by Emmanuel CHOISNEL

Anyone can see how the hazards of climate affect the output of farms, even if only through articles in the press or the disastrous effects on some countries' economies. Exceptionally heavy rain which floods land and even claims lives is the most spectacular example of this, and - in the tropics there are hurricanes too. But drought, more insidious but equally terrible, is also a bringer of hunger, famine and death.

Weather goes wrong in different ways in different parts of the globe. In the medium latitudes, or temperate zones, drought is unlikely and it would in any case be very unusual for the same climatic anomaly to be repeated two years running. But things are different in the Sahel, where rainfall stayed stubbornly below the normal level (established over a period of 60 years,) for 20 years, from 1968 to 1987.

In South East Asia, a late monsoon or temporary break in the rainfall can badly affect agricultural output on the Indian sub-continent or the islands of Indonesia. It happened in 1918, 1972 and 1987 and there was the recent El Nino episode, which persisted from June 1982 to April 1983, triggering drought in some places (North East Brazil, Venezuela, Australia, East and West Africa and China), and catastrophic rainfall over some parts of the Pacific Coast.

In this article, we look at an even more general prospect-that of an irreversible change in world climate and its potential impact on plant life. We shall see which are the relevant climatic variables to look at, investigate what a global weather change involves and outline the scenarios of change which might apply to the different ecosystems.

Relevant climatic variables

Plants react to many of the variables of the natural environment- the parts above ground mainly to atmospheric variables and the roots mainly to the variables of the soil-and this response may be reflected in one or more physiological functions such as growth, photosynthesis, development and so on.

The general circulation models suggest that a higher concentration of carbon dioxide or greenhouses gases in general (methane, CFCs etc) will bring about a change in the different basic variables of the climate.

So we need to look at the direct effects of carbon dioxide concentration, of course, and at the individual and, above all, combined effects of climatic variables on plant physiology and, possibly, at any changes in the nitrogen cycle in the atmosphere caused by gas emission from burning savannah (to clear land for crops) and forests.

Without prejudice to the quality of the forecasting in the circulation models usually used to diagnose global weather changes, we should start with the following climatic variables-flows in global sunshine and more specifically in the range of active wavelengths vis-à-vis plant photosynthesis, air temperature and its many physiological effects (frost or heating and changes to the rate of development and growth), rainfall, windspeed and atmospheric humidity.

A proper evaluation of soil water deficits and their effects on the water available to plants also involves looking at the level of, and any variations in, the potential evapotranspiration - a variable which is a combined function of air temperature, sunshine, atmospheric humidity and windspeed.

What is climatic change?

Day-to-day observation of the weather shows that the meteorological conditions in any given place will vary locally in a manner which appears arbitrary but which in fact is dictated by large-scale atmospheric movements. Some paroxysmal atmospheric situations may even trigger dangerous meteorological phenomena such as hail, tornadoes and hurricanes, which may only last a short while but can be highly destructive nonetheless. The notion of climate is something else. First of all, it involves taking a particular period of time (fortnight, three weeks, month, season, year etc) and then observing it over a number of years-30 consecutive years is the World Meteorological Organisation's standard - to characterise the climate of a particular place.

What about fluctuations in climate?

First of all, does this mean the climate of a region, a country, part of a continent (Western Europe, for example), a whole continent or the whole world?

A definition of the average climate is the starting point, using the average values, month by month, of the main components of the climate, calculated in the light of the 30-year climatic reference period. The average present climate is defined by reference to 1951-1980.

The climate varies around this average state to reflect the statistical dispersion of weather components around the average values. This climatic variability is natural and must be considered as intrinsic.

Nonetheless, each climatic anomaly (i.e. abnormal air temperatures, sunshine, or rainfall for three weeks or more) may cause agricultural output to decline in the I region in question if it happens at a critical moment in the crop cycle-as with, say, a drought in the temperate zones in the northern hemisphere in July. And if the anomaly lasts for two or three seasons in a row, as occurred when I Europe had a drought lasting from December 1975 to August 1976, the consequences on production may be disastrous.

A global climatic anomaly, i.e. a series of anomalies affecting different parts of the world, occurred in 1972. France had a cold, wet summer, the monsoon failed in India and Pakistan and there was drought in the European parts of the former USSR, disastrous drought in the Sahel and El Nino off the coast of Peru. There was a huge protein shortfall on the food and agriculture market as a result and a period of marked variation in grain prices began. There has been nothing like it since.

But it was not a real climatic change. A real climatic change is a change in the energy balance of the land-sea-atmosphere-cryosphere-biosphere leading to an ! irreversible change in the average climate. It is something which affects the whole globe and it can be caused by a heightening of the greenhouse effect through increased concentration of the so-called greenhouse gases (carbon dioxide, CFCs etc).

The consequences of a global climatic change will vary according to latitude, to season and even to where a region is located within the continent (on the coast, inland, on the east or west of the land mass etc).

Simulations of changes in world climate connected to a doubling of the relevant CO2 levels using general atmospheric circulation models give a new level for the average climate season by season. A figure has been put on the rise in the annual temperature increase of the globe, but the models disagree about the scenarios for regional trends in rainfall. There is a great deal of uncertainty as to trends in cloud cover-- on which any change in potential evapotranspiration (i.e. the climatic demand for evaporation on the vegetation) depends.

Furthermore, the climate modellers have not yet made any firm statement on a possible change in the degree of variability of the climate from one year to another because of the difficulty of I interpreting the statistics on the results of the models. Yet this is the first thing the agrometeorologist asks when looking at the possible effect of global change in climate on agricultural production. An increase in the annual variability of rainfall (if a diagnosis were made) would have major repercussions on the annual variations in agricultural yield.

Diagnosis of the present situation -climate, arable zones and natural and cultivated ecosystems

Weather is fairly well understood today and it is described in all areas which have proper weather stations - i.e., roughly, the continental and non-desert zones with high- and medium-density populations. Ocean zones are still not so well understood, despite the fact that they play a vital part in the energy balance of the climatic system mentioned earlier.

The general atmospheric circulation . models used to simulate the present climate reproduce, roughly, the main features of the different weather zones- although they over-estimate the annual average rainfall in some parts of the globe I (Burma and Malaysia in South East Asia, for example) by a factor of two.

The world's various climates can be described in a number of ways. They can be classified by the Koppen method, the oldest system but still an appropriate one (although specification is needed for the Europe-Middle East area), or by the Holdridge method (which uses a diagram combining an annual thermal index, total annual average rainfall and the ratio of potential evapotranspiration to annual rainfall), which is currently favoured by the modellers.

The earth's 16 million km² of arable zones take up very little more than 10% of its land (grazing land not included). The potential fragility of world agro-food resources is obvious when it is realised that the desert and semi-desert zones take up three times more-upwards of 30% of the continental land masses. More than half the arable land, the FAO suggests, is used to grow grain, 80% of it wheat, maize or rice.

What about the natural ecosystems? Almost a quarter of the land is forest, divided into two, clearly distinct groups:

-humid equatorial forests;
-northern forests.

Scenarios for trends in plant ecosystems

An in-depth investigation of this involves:

1) identifying the principal types of ecosystem to be taken into consideration;
2) knowing how plants and trees respond to a given variation in (one of the) climatic variable(s);
3) dividing continental zones into major climatic types;
4) looking at scenarios for climatic trends for each of these types.

The effect of climatic change on ecosystems may be apparent in a number of ways-in the response of plant communities, in the effects on crop yield, in changes in the range of species present in a given ecosystem or in the spread of that ecosystem. We shall deal with each of these in turn.

Different types of ecosystem

We use the word 'biome' for the major sets of ecosystems occupying the continents. The generic names for these are tropical forest, rain forest, savannah, Mediterranean forest, temperate forest, northern forest, steppe, tundra, desert etc. The word 'ecosystem' tends only to be used for natural vegetation. In cultivated parts of the globe, a classification by biome would be based on the vegetation which would occupy the land if there were no crops. By extension, we use the term 'cultivated ecosystem' here to mean agricultural areas which are priorities from the point of view of the world food supply. The distribution of biomes roughly reflects the major division of the continental land masses into principal climate types-illustrating the crucial importance of a global climatic change to the spread of plant ecosystems over the surface of the globe. So, to use Koppen's terminology, tropical forest and savannah are associated with a humid tropical climate, steppe and desert with a dry climate, northern forests with a sub-Arctic climate and tundra with a polar climate.

However, there are other subdivisions of ecosystems which may be useful to us here. For example, do they have annual or perennial crops and is the plant cover closed or not?

When it comes to global climatic change, the experts focus more on forest ecosystems than agricultural systems, on the grounds that the former are, on the face of it, more vulnerable than the latter and that, in the case of the latter, the type of crops can be changed when the time comes.

Plant sensitivity and response to variations in climate

Basic knowledge in this field is obtained by studying the way a given plant reacts to a single element of the climate in isolation, in a strictly controlled laboratory environment.

The problem is in the natural climatic environment where a number of variables may fluctuate simultaneously. And the concomitant variations in any one region or season are marked by a relative interdependence of variables-a milder winter in a temperate ocean climate, for example, will usually go hand in hand with more rain.

The problem is complicated by the fact that the change in climate which concerns us here is also linked to an increase in atmospheric CO2, which directly affects photosynthesis. Agricultural results suggest that this initially encourages the growth of plant life-but there again, this is a situation in which CO2 alone varies and the other factors (temperature, water and nutritive elements) are at optimum levels.

In fact, the important thing in diagnosing plant response to climatic change in the light of scenarios for trends in climatic variables deduced from general circulation simulation models is to find out whether one element of the climate can constitute a limiting factor or not.

There are three limiting factors:

-light and sunshine;
-temperature;
-water.

Light, or, more precisely, the fraction of sunshine between, roughly, 0.4 and 0.7 µm, affects photosynthesis.

Because of its effect on photosynthesis, respiration and the translocation of carbohydrate, temperature affects both the running of the plant's timetable of biological development (phenological stages such as flowering, maturity etc) and its growth. These are what might be called the cumulative effects of temperature.

Temperature thresholds must also be taken into account here. Exposure to heat and exposure to cold can both be limiting factors if the temperatures reached provoke a lethal effect on the plant.

Availability of water also affects the various physiological functions of plant life. Water restriction (water stress) can limit or inhibit growth, compromise the success of bedding and speed up the development of the plant.

Lastly, CO2 concentration affects photosynthesis, photorespiration, nocturnal respiration and morphological development.

Scenarios for trends in world climate

These are given for a doubling of the CO2 equivalent concentration, i.e. assuming an increase in the concentration of this gas itself and translating the effect of an increase in the other new greenhouse gases into an equivalent increase in CO2 from the point of view of their contribution to boosting the greenhouse effect.

The reference for CO2 concentration is the so-called pre-industrial yardstick, i.e. the 270 ppmv of around 1880. So when it comes to the date at which doubling will occur, the scenarios depend on trends in industrial and human activity.

The first and principal figure is for the rise in the annual average temperature of the whole globe-assuming that a fresh state of land-sea-atmosphere-biosphere equilibrium gas been achieved (figure 1). The US Environmental Protection Agency published its estimated figures for this in 1983- + 3°C to ± 1.5°C, which gives a range of + 1.5°C to + 4.5°C. The uncertainty is linked to poor understanding of the way ocean surfaces respond to thermal forcing from the atmosphere and to doubts about the way the cloud cover has been modelised.

Experts then came up with scenarios for regional trends. In 1987, the conclusions of a group of experts set up under the aegis of the World Meteorological Organisation and the UN Environmental Programme gave the following evaluation (northern hemisphere only):

-an increase in maximum air temperatures at high latitudes and in winter which is about double that of the annual average global temperature increase;
-an increase in the temperature of the tropics which is smaller than the increase in the global temperature, and without any seasonal changes;
-far less reliability of scenarios for rainfall trends. If rising temperature is not cushioned by increased cloud cover, then it will boost evaporation over the oceans. Experts predict that the atmospheric water cycle will speed up world wide. An increase in total annual rainfall without indication as to seasonal breakdown is of little interest when it comes to diagnosing the consequences on farming, because the crucial thing is evaluating water resources in summer, since plants' water intake depends on the level of potential evapotranspiration. Furthermore, agriculture wants maximum regularity of rainfall, since heavy rainfall over a short period can be damaging or just plain destructive and is more difficult for the soil to absorb.

Table 1 sums up the experts' scenarios for the two extreme seasons in three sections of latitude in the northern hemisphere.

The latest evaluation of the rise in average world temperature with a doubling of CO2 is + 1.8°C, with a higher figure in the northern hemisphere (+ 2.5°C) than in the southern hemisphere (+ 1.0°C). If industrial activity neither declines nor speeds up, this doubling is expected to have taken place by about the year 2030.

Figure 1: Evolution of Global World Temperature

(Source: Jaeger, WMO-UNEP, 1988)

In September 1990, the intergovernmental group on trends in the climate came up with various scenarios for regional trends in five zones which had scored fairly similar results in the various general circulation models used to simulate the climatic consequences of the doubling of the CO2 equivalent concentration. They were the central part of North America, South Asia, the Sahel, Southern Europe and Australia. The scenarios are set out, with reservations (due to a low level of confidence), in Table 2. For the purposes of interpretation, note that the data:

-are for extreme seasons (i.e. winter and summer);
-correspond to an average for the whole zone in the given ranges of latitude and longitude and therefore mask possible differences within those zones. 'Southern Europe', for example, goes from the near Atlantic to Turkey and covers a wide range of what are currently very different climatic types;
-are to be considered es 'best estimates'. For example, they may be cut by 30% for a low hypothesis or increased by 50% for a high hypothesis to reflect the range of results in the different models;
-contain no indication of any change in the extent of the variability of the climate.

The most recent conjecture here suggests that the annual variability of temperatures will decline and the annual variability of rainfall will go up. But in the absence of any season-by-season regional scenarios, it is not possible to say what the consequences on agriculture will be.

Possible effect of a global climatic change on farming and forests

Here we have to consider not just the effect of the change in the range of variation of an individual component of the climate (temperature or rainfall), but also-and this is most important-of that effect combined with concomitant changes in temperature, rainfall and cloud cover affecting various physiological functions (growth, photosynthesis, transpiration, phenology, reproduction, mineral nutrition etc) and ecological processes. So it is an exercise which is full of pitfalls.

We also need a diagnosis of the current spatial distribution of crops across the I world and of their suitability for the climate. In the southern reaches of the temperate zones, we have to identify the crops which have already reached the upper limits of adaptability as far as heat is concerned-on the assumption that warming is to occur.

We shall look at both the possible effects on farming and forests and- . naturally, given what has already been said about the relations between plant distribution and the major climatic zones and on the different scenarios for climatic change in different latitudes-at the effect by section of latitude. The comments relate mainly to the northern hemisphere, partly because it contains more continental land masses than the southern hemisphere and partly because climatic simulations suggest it is more affected by the consequences of a doubling in the concentration of CO2 equivalent.

Table 1: Scenarios of climatic change by section of latitude by season in the northern hemisphere

Regional scenarios for trends in climate in five parts of the globe

The tropics

Temperature is not considered to be a limiting factor as far as farming is concerned and the higher temperature scenarios announced in case of climatic change tend to be lower than those in other sections of latitude. So rainfall is the real key variable here.

Here, where the main crop is rice (apparently the staple food of more than half of mankind), the length of the rainy season, closely tied up with the arrival of the monsoon, determines the year's agricultural output. The most sensitive regions are those where feeding an expanding population has forced people to grow crops in marginal areas where the soil is poor and less water is available. The scenario for South Asia (essentially the Indian sub-continent), in Table2 involves an increase in the summer monsoon.

The irregular arrival of the monsoon has a lot to do with variability in the climate-a point on which the model makers are silent. Excessive rainfall at harvest-time can also be damaging.

The most immediate concern with the tropical rainforests is the present rate of deforestation-at least 11 million hectares are cleared every year, with a boost to the greenhouse effect too, when bare earth and the rotting waste from clearance release masses of CO2 into the atmosphere.

However, the diversity of species making up what is left means that the ability for change is relatively good-in that rainfall trend scenarios are not geared downwards because, on the face of it, the species there are adapted to globally non-limiting hydric conditions.

In the semi-arid parts of the world, the length of the period of vegetation depends strictly on how long the rains last and any delay in the planting dates will reduce this period of growth and restrict production. The Sahel had 20 consecutive years of sub-normal rainfall ('normal' being established over a much longer period of about 60 years) and the decisive climatic factor constituted by this persistent downward trend in (roughly) 1968-1987 was a major factor of desertification there. So an analysis of the problem must take account of how the interaction of trends in the plant cover and surface energy processes help determine climate.

Temperate zones in medium latitudes

These are the main grain-producing zones, providing more than 75% of the world's wheat and maize output. The thermal effects as such on farming tend to be seen as positive since they extend the crops' potential growing period-providing, of course, that lethal high-temperature thresholds are not reached. However, higher summer temperatures combined with unchanged or diminishing cloud cover increase the potential level of evapotranspiration and thus help push up the crops' demand for water (with summer rainfall constant). Even with ' sound water resources, an earlier start to l irrigation will put extra constraints on the farmer.

Both direct and indirect effects should be considered when looking at water resources. Lower summer rainfall is, of course, a directly limiting factor as far as non-irrigated crops are concerned, particularly for maize, which has a critical period in July. One worrying indirect effect is that a larger percentage of liquid precipitation in mountain areas than at present would reduce the water stored in the snow cover by an equivalent amount,; thereby reducing the water resources of crops lower down in spring and summer.

The temperate forests, all of which are in the northern hemisphere, will apparently suffer from a northwards shift of the zone in which the species which make them up could spread. A crisis situation would be difficult to avoid if the annual average temperature rose by more than, one degree in less than a century. The natural speed of seed migration deduced from an analysis of the distribution of species at the end of the Quaternary in Europe and North America is about 1050 km per century for most forest species (it happened at the time of postglacial recolonisation). Natural regeneration can only occur through seed migration, with a 1°C drop corresponding to a shift of 100-150km northwards. Changes in the productivity of forests and in the rate at which forest cover replaces itself naturally are in any case to be expected.

High latitude regions

The scenarios for these regions involve, typically, a change in winter climates, with an increase of as much as 4°C or more in temperature and extra precipitation. The thermal effect would shift the | possible zone of extension of plants and l shrubs northwards and extend the growing period-which would push up grain output in these places and make more productive strains a possibility. Winter wheat could be substituted for spring wheat in some parts, particularly Canada.

More rain in winter would be a help in zones with deep soil and heavy water storage potential in autumn and winter- although excessive rainfall could well have an adverse effect on farming methods.

Northern forests can be expected gradually to colonise land further north, but the thermal effect could well adversely affect the southern reaches. Increased precipitation would improve the growth of forest species growing on deep soil, but shallow-rooted trees would suffer more from the general increase in evapotranspiration throughout the year. The risk of organic matter in the tundra breaking down faster under the effect of warming is a cause of concern in monitoring atmospheric CO2.

These prospects of climatic change in different parts of the globe make the need for forward-looking coordination of agro-food resources and plant life in general more urgent than ever. Constant monitoring of the ecosystems-along the lines of NASA's Earth Observing System-has to be set up by the end of the century.

Strategically speaking, the creation of seed banks could be an asset when it comes to maintaining a range of vital genetic material to cater for changing climatic and agro-food situations across the world. It is important to preserve the intraspecific genetic diversity of the wild species in the forest ecosystems-knowing that a forest with multiple species will, given all the faculties of adaptation of each one, withstand climatic change better than a forest composed of a single species of tree. E.C.