SCOPE 13 - The Global Carbon Cycle

ABSTRACT

The pools and fluxes of carbon in terrestrial biota form important links in the total biogeochemical cycle. This study evaluates some of the earlier estimates while an attempt is also made for a new assessment in the light of many recent data on production and phytomass becoming available through the efforts of the International Biological Programme. The problems of classifying ecosystem types and of obtaining their real surface areas are being discussed. Recent surface areas are evaluated and used for calculating total production and phytomass. Total net primary production of terrestrial biota is estimated to be 60 x 10¹⁵ g C/year produced by 560 x 10¹⁵ g C living phytomass. The total amount of standing dead is estimated as 30 x 10¹⁵ g C, and the total amount of litter as 60 x 10¹⁵ g C. An attempt is made to assess the total litterfall, being the flux from living to the dead compartment in the biota. The value of total litterfall is estimated to be between 45 and 50 x 10¹⁵ g C. Soil organic matter amounts to 16002000 x 10¹⁵ g C, a value much lower than the most recent one published, but higher than any of the earlier estimates. Man's activities influencing the carbon cycle are discussed, including urbanization and industrialization. Potential production of ecosystem in relation to world population and food demands is briefly discussed.

5.1 INTRODUCTION

Primary productivity is an important link in the carbon cycle, since it is the main flux from the atmosphere to the biota. Primary or basic productivity of an ecological system, community, or part can be defined as the rate at which radiant energy is stored by photosynthetic and chemosynthetic activity of producerorganisms, chiefly green plants, in the form of organic substances which can be used as food materials (Odum, 1971). In the process, CO₂ from the atmosphere is converted into glucose (C₆H₁₂O₆), the foundation stone for any further biochemical synthesis of new complex compounds. A distinction must be made between gross primary productivity, which is the total rate of photosynthesis including the organic matter (mainly C₆H₁₂O₆) used up in respiration (R_A) during the measurement period, and net primary productivity (NPP), which is the rate of storage of organic matter in plant tissues in excess of respiration; it is also called ^,apparent photosynthesis' or `net assimilation' (Odum, 1971). The terms 'productivity' and `rate of production' are interchangeable here. The term `net production' is used to designate a total of accumulated organic matter. Although not strictly necessary, a time element is always used as a reference scale, such as one year when speaking of agriculture. In this sense the term `production' is employed in this study. It should be noted that in forestry the terms `net productivity' and ^'net production' are only used to designate the annual wood growth or even the usable above-ground wood increase. Sometimes the increase in phytomass during a certain period of time is called net production, not taking into account loss in material caused by dying off or grazing.

The assimilation of CO₂ can be estimated, to a certain extent, in terms of the reserves of phytomass, i.e. the total reserve of living organic matter in the aboveground and underground spheres of plant communities. For the estimates, a distinction should be made in the structural elements of the phytomass, such as perennial and annual above-ground parts, non-green and green assimilating parts, etc. Account must also be taken of dead organic matter in the ecosystem in the form of litter or peat deposits (see Bazilevich, 1974). For an estimate of the total amount of carbon in terrestrial ecosystems, the soil organic matter (humus) should also be considered. A measure of the returned organic matter is the annual litter fall, including dead parts of plants or entire plants above and below ground in the communities, and the decomposition rate (the rate at which the litter disintegrates through microbial activities and becomes incorporated in the soil as humus). Decomposition causes the main flux of carbon from the biosphere to the atmosphere. In the process of the life activity and of dying off of plants, chemical elements are returned to the atmosphere and lithosphere. This is the essence of the biogeochemical work of the living matter, or the so-called small biological cycle of chemical elements.

The magnitude of the world's terrestrial production and phytomass can be estimated in two different ways, either by classifying the biosphere into ecosystem types and estimating averages and total values for each of these, or by modelling the effects of environmental factors on productivity and phytomass and integrating the results of the model for the earth surfaces. The first approach is used in this study (see Whittaker and Likens, 1973a, 1973b, 1975). Details on the second approach can be found in Lieth (1972, 1975).

Numerous estimates have already been made of the total world net primary production. Table 5.1 lists the various estimates in chronological order, with land and marine biota treated separately (Bazilevich et al., 1971; Whittaker and Likens, 1973b). One of the most recent estimates of total terrestrial production is 53 x 10¹⁵ g C/year (117.5 x 10¹⁵ g dry matter) (Whittaker and Likens, 1975). The highest estimate of production comes from Bazilevich et al. (1971, 1974), who give 72x 10¹⁵ g C.

Table 5.1 Estimates of world net primary production. (Adapted from Whittaker and Likens, 1973b)

	Amounts estimated x 1015 g
References*	Date	land carbon	sea carbon		total carbon		DM(= C x 2.2)
Liebig	1862				115.9		255
Ebermayer	1882				24.6†		54
Schroeder	1919	16.3	20		36.3		79.9
Noddack	1937	15.1	28.6		43.7		96
Riley‡	1944	10	22102		83		183
Steemann-Nielsen	1958		15
Fogg‡	1958	12	16		28		62
Ryther	1959		53
Müller	1960	10.3	25		35		77
Deevey	1960	56.4	33.4		90		200
Duvigneaud	1962	15.6
Vallentyne	1965	2232	2228		4460		115
Lehninger	1965	16.6	16.6		33		73
Bowen	1966	106	29		135		357§
Whittaker and Likens	1969	49.5	25		74.5		164
Ryther	1969		20		74		162.8
Olson	1970	54
Koblentz-Mishke et al.	1970		23
Bazilevich et al.	1970	78	27		105		233
SCEP	1970	56	22		78		172
Lieth	1972	45.5	25		70.5		155
Golley	1972	40.5	25		65.5		144
Lieth	1973	45.5
Whittaker and Likens	1973	48.2	25		73.6		162
Whittaker and Likens	1975	53.4	25		78.4		172.5
*For references till 1972 see Whittaker and Likens (1973b).
†Based on CO2 consumption.
‡These authors estimated gross production. It is assumed that net production is 50% of gross production.
§Bowen used different factors for the conversion of carbon to dry matter; for land biota 2.16; for sea biota 4.42.

Phytomass estimates are less numerous. The following values are found in the literature and are often quoted: 450 x 10¹⁵ g C (Bolin, 1970); 480 x 10¹⁵ g C (Garrels et al., 1973); 518 x 10¹⁵ g C (Bowen, 1966); 680 x 10¹⁵ g C (Baes et al., 1976), 700 x 10¹⁵ g C (Waksman, 1938), 826 x 10¹⁵ g C (Whittaker and Likens, 1975); 1000 x 10¹⁵ g C (Garrels and Mackenzie, 1972); and 1080 x 10¹⁵ g C (Bazilevich et al., 1971; Rodin et al., 1975). The values given by Waksman and by Garrels and Mackenzie refer to all living matter in the biosphere (i.e. including marine organisms). Bazilevich et al. (1971) calculated the plant mass reserves and annual production for a reconstructed plant cover of the earth, i.e. without correction for agricultural lands, cut forest land, etc., whereby a climax vegetation was assumed for each of the bioclimate zones distinguished. They were primarily interested in evaluating the potential biological resources of the earth. More realistic are the estimates by Lieth (1975) and by Whittaker and Likens (1973a, 1973b, 1975), whereby measured production and phytomass values of various ecosystem types are used for extrapolation.

Table 5.2 Net primary production and other characteristics related to productivity (After Whittaker and Likens, 1973b, 1975, and Whittaker, 1975)

	Net primary production (DM)				Biomass (DM)
		Normal			Normal			Litter	Animal	Animal	Animal
	Area	range	Mean	Total	range	Mean	Total	mass	consumption	production	biomass
Ecosystem type	(1012 m2)	(g/m2 yr)	(g/m2 yr)	(1015 g/yr)	(103 g/m2)	(103 g/ m2)	(1015 g)	(1015 g)	(1012 g/yr)	(1012 g/yr)	(1012 g)
Tropical rain forest	17.0	10003500	2200	37.4	680	45	765	3.4	2600	260	330
Tropical seasonal forest	7.5	10002500	1600	12.0	660	35	260	3.8	720	72	90
Temperate evergreen forest	5.0	6002500	1300	6.5	6200	35	175	15.0	260	26	50
Temperate deciduous forest	7.0	6002500	1200	8.4	660	30	210	14.0	420	42	110
Boreal forest	12.0	4002000	800	9.6	640	20	240	48.0	380	38	57
Woodland and shrubland	8.5	2501200	700	6.0	220	6	50	5.1	300	30	40
Savanna	15.0	2002000	900	13.5	0.215	4	60	3.0	2000	300	220
Temperate grassland	9.0	2001500	600	5.4	0.25	1.6	14	3.6	540	80	60
Tundra and alpine	8.0	10400	140	1.1	0.13	0.6	5	8.0	33	3	3.5
Desert and semidesert scrub	18.0	10250	90	1.6	0.14	0.7	13	0.36	48	7	8
Extreme desert (rock, sand, ice)	24.0	010	3	0.07	0-0.2	0.02	0.5	0.03	0.2	0.02	0.02
Cultivated land	14.0	1004000	650	9.1	0.412	1	14	1.4	90	9	6
Swamp and marsh	2.0	8006000	3000	6.0	350	15	30	5.0	320	32	20
Lake and stream	2.0	1001500	400	0.8	00.1	0.02	0.05	-	100	10	10
TOTAL CONTINENTAL	149		782	117.5		12.2	1837	111	7810	909	1005
Open ocean	332.0	2400	125	41.5	00.005	0.003	1.0		16 600	2500	800
Upwelling zones	0.4	4001000	500	0.2	0.0050.1	0.02	0.008		70	11	4
Continental shelf	26.6	200600	360	9.6	0.0010.04	0.001	0.27		3000	430	160
Algal beds and reefs	0.6	5004000	2500	1.6	0.044	2	1.2		240	36	12
Estuaries (excluding marsh)	1.4	2004000	1500	2.1	0.014	1	1.4		320	48	21
TOTAL MARINE	361		155	55.0		0.01	3.9		20 230	3025	997
FULL TOTAL	510		336	172.5		3.6	1841		28 040	3934	2002

An increased interest in the production of natural ecosystems, particularly stimulated through the various projects carried out within the framework of the International Biological Programme (I.B.P.-News, 1967, 1969), has resulted in a flow of publications, a number of national and international meetings, which in turn resulted in various Proceedings and compilation works, such as Wiens (1972), of the biosphere. Worldwide totals by ecosystem types and for the earth's surface. Young (1968), Human Ecology (1973), Woodwell and Pecan (1973), Reichle et al. (1975), Lieth and Whittaker (1975), Cooper (1975), and works on more specified ecosystem types, such as Eckardt (1968), Reichle (1970), Duvigneaud (1971), Golley and Golley (1972), Farnworth and Golley (1973), Golley et al. (1975), Golley and Medina (1975), Ulrich et al. (1974), Young (1974, 1976). In the study by Lieth and Whittaker (1975) of the history of productivity, methods of measurement as well as patterns in productivity and some application in research are dealt with.

This study is an attempt to evaluate the various data on NPP and phytomass of terrestrial ecosystems found in the literature and to use this data for a new assessment of total production and phytomass, whereby recent results of production studies are taken into account. In particular, more attention is paid to realistic surface area measurements and the relation of organic matter with the environment.

Table 5.3 Various estimates of surface areas of world terrestrial ecosystems. Updated from Golley (1972). (Areas in 10⁶ km²)

										Whittaker
										and	Bazilevich
		Shantz	Deevey	Stamp*	Lieth	Bowen	Schmitt	Whittaker	Golley	Likens†	et al. ‡	This
	Ecosystem type	1954	1960	1960	1964	1966	1965	1970	1972	1973, 1975	1971	study
1.	Forests	53.3	44.41	54.4	43.6	44.0	45.0	50.0	36.0	48.5	64	31.0
	A. Tropical rain	9.7	20.25	19.0	14.7	14.7	20.0	20.0	20.0	17.0	25	10.0
	B. Tropical seasonal	5.2								7.5	9	4.5
	C. Temperate evergreen	1.4	14.6	8.6			6.0	18.0	6.0	5.0	12	3.0
	D. Temperate deciduous	16.6	5.66	26.8	4.9	4.9				7.0		3.0
	E. Boreal	19.5	3.9		10.0	10.0	4.0	12.0	10.0	12.0	18	9.0
	F. Other forests	0.9			14.0	14.4	15.0					1.5
2.	Woodland and shrubland	3.8						7.0		8.5	4	4.5
3.	Savanna	23.0		12.9				15.0	15.0	15.0	15	22.5
4.	Grasslands	9.0	36.9	22.5	25.7	27.0	37.0	9.0	25.0	9.0	9	12.5
5.	Tundra	11.2	8.5	9.9		11.0	9.0	8.0	10.0	8.0	8	9.5
6.	Desert and semidesert (scrub)	27.1	22.4	25.2	52.0	33.0		18.0	25.0	18.0	16	21.0
7.	Extreme desert (rock, sand, ice)	6.2	19.7	12.4		11.0	41.0	24.0	15.0	24.0	23	24.5
8.	Cultivated land		13.31	7.9	14.0	23.0	10.0	14.0	15.0	14.0		16.0
9.	Swamp and marshes		3.3					2.0		2.0	4.0	2.0
10.	Lake and stream							2.0		2.0	2.0	2.0
11.	Human area											2.0
12.	Others								7.0		4.3	1.8
Terrestrial total		133.6	148.5	145.2	148.0	149.0	142.0	149.0	148.0	149.0	149.3	149.3
*Based on world soil groups.
†The estimate of woodland and lake and stream in 1975 differ slightly from 1973.
‡Approximate values. Bazilevich et al. distinguish 106 different ecosystem types which proved to be difficult to group according to the system adopted in this study.

The studies by Whittaker and Likens (1973b, 1975) have formed the basis for the new assessment. Table 5.2 gives the principal data on NPP, phytomass, and related characteristics as presented by these authors. In addition to NPP and phytomass, attention is paid to dead organic matter. Changes in some ecosystem types, induced by man's activities, are discussed in relation to productivity and phytomass. This part is dealt with in much more detail by Bramryd (Chapter 6, this volume).

5.2 CLASSIFICATION OF ECOSYSTEM TYPES

The earth's surface is naturally a mosaic of different kinds of vegetation, associated with different environmental features. All these differences are expressed in net primary productivity variations, which are of vital importance for the self-maintenance or management of the respective ecosystems. The understanding of these variations and their causes is therefore of prime importance for the optimal use of individual types of ecosystems (Lieth and Whittaker, 1975).

In the earlier estimates of total production and phytomass, only a very few vegetation types were distinguished (Lieth, 1975), due to the limited amount of data available. This changed when more data were published as a result of an increase in the number of vegetation maps and an increased interest in the productive dimensions of the biosphere. For example, Bazilevich et al. (1971), and Rodin et al. (1975) give NPP and phytomass data for 106 different vegetation units. Usually, only about 20 units are distinguished (Lieth, 1972, 1975; Whittaker, 1970; Whittaker and Likens, 1973a, 1973b, 1975).

Various classification systems have been designed, based on different criteria, and remain to some degree subjective (Rübel, 1930; Ellenberg and Müller-Dombois, 1967; Schmithüsen, 1968; Schmidt, 1969; Walter, 1973, 1976; UNESCO, 1973.

The terms `vegetation unit', or `ecosystem type', are applied to any grouping of plants and are not limited. They are, therefore, perfectly safe terms to use to designate a band of tropical forest or a marsh vegetation. However, in studying any of these, it is desirable to recognize criteria for comparison. Therefore, in this study, the same classification was adopted as that used by Whittaker and Likens and presented in Table 5.3, which is, in fact, the UNESCO scheme in its modified form. However, some subdivisions have been made in a few ecosystem types, when considered necessary in the light of recent production of surface area data, for instance, the division in several savanna vegetations, and the splitting of tundra biome into polar desert, high arctic, and low arctic tundra. It is beyond the Scope of this paper to give detailed descriptions of each of the ecosystem types. The reader is referred to Ellenberg and Müller-Dombois (1967), Schmidt (1969), Schmithüsen (1968), and Walter (1973, 1976).

It is beyond doubt that, although a classification has been adopted, it still remains almost impossible to draw a sharp distinction between any of these ecosystem types or their subdivisions. Each division is an arbitrary separation. Figure 5.1 represents the distribution of the main biomes of the biosphere. Going from the equator to the North or South Poles there is a zonation of vegetation units, chiefly based on climatic factors. Gradually, one unit is replaced by another (Walter, 1973). Only the tundra and the boreal forest have some continuity throughout the northern hemisphere.

Other biomes of the same type (e.g. tropical rain forests, temperate grassland) are isolated in different biogeographical regions, and therefore may be expected to have ecologically equivalent, but often taxonomically unrelated species.

Whenever possible, every effort has been made to conform to the adopted classification. However, because of incompleteness of the available information, mistakes in classifying local ecosystem terminologies could not be avoided. Within each main biome there exist various opinions as to its subdivisions. The terms `forest', `woodland', `grassland', and `savanna' have proved especially difficult to specify, as each author has his own concept of what they constitute. For example, `woodland' may be applied to a vegetation which lacks a continuous tree canopy, but the total vegetation coverage is continuous, while it can also be applied as a general term for forest (Ovington, 1965). On the other hand, the term `forest' is used for real closed forests as well as for more open woodlands. In this study, the term `woodland' has only been used for a group of unclassifiable, woody types of vegetation in temperate zones.

The definition of savanna is elusive. It involves a discontinuous canopy of trees and a continuous cover of herbaceous plants. The savanna grades into woodland savanna and parklands on the one hand, and into open grassland on the other. The question of how many trees make a savanna, or how many woody plants form a woodland, cannot be answered (Laubenfels, 1975). The savanna is characteristically a tropical grassland, often disturbed by fire, with gallery forest along streams and with scattered groves. Palms are not infrequent, and the scrubby growth tends to be thorny. Typical thorn scrub and thorn forest usually appear adjacent to either savanna, or light tropical forest and often develop out of this type after disturbance, especially overgrazing.

Figure 5.1 Schematic map of the major biomes of the world. (After Odum, 1971. Reproduced by permission of Saunders Co., London.)

The term `grassland' has only been applied to temperate grasslands, which include such vegetation as prairies: Great Plains in North America, steppes in Inner Asia, pampas in South America, and the velds in South Africa. However, it is difficult to separate semi-natural grasslands from the permanent pastures established and managed by man.

`Tundra' is defined as the treeless regions beyond timberlines in the north (Arctic tundra), and on high mountains (alpine tundra) (Webber, 1974). This is the usual definition, which is less limited than the original meaning (Lapland treeless plain). Polar desert, which to some authors is not part of the tundra zone, is also included. Oceanic moorland areas in cool temperate climates, which are sometimes included in the definition, are not treated as tundra in this study. For the various opinions on the arctic and alpine zonations in North America and Eurasia refer to Alexandrova (1970), Barry and Ives (1974), and Blüthgen (1970).

Bogs (peatlands) and `human area' are ecosystem types which have never yet been treated as separate entities. They have been included because for bogs an estimate of their surface area could be made, while it was realized that human area can no longer be neglected, due to the rapid expansion of human settlements, cities, and industries. Urbanization mainly takes place in areas which have high biological productivity, replacing it with low-productive sites, thus diminishing NPP and phytomass.

5.3 ASSESSMENT OF SURFACE AREAS OF THE VARIOUS ECOSYSTEM TYPES

Various estimates of surface areas of the main ecosystem types of the world have been made in the past. Table 5.3 represents a summary of these estimates. The great variation reflects the difficulties in classification, as discussed above, and also the fact that no biome has ever been measured on a global scale. The estimates are often based on figures given by FAO in their various statistic publications, which, in the absence of more accurate information, are the best available source.

The values given by Whittaker and Likens (1973, 1975) are derived from Lieth (1972, 1975) and are assumed to apply to the situation in 1950. In the present study, these values have been updated by making use of recent reports and detailed vegetation maps, as well as various assumptions on human interference in natural vegetation.

All data are given in round figures to avoid the pretention that the new assessment is complete and consistent. Although many new sources have been consulted, some of the sources were incomplete in themselves. In addition, it would have taken a long period of library study to trace and understand all background data of the sources, particularly concerning classification and methodology.

The reassessment of surface areas was mainly based on the following considerations and assumptions: as a result of clearing and fire, the tropical humid forests are gradually replaced by secondary forests, derived savanna, cultivated plots, and roads and settlements. Likewise, dry forest is giving way to savanna-like vegetations, settlements and cultivation, while temperate rain forests are being replaced by mountain grasslands and meadows. Destruction of savanna by fire and overgrazing leads to degradation and a semidesert-like plant-cover, while semideserts turn into deserts through human and climatic influences.

Taking recent estimates for surface areas of forests, such as those given by Persson (1974), Brünig (1977), Synnott (1977), Reinbek-Weltforstatlas, and using many other references, the areas of the remaining ecosystem types were estimated. The new surface areas are presented in Table 5.5.

Persson (1974) has estimated the area of the closed forests in the world as 28 x 10¹² m² or 22% of land area, while the area of open woodlands of different types is closer to 10 x 10¹² m². The area of closed boreal forests is given as 6.7 x 10¹² m². Although definitions of closed forests and open woodland are given, they are rather flexible and Persson (1974) stated that they need to be improved. According to Brunig (1977), the area covered by tropical closed forests and open woodland is roughly 20 x 10¹² m², of which half is closed forest. The term `tropical closed forest' is used synonymously with `tropical moist forest' and excludes the dry tropical forest.

Brünig's estimate was adopted in this study. Other estimates for tropical moist forests are: 16 x 10¹² m² (Budowski, 1956 for humid and monsoon forests together); approximately 9 x 10¹² m² (Persson, 1974); 7.75 x 10¹² m² (UNESCOMAB, 1977); and 9.35 x 10¹² m² (WWF cited in Woodwell et al., 1977). Although it is difficult to judge to what extent the values given by Whittaker and Likens (1975) for 1950 reflect the real situation, they nevertheless form a basis for comparison.

If the present estimate of 10 x 10¹² m² for tropical moist forests is close to the truth, this would mean a decrease of 42% since 1950. Assuming the same rate of decrease for tropical dry forests, the present area can be estimated at 4.5 x 10¹² m², a value which coincides with the estimate of Persson (1974). Our surface area of closed boreal forests is slightly higher than that given by Persson (1974). The surface area for open boreal forests, or forest tundra, is similar to that given by Mikola (1970).

It is difficult to judge the presented surfaces for temperate forests. Both are based on approximately the same rate of decrease since 1950 as for tropical forests. Compared with the data given by Persson (1974), the values appear to be too high.

Figures on surface areas of savanna vegetations are scarce. Malaisse et al. (1975) gives 3.8 x 10¹² m² for savanna woodland in Africa (forêts claires); this includes Miombo (which is thorny forest). Synnott (1977) gives 1.7 x 10¹² m² for Miombo. Hueck (1966) reports 0.85 x 10¹² m² for thorny `forests' (caatinga) in Brazil. Our figure of 3.5 x 10¹² m² does not seem an overestimate.

The total surface area of permanent pastures and meadows given by FAO (1974) is 30 x 10¹² m² , but probably does not include all savanna from Whittaker and Likens (24 x 10¹² m²). The present value is close to that given by FAO. As grasslands, in the widest sense of the word, form the world's second most important ecosystem type, it is evident that more attention should be paid to the classification and mapping of grasslands.

The total peatland area of predominantly temperate regions is 2.31 x 10¹² m² (Moore and Bellamy, 1973), which the authors consider a low estimate. Peatland areas in the tropics cover 0.38 x 10¹² m² (Soepraptohardjo and Driessen, 1976; Pons, 1977). If a correction is made for the overlap in both estimates, the total peatland area is 2.56 x 10¹² m². The largest areas are located in the boreal zone of North America, the U.S.S.R., and of some countries in Northern Europe. The distribution of tropical peatlands is limited (Pons, 1977); they are concentrated in South East Asia, around the Sunda Flat in Indonesia and Malaysia where a wet climate prevails (Pons, 1977).

Many of the peatlands are drained for agricultural use, afforestation or exploitation of the peat for horticultural purposes or fuel (see Chapter 6, this volume). Therefore, it was assumed that 1.5 x 10¹² m² are still unexploited, and mainly consist of bogland.

It might well be possible that there is an overlap between peatlands and the category of swamps and marshes. The surface area for the latter was taken from the literature (Whittaker and Likens, 1975). This value probably needs revision as many marsh areas are being drained, particularly in the tropics.

The estimate of mangrove area was based on area data given by Chapman (1976) for about one-third of the total mangroves. The remainder was calculated by measuring the length of coastlines where mangroves are found and assuming an average width of 80 m.

The subdivision into annual and perennial crops has been made using FAO Production Yearbook data (FAO, 1974). Their figures indicate that perennial crops, such as fruit trees, coffee, tea, and rubber plantations, cover less than 8% of the total cultivated areas. The chosen figures remain rather subjective, as is the division between temperate and tropical.

Compared with the surface area of cultivated land, which is usually given as 14 000 x 10⁹ m², the present area reflects an increase of less than 0.5% per year. This seems very low considering the fact that the increase in food production is chiefly a result of expanding cultivated land and partly a result of increased yield per unit area. However, at the same time, a decrease in cultivated area due to desertification, erosion, or other factors has to be taken into account. According to FAO statistics (FAO, 1974) cultivated land covers approximately 15 000 x 10⁹ m². Much of FAO data is out of date. In addition, it refers only to those plots of which the crops reach the national or international market; smallholdings are probably not included.

The term `human area' has been introduced, defined as the area occupied by man for housing, schooling, utilities, industries, transport, etc., and is not restricted to urban areas. Because of the rapid expansion of urban-industrial development and the large diversity of its impact on the environment, an attempt has been made to assess the extension of total human area. Published data are scarce and, if available, mainly for developed countries. Evdokimova et al. (1976) report that 10% of the forest zone of the U.S.S.R. is occupied by cities, roads, and settlements. In the forest-steppe zone this area is 4%, while in the steppe zone it amounts to 20%. According to Maier-Bode (1959) 7.7% of the land surface in West Germany was already being used as cities, villages, roads, etc., in the 1950s. In the Netherlands, cities, villages, roads and industries occupy 9.2% of the country (Statistisch Zakboek, 1977). In Japan, 48% of the total population lives in built-up areas, which cover 1.25% of the total land area. Of course no extrapolation can be made globally, but these high figures indicate that the surface of human area might be rather high and can no longer be neglected. Population data of about 3000 metropolitan areas with known surfaces (Rand McNally, 1972) has revealed that about 25% of the world population live on 617 x 10⁹ m². Roads and railways probably cover some 180 x 10⁹ m² (World Road Statistics, 1975). Total human area was assumed to be 1.5% of the land surface (without perpetual ice) or 2000 x 10⁹ m². This might be too low an estimate, as a recent and more detailed study gives 1.82% (Ajtay, unpublished). Such a relatively small area has not, of course, a great influence on our estimate of NPP and phytomass. 

5.4 ASSESSMENT OF NET PRIMARY PRODUCTIVITY AND LIVING PHYTOMASS 

All data on productivity, phytomass, litter, and soil organic matter are commonly expressed in grams of dry matter, grams of carbon, or sometimes both. The relationship of dry matter to carbon is variable. Woody parts and roots normally have a higher carbon content than foliage. During the process of decomposition, the carbon content can change. Kimura (1963) found that the carbon content in the needle litter remained fairly constant with the progress of decomposition (52.7%), while for branch litter it gradually increased up to 58%. Table 5.4 shows the variation in carbon content of plant parts, and of different components in forest ecosystems. For reasons of comparison, the following conversion factors were used in this study: NPP, phytomass, and litterfall 45% (Lieth and Whittaker, 1975; Larcher, 1976); litter 50% (Kimura, 1963); and humus 58% (Waksman, 1938; FitzPatrick, 1974).

For the assessment of net primary productivity and phytomass, the values given by Whittaker and Likens (1975) were evaluated and adjusted in the light of much new data (mainly from IBP-studies) published in various forms. Earlier and recent compilation works dealing solely or partly with productivity were consulted, such as Lieth (1972), Rodin and Bazilevich (1967), Young (1968), Wiens (1972), Cavé (1974), Van Dobben and Lowe-McConnell (1975), Cooper (1975), Lieth and Whittaker (1975), Reichle et al. (1975), as well as many studies on specific ecosystem types. For tropical ecosystems, the following references are mentioned:

Table 5.4   Average carbon content of plants and parts of plants on DM-basis. (Number of samples given in parentheses)

Misra and Gopal (1968), Golley and Golley (1972), Farnworth and Golley (1973), Golley and Medina (1975), and Golley et al. (1975). Important studies on forest production include Reichle (1970), Duvigneaud (1971), Ellenberg (1971), Ulrich et al. (1974), Young (1974, 1976), and Brünig (1977). Valuable data on grasslands can be found in Coupland and Van Dyne (1970), French (1971), Breymeyer (1971), Rychnovska (1972), Coupland (1973a, 1973b), Whyte (1974), César and Menaut (1974), Numata (1975) and Pandeya et al. (1977). Useful data on tundra ecosystems are available in Alexandrova (1970), Webber (1974) and Wielgolaski (1975a). It is impossible to list all the smaller publications which were consulted, but a few should be mentioned, such as Klinge (1973a, 1973b, 1973c, 1976), Klinge and Rodrigues (1973), and Lemée and Huttel (1975). 

Since the data were drawn from so many different sources, a considerable variation in methodological approach was to be expected. All data from literature was converted to dry matter and to total production and phytomass, i.e. above plus below ground. When there was a questionable figure, the orginal paper from which data were quoted were consulted; locations were checked on vegetation maps when there was any doubt about classification.

The adjusted means for NPP and phytomass per ecosystem type are given in Table 5.5, together with total production and total phytomass; the latter values were calculated for the new estimates of the surface areas. In some cases, the means are approximate averages of published values (forest, grassland, savanna forest, tundra), but in many other cases they have been chosen subjectively as possible values (Whittaker and Likens, 1975). It proved particularly difficult to place the typical local vegetations, of which NPP and phytomass data were available, into the adopted classification. Compared with the values of Whittaker and Likens, the main differences in NPP data are for savanna, temperate grasslands, and tundra, which are all higher. Forest productivity seems only slightly higher than was previously reported.

Human area is not entirely unproductive, for parks, small woodlands, gardens, or other quarters with vegetation still fix carbon by photosynthesis. The nonproductive site is an energy-consuming system in the form of fossil fuels, thereby releasing large quantities of carbon to the atmosphere, rivers, and oceans. It was assumed that 40% of the total human area is still productive. According to Abrams (1965), 18% of the central-city part of some North American cities was open space. He found a positive relation between increasing population, occupied area, and open space for some of the cities. Duvigneaud et al. (1977) found for Brussels that up to 50% was open space, with a relatively high productivity.

The phytomass values per unit area are meant to be annual average values. However, due to limited data on phytomass and its fluctuation per year, the presented values are probably less accurate than the NPP values. In the case of annual crops of cultivated land, an average phytomass was calculated by dividing the NPP by 12 (the number of months per year). Maximum standing crop usually applies to the time of harvest and is slightly lower than NPP due to dying of plant parts as the growing season proceeds. Even if maximum standing crop was taken as an average, it would not have changed the total phytomass by very much. The new estimates for total net primary production and total living phytomass are 133 x 10¹⁵ g dry matter and 1243.9 x 10¹⁵ g respectively (i.e. 60 and 560 x 10¹⁵ g C).

The main share in the total production has the forest ecosystems, followed by the different savanna systems, cultivated lands, and temperate grasslands. The total production value is more than 10% higher than the estimate of Whittaker and Likens, and higher than any of the earlier estimates (see Table 5.1), except that of Bazilevich et al. (1971). The much higher values of savanna/grassland for NPP are explained by the fact that in the past years more attention has been paid to root production in these ecosystems. Root production appears much higher than was originally supposed (Numata, 1975; Pandeya et al., 1977). However, no conclusion may be drawn about changes. Not only do the surface areas differ, but the estimates for NPP and phytomass sometimes differ considerably from author to author; in addition, there is the difficulty of classification.

Table 5.5   Surface areas, net primary productivity, and phytomass of terrestrial ecosystems of the biosphere. Adopted conversion factor from DM to carbon is 0.45

To make a better comparison between the new assessment and that of Whittaker and Likens, the values for NPP of Table 5.5 were multiplied with the surface areas used by Whittaker and likens (column 2, Table 5.2), assuming NPP of natural ecosystems has not changed since 1950. For the agricultural systems, only a slightly higher average NPP was taken (800 g/m² ). Calculated in this way, total net production amounts to 142 x 10¹⁵ g dry matter or 64 x 10 ¹⁵ g C. Although there still remain many incongruencies which bias any comparison, e.g. some ecosystem types are not listed by Whittaker and Likens, it could be concluded that total net primary production has declined since 1950. However, this cannot be proved.

The total amount of phytomass (560 x 10 ¹⁵ g C) is much lower than the estimate by Whittaker and Likens. Major differences exist between the estimates of phytomass per unit area of the forest, the savanna, and the tundra. The former are somewhat lower than those of Whittaker and Likens, while the latter are much higher. Forests contribute up to 75% of the total phytomass, followed by woodlands and savanna. The above comments on comparison of the data also apply to the phytomass data. Nevertheless, it seems reasonable to conclude that, compared with the data of Whittaker and Likens, total phytomass is likely to have decreased over the past decades. This decline can chiefly be traced to the continuous decrease in forest-phytomass as a result of intensive exploitation and clearcutting. The cleared areas are replaced by systems with a much lower phytomass per unit area, so that the loss is only partly compensated for. Compared with other estimates, the present estimate for total phytomass is higher than that given by Bolin (1970) (450 x 10¹⁵ g C), and by Bowen (1966) (510 x 10¹⁵ g C), which makes it all the more difficult to draw conclusions about changes.

The values for total NPP and phytomass depend mainly on the extrapolation on single data or sets of average values of relatively small research plots to large uncertain surface areas, and they are, therefore, burdened by a great degree of inaccuracy. Existing vegetation maps of the world or individual continents present the opitmal situation, without taking human interference into account. On the other hand, large-scale vegetation maps, representing detailed local vegetation units, are not suitable to be used for extrapolations to world scale but could be used for NPP and phytomass estimates of regions (Sharp, 1975; Sharp et al., 1975). Without a good international usable classification of ecosystem types, and without maps representing the real situation which can be used for obtaining real surface areas, it remains impossible to make estimates about NPP production and phytomass with a certain degree of accuracy.

As the plant cover of the earth surface is rapidly changing due to man's interference, continuous monitoring is needed, which could be done by satellite and ground observations. In addition, more field observations about NPP and phytomass of natural ecosystems are needed. In preparing the present study, we mapped each location of which NPP and phytomass data were available. We must conclude that, despite the efforts of the IBP, there still exist large gaps of knowledge on productivity of ecosystem types, particularly in South America, Asia, Africa, and Australia.

5.5 DEAD ORGANIC MATTER

The dead organic matter in land biota consists of the total amount of organic matter incorporated in trees and shrubs that are dead but still standing, of dead organs (dry but not yet fallen branches of trees and shrubs, dry stems of herbaceous plants), the matter accumulated as litter, steppe matting or turf horizon of the soil, and the organic matter accumulated in the soil as humus (Rodin and Bazilevich, 1967).

5.5.1 Standing Dead

Limited data are available on the amount of dead plant parts, still attached to living plants. Data are mainly available on forests and tundra ecosystems. Table 5.6 lists some values of above-ground dead phytomass in various ecosystems. Compared with the values for total living phytomass (Table 5.5) the standing-dead component is only very small, except for savanna forest where it can equal about 25% of the living phytomass. The data are too scanty to make any extrapolation, but it seems reasonable to assume that total standing dead, on average, equals about 5% of the total living phytomass, i.e. about 60 x 10¹⁵ g DM. With an average carbon content of 50%, this would mean 30 x 10¹⁵ g C, a value which is much lower than that given by Bazilevich (1974) for standing dead and dry trees, namely 75 x 10¹⁵ g C. 

5.5.2 Litter and Litterfall

The term `litter' is used in ecology with the following two meanings: the layer of dead plant material, which may be present on the soil surface; and dead plant materials which are not attached to a living plant. These are not, however, satisfactory as definitions for the ecologist concerned with the functioning of ecosystems. The litter layer may be clearly distinguishable from an underlying mineral layer or there may be no sharp boundary between a layer containing recognizable plant structures and a layer containing only amorphous organic material.

The presence of tree trunks, in the English Pennines, 7000 years old, under peat 4 m deep, illustrates the problem of defining a litter layer by the same criteria in all habitats. Also, tree trunks above ground are a form of litter, but are often treated separately.

The problems are no less severe when litter is defined as dead plant materials which are not attached to a living plant. Plant organs neither die instantly nor, when dead, fall instantly. Abscission of a leaf follows a more or less prolonged senescence when much of the mineral content is withdrawn to the stem and the phylloplane fungi are already decomposing the carbohydrates. The argument may be extended to recognize that there is a turnover of molecules in all living matter, and that death begins prenatally.

Table 5.6 Average standing crop of dead phytomass in some ecosystem types. Values in g/m² dry weights. (Number of measurements in parentheses)