would give sufficient time for many trees to reach a large size. Heinselman (1978) reaches a similar conclusion, suggesting a ground fire (though not the more severe crown fire) cycle of 200-300 years for the mixed northern margin forests of the eastern USA/south-eastern Canada, affecting stands of forest damaged by wind or severe frost. Heinselman notes that such species as red oak, black cherry (Prunus serotina) and white pine which are often present in these forests, and that these seem to require occasional ground fires for their regeneration. Given the high rate of accumulation of carbon in forest biomass (see below), it seems unlikely that a fire frequency of this order would be sufficient to keep the natural forests down to the sort of biomass level which is generally observed in such anthropogenically disturbed forests at present.

Forests which are dominated by hardwoods would seem to burn still less frequently than this. It is of interest to note that the 'witness tree' data of Lambert (1991) (see below) from presettlement Illinois, suggests very high biomass even in climatically marginal NW forests close to the prairie zone. Thus it seems that crown fire was not frequent enough to have any major effect in suppressing biomass.

Also to be considered at least as far back as the early Holocene is the possible role of pre-agricultural human communities in creating clearings to encourage prey animals to graze. Agriculture was present at low densities through most of Europe by 6,000 years ago. There are signs of forest burning for such purposes in the mesolithic and early neolithic of Europe. Such disturbance factors may have tended to depress biomass. However, in the pre-agricultural early-to-mid Holocene it is unlikely that population densities were high enough to significantly affect overall forest biomass and carbon storage. However, Willis and Bennett (1994) argue that there is very little evidence for any real human impact on forest structure in most parts of Europe before about 4,000 years ago, and amongst Amerindian groups of the temperate forest zone (living partly as farmers and partly as hunter-gatherers) in North America, agriculture does not seem to have reached the cool and warm temperate forests of the Mississippi basin and the southeast until between 2000 and 1300 years ago (Hurt 1987).

Travellers' records from the late 1700's and early 1800's in the north-eastern USA seem to confirm that the forest canopy was essentially unbroken except where European settlers had colonized an area (Perlin 1991).

Suggested summary values for the preanthropogenic state

Storage (tC/ha) Ecosystem component
25 tC/ha Dead standing trees, coarse woody litter, leaf litter and other debris
160 tC/ha Above and below-ground vegetation
140 tC/ha Soil organic carbon
325 tC/h Total carbon storage



Storage (tC/ha) Location Author(s)
25 tC/ha (1.) Coarse woody debris. Deciduous forest M. Harmon pers. comm.
3.45-8.1 tC/ha (2.) Coarse debris. China Harmon & Hua (1992)
20 tC/ha (3.) Coarse debris. North-eastern USA. Harmon pers. comm.
27.5 tC/ha (4.) Debris+litter. Mixed forest, USSR Kolch.& Vinson (1994)

(1.) Based on his impression from various site studies, and allowing for the role of anthropogenic disturbance and harvesting of timber even in old-growth forests.

(2.) Chinese deciduous old-growth forests, various site studies; the figures for individual site studies vary between 3.45 and 8.1 t C/ha coarse woody debris.

(3.) North American (New England) deciduous old-growth forests, usually 20 t C/ha in coarse woody debris (towards the south, this figure falls to around 10 t/ha C). (Harmon pers. comm.).

(4.) Kolchugina & Vinson 1993, based on a compilation of data on mixed-deciduous forests in the former Soviet Union. Many of these forests have been subject to recent wood harvests and maintainance, so their stock of coarse woody debris can be expected to be considerably lower than would be the case in natural forests. The figures are as follows; 17.5 tC/ha woody debris, 10 tC/ha in litter (Total = 27.5 tC/ha).

Conclusion; It is striking that the less human influence there is on a cool temperate forest, the greater the amount of coarse woody debris that accumulates. For instance, at the time of the spread of European agriculture through the eastern North American forests during the 1850's, photographs and paintings from the still 'virgin' forest surrounding new swidden sites generally record a dense forest canopy punctuated by frequent standing dead trees (e.g. some of the paintings shown in Sears 1994). One can contrast these forests with their modern replacements, in which dead standing trees and fallen boughs are almost absent. It is appears that in these forests at least, coarse woody debris was a major store of carbon in the preanthropogenic state.

Based on the few observations cited above, an overall value of about 25 tC/ha is suggested for litter and coarse woody debris combined, in the pre-anthropogenic state. No specific data are available on the subcategory of swamp forest, but presumably the debris pool would be much greater in this form of forest.



Storage (tC/ha) Location Author(s)
100 tC/ha (1.) Deciduous/mixed forests Olson et al. (1983)
170 tC/ha (2.) Deciduous forests >100 yrs Cannell (1982)
172 tC/ha (3.) Old-growth forest, China. Harmon & Hua (1992)
174 tC/ha (4.) Old-growth forest? Rodin et al. (1975)
150 tC/ha (5.) Nothofagus, NZ >300 yrs. Cannell (1982)
190 tC/ha (6.) Himalayan forests Singh et al. (1994)
304 tC/ha (7.) Upland forests. Illinois presettlement Lambert (1991)
499 tC/ha (7.) Bottomland forests. Illinois presettlement Lambert (1991)
305 tC/ha (7.) Inundated forests. Illinois presettlement Lambert (1991)
106 tC/ha (8.) 80-120 yr-old forest, Tennesee Hanson et al. (in press)
471 tC/ha (9.) 120 yr old forest, corn belt of USA Birdsey (1996)
146 tC/ha (10.) 120 yr old birch-maple-beech forest, NE USA. Birdsey (1996)

(1.) Based on a review of literature. 80 (low) - 100 (medium) - 140 (high) tC/ha. One of the major sources cited is Cannell's book (below) which is based largely on figures for woodland that have either been planted, partly harvested or allowed to regenerate from clear-felling within the past century. Clearly, such figures will give a lower biomass than would be representative of deciduous/mixed forests in a non-anthropogenic state. It appears from the literature mentioned below that the few surviving non-logged sites have much higher carbon storage values, on average, which emphasizes the inappropriateness of using Olson et al.'s figures for the past.

(2.) Cannell (1982) gives biomass values for temperate forest stands of various ages, gathered from forest inventories from around the world. Here, values are selected for stands of semi-natural forest or plantation that are over 100 years old since the last clear-felling or major harvesting event. Adding an extra 30% for root mass (= 0.25 roots:shoots), and multiplying the biomass figures by 0.475 to convert to carbon, the following figures were obtained; Czechslovakia 175, Denmark 166, Belgium 204, Belgium 185, France 112, Netherlands 147, Poland 138, Poland 141, USSR 180, USSR 229. This gives a mean value of 170 tC//ha for European broadleaved deciduous forest. Note that the indications amongst this dataset are of the older forest sites increasing in biomass well after the first century, so these figures may not represent the full biomass reached by many forests before human intervention, even with a background disturbance of occasional storm-throw events.

(3.) Harmon & Hua (1992) cite a study of areas of Chinese (Changbai province) deciduous old-growth forest >160 years old having above-ground living material ranging from 103 to 165 tC/ha. There is a need to add an extra % for belowground roots; adding an extra 30% to the total this would give an overall carbon mass of 130-214 tC/ha (simple mean of these two figures = 172 tC/ha).

(4.) Rodin et al. concluded their figure from a summary of IBP evidence. However, IBP studies in general have been criticized for concentrating on high biomass sites, although for present purposes these may also have been representative old-growth sites.

(5.) Cannell (1982) presents data for Lowland Nothofagus stands, >300 yrs old. New Zealand 155, New Zealand 145 Mean for these two sites = 150 tC/ha. Whilst evergreen, Nothofagus forests mostly exist under cool-temperate conditions in uplands on North Island and on South Island, where they replace the original forests cleared by Maoris within the last 1000 years (McGlone 1994).

(6.) Based on the figures discussed for samples of Himalayan temperate forest (see section on Warm Temperate Forest; ecosystem type 10).

(7.) Lambert (1991) and Lambert & Brown (in press) have derived carbon mass estimates from the extensive 'witness tree' data (several hundred samples) throughout the state of Illinois (USA) at the time of first European settlement. The data, from trees over 5cm diameter that were in place at the corner of regularly spaced land plots, included species and diameter at breast height. Lambert used these data, together with standard forest allometric equations, to deduce the carbon mass aboveground, then assumed 25% root mass and added an extra 18.3 tC/ha for understory vegetation.

These figures are evidently far higher than present-day carbon storage in forested areas of Illinois at present. Since this is a relatively 'marginal' area for forest, with large areas of prairie vegetation occurring in parts of the state that contributed to these high carbon storage estimates, it may be taken to suggest that despite being close to its dry climatic margins the northern temperate forest did not in fact suffer unduly from fire events (despite the previous Indian population using fire as a hunting aid), nor did climatic dryness reduce the forest biomass below what one would consider a very high biomass for a temperate deciduous forest.

Also of interest is the fact that the seasonally inundated forests on relatively swampy ground also had a very high carbon mass in vegetation.

(8.) This figure is for an intensively studied area of upland Quercus/Carya deciduous forest in the Walker Branch watershed of east Tennessee. The carbon mass (derived from the biomass figure by multiplying by 0.475) is estimated from DBH and root stump dimensions, using a forestry regression equation. This forest is one of the sites which contributed to the relatively low eastern USA forest average of Olson et al. In 1967, when the IBP measurement was taken, the carbon mass of the same area of forest was estimated to be only 76 tC/ha using the same regression equation; the estimate of the extra carbon mass (about 25% extra) from the equation is due to the ongoing diameter growth of the trees and illustrates how rapidly carbon may accumulate in forests with age.

(9.) Birdsey extrapolates from younger stands using stand accumulation models. The fact that this is based on a model opens up the possibility of errors in the estimated carbon storage accumulation over time.

This carbon storage figure is estimated as 'carbon in live vegetation' (including understory) from growth and yield studies for managed mixed hardwood timberland in the USA corn belt (NW eastern temperate forest zone). A table of five-year intervals is given, with a gradual build-up to this total. Note that the total seems unusually high by comparison with most present-day forest in the USA, but is based on multiple sites and represents a study carried out by an authoratative source.

(10.) A figure of 146 tC/ha for 120-year old NE USA beech-maple-birch forest is given as the end product of a table of five-year increments based on extrapolation from observations of younger forest sites. It is unlikely however that a 120-year age represents an end-point in carbon accumulation in forest stands. An additional 3 tC/hacarbon is present in understory vegetation, bringing the total to 149 tC/ha. Other totals given by Birdsey for 120-year old stands include: Lake states aspen-birch, 137 tC/ha; Lake States bottomland hardwoods 242 tC/ha; Central States (western edge of eastern forest belt and west of Mississippi) oak-hickory, 145 tC/ha; northeast red pine, 328 tC/ha; Appallachian loblolly pine plantation, 277 tC/ha; Lake States white/red pine; 308 tC/ha. It is noticable that stands older than 100 years have much higher carbon content than the younger forest that prevails over most of the eastern USA. Presumably, carbon would continue accumulating for some time after 120 years in natural forests, since the average disturbance rate (see above) would be less than this. Thus, these figures are relatively conservative as estimates of the forest carbon storage of 'natural' or pre-European settlement forests, and suggest that the true figure of the cool temperate eastern USA forests would have been in the range 150-350 tC/ha.


The figures given above show a striking discrepancy between the 'present-natural' carbon storage values of around 80 tC/ha in biomass for eastern USA forests (Brown et al. 1997), and the values suggested from old forest inventories and from local surviving stands of old growth forest (e.g. several of the values from the table above). As Brown et al. (1997) suggest, it is likely that the present-actual carbon storage in vegetation in the case of eastern USA forests is no more than half of what it would be in the potential steady-state condition. The indications are that cool temperate deciduous/mixed forest tends towards an equilibrium biomass of at least 160 tC/ha, probably nearer 250 tC/ha, in the presence of a natural disturbance regime (destruction only every 200-300 years, or less frequently) but in the absence of intense anthropogenic disturbance. Swamp forest would possibly have a somewhat lower biomass (though many willows - Salix - and poplars - Populus - can reach a very considerable size, so it might not be very much lower), due to more frequent channel disturbance on floodplains. However, the 'witness tree' data discussed above suggest that even floodplain forests and climatically marginal north-western forests had a very high carbon storage compared with the present-day forests of the eastern USA.

Such conclusions are further backed up by other more general observations: though it is only in North America that we can through historic writings glimpse the northern temperate forests as they would have been in a primaeval state. It is apparent from such early writings that dense, high forest blanketed virtually all of the north-eastern USA before artificial forest clearance by settlers (Perlin 1991). 'Natural' clearings occurred but were not an important part of the landscape, and even these may have in part been maintained by Indians (Perlin 1991). By the early 1800s, when writings by travellers and naturalists became more common, the forests to the east of the Appallacians were already extensively cleared and selectively logged, much of the largest lumber having been extracted for ship-building.

Some of the few clues that we can reasonably obtain to the state of these forests come from travellers in the Ohio region, where settlement was not yet extensive. At present, most of this area has been deforested, and the forests which remain are all in a relatively 'young' successional state due to logging. In the early 1800's, however, the forests clearly contained a large proportion of very big trees, including common species which are nowadays not generally seen as growing to a large size (presumably because all large individuals of their populations have been felled in the recent past). Perlin (1991) summarizes what the travellers saw when they journeyed to the west of the northern Appalacians (p.325-326);

Edmund Dana wrote that "the forest trees of the west (i.e. west of the Appalacians) grow to an uncommon height." Measurements taken of various species by early visitors to the region proved Dana's assessment correct. The oak commonly had "a straight trunk without a single branch for seventy feet; and from that point to the upper branch it has measured seventy feet more," according to Francis Baily, a visitor to the Ohio country at the end of the eighteenth century. The girth of many of the trees was quite impressive too. Martin Birbeck a curious and adventuresome Englishman, reported in his diary in 1818, "Yesterday I measured a walnut tree almost seven feet in diameter, and just by (its side), were rotting...two sycamores of nearly equal dimensions...I (also) measured a white oak, by the roadside, which at four feet from the ground was six feet in diameter." Sycamores growing on the banks of the Ohio were judged as the "loftiest and largest trees of the United States." A French botanist collecting samples in the Ohio area between 1800 and 1810 reported finding a member of the species with a circumference of over forty feet.



Storage (tC/ha) Location Author(s)
152 tC/ha (1.) Cool deciduous forest Zinke et al. (1984)
130 tC/ha (2.) Cool mixed (conifer/bdlvd) woods Zinke et al. (1984)
134 tC/ha (3.) Temperate deciduous forest Schlesinger (1985)

(1.) Based on a review of global standardized soil data. 105 samples, with a standard deviation of 12.6. This dataset is probably heavily biased towards anthropogenically disturbed forest sites, so is probably an underestimate of the soil carbon of such forests in a non-anthropogenic state.

(2.) Based on a review of global standardised soil data. 56 samples, standard deviation of 9.5. This dataset is probably heavily biased towards anthropogenically disturbed forest sites, so is probably an underestimate of the soil carbon of such forests in a non-anthropogenic state.

(3.) Based on literature review. Again, these values are probably depressed by anthropogenic disturbance of the sites.

Conclusion; From the extensive amount of standardised data that has been gathered, it seems that a value of 140 tC/ha for soils would be a reasonable (if conservative) value for non-anthropogenic forests on non-swamp soils. However, this value is probably much too low for swamp forests (14b), which tend to accumulate fen peats.