ACTA AGRONOMICA SINICA

Volume 34, Issue 9, September 2008

Online English edition of the Chinese language journal

Cite this article as: Acta Agron Sin, 2008, 34(9): 1588–1597. RESEARCH PAPER

Impacts of Changes in Climate and Its Variability on Food

Production in Northeast China

JIN Zhi-Qing* and ZHU Da-Wei

Institute of Agricultural Resources and Environment, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, China

Abstract: On the basis of the output of 3 General Circulation Models (GISS, GFDL, and UKMO GCMs), combined with the local
current daily weather data from 1961 to 2000 (Baseline) at 19 sites and the 3 hypotheses about increase in climatic variability (CV) in the
future, 9 scenarios of (CC+!CV) involving both climate change (CC) and its variability (!CV) were generated at the 19 sites in 3
agro-ecological zones in Northeast China using the Weather Generator WGEN as a tool. Four crop models, SOYGRO, CERES-Maize,
CERES-Wheat, and CERES-Rice in DSSAT, were used as effect models and their parameter modification, validation, and sensitivity
analyses were carried out using the baseline weather, statistical yield data of the 4 crops, and the local typical soil data. When running the
effect models under both baseline and various (CC+!CV) scenarios, with the comparison between the simulated outputs, the potential
impacts of changes in both climate and its variability on regional food production with a doubling of CO2 were assessed. The results
showed that the 4 effect models were available in the 3 agro-ecological zones and could be used as tools in climate impact studies.
Generally, climate change would be favorable for soybean [Glycine max (L.) Merr.] and rice (Oryza sativa L.) productions in the studied
regions, especially in the Northern Cold Zone and Eastern Wet Zone. In contrast, it was unfavorable for maize (Zea mays L.) and spring
wheat (Triticum aestivum L.) productions. Particularly, the maize yields would reduce significantly under all scenarios. With the increase
of CV, the rainfed crops, such as soybean, maize, and spring wheat, were negatively affected in decreases of yield and yield stability,
whereas, the irrigated rice was in less influence.

Keywords: Northeast China; food production; climate change; climate variability; modeling

Northeast China, where abounds in soybean [Glycine max (L.) predicted to be 6 to 7°C higher than that at present based on the
Merr.], maize (Zea mays L.), spring wheat (Triticum aestivum L.) 3 GCMs (GISS, GFDL, and UKMO) with a doubling of CO2
and rice (Oryza sativa L.), is one of the biggest bins and the (555 !mol mol−1) [5, 6]. Thus, farmers would experience an
largest commodity grain bases in China. In the past 3 decades, unprecedented high temperature in the future century. The
there has been a steady increase in crop yields due to progress of potential effects of such climate change and its variability on
science and technology, as well as enhanced input of the local food production have aroused great attentions of central and the
governments [1]. On the other hand, Northeast China is one of the local governments, scientists and food producers.
most susceptive areas to climate change in the country. It has The possible geographical shift of the safe northern limit for
been reported that there was an increase in mean temperature of winter wheat in Northeast China has been assessed using
0.34°C per 10 years in the past 5 decades [2]. The meteorological CERES models combined with certain climatic indices and
disasters, such as droughts, floods, and cold damage, have outputs of the 3 GCMs, and several adaptation strategies for
become more frequent with increase in climatic variability (CV), local food crop production have been suggested [7]. Although the
which make Northeast China to be one of the areas with the average changes in climatic elements have been studied in
greatest fluctuation in grain yields in China [3]. According to generating the climate change scenarios in recant years [8–11], the
Intergovernment Panel on Climate Change (IPCC)ˈthe increase CV change is neglected. In fact, there is a close relationship
in mean earth temperature would be 1.8–4.0°C at the end this between CV and meteorological disasters, which are extremely
century [4]. The mean annually temperature in Northeast China is important to crop production. Moreover, in the previous studies,

Received: 11 January 2008; Accepted: 25 March 2008.

* Author for English translation and correspondence. E-mail: [email protected]
Copyright © 2008, Crop Science Society of China and Institute of Crop Sciences, Chinese Academy of Agricultural Sciences. Published by Elsevier BV. All rights reserved.
Chinese edition available online at http://www.chinacrops.org/zwxb/
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

the genetic parameters of CERES models were modified Climate Center of China Meteorological Bureau. Soil data,
according to field experimental data of more than 3 successive including soil classification, pH value, organic carbon content,
years with local representative cultivars using the Trial and Error grain size, and bulk density, were obtained from the State Soil
method. However, the rationality is suspect for representation of Survey Service of China (1993) [13] and used as inputs of the
the data and the method to modify parameters. In this study, the models. The mean annual yields (1981–2000) of soybean, maize,
possible changes in CV were added to the regional climate wheat, and rice at the 19 sites were collected as crop data from
change scenarios, and field-experimental data of short years was the statistical data of Chinese Academy of Agriculture Sciences.
replaced by long serial regional statistical data, which had been Other data, such as the growth duration in days, planting density,
deducted the influence from science and technology and row spacing [14] were used to regulate the genetic parameters
advancement, to ensure the evaluation results more scientific and and validate the models.
reasonable.
1.3 Generation of the local climate change scenarios
1 Materials and methods There were 2 types of doubled CO2 climate change scenarios
in this study, namely CC and CC+!CV scenarios. The CC
1.1 Division of agro-ecological zones and sites selected scenarios considered only climate change, whereas CC+!CV
In Northeast China, soybean, maize, wheat, and rice are all scenarios considered both changes in climate and its variability.
single crops and their growing seasons are similar. Therefore, it These 2 types of scenarios possess 40 years data in length, the
is reasonable to employ the classification for spring wheat [12] to same with the baseline weather data. The CC scenarios were
divide the area into 3 agro-ecological zones, and 19 generated to combine outputs of the 3 GCMs (GISS, GFDL, and
representative sites were selected (Fig. 1). When running the UKMO) with local baseline weather [15]. The CC+!CV
effect models under both the baseline weather and the climate scenarios were created using WGEN [16, 17] in DSSAT [17, 18] as a
change scenarios, the rain-fed pattern was selected for the dry tool in combination with the CC scenarios and 3 hypotheses
farming crops (soybean, maize, and spring wheat), because it is concerning change in CV with a doubling of CO2, i.e., CV
very common in Northeast China, and the automatic irrigated would increase respectively 0 (!CV = 0, the control), 10%
pattern was selected for paddy rice, due to rice cultivation can be (!CV = 10%), and 20% (!CV = 20%), compared with that at
found only in those areas where the irrigation conditions are present. Three steps to generate the CC+!CV scenario were as
favorable. follows: first, to input the CC scenario data at each site into
WGEN for obtaining the initial variances for both temperature
1.2 Data sources and precipitation; second, to multiply the initial variance of
temperature by 100%, 110%, and 120%, respectively for getting
The historic climatic data (Baseline, 1961–2000) to generate
the new temperature variances corresponding to the 3
scenarios comprised the daily maximum and minimum air
hypotheses; finally, to put the new variances into WGEN again
temperature, sunshine duration in hour, and precipitation, which
for generating the CC+!CV scenarios at each site. However, the
were taken from 19 meteorological stations, provided by the
second step to obtain the new variance for precipitation was
more complicated, because precipitation was assumed to obey
the incomplete gamma (") distribution in WGEN, and there
were 2 parameters, # (shape parameter) and $ (scale parameter),
required to be estimated [19]. According to the description by
Chen et al. [20], values of # and $ in the new variances could be
obtained by dividing (or multiplying) that in the initial variances
by square of 100%, 110%, and 120%, respectively.

1.4 Effect models

To assess the potential impacts of changes in climate and its
variability on food production in the studied region, 4 crop
models in DSSAT, i.e., SOYGRO, CERES-Maize, CERES-
Wheat, and CERES-Rice, were selected as the effect models.
These models are strong in mechanism and can be used to
simulate the influences of climate, soil, genotypes, and field
managements on crop growth, development, and yield formation.
Fig. 1 Three agro-ecological zones and 19 representative In the past 20 years, these models have been widely validated
sites in Northeast China and applied in many countries [21, 22]. Since 1990, the DSSAT
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

models have been chosen as the effect models to evaluate data in odd years. Because the genetic parameters modified in
impacts of global climate change on crop production. this way were based on the yield data at a county level, they
The effect models have been modified to simulate the no longer described a given crop variety grown in a specific
physiological effects of CO2 enrichment on crop photosynthesis field. Thus, such genetic parameters were regarded as
and transpiration, because crops growing in higher CO2 “regional parameters” for a crop [7].
concentration exhibit increase in photosynthetic rates and According to validation results for the 4 crops (Fig. 2), the
closure of stomata, leading to raising yields and reducing correlation coefficients between the simulated yields and
transpiration and improving water use efficiency [23, 24]. However, statistical yields are significant at 1% statistical level, and the
there is a distinct difference between C3 and C4 crops in response fit equations are all in perfect slopes (approximately equal to 1)
to CO2 enrichment, due to their distinctions in biochemistry and with small intercepts. Moreover, there is a great agreement in
anatomy [24]. In this study, such difference was treated using growth duration in days between the observations and
“coefficient correction” method, in which the photosynthetic (or simulations with an error less than 4 d (data not shown).
transpiration) rate under doubled CO2 was equal to that under These results indicate that the 4 effect models are suitable in
330 !mol mol−1 to multiply the coefficient of correction C (or R). the studied region.
The C values for different crops were determined with the ratios
1.6 Sensitivity analysis of the effect models
of photosynthetic rates at the 2 levels of CO2 [22]. The R values
were taking the ratio of the stomata resistances (%) at the 2 levels The effect models were further tested by analyzing their
of CO2 [25, 26] and could be calculated by Penman-Monteith sensitivities to changes in temperature, precipitation, and their
equation. Table 1 gives the values of C and R for the 4 crops. It variability in several sites. WGEN was used again to create
can be seen that soybean benefits the best in photosynthesis from the new weather files with 2 or 4°C increase in temperature,
the elevated CO2, followed by rice and wheat, and maize was the ±20% changes in precipitation, as well as increase in 10% or
least. However, the ratio of stomata resistances (%) for maize (C4 20% in climatic variability. Then the effect models were run
plant) is much higher than that of the other 3 crops, indicating respectively under both the new weather files and baseline
that maize would gain more profit than the others in increasing weather, and finally comparisons were made between the
water use efficiency. simulation results in these 2 cases (Tables 2 and 3). As shown
in Table 2, the simulated yields of soybean at Fujin were
1.5 Modification and model validation of sensitive to both temperature and precipitation. When !P = 0
genetic parameters and !T = 2°C, the yields increased by 2% due to improvement
Genetic parameters in the effect models were modified of heat condition; when !P = 0 and !T = 4°C, the yields
using the Trial and Error method referencing DSSAT User’s reduced by 4% because of significant shortening of growth
Guide [6, 17]. First, modifying the genetic parameters related to duration. There were only a little yield fluctuation when !T =
growth duration in days for each crop at a single site using the 0 and !P = −20%, but a large increase in temperature (!T =
phenological and soil data obtained from the local field 4°C) plus a sharply decrease in precipitation (!P = −20%)
experiments and the daily weather data during the same would lead to a significant yield drop. The results for spring
period, and then fixed them for each site and crop. Second, wheat at Mohe were similar with that for soybean, but the
modifying the other genetic parameters related to crop yield simulated yields were more sensitive to precipitation, due to a
using the local statistical yield data (1981–2000) in even years large amount of water requirement. The maize simulated
after deducting the trend yields resulted from science and yields at Baicheng seemed to be more sensitive to temperature,
technology advancement and the synchronous weather data. but insensitive to precipitation. A perhaps reason was that
Finally, validating the effect models using the statistical yield maize had a better environmental adaptation to water stress.
As to the paddy rice at Tonghua, the simulated yields were
only sensitive to temperature, but extremely insensitive to
Table 1 Photosynthetic ratio (C) and stomata resistances (%) precipitation, because the automatic irrigation pattern was
used in the 4 effect models to simulate physiological selected when running CERES-Rice. The results for 4 crops
effects of CO2 enrichment were all coincident with the local crop practice.
Photosynthetic ratio Stomata resistances (s m−1) To analyze the model’s sensitivities to changes in
Crop
A B A B variability of temperature [!CV (T)] and precipitation [!CV
Soybean 1.00 1.21 34.4 49.7 (P)], the effect models were respectively run under different
Maize 1.00 1.06 55.8 87.4 !CV (T) and !CV (P) scenarios given by WGEN. In the case
Spring wheat 1.00 1.17 34.4 49.7 of !CV (P) =0, the simulated yields for the 4 crops decreased
Rice 1.00 1.17 34.4 49.7 with increase in !CV (T). However, in the case of !CV (T) =
A: CO2 concentration of 330 !mol mol−1; 0, only the dry farming crops (soybean, wheat, and maize)
B: CO2 concentration of 550mol mol−1. behaved significant yield reductions with increase in !CV (P).
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

Simulated yield (t ha−1)

Simulated yield (t ha−1)

Simulated yield (t ha−1)

Simulated yield (t ha−1)

Fig. 2 Comparisons between simulated yield and statistical yield at 19 sites in Northeast China

Table 2 Sensitivity analysis of the 4 effect models to changes in temperature and precipitation in Northeast China (%)
Change in temperature Change in precipitation
Crop Site and agro-ecological zone Effect model
(°C) !P = 0 !P = +20% !P = −20%
Soybean Fujin, Eastern Wet Zone SOYGRO 0 0 2 −3
2 2 4 0
4 −4 −1 −6
Maize Baichen, Western Arid Zone CERES-Maize 0 0 1 −2
2 −7 −5 −9
4 −17 −14 −20
Spring wheat Mohe, Northern Cold Zone CERES-Wheat 0 0 8 −12
2 2 9 −6
4 −7 0 −17
Rice Tonghua, Eastern Wet Zone CERES-Rice 0 0 0 0
2 −6 −6 −6
4 −11 −11 −11

When !CV (T) and !CV (P) increased together, the simulated elements would change, but also the meteorological disasters,
yields of the 4 crops would further decreases (Table 3). These such as droughts, floods, and extremely temperature events,
results showed that increase in climatic variability would be would become more frequently with increase in CV, which
unfavorable to crop production, and the effect models could would bring a great shock to the local food production.
be adapted as tools in this climate impact study. Although there are few heat injuries at present in Northeast
China, it is predicted to be one of the most crucial disasters in
1.7 Indices of weather disasters
the future. To estimate the possible trend, extent, and degree of
When CO2 is doubled, not only the average climatic meteorological disasters, 4 indices of disastrous weather were
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

Table 3 Sensitivity analysis of the 4 effect models to change in temperature and precipitation variability in Northeast China (%)
Change in temperature Change in precipitation variability
Crop Site and agroecological zone Effect model
variability (%) !CV (P) = 0 !CV (P) = 10% !CV (P) = 20%
Soybean Fujin, Eastern Wet Zone SOYGRO 0 0 −2 −5
10 −2 −5 −10
20 −4 −10 −13
Maize Baichen, Western Arid Zone CERES-Maize 0 0 −3 −9
10 −8 −10 −11
20 −12 −14 −21
Spring wheat Mohe, Northern Cold Zone CERES-Wheat 0 0 −2 −4
10 −2 −7 −8
20 −4 −9 −10
Rice Tonghua, Eastern Wet Zone CERES-Rice 0 0 −1 −1
10 −2 −3 −4
20 −3 −5 −6

selected as follows: increase of !CV, except for the Western Arid Zone and the
& A serious low-temperature year is defined as the year Eastern Wet Zone in the GISS scenario, but the differences
in which the sum of mean monthly temperature during between scenarios and the baseline weather did not reach the
crop growing season departure from the mean is less 5% probability level. As to the seasonal drought, it would
than or equal to –4°C [27]; abbreviated as LTY; decrease under all the !CV0 scenarios compared with the
& A hard rain day is defined as the day with daily baseline, because more precipitations were observed in the 3
precipitation greater than or equal to 50 mm [28]; zones according to the outputs of 3 GCMs. However, it was
& A seasonal drought is defined as the period in which insufficient to conclude that the drought threat would slow
the sum of daily precipitation in 20 successive days is down in the studied regions owing to the simultaneous
less than or equal to 5.0 mm; and acceleration of crop evapotranspiration led by high temperature.
& A high-temperature day is defined as the day with On the contrary, the frequency of seasonal drought would be
maximum daily temperature greater than or equal to accelerated when !CV was enlarging. In summary, the heat
35°C, abbreviated as HTD. injury would be the primary climatic adversity in the future.
Using these indices, it was possible to account for the High temperature would not only shorten crop growth
number of days for HTDs, LTYs and occurrence probabilities duration, but also affect pollen fertilization and finally result
for hard rains and seasonal droughts from both the baseline in yield reduction.
weather and scenario files.
2.2 Impacts of climate change and its variability on
2 Results soybean yield

When climate change was considered alone, the soybean

2.1 Trends of climate disasters under (CC+!CV) scenarios
yields simulated under the 3 GCM-based scenarios in both
If only consider the case of climate change alone (!CV0 Northern Cold Zone and Eastern Wet Zone were remarkably
scenario), the HTDs in the 3 agro-ecological zones would higher than that under the baseline weather with a 100% and
significantly increase under the 3 GCM-based climate change 30% increase, respectively (Fig. 3). The reasons for yield
scenarios. The HTDs in Western Arid Zone were much more increase might be CO2 enrichment and the improvement of
than that in Northern Cold Zone and Eastern Wet Zone. And the thermal condition, especially in the Northern Cold Zone
HTDs under the UKMO scenario were much more than that where heat condition is obviously deficient at present. It was
under the GISS and GFDL scenarios. Moreover, the HTDs more complicated in the Western Arid Zone where soybean
under all the 3 scenarios would further increase with increasing yields increased by 39% on an average under the GISS-based
of !CV. There would be no significant change in LTYs, scenarios, while decreased by 26% and 37% under the GFDL
compared the (CC+!CV) scenarios with the baseline weather, and UKMO-based scenarios, respectively. These results were
indicating that chilling damage would exist in the future despite related to the changes in temperature and precipitation
of its rather low probability (Table 4). predicted by various GCMs. Under the GISS scenario, air
When considering the case of climate change alone, the temperature elevated 2.9°C and the precipitation increase by
frequency of hard rains showed an ascendant trend with 40%, thus the simulated soybean yields evidently improved.
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

Table 4 Frequency of weather disasters under BASELINE and 3 GCM-based CC+!CV scenarios in Northeast China
GISS GFDL UKMO
Agroecological zone Weather disaster Baseline
!CV0 !CV10% !CV20% !CV0 !CV10% !CV20% !CV0 !CV10% !CV20%
Northern Cold Zone Heat injury 0.8 5.5 7.6 9.3 8.0 9.9 11.7 32.1 34.7 36.8
Chilling damage 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
Hard rain 0.6 0.5 0.8 1.0 0.5 0.7 0.8 0.5 0.8 1.0
Seasonal drought 1.0 0.5 0.6 0.7 0.5 0.5 0.8 0.5 0.5 0.7
Eastern Wet Zone Heat injury 0.4 6.2 8.1 10.1 9.3 10.9 13.0 29.6 31.9 34.0
Chilling damage 0.1 0.1 0.1 0.1 0.0 0.1 0.0 0.0 0.1 0.1
Hard rain 0.8 1.2 1.4 1.8 1.0 1.2 1.5 0.8 1.0 1.4
Seasonal drought 0.6 0.4 0.5 0.5 0.4 0.5 0.6 0.4 0.5 0.6
Western Arid Zone Heat injury 4.0 17.2 20.4 21.5 28.4 31.3 31.9 49.8 52.7 52.7
Chilling damage 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.1 0.1
Hard rain 0.8 1.7 2.0 2.3 0.5 0.8 1.0 0.9 1.0 1.5
Seasonal drought 1.3 0.8 0.9 0.9 1.0 1.1 1.2 0.9 1.0 1.2
Weather disasters are indicated with indices of days for maximum daily temperature '35°C, frequency of serious low temperature year, days of hard
rain, and frequency of seasonal drought, respectively.

120 Northern Cold Zone 60 Eastern Wet Zone 60 Western Arid Zone

100 40
Change in yield (%)

80 40 20

60 0

40 20 -20

20 -40

0 0 -60
GISS GFDL UKMO GISS GFDL UKMO GISS GFDL UKMO
General Circulation Model (GCM)

!CV0 !CV10% !CV20%

Fig. 3 Percentage change in soybean yields simulated under various CC+CV scenarios based on 3 GCMs
compared with the base yields in the 3 agro-ecological zones of Northeast China

Under the GFDL and UKMO scenarios, temperature precipitation was uneven, for example, the average days for
increased 4.4 to 7.3°C compared with the Baseline weather, hard rains in Northeast China was 0.9 d annually under the
and the number of days for maximum daily temperature !CV0 scenario, whereas it was 1.1 and 4.0 d under the
'35°C increased from current 4.0 to 28.4–49.8 d y−1 with !CV10%, and !CV20% scenarios, respectively. On all
effective doubling of CO2. In addition, the increase in accounts, with the increase of !CV, weather disaster events,
precipitation predicted was less than 10%. All of these could such as extremely high temperature, hard rain, and seasonal
accelerate soybean development, result in shortening drought, tended to be more frequent and led to negative
photosynthetic time and insufficient of seed filling. Moreover, influences on soybean yield. However, climate change was
the enhanced evapotranspiration would aggravate the seasonal generally favorable for soybean production in Northeast
droughts and finally lead to serious yield reduction. China, because the average simulated soybean yield could
Under the case of both CC and !CV, the simulated soybean increase 51, 29, and 9% than the base yield, under GISS,
yields decreased compared with the control under !CV0 GFDL, and UKMO with !CV20% scenario, respectively.
scenario (Fig. 3). The simulated yields reduced averagely
2.3 Impacts of climate change and its variability on
from 2 to 11% under the !CV10% scenario and from 3 to
maize yield
29% under the !CV20% scenario, respectively. This was
because that the days of maximum daily temperature '35°C The simulated maize yield decreased throughout the studied
increased from 19.0 to 21.4 (!CV10%) and 23.0 d (!CV20%) region when !CV was ignored, which was completely
and during the growing season and the distribution of contrary to soybean yield (Figs. 3 and 4). This could be
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

explained with the growth duration and photosynthetic such disaster (Table 3).
characteristic of maize. The higher temperature could shorten
2.4 Impacts of climate change and its variability on
the growth duration of crops, resulting in the insufficiencies of
spring wheat yield
photosynthesis and grain filling. As being a C4 crop, maize
could rarely get direct benefits from CO2 enrichment. Among Under the condition that !CV was ignored (only climate
the 3 agro-ecological zones, the Western Arid Zone was the change was considered), the simulated yields of spring wheat
largest loser in maize yields with the percentage of 17%, sharply decreased in the 3 agro-ecological zones under the
followed by the Northern Cold Zone with 9% and the Eastern UKMO scenario with the largest percentage of 56%. The
Wet Zone with 6%. The large increase in temperature was wheat yield decreased by 7% under the GFDL scenarios
regarded as the main reason for maize yield reduction in the although there were small increases in some cases. However,
Western Arid Zone and the change in precipitation played under the GISS scenario, wheat yield increased slightly in
unimportant role, because maize could close the stomates to most cases. The average yield reduced the least in the Eastern
weaken leaf transpiration under a doubled CO2 concentration, Wet Zone, followed by the Northern Cold Zone. In the
thus, the water use efficiency was improved greatly. A good Western Arid Zone, yield decreased under most scenarios
example was the 5% reduction of maize yield under the GISS except for the GISS-based one. The regional differences based
scenario even though precipitation increased by 40% during on various scenarios were not only because of the regional
the growing season (Fig. 4). The Northern Cold Zone and disagreements on the increase amplitude of temperature and
Eastern Wet Zone showed smaller yield reductions than the the precipitation but also in relation to the strict and large
Western Arid Zone subject to less increases in temperature in requirement of water in wheat. In the Eastern Wet Zone, the
the 2 zones. Among the 3 GCM-based scenarios, yield sufficient precipitation with good soil moisture was favorable
reduction was the largest under the UKMO scenario (16%), for wheat growth, thus, the yield increased or decreased
followed by the GFDL and GISS scenarios with reduction slightly even under the high temperature condition (Fig. 5). In
percentages of 11% and 5%, respectively (Fig. 4). The increase the Northern Cold Zone, yield reduced slightly owing to
amplitude of temperature might be the reason of such relatively sufficient soil moisture and low evapotranspiration.
differences. However, in the Western Arid Zone, in addition to rare
Considering the case of CC+!CV, the simulated maize precipitation, high temperature strongly enhanced evapo-
yields further reduced with the increase of !CV (Fig. 4). transpiration and caused a serious yield reduction under the
Compared with the control (!CV0 scenario), the average “extreme” UKMO scenario. The yield increase in this zone
maize yield decreased by 2% and 4% under !CV10% and under the GISS scenario was probably related to predicted
!CV20% scenarios, respectively. These results were increase of precipitation.
coincident with the conclusion by Lal et al. [9], which was In this paper, a concept called evapotranspiration ratios ($)
from a study on the simulated influences of !CV on maize was employed to illuminate the key role of water condition on
yield in Madhya of India using CERES-Maize model. The yield formation in spring wheat [26], which was defined as the
further reduction of maize yield was mainly on account of the ratio of actual evapotranspiration to potential evapotranspiration
increases of days of maximum daily temperature '35°C, day during the growing season. The results showed that the $ values
of hard rain, and frequency of seasonal drought. Moreover, was approximately 0.79 under the GISS and GFDL scenarios
chilling damage also existed in spite of low probability of for maintaining the growth and development of wheat; whereas

General Circulation Model (GCM)

GISS GFDL UKMO GISS GFDL UKMO GISS GFDL UKMO
0 0 0
Change in yield (%)

-10 -10 -10

-20 -20 -20

Northern Cold Zone Eastern Wet Zone Western Arid Zone

-30 -30 -30

!CV0 !CV10% !CV20%

Fig. 4 Percentage change in maize yields simulated under various CC+CV scenarios based on 3 GCMs
compared with the base yields in the 3 agro-ecological zones of Northeast China
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

20 20 20
Change in yield (%)
0 0 0

-20 -20 -20

-40 -40 -40

-60 -60 -60

Northern Cold Zone Eastern Wet Zone Western Arid Zone
-80 -80 -80
GISS GFDL UKMO GISS GFDL UKMO GISS GFDL UKMO

General Circulation Model (GCM)

!CV0 !CV10% !CV20%

Fig. 5 Percentage change in simulated yields of spring wheat under various CC+CV scenarios based on 3 GCMs
compared with the base yields in the 3 agro-ecological zones of Northeast China

it was only 0.55 under the UKMO scenario, which could not with the base yields. On the basis of agro-ecological zones,
meet the requirement of spring wheat. the average yields simulated under the 3 GCM-based
Increases in !CV would bring a negative effect on yield of scenarios increased by 17% in the Northern Cold Zone and
spring wheat (Fig. 5). The average simulated yields decreased both 11% in the other 2 zones. Rice is a warm-favorite crop,
respectively by 4–6% and 7–9% under the !V10% and therefore, increase in temperature might be advantageous to
!V20% scenarios compared with the control (!CV0 scenario). rice yield production. The CO2 enrichment in the air has
This might be resulted from more frequent weather disasters, also a positive effect on rice yields, because of the
such as droughts, floods, and extreme high temperature characteristics of photosynthesis in C3 crops. These positive
events. effects could compensate the negative effects due to
shortened growth duration resulted from temperature
2.5 Impacts of climate change and its variability on
elevation.
rice yield
With the increase of !CV, the rice yields simulated
If considering the case of climate change alone, the decreased under all the CC+CV scenarios in the 3
simulated rice yields increased under all scenarios agro-ecological zones of Northeast China. Compared with the
compared with the base yields, indicating that climate control (!CV0) scenario, the rice yields simulated decreased
change has a positive effect on rice yields in the studied averagely by 1% and 2% under the !CV10% and !CV20%
region, if irrigation water is sufficient. On the basis of the scenarios, respectively (Fig. 6). However, paddy rice was
scenarios, the simulated rice yields increased by different from the other dry farming crops, it would be less
approximately 15% under both the GISS and GFDL influenced by increase in !CV, due to its stronger tolerance to
scenarios and 10% under the UKMO scenario, compared high temperature and perfect irrigation condition.

25 Northern Cold Zone 25 Eastern Wet Zone 25 Western Arid Zone

Change in yield (%)

20 20 20

15 15 15

10 10 10

5 5 5

0 0 0
GISS GFDL UKMO GISS GFDL UKMO GISS GFDL UKMO
General Circulation Model (GCM)

!CV0 !CV10% !CV20%

Fig. 6 Percentage change in rice yields simulated under various CC+CV scenarios based on 3 GCMs
compared with the base yields in the 3 agro-ecological zones of Northeast China
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

2.6 Analysis of yield stability bring a big loss to food production. However, it is very
difficult to predict CV change with a changed climate,
In the case of only climate change considered, the yield
because there are a complicated and uncertain relationship
variability coefficient (YVC) for spring wheat was 36% in an
between changes in average climate elements and occurrence
average (31%, 33%, and 44% under 3 GCM-based scenarios,
of extreme climatic events. In this paper, change in CV is
respectively), which was the highest among the 4 crops. The
taken into account in the regional climate change scenarios by
other 3 crops ranked as soybean of 25% (21%, 27%, and 28%
several hypotheses. The results will help to understand the
under 3 GCM-based scenarios), maize of 22% (14%, 28%,
directions and extents of impacts of climate change and its
and 27% under 3 GCM-based scenarios), and rice of 12%
variability on food production in Northeast China though
(11%, 12%, and 14% under 3 GCM-based scenarios). Rice
some subjective factors cannot be avoided completely.
had the lowest YVC due to sufficient irrigation water. Among
In this impact study, agriculture techniques at present are
the 3 agro-ecological zones, the mean YVC for 4 crops in
assumed to be maintained in the future, i.e., there would be no
Western Arid Zone was the largest (31%), followed by
changes in cropping systems, varietal dispositions and field
Northern Cold Zone (22%) and Eastern Wet Zone. (20%)
managements, no obvious soil problems and the insects,
With respect to GCMs, average YVC under UKMO, GFDL,
diseases, and weeds would be in good control. Doubtlessly,
and GISS scenario were 28%, 25%, and 19%, respectively.
the farmers in future will adopt some new techniques and
For the 3 rainfed crops, the average YVC were obviously
measures for adapting to climate change. However, it is
higher under !CV20% than that under !CV10% scenario,
difficult even impossible to exactly estimate changes in every
which were 3% and 7% larger than that under !CV0 scenario
issue in the future. Thus, the agricultural technologies were
(control), respectively. However, irrigated rice was less
assumed to be unchanged in most impact studies. These
influenced by !CV with only 1% and 2% of YVC increase
assumptions aim at making the comparison between current
than the control (Table 5).
and predicted status feasible and convenient.
The 3 GCMs used in this study are equilibrium models which
3 Discussion
are commonly used in the world. The equilibrium here is
A changed CV associated with climate change is closely defined as an instantaneous state of global climate in future,
related to the probability of weather disasters, which often corresponding with an effective doubling of CO2. The purpose

Table 5 Coefficients of yield variability for the 4 crops under various CC+CV scenarios
based on 3 GCMs in the 3 agro-ecological zones of Northeast China
GISS GFDL UKMO Average of models
Crop
!CV0 !CV10% !CV20% !CV0 !CV10% !CV20% !CV0 !CV10% !CV20% !CV0 !CV10% !CV20%
Soybean
NCZ 17 17 29 16 21 28 18 21 29 17 20 29
WAZ 28 28 31 45 49 48 35 37 39 36 38 39
EWZ 18 20 22 21 23 25 31 35 37 23 26 28
Mean 21 22 27 27 31 34 28 31 35 25 28 32
Maize
NCZ 8 9 13 9 14 12 25 27 33 14 17 19
WAZ 19 22 25 52 61 70 32 44 46 34 42 47
EWZ 15 17 21 22 24 26 25 31 36 21 24 28
Mean 14 16 20 28 33 36 27 34 38 23 28 31
Spring wheat
NCZ 27 31 36 24 29 30 41 44 50 31 35 39
WAZ 39 42 43 47 51 54 48 55 54 45 49 50
EWZ 26 29 33 27 29 31 42 46 50 32 35 38
Mean 31 34 37 33 36 38 44 48 51 36 40 42
Rice
NCZ 14 18 20 14 15 21 21 22 23 16 18 21
WAZ 8 8 10 8 8 8 7 8 9 8 8 9
EWZ 11 11 12 13 14 14 14 14 16 13 13 14
Mean 11 12 14 12 12 14 14 15 16 12 13 15
Average of crops 19 21 25 25 28 31 28 32 35 24 27 30
NCZ: Northern Cold Zone; WAZ: Western Arid Zone; EWZ: Eastern Wet Zone.
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

to develop equilibrium models is mainly to avoid the speed of References

global warming, which involves not only the complicated
climate systems, but also a large amount of affairs in economy, [1] He X L, Zhang P Y, Liu W X. Time serial change of grain
society and diplomacy, with a big uncertainty and dispute. yield per hectare and its impact factors in Northeast China. Res
Whenever the global warming rate must be considered, many Agric Modernization, 2006, 27: 360–363 (in Chinese with
English abstract)
transient GCMs, such as GISS Transient B Run, A2, and B2,
[2] Sun F H, Yang X Q, Lu S, Yang S Y. The contrast analysis on
might be more feasible, in which even different plans about
the average and extreme temperature trend in Northeast China.
emission reduction of greenhouse gases were taken into
Sci Meteorol Sin, 2006, 26: 157–162 (in Chinese with English
account. abstract)
Although the 3 equilibrium GCM-based scenarios give [3] Cheng Y Q, Zhang P Y. Regional patterns changes of Chinese
similar results at global level, they predict quite differently at grain production and response of commodity grain base in
regional level. According to the current knowledge, it is Northeast China. Sci Geogr Sin, 2005, 25: 513–520 (in Chinese
impossible to judge which one is accurate in assessment. Thus, with English abstract)
output of the 3 GCMs were used in this study, combined with [4] Intergovernmental Panel on Climate Change (IPCC). Global
the local historic data and 3 hypotheses to generate the regional Climate Projections: the Physical Science Basis. Cambridge
scenarios involving changes both in climate and its variability University Press, 2007. p 749
in the future. The simulated results under these scenarios were [5] Jin Z Q, Chen H, Ge D K, Fang J, Zheng X L. To generate
climate scenarios in China with doubled CO2 by using GCMs
compared each other. Once consistent trend are obtained, it
outputs and historic weather data. Chin J Agrometeorol, 1992,
could be regarded that the occurrence probability of this result
13: 13–21 (in Chinese with English abstract)
might be higher in the future, and vice versa.
[6] Jin Z Q. Impacts of Global Climate Change on Food
Of late, most successful crop growth models were designed Production in China: A Modeling Study. PhD Dissertation of
for crop management under normal climatic conditions. Nanjing Agricultural University, 1996. pp 75–159 (in Chinese
Generally speaking, they could not response rationally to with English abstract)
extreme climatic events, such as floods, hard rains, droughts, [7] Jin Z Q, Ge D K, Shi C L, Gao L Z. Several strategies of food
and heat waves. This needs to be emphasized in further crop production in the Northeast China plain for adaptation to
studies. As expiations, several indices of weather disasters global climate change: A modeling study. Acta Agron Sin,
have been selected based on CC+!CV scenarios in this study 2002, 28: 24–31 (in Chinese with English abstract)
to analyze the possible directions and extents of disasters in [8] Alexandrov V A, Hoogenboom G. The impact of climate
the future. variability and change on crop yield in Bulgaria. Agric For
Meteorol, 2000, 104: 315–327
[9] Lal M, Singh K K, Srinivasan G, Rathore L S, Naidu D,
4 Conclusions
Tripathi C N. Growth and yield responses of soybean in
Madhya Pradesh, India to climate variability and change. Agric
The 4 effect models, SOYGRO, CERES-Maize, CERES-
For Meteorol, 1999, 93: 53–70
Wheat, and CERES-Rice, are suitable in Northeast China and [10] Mearns L O, Rosenzweig C, Goldberg R. Mean and variance
it is reasonable to use these models as tools to assess the change in climate scenarios: methods, agricultural applications,
impacts of climate change and its variability on crop and measures of uncertainty. Climate Change, 1997, 35: 367–396
production. When CO2 is doubled, heat injury would be the [11] Zhao X Y, Zhang H Y. The sustainability of Chinese
major climatic disaster during crop growing season in agricultural yield affected by global climate change. Sci Geogr
Northeast China. With increase in !CV, other weather disasters, Sin, 2000, 20: 224–229 (in Chinese with English abstract)
such as hard rain and seasonal droughts, would be more [12] Jin S B. China Wheat. Beijing: China Agriculture Press, 1996.
frequent. Simultaneously, chilling damage could not be p 34 (in Chinese)
neglected although its occurrence probability seems rather low. [13] State Soil Survey Service of China. China Soil Series, Vol. 5.
Beijing: China Agriculture Press, 1993. pp 2–389 (in Chinese)
Generally, climate change would be favorable for soybean and
[14] Cui D C. Images of Climate Resource for Chinese Major Crops.
paddy riceˈbut unfavorable for maize and spring wheat
Beijing: China Meteorology Press, 1984. pp 122–140 (in
productions in Northeast China in the future With increase in
Chinese)
!CV, the yields would be further reduced, excepting irrigated [15] Rosenzweig C, Parry M L. Potential impact of climate change
rice. on world food supply. Nature, 1994, 367: 133–138
[16] Richardson C W, Wright D A. WGEN: a model for generating
Acknowledgment daily weather variables. U.S. Department of Agriculture,
Agricultural Research Service, ARS-8, 1984. pp 1–83
This study was supported by the National Natural Science [17] IBSNAT Project. DSSAT User’s Guide. University of Hawaii,
Foundation of China (30370815, 30470332). Honolulu, 1992
 JIN Zhi-Qing et al. / Acta Agronomica Sinica, 2008, 34(9): 1588–1597

[18] Ritchie J T, Alocilja E C, Singh U, Uehara G. IBSNAT and the Potential Effects of Global Climate Change on the United
CERES-Rice Model. Weather and Rice. Proceedings of States, Appendix C. Report to Congress. US Environmental
international workshop on the Impact of Weather Parameters Protection Agency, EPA-230-05-89-050, Washington DC, USA,
on Growth and Yield of Rice. International Rice Research 1989
Institute, Los Banos, Philippines, pp. 271-281 [24] Jin Z Q. Different responses of C3 and C4 crops to increase of
[19] Hanson C L, Osborn H B, Woolhiser D A. Daily precipitation CO2 concentration in the atmosphere. World Agric, 1995, (9):
simulation model for mountainous areas. Transactions of the 36–37 (in Chinese)
ASAE 32, 1989. pp 865–873 [25] Jin Z Q, Ge D K, Zheng X L, Chen H. Assessing the potential
[20] Chen C, Jin Z Q, Zheng Y F, Liu J D. Assessing the impacts of impacts of global climate change on maize production in China.
doubled CO2 climate change and its variability on winter wheat Acta Agron Sin, 1996, 22: 513–524 (in Chinese with English
production in the Huang-Huai-Hai Plain of China. Jiangsu J abstract)
Agric Sci, 2004, 20: 7–12 (in Chinese with English abstract) [26] Jin Z Q, Ge D K, Chen H, Zheng X L. Positive and negative
[21] Greenwald R, Bergin M H, Xu J, Cohan D, Hoogenboom G, effects of global climate change on soybean production in
Chameides W L. The influence of aerosols on crop production: China. Soybean Sci, 1994, 13: 302–311 (in Chinese with
a study using the CERES crop model. Agric Syst, 2006, 89: English abstract)
390–413 [27] Wang X Q, Ma S Q, Xi Z X, Wang S L. Long-term forecast of
[22] Smith J B, Tirpak D A. The Potential Effects of Global Climate summer’s chilling damage in Northeast China. J Nat Disasters,
Change on the United State: Report to Congress. Washington 2006, 15(3): 42–45 (in Chinese with English abstract)
DC: U S Environment Protection Agency, 1989. pp 2–54 [28] Huang D Y, Wang K L, Huang M, Chen H S, Wu J S, Zhang G
[23] Peart R M, Jones J W, Curry R B, Boote K, Allen Jr L H. P, Peng T B. Seasonal drought problems in the red soil hilly
Impact of climate change on crop yield in the southeastern region of the middle subtropical zone of China. Acta Ecol Sin,
USA: A simulation Study. In: Smith J B, Tirpak D A eds. The 2004, 24: 2516–2523 (in Chinese with English abstract)