Physiologia Plantarum 133: 705–. 2008 Copyright ª Physiologia Plantarum 2008, ISSN 0031-9317 REVIEW Impact of climate change on crop nutrient and water use efficiencies Sylvie M. Brouder* and Jeffrey J. Volenec Department of Agronomy, Lilly Hall of Life Sciences, Purdue University, 915 W. State Street, West Lafayette, IN 47907-2054, USA Correspondence *Corresponding author, e-mail: [email protected]Received 15 November 2007; revised 1 May 2008 doi: 10.1111/j.1399-3054.2008.01136.x Implicit in discussions of plant nutrition and climate change is the assumption that we know what to do relative to nutrient management here and now but that these strategies might not apply in a changed climate. We review existing knowledge on interactive influences of atmospheric carbon dioxide concen- tration, temperature and soil moisture on plant growth, development and yield as well as on plant water use efficiency (WUE) and physiological and up- take efficiencies of soil-immobile nutrients. Elevated atmospheric CO 2 will increase leaf and canopy photosynthesis, especially in C3 plants, with minor changes in dark respiration. Additional CO 2 will increase biomass without marked alteration in dry matter partitioning, reduce transpiration of most plants and improve WUE. However, spatiotemporal variation in these attributes will impact agronomic performance and crop water use in a site-specific manner. Nutrient acquisition is closely associated with overall biomass and strongly influenced by root surface area. When climate change alters soil factors to restrict root growth, nutrient stress will occur. Plant size may also change but nutrient concentration will remain relatively unchanged; therefore, nutrient removal will scale with growth. Changes in regional nutrient requirements will be most remarkable where we alter cropping systems to accommodate shifts in ecozones or alter farming systems to capture new uses from existing systems. For regions and systems where we currently do an adequate job managing nutrients, we stand a good chance of continued optimization under a changed climate. If we can and should do better, climate change will not help us. Introduction Climate change variables including precipitation (amount and distribution), temperature and atmospheric CO 2 con- centrations are expected to alter agricultural productivity patterns worldwide. Carbon dioxide is a plant nutrient, and atmospheric enrichment has the potential to enhance plant productivity. Schimel (2006) observed that, at least in some regions, agriculture may be one of the bright spots, ‘the silver lining in the climate change cloud’. But higher global temperatures and altered precipitation patterns are expected to accompany the higher CO 2 levels, and these factors may lessen or negate any pro- duction increases or even depress production below current levels. The myriad of modeling studies attempting to project the short- and long-term impacts of climate change on agriculture are consistent only in highlighting that the nature of the productivity change itself will vary. Realized yield changes will reflect differences in local environments as well as differences in access to seed and management technologies that may offset negative climate change impacts. Regardless, with any potential changes in agricultural productivity comes a potential for associated changes in crop nutrient use. Local potential yield levels are Abbreviations – AE, agronomic efficiency; FACE, free-air concentration enrichment; PE, physiological efficiency; Ps, net photosynthesis; Rd, dark respiration; UE, uptake efficiency; WUE, water use efficiency. Physiol. Plant. 133, 2008 705
20
Embed
Impact of climate change on crop nutrient and water use ... · Impact of climate change on crop nutrient and water use efficiencies Sylvie M. Brouder* and Jeffrey J. Volenec ... knowledge
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Physiologia Plantarum 133: 705–. 2008 Copyright ª Physiologia Plantarum 2008, ISSN 0031-9317
REVIEW
Impact of climate change on crop nutrient and water useefficienciesSylvie M. Brouder* and Jeffrey J. Volenec
Department of Agronomy, Lilly Hall of Life Sciences, Purdue University, 915 W. State Street, West Lafayette, IN 47907-2054, USA
Izaurralde et al. (2003) remark that their results for US
wheat and maize are more favorable than the earlier
projections of Brown and Rosenberg (1999) who foundonly small increases in yield with a temperature increase
of 2.5�C and large decreases in yield with a temperature
increase of 5�C. The relative merits of the different results
can be difficult to discern. Changes in assumptions
pertaining to critical drivers such as the interannual
variability of precipitation (intensity and occurrence) and
temperature (extremes and their duration) can drastically
alter model outcomes (Porter and Semenov 2005). Someregions of the world such as the central United States
appear predisposed to respond more beneficially than
others. Furthermore, the extent to which a given agricul-
tural region is vulnerable to negative impacts of climate
change reflects social and economic variables; projec-
tions vary according to assumptions about levels of
available technology and market forces (Reilly and
Schimmelpfennig 1999). For example, Darwin et al.(1995) predict a 20–30% reduction in global cereal
production without technology and market factors, but
a 0.2–1.2% increase with these factors optimized. In
a similar study, Rosenzweig and Parry (1994) specifically
highlight the effect of differential access to technology in
developing countries where available technologies may
not overcome the negative impacts on global climate
change. Finally, new results from free-air concentration
Table 1. Present-day and future (2095) regional temperatures, precipitation and associated crop production-influencing factors as estimated by the
Hadley Centermodel. Adapted from Izaurralde et al. (2003). aUS Cornbelt states areOhio, Indiana, Illinois, Iowa, andMissouri; U.S. Southern Plains states
are Texas and Oklahoma. bWUE is plant or crop WUE.
and inorganic pools is strongly influenced by moisture
and temperature, and thus, global climate change may
strongly influence solution concentrations of N as well as
S. Some have speculated that soil C pool size will not
change as increased soil respiration and decompositioncaused by soil warming will be moderated by the
increased C supply belowground (Kirschbaum 2000).
Others, however, note that interactive and indirect effects
of water and soil nutrient availability may lead to
unexpected outcomes as uncertainties abound in our
understanding of key feedback processes (Pendall et al.
2004). For example, many expect elevated CO2 to in-
crease belowground C that will, in turn, enrich micro-bial C, but Zak et al. (2000) reviewed the literature
on microbial C and N responses to elevated CO2 and
found reports of increases, decreases and no change. For
N, the review of Pendall et al. (2004) suggests that
increased CO2may not exert a significant direct effect on
N mineralization per se but associated warming can
cause increased N mineralization, leading to increased
solution-phase N. While few, if any, studies have ex-amined impacts of elevated CO2 on solution-phase
concentrations of nutrients such as K whose availability
is not strongly controlled by biological activity, theory
suggests that any impacts will also be indirectly mediated
by temperature and moisture changes. Rates of adsorp-
tion/desorption reactions will accelerate with increased
temperature, and changes in soil moisture may further
modify reactions by altering the ionic strength of the soilsolution. However, uncertainties surrounding the magni-
tude of temperature increases coupled with the spatial
and temporal variation in soil moisture make it challeng-
ing to predict how climate change will impact plant K
availability.
Almost 50 years ago, Barber proposed that nutrient
transport through the soil matrix toward roots occurred
by two simultaneous processes: mass flow and diffusion(Barber 1962). As a plant transpires, solution-phase nutri-
ents are transported in the convective movement of water
in the bulk soil toward root surfaces. Quantitatively, mass
flow contributions to a nutrient’s acquisition are the
product of the volume of water transpired (v0) and the
mean solution-phase concentration (Cl, Table 3). For
nutrients that are highly buffered and maintained at low
solution-phase concentrations, mass flow does notdeliver sufficient quantities to the root surface. Therefore,
in the presence of a growing root, concentrations of these
nutrients in the solution immediately adjacent to the root
surface will be depleted. Movement by diffusion is
a function of an ion’s diffusivity in water, the water
content of the soil, the tortuous nature of the pathway an
ionmust travel to reach the root, the buffer power and the
concentration gradient created by root uptake (Table 3).
Barber models and other, similar single- and multi-root
models integrate equations for mass flow and diffusive
flux with equations that characterize development of the
root system and transport across the root membrane, the
latter often based on Michaelis–Menton kinetics tocharacterize plant uptake as a function of ion concentra-
tion at the root surface (see Silberbush 2002 for a brief
review of models and their features).
Root surface area and diffusive flux
Previously and again today, as we consider the impacts of
climate change, the value of these models is that theyallow us to explore specific aspects of UE, factors and
processes that are complex, concomitant and non-linear
and that are time consuming, expensive and extremely
difficult to assess with direct experimentation. Indeed,
questions of the impacts of temperature and moisture on
nutrient availability are not new, even if the specific
condition of elevated CO2 has yet to be explicitly
addressed. Ching and Barber (1979) examined the effectof increasing root zone temperature from 15 to 30�C on
availability and uptake of K by maize seedlings. Raising
root zone temperature increased nutrient uptake in both
fertilized and unfertilized treatments (Fig. 2A). They also
observed a positive impact on both availability and
uptake factors. At 30�C, root surface area was increased
approximately 70% at high and low K fertility; K diffusive
flux increased 160 and 50% in low- and high-fertilitytreatments, respectively. Mackay and Barber (1984)
observed a similar effect on maize P accumulation with
a more moderate temperature comparison (19 vs 25�C;Fig. 2B). Again, marked increases in root surface area at
both high- and low-fertility levels accompanied one- to
two-fold increases in nutrient uptake. Temperature
impacts on diffusive flux were again apparent, although
much smaller in magnitude than the changes in rootsurface area. While changes in temperature with global
climate change are expected to be substantially smaller
than the experimental treatments used in these studies,
there is no reason to expect responses to be different other
than in magnitude. Mackay and Barber (1985) also
examined the effect of drought on P uptake and avail-
ability and found reduced nutrient uptake, root surface
area and ion diffusivity with moisture stress for bothhigh and low fertility (Fig. 2C). In this study, the mois-
ture treatments are directly meaningful in the context
of climate change scenarios.
For many, the observation that increasing soil moisture
and temperature from suboptimal to optimal conditions
increases nutrient diffusion and root growth will
seem obvious. Following the Stokes–Einstein equation,
714 Physiol. Plant. 133, 2008
diffusion of an ion in water is a direct function of both
temperature and viscosity; viscosity itself is a function oftemperature (Barber 1995). At 15�C, the rate of diffusion
is only 78% of the rate at 25�C (Weast 1982). Ion dif-
fusivity rates in soil are a direct function of ion diffusiv-
ity inwater and soilmoisture content. At low soilmoisture
content, the diffusion pathway becomes longer as ions
must travel around expanded air pockets. Likewise, cell
expansion requires adequate water, and species-specific
temperature optimums for root growth have beenextensively documented (for a review, see McMichael
and Burke 2002). However, moving beyond the obvious
effects of temperature and moisture on availability and
acquisition, the more difficult and relevant question
concerns the extent to which a specific factor or suite of
factors contributes to observed reductions in nutrient
uptake. In their study on soil moisture and P, Mackay and
Barber (1984) reported a strong, linear relationship(r2 ¼ 0.96) between root surface area and P uptake
across three soils and three moisture levels. They did not
report the relationship between diffusive flux and P
uptake but plotting their tabular data finds amuchweaker
relationship (r2 ¼ 0.36, P ¼ 0.053; data not shown),
suggesting that root surface area reductions may be more
directly important for P uptake. Separate sensitivity
analysis for model performance in predicting both P andK uptake supports this conclusion (Silberbush and Barber
1983a, 1983b). For both nutrients, varying one model
parameter while holding all others constant identified
root growth rate as the single most influential factor
governing nutrient uptake. Increasing diffusivity did not
greatly increase uptake but, within the parameter ranges
explored, proportional decreases in diffusivity reduced
uptake as much as corresponding changes in root surface
area.Certainly, such sensitivity analyses have their short-
comings. In its most simple form, the approach overlooks
parameter interdependence, and not all parameters
are equally amenable to change. But a more thoughtful
tinkering with parameters coupled with targeted exper-
imentation over widely varied plant–soil systems can
produce solid working hypotheses. As reviewed by
Brouder (1999), investigations of K accumulation byflooded rice (Teo et al. 1992), slash pine seedlings grown
alone and in combination with other species (Van Rees
et al. 1990) and cotton grown in a range of soil conditions
(Brouder and Cassman 1994a) also identified root geo-
metry (length and diameter) as highly sensitive and a
potentially dominating parameter controlling K accumu-
lation. Direct evaluations of genotypic differences in root
geometry and K acquisition efficiency of soybean (Silber-bush and Barber 1984), corn (Schenk and Barber 1980)
and cotton (Brouder and Cassman 1990, 1994b) serve to
further substantiate the relative importance of root growth
when compared with other nutrient availability and
acquisition factors for uptake of relatively immobile
nutrients.
These observations on the importance of root explora-
tion of the soil by enhanced root surface areamay seem tobode well for a changed climate where CO2 fertilization
could increase C available for building additional fine
roots. If root:shoot ratios remain constant but the overall
plant is bigger (as discussed above), there may be more
potential for an enlarged root system to capture the
relatively immobile nutrients. The environment of the
root system is extremely heterogeneous in time and
Fig. 2. The effect of temperature or moisture on nutrient uptake of maize and on selected soil availability and root acquisition parameters. Data are
shown as percentage change from the baseline condition. Data are adapted from experiments where (A) root zone temperatures were increased from15
to 29�C in an unfertilized soil and a soil receiving 500 mg K g21 soil (Ching and Barber 1979), (B) root zone temperatureswere increased from18 to 25�Cin a low- and high-P fertility soil (Mackay and Barber 1984) and (C) root zone soil moisture was reduced in soil water potential from233 to2170 kPa in
a low- and high-P fertility soil (Mackay and Barber 1985).
Physiol. Plant. 133, 2008 715
space; the adaptation of extreme phenotypic plasticity to
exploit such an environment is a key attribute of success
(Fitter 2002). Crop plants are expected to be particularly
plastic in response to patchy nutrient availability as such
plants were initially not only adapted to but also
improved in their ability to be strongly responsive toenhanced nutrient supply. As documented in classic
experiments by Drew, fine roots proliferate in zones
enriched with nutrients, particularly NH14 , NO2
3 and P
(Drew 1975, Drew and Saker 1975, 1978, Drew et al.
1973). This phenomenon has been repeatedly shown
both in controlled environment and in the field for many
major crop species [e.g. sorghum-sudangrass (Pothuluri
et al. 1986), winter wheat (Newman and Andrews 1973),corn (Zhang and Barber 1993) and cotton (Brouder and
Cassman 1994b)]. As summarized in several reports
(Lynch and St Clair 2004, Pendall et al. 2004), a few
studies have suggested that root architecturemay respond
to elevated CO2. For example, Pritchard and Rogers
(2000) proposed that elevated CO2 would stimulate
lateral branching, particularly in surface horizons. But
such responses and/or their benefits may be conditionalupon other climate change variables and the quantity and
distribution of nutrients. Some studies have suggested that
elevated CO2 may help negate the impact of increased
temperatures that exceed the optimum for root growth
(Bassow et al. 1994, Wan et al. 2004). In studies of
amodel grassland,Maestre and Reynolds (2006) reported
that belowground biomass increased in response to high
CO2 but only if high levels of nutrients were provided;root proliferation into nutrient patches increased with
increasing nutrient availability but was not influenced by
ambient CO2. As can readily be seen with modeling
studies, root proliferation is of no benefit if roots are
competing with each other. Thus, we may be headed
toward a not too surprising conclusion that growing
a bigger plant with CO2 fertilization may require en-
hanced nutrient inputs.
Water influx and mass flow
In general, model simulations for immobile nutrients
have not been very sensitive to changes in the rate of
water influx into the root (Silberbush and Barber 1983a,
1983b), a point relevant to discussions of the positive
influence of CO2 fertilization on plant WUE. Underconditions of reduced transpiration, some have theorized
that acquisition of nutrients that travel frombulk soil to the
root surface primarily by mass flow will be negatively
affected, resulting in nutrient deficiency (Lynch and St
Clair 2004). Nutrients long considered to be delivered
primarily bymass flow include soil-mobile nutrients such
as nitrate and sulfate and soil-exchangeable nutrients
such asMg andCa that are abundant in the solution phase
but required in relatively small quantities by the plant
(Barber 1995). However, reducing mass flow to a point
where it restricts nutrient delivery but does not cause
a more direct effect of water stress (e.g. reduced root
growth) seems unlikely. Diffusion and mass flow are notmutually exclusive deliverymechanisms; the process that
dominates is not an attribute of the nutrient itself but
a reflection of root zone conditions. When the product of
water uptake per unit root surface area (v0) and ion
concentration in the soil solution (Cl) are equivalent to the
plants needs (Imax, Table 3), mass flow will clearly be
the dominant mode of solute transport to the root (in the
context of the Cushman–Barber model, v0Cl ¼ Imax), butwhen v0Cl < Imax, diffusion contributes to nutrient trans-
port. The Ching and Barber (1979) study discussed above
can be used to illustrate this point. Adding 500 mg K g21
soil increased v0Cl from 3.2 � 1028 to 4.8 � 1026 mmol
cm22 s21, while Imax remained constant at 5.6 � 1027
mmol cm22 s21, switching the dominant transport pro-
cess from diffusion to mass flow (at 15�C, calculatedfrom Ching and Barber 1979). Rerunning the model(Version 1.1, Oates and Barber 1987) and reducing v0to 1 � 1027 cm s21, a >85% reduction does not effect
simulated K uptake for either fertility treatment (<1.5%
uptake reduction; model output not shown). Indeed, Van
Vuuren et al. (1997) showed this phenomenon with
wheat grown in elevated ambient CO2 under conditions
of ample and restricted soil moisture. Transpiration was
repressed at 700 mmol CO2 mol21 but plant P acquisi-tion was not impacted by dry soil conditions. Thus, in
terms of transport, any nutrient stress resulting from
reduced transpiration would likely reflect the failure of
the secondary process of diffusion to deliver adequate
nutrients to the root surface.
Uptake kinetics
Model simulations of uptake of Pand K have also not been
particularly sensitive to changes in kinetic aspects of
acquisition, but this approach may not be sufficiently
rigorous to address the importance of variation in kinetics
to UE. Recent advances in molecular genetics permit
a more critical evaluation of questions focused on the
limitations imposed by kinetic parameters of nutrient
uptake than was previously possible. Genes and compli-mentaryDNAs for dozensof high-affinity nutrient-specific
ion carriers have been cloned and characterized. Expres-
sion of these genes can be driven to high levels by root-
specific and constitutive promoters. Working in model
systems, Misson et al. (2005) identified genes related to
high-affinity P transport across membranes, genes that
Raghothama (2000) had proposed would be critical to
716 Physiol. Plant. 133, 2008
root acquisition of P from low P soils. BassiriRad (2000)
proposed that altering nutrient uptake to meet plant needs
in a changing environment would be best accomplished
by focusing on high-affinity nutrient transporters and their
kinetic parameters. In theory, elevated CO2 should permit
upregulation of transporters as there would be a higheravailability of carbohydrates to meet transporter energy
the effectiveness of molecular engineering the kinetic
aspects of nutrient uptake to negate the consequences of
global change has not been critically evaluated.
The advent of molecular techniques has made it
possible to examine the importance of gene expression
for regulation of nutrient uptake across the cell mem-brane. We explored the impact of expression of high-
affinity P transporters on tobacco (Nicotiana tabacum L.)
growth and P uptake (A.S. Berg, 2004, Thesis, Purdue
University, West Lafayette, IN, USA). Expression of high-
affinity P transporter genes from yeast and Arabidopsis,
driven by a constitutive promoter and measured as
transcript abundance, was very high in both root and
shoot tissues. Two control groups were included: trans-genic plants containing the transformation vectorwithout
a P transporter insert and a commercial tobacco cultivar,
W-38. Plants were grown for 7 weeks in soils that had
either low or high soil test P concentrations, and dry
weights and plant P content were measured 4, 5, 6 and
7 weeks after transplanting. As expected, plant growth
andP uptakeweremuchgreater in the high-P soil than the
low-P soil (Fig. 3). Growth and P uptake of the transgenicplants containing the high-affinity P transporters were
similar to the transgenic control plants without the P
transporter insert in both soils. At week 4 in high-P soil,
growth of the commercial cultivar W-38 was less than
both plants transformed with P transporters, and P uptake
of W-38 was reduced when compared with the trans-
formed control plants. However, by week 7, P uptake of
W-38 in the high-P soil was greater than that of the otherplants. There was no influence of overexpression of either
yeast or Arabidopsis P transporter gene on P uptake and
plant growth in the low-P soil.
To date, our study is one of only a very limited number
of studies where transgenic approaches to improve
nutrient uptake have been examined in soils. Recently,
Park et al. (2007) have reported that constitutive ex-
pression of a high-affinity P transporter from tobaccocould increase tissue P concentrations of rice. Although
these authors observed higher instantaneous uptake rates
of 32P in transgenic plants compared with untransformed
control plants, total P uptake was not reported because
tissue mass data were not assessed. Therefore, the re-
ported growth reductions (qualitative results only) were
possibly confounded with observations of higher tissue P
concentrations. Surprisingly, a comprehensive survey of
the literature revealed no published reports focusing on
upregulation of K transporters and its impact on K uptake
from soil. Numerous studies have reported induction of
K transporters in roots exposed to low media (not soil) K
concentrations (Ashley et al. 2006 and references cited
therein) and imply that these changes are essential to
maintain rapid K uptake as solution K concentrationsdecline. However, Garciadeblas et al. (2007) recently
suggested that K transporters may have broader functions
in plants including high-affinity K uptake, K efflux into the
media to reduce tissue K concentrations and as a link
between K nutrition and root morphogenesis. This sug-
gests that the roles of K transporters may go beyond
merely facilitating K uptake across the plasmamembrane
at low K concentrations. Clearly, even without climatechange as an additional variable, we posses only a rudi-
mentary understanding of the role of transporters in
nutrient uptake from soil.
Fig. 3. Trends in total biomass and plant P uptake of tobacco lines grown
in high- and low-P soils during weeks 4–7 posttransplanting. Two lines
contained constitutively expressed high-affinity P transporter genes from
yeast and Arabidopsis, one control line was transformed with the vector
alonewithout a P transporter gene and the fourth linewas the commercial
cultivar W-38. Asterisks indicate dates where significant differences
between W-38 and the other lines occurred (see text for details).
Physiol. Plant. 133, 2008 717
Under P-limited conditions, upregulation of P trans-
porters is just one of several known physiological mech-
anisms plants can use to enhance P uptake. A key
additional physiological mechanism is the secretion of
organic acids (Sanchez-Calderon et al. 2006), also an
important factor for mobilizing other, relatively insol-uble nutrients including Fe (Lynch and St Clair 2004).
Theoretically, enhanced allocation of C belowground as
a result of global climate change could alter the quantity
and quality of exudates that may benefit nutrient uptake
in soils where acidity or alkalinity limit nutrient solubility.
As reviewed by Lynch and St Clair (2004), only a few
studies have critically examined this hypothesis, and
results to date have been mixed; Norby et al. (1987),Hodge (1996) and Uselman et al. (2000) found no effect
of elevated CO2 on exudates, while Hodge and col-
leagues observed reduction in volume and changes in
composition of exudates (Hodge and Millard 1998,
Hodge et al. 1998). As repeatedly remarked in the liter-
ature, the area of root uptake responses to global climate
change are understudied and requiremuchmore intensive
study (BassiriRad 2000, Lynch and St Clair 2004, Pendallet al. 2004).
Conclusions: managing plant nutrition ina changed climate
What are the practical implications of the above
discussions? First and foremost is the concept that crop
plants may be bigger, smaller or similar in size whencompared with today’s specimens, but their nutrient
content and PE will be scaled according to size. To date,
experimentation on crop plants has not found conclusive
evidence that PE is altered in high CO2 environments
(Long et al. 2006). The observation of Schimel (2006) that
‘Some set of biological processes appears to operate to
reduce the impact of CO2 on realized gains in biomass
and yield below that expected from the effects ofphotosynthesis.’ can be viewed as a simple restatement
of the Law of the Minimum within the context of global
climate change. Clearly, nutrient stress has the potential
to reduce growth stimulation by elevated CO2 (Campbell
and Sage 2006, Lynch and St Clair 2004). Modest, crop-
specific benefits in agricultural yieldsmay be realized but
only where nutrient availability can be optimized and
where climate change increases temperatures to a spe-cies-specific optimum and changes precipitation patterns
to reduce water stress (drought or flooding) days. C3
species may also accrue a direct benefit from CO2
fertilization. Nutrient recommendations for a changed
climate will operate on the same premise as current
recommendations – an understanding of the PE that is
specific to the crop and of the UE that is specific to the
unique combination of crop and soil. Simple, empirical
models will continue to be used to translate this in-
formation from theory into practice. We anticipate that
major portions of today’s soil fertility/plant nutrition
recommendations will remain viable irrespective of
climate change.
Implications for nutrient management
Nutrient replacement is a core tenet of many existing
recommendations for sustainable management of rela-
tively immobile nutrients. If plants produced under
elevated CO2 are simply bigger, but otherwise the same
in their gross nutrient content per unit biomass, thenpresent-day nutrient balance calculations for fertilizer
recommendationswill remain applicable. In crop species
that have been extensively improved for agriculture,
nutrient concentrations, especially in grain, can be
relatively constant when yields are not limited by other
factors. Dobermann et al. (1996a) examined irrigated rice
yields and grain composition in the Philippines, Indo-
nesia, Vietnam, China and India and determined that theK concentrations of modern rice varieties were fairly
constant across environments. Fifty percent of all samples
analyzed ranged from 2.5 to 3.3 mg kg21. In a 6-year
study conducted at five locations on widely varying soils,
we have also documented relatively constant nutrient
concentrations in high-yieldingmaize and soybean grain.
Across site-years, P, K and Mg concentrations in maize
grain (yields >10 000 kg ha21) averaged 3.3, 3.9 and1.3 mg kg21, respectively; P, K andMg concentrations in
soybean grain (yields>3500 kg ha21) averaged 5.5, 18.3
and 2.4 mg kg21, respectively (Table 4). Standard devia-
tions in nutrient concentrations were relatively similar
between species, although coefficients of variation
tended to be lower in soybean, reflecting its higher
concentration values. When these average removal
values are used to estimate actual crop removal over thefull range of yielding environments, the relationship
between predicted and measured values is very strong.
For example, the predicted:measured relationship for
yearly K removal is close to unity for both crops (Table 4,
Fig. 4A). The predicted:measured relationship for 6-year
cumulative removal of a maize–soybean rotation has
a 1:1 relationship across all locations (Fig. 4B).
Therefore, provided we continue to pursue a nutrientreplacement philosophy, changes in regional input re-
quirements will be most remarkable where we alter the
cropping system to accommodate shifts in crop ecozones
or alter the farming system to capture new uses from
existing systems (e.g. use of whole-plant maize for bio-
fuels). Climate change may disproportionately increase
the risks of growing one crop species when compared
718 Physiol. Plant. 133, 2008
with an acceptable alternative. Southworth et al. (2000)suggest that variation by 2050 may increase risks
associated with growing maize in southern regions of
the Cornbelt, and growers may elect to modulate risk by
growing a different crop that is better suited to the
emerging ecozone. These authors suggest that growers
may benefit fromplanting indeterminant crop species like
soybean in place of maize to deal with the greater
potential risks associatedwith increased climate variationand, at the same time, derive benefit from increased CO2
that occurs when growing a C3 crop species.
Changes in demand for agricultural products may also
cause dramatic changes in regional requirements for
nutrient inputs. Shortages of fossil fuels and an aggressive
bioenergy agenda shifted large areas of the US Cornbelt
from a maize–soybean rotation to continuous maizeproduction in 2007. Despite lower grain concentrations
of all nutrients (Table 4), maize’s higher yields and lack of
N2 fixation will significantly increase input requirements
for P and N, although K input requirements will be
reduced. In grain crops where cellulosic biomass may
eventually also be harvested, nutrients removed in
residue will need to be replaced and this could require
significant new inputs. For the 74 million ha of irrigatedrice in Asia, Dobermann et al. (1996a) estimate that
harvesting straw for fuel will increase crop K removal
at least five-fold from 0.9–1.2 million t year21 to 5–9
million t year21. Furthermore, residue removal itself
may reduce soil nutrient supply as residue return both
protects against soil erosion loss and replenishes soil
Table 4. Nutrient concentrations in high-yielding maize and soybean grown in Indiana, USA. For all observations, maize (n ¼ 358) and soybean
(n ¼ 474) yields exceeded 10 and 3.5 Mg ha21, respectively. Regression relationship is for all observations in a 6-year, five location (60 plots location21)
study ofmaize–soybean rotations. Predicted values are the product of yield andmean nutrient concentration; observed values are the product of yield and
the measured concentration in subsamples from each plot-year. NS, P > 0.05.
Nutrient
Grain nutrient concentration Nutrient removal regression: observed vs predicted
Fig. 4. Relationship between measured K removal by maize and soybean crops and predicted K removal based on crop yields and an assumed constant
unit removal value (3.9 and 18.3 mg K kg21 grain dry weight for maize and soybean, respectively; Table 4). Data are from a 6-year, five location, 60 plots
location21 study conducted in Indiana, United States. Data shown are for (A) annual crop removal in each experimental plot and (b) 6-year cumulative
removal in each experimental plot by the maize–soybean rotation. Different symbols are used to identify crop (A) or experimental location (B).
Physiol. Plant. 133, 2008 719
organic C. As discussed above, soil organic C is an im-
portant source of nutrients such as N and helps retain
availability of nutrients such as Fe that can form organic
chelates. Limited research has shown that maize stover
removal can lower grain and stover yields of subsequent
crops and also soil C pools (Wilhelm et al. 1986). Whilethe dynamics of governing biomass conversion to soil
organic C is not well understood and is a subject of
intensive ongoing research (Wilhelm et al. 2007), residue
removal drives changes in soil energy balance. Bare soils
can be >5�C warmer with much higher surface evapo-
transpiration than residue covered soils (reviewed by
Wilhelm et al. 2004), resulting in altered rates of min-
eralization and nutrient diffusion.For regions and systems where we currently do an
adequate jobmanaging nutrients, we stand a good chance
of continuing to optimize nutrient use under a changed
climate. If we can and should do better, climate change
will not help us. To this end, the irrigated rice study of
Dobermann et al. (1996a, 1996b) not only serves
a cautionary warning but also highlights a key aspect of
nutrient management in need of improvement. Theyconclude that current recommendations for K fertilizer
additions in most intensive irrigated rice domains do not
replace the K removed in present-day yields; they remark
that with either increased yields from technology or straw
removal without any increase in yield, K removal will far
exceed the present fertilizer levels and deplete soil K
reserves, ultimately degrading the soil resource. Driving
this imbalance is a lack of appreciation or perhapsknowledge of the K-supplying power of the soil, that
specific combination of the crop and soil that governs UE
(Dobermann et al. 1996b). Review of existing recom-
mendations for the US Cornbelt suggests that this problem
is not unique to Asian rice production. Despite extensive
scientific study and available tools (e.g. high-resolution
soil surveys and spatially and temporally intensive soil
testing results), current recommendations are not welltailored to knownsoil- and crop-specific differences inUE.
Long-term studies in Indiana suggest that additional rates
of 7.5–20 kg ha21 are required to increase available K in
actively farmed soils by 1 mg kg21 for a range of major
agronomic soils (Li and Barber 1988 and ongoing studies)
but recommendations call for only 4.5–7.5 kg ha21
(Vitosh et al. 1995) to achieve this change. The reason
for this clear disconnection between the recommenda-tions for K management and the observations of local soil
responses has been difficult to discern. Thus, while the
empirical model that addresses nutrient replacement is
good, the empiricalmodel for soilmanagement appears to
require significant improvements in at least a few major
agronomic regions if we are to achieve optimum AE in
both present and future production.
Implications for crop improvement
Finally, in our discussions of plant growth and nutrient
needs in a changed climate, we should not overlook the
combined forces of crop improvement and genetic
variation/natural selection. To date, most experimenta-
tion on the effects of elevated CO2 on plant production,
including the elaborate FACE studies, has been con-
ducted by imposing elevatedCO2 levels on plantmaterial
adapted to current atmospheric composition. Genotypic
variation in traits influencing phenotypic expression and
plasticity in important plant attributes such as root
architecture and exudation will allow continued drift
toward form and function adapted to changed conditions.
The Cook et al. (1998) study of evolution of N. strictus
ecotypes under 790 mmol CO2 mol21 is a persistent re-
minder that we should be cautious in drawing conclu-
sions when skipping a 100 years of selection pressure.
Crop improvement efforts will only hasten the process as
suggested by a recent analysis of shifting agroecozones in
the United States. In a study initially designed to examine
the effect that climate change has had to date on cropping
patterns, Reilly et al. (2003) analyzed the geographic
centers of production for maize, soybean and wheat over
the last 100 years. They found a significant north and
westward shift in centroids for both maize and soybean
production, and this shift was accompanied by a 4�Cdecrease in temperature despite an estimated US warm-
ing trend of 0.6�C. This shift reflects management and
genetic technologies including development of new
varieties of soybean that are adapted to longer photo-
periods and earlier maturing maize hybrids that
decreased risk because of early frost. The authors remark
that in the last 100 years, we have seen adaptation to the
magnitude of temperature change that we expect for the
coming century, albeit in the opposite direction.
As noted in the beginning of the paper, pursuit of
adaptive technologies will certainly mitigate negative
impacts and enhance advantages for future plant growth.
While the promise of enhanced nutrient uptake through
transgenic manipulation of transports has yet to be
realized and more research is needed, morphological
traits may be as or more promising crop improvement
targets. As summarized by Lynch (2007), these include