CHIN ESE GEO GRAPHICAL SCIENCE
Volume 8, N umber 4, pp. 309- 316, 1998
Science Press, Beijing, China
IMPACT OF CLIMATE ON GRAIN YIELD PER
UNIT AREA IN CHINA DURING THE
YEARS OF 1949- 1992
Xie Yun( 谢云)
L iu Jidong( 刘继东)
T he D ep ar tment of Resour ce and Environmental Science, Beij ing Nor mal Univer sity ;
T he Op en Resear ch Labor atory of Env ir onmental Change and N atur al Disaster s
of the S tate Education Commission, Beij ing 100875, P . R . China
( Received 8 M ay 1998)
ABSTRACT: T he g rain yield per unit area in China has been increasing from 1949 to 1992 w ith annual fluctuations, w hich was caused by both policies and monsoon climate. T he total grain yield is d-i
vided into trend yield, which r eflects the influence of economic factors, and climate yield, w hich reflects the impact of climate factors. T he results of grain and climate yield by EOF ( empirical o rthog onal function) method show t hat t he g rain yield in the eastern China is higher and mor e fluctuating
t han t hat in the w estern region. T he spatial and temporal distributions of climate yield were div ided
into four patter ns. Regional unanimity of climate- yield variation is the first and main spatial pattern,
and then ther e are the spatial patterns of north- south difference and east- west difference in some
years. Generally, climate condit ions are good fo r grain yield dur ing the 1950s and 1980s all over the
country .
KEY WORDS: grain yield, climate yield, empir ical ort hogonal function, spatial and tempor al distr-i
bution
Since t he year of 1949, t he product ion of g rain in China has experienced a great increase,
w hich indicates that not only the f ood problem of 1/ 5 of t he w orld populat ion has been solved,
but also China can ex port g rain into t he world g rain m arket. H owever, it should be seen that ,
at present , product ion of agricult ure in China depends on natural condit ions great ly, especially
climate. T he climat e in China is monsoon climat e, w hich is charact erized by the great annual
v ariat ion of precipit at ion and severe nat ural disaster such as flood and draught and makes grain
product ion more unst able. A lot of researches have focused on analysis of the im pact of climat e
change in the fut ure on China s agriculture product ion, w hich includes planting system,
change of quant it y and qualit y of crops, change of t he lengt h of t he growing season, etc( Guo et
Support ed by the N at ional N at ural Science Fundat ion of China.
309
al . , 1993; Guo et al . , 1996) . As regard to the impact of modern climat e on agriculture,
there have been many reg ional researches. T his paper at tempt s to research t he ef fects of climat e
factors on the f luctuat ions of the grain y ield and it s regional variance throug h analyzing the t emporal and spat ial variation of t he grain yield per unit area in China and clim at e change for more
than 40 years. It is important for the sustainable development of Chinese agricult ure.
I. DAT A AND M ETH OD
St at e St at istic Bureau s dat a of g rain yield of 1949- 1992 f or each province, municipalit y,
and autonomous reg ion were used. Because t he dat a in T ianjin and the Xizang Autonomous Reg ion are incomplete and there are no data f rom T aiw an P rovince, they are not included in this
st udy, and t he data of new province H ainan is st ill included in Guangdong Province.
T he product ion is t he result of both social economic and nat ural fact ors. T he g rain yield
has been increasing cont inuously w ith annual fluct uat ions. So it can be divided into tw o parts:
trend yield and clim at e y ield. T he f ormer w as det ermined by social econom ic input and the development of t echnolog y. T he lat ter w as mainly reflect s the influence of climate factors w hich
are still out of man s control. T his can be ex pressed as follow s:
Y = Yt + Yc
w here Y is actual grain y ield per unit area; Y t is the t rend yield, represent s t ime tendency; Y c
is climate y ield, w hich refers to the fluct uat ion of grain yield. T his paper analyzes the climat e
yield mainly . According t o our research w e use a polynomial equation to sim ulat e the trend
yield as follow s:
Y t = b0 +
b i ti
w here t is t ime sequence; b is coef ficient; and i is power w hich equal t o 4 in t his st udy . T he
regression result of t he selected equat ion is at 1% signif icant level. T hen
Yc = Y - Yt
In order t o eliminate t he inf luence caused by great dif ference of numbers scale among diff erent regions and make the yields comparable, climat e yields in each province, municipalit y,
aut onomous region have been normalized. Normalized climat e yield data w as disint egrat ed by
empirical ort hogonal funct ion ( EOF) , the results w ere used t o analyze t he spat ial dist ribution
and t emporal chang e of climate yields and the impact s of climate on g rain product ion in China.
II. T EM PORAL AND SPAT IAL CHARACT ERIST IC OF GRAIN YIELD
T he level of grain y ield per unit area indicat es a reg ion s agricult ure production level, especially in China, which has a large amount of population but insuf ficient land resource. T he
food supply mainly depends on t he increase of grain yield per unit area. In China, the grain
310
yield has been increasing g radually during more t han 40 years, from 1020 kg/ ha in 1949 to
4005 kg / ha in 1992. T he average increase rate w as about 67. 5 kg/ ha per year, but it was diff erent in different periods ( T able 1) .
T able 1
T he change of grain yields in different per iods
Years
1950s
1960s
1970s
1980s
1949- 1992
Increase( kg/ ha)
39
79. 5
81
111
67. 5
T he levels of grain yield per unit area demonst rat e st rong reg ional diff erence. T he areas
w it h higher productivity of grain production are most ly locat ed in east China. T hey are Nort heast Plain, Nort h China Plain, Changjiang River M iddle and Low er Reaches P lain and Pearl
River Delta and the g rain yields in these regions are above 4500 kg/ ha. T he g rain yields in
most of provinces in nort hw est China, such as Shanx i, Shannx i, Inner Mong olia, Gansu,
N ingxia, Qinghai provinces( reg ions) , are just below 3000 kg/ ha. T he grain yields of ot her
places in China are betw een 3000- 4500 kg/ ha. Generally speaking, t he f luct uat ion range in
east China wit h higher product ivit y is greater t han t hat in other reg ions wit h com parably low er
product iv it y ( T able 2) . T his is because east China is impacted by monsoon climate w it h a great
annual variat ion of climat e and f requent climat e disast ers.
T able 2 T he level and chang e of grain yields( kg/ ha)
G rain yield( kg/ ha)
A bsolute variat ion of grain yield( kg/ ha)
R elat ive variat ion of grain yield( % )
< 3000
3000- 4500
595. 5
798. 0
1219. 5
4500
21. 9
21. 2
22. 6
III. T HE T EM PORAL AND SPAT IAL VARIAT ION OF CL IMAT E YIELD
1. Basic Characterist ics
Climat e yield changed from year to year and the change scale w as different among diff erent
regions. T he sm allest change of climate yield was 552 kg/ ha w hile t he larg est w as 2850 kg/ ha.
Liaoning, Jilin and Shandong provinces have t he great est range of change, which is beyond
1500 kg/ ha. Average national climate y ield w as 37. 5 kg/ ha during the 1950s, it w as decreased
during the 1960s and 1970s, and then t ended to increase during the 1980s ( T able 3) . T he dat a
of climate yield shows three diff erent patt erns in temporal variation ( Fig. 1) .
T able 3 Climate y ield in different per iods
D ecades
1950s
1960s
1970s
1980s
Climat ic yield( kg/ ha)
37. 5
- 12. 6
- 27. 6
6. 1
( 1) Double peak w ave t ype, which is m ainly represented by Hebei, Shandong, Henan,
311
Anhui in H uang- Huai Plain; Middle and low er reaches of t he Chang jiang River including
Shanghai, Jiangsu, Zhejiang, Jiang xi, H ubei, Hunan, Sichuan, Fujian, Guang dong, Guangx i
prov inces ( regions) in sout h China ( F ig . 1a) .
( 2) Gentle t ype, w it h less change, w hich includes Beijing , Inner Mongolia, Shanxi,
Shaanx i, Ningx ia.
( 3) Sing le depress type w hich can be f ound in Gansu, Qinghai, Xinjiang in northwest
China and Yunnan, Guizhou in sout hw est China and Heilong jiang in nort heast China. T he
low est value appeared in the 1960s and aft er t hat t he fluct uation becam e minor ( F ig. 1 C1 ) . In
Jilin and Liaoning t he lowest value appeared in t he 1980s and t here had minor fluctuat ion before
( Fig. 1 C2 ) .
a. double peak w ave type
b. gent le t ype
c. singl e valley t ype
F ig. 1 Different types of climate yield chang e
2. Empirical Ort hogonal Funct ion Analysis for Climat e Yield
T he cont ribut ion of accum ulat ed variance of the f irst 4 special vect ors of t he EOF is
62. 9% . T hey show the main spatial variat ion of clim at e y ield ( F ig . 2) .
T he first special vector s contribut ion is 35. 0% , w hich is t he most important spatial var-i
ation patt ern. All f irst vect or values are greater than zero, and it reflects t he uniform feat ure all
over t he country in t he variat ion of t he climate yield. When its corresponding time coeff icient is
g reat er t han zero, climat e yield all over the country increases, such as the years of 1958, 1966,
1970 and 1984. Otherw ise, climat e yield decreases, such as 1961, 1977 and 1989. T he hig her
v alues of this pat tern are in sout h China lying to t he south of t he Qinling M ount ain and the
312
Fig. 2
Distr ibution of the first 4 EOFs ( EOF1 - EO F4)
Huaihe River, w hich indicat es climate yield has great er f luctuat ion there. T he lower values are
in nort h China w here t he fluct uat ion of climat e y ield is relat ively minor. T he t ime coef ficient
( F ig. 3) f luct uat es around zero, w hich is posit ive in the 1950s, lat er 1960s, earlier 1970s, and
middle of 1980s, in t he ot her t imes t he values are neg at ive. T he low est and highest values appear in the beginning of t he 1960s and t he middle of t he 1980s respect ively, which corresponded to the most severe natural disast er year and the most f avorable w eather y ear since 1949.
During the 44 years from 1949 to 1992, t he climate yield increased for 25 years but decreased
for 19 years.
T he second special vect or of EOF cont ribut es 10. 4% variance. T w o obviously diff erent
regions can be indicated according to t he line of w est or nort h boundaries of Hebei, Shandong,
Anhui, Hubei, Hunan and Guang xi provinces. T he vector value is negat ive in sout heast region
w it h the lowest center in south China coast area , w hile t he northwest region has a positive
v alue w it h t he highest center of Inner M ongolia , Gansu , and Qinghai prov inces . When its
313
F ig. 3
T ime coefficient of the first 4 EOFs ( EOF1 - EOF 4)
corresponding t ime coeff icient is greater t han zero, the climat e yield decreases in the region
w here t he vect or value is neg at ive but increases in the reg ion w here t he vect or is posit ive. When
time coeff icient is below zero, the sit uation is opposit e. For example, during t he 1950s and
1970s the time coef ficient w as positive, so t he clim at e yield increased in northw est China but
decreased in southeast China. Oppositely , during the 1960s and 1980s, t he climat e yield increased in southeast China but decreased in nort hw est China. During t he 1990s t he situation
w as t he same as t hat in the 1950s and 1970s ( Fig. 3) . What is noticeable is that southeast area
w it h the neg at ive vector value is the main area of t he grain product ion, so the change of grain
yield in this area w ill strong ly influence t he yield of t he whole count ry. In short, the change
tendency of t he southeast showed by EOF2 is similar t o that show ed by EOF1 for the w hole
count ry: w hen yield was increased in southeast China during the 1960s and 1980s, it w as the
same as the w hole country. T he symbol of EOF2 is value of sout heast area is opposit e t o that of
EOF1 is, and t he change of their t ime coeff icient s is opposite t o each ot her eit her.
EOF3 is cont ribution ratio of variance is 8. 6% . T he spat ial pat tern of EOF 3 shows t he diff erence from sout h t o north divided by the Changjiang ( Yangt ze) River. T he vector value in
the sout h is negative but posit ive in the nort h. When t he t ime coef ficient is larger t han zero,
the climate yield decreases in t he south but increases in t he nort h. T he spat ial pat t ern of EOF3
is not remarkable before t he 1970s, but rem arkable since t he 1970s, and t his shows that the
difference of climat e yield bet ween sout h and nort h has increased.
EOF4 is contribut ion rat e of variance accounts f or 8. 0% . It show s the dif ferent change be314
t w een east and west. T he vect or value in the east region is negat ive but positive in the w est reg ion. When the time coef ficient is larger t han zero, t he climate yield decreases in the east but
increases in t he w est. T he different change of climat e yield betw een east and west reg ion w as
remarkable in the 1960s.
T he t ypical years of each EOF during 1949- 1992 is given in T able 4.
T able 4
EO F
Y ears
( t im e coeff icient > 0)
Change of the first 4 EOFs from 1949 to 1992
Pat t ern
Y ears
( t ime coeff icient < 0)
Pat t ern
EO F1
1952, 1953, 1955, 1958 1965,
1966, 1967, 1970, 1971 1973,
1982, 1983, 1984, 1992
Increase of climat e yield
in t he w hole count ry
1949, 1960, 1961, 1962
1963, 1977, 1978, 1981
1988, 1989
D ecrease of climat e
yield in t he whole count ry
EO F2
1954, 1974, 1975
D ecrease of climat e
yield in t he southeast
but increase in t he
nort hw est
1951, 1968, 1969, 1985
1986, 1987
Increase of climat e yield
in t he sout heast but decrease in t he nort hw est
EO F3
1976, 1990
D ecrease of climat e
yield in t he sout h but
increase in t he nort h
1980 , 1991
Increase of climat e yield
in t he sout h but decrease in t he nort h
EO F4
1956, 1957, 1964, 1972
D ecrease of climat e
yield in the east but increase in t he w est
1950, 1959, 1979
Increase of climat e yield
in t he east but decrease
in t he west
3. Analysis of t he Impact of Climate Change on Climate Yield
Compared to the st udies of the t emporal and spat ial variat ion of t emperat ure and precipit at ion in China( Chen et al . , 1996) , t he relat ionship betw een the change of climat e and climat e
yield is obvious. T he result of EOF 1 , w hich shows the great increase during the 1950s and
1960s but decrease during t he 1970s and 1980s in the w hole country, is consistent to t he decrease t endency of precipit at ion in China f rom t he 1970s t o 1980s ( Wang, 1996) . T his shows
the import ant role of precipitat ion change in climat e yield change in China. T he pat t ern of
EOF1 mainly reflect s the role of precipit at ion in clim at e yield. T he decrease tendency of precip-i
t at ion in China from the 1970s w ill be harmful t o grain yield. Alt hough t he precipit at ion increases since t he 1980s, it is still m uch low er t han t hat in t he 1950s. T he result of EOF2 m ainly ref lects t he im pact of reg ional precipit at ion change on climate yield. In the 1950s and the lat e
of t he 1970s, t he clim at e yield increased w ith t he increasing of precipit at ion in nort h China,
but in the 1960s and 1980s, t he same phenomenon happened in t he southern China. T he result
of EOF3 mainly reflects the impact of t he t emperat ure change on g rain y ield. During last 40
years, temperat ure t ends to rise in t he north, but in t he south it did not increase or even a lit tle
decrease ( T ang, 1996) . So the climat e yield increased in north China but decreased in the
sout h. It is not iceable t hat t he pat tern of EOF3 is more t ypical since the 1970s wit h t he tem per315
ature rising, this show s t hat the impact of t emperat ure change on climat e yield w ill be great er
w hen t he climat e becom es w armer and w armer.
IV. CONCL USIONS AND DISCU SSIONS
Several conclusions could be drawn based on the above analysis:
( 1) T he climate yield, t aking out t he trend yield from t he tot al g rain yield, w ill indicat e
clearly t he influence of t he climate chang e on grain yield. In the most of the cases t he im pact of
climate chang e on grain y ield is t he uniform in t he w hole country, t he grain y ield increases
some years but decrease in some ot her y ears all over the country. Some years it is different f rom
different regions, climate yield increase in some regions but decrease in t he ot hers.
( 2) T he f irst 4 EOF accumulat ed variance is 62. 9% , which basically reflects the main
temporal and spat ial variat ion of the climate y ield change.
( 3) T he climate yield chang e is mainly affect ed by the change of precipit at ion for the more
than 40 years, however, aft er the 1980s, t he influence of temperature change on climate yield
w as greater because of t he t emperat ure rising . EOF 1 and EOF2 show the ef fects of the change
of precipitation, and EOF3 indicat es the eff ect of t he change of t em perature.
( 4) T he grain y ield is t he result of bot h social econom ic and natural condit ions. It can be
div ided into trend yield and climat e yield. T he trend y ield is caused by t he social and economic
factors, w hich is also influenced by t emperat ure rising, and t he climate yield ref lects t he fluct uation of climate, w hich could be also influenced by policy change.
R EF ERENCES
Ch en Y uqong, Li H ui, 1996. The spat ial and t emporal distribut ion of annual and seasonal t emperat ure in about 40 years of Chi
na by EO F method. In: The S t udy of Cl i mate Change a nd M od els, V ol. 1. Beijing: M eteorological Press. 207- 214. ( in
Ch inese)
G ao S uhua et al . , 1993. T he possible green house impact of atmospheric CO2 content in creasing on t he agricultural product ion
in fut ure in China. Scie nti a A t mospher ica S ini ca. 17( 5) : 584- 591. ( in Chinese)
G uo Jianping et al . , 1996. Th e impact of atmospheric CO 2 cont ent double increasing on t he eff iciency of agricult ural climate resources. In: Wang Futang ( Ed. ) . The S t udy Impact of Cli mat e Change on Agri cul tur e in China . Beijing: M et eorological
Press. 77- 81. ( in Chinese)
T ang Guoli, 1996. The analysis of t emperat ure normal sequence trend in China. In: The S tudy of Cl imat e Change and M odel s
Vol. 1. Beijing: M et eorological Press. 196- 199. ( in Ch inese)
W ang Bomin, 1996. The analysis of precipit at ion normal sequence trend in China. In: The S tudy of Cl imat e Change and M odels, V ol. 1. Beijing: M et eorological Press. 200- 206. ( in Chinese)
316