The Effects of Increased CO2 and Climatic Change on Terrestrial Ecosystems
|Global Perspectives, Aims and Issues|
|R. A. WARRICK, H. H. SHUGART, M. JA. ANTONOVSKY, WITH J. R. TARRANT AND C. J. TUCKER|
|8.1 TERRESTRIAL ECOSYSTEMS, CLIMATE AND MAN|
|8.2 FOCUS AND AIMS|
|8.3 CO2 AND CLIMATIC CHANGE: INFORMATION FOR IMPACT STUDIES|
|8.3.1 Scenarios of Climatic Change|
|8.3.2 Applications and Improvements|
|8.4 GLOBAL AGRICULTURE AND CLIMATIC CONSTRAINTS|
|8.4.1 Semi-arid Tropics|
|8.4.2 Humid Tropics|
|8.4.3 The Temperate Regions|
|8.4.4 Global Food Trade|
|8.4.5 Implications and Issues|
|8.5 GLOBAL FOREST ECOSYSTEMS AND CLIMATIC CONSTRAINTS|
|8.5.1 Climatic Change and the Areal (Macroscale) Response of Forests|
|8.5.2 Local Changes in Forest Composition|
|8.5.3 Possible Impacts on Forest Productivity|
|8.6 SOME FURTHER IMPLICATIONS OF ECOSYSTEM CHANGE: SUMMARY AND CONCLUSIONS|
|NOTE ON AUTHORSHIP AND ACKNOWLEDGEMENTS|
It is safe to begin with the premise that a climatic change within the range indicated by climate models for a doubling of atmospheric CO2 could have profound effects on global ecosystems. A cursory, global survey of natural systems reveals an unmistakable correspondence between the broad features of regional climates and the major characteristics of the world's biomes. The transitions from tundra, boreal forests, temperate forests, deserts, grasslands and tropical forests vary systematically with global temperature and precipitation patterns and produce the areal distributions illustrated in Figure 8.1. That climate is, in fact, a primary determinant of the composition and spatial patterns of the major biomes is beyond doubt, although certainly there are feedbacks in the other direction, as through the albedo or hydrologic characteristics of regional plant assemblages and associated soil types. Major changes in the global climate could be expected to alter natural rates of ecosystem change in and between these biomes, particularly within the marginal zones of transition.
If, indeed, the changes in climate that may result from increasing concentrations of greenhouse gases fall in the middle range of estimates provided by climate models (1.5 to 5.5 °C for a CO2 doubling, as discussed in Chapter 5), we will experience a global climate that is probably at least as warm, or warmer, than at any time within the last 200,000 years. During that time, there were a number of fluctuations of climate, each of which was associated with geographical patterns of terrestrial ecosystems that were markedlyin some cases dramaticallydifferent from those evident today. For example, during the Medieval Warm Epoch (800 to 1200 AD) when average temperatures were perhaps l °C warmer than present (at least around the region of the North Atlantic), Canadian boreal forests extended well north of the current timberline and cereal cultivation flourished in Iceland and Norway up to 65º N latitude (Lamb, 1977). The warm period of the early Holocene (9000 to 6600 BP) witnessed extraordinarily wetter conditions in the vast subtropical dry zones extending from West Africa to the Indus ValleyRajasthan area; thriving savannah grasslands existed in large areas that are now unproductive desert (Hare, 1979; Flohn, 1980). The last glacial period (70,00010,000 years ago) was associated with tundra in Central Europe and with a drastic shrinkage of the tropical rainforest, which quickly expanded with post-glacial warming. The earliest warm peak (Eem) of the last interglacial (around 120,00080,000 years ago) was perhaps 23 °C warmer in the mid- to high-latitudes than today, and broad expanses of deciduous forests extended north into areas now occupied only by nondeciduous species (for reviews, see Flohn, 1980; Gerasimov, 1979). Solely on the evidence of the distant past, the potential for ecosystem change in a warmer future is enormous.
Figure 8.1 The global distribution of major ecosystem types, based largely on
Whittaker (1970) (from Bolin, 1980)
Past changes in climate were also accompanied by variations in the atmospheric concentration of CO2 . However, it is not clear to what degree these CO2 variations were involved directly in global-scale shifts of terrestrial ecosystems in the past, if at all, notwithstanding the fact that in experimental situations higher CO2 concentrations have been shown repeatedly to stimulate plant growth and productivity. In this case, the distant past provides few clues for the future.
A fundamental difference between changes in global ecosystems of the past and those of the future is the dominating influence of human intervention in the natural environment. Just as human activity may be responsible for altering the state of the global climate, so, too, are humans capable of manipulating the global biota to a considerable degree. Deforestation by man was a major contributor to past increases in the atmospheric CO2 concentration, as discussed in Chapter 3; deforestation continues at an alarming rate in many areas of the tropics. In only a matter of centuries, one human activity, agriculture, has added another major terrestrial 'ecosystem', largely at the expense of grasslands and forests. This substitution still continues. Today, cultivated land (excluding pasture and grazing land) occupies about 10% of the world's land surface, an area approximately equivalent to one-third of that occupied by forests (Figure 8.2).
The capacity for human interference may be appliedpurposefully or inadvertentlyto counteract, retard or accentuate impacts on global biomes that would otherwise occur naturally. For instance, there is strong evidence to suggest that overgrazing and other human activities that are destructive to vegetation have exacerbated the processes of desertification in the subtropical semi-arid lands (Hare, 1983; 1984). Although the future direction of this influence is difficult to foresee, man's role in the process of ecosystem change is so large that any thoughtful assessment of the impacts of increased atmospheric CO2 , other trace gases and future climatic change must eventually take human management and response into account.
Figure 8.2 World land use in 1977 (From Crabbe and Lawson, 1981)
For this reason, we find it difficult to speak of `managed' and `natural' ecosystems in the context of global climatic change. It can be argued that no terrestrial ecosystem is left untouched by the human hand; the ubiquitous nature of the global CO2 increase is the case in point. It is perhaps more meaningful to refer to a continuum of management, from least managed to most managed. Certainly, agricultural systems fall at the latter end of the spectrum. From a global perspective, forest ecosystems tend to fall toward the other end of the spectrum (although one can cite examples of almost any degree of management) and have important implications for the global carbon cycle as well as economic activity. In Chapters 8, 9 and 10 we thus focus specifically on agriculture and forest ecosystems as representative of the management spectrum, hoping to achieve balance while remaining mindful that many more interrelated environmental systems and resources (e.g. the hydrologic cycle) are bound to be affected as well.
In the pages that follow, the human element causes us to tread a difficult path. Our charge is to address the issue of the 'primary' impacts of CO2 and climatic change on ecosystemsthat is, plant response, yield effects, changes in ecosystem composition or areal extentleaving 'secondary', or social and economic, impacts for subsequent scrutiny. However, especially for agriculture (which, after all, by definition is an economic activity) we feel no compunction in occasionally crossing interdisciplinary boundaries in order to incorporate relevant socio-economic factors.
The basic premise of this and subsequent chapters is that at present we do not have available reliable estimates of future climatic change at a regional scale which would allow us to predict changes on global ecosystems. Nevertheless, this should not deter the scientific community from examining the sensitivity of such systems to changes in climatic variables. This requires an explication of the range of approaches which have been, or could be, taken to do so. Our aim is just
to present the major issues and research questions;
to outline the methodological approaches by which they could be addressed and to exemplify each through existing studies;
to summarize what is known concerning the sensitivity of ecosystems to increased CO2 and climatic change;
to suggest promising areas of needed research or methodological refinement.
In the remainder of this chapter we provide an overview of the major issues and methodological considerations in assessing impacts in the global context as a prelude to more detailed discussions of agriculture and forest ecosystem in Chapters 9 and 10, respectively. What we cannot provide, unfortunately, are estimates of future changes in ecosystems with a degree of confidence sufficient for informed policy formulation and strategy choiceonly possibilities.
There are two ways in which ecosystems can possibly be affected by the rising levels of greenhouse gases, and they are not necessarily mutually exclusive. The first is the direct effect of higher ambient CO2 concentrations on plant growth and development. The direct CO2 effect has been the subject of numerous research investigations, particularly on agricultural crop species. Understanding of plant response has come about primarily by controlled greenhouse experiments and, to a far lesser extent, by the use of crop-growth simulation models based largely on such experimental results. There are very few data concerning the effects of enhanced CO2 in actual field conditions. Any direct effects on ecosystems from increasing concentrations of atmospheric CO2 in this century have yet to be detected convincingly, although some scientists believe that the productivity of natural vegetation and, especially, grain yields has increased and will increase considerably in the future (e.g. Budyko, 1982).
The literature has been extensively reviewed and evaluated (e.g. Lemon, 1983),
most recently by the U.S. Department of Energy (1984). Overall, experimental
research consistently demonstrates that, directly, CO2 is
in terms of crop yields and possibly forest productivity. In general, higher
ambient CO2 stimulates greater net photosynthesisthe so-called
'fertilization' effectand decreases transpiration through a partial stomatal closure, resulting
in greater water use efficiency in plants, at least at the microscale. However,
the extrapolation from individual plants to dynamic ecosystems is highly
tenuous. There are considerable differences between plants regarding their
response characteristics. Competition between plants, and the interactions
between plants, animals and microbes, are likely to change. For some organisms
in terrestrial ecosystems, the net effect of higher CO2 concentrations in the
long-term may be negative, perhaps fatal.
The second way in which ecosystems can be affected by increasing concentrations of greenhouse gases is through changes in climate. Here, too, uncertainty abounds: in relation to the changes in regional climate, to their effects on plants and ecosystems, and to the ways in which enhanced CO2 will modify the effects of climate on plants. As discussed in Chapter 5, the extent of our knowledge concerning changes in climate can only be expressed confidently in terms of averages at the global scale. Confidence drops precipitously as spatial resolution increases. Currently, GCMs can replicate the major features of the general circulation but are unable to produce realistic simulations of present-day climate at a regional level that is appropriate for impact analysis (at least 200300 km for agriculture). Furthermore, higher confidence can be placed on GCM predictions of changes in temperature than on changes in precipitation or other climate variables like humidity or radiation. In most cases this situation precludes the possibility of actually predicting the impacts of climatic changes on ecosystems.
8.3.1 Scenarios of Climatic Change
In lieu of predictions, scenarios of changes in regional climate can be used to investigate the sensitivity of ecosystems to climatic variation, and to test and refine the methods of impact analysis. Three approaches have been followed in scenario development (WMO, 1985). The first is to use GCM simulations of the effects of CO2 on climate, despite the lack of realism at the regional scale. The major advantage of GCMs is that, potentially, all required meteorological variables can be generated with true global coverage. Notwithstanding the problems, wide use has already been made of some GCM CO2 experiments for impact analysis (see Sections 9.4 and 10.4 below).
The use of past climate data is another way of generating scenarios. Palaeoclimatic reconstructions of past warm periods (e.g. the Altithermal) have been suggested as analogues for a CO2 -related climatic change (e.g. Kellogg and Schware, 1981), but they, too, suffer from lack of regional detail, among other things. Lately, work has focused on the use of recent changes in climate observed in the instrumental record as analogues for the future (Lough et al., 1983). Although the magnitudes of climatic changes during the past 100 years are considerably less than those suggested by 2 x CO2 model experiments, they can be used directly as scenarios of climatic conditions en route to doubled CO2 . The use of instrumental data in this case means that internal consistency is virtually assured and that scenarios with fine spatial resolution can be developed. This represents a clear advantage over other methods. However, one must remain suspicious of the realism of such instrumentally based scenarios as CO2 predictions (if such claims are made), since the spatial character of climatic change may not be independent of the method of forcing (see Chapter 3). Moreover, the dependence on observational data means that, for many parts of the world, the length, availability or reliability of records of critical meteorological variables may not be adequate for scenario construction.
The third approach is simply to select arbitrary changes in climate variables. For example, a frequently used scenario is a 1 °C temperature rise and a 10% decrease in precipitation, since this has some correspondence to general climatic changes suggested by GCMs for the not-so-distant future, particularly for North America. The approach is recommended by its simplicity and ease by which it can be uniformly applied in comparative analyses. The main disadvantages are that changes in climate are unlikely to be seasonally constant, and that the arbitrary changes may be inconsistent with local meteorological conditions and observed correlations between climate variables.
8.3.2 Applications and Improvements
In impact studies all three kinds of scenario have been employed, but in a somewhat indiscriminate, haphazard fashion. This is not surprising given the diverse nature of impact studies and the research proclivities of impact analysts themselves. The consequence, however, is that existing studies and methodological tools are difficult to compare and evaluate.
Ultimately, the solution to this problem is the improvement of GCMs to the point where detailed, regional climates can be simulated with realism. For analyses of impacts on agriculture and forests in both the high and low latitudes, this requires knowledge of possible changes in precipitation patterns as well as temperature. But this is a long-term goal and vast improvements of this nature are not foreseeable within the next 510 years, if ever. In the very long term, it may eventually be possible to develop interactive models capable of describing the interplay between ecosystems and climate. In the meantime, the situation could be improved by the development of sets of regional scenarios of climatic change that could be uniformly applied, thereby increasing the credibility and comparability of climate impact studies.
As models for impact analysis improve, it is becoming increasingly apparent that climate scenarios must provide information on the frequency distributionsthe variability and extreme events, as well as the central tendenciesof key meteorological variables. This holds for forest ecosystems as well as agriculture. Short-term occurrences of unusual climatic anomalies during specific stages of the life cycle of forest stands, for example, can have far-reaching, long-term consequences for future forest composition and productivity (Chapter 10). This is certainly the case for agriculture in which changes, for example, in the frequencies of early- or late-season frosts can affect average yield trends, both through the direct impacts on plant growth, and indirectly by, say, the beneficial effects of killing overwintering pests (Chapter 9).
The prospect of developing and applying scenarios of climatic change for impact analyses at the international scale is daunting, to say the least. We must begin by outlining the dimensions of the problem from the global perspective. Let us consider agriculture first.
The genetic foundation of modern world agriculture is surprisingly narrow. As pointed out by Swaminathan (1979), from the beginning of the Neolithic the domestication of plants and animals has involved a selective filtering of the species (and a reduction in the number of their varieties) upon which mankind depends for agricultural production. On a worldwide basis, there are only 30 crop species whose individual production exceeds 10 million tonnes (Mt) annually. Cereals account for over half of the world's arable land use, and only three cropswheat, maize and riceaccount for 80% of total cereal production (FAO, 1983a). Similarly, only two animal products, beef and pork, make up approximately three-quarters of the world's total animal production.
At local and regional scales, it is unclear whether this 'simplification' increases the vulnerability of agricultural production to climatic change. Some have expressed fears that this indeed may be the case (Swaminathan, 1984) and have argued for a greater effort to introduce species or varieties that are better suited to local environments as, among other things, a hedge against climatic change and variability (NAS, 1972; 1976). On the other hand, new varieties from crop research institutes tend to be more adaptable over a broad range of climatic conditions than the traditional varieties (for rice, say, there are thousands of varieties) that are being replaced and that are often very local in their use. It is possible, therefore, that plant research has actually made it easier to adapt to changing climate (Wortman and Cummings, 1978).
At the global scale, the differences between agricultural regions are immense, which is one reason why total world food production is remarkably stable from one year to the next. The differences pertain not only to environmental and climatic characteristics, but to levels of economic development, technology and human living standards as well. These regional differences are vital to modern patterns of agricultural production, so much of which is based on comparative advantage production and trade. For example, the same major cereal crops are grown and/or consumed in countries as diverse as the USA, Australia, Spain, Western Europe and the USSR; differential regional effects of increased CO2 and climatic change could tip the balance, with worldwide repercussions. At the broadest scale, the regional differences are most pronounced between the tropics and the temperate regions and their semi-arid and humid zones.
8.4.1 Semi-arid Tropics
Within many areas of the tropics, agricultural production is low and unstable compared to the temperate regions of the world. To a large extent, this is due to the broad problems of underdevelopment, struggling economies and low levels of agricultural technology that are prevalent in the low latitudes. It has been argued often that the potential for increased production is very large given higher (and more expensive) levels of input (e.g. for a recent analysis see Shah and Fischer, 1984). Technology, however, is only one factor among a host of socio-economic and environmental constraints.
The climate of the tropics influences the patterns of agricultural activity and contributes considerably to the persistent agricultural problems. In general, rainfall is the climate variable of primary importance in shaping the spatial and temporal variations of agriculture in the tropics. Temperature is secondary but becomes increasingly important further polewards.
This is clearly evident in the semi-arid tropics, a region that contains only 13% of the world's lands, 15% of its people (it includes 48 less developed countries) and a small proportion of its food production; sorghum and millet are the principal crops (Swindale et al., 1981). Here the relationship between rainfall and agriculture is finely tuned: the seasonal cycle of rainfall directly determines the tempo and rhythm of agriculture through its limitation on the length of the growing season (Mattei, 1979). The rainy season is the growing season and, over much of the tropical world, is synonymous with the monsoon. Unlike the semi-arid temperate regions, where rainfall is more evenly distributed over the year and where soil moisture reserves can accumulate, the semi-arid tropics receive rain only during concentrated months of the year. The planting date depends critically on the first rains, and delays in the planting date have major effects on end-of-season yields. Thus, even relatively small variations in the amount and timing of rainfall cause high variability in interannual yields.
In many areas, changes in climate could magnify existing problems. Traditional cropping patterns throughout the semi-arid region are finely adjusted to the spatial characteristics of climate (Mattei, 1979). Crop varieties, planting dates and management practices vary markedly along environmental gradients in accordance with climatic expectations. In many cases, even minor alterations in the rainfall regime could disrupt this delicate adjustment and have major repercussions on agricultural productivity. In addition, the inherent environmental variability, along with the limited reserves and access to capital, means that traditional agriculturalists place high priority on minimizing the risks of loss in dry years in lieu of maximizing their gains in the wet years (Swindale et al., 1981). This, in turn, affects long-term average yields and production. Increases in the frequency of dry years could further consume the limited resources of semi-arid farmers of the tropics and have serious impacts on production in countries that already (and increasingly) depend on food imports and suffer from problems of malnourishment. The current plight in Africa of high drought incidence, desertification, high population growth and declining per capita food production may be symptomatic of this process. If, on the other hand, the frequency of dry years decreases in the future, many of the problems could be ameliorated. The direction of precipitation change is highly uncertain.
8.4.2 Humid Tropics
In the humid tropics, rainfall is also a major growth-limiting factor due to its variability and the high potential evapotranspiration. The close dependence of the growing season on rainfall throughout the entire tropics is depicted schematically in Figure 8.3. Food production depends not only on the timely appearance of the monsoon, but also on its strength and reliability throughout the growing season. Indeed, late heavy rains are often just as disruptive as the late arrival of rain, causing flood damage to established crops at a time when it is too late to replant. This applies even to rice which has high water requirements.
Rice is the principal crop of the Asian humid tropics and is the staple food for perhaps 60% of the world's population (van Keulen, 1978). Production of rice in lowland locations is suited to the humid tropics, particularly if well-developed water control and/or irrigation networks are in place. Nevertheless, even lowland rice is susceptible to rainfall variability. Tanaka (1978) has shown a positive relationship between rice yield and rainfall throughout monsoon Asia. The relationship is strongest in areas where water control and irrigation are least developedIndia, Burma, Bangladesh, Republic of Korea. The effects of rainfall variability are seen clearly in Sri Lanka where marked differences in rice production occur over short distances between the wetter and drier zones; the Dry Zone is chronically hindered by rainfall fluctuations, which discourage further investment in water control, create financial insecurities and accentuate the impacts of climate (Gooneratne, 1978).
Under optimally irrigated conditions, radiation and temperature influence rice yields through their effects on photosynthesis and respiration rates, respectively (van Keulen, 1978). For instance, in Japan (a temperate region), irrigation has greatly reduced the susceptibility of lowland rice to rainfall variation (Toriyama, 1978). Rather, coolness hampers yields, especially in the northern areas of Japan, a problem which is being solved in part through crop breeding for cool temperature tolerance (Satake, 1978; Nakagawa, 1978).
Expanding use of marginal lands may be increasing vulnerability to climatic change. In most of the developing countries of the low-latitudes, aggregate food production has managed to keep abreast of population growth over the last few decades. Per capita food production has risen slowly, at about 0.4% per year (Pino et al., 1981), with a few exceptions (like large portions of Africa) where actual declines have occurred. In some areas, yield increases have been responsible for the gains, as in the case of rice production throughout most of south and south-east Asia, or in India where higher yielding varieties of wheat have been adopted. However, in many regions of the developing world, the expansion of total land under cultivation has mainly accounted for production gains. In other areas, a switch to more intensive cultivation practices or crop type has taken place. These areas of new production or rapidly changing agricultural practices often tend to be marginal lands (the best lands are taken first) which are less productive and more susceptible to climatic fluctuations.
Figure 8.3 A schema of the relationship between mean annual precipitation and growing season length in tropical climates (from Newman, 1977)
For example, Fukui (1979) argues that in South Asia there are three types of agricultural pursuits involving annual crops other than lowland rice that are particularly susceptible to climatic variations. These are currently widespread in the humid tropics and are expanding, and include: (1) shifting cultivation in which traditional fallow cycles are shortened, leading to declines in soil fertility and erosion; (2) continuous cropping (of maize, sorghum, upland rice, cassava, peas, beans, etc., primarily for subsistence); and (3) cultivation of feed crops for export, an activity which has increased enormously in recent years (e.g. maize production in Thailand). These activities are typically relegated to upland areas that are subject to soil deterioration and that are hydrologically marginal when intensively utilized for agricultural production. Population increases, resettlement programmes, urbanization, expanding markets for feed crops and other factors are encouraging greater use of the upland areas. These processes thus may be accelerating environmental degradation and increasing the risks from climatic variability (Fukui, 1979).
Moreover, there is a strong social, political and economic component to the increasing agricultural occupation of marginal lands. The pressures to relocate are felt greatest by the poorer cultural groups in many parts of the developing world, those who stand on the fringe of the growing market economies. The better lands are often converted to cash crop production by those with capital and expertise to do so, while the disadvantaged are forced to accept less than environmentally favourable locationsa 'marginalisation' process (O'Keefe and Wisner, 1975; Susman et al., 1983). Thus, in considering the notion of marginal land, one is forced to consider 'marginalised' people: `where' and `who' are inextricably bound.
8.4.3 The Temperate Regions
In this and other respects, the temperate regions of the world differ sharply from the tropics. The mid- and high-latitude zones are the centres of grain production other than rice. North America, Europe, the USSR and China account for over threequarters of the world's maize andalong with Argentina and Australia in the Southern Hemisphere-wheat production (USDA, 1981). The potential climatic range of most of the important crops grown in the temperate regions is quite large. Wheat and potatoes, for instance, can be grown in any state of the USA (Wittwer, 1980).
In the mid- and high-latitudes, temperature is an important factor in shaping the spatial and temporal variations in agriculture. In contrast to the tropics, the growing season in the temperate zones is generally defined by temperature. Polewards, the geographic extent of crop production is ultimately set by temperature, and the length of the frost-free growing season determines the spatial limits of various agricultural activities. One research issue is whether a warmer climate would allow the geographic extension of cereal production into higher latitudes and elevations, as in expanding the northern limits of wheat and barley production in Canada (Williams, 1975). In areas further equatorward, a warmer climate might present opportunities for an additional crop per year.
But this is not to say that precipitation does not exert a strong, often dominant, force, particularly in the drier, mid-continental locations. In the USA, dry-land wheat growing merges into livestock grazing at its western-most boundary as a result of low and variable precipitation. In drier areas of the US Great Plains, wheat growing is only feasible with groundwater irrigation, as in the High Plains of Texas. Even in the northern cereal growing regions of the USSR and Canada where average temperatures define the spatial extent of crop production, crop yields within the core region appear to be most sensitive to soil moisture conditions (Stewart, 1985). There, the principal yield response of a change in temperature is often through the effect on evapotranspiration and, hence, soil moisture availability. Thus, in the mid-to high-latitudes, temperature and precipitation have complex and interactive effects on crop production.
The role of climate in the interannual variability of grain production is a major concern. There are some major differences between regions. For instance, the large interannual variability in total wheat production in the Soviet Union is weather-dominated, whereas in the United States the fluctuations are smaller and explained largely by government policy regarding land utilization (Tarrant, 1984). Yield variations at smaller scales are related to climate in complex ways, as attempts at modelling cropclimate relationships have shown (see Chapter 9.4). Large annual fluctuations in region-wide grain yields can result from specific short-term weather anomalies occurring during critical periods of plant growth. For example, in the North American corn (maize) belt, flowering and pollination occur in July, during which time the crop is particularly susceptible to high temperatures. High temperatures can promote premature flowering and the tassels throw their pollen before the silks are formed, with the result of poorly formed ears and low yields. Such was the case in 1980 (McQuigg, 1981). In 1974, the corn yield suffered a large drop from late-season frost, another threat to regional production.
In contrast to the tropics, the large increases in total production during
recent decades in the temperate, grain-producing regions have resulted largely
from intensification, rather than from expansion of cropped area. (An important
exception has been the opening of new lands for the extension of wheat
production into Kazakstan and Siberia in the Soviet Union.) Development of new
grain varieties, irrigation, increasing applications of fertilizers and
pesticides, research and improved managementbroadly,
contributed to significant increases in average yields, particularly in the
post-war period. In the United States, for example, maize yields have
approximately tripled and wheat yields have doubled since 1950. Some countries
in Europe have reached average wheat yields that are two and three times those
produced in North America (USDA, 1981). For the United States, the
government-supported technological advances have created a production capacity
that has outstripped demand. A chronic problem has been, and continues to be,
one of overproduction and of finding ways of marketing the surpluses.
Withholding landoften marginal landsfrom production has been one policy
response to this problem.
Consequently, the climate-related issues of the higher latitudes, though fundamentally similar to those of the lower latitudes, are expressed slightly differently. There are two sets of issues. First, will technology be able to maintain increases in average yields in the future? Will the rates of increase from technology or possibly higher CO2 concentrations be sufficient to counterbalance or overcome any adverse impacts on average yields from climatic change? From the long-term perspective, one concern is whether future rates of agricultural change and adaptability will diminish as yields approach the theoretically possible maxima. Some have argued, for instance, that US grain production is reaching a plateau in yield gains (e.g. Thompson, 1975). Others see the yield trends continuing with no anticipated problems in adjusting to climatic change (CIA, 1974; Waggoner, 1983). This issue is further clouded by the uncertainties over continued government support for agriculture in the US and EEC in the face of budget deficits and financial problems. The answers are not clear.
Second, has technology also been effective in reducing the interannual variability of yields resulting from the detrimental year-to-year variations in weather? In other words, is grain production today less vulnerable to the risks of droughts, frosts, freezes, etc., than in the past? And, therefore, do we have less to worry about regarding climatic change and variability in the future? These issues have been subject to considerable debate over the last decade and are still largely unresolved (Schneider and Londer, 1984). Of course, since absolute yields have increased one would expect that the absolute yield variability has also increased. This supposition is borne out by the grain yield statistics in North America (Newman, 1978) and in England and Wales (Dennett, 1980), for instance. The more controversial issue revolves around the relative variability, that is, the departures as a per cent of average (or trend) values. In the context of North America, several studies claim that technology has been triumphant in reducing yield variability (e.g. Newman, 1978; USDA, 1974). Others argue that if one accounts for the fact that from the mid-1950s to the mid-1970s climate was less variable, then grain yields show no decrease in sensitivity to climatic variations and, in fact, the sensitivity may be increasing (NOAA, 1973; Haigh, 1977; Hazell, 1984). The issue so far remains unresolved, principally because of the difficulty in separating the effects of weather and technology.
8.4.4 Global Food Trade
At a global scale, problems of climate and agriculture are superimposed upon a world that is increasingly interconnected through trade, especially cereal trade. From 1960 to 1982, the cereal trade expanded considerably, from about 67 Mt to 221 Mt (FAO, 1961; 1983b). Most of the interaction occurs between the developed countries, those that have the surpluses to sell and those that have the financial ability to buy (Figure 8.4). Wheat dominates the world trade, whereas trade in rice is very small in relation to annual world production.
Figure 8.4 Major wheat exporting countries and their markets. Data from the
Annual Report of the International Wheat Council (1983) (from Tarrant, 1985)
Imports by developed nations have increased fourfold over these twenty-two years. The importers of the developed countries enter the world market largely to acquire feed grains to accommodate rising domestic demands for higher dietary standards as a result of increasing incomes and policy decisions, rather than population growth. The largest importers of cereals are the Soviet Union and Japan, who accounted for 29% of total imports in 1980 (FAO, 1982). Although the developing nations import less than developed nations, their rate of increase of imports has been slightly higher over the last 20 years. Of the developing countries, it is the middle income countries that account for the larger share of the purchases (Wagstaff, 1982). Grain imports of some OPEC countries have increased sharply following the rise of national incomes during the last decade (Parker, 1978). In contrast, the poorer countries, who, nutritionally, have greatest need for food imports, often simply cannot afford the price.
Exports of cereals originate from only a handful of countries in the mid-latitudes. In 1982, exports from the USA and Canada tallied 127 Mt, or about 58% of the world total, with Argentina, Australia and France contributing a large portion of the remainder (FAO, 1983b). Furthermore, while exportable surpluses have become concentrated in fewer countries, the quantities have increased dramatically. In 1960, North America's share of net grain exports stood at only 39 Mt. Meanwhile, from 1960 to 1982, net imports to Asia increased fourfold and, after entering the world market as a major importer in the early 1970s, the Soviet Union (and Eastern Europe) have increased imports to around 44 Mt in 1982 (FAO, 1983b).
One of the major uncertainties in this interregional reliance is climate. The impairment of grain production in one region of the world is of increasing importance to regions elsewhere. Moreover, the trend toward concentration of the centres of both grain supply and demand does not bode well for lower income countries. Of particular concern is the probability of simultaneous crop failures in several major producing or importing countries, as in North America and the Soviet Union. Currently, the probabilities are small (Sakamoto et al., 1980) and production losses in one region are usually offset by gains elsewhere, hence the low interannual fluctuations in aggregate world production. If, however, climatic change in the mid-latitudes proves detrimental to cereal production and the year-to-year risks of shortfalls increase, then the major sources of both supply and demand for grain exports would suffer. The position of the less developed countries with respect to their ability to purchase food could decline drastically under conditions of `shortage' and higher prices on the world market.
8.4.5 Implications and Issues
In sum, we have made the following broad observations regarding climate and agriculture at the global scale:
Rainfall is a principal constraint to agriculture in the semi-arid and humid tropics. The effects of increasing atmospheric CO2 on precipitation patterns and radiation in the tropics (and elsewhere) are highly uncertain.
Due to climatic, agro-technological and socio-economic factors, yields and food production in many developing countries of the tropics and subtropics tend to be low and unstable. Current interannual variability of yields is an acute and pressing problem. With few food security stocks, any downward turn in production could be critical.
In the temperate and higher latitudes, temperature and precipitation interactively influence the spatial and temporal variations in agriculture. It is believed that a CO2 -induced warming would be pronounced at these latitudes.
'Technology' has steadily increased average grain yields in the developed countries of the temperate latitudes. Will the technological progress continue? Will technology or possibly high CO2 concentrations compensate for any adverse effects of climatic change? Has technology reduced or increased the interannual variability in crop yields due to weather fluctuations?
Food production at the margins of crop regions may be particularly sensitive to climatic change. Increasing use of marginal lands for intensive cultivation has contributed to, and may be increasing, the sensitivity of production to climatic change and variability.
Increased interregional reliance from expanding grain trade has made the climate of one region the concern of another. Both key centres of supply and demand for exportable surpluses are located in the temperate regions. Any adverse effects of climatic change could have disproportionately greater impacts on developing countries of the lower latitudes whose purchasing power cannot always compete during times of scarcity.
How the global systems of cereal production would, or could, respond over time to gradual changes in the regional climate, including CO2 concentrations, is the key issue. As a starting point, this requires fundamental knowledge of the sensitivity of crop yields in the core regions of production, either as it would affect supply in the exporting countries or demand in the importing countries. It also requires knowledge of how regional food production might be altered through possible changes in cropping patterns at the margins of production (WMO, 1984). Finally, it requires understanding of the dynamic response of the agricultural system as it re-adjusts over time through changes in technological inputs, management practices, pricing mechanisms, government policies, food security stocks and the like. We know there exists considerable flexibility and resilience in agricultural production in the face of climatic change and variability. The question is how much and for whom.
Against this background we shall address the specific issues and approaches to assessing the impacts of increasing global CO2 and climatic change in Chapter 9.
The relatively 'natural' appearance of many forested landscapes can give one the false impression that forest ecosystems are largely unmanaged. In fact, the world's forests are subjected to a wide range of management levels. Almost all forests are managed to some degree for purposes ranging from intensive commercial extraction to extensive resource conservation. Therefore, as in agriculture, the potential impacts of increasing concentrations of atmospheric CO2 and climatic change must ultimately be examined in the context of human use and manipulation of the natural system.
In the most extreme case, there has been some experimentation with intensively managed biomass plantations in which trees are irrigated, fertilized and harvested in short (ca 2 to 5 years) rotations. This form of cellulose production is the type of forestry that most closely resembles intensive agriculture. In more traditional forestry, forest management involves the regeneration of a commercially valuable tree species by altering sites, planting seedling trees at appropriate spacings, thinning the trees and harvesting the tree crop. In favourable environments, some of these activities are left to natural processes. For example, if a commercial tree species has vigorous regeneration in a given environment, the site preparation or planting management steps can be eliminated, and thinning and harvesting of trees become the only concern. In less intensive forestry, trees are periodically harvested, but the thinning of trees to optimize the forest productivity is omitted. In extensive forms of management, forests are maintained to protect watersheds, wildlife habitats or recreational environments. Even in the most remote forests, there is a degree of management that stems from human intervention with natural processes: for example, reducing the frequency of wildfires in wilderness areas.
Because of this gradient of management intensity, global environmental change could manifest itself in radically different ways. In more intensive forms of forestry, a change in growth and regeneration rates could affect management costs or the techniques used to extract wood products. In the less intensively managed forests, an environmental change might actually change the structure, composition and areal distributionand consequently the functionof the forest ecosystem.
There are two aspects of the behaviour of forest systems that should be considered in assessing the impacts of environmental change. First, there is a considerable degree of spatial heterogeneity in the potential response of the world's forests to changes in climate, as discussed above. Second, at any given place there is a wide range of temporal scales over which forests will respond dynamically. Unlike the vast majority of agricultural systems, forested systems are dominated by long-lived organisms (trees) that can respond to stress or change at several different time scales. Problems in assessing the response of forest systems with respect to any alteration of environmental conditions are made complex by these multi-level responses (discussed in detail in Chapter 10).
There are three major interrelated issues concerning the impacts of increasing CO2 and climatic change that, taken together, transcend the spectrum of forest management and combine the considerations of spatial and temporal dynamics: What are the possible impacts on the areal distribution of the world's forests? At local scales, what are the potential changes in forest composition? What are the possible changes in forest productivity? Let us examine each of these issues in turn.
8.5.1 Climatic Change and the Areal (Macroscale) Response of Forests
The clearly observable correspondence between the distributions of global climates and the spatial patterns of vegetation leads one to expect that a change in the former should eventually produce a response in the latter. What climatic factors should be considered in evaluating the potential change in areal distribution of the world's forests?
Some guidance is provided by Holdridge (1947; 1964), who developed a systematic classification of the expected vegetation under differing temperature and moisture conditions (Figure 8.5). The Holdridge diagram is similar to other climate/vegetation mapping systems in that it explicitly recognizes the variables of temperature (expressed in this case as 'biotemperature' which is computed as a heat index for periods during which plants can be photosynthetically active) and moisture (expressed as either rainfall or evapotranspiration). The Holdridge diagram illustrates several relationships that, while perhaps oversimplifying, provide perspective for understanding the response of the global vegetation to climatic change. First, there is a parallel between the latitudinal zonation of the earth (boreal, tropical, etc.) and the zonation of vegetation at different altitudes on mountains (montane, alpine, etc.). Second, the responses to temperature and moisture or precipitation changes depend on relative, rather than absolute, changes. A small absolute increase in temperature could be expected to cause a large response in the ecosystems of the cooler climates of high altitudes or latitudes. Similarly, a small absolute increase in moisture could have a profound effect in an arid region. To cause a vegetational change of comparable magnitude in a wet, warm region, the environmental changes would need to be much greater.
Figure 8.5 The Holdridge Life-Zone Classification System (Holdridge, 1947; 1964)
If one plots the present global cover of forest as a function of latitude (Figure 8.6), the forested regions of the world resolve into two major forest systems. First, there is a considerable extent of forest in the higher latitudes that tends to be dominated by evergreen coniferous treesthe circumpolar boreal forest. Second, at low latitudes a variety of evergreen and raingreen tropical forests form a second great forest system. The deciduous forests of the middle latitudes which once covered large areas in Europe, China and the United States have been reduced greatly in areal extent by land conversion. For the reasons cited above, of the two great forests now covering the earth, one would expect the boreal forests to be most sensitive to a warming. To determine the response of tropical forests to climatic change would necessarily require spatial information on the change in moisture, if we generalize simply from Figures 8.5 and 8.6.
The amount of time that might be required for the areal distribution of forests actually to respond to a change in global climate is largely a matter of conjecture. The time needed for trees to migrate into a region can vary widely according to the species (e.g. from 25 meters per year in Fraxinus ornus to 2000 meters per year in Acer species; Huntley and Birks, 1983). The ranges of many important tree taxa in both North America and Europe have been moving since the large alteration of the global vegetation pattern that accompanied the last glaciation and may still be moving.
Figure 8.6 Latitudinal distribution of forested land of the earth in 5-degree zones. Percentages are related to the total area of each zone. The distinct bimodality of the distribution corresponds with the boreal forests in the higher northern latitudes and the tropical forests in the equatorial zones (from Baumgartner, 1979)
8.5.2 Local Changes in Forest Composition
Along with the potential areal response of the world's forests (and other vegetation types), there is a smaller scale problem of how forest composition at given locations might change. Results of computer models and of observations of actual forests indicate that, in forest compositional dynamics, the inertia is of the order of the generation of a tree (ca 100 to 200 years). A closer inspection of the processes controlling the structure and composition of a forest reveals that the mechanisms which could be altered by climatic change are numerous and of great potential importance. Changes in climate could be expected to alter differentially the regeneration success and the growth and mortality rates of tree species. The alteration of the competitiveness of the various taxa most likely would be manifested as a change in the forest composition. Changes in moisture conditions (and perhaps thunderstorm frequency) could alter fire probabilities and rates of wildfire spread. Warmer winters could decrease the mortality of overwintering insect pests and thus increase the likelihood or perhaps the intensity of insect outbreaks.
The magnitude of these changes could, in some cases, be quite large, even with relatively small changes in climate. Moreover, these effects are apt to be case specific; it is difficult to generalize across all forests. For instance, a warming in one region could produce increased forest productivity and dominance of a valuable commercial tree species, while the same warming in another region could increase pest populations or wildfire frequency and reduce the extent of commercial forest. Existing computer simulation models for forest phenomena at these time and space scales may prove useful for assessing a number of these local effects and for this reason are reviewed in Chapter 10.
8.5.3 Possible Impacts on Forest Productivity
The immediate responses that one might expect to occur from increased CO2 or climatic change involve modification of forest productivity. Again, there are numerous differences in productivity and in the factors that could modulate the effects of climatic change from one location to the next. For example, a warming could be expected to do little to increase the productivity of a nutrient-limited forest system. Nonetheless, across a broad range of forests, there are positive relationships (with considerable variability) between the temperature and either the total biomass or the net productivity of forests. Given an adequate supply of water and nutrients, one would expect a global warming generally to enhance the forest productivity. That most of the world's forests, however, are to some degree nutrient- and water-limited with respect to their growth rates is sufficient ground for caution with regard to this generalization.
It may be possible to monitor changes in productivity through the use of remote imagery. For instance, recent research has shown that l- and 4km advanced very high resolution radiometer (AVHRR) data from the U.S. National Oceanic and Atmospheric Administration (NOAH) polar-orbiting, sun-synchronous series of operational satellites have immediate utility for repetitive global monitoring of terrestrial vegetation. Daily, the AVHRR sensor records the green leaf dynamics (measured as the photosynthetically active radiation intercepted by terrestrial vegetation) for the entire planetary surface. Daily coverage provides a means to minimize the effects of cloud cover by selection of the most cloud-free data over a given location over several days. Once cloud free data are available at selected time intervals, these data can be used as time series to infer primary production (Goward et al., 1985; Tucker et al., 1985a), to classify continental land cover (Tucker et al.,, 1985b; Townshend et al., 1985) and to study phytophenological phenomena (Justice et al., 1985).
As an example of the global patterns that can be observed using these satellites, Plate 1 shows the integral of the photosynthetically active radiation absorbed by the terrestrial vegetation for a 12-month period. It represents net primary production (or an index highly correlated with net primary production, after Kumar and Monteith, 1982, and Monteith, 1977). Areas of desert and tundra have integrated values (and hence productivity values) close to zero and are represented by the tan colours (i.e. the Sahara, Arabia, portions of southwest Asia, Namibia, coastal South America, Arctic areas and Antarctica). The orange-tan colours represent areas with low amounts of primary production such as semi-arid grasslands and steppes (i.e. Sahel zone, Patagonia, the interior of Australia, etc. and the more northern areas of the boreal zone and southern areas of the tundra). The purple colours represent areas of highest primary production such as much of South America, Central America, the West coast of the USA, equatorial Africa, southeast Asia, etc.). From the patterns in Plate 1, one can clearly see the general pattern of the earth's vegetation and the location of the forests.
Through the use of remote sensing techniques a potential for global environmental monitoring of productivity now exists. These methodologies also have the potential to increase our understanding of the interactions between the biospheric productivity and the atmospheric systems. For example, latitudinal patterns of primary production inferred from AVHRR data seem to be inversely correlated with annual latitudinal fluctuations in atmospheric CO2 concentrations (Tucker et al., 1985c). The exact methods or techniques needed to extract quantitative information crucial to many global environmental questions will likely require additional development, but the continued availability of the AVHRR data will allow future comparisons to be made from actual data. Data for such analyses have existed in archive form since July 1981 (with the launch of the NOAA-7) and continued data collection is guaranteed through the mid-1990s
Plate 1 The global pattern of the integral of the photosynthetically active radiation absorbed by terrestrial vegetation for 1982 obtained from the NOAA-7 advanced very high resolution radiometer sensor system. The intergral image approximately represents net primary production for the year
There are no firm grounds for believing that the net effects of increased CO2 and climatic change will be adverse rather than beneficial. At the extreme, some assessments like the Global 2000 Report to the President (U.S. Council on Environmental Quality and Department of State, 1980) see future changes in climate coinciding with deteriorating conditions in agriculture, forests and other resources, and thus paint a very gloomy picture indeed. In contrast, Simon and Kahn (1984) examine the same issues and, in a strongly optimistic tone, reach just the opposite conclusions. In fact, at a global scale the uncertainties that are involved in both sets of analyses are large enough to accommodate both views.
If, in this (and subsequent) chapters, we tend to emphasize the potential negative impacts, it is only because those are the ones which are of most immediate concern to society and which the scientific community should hope to identify and predict. The list of possible adverse consequences of climate-related ecosystem changes is long and speculative, and the following represent a sample.
Conservation. There are many natural parks and reserves that are refuges for rare and endangered plant and animal species. Often these parks occupy a relatively small area in a setting of non-park land. If an environmental change made such parks unsuitable as habitats for these species, it is uncertain whether alternative refuges could be found or whether it would even be possible to transport species to new sites. The risk of widespread extirpation of rare species (particularly those with local distributions) could be high as a result of climatic change.
Forestry. Forestry as a predictive science used in a management context is highly dependent on data or local knowledge of forest response to specific management treatments. Under a sufficently large change in climate, this local knowledge base would have to be used outside its calibration range and the consequences of management actions would be less certain.
Related ecological processes. The global pattern of many of the ecological processes in natural systems could be altered if the climate changed. Insect pests, pathogenic organisms and wildfire frequencies could all change. While the prediction of such changes is highly uncertain, their potential impacts are quite large.
Hydrological systems. The impact of climatic change on regional ecosystems (particularly forests) could alter the hydrological characteristics of watersheds. Decreases in transpiration rates from the direct effects of increased CO2 on vegetation might increase runoff, for instance, and enhance the effects of precipitation increases or offset the effects of precipitation decreases (Wigley and Jones, 1985). Changes in flooding and river flow rates could have pronounced effects on the rivers themselves, on the ecosystems adjoining the rivers, and on the various human activities that depend on reliable quantity and quality of water.
If, on the other hand, the impacts of increased CO2 , trace gases and climatic change on agriculture, forests and other ecosystems prove, on balance, favourable, all the better.
In summary, this chapter has set the stage for a more detailed examination of
the effects of increased CO2 and climatic change by outlining the major issues
and dimensions of the problem in the global context, with an emphasis on
agriculture and forest ecosystems. For both agriculture and forests the basic
questions are similar: How would crop yields or forest productivity change? How
might crop types or forest composition be altered, particularly at the margins
of production or at ecological transition zones? How might these effects,
integrated over time and space, change global patterns of forests or
agricultural production, taking into consideration the interactive natural and
human processes that make both systems very dynamic?
In order to derive meaningful, credible answers, it is necessary to interface scenarios of environmental change with research procedures or models capable of testing the sensitivity of the systems. What approaches are available? How have they been used and what questions have been asked of them? What have we learned so far from their specific applications to problems of increased CO2 concentrations and climatic change? We turn now to consider these questions in the context of agriculture and, in the subsequent chapter, forests.
NOTE ON AUTHORSHIP AND ACKNOWLEDGEMENTS
Primary responsibility for writing the sections concerned with the general ecosystem effects and with agriculture (8.1 to 8.4) was assumed by R. A. Warrick. H. H. Shugart and M. Ja. Antonovsky concentrated on forest ecosystems (Section 8.5). The collaborating authors, J. R. Tarrant and C. J. Tucker, contributed substantially to Sections 8.4 and 8.5.3, respectively. T. M. L. Wigley also wrote portions of Section 8.3, for which we are grateful.
The authors are indebted to the following individuals for their review comments and constructive suggestions: M. A. Ayyad, W. Böhme, B. Bolin, W. Degefu, B. R. Döös, H. W. Ellsaesser, H. L. Ferguson, D. P. Garrity, R. M. Gifford, F. K. Hare, J. Jäger, M. El Kassas, P. M. R. Kiangi, H. H. Lamb, H. E. Landsberg, J. A. Laurmann, A. de Luca Rebello, J. Mattsson, W. J. Maunder, A. S. Monin, O. Preining, A. Rapp, B. R. Strain, M. S. Swaminathan, P. E. Waggoner, G. F. White, T. Woodhead, M. M. Yoshino.
Bolin, B. (1980) Climatic Changes and their Effects on the Biosphere, WMO-No. 542, Geneva, WMO.
Budyko, M. I. (1982) The Earth's Climate: Past and Future, New York, Academic Press.Crabbe, D., and Lawson, S. (1981) The World Food Book, London, Kogan Page.
Central Intelligence Agency (CIA) (1974) Potential Implications of Trends in World Population, Food Production and Climate, OP-401, August.
Dennett, M. D. (1980) Variability of annual wheat yields in England and Wales, Agricultural Meteorology, 22, 109-111.Flohn, H. (1980) Possible Climatic Consequences of a Man-Made Global Warming, Laxenburg, Austria, International Institute for Applied Systems Analysis.
Food and Agriculture Organization (FAO) (1983a) Production Yearbook, Rome, FAO.Food and Agriculture Organization (FAO) (1961, 1982, 1983b) Trade Yearbook, Rome, FAO.
Fukui, H. (1979) Climatic variability and agriculture in tropical moist regions, in WMO, Proceedings of the World Climate Conference, WMO-No. 537, pp. 426-474. Geneva, WMO.Gerasimov, I. (1979) Climates of past geological epochs, in WMO, Proceedings of the World Climate Conference, WMO-No. 537, pp. 88-111. Geneva, WMO.
Gooneratne, W. (1978) Recent climatic fluctuations and food problems in Sri Lanka, in Takahashi, K., and Yoshino, M. M. (eds) Climatic Change and Food Production, pp. 111-123. Tokyo, University of Tokyo Press.
Goward, S. N., Tucker, C. J., and Dye, D. G. (1985) North American vegetation patterns observed with NOAA-7 Advanced Very High Resolution Radiometer, Vegetacio (in press).Haigh, P. (1977) Separating the Effects of Weather and Management on Crop Production, Contract report, C.F. Kettering Foundation, Grant No. ST-77-4, Columbia, Missouri.
Hare, F. K. (1979) Climatic variation and variability: empirical evidence from meteorological and other sources, in WMO, Proceedings of the World Climate Conference, WMO-No. 537, pp. 51-87. Geneva, WMO.Hare, F. K. (1983) Climate and Desertification, WCP-44, World Climate Programme, Geneva, WMO.
Hare, F. K. (1984) Recent climatic experience in the arid and semi-arid lands, Desertification Control Bulletin 10, May, Nairobi, United Nations Environment Programme.Hazell, P. B. R. (1984) Sources of increased instability in Indian and U.S. cereal production, Am. J. of Agric. Econ., 66, no. 3, 302-311.
Holdridge, L. R. (1947) Determination of world plant formations from simple climatic data, Science, 105, 367-368.Holdridge, L. R. (1964) Life Zone Ecology, San José, Costa Rica. Tropical Science Center.
Huntley, B., and Birks, H. J. B. (1983) An Atlas of Past and Present Pollen Maps for Europe: 0-13000 Years Ago, Cambridge, Cambridge University Press.Justice, C. O., Townshend, J. R. G., Holben, B. N., and Tucker, C. J. (1985) Phenology of global vegetation using meteorological satellite data, Int. J. Remote Sens. (in press).
Kellogg, W., and Schware, R. (1981) Climate Change and Society, Boulder, Colorado, Westview Press.Kumar, M., and Monteith, J. L. (1982) Remote sensing of plant growth, in Smith, H. (ed), Plants and the Daylight Spectrum, London, Pitman.
Lamb, H. H. (1977) Climate: Present, Past and Future, Vol. 2, London, Methuen.
Lemon, E. R. (ed) (1983) CO2 and Plants: The Response of Plants to Rising Levels of Atmospheric Carbon Dioxide, AAAS Selection Symposium No. 84, Boulder, Colorado, Westview Press.Lough, J. M., Wigley, T. M. L., and Palutikof, J. P. (1983) Climate and climatic impact scenarios for Europe in a warmer world, Journal of Climatology and Applied Meteorology, 22, 1673-1684.
McQuigg, J. D. (1981) Climate variability and crop yield in high and low temperate regions, in Bach, W., Pankrath, J., and Schneider, S. H. (eds), FoodClimate Interactions, pp. 121-137. Dordrecht, D. Reidel Publishing Company.Mattei, F. (1979) Climatic variability and agriculture in the semi-arid tropics, in WMO, Proceedings of the World Climate Conference, WMO-No. 537, pp. 475-509. Geneva, WMO.
Monteith, J. L. (1977) Climate and the efficiency of crop production in Britain, Phil. Trans. R. Soc., B-281, 277-296.Nakagawa, Y. (1978) Occurrence and prevention of drought injury to crops in western Japan, in Takahashi, K., and Yoshino, M. M. (eds), Climatic Change and Food Production, pp. 231-236. Tokyo, University of Tokyo Press.
National Academy of Sciences (NAS) (1972) Genetic Vulnerability of Major Food Crops, Washington, D.C., NAS.National Academy of Sciences (NAS) (1976) Climate and Food: Climatic Fluctuation and U.S. Agricultural Production, a Report of the Committee on Climate and Weather Fluctuations and Agricultural Production, National Research Council, Washington, D.C., NAS.
National Oceanic and Atmospheric Adminstration (NOAH) (1973) The Influence of Weather and Climate on United States Grain Yields: Bumper Crops or Drought, Washington D.C., U.S. Dept. of Commerce.Newman, J. E. (1977) Growing seasons as affected by climatic change, in Proceedings of the Conference on Climate Change and European Agriculture, Centre for European Agricultural Studies, Wye College, University of London, Ashford, Kent, UK.
Newman, J. E. (1978) Drought impacts on American agricultural productivity, in Rosenberg, N. E. (ed), North American Droughts, pp. 59-79. Boulder, Colorado, Westview Press.O'Keefe. P., and Wisner, B. (1975) African drought: the state of the game, in Richards. P. (ed), African Environment, London, International African Institute.
O'Neill. R. V., and DeAngelis, D. L. (1980) Comparative productivity and biomass relations of forest ecosystems, in Rechle, D. E. (ed), Dynamic Properties of Forest Ecosystems, Cambridge, Cambridge University Press,
Parker. J. B. (1978) US Farm exports to Saudi Arabia seen doubling this year, Foreign Agriculture, 16 (41), 2.Pino. J. A., Cummings, Jr. R. W., and Toenniessen, G. H. (1981) World food needs and prospects, in Bach, W., Pankrath, J., and Schneider, S. H. (eds), Food-Climate Interactions, pp, 47-68, Dordrecht, D, Reidel Publishing Company,
Sakamoto. C., LeDuc, S., Strommen, N., and Steyaert. L. (1980) Climate and global grain yield variability, Climatic Change, 2(4), 349-361.Satake. T, (1978) Sterility-type cool injury to rice plants in Japan, in Takahashi, K., and Yoshino, M. M. (eds), Climatic Change and Food Production, pp. 245-254, Tokyo, University of Tokyo Press.
Schneider. S. H., and Londer, R. (1984) Coevolution of Climate and Life, San Francisco, Sierra Club Books,Simon. J. L., and Kahn. H. (eds) (1984) The Resourceful Earth, Oxford, Basil Blackwell.
Shah. M., and Fischer. G. (1984) People, land, and food production: potentials in the developing world, Options, 2, 1-5,Stewart. R. B. (1985) The impacts on cereal production, in Part III: The impacts of climatic variations on agriculture in the Canadian Prairies: the case of Saskatchewan, in Parry, M. L,, Carter, T, R,, and Konijn, N, (eds), Assessment of Climate Impacts on Agriculture, Vol. 1 in High Latitude Regions, Dordrecht, D. Reidel Publishing Company (Forthcoming),
Susman. P. O'Keefe, P., and Wisner. B. (1983) Global disasters, a radical interpretation, in Hewitt, K. (ed), Interpretations of Calamity, pp. 263-283. Boston, Allen & Unwin Inc.Swaminathan, M. S. (1979) Global aspects of food production, in WMO, Proceedings of the World Climate Conference, WMO-No. 537, pp, 369-405, Geneva, WMO.
Swaminathan. M. S. (1984) Climate and agriculture, in Biswas, A.K. (ed), Climate and Development, Natural Resources and the Environment Series Vol. 13, pp, 65-95. Dublin, Tycooly International Publishing Limited,Swindale. L. D., Virmani. S. M., and Sivakumar. M. V. D. (1981) Climatic variability and crop yields in the semi-arid tropics, in Bach, W., Pankrath, J., and Schneider, S, H, (eds), FoodClimate Interactions, pp. 139-166, Dordrecht, D, Reidel Publishing Company.
Tanaka. M. (1978) Synoptic study on the recent climatic change in Monsoon Asia and its influence on agricultural production, in Takahashi, K,, and Yoshino, M. M. (Eds), Climatic Change and Food Production, pp, 81-100, Tokyo, University of Tokyo Press.Tarrant. J. R. (1984) The significance of variability in Soviet cereal production, Trans. Inst. Br, Geogr. N.S., 9, 387-400.
Tarrant. J. R. (1985) A review of international food trade, Progress in Human Geography, June (in press).Thompson. L. M. (1975) Weather variability, climatic change and grain production, Science, 188, 535-541.
Toriyama. K. (1978) Rice breeding for tolerance to climatic injury in Japan, in Takahashi. K., and Yoshino. M. M. (eds), Climatic Change and Food Production, pp. 237-243, Tokyo, University of Tokyo Press.Townshend. J. R. G., Justice. C. O., and Kalb. V. (1985) South American land cover classification (submitted to Interciencia).
Tucker. C. J., Vanpraet. C. L., Sharmon. M. J., and Von Ittersum. G. (1985a) Satellite remote sensing of total herbaceous biomass production in the Senegalese Sahel: 1980-1984, Remote Sens. Environ.. 17 (in press).Tucker. C. J., Townshend. J. R. G., and Goff. T. I.. (1985b) African land cover classification using satellite data, Science. 277, 369-375.
Tucker. C. J., Fung. I. Y.. Keeling. C. D., and Gammon. R. H. (1985c) The relationship of global green leaf biomass to atmospheric CO2 concentrations (submitted to Nature).U.S. Council on Environmental Quality and Department of State (1980) The Global 2000 Report to the President: Entering the Twenty-First Century, Vols. I-III, U.S. Government Printing Office, Washington, D.C.
U.S. Department of Agriculture (USDA) (1981) Agricultural Statistics, Washington, D.C., Government Printing Office.U.S. Department of Agriculture (USDA) (1974) The world food situation and prospects to 1985, Foreign Agriculture Report no. 98, Washington, D.C., U.S. Government Printing Office.
U.S. Department of Energy (DOE) (1984) Carbon dioxide and climate: summaries of research in FY 1983 and FY 1984, Office of Energy Research, DOE/ER-0202, Springfield, Maryland, National Technical Information Service.Van Keulen. H. (1978) Simulation of influence of climatic factors on rice production, in Takahashi. K., and Yoshino. M. M. (eds). Climatic Change and Food Production, pp. 345-358. Tokyo, University of Tokyo Press.
Waggoner. P. E. (1983) Agriculture and a climate changed by more carbon dioxide, in NRC, Changing Climate, Report to the Carbon Dioxide Committee, Board of Atmospheric Sciences and Climate, pp. 383-418. Washington, D.C., National Academy Press.Wagstaff. H. (1982) Food imports of developing countries, Food Policy, 7 (1),57-65.
Whittaker. R. H. (1970) Communities and Ecosystems, New York, Macmillan.
Wigley. T. M. L., and Jones. P. D. (1985) Influences of precipitation changes and direct CO2 effects on streamflow, Nature, 314, 149-152.
Williams. G. D. V. (1975) Assessment of the impact of some hypothetical climatic changes on cereal production in Western Canada, in Proceedings of the Conference on World Food Supply in Changing Climate, Sterling Forest, N.Y., Dec 2-5, 1974.
Wittwer. S. H. (1980) Overview: report of Panel III on the environmental effects on the managed biosphere, in U.S. Dept. of Energy, Workshop on Environmental and Societal Consequences of a Possible CO2-induced Climatic Change, CONF-79044143, pp. 46-48. Washington, D.C., U.S. Department of EnergyWorld Meteorological Organization (WMO) (1984) Report of the Study Conference on Sensitivity of Ecosystems and Society to Climate Change, Villach, Austria, 19-23 September, 1983, WCP-83, Geneva, WMO.
World Meteorological Organization (WMO) (1985) The Reliability of CropClimate Models for Assessing the Impacts of Climatic Change and Variability, Report of the WMO/UNEP/ICSU-SCOPE Expert Meeting, Geneva, May 19-24, 1984 (in press).
Wortman. S., and Cummings. Jr. R. W. (1978) To Feed This World: The Challenge and the Strategy, Baltimore, Johns Hopkins University Press.
The electronic version of this publication has been