TY - JOUR
AU1 - Ramirez-Villegas, Julian
AU2 - Watson, James
AU3 - Challinor, Andrew J.
AB - Abstract Genotypic adaptation involves the incorporation of novel traits in crop varieties so as to enhance food productivity and stability and is expected to be one of the most important adaptation strategies to future climate change. Simulation modelling can provide the basis for evaluating the biophysical potential of crop traits for genotypic adaptation. This review focuses on the use of models for assessing the potential benefits of genotypic adaptation as a response strategy to projected climate change impacts. Some key crop responses to the environment, as well as the role of models and model ensembles for assessing impacts and adaptation, are first reviewed. Next, the review describes crop-climate models can help focus the development of future-adapted crop germplasm in breeding programmes. While recently published modelling studies have demonstrated the potential of genotypic adaptation strategies and ideotype design, it is argued that, for model-based studies of genotypic adaptation to be used in crop breeding, it is critical that modelled traits are better grounded in genetic and physiological knowledge. To this aim, two main goals need to be pursued in future studies: (i) a better understanding of plant processes that limit productivity under future climate change; and (ii) a coupling between genetic and crop growth models—perhaps at the expense of the number of traits analysed. Importantly, the latter may imply additional complexity (and likely uncertainty) in crop modelling studies. Hence, appropriately constraining processes and parameters in models and a shift from simply quantifying uncertainty to actually quantifying robustness towards modelling choices are two key aspects that need to be included into future crop model-based analyses of genotypic adaptation. Climate change, crop models, genotypic adaptation, ideotypes, impacts. Introduction Agriculture is one of the most vulnerable sectors to changes in climates, due to its reliance on adequate environmental conditions for achieving high productivity (Huntingford et al., 2005). Crops are affected by shortages or excesses of water or excessively high or low temperatures during key periods of their growing cycle (Porter and Semenov, 2005). The effects of adverse environmental conditions have largely been studied and reported by several authors, using combinations of models and data (Allen et al., 2005; Boote et al., 2005). This understanding, in addition to well-constrained and skilful simulation models, can provide insights on what could happen under future climate scenarios of higher temperatures, changing precipitation patterns, and increased likelihood of extremes. Although the figures are varied, recent literature indicates that negative impacts are expected to affect the basic food basket (i.e. wheat, rice, maize, and grain legumes), as well as major cash crops (i.e. sugarcane, coffee, cocoa) at moderate or low (≤+3 ºC) levels of warming if no adaptation actions are taken (Lobell et al., 2008; Challinor et al., 2014b; Porter et al., 2014). Evidence from regional and local studies as well as global meta-analyses of modelling studies indicates that adaptation strategies are critical in countering any negative and/or capitalizing positive effects that may arise as a result of climate change (Claessens et al., 2012; Challinor et al., 2014b). Adaptation strategies are probably the only means by which food availability and stability can be maintained and/or increased so as to meet future food security needs. In fact, recent model-based global estimates indicate that even incremental adaptation strategies could result in mean yield increases of ~7% at any level of warming (Challinor et al., 2014b; Porter et al., 2014). This suggests that substantial opportunities may exist if deeper (i.e. systemic and transformational) changes in cropping systems are implemented. This review focuses on one such strategy, namely, genotypic adaptation. Genotypic adaptation involves the incorporation of novel traits in crop varieties so as to enhance food productivity and stability and, more broadly, also the design of crop ideotypes (i.e. crop plants with ideal traits) for future climates (Donald, 1968; Semenov and Stratonovitch, 2013). Specifically, the use of models for the development of genotypic adaptation options is reviewed. Some important crop responses to key environmental factors are examined first. Secondly, two aspects of climate impacts research are examined: (i) the different approaches to climate change adaptation, and (ii) the importance of models for developing adaptation options. Existing models are then described and recommendations provided so as to capitalize on the potential of using crop model ensembles for understanding crop responses and adaptation options under future climate scenarios. Finally, how crop-climate models can help focus the development of future-adapted crop germplasm in breeding programmes is described. In doing so, past experiences and recent trends in the crop modelling literature are reviewed. To conclude, a framework is proposed that mainstreams crop model-based analyses into future breeding strategies. Key plant processes and crop responses to varying environmental factors In large areas, climate signals are discernible for many crops and regions even when aggregated growing season information is used (Fig. 1). Signals in such areas reflect crop plant responses to variations in weather and climate at a local scale. Some of these responses are discussed in detail below. Fig. 1. View largeDownload slide Percentage variance in historical crop yields explained by seasonal mean temperature and seasonal total precipitation across (A) crops and (B) regions. Variance explained is measured using the coefficient of determination (R2) as derived from the statistical models in Lobell et al. (2008). Both panels show the same data, but pooled differently. Variation for each crop in (A) reflects differences between regions and variation for each region in (B) reflects differences between crops. Thick lines are the medians, boxes represent the interquartile range, whiskers extend to 5–95% of the data, and the dots are outliers. Fig. 1. View largeDownload slide Percentage variance in historical crop yields explained by seasonal mean temperature and seasonal total precipitation across (A) crops and (B) regions. Variance explained is measured using the coefficient of determination (R2) as derived from the statistical models in Lobell et al. (2008). Both panels show the same data, but pooled differently. Variation for each crop in (A) reflects differences between regions and variation for each region in (B) reflects differences between crops. Thick lines are the medians, boxes represent the interquartile range, whiskers extend to 5–95% of the data, and the dots are outliers. A balance exists in the plant–soil–atmosphere interaction so as to allow enough carbon uptake for plant growth, prevent desiccation due to excess transpiration, and maintain canopy and leaf temperatures at near-optimum levels (Huntingford et al., 2005; Lobell et al., 2013). Stomatal conductance, a key factor regulating plant growth, is highly correlated with net photosynthesis (Wong et al., 1979) and is affected by air moisture deficit (i.e. vapour pressure deficit; VPD), radiation intercepted, leaf temperatures, ambient CO2 concentrations, and soil moisture. However, both temperature and air and soil moisture conditions operate against plant growth and also against each other in ways that are often difficult to understand. Mean air temperatures drive canopy and leaf temperatures, which are determinant for photosynthesis. Photosynthetic efficiency varies with temperature in all crop species because it affects RuBisCO (ribulose 1,5 biphosphate carboxylase oxygenase) activity and, in turn, intercellular CO2 concentration and stomatal conductance (Hew et al., 1969; El-Sharkawy, 2014). The response of photosynthesis to temperature varies by species (Fig. 2A). Mean temperatures also drive crop development rates and thus define crop duration (Fig. 2B) which, in turn, affects total photosynthetically active radiation (PAR) intercepted—linearly related to total biomass production. Daily extremes of temperature reduce crop yield mostly through damage to the plant reproductive organs (Fig. 2C) (Peng et al., 2004) and hastened senescence (Asseng et al., 2011). However, complex responses and interactions occur throughout the cropping cycle. For an example: under optimal temperatures and water availability, photosynthesis and transpiration from leaves occur at normal rates; however, under high temperatures plants open their stomata to avoid heat stress, which increases within-leaf CO2 concentrations and thus biomass accumulation (exception being made under high VPD conditions, i.e. dry air as, in such a case, stomata would remain closed to avoid excessive transpiration). If the available soil water is limited, this induces desiccation and stomata are then closed. Drought causes desiccation and stomatal closure, but at the same time water is a direct input of photosynthesis and so the effects on carbon fixation are more direct than those of temperature. In addition, stomatal closure causes within-leaf CO2 concentrations to decrease, thus decreasing inputs to photosynthesis, in some cases also increasing photorespiration (Kobza and Edwards, 1987). This causes lower biomass production and limits growth (Hew et al., 1969; Huntingford et al., 2005). Low light incidence (i.e. solar radiation) also reduces photosynthesis, whereas winds increase transpiration. Drought stress may be induced by increased osmotic pressure in saline soils. Many limiting conditions can occur simultaneously in a given site (see for example Trnka et al., 2014), thus making any prediction of their effect a challenging task. Fig. 2. View largeDownload slide Responses of (A) net photosynthesis to leaf temperature, (B) development rates to mean daily air temperature, and (C) crop yield to temperature during the reproductive period. Data in (A) have been derived from the study of Nagai and Makino (2009) for wheat and rice, and from Bird et al. (1977), Schmitt and Edwards (1981), Crafts-Brandner and Salvucci (2002), and Labate et al. (1990). Solely for illustrative purposes, maize data were fitted to a spline curve with 5 degrees of freedom. Rice and wheat data were fitted to 3rd order polynomials as in Nagai and Makino (2009). Curves in (B) were plotted following Parent and Tardieu (2012). Development rates at each temperature in their models have all been normalized by development rates at 20 °C. Data from (C) were derived from Peng et al. (2004) for rice (hence the x-axis for rice is the minimum growing-season temperature), from Gibson and Paulsen (1999) for wheat (hence the x-axis is the mean temperature during grain-filling), and from Wilhelm et al. (1999) for maize (hence the x-axis is the mean temperature post-anthesis). For (C), all data were linearly scaled so that the maximum yield corresponded to a value of 1. Fits in (C) all follow a linear regression except for rice where the original 2nd degree polynomial of Peng et al. (2004) was used. Fig. 2. View largeDownload slide Responses of (A) net photosynthesis to leaf temperature, (B) development rates to mean daily air temperature, and (C) crop yield to temperature during the reproductive period. Data in (A) have been derived from the study of Nagai and Makino (2009) for wheat and rice, and from Bird et al. (1977), Schmitt and Edwards (1981), Crafts-Brandner and Salvucci (2002), and Labate et al. (1990). Solely for illustrative purposes, maize data were fitted to a spline curve with 5 degrees of freedom. Rice and wheat data were fitted to 3rd order polynomials as in Nagai and Makino (2009). Curves in (B) were plotted following Parent and Tardieu (2012). Development rates at each temperature in their models have all been normalized by development rates at 20 °C. Data from (C) were derived from Peng et al. (2004) for rice (hence the x-axis for rice is the minimum growing-season temperature), from Gibson and Paulsen (1999) for wheat (hence the x-axis is the mean temperature during grain-filling), and from Wilhelm et al. (1999) for maize (hence the x-axis is the mean temperature post-anthesis). For (C), all data were linearly scaled so that the maximum yield corresponded to a value of 1. Fits in (C) all follow a linear regression except for rice where the original 2nd degree polynomial of Peng et al. (2004) was used. The effects of increased CO2 are beneficial for almost any food crop, with increased CO2 concentrations thought to increase dry matter and thus yield (Leakey et al., 2009). However, there is contrasting experimental evidence on crop responses to enhanced CO2 concentrations across varying degrees of soil water and air moisture availability (Long et al., 2006; Tubiello et al., 2007; Ainsworth et al., 2008), despite advances in theoretical understanding (Ghannoum et al., 2000; Leakey, 2009). Underlining experimental evidence on crop responses to elevated CO2 concentrations is therefore needed, since most models incorporate effects in a fairly basic fashion—mainly through empirical factors to enhance assimilation. Particular attention must be placed on understanding the interactions between enhanced [CO2] and other environmental controls (particularly drought and high temperatures), as these remain only partially understood (White et al., 2011; Asseng et al., 2013). A large number of other factors exert control on plant growth and, particularly, on photosynthesis, biomass accumulation, and yield. Leaf nitrogen (N) content is strongly and positively associated with carbon exchange rates (CER), radiation use efficiency (RUE), and total plant biomass (Sinclair and Horie, 1989). Similarly, low phosphorus (P) and potassium (K) contents can also lead to limited CER and biomass production (Longstreth and Nobel, 1980; Fredeen et al., 1990). Limited availability of other nutrients (e.g. calcium, magnesium, sulphur, zinc, and iron, among others) can limit plant growth and reduce the nutritional quality of the harvested product, but research on their effects on plant processes is sparse. Responses to ozone concentrations (O3) are expected to affect negatively leaf area dynamics, light interception, and biomass allocation and accumulation, but data scarcity has precluded an accurate simulation of this process (Ewert and Porter, 2000). Understanding, parameterizing, and evaluating many of these responses in models is essential for impacts science. Approaches for assessing climate impacts Methods to assess impacts can be classified in projection-based approaches and utility-based approaches. Utility-based approaches (also known as decision-based approaches) focus on making decisions that are robust against the known uncertainties. This is usually done by exploring the outcomes of decisions under a number of plausible scenarios and then choosing those decisions whose outcomes are not affected by the underlying uncertainties (Vermeulen et al., 2013). Projection-based approaches (also known as predict-then-act approaches) are based on the use of models and data to produce projections of a given system’s future state that can be used by decision-makers. Projection-based approaches therefore focus on reducing uncertainties in order to provide decision-makers with information that can be directly used to make a decision. As with most of the modelling literature, this review focuses on projection-based approaches. In the following sections, a summary of related methods is provided. For further discussion on decision-based approaches the reader is referred to Vermeulen et al. (2013). In projection-based frameworks, global climate model projections for one or more given forcing scenarios are, typically, first scaled and/or bias-corrected to produce climate scenarios. Crop models are then forced using these climate scenarios to produce a range of projections that are then used to conceptualize and develop adaptation strategies to be tested or implemented at different scales (from global to the field) (Fig. 3). Modelling choices across the framework shown in Fig. 3 are thus varied and can produce differing responses, thus causing uncertainty. It is expected for almost all the steps in the impact assessment process that uncertainty will increase, although it can be reduced via model calibration and evaluation. The global meta-analysis of Challinor et al. (2014b) is particularly useful in portraying some of the uncertainties to which impact projections are subjected. Fig. 3. View largeDownload slide Ways in which an impact assessment is typically approached in projection-based frameworks. The arrows indicate the flow of information. The hollow arrow at the bottom shows that, as long as more information is derived from climate projections, uncertainties are likely to increase as a result of what is known as ‘cascade of uncertainties’. Fig. 3. View largeDownload slide Ways in which an impact assessment is typically approached in projection-based frameworks. The arrows indicate the flow of information. The hollow arrow at the bottom shows that, as long as more information is derived from climate projections, uncertainties are likely to increase as a result of what is known as ‘cascade of uncertainties’. The role of process-based models in estimates of climate change impacts and adaptation The choice of both crop models and climate model projection types for climate change impact assessment varies substantially across modelling studies (White et al., 2011). Nevertheless, the vast majority of projection-based studies focus on a site-specific scale and use process-based simulation models, though a recent trend exists for regional-scale studies that use simple (yet process-based) or statistical models (Ramirez-Villegas and Challinor, 2012). Rivington and Koo (2011) report the existence of 122 crop models—from which roughly one-half are process-based. Due to the focus of this review, in this section, emphasis is placed on process-based models. Process-based models are both the most diverse and the most complex of the two model types reviewed here and can themselves be divided into two categories according to scale and level of complexity: (i) regional-scale and (ii) field-scale. Regional-scale models have been designed to capitalize on large-scale crop–climate relationships and thus operate at scales commensurate with those of global and regional climate models (i.e. 25–100 km). Despite their reduced complexity, regional-scale models retain enough mechanistic detail in plant growth processes to be used with reasonable confidence under future climate scenarios, including increased CO2 concentrations and higher rates of extreme temperature and drought events (Challinor et al., 2004, 2007). Conversely, field-scale crop models are tools aimed at simulating growth processes in plants so that technological changes and environmental effects at the farm level can be assessed (El-Sharkawy, 2005). Initially, field-scale models were conceived with the objective of being perfect and comprehensive and able to reproduce all plant functions (the ‘universal model’ myth: see Sinclair and Seligman, 1996), though they rapidly evolved into approaches that were theoretically coherent, yet different in their implementation and purpose (Affholder et al., 2012). While the choice of which processes to represent in detail, and the level of detail achieved for a given process, is limited by an understanding of crop physiology derived from available data (Craufurd et al., 2013), it is also governed by research focus and intended model use. The guiding principle for designing abstractions in such models is to: ‘Use the right level of description to catch the phenomena of interest. Don’t model bulldozers with quarks’ (Goldenfeld and Kadanoff, 1999). Designing models for extensibility and correctness There are three key aspects involved in the development and use of well-established process-based crop models: (i) the modelling of biophysical processes, (ii) the selection and maintenance of technical methodologies, and (iii) collaborative community support. Modelling biophysical processes involves choosing the right abstractions to map the interactions of genotype, management, and environment to phenotypic traits of interest. The selection of technical methodology involves choosing programming languages, software environments, data formats, collaboration software, computing hardware, and protocols for maintaining model quality (e.g. automated testing) and uncertainty (e.g. model ensembles). Collaborative community support includes communication between developers of the model, between the modelling team and other expert modelling groups, and between model developers, users, and the wider community of stakeholders (such as farmers, consultants and policy-makers). These key modelling aspects have traditionally been undertaken within individual research groups, often using ad hoc procedures—although with exceptions (e.g. the International Consortium for Agricultural Systems Applications, ICASA; White et al., 2013). However, two relatively recent developments have had a significant impact on the design and development of process-based crop models. First, a significant increase in available computer processing power has enabled ever-increasing complexity in the processes being modelled. ‘Next generation’ frameworks spanning processes from gene expression to climate change are becoming available (Holzworth et al., 2014). Second, the rapid adoption of online tools has enabled global collaborative model development (McLaren et al., 2009) and inter-comparison (AgMIP; Rosenzweig et al., 2013) and changed expectations regarding the availability of model source code and data. Contemporary process-based crop models are increasingly being used to combine sub-components (such as different crop types and genotypic traits) in novel ways. These models are typically not developed in isolation, but are the refinement and integration of pre-existing algorithms, data, and models (see Fig. 1 in Holzworth et al., 2014). In addition, they are developed and tested in a variety of programming languages and computing environments, utilizing agronomical and climate data provided in a wide variety of formats. This increased complexity of process-based crop modelling, and the global, cross-disciplinary nature of model development, assessment, and use, has led to modelling groups adopting more formal techniques to support their research. In particular, to facilitate scientific reproducibility, sharing, inter-comparison, and the integration of sub-models and data, the crop modelling community is increasingly relying on tools and techniques from the software development community. The use of support tools such as wikis, source code version control, and issue tracking (as in the GLAM, DSSAT, and APSIM communities), online user interfaces (Hochman et al., 2009), and the adoption of modular source code frameworks, is becoming more frequent. For example, the current APSIM process-based crop modelling framework (Holzworth et al., 2014) employs (i) a modular software structure that allows components to be combined in novel ways at runtime and to be improved and tested in isolation, (ii) XML configuration files allowing model parameters and custom logic to be shared in a standardized way, and (iii) the integration of scripting language control (including the R and C# languages) that facilitates quick prototyping and sharing of model logic. While such developments are significant steps towards improved model sharing, uncertainty analysis, and code correctness, more work needs to be done. Automated testing, source code version control, and modular model structure are not yet ubiquitous process-based modelling practices. Standardization of common parameter names and their definitions would facilitate more complete model inter-comparisons. Significant gains can be achieved through the adaptation of the software design patterns process (Gamma et al., 1994) to document key crop modelling components such as biophysical processes, model structure, ensemble design, and model inter-comparison, in a form independent of any specific implementation or programming language. The development of such patterns would help reduce the reinvention of solutions, encourage the use of state-of -the-art procedures, and provide a community platform for crop model improvement. The use of ensembles for informing impacts and adaptation The aforementioned increase in the complexity and number of models, along with significant advances in the climate models used to drive regional-scale yield projections, has led to greater confidence in our model projections. However, increasing model detail has meant that uncertainty in projections is not being reduced (see, for example, Knutti and Sedláček, 2012). In addition, model simplifications (such as regional scale process-based, statistical, and niche-based models) have introduced their own uncertainties in terms of spatio-temporal scaling and specificity, and the interrelated lineage of process-based crop models complicates assessments of model uncertainty. As a result, an emphasis on quantifying the uncertainty in projected yields has become prevalent (Iizumi et al., 2009; Asseng et al., 2013). Crop predictions based on single parameter sets or single model output values are no longer good enough. Consequently, projecting crop responses under future climate scenarios requires careful treatment of issues related to parameter uncertainty, structural uncertainty (model discrepancy), algorithmic uncertainty (code uncertainty), parametric variability, experimental uncertainty (observation error), and interpolation uncertainty (Kennedy and O’Hagan, 2001; Challinor et al., 2009a). While accounting for all of these uncertainty sources is critical for the robust use of environmental models in general, the tendency for crop models to be developed using information from one spatial scale, and applied at another, means that crop modellers must pay particular attention to parameter, structural, and interpolation uncertainty. An assessment of 178 published studies on climate-change impacts (sourced by searching the keywords ‘climate change impacts’ in http://scholar.google.com in June 2014) indicates that field-scale, regional-scale process-based models, and statistical models are used at a variety of spatial scales (Fig. 4). For field-scale process-based models, the fact that c. 50% of studies use the models at scales other than those for which the models were originally designed suggests some potential for model versus study scale mismatch or even model misuse (Fig 4A). While, mathematically, one-dimensional models can be used across different spatial scales, remarkably, virtually no study using field-scale process-based models at scales beyond individual fields assesses parameter uncertainty or parameter scaling issues (Fig. 4B; Iizumi et al., 2014). More importantly, the implications of model misuse, including the use of models that lack key processes and scale mismatches, may impact further estimates of adaptation (Challinor et al., 2014a; Lobell, 2014). This is of particular importance since about one in every three studies does not conduct model evaluation regardless of the type of model used (Fig. 4C). Fig. 4. View largeDownload slide Use and misuse of crop models, based on 178 model results published in climate-change impact studies between 1994 and 2014, and disaggregated by model type. (A) Fraction of results that perform simulations at the scale for which the model was designed; (B) fraction of results (at scales other than field) for each model type that use multiple parameter sets (i.e. account for parametric uncertainty); and (C) fraction of studies that state model evaluation procedures for their locations or areas of interest. Model types are as follows: CSM-FS: field-scale crop growth simulation model; CSM-RS: regional-scale crop growth simulation model; E/S: empirical and/or statistical. Note that field-scale models are used above field scale in roughly 50% of the cases. Fig. 4. View largeDownload slide Use and misuse of crop models, based on 178 model results published in climate-change impact studies between 1994 and 2014, and disaggregated by model type. (A) Fraction of results that perform simulations at the scale for which the model was designed; (B) fraction of results (at scales other than field) for each model type that use multiple parameter sets (i.e. account for parametric uncertainty); and (C) fraction of studies that state model evaluation procedures for their locations or areas of interest. Model types are as follows: CSM-FS: field-scale crop growth simulation model; CSM-RS: regional-scale crop growth simulation model; E/S: empirical and/or statistical. Note that field-scale models are used above field scale in roughly 50% of the cases. In the last ten years, the critical task of quantifying and accounting for the full range of uncertainty sources in models has been recognized by the weather, climate, and hydrological communities (Stainforth et al., 2005; Beven, 2006). However, there has been limited applied appreciation for these issues in the crop modelling community besides quantifying parameter (Iizumi et al., 2009; Tao and Zhang, 2013) and structural uncertainty in impacts projections (Ruane et al., 2013; Asseng et al., 2013). While many crop–climate impact studies include some treatment of modelling uncertainty (e.g. by using various future climate projections, crop parameters, and crop models), sampling of the entire model and parameter space is fundamentally incomplete and is likely to underestimate the importance of uncertainty in model-based projections of impacts and adaptation. Therefore, in order for the crop modelling community to move towards ensembles that better sample the uncertainty space and provide useful information for food security assessments, platforms that allow model, parameter, and input transferability between groups and regions so as to facilitate ensemble simulations for both site- and regional-scale assessments need to be developed (also see the section ‘Designing models for extensibility and correctness’). In addition, characterizing the crop model space (Angulo et al. (2013) and better understanding of parameter and process scaling (Iizumi et al., 2014) will ultimately allow for a better understanding and sampling of the model and parameter uncertainty space. Design of genotypic adaptation strategies using crop models The importance of genotypic adaptation Genotypic adaptation is expected to be one of the most important adaptation strategies to future climate change (Challinor et al., 2009b; Semenov and Stratonovitch, 2013). For instance, Challinor et al. (2014b) indicated that switching from currently grown to better-adapted varieties that are cultivated elsewhere or stored in gene banks (‘cultivar adjustment’) is a more effective adaptation strategy than adjusting planting dates, improving irrigation, and enhancing fertilization (Fig. 5). In addition, increased evidence exists that climate-change stresses can, to a large extent, be managed or completely offset through the breeding of new ‘climate-smart’ cultivars with improved yield potential and stability (Ortiz et al., 2008; Semenov and Stratonovitch, 2013). Progress in crop breeding demonstrates the scales of potential yield gains. In Africa, two decades of maize breeding have led to mean genetic gains of 14kg ha–1 year–1 under drought and 40kg ha–1 year–1 under optimum conditions (Badu-Apraku et al., 2013). Similarly, mean wheat breeding gains in hot (seasonal mean temperature >25 °C) environments in the last 25 years are about 100kg ha–1 year–1 under drought and 25kg ha–1 year–1 under well-watered conditions (Gourdji et al., 2013). For rice, genetic gains have been estimated as 45kg ha–1 year–1 for Brazilian upland systems in the period 2002–2009 (Breseghello et al., 2011), whereas, in irrigated rice in Asia, the release of the semi-dwarf rice variety IR8 alone produced an increase of almost 70% in rice potentials during the 1950s and 1960s (Peng et al., 2008). Fig. 5. View largeDownload slide The benefit of different adaptation practices expressed as percentage change, from the baseline, in yield change with adaptation minus yield change without adaptation (adapted from Challinor et al., 2014b by permission from Macmillan Publishers Ltd: Nature Climate Change4, 287–291. Challinor AJ, Watson J, Lobell DB, Howden SM, Smith DR, Chhetri N. A meta-analysis of crop yield under climate change and adaptation. © 2014). Data in this figure consist of yield changes from 32 simulation studies for various crops as described in Challinor et al. (2014b). Bars are means for each category and lines indicate the standard error. Note that the vast majority of data in the second category come from a single study (Deryng et al., 2011). Fig. 5. View largeDownload slide The benefit of different adaptation practices expressed as percentage change, from the baseline, in yield change with adaptation minus yield change without adaptation (adapted from Challinor et al., 2014b by permission from Macmillan Publishers Ltd: Nature Climate Change4, 287–291. Challinor AJ, Watson J, Lobell DB, Howden SM, Smith DR, Chhetri N. A meta-analysis of crop yield under climate change and adaptation. © 2014). Data in this figure consist of yield changes from 32 simulation studies for various crops as described in Challinor et al. (2014b). Bars are means for each category and lines indicate the standard error. Note that the vast majority of data in the second category come from a single study (Deryng et al., 2011). Under future climate scenarios, ideotype design appears as a key strategy to drive breeding decisions, since breeding towards a crop ideotype is more efficient than breeding to remove undesired characteristics one at a time (Peng et al., 2008). Crop ideotypes are idealized plant types that have the greatest effectiveness in producing dry matter and yield under given environmental conditions (Donald, 1968). Defining a crop ideotype involves a definition of the physical-morphological (e.g. height, maximum leaf size, leaf thickness, and positioning) and physiological (e.g. stomatal conductance, photosynthetic efficiency) characteristics of a given crop plant, that would allow such a plant to respond well under certain conditions (e.g. in a drought-prone environment). Breeding programmes are currently challenged with having to set priorities based on climate-change impacts projections (Cairns et al., 2013). Decisions of which traits to breed and by when varieties would need to hold such traits are expected to be largely influenced by the type (e.g. increase in mean, increase in extreme events), direction (e.g. drier and warmer, wetter and warmer), and extent (how much warmer, how much drier) of the projected climatic changes in a given area (Stamp and Visser, 2012). Many breeding programmes, however, already work towards achieving crop ideotypes for different agro-environmental zones (Berry et al., 2007; Peng et al., 2008). Hence, progress towards better future food security prospects of increased food availability and stability through breeding better-adapted crop varieties seems, at least in principle, possible to achieve. The potential role of crop models for developing genotypic adaptation options Process-based crop models can help make informed decisions with regard to genotypic adaptation options and ideotype design both under current and future climates (Baenziger et al., 2004; Banterng et al., 2004). The main challenge, however, is carefully to interpret modelling outcomes so as to provide information that is of use for breeders. Recent experiences in the use of crop model simulated ideotypes for crop breeding in rice as well as existing model-based investigations of genotypic adaptation and ideotype design reveal encouraging results with regard to increasing food availability and stability in the context of climate change adaptation. Under current climates, probably the most notable example of ideotype design for increasing yield potential is the New Plant Type (NPT) proposed and developed by the International Rice Research Institute (IRRI) and the subsequent establishment of the super rice programme in China inspired by the NPT (Cheng et al., 2007; Peng et al., 2008). IRRI’s NPT had its origins in the work of Dingkuhn et al. (1991), who used a process-based growth simulation model to propose a rice ideotype. Based upon model simulations, they hypothesized that 25% productivity gains could be achieved by increasing the length of the grain-filling phase, maintaining high concentration of nitrogen in the leaves, increasing the vertical gradient of nitrogen in the foliage (so that top leaves have more N, and lower leaves have less), enhancing leaf growth in the early stages with reduced leaf growth in later stages, larger panicles but reduced tillering capacity (i.e. a lower number of panicles), more assimilates in the stems, and a longer life span and larger size of the flag leaves (Dingkuhn et al., 1991). Since morphological characteristics are easier to select for in breeding trials, a more precise definition of these was done in a subsequent study (Khush, 1995) (Fig. 6). Two breeding cycles then led to the development of NPT varieties that out-yielded check varieties (Peng et al., 2008). Following IRRI’s promising results, the super rice programme in China was established (Cheng et al., 2007). In addition to what had been proposed by IRRI’s NPT breeding programme, a more specific definition of the position and size of the flag leaves and an optimization of photosynthetic efficiency were done. Newly developed super rice varieties reportedly out-yielded commonly cultivated rice hybrids by 15–25% in many regions of China (Peng et al., 2008). Further research and development of ideotype rice cultivars and hybrids is currently being pursued both in China and internationally by IRRI. Fig. 6. View largeDownload slide Different plant types of rice. Left: tall conventional plant type. Centre: improved high-yielding and high-tillering plant type typical of the green revolution. Right: low-tillering ideotype (new plant type) with larger sink capacity (larger panicles and grains) and sturdier stems. Taken from Khush (2001). Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Genetics2, 815–822. Khush GS. Green revolution: the way forward. © 2001. Fig. 6. View largeDownload slide Different plant types of rice. Left: tall conventional plant type. Centre: improved high-yielding and high-tillering plant type typical of the green revolution. Right: low-tillering ideotype (new plant type) with larger sink capacity (larger panicles and grains) and sturdier stems. Taken from Khush (2001). Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Genetics2, 815–822. Khush GS. Green revolution: the way forward. © 2001. Under future climates, by contrast, to the knowledge of the authors, no breeding programme is currently breeding a model-designed plant type; although the WHEAt and barley Legacy for Breeding Improvement (WHEALBI) appears as a new (started in early 2014) promising initiative. Nevertheless, the number of model-based studies investigating genotypic adaptation through the introduction of novel traits and ideotype design has been increasing in the last decade. All these studies point in the same direction: that, in most situations, genotypic adaptation can offset climate-change-related losses and even boost crop yields. For instance, studies on wheat indicate that climate-ready varieties would out-yield currently cultivated varieties by 25–65% under future climates (Semenov et al., 2014), and similar figures have been reported for other crops such as groundnut, sorghum, and maize (Fig. 7). These figures are, however, contingent on two key modelling aspects. Fig. 7. View largeDownload slide Simulated future potential benefits from genotypic adaptation (including ideotype design) as derived from available modelling studies for four different crops in different sites. Studies are as follows: Semenov and Stratonovitch (2013) and Challinor et al. (2010) for wheat; Singh et al. (2014) for sorghum; Singh et al. (2012, 2013) and Challinor et al. (2007, 2009b) for groundnut; and Lobell et al. (2013) for maize. The benefit of genotypic adaptation has been calculated as the difference between yield changes under adaptation and that under no adaptation, except in the case of Challinor et al. (2010) for which the relative change in crop failure rate between adaptation and no-adaptation results was used. Thick lines are the medians, boxes represent the interquartile range, whiskers extend to 5–95 % of the data, and the dots are outliers. Fig. 7. View largeDownload slide Simulated future potential benefits from genotypic adaptation (including ideotype design) as derived from available modelling studies for four different crops in different sites. Studies are as follows: Semenov and Stratonovitch (2013) and Challinor et al. (2010) for wheat; Singh et al. (2014) for sorghum; Singh et al. (2012, 2013) and Challinor et al. (2007, 2009b) for groundnut; and Lobell et al. (2013) for maize. The benefit of genotypic adaptation has been calculated as the difference between yield changes under adaptation and that under no adaptation, except in the case of Challinor et al. (2010) for which the relative change in crop failure rate between adaptation and no-adaptation results was used. Thick lines are the medians, boxes represent the interquartile range, whiskers extend to 5–95 % of the data, and the dots are outliers. (i) The ability of the model correctly to simulate processes that are relevant in future climate scenarios The fact that all existing models have been subjected to varying degrees of evaluation mostly against agronomic trial data (Asseng et al., 2013; Bassu et al., 2014) and many individual model components (e.g. water balance, photosynthesis response) are often assessed independently has increased confidence on the capabilities of models to simulate crop responses under varying environmental conditions, including climate change. Recent literature, however, indicates shifting climate distributions and an increased likelihood of extreme events (Battisti and Naylor, 2009; Trnka et al., 2014) and this may result in additional and/or different processes constraining future crop yields compared with present-day conditions. Indeed, a recent review identified that only a handful (≤6) of crop models currently used in impact and adaptation studies simulate CO2 impacts on canopy temperature (by computing a soil–plant–atmosphere energy balance), a key process under climate change (White et al., 2011). It is thus not clear whether models already include sufficient detail so as to simulate any additional processes that may arise from projected climate change. This has, in turn, resulted in the need for additional field experiments in which novel conditions and their interaction are evaluated and then tested in multi-model inter-comparison frameworks (Rosenzweig et al., 2013). While these initiatives are clearly a way forward, individual-study assessments of processes and their interactions in single-model and multi-model ensemble simulations as well as more complete descriptions of model limitations with respect to key missing processes are warranted in future genotypic adaptation studies (see Fig. 1 in Singh et al., 2014). Achieving a better representation of future-climate relevant processes will ensure that model-based analyses are more realistic. (ii) The correct separation between model parameters that influence yield as a function of crop physiology and those with a large impact on simulated yield only due to model specification That is, the possibility of relating model parameters to the effect of alleles on given loci or genes controlling key traits (Luquet et al., 2012). Importantly, there is a tight link between such a relationship and model complexity—an overarching issue in climate impacts prediction, because overly simplistic models are unlikely to capture physiological responses with enough level of detail for use in crop breeding (Luquet et al., 2012), but overly complex models are more difficult to constrain at the scales typical of climate prediction frameworks (Challinor et al., 2009a). Work towards linking quantitative trait loci information and process-based crop-growth modelling, however, shows promise. For example, Chenu et al. (2009) used a gene-to-phenotype modelling approach that included a genetic model and a process-based crop model to simulate the impact of leaf and silk elongation traits (as derived from genetic data) on maize yield across different environments. Despite some success, however, the lack of a more thorough consideration of genetic effects (beyond those related to crop development; Messina et al., 2006; Challinor et al., 2009b) on yield and genotype-by-environment interactions in genotypic adaptation studies suggests that appropriate frameworks need to be established (Cooper et al. 2005; Chenu et al., 2009). In order to mainstream crop-model-based analyses of genotypic adaptation into breeding programmes, more research, as well as a framework on the coupling of crop and genetic models, is needed. Figure 8, which based on the work of Chenu et al. (2009) and Cooper et al. (2005), is an attempt at such a framework, through which it is expected that model-based analyses of genotypic adaptation can incorporate genetic information from breeding programmes and, in turn, retrieve ex-ante assessments of genotypic responses across environments (Yin et al., 2003, 2004). As a starting point, traits that have constant QTLs (and hence constant model parameters) across environments have to be determined. Modular crop models can then be coupled with ‘plug-and-play’ parameterizations of relevant characteristics for which genetic information is available, with appropriate sensitivity testing to ensure realism. Genetic model simulations of crosses between promising parental lines can then yield crop model parameters and be run through an ensemble of crop models in one or more environments. The resulting crop model simulations can then be used to select promising phenotypes and the process repeated for various steps in the breeding cycle (Fig. 8). Fig. 8. View largeDownload slide Proposed framework for incorporating genetic information into simulation studies of genotypic adaptation. The figure is derived from the practical example of Chenu et al. (2009). The dashed line that links the genetic portion of the diagram with the environment indicates that analyses are needed to identify traits whose QTLs are constant across environments. Fig. 8. View largeDownload slide Proposed framework for incorporating genetic information into simulation studies of genotypic adaptation. The figure is derived from the practical example of Chenu et al. (2009). The dashed line that links the genetic portion of the diagram with the environment indicates that analyses are needed to identify traits whose QTLs are constant across environments. In addition, the simulation of genotypic adaptation (including ideotype design) for projected weather conditions of an uncertain nature means that additional principles may be needed in order to develop robust projections of adaptation. In particular, appropriately constraining processes and parameters in models across scales (Iizumi et al., 2014) and a shift from simply quantifying uncertainty to actually quantifying robustness (i.e. the relationship between uncertainty and the climate change signal) towards modelling choices (Ramirez-Villegas, 2014) are two key aspects that need to be included into crop model-based analyses of genotypic adaptation. Two key initiatives toward these aims include the AgMIP (Rosenzweig et al., 2013) and FACCE-MACSUR (http://www.macsur.eu) projects. Conclusions The challenges ahead with regard to developing genotypic adaptation strategies that can then be implemented in breeding programmes are substantial. On the one hand, climate change impacts are projected to pose significant challenges to agriculture and genotypic adaptation strategies are critical for responding to such challenges. On the other hand, uncertainties in climate and crop modelling are substantial and poorly explored in studies of genotypic adaptation to future climates that use process-based simulation models, particularly at field scales. While uncertainties need to be better understood and quantified (see the section ‘Design of genotypic adaptation strategies using crop models’), it is important to note that a shift in focus from solely quantifying output variance to quantifying robustness is required in order so as to facilitate assessments and interpretation of confidence levels in crop model-based projections of genotypic adaptation. In addition to this, it is critical that genotypic adaptation options are grounded in genetic and physiological knowledge that can be mainstreamed in real-world breeding programmes. To this aim, while recently published studies have demonstrated the potential of genotypic adaptation strategies and ideotype design, two main goals need to be pursued in future studies: (i) a better understanding of driving processes under future climate change; and (ii) a coupling between genetic and crop growth models—perhaps at the expense of the number of traits analysed. Importantly, the latter may imply additional complexity (and likely uncertainty) in crop modelling studies. Therefore, modularity in crop models as well as individual component testing against observational data would be critical components in any attempts to simulate crop-breeding strategies under future climate scenarios. Acknowledgements This work was supported and funded by the CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). The authors thank members of the Climate Impacts Group at the University of Leeds for insightful discussions during the course of the review. JRV thanks Dr David B Lobell from the University of Stanford for providing the data presented in Fig. 1, Dr Camila Rebolledo from the International Center for Tropical Agriculture (CIAT) for literature and discussion on crop model parameter relationships with phenotypic plasticity and genetics, Camilo Barrios for pointing out key literature on gene-to-phenotype models, and Dr Steve Prager from CIAT for insightful comments on an earlier version of the manuscript. References Affholder F Tittonell P Corbeels M Roux S Motisi N Tixier P Wery J . 2012. Ad hoc modeling in agronomy: what have we learned in the last 15 years? Agronomy Journal  104, 735– 748. Google Scholar CrossRef Search ADS   Ainsworth EA Leakey ADB Ort DR Long SP . 2008. FACE-ing the facts: inconsistencies and interdependence among field, chamber and modeling studies of elevated [CO2] impacts on crop yield and food supply. New Phytologist  179, 5– 9. Google Scholar CrossRef Search ADS PubMed  Allen LH Boote KJ Vara Prasad P V Thomas JMG Vu JCV . 2005. Hazards of temperature on food availability in changing environments (Hot-Face): global warming could cause failure of seed yields of major food crops  . Broomfield, Colorado, USA. Angulo C Rötter R Trnka M Pirttioja N Gaiser T Hlavinka P Ewert F . 2013. Characteristic ‘fingerprints’ of crop model responses to weather input data at different spatial resolutions. European Journal of Agronomy  49, 104– 114. Google Scholar CrossRef Search ADS   Asseng S Ewert F Rosenzweig C et al.   . 2013. Uncertainty in simulating wheat yields under climate change. Nature Climate Change  3, 827– 832. Google Scholar CrossRef Search ADS   Asseng S Foster IAN Turner NC . 2011. The impact of temperature variability on wheat yields. Global Change Biology  17, 997– 1012. Google Scholar CrossRef Search ADS   Badu-Apraku B Oyekunle M Menkir A Obeng-Antwi K Yallou CG Usman IS Alidu H . 2013. Comparative performance of early-maturing maize cultivars developed in three eras under drought stress and well-watered environments in West Africa. Crop Science  53, 1298– 1311. Google Scholar CrossRef Search ADS   Baenziger PS McMaster GS Wilhelm WW Weiss A Hays CJ . 2004. Putting genes into genetic coefficients. Field Crops Research  90, 133– 143. Google Scholar CrossRef Search ADS   Banterng P Patanothai A Pannangpetch K Jogloy S Hoogenboom G . 2004. Determination and evaluation of genetic coefficients of peanut lines for breeding applications. European Journal of Agronomy  21, 297– 310. Google Scholar CrossRef Search ADS   Bassu S Brisson N Durand J-L et al.   . 2014. How do various maize crop models vary in their responses to climate change factors? Global Change Biology  20, 2301– 2320. Google Scholar CrossRef Search ADS PubMed  Battisti DS Naylor RL . 2009. Historical warnings of future food insecurity with unprecedented seasonal heat. Science  323, 240– 244. Google Scholar CrossRef Search ADS PubMed  Berry PM Sylvester-Bradley R Berry S . 2007. Ideotype design for lodging-resistant wheat. Euphytica  154, 165– 179. Google Scholar CrossRef Search ADS   Beven K . 2006. A manifesto for the equifinality thesis. Journal of Hydrology  320, 18– 36. Google Scholar CrossRef Search ADS   Bird IF Cornelius MJ Keys AJ . 1977. Effects of temperature on photosynthesis by maize and wheat. Journal of Experimental Botany  28, 519– 524. Google Scholar CrossRef Search ADS   Boote KJ Allen LH Vara Prasad PV Baker JT Gesch RW Snyder AM Pan D Thomas JMG . 2005. Elevated temperature and CO2 impacts on pollination, reproductive growth, and yield of several globally important crops. Japanese Journal of Meteorology  60, 469– 474. Google Scholar CrossRef Search ADS   Breseghello F de Morais OP Pinheiro PV et al.   . 2011. Results of 25 years of upland rice breeding in Brazil. Crop Science  51, 914– 923. Google Scholar CrossRef Search ADS   Cairns JE Hellin J Sonder K Araus JL MacRobert JF Thierfelder C Prasanna BM . 2013. Adapting maize production to climate change in sub-Saharan Africa. Food Security  5, 345– 360. Google Scholar CrossRef Search ADS   Challinor AJ Osborne T Shaffrey L Weller H Morse A Wheeler T Vidale PL . 2009a . Methods and resources for climate impacts research. Bulletin of the American Meteorological Society  90, 836– 848. Google Scholar CrossRef Search ADS   Challinor AJ Parkes B Ramirez-Villegas J . 2014a . Land-use heterogeneity causes systematic bias in crop yield response to climate change. Global Change Biology  doi: 10.1111/gcb.12808. Challinor AJ Simelton ES Fraser EDG Hemming D Collins M . 2010. Increased crop failure due to climate change: assessing adaptation options using models and socio-economic data for wheat in China. Environmental Research Letters  5, 34012. Google Scholar CrossRef Search ADS   Challinor AJ Watson J Lobell DB Howden SM Smith DR Chhetri N . 2014b . A meta-analysis of crop yield under climate change and adaptation. Nature Climate Change  4, 287– 291. Google Scholar CrossRef Search ADS   Challinor AJ Wheeler TR Craufurd PQ Ferro CAT Stephenson DB . 2007. Adaptation of crops to climate change through genotypic responses to mean and extreme temperatures. Agriculture, Ecosystems & Environment  119, 190– 204. Google Scholar CrossRef Search ADS   Challinor AJ Wheeler TR Craufurd PQ Slingo JM Grimes DIF . 2004. Design and optimisation of a large-area process-based model for annual crops. Agricultural and Forest Meteorology  124, 99– 120. Google Scholar CrossRef Search ADS   Challinor A Wheeler T Hemming D Upadhyaya H . 2009b . Ensemble yield simulations: crop and climate uncertainties, sensitivity to temperature and genotypic adaptation to climate change. Climate Research  38, 117– 127. Google Scholar CrossRef Search ADS   Cheng S-H Cao L-Y Zhuang J-Y Chen S-G Zhan X-D Fan Y-Y Zhu D-F Min S-K . 2007. Super hybrid rice breeding in China: achievements and prospects. Journal of Integrative Plant Biology  49, 805– 810. Google Scholar CrossRef Search ADS   Chenu K Chapman SC Tardieu F McLean G Welcker C Hammer GL . 2009. Simulating the yield impacts of organ-level quantitative trait loci associated with drought response in maize: a ‘gene-to-phenotype’ modeling approach. Genetics  183, 1507– 1523. Google Scholar CrossRef Search ADS PubMed  Claessens L Antle JM Stoorvogel JJ Valdivia RO Thornton PK Herrero M . 2012. A method for evaluating climate change adaptation strategies for small-scale farmers using survey, experimental and modeled data. Agricultural Systems  111, 85– 95. Google Scholar CrossRef Search ADS   Cooper M Podlich DW Smith OS . 2005. Gene-to-phenotype models and complex trait genetics. Australian Journal of Agricultural Research  56, 895– 918. Google Scholar CrossRef Search ADS   Crafts-Brandner SJ Salvucci ME . 2002. Sensitivity of photosynthesis in a C4 plant, maize, to heat stress. Plant Physiology  129, 1773– 1780. Google Scholar CrossRef Search ADS PubMed  Craufurd PQ Vadez V Jagadish SVK Prasad PVV Zaman-Allah M . 2013. Crop science experiments designed to inform crop modeling. Agricultural and Forest Meteorology  170, 8– 18. Google Scholar CrossRef Search ADS   Deryng D Sacks WJ Barford CC Ramankutty N . 2011. Simulating the effects of climate and agricultural management practices on global crop yield. Global Biogeochemical Cycles  25, GB2006. Google Scholar CrossRef Search ADS   Dingkuhn M Penning de Vries FWT De Datta SK van Laar HH . 1991. Concepts for a new plant type for direct seeded flooded tropical rice. In: Direct seeded flooded rice in the tropics  . Los Baños, Philippines: International Rice Research Institute, 17– 38. Donald CM . 1968. The breeding of crop ideotypes. Euphytica  17, 385– 403. Google Scholar CrossRef Search ADS   El-Sharkawy M . 2005. How can calibrated research-based models be improved for use as a tool in identifying genes controlling crop tolerance to environmental stresses in the era of genomics—from an experimentalist’s perspective. Photosynthetica  43, 161– 176. Google Scholar CrossRef Search ADS   El-Sharkawy MA . 2014. Global warming: causes and impacts on agroecosystems productivity and food security with emphasis on cassava comparative advantage in the tropics/subtropics. Photosynthetica  52, 161– 178. Google Scholar CrossRef Search ADS   Ewert F Porter JR . 2000. Ozone effects on wheat in relation to CO2: modelling short-term and long-term responses of leaf photosynthesis and leaf duration. Global Change Biology  6, 735– 750. Google Scholar CrossRef Search ADS   Fredeen AL Raab TK Rao IM Terry N . 1990. Effects of phosphorus nutrition on photosynthesis in Glycine max (L.) Merr. Planta  181, 399– 405. Google Scholar CrossRef Search ADS PubMed  Gamma E Helm R Johnson R Vlissides J . 1994. Design patterns: elements of reusable object-oriented software  . Reading, UK: Addison-Wesley. Ghannoum O Caemmerer S Von, Ziska LH Conroy JP . 2000. The growth response of C4 plants to rising atmospheric CO2 partial pressure: a reassessment. Plant, Cell and Environment  23, 931– 942. Google Scholar CrossRef Search ADS   Gibson LR Paulsen GM . 1999. Yield components of wheat grown under high temperature stress during reproductive growth. Crop Science  39, 1841– 1846. Google Scholar CrossRef Search ADS   Goldenfeld N Kadanoff LP . 1999. Simple lessons from complexity. Science  284, 87– 89. Google Scholar CrossRef Search ADS PubMed  Gourdji SM Mathews KL Reynolds M Crossa J Lobell DB . 2013. An assessment of wheat yield sensitivity and breeding gains in hot environments. Proceedings of the Royal Society B: Biological Sciences  280, 1752. Hew C-S Krotkov G Canvin DT . 1969. Effects of temperature on photosynthesis and CO2 evolution in light and darkness by green leaves. Plant Physiology  44, 671– 677. Google Scholar CrossRef Search ADS PubMed  Hochman Z van Rees H Carberry PS et al.   2009. Re-inventing model-based decision support with Australian dryland farmers. 4. Yield Prophet® helps farmers monitor and manage crops in a variable climate. Crop and Pasture Science  60, 1057– 1070. Google Scholar CrossRef Search ADS   Holzworth DP Huth NI DeVoil PG et al.   . 2014. APSIM—evolution towards a new generation of agricultural systems simulation. Environmental Modelling & Software  62, 327– 350. Google Scholar CrossRef Search ADS   Huntingford C Hugo Lambert F Gash JHC Taylor CM Challinor AJ . 2005. Aspects of climate change prediction relevant to crop productivity. Philosophical Transactions of the Royal Society B: Biological Sciences  360, 1999– 2009. Google Scholar CrossRef Search ADS   Iizumi T Tanaka Y Sakurai G Ishigooka Y Yokozawa M . 2014. Dependency of parameter values of a crop model on the spatial scale of simulation. Journal of Advances in Modeling Earth Systems  (in press). Iizumi T Yokozawa M Nishimori M . 2009. Parameter estimation and uncertainty analysis of a large-scale crop model for paddy rice: application of a Bayesian approach. Agricultural and Forest Meteorology  149, 333– 348. Google Scholar CrossRef Search ADS   Kennedy MC O’Hagan A . 2001. Bayesian calibration of computer models. Journal of the Royal Statistical Society: Series B (Statistical Methodology)  63, 425– 464. Google Scholar CrossRef Search ADS   Khush GS . 1995. Breaking the yield frontier of rice. GeoJournal  35, 329– 332. Google Scholar CrossRef Search ADS   Khush GS . 2001. Green revolution: the way forward. Nature Reviews Genetics  2, 815– 822. Google Scholar CrossRef Search ADS PubMed  Knutti R Sedláček J . 2012. Robustness and uncertainties in the new CMIP5 climate model projections. Nature Climate Change  3, 369– 373. Google Scholar CrossRef Search ADS   Kobza J Edwards GE . 1987. Influences of leaf temperature on photosynthetic carbon metabolism in wheat. Plant Physiology  83, 69– 74. Google Scholar CrossRef Search ADS PubMed  Labate CA Adcock MD Leegood RC . 1990. Effects of temperature on the regulation of photosynthetic carbon assimilation in leaves of maize and barley. Planta  181, 547– 554. Google Scholar CrossRef Search ADS PubMed  Leakey ADB . 2009. Rising atmospheric carbon dioxide concentration and the future of C4 crops for food and fuel. Proceedings of the Royal Society B, Biological Sciences  276, 2333– 2343. Google Scholar CrossRef Search ADS   Leakey ADB Ainsworth EA Bernacchi CJ Rogers A Long SP Ort DR . 2009. Elevated CO2 effects on plant carbon, nitrogen, and water relations: six important lessons from FACE. Journal of Experimental Botany  60, 2859– 2876. Google Scholar CrossRef Search ADS PubMed  Lobell DB . 2014. Climate change adaptation in crop production: beware of illusions. Global Food Security  3, 72– 76. Google Scholar CrossRef Search ADS   Lobell DB Burke MB Tebaldi C Mastrandrea MD Falcon WP Naylor RL . 2008. Prioritizing climate change adaptation needs for food security in 2030. Science  319, 607– 610. Google Scholar CrossRef Search ADS PubMed  Lobell DB Hammer GL McLean G Messina C Roberts MJ Schlenker W . 2013. The critical role of extreme heat for maize production in the United States. Nature Climate Change  3, 497– 501. Google Scholar CrossRef Search ADS   Long SP Ainsworth EA Leakey ADB Nösberger J Ort DR . 2006. Food for thought: lower-than-expected crop yield stimulation with rising CO2 concentrations. Science  312, 1918– 1921. Google Scholar CrossRef Search ADS PubMed  Longstreth DJ Nobel PS . 1980. Nutrient influences on leaf photosynthesis. Plant Physiology  65, 541– 543. Google Scholar CrossRef Search ADS PubMed  Luquet D Soulié JC Rebolledo MC Rouan L Clément-Vidal A Dingkuhn M . 2012. Developmental dynamics and early growth vigour in rice. 2. Modelling genetic diversity using ecomeristem. Journal of Agronomy and Crop Science  198, 385– 398. Google Scholar CrossRef Search ADS   McLaren CG Metz T van den Berg M Bruskiewich RM Magor NP Shires D . 2009. Chapter 4: Informatics in agricultural research for development. Advances in Agronomy  102, 135– 157. Messina CD Jones JW Boote KJ Vallejos CE . 2006. A gene-based model to simulate soybean development and yield responses to environment. Crop Science  46, 456– 466. Google Scholar CrossRef Search ADS   Nagai T Makino A . 2009. Differences between rice and wheat in temperature responses of photosynthesis and plant growth. Plant and Cell Physiology  50, 744– 755. Google Scholar CrossRef Search ADS PubMed  Ortiz R Sayre KD Govaerts B Gupta R Subbarao GV Ban T Hodson D Dixon JM Iván Ortiz-Monasterio J Reynolds M . 2008. Climate change: can wheat beat the heat? Agriculture, Ecosystems and Environment  126, 46– 58. Google Scholar CrossRef Search ADS   Parent B Tardieu F . 2012. Temperature responses of developmental processes have not been affected by breeding in different ecological areas for 17 crop species. New Phytologist  194, 760– 774. Google Scholar CrossRef Search ADS PubMed  Peng S Huang J Sheehy JE Laza RC Visperas RM Zhong X Centeno GS Khush GS Cassman KG . 2004. Rice yields decline with higher night temperature from global warming. Proceedings of the National Academy of Sciences, USA  101, 9971– 9975. Google Scholar CrossRef Search ADS   Peng S Khush GS Virk P Tang Q Zou Y . 2008. Progress in ideotype breeding to increase rice yield potential. Field Crops Research  108, 32– 38. Google Scholar CrossRef Search ADS   Porter JR Semenov MA . 2005. Crop responses to climatic variation. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences  360, 2021– 2035. Google Scholar CrossRef Search ADS   Porter JR Xie L Challinor AJ Cochrane K Howden M Iqbal MM Lobell DB Travasso MI . 2014. Chapter 7. Food security and food production systems. Climate change 2014: impacts, adaptation and vulnerability. Working Group II Contribution to the IPCC 5th Assessment Report  . Geneva, Switzerland. Ramirez-Villegas J Challinor A . 2012. Assessing relevant climate data for agricultural applications. Agricultural and Forest Meteorology  161, 26– 45. Google Scholar CrossRef Search ADS   Ramirez-Villegas J . 2014. Genotypic adaptation of Indian groundnut cultivation to climate change: an ensemble approach  . PhD thesis, University of Leeds, UK. Available at: http://etheses.whiterose.ac.uk/6372/ Rivington M Koo J . 2011. Report on the meta-analysis of crop modelling for climate change and food security survey  . CGIAR Research Program on Climate Change, Agriculture and Food Security. Rosenzweig C Jones JW Hatfield JL et al.   . 2013. The Agricultural Model Intercomparison and Improvement Project (AgMIP): protocols and pilot studies. Agricultural and Forest Meteorology  170, 166– 182. Google Scholar CrossRef Search ADS   Ruane AC Cecil LD Horton RM Gordón R McCollum R Brown D Killough B Goldberg R Greeley AP Rosenzweig C . 2013. Climate change impact uncertainties for maize in Panama: farm information, climate projections, and yield sensitivities. Agricultural and Forest Meteorology  170, 132– 145. Google Scholar CrossRef Search ADS   Schmitt MR Edwards GE . 1981. Photosynthetic capacity and nitrogen use efficiency of maize, wheat, and rice: a comparison between C3 and C4 photosynthesis. Journal of Experimental Botany  32, 459– 466. Google Scholar CrossRef Search ADS   Semenov MA Stratonovitch P . 2013. Designing high-yielding wheat ideotypes for a changing climate. Food and Energy Security  2, 185– 196. Google Scholar CrossRef Search ADS   Semenov MA Stratonovitch P Alghabari F Gooding MJ . 2014. Adapting wheat in Europe for climate change. Journal of Cereal Science  59, 245– 256. Google Scholar CrossRef Search ADS PubMed  Sinclair TR Horie T . 1989. Leaf nitrogen, photosynthesis, and crop radiation use efficiency: a review. Crop Science  29, 90– 98. Google Scholar CrossRef Search ADS   Sinclair TR Seligman NG . 1996. Crop modeling: from infancy to maturity. Agronomy Journal  88, 698– 704. Google Scholar CrossRef Search ADS   Singh P Boote KJ Kumar U Srinivas K Nigam SN Jones JW . 2012. Evaluation of genetic traits for improving productivity and adaptation of groundnut to climate change in India. Journal of Agronomy and Crop Science  198, 399– 413. Google Scholar CrossRef Search ADS   Singh P Nedumaran S Ntare BR Boote KJ Singh NP Srinivas K Bantilan MCS . 2013. Potential benefits of drought and heat tolerance in groundnut for adaptation to climate change in India and West Africa. Mitigation and Adaptation Strategies for Global Change  19, 509– 529. Google Scholar CrossRef Search ADS   Singh P Nedumaran S Traore PCS Boote KJ Rattunde HFW Prasad PVV Singh NP Srinivas K Bantilan MCS . 2014. Quantifying potential benefits of drought and heat tolerance in rainy season sorghum for adapting to climate change. Agricultural and Forest Meteorology  185, 37– 48. Google Scholar CrossRef Search ADS   Stainforth DA Aina T Christensen C et al.   . 2005. Uncertainty in predictions of the climate response to rising levels of greenhouse gases. Nature  433, 403– 406. Google Scholar CrossRef Search ADS PubMed  Stamp P Visser R . 2012. The twenty-first century, the century of plant breeding. Euphytica  186, 585– 596. Google Scholar CrossRef Search ADS   Tao F Zhang Z . 2013. Climate change, wheat productivity and water use in the North China Plain: A new super-ensemble-based probabilistic projection. Agricultural and Forest Meteorology  170, 146– 165. Google Scholar CrossRef Search ADS   Trnka M Rötter RP Ruiz-Ramos M Kersebaum KC Olesen JE Žalud Z Semenov MA . 2014. Adverse weather conditions for European wheat production will become more frequent with climate change. Nature Climate Change  4, 637– 643. Google Scholar CrossRef Search ADS   Tubiello FN Amthor JS Boote KJ Donatelli M Easterling W Fischer G Gifford RM Howden M Reilly J Rosenzweig C . 2007. Crop response to elevated CO2 and world food supply. European Journal of Agronomy  26, 215– 223. Google Scholar CrossRef Search ADS   Vermeulen SJ Challinor AJ Thornton PK et al.   . 2013. Addressing uncertainty in adaptation planning for agriculture. Proceedings of the National Academy of Sciences, USA  110, 8357– 8362. Google Scholar CrossRef Search ADS   White JW Hoogenboom G Kimball BA Wall GW . 2011. Methodologies for simulating impacts of climate change on crop production. Field Crops Research  124, 357– 368. Google Scholar CrossRef Search ADS   White JW Hunt LA Boote KJ Jones JW Koo J Kim S Porter CH Wilkens PW Hoogenboom G . 2013. Integrated description of agricultural field experiments and production: the ICASA version 2.0 data standards. Computers and Electronics in Agriculture  96, 1– 12. Google Scholar CrossRef Search ADS   Wilhelm EP Mullen RE Keeling PL Singletary GW . 1999. Heat stress during grain filling in maize: effects on kernel growth and metabolism. Crop Science  39, 1733– 1741. Google Scholar CrossRef Search ADS   Wong SC Cowan IR Farquhar GD . 1979. Stomatal conductance correlates with photosynthetic capacity. Nature  282, 424– 426. Google Scholar CrossRef Search ADS   Yin X Stamb P Kropffa MJ Schapendonkc AHCM . 2003. Crop modeling, QTL mapping, and Their complementary role in plant breeding. Agronomy Journal  95, 90– 98. Google Scholar CrossRef Search ADS   Yin X Struik PC Kropff MJ . 2004. Role of crop physiology in predicting gene-to-phenotype relationships. Trends in Plant Science  9, 426– 432. Google Scholar CrossRef Search ADS PubMed  © The Author 2015. Published by Oxford University Press on behalf of the Society for Experimental Biology. All rights reserved. For permissions, please email: journals.permissions@oup.com
TI - Identifying traits for genotypic adaptation using crop models
JF - Journal of Experimental Botany
DO - 10.1093/jxb/erv014
DA - 2015-03-07
UR - https://www.deepdyve.com/lp/oxford-university-press/identifying-traits-for-genotypic-adaptation-using-crop-models-1OF2y6LdmR
SP - 3451
EP - 3462
VL - 66
IS - 12
DP - DeepDyve
ER -