Introduction
Terrestrial hydrological and carbon cycles are intimately coupled via
transpiration and photosynthesis processes which are regulated by plant leaf
stomata. Due to land use/cover changes, intensified agricultural management,
and climatic change, terrestrial eco-hydrological processes have been
noticeably shifted on multiple spatiotemporal scales (Tian et al., 2011;
Douville et al., 2013); for example, prevailing irrigation and application of
chemical fertilizers have raised soil moisture, evapotranspiration (ET), and
crop productivity, etc. In some regions the effects of human activities are
the same magnitude as, or even exceed, the impacts of global warming on the
productions of agro-ecosystems (Haddeland et al., 2014). In the last decades,
global consumptive water use and carbon fixation by terrestrial ecosystems
have been demonstrated to slightly increase with more efficient water use,
corresponding to changes in climatic factors and the fertilization effect of
elevated atmospheric CO2 concentration (Yan et al., 2013; Nayak et al.,
2013). Consequently, spatiotemporal patterns of water and carbon fluxes on
the regional scale are changing under global change (Zeng et al., 2014; Liu
et al., 2012).
As ET is the major component of water budget in the water-limited basins, its long-term tendency
has been taken as a critical indicator for diagnosing the intensification of
the regional water cycle. The complementary relationship between actual and
potential ET may reveal some clues to hydrological changes. Observations
illustrated that potential evaporation rates (ETp) (represented as
pan evaporation) decreased in Europe, the US, China, India, and Australia in
recent decades (Brutsaert, 2006; Katul et al., 2012), implying the decline of
available energy and aerodynamics devoted to latent heat flux over the land
surface. The climatic factors dominating ETp change are usually
diverse. For example, over the North China Plain (NCP) the changes in
ETp are mainly attributed to declines in global radiation and
near-surface wind speed (Tang et al., 2011; Song et al., 2009). However, in
southern Turkey a noticeable decline in ETp was attributed to
enhanced air humidity associated with the expansion of irrigation acreage and
more water evaporated into the atmospheric boundary layer (Ozdogan and
Salvucci, 2004). Burn and Hesch (2007) revealed that decreasing wind speed
and raised water vapour deficit responded to the trend of ETp in
the Canadian Prairies. On a large scale, precipitation is usually the
principal factor determining actual ET change; for example, Qian et
al. (2007) showed that an increase in ET in the Mississippi River basin
followed precipitation propensity, while the effects of solar radiation and
air temperature changes were minor.
Terrestrial eco-hydrological processes are driven by climate and modulated by
human activities. Generally climate warming enhances atmospheric evaporative
demand, while CO2 fertilization stimulates photosynthesis and inhibits
leaf stomatal conductance, leading to more biomass accumulation and higher
water productivity (Field et al., 1995; Buckley and Mott, 2013).
Simultaneously, land use change and land management also noticeably affect
the ecosystem production and hydrological fluxes (Shi et al., 2011).
Separating the contributions of climatic change, CO2 enrichment, and
human activities to the long-term trends of water and carbon cycles is
critical for assessment of ecosystem responses and resilience to
environmental changes. Some researchers have explored the relative
contributions of climate change and vegetation dynamics to changes in global
land surface evapotranspiration and river runoff (Betts et al., 2007; Piao et
al., 2007; Alkama et al., 2010; Liu et al., 2012; Chen et al., 2014, 2015;
Banger et al., 2015), but the conclusions are still inconsistent. Climate
change dominated the inter-annual variability of ET, while land use changes
and agricultural practices and techniques exerted more discernable effects on
the water cycle in the long term (Liu et al., 2012). However, Alkama et
al. (2010) and Shi et al. (2011) demonstrate that climate change is the
predominant driver of the global ET tendency in the 20th century. Commonly,
the contributions of climate change to vegetation productivities on a large
scale are quantified by the ecosystem models or statistical models. For
example, Piao et al. (2015) documented that elevated atmospheric CO2 and
nitrogen deposition were the critical contributors to terrestrial greening
over China; Baker et al. (2010) figured out that climate anomalies in
springtime were the most frequent drivers of annual GPP variations in North America; Nayak et al. (2013)
reported that climate change had a relatively small but significant control
(15 %) on the trend of terrestrial net primary production (NPP) over
India. In the crop ecosystems, contributions of climate change, cultivar
renewal, and agronomic management to change in crop yield have been separated
with the crop or statistical models (Yu et al., 2012; Song et al., 2014; Bai
et al., 2016; Guo et al., 2014; Z. Wang et al., 2016). The impacts of climate
change on crop yield may be positive or negative in different regions,
depending on the tendencies of the dominant factors (Ewert et al., 2015).
As one of the granaries in China, the North China Plain (NCP) is
confronted with challenges of
agriculture sustainability due to global change and social and economic
development. Thereby, it is crucial to understand the impacts of climate
change on the productions of cropping systems. Over the plain, a winter
wheat–summer maize double cropping system prevails, supported by irrigation,
fertilizer, and agronomic techniques. In situ measurements, agricultural
annals, and the regional remotely sensed vegetation index dataset all
illustrate that both wheat and maize productivities have accrued remarkably
since the 1970s (Mo et al., 2009; Yuan and Shen, 2013); correspondingly, the
seasonal water consumption and water use efficiency are also slightly
improved (Zhang et al., 2011). The achievement of long-term increasing grain
production is related to the active adoption of new varieties for
stabilization, extending the length of the crop growth period, as well as
agronomic technique advancement (Liu et al., 2010; Sacks and Kucharik, 2011).
Currently, water amount for food production consists of 65 % of total
water consumption here. Further, along with the gradually augmented domestic
and industrial water requirement, groundwater in some areas of the plain has
been overexploited, and the environmental water requirement is generally
under deficit conditions (e.g. MWR, 2010). Faced with the rapid deteriorating
agricultural environment, some critical issues are still unclear, such as
what mechanisms drive the evolutions of eco-hydrological processes over the
plain? What are the impacts of climate change on the cropping systems?
In this study, the VIP eco-hydrological dynamic model integrated with the
NOAA-AVHRR remotely sensed normalized difference of vegetation index (NDVI)
is employed to assess the spatiotemporal evolutions of ET and vegetation GPP
over the NCP from 1981 to 2013. By numerical experiments with the VIP model
and the factor separation method, the contributions of climate change,
fertilization of atmospheric CO2 enrichment, agronomic practices, and
technological advancement to crop water consumption and productivity are then
analysed and the relevant mechanisms are discussed.
Method and materials
Study region
The NCP is one of the country's granaries, extending from latitude
32∘00′ to 40∘24′ N and longitude 112∘48′ to
122∘45′ E (Fig. 1a, b). It is located in the eastern part of
China, with an area of 40 × 104 km2, which is an alluvial
plain developed by the intermittent flooding of the Huang, Huai, and Hai
rivers, and 85 % is cultivated as farmland. The warm temperate climate
varies gradually from sub-humid in the southern to semi-arid in the northern
parts. The annual precipitation ranges from 500 to 1000 mm, occurring
irregularly among the seasons, and more than 70 % falls in summer. Soil
moisture deficit happens widely during the spring and early summer period.
Besides soybean/millet/sorghum/cotton, the double cropping system of winter
wheat–summer maize prevails in the plain, where wheat and maize are the most
common harvest crops in the summer and autumn seasons, respectively. Due to
insufficient precipitation, the spring crops (such as wheat) usually need
supplemental irrigation to form favourable production.
Land use/cover (a) and soil texture (b) of the
North China Plain (NCP) (DBF: deciduous broadleaf forest; BNF: broadleaf and
needleleaf mixed forest; ENF: evergreen needleleaf forest; EBF: evergreen
broadleaf forest).
![]()
The VIP eco-hydrological model
The physical-process-based VIP (Vegetation Interface Processes)
eco-hydrological model is designed to simulate the exchanges of energy,
water, and carbon between the terrestrial ecosystem and the atmosphere (Mo et
al., 2014). In the model, ET is termed the summation of canopy transpiration
and evaporation from the canopy intercept and the soil surface, computed
separately with the Penman–Monteith equation. Transpiration and
photosynthesis processes are coupled through the Ball–Berry relationship
between leaf stomatal conductance and net assimilation rate. Regarding the
carbon cycle aspect, leaf carbon fixations on sunlit and shaded leaves
are predicted with the biochemical schemes for C3 (Farquhar et al., 1980) and
C4 plants (Collatz et al., 1992). In the radiation budget scheme, shortwave
radiation transfer in the canopy distinguishes the leaf spectral properties
of visible and near-infrared radiation, as well as the fraction of direct
beam and diffusive irradiance in global radiation. Precipitation
throughfall–runoff generation over the land surface is calculated with a
curve-number (CN) type equation on a daily scale, using the daily net
precipitation and the moisture deficit of the upper soil layer. Simulation of
soil water movement in the root zone is carried out with a discrete Richards
equation in three layers. The crop and natural vegetation growth modules are
also embedded in the model to simulate the biomass accumulation and
carbon cycle.
Data
The VIP model input data include land use/cover, soil physical properties,
and atmospheric forcing variables. The GIMMS AVHRR 15-day normalized
difference of vegetation index time series (NDVI3g; values are scaled by 1000)
(https://nex.nasa.gov/nex/projects/1349/wiki/general_data_descriptionand_access/)
(Pinzon and Tucker, 2014) from 1981 to 2013 is used to retrieve the
vegetation leaf area index and other land surface characteristics. The land
use classification originates from both Landsat TM images
(www.resdc.cn) and MODIS remote sensing products, in which the farmland
is classified as rice paddy and dryland. Soil textural data are available on
a scale of 1 : 1 000 000 represented as fractions of sand, silt, and
clay, by which the parameters of soil porosity (θsat) and
saturated hydraulic conductivity (Kws, mm s-1) are estimated
as in Bonan (1996). Daily climate variables (air temperature, water vapour
pressure, wind speed, sunshine duration, and precipitation) recorded at 87
climatic stations (http://data.cma.cn) in and around the study area are
available for generating the spatial atmospheric forces. The NDVI data are
error-checked and the erroneous data are replaced by interpolation with the
preceding and subsequent values according to the time series by the
Savitzky–Golay (SG) filter (Savitzky and Golay, 1964), and then the daily
values are derived with the Lagrange polynomial. Vegetation leaf area index
(LAI) is retrieved from the NDVI with empirical relationships for different
plant functional types.
The data used for model validation are field flux measurements with an eddy
covariance technique at the Yucheng (116∘38′ E, 36∘57′ N), Daxing (116∘25′ E, 39∘37′ N), Miyun
(117∘19′ E, 40∘38′ N), and Guantao
(115∘8′ E, 36∘31′ N) sites over the plain. The
cropping systems at the Yucheng, Daxing, and Guantao sites are all rotation
of winter wheat–summer maize, while
land cover is dwarf shrub at the Miyun site. The eddy covariance data are
processed with general procedures (Liu and Xu, 2013; Liu et al., 2013). GPP
data are available only at the Yucheng site. In addition to eddy covariance
fluxes, grain yield records of wheat and maize in county statistics are also
used to verify the GPP predictions at the regional scale. The method of crop
yield converted to GPP follows Xin et al. (2015).
Model implementation and numerical experimental design
Simulation setup
The model simulations were conducted at 8 km spatial resolution and a
half-hour time step. The cropland practices are classified into wheat–maize
and wheat–rice rotations. Atmospheric driving forces are interpolated from
daily meteorological variables recorded at the climatic stations to grid
cells with a gradient inverse distance square method (GIDS), which accounts
for the effects of elevation, latitude, and longitude (Nalder and Wein,
1998). Estimated with sunshine duration in a linear relationship, the global
radiation is subdivided into direct visible and near-infrared parts, as well
as direct beam and diffusive components with Weiss and Norman (1985). The
daily air temperature is extended to hourly values with a sinusoidal function
based on the daily maximum and minimum temperatures (Campbell and Norman,
1998). During the winter wheat growing period, irrigation water is supplied
when water storage in the root zone is below 60 % of the field capacity.
Summer maize is set to be irrigated not more than one time in its growth
period. The simulation is conducted with prescribed daily LAI series
retrieved from remotely sensed NDVI series for eco-hydrological prediction
from 1980 to 2013, in which the first year is taken as warming up.
Separation of climate change and management effects
By using a general function, f, the scalar fluxes (water vapour and carbon)
between land surface and the atmosphere are determined by climate factors
(M), fertilization of elevated atmospheric CO2 (C), and agronomic
management and technological advancement (in this study we assume the
long-term trend of leaf area index (LAI) may represent the effects of human
activities on crop and natural ecosystems. Human activities
(also referred to as management practices) in the agro-ecosystem include
renewals of cultivars, irrigation facility improvement, fertilizer
application, soil quality amelioration, etc.), namely,
f=FM,C,LAI,….
The changes in f contributed by a single factor (expressed as fi) and
its interaction with another factor (expressed as fij) can be decomposed
by the Taylor expansion as
fi=∂F∂xiΔxi+12!∂2F∂xi2Δxi2+⋯+1n!∂nF∂xinΔxin,fij=fi+fj+∂2F∂xi∂xjΔxiΔxj+⋯,
where xi and xj (i≠j) represent M, C, and LAI,
respectively. The factor separation methodology from Stein and Alpert (1993)
and Alkama et al. (2010) is used to categorize the contributions of climate
change, CO2 fertilization, and LAI, as well as their interactions to
long-term trends of ET and GPP. Similar to Alkama et al. (2010), the total
effect, f123, is expressed as
f123=f1+f2+f3+f12+f13+f23+f123,
with
f12=f12-f1-f2,f13=f13-f1-f3,f23=f23-f2-f3,
where f1, f2, and f3 are the direct contributions of climate
change, atmospheric CO2 enrichment fertilization, and agronomic
management, respectively; f12 is the contribution of interactions of
climate change and CO2 enrichment; f13 is the contribution of
interactions between climate change and agronomic management; f23 is the
contribution of interactions between CO2 fertilization and agronomic
management; f123 is the contribution of interactions between climate
change, CO2 fertilization, and agronomic management.
Seven numerical experiments designed to fully distinguish the contributions
of climate change, CO2 fertilization, and agronomic management are
conducted by the VIP model over the NCP from 1980 to 2013. The experiments
are as follows:
f123 (“all factors”): current climate, CO2, and LAI
spatiotemporal pattern;
f1 (“climate change effect”): current climate, but the
atmospheric CO2 concentration is fixed at year 1981, and the LAI pattern
is set as the multi-year average;
f2 (“CO2 fertilization effect”): climate and LAI are fixed
at a specific year, but with the current CO2 concentration;
f3 (“management effect”): climate and CO2 concentration
are fixed at a specific year, but the current LAI pattern is used;
f12 (“climate change and CO2 fertilization effects”):
LAI pattern is fixed, but the current climate and CO2 concentration are
used;
f13 (“climate change and management effects”): CO2
concentration is fixed, but the current climate and LAI pattern are used; and
f23 (“CO2 fertilization and management effects”):
climate is fixed at 1981, but current CO2 and LAI are used.
The trends of annual ET and GPP in the above experiments are calculated,
and then, according to Eqs. (4) to (7),
the contributions of climate change and CO2 fertilization and management
practices to ET and GPP long-term trends are then separated.
Comparisons of the
simulated daily ET with eddy covariance measurements.
![]()
Comparison of the simulated
GPP with eddy covariance measurements at the Yucheng site.
![]()
Comparisons of the
predicted GPP and statistic yield derived GPP at county scale in 2000 and
2005.
![]()
Discussion
Our simulations suggested that annual ET and vegetation GPP increased over
the North China Plain from 1981 to 2013. Climate change contributed
positively to ET intensification, but negatively to GPP enhancement.
Agronomic management and technological advancement are the dominant factors
in promoting GPP increase. The use of remote sensing NDVI series has greatly
improved the reliability of vegetation water consumption and productivity
prediction at spatial and temporal scales, even if there were uncertainties
in vegetation characteristics retrieved from the NDVI dataset. The results in
this study were supported and
consistent with the most relevant studies on field and regional scales.
Is the trend of ET upward or downward over the NCP?
Although the crop productivities are steadily increasing, whether the actual
ET over the NCP is increasing or decreasing during the last 3 decades is
controversial from the reported literatures. By using the complementary
relationship models (Brutsaert and Stricker, 1979), actual ET and potential
ET both decreased (Cao et al., 2014; Gao et al., 2012). However, ET
increased, predicted by the process-based VIP model from 1981 to 2013, which
was consistent with the increasing trend of terrestrial greenness (S. Wang et
al., 2016). Yuan and Shen (2013) found that in the northern part of the NCP
(Hebei Province), ET was positively correlated with recorded crop grain
yield, and agricultural water use has increased. Field measurements under
well-watered fields also showed that seasonal ET rates of both winter wheat
and summer maize increased (Zhang et al., 2011). Brutsaert (2006)
acknowledged that decreasing pan evaporation was evidence of increasing
terrestrial evaporation. As is generally known, the sign of ET change should
be the same as that of vegetation greenness. Over the NCP a positive trend of
ET was more believable, in the light of a significantly increasing NDVI over
the growing season, especially in spring. The positive effects of warming
with higher water vapour deficit on ET might be offset by the negative effect
of declining solar radiation and wind speed on potential evaporation. In view
of the complementary relationship hypothesis (Hobbins et al., 2001),
alteration of available energy partitioned into latent heat flux (or ET) is
dominated by the atmospheric water vapour deficit, namely, while more vapour
is evaporated into the atmosphere boundary layer, its water vapour deficit is
then accordingly relaxed, resulting in a lower rate of ETp.
However, while declining global radiation is the dominant factor in the ET
trend, actual ET (ETa), wet surface ET (ETw), and
potential ET (ETp) trace the trend of available energy (net
radiation). In this study case, net radiation declined at a rate of
-5.58 MJ yr-1 (r= 0.56, p < 0.01) over the NCP. As
the trends of both ETp and ETw were dominated by the
radiation trend, the
ETa estimated from the complementary relationship definitely
followed the negative trend of radiation, because the positive trend of
aerodynamic evaporation was weak as a tradeoff of the positive effect of
rising water vapour deficit and the negative effect of decreasing wind speed.
However, the declining ET trend which resulted from the reduced radiation has actually been reversed by
increasing green leaf area, which would reduce land surface albedo and
temperature, etc. Consequently, ET and GPP showed a slight increase. This
study case also confirmed limitations of the complementary relationship for
diagnosing the evaporation trend under the condition of declining radiation.
Effects of climate change and CO<inline-formula><mml:math id="M310" display="inline"><mml:msub><mml:mi/><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:math></inline-formula> on the cropping systems
Without adaptation measures, climate change is shown to exert negative
effects on the productivity of cropping systems in the NCP during the last 3
decades (Mo et al., 2013; Liu et al, 2010). Changes in individual climate
variables affected specific cropping systems differently, which was
associated with crop type and growing season. Climate warming in the winter
and spring seasons was beneficial to vegetative growth of winter wheat (Mo et
al., 2013). Although air warming has shortened the growing length, the
autonomously adopted cultivars with higher thermal requirements usually
maintained the crop growth length and accumulated more photosynthesis
products, which may outweigh the extra respiration consumption under warmer
climates (Wang et al., 2012). So far, the effects of global warming on wheat
production have been inhomogeneous in the North China Plain, which were
positive in the northern part but negative in the southern part (Zhang et
al., 2013). The reasons are that during the wheat growth period air
temperature is still below the favourable conditions in the high-latitude
part of the plain; therefore, recent global warming is benign to the wheat
growth, whereas air warming may be detrimental in the southern part,
especially for rainfed wheat (Xiao and Tao, 2014).
However, dominated by summer monsoon in the North China Plain, the climate is
hot in the summer maize growth period. Due to maize being a tropically
originated species with a high thermal requirement, it can tolerate
relatively high air temperature. Our study showed that maize has not suffered
noticeably from air warming in the recent decades, confirmed also by Xiong et
al. (2012). However, Guo et al. (2014) reported that effect of air warming on
maize was adverse with an Agro-Ecological Zones model, and the decreased
daily temperature range (DTR) may be detrimental to crop yields.
Nevertheless, adaptation measures may boost the production, such as harvest
time delay (Wang et al., 2012) or planting date advancement (Sacks and
Kucharick, 2011).
As atmospheric aerosols from industrial production and combustion have
increased, global radiation has declined in many parts of the world, in which
direct components decreased but diffuse components increased, so-called
“global dimming” (Liepert, 2002; Ren et al., 2013). The decline of global
radiation has resulted in less pan evaporation and carbon assimilation in
crop and natural vegetation communities (Xiong et al., 2012; Xiao and Tao,
2014); nevertheless, plant canopies can use the diffuse radiation with higher
efficiency than direct beams (Gu et al., 2002). In springtime, while the
atmospheric circulation is shifting from continental to ocean monsoon in East
Asia, the wind speed is relative high, and consequently air pollution and
aerosols are usually low; therefore, global radiation is not reduced
obviously in the wheat growth period (Table 1), illustrating that air warming
and precipitation variability other than radiation decrease are the principal
climate factors contributing to the tendency of wheat production. In
contrast, global radiation declines significantly in the summer season in the
North China Plain (Table 1); as a result, productivities of autumn harvest
crops such as maize are mainly affected. For example, Guo et al. (2014)
reported that maize potential productivity was reduced by 20 kg hm-2
due to global radiation decline in the last decades over China.
During the study period, atmospheric CO2 concentration increased from
340 to 396 ppm, which contributed to enhancement of crop productivity.
However, most studies with statistical analysis models neglected the
contribution of CO2 fertilization (e.g. Lobell and Burke, 2010; Song et
al., 2014). As confirmed by FACE experiments, elevated atmospheric CO2
concentration is accelerating plant photosynthesis and reducing
transpiration, whose fertilizer effect is 0.065 % per ppm increment for
C3 plants (Field et al., 1995; Long et al., 2006; Ainsworth et al., 2008). In
this study, the contribution of CO2 to GPP is positive, but negative for
ET. The positive CO2 effect on GPP almost compensates for the negative
effects of climatic variable changes. However, it should be borne in mind, as
Ainsworth et al. (2008) pointed out, that the CO2 fertilizer effect may
be overestimated by the process-based crop/ecosystem models.
Effects of agronomic practice, technique advancement and other
factors
The remotely sensed NDVI is an excellent indicator of long-term changes of
vegetation covers. Here we assumed that climate change did not modify the
tendencies of vegetation covers, but dominated the inter-annual variation of
vegetation density. Renewal of crop cultivars, applications of synthetic
fertilizer and irrigation, as well as conservancy tillage and nitrogen
deposition all contribute to the crop and/or natural productivity
improvements (Yu et al., 2012; Bai et al., 2016; Piao et al., 2015). National
statistical records of grain yields on county scale showed a rapid increase
from 1980 to 1990 and a moderate increase in the 2000s. Enhancement of crop
yields mainly stemmed from more biomass accumulation and a higher harvest
index than previous varieties (Zhang et al., 2013). In our simulations the
upward trend of GPP was more significant than that of ET, which was
consistent with the increasing trend of the cumulative NDVI. Agronomic
practices and technology advancement contributed 103 % of GPP changes in
our study. By using crop models, Yu et al. (2012) and Song et al. (2014)
reported, respectively, the relative contributions of 92 and 62 % by
agronomic management and renewal of cultivars for rice, and Guo et al. (2014)
showed that 99.6 to 141.6 % of maize yield increases was contributed by
technological advancement in China since the 1980s. The previous studies
showed that, if no adoption measures were taken, climate change generally
contributed negatively to crop productivities in the mid-latitude areas, but
the negative effects were usually compensated for by genetic improvements,
applications of fertilizer and irrigation, pest and weed control, as well as
CO2 and nitrogen deposition effects (Liu et al., 2010; Lobell et al.,
2011; Guo et al., 2014; Bai et al., 2015). Under warming climate conditions,
it is expected that water requirement by crops and natural plants will
increase, but the intensified ET may be limited by insufficient soil moisture
availability. Therefore, the sustainability of crop productions greatly
depends on the improvement of agronomic management and technological
advancement on variety breeding.
Summary and conclusions
Climate change and human activities have greatly altered the hydrological
regime and crop productivity in the North China Plain, with warm temperate
climate during the recent 3 decades. The VIP ecological model integrated with
the NOAA-AVHRR NDVI data series predicted that spatial average annual actual
ET weakly increased, while vegetation primary productivity (GPP)
significantly increased (p < 0.01) from 1981 to 2013,
consistent with the remotely sensed NDVI trend. The increases in actual ET
and GPP mainly occurred in spring, while ET and GPP were obviously
decreasing in August owing to global radiation diminishing.
Climate change, elevated atmospheric CO2 fertilization, and agronomic
management (new cultivars, irrigation, and chemical fertilizer) all
contribute to the inter-annual trends of ET and crop GPP. The relative direct
contributions of climatic change, CO2 fertilization, and agronomic
management to ET increase are 56, -28, and 68 %, while the
contributions to GPP tendency are -32, 25,
and 103 %, respectively. Air warming intensifies the crop water
requirement and enhances the production of crops harvested in summer. The
decline of global radiation resulting from exaggerated aerosol concentration
and air pollution is considered to be the main cause of GPP reduction in
August. The study confirms the necessary for imminent control of air
pollution and aerosol to sustain the productivity of the agricultural system.