United States

Figure 34.1 shows different yield anomalies, (y — yt)/yt, for the U.S. states Colorado, North Dakota, and Kansas. These curves show the important regularity that, despite significant distinctions in meteorological conditions in these three U.S. states and their remoteness from each other, the principal features of their annual variability are similar. The five-year smoothed averaged values of relative anomalies were mainly positive during 190017, 1940-48, and negative during 1918-39.

In addition to the annual or five-year mean variability, yield variability can also be studied using the hydrothermal coefficient (HTC) and S index that are widely used for drought monitoring in the former Soviet Union (chapter 15). The variations in these indices for the above U.S. states are shown in figure 34.1. The meteorological data used for computing these indices were collected from three meteorological stations: Denver (for Colorado), Bismarck (for North Dakota), and Kansas City (for Kansas) (http://lwf.ncdc.noaa.gov/noaa/climate/).

Figure 34.1 The historical variability in wheat yields in Colorado, North Dakota, and Kansas, USA. (a) Annual variability in yield (metric tons per ha) and its trend (solid line), (b) yield anomalies from five-year means (solid line); two horizontal dashed lines show the 15% yield anomalies; (c) S-aridity index and its five-year means (solid line), and (d) hydrothermal coefficient (HTC) and its five-year means (solid line); dashed line shows the HTC normal.

Figure 34.1 The historical variability in wheat yields in Colorado, North Dakota, and Kansas, USA. (a) Annual variability in yield (metric tons per ha) and its trend (solid line), (b) yield anomalies from five-year means (solid line); two horizontal dashed lines show the 15% yield anomalies; (c) S-aridity index and its five-year means (solid line), and (d) hydrothermal coefficient (HTC) and its five-year means (solid line); dashed line shows the HTC normal.

Europe

Figure 34.2 shows wheat yield anomalies for Bulgaria for 1960-2001 after removing the linear (assumed) technological trend (i.e, the yield increase on account of fertilizers, mechanization, and the introduction of high-yield varieties). Such trends were first developed for the United States and the former Soviet Union (Menzhulin and Nikolayev, 1987; Menzhulin et al., 1987). For this purpose fertilizer used (kg/ha), combine harvesters

1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 Figure 34.1 continued

and tractors (relative numbers/ha) available, electric power, and pesticides and herbicides used have been included in the calculations. In the case of Bulgaria the correlation coefficient between n5 (5-year smoothed relative yield anomaly) and S5 (five-year smoothed S index) is 0.51. Wheat yield anomaly was positive from 1970 to 1992. The deviations of annual yields from the technological trend during this period were mainly positive. In one unfavorable year, 1985, the wheat yield in Bulgaria fell to 11% below the trend. After 1983 the yield anomaly trend entered into a declining phase, and after 1992 the wheat yield anomalies were mainly negative. In 1996 a serious failure in wheat production took place, when yield was 47% below the trend.

Figure 34.3 shows wheat yield anomalies during 1960-2001 for several other European countries. By studying the trends in anomalies one can conclude that the 20-year period, from the beginning of the 1970s until

CLIMATE CHANGE, GLOBAL WARMING, AND DROUGHTS 437 Wheat, Kansas, USA

1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Figure 34.1 continued

1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Figure 34.1 continued the beginning of the 1990s, was favorable for many European countries with positive anomalies. Like for Bulgaria during this period, drought (causing 15% or more reduction in yields) was practically absent. The exceptions were only five years: 1979 for Austria, 1991 for Albania, 1993 for Hungary, 1990 for Greece, and 1974 for Malta. Yield anomalies were negative after the mid-1990s.

African Mediterranean and Near East

Figure 34.4 shows wheat yield variation for some African Mediterranean and Near East countries during 1967-95. The yield trend was characterized mainly by negative anomalies. Frequent droughts (causing yield reduction exceeding 20% of the normal) occurred in 1970, 1978, 1983, 1984, and 1985 in Yemen; in 1970,1971,1973,1977,1979,1984, and 1989 in Syria;

Figure 34.2 Variability in wheat yield in Bulgaria as expressed by (a) S-aridity index and (b) annual variability; bold solid lines represent trends.

in 1968, 1969, 1971, 1978, 1979, 1983, 1984, and 1985 in Lebanon; in 1970, 1973, 1975, 1976, 1977, 1978, 1979, 1981, 1982, 1985, and 1986 in Jordan; in 1970, 1973, 1977, 1982, 1982, 1986, and 1991 in Cyprus; and in 1970,1976,1979,1982, and 1984 in Sudan. However, in Egypt and Iran, a severe drought occurred only once, although the average anomaly of the wheat production in these countries was also negative.

Southern Hemisphere

Figure 34.5 shows the results of the yield anomalies for some Southern Hemisphere countries for last 40 years. It is interesting to note that in these main grain-producing areas of the Southern Hemisphere the changes in the wheat yields followed a common pattern. In South Africa, Namibia, Australia, New Zealand, Argentina, and Uruguay, the beginning and end of the negative phase of the wheat yield anomalies fall approximately in the same years: 1970 and 1995. During this period, drought (causing yield reduction of 20% or more from the normal) occurred in 1978,1980,1983,

1985, 1989, 1990, and 1995 in South Africa; in 1977, 1978, 1985, 1986, 1987, and 1988 in Namibia; in 1972, 1977, 1980, 1982, and 1994 in Australia; and in 1970 and 1989 in New Zealand. In Argentina negative wheat yield anomalies occurred during 1970-95, and droughts occurred in 1968 and 1981. In Uruguay drought occurred in 1976, 1977, 1978, 1985,

1986, and 1993.

Having noted that the regularities in the variation of wheat production anomalies in different geographical regions in recent years, we now turn our attention to the question of whether the results obtained could be used for prediction purposes—to extrapolate these trends for the years ahead.

0 0

Post a comment