Keywords

3.1 Introduction

Maize is the key crop for food and feed security and income generation for millions of smallholder farmers in sub-Saharan Africa (SSA) , Asia , and Latin America . It is a major source of calories in the diets of nearly 230 million inhabitants of developing countries—81 million in SSA, 141 million in South Asia, and 8 million in Latin America. Annual per capita maize consumption averages 36, 10, and 23 kg, respectively, in these regions, but this masks significant variation and per capita food consumption of maize. In Mesoamerica, annual maize consumption exceeds 80 kg per capita in Guatemala, Honduras, and El Salvador, rising to 125 kg in Mexico (Shiferaw et al. 2011).

Maize accounts for almost half of calories and protein consumed in eastern and southern Africa (ESA) and one-fifth of calories and protein consumed in West Africa. Maize consumption levels exceed 130 kg per capita per year in Lesotho, Malawi, and Zambia. The highest amounts of maize consumed are found in southern Africa, at 85 kg/capita/year, as compared to 27 kg/capita/year in East Africa and 25 kg/capita/year in West and Central Africa (Shiferaw et al. 2011). In South and South-east Asia, where direct maize consumption on an annual average is estimated to be only 6 and 16 kg per capita, respectively, there are several areas (especially in the highlands and tribal regions) where maize is consumed directly at much higher rates (Babu et al. 2013).

In Asia, countries such as China, India, and Indonesia have recorded impressive growth rates in maize production (in the range of 5–6 % per year). Maize is now the crop with the largest cultivated area in China, with nearly 33.5 million hectares (Hu and Zimmer 2013); globally, China’s maize acreage and production are next only to the USA. Maize yields have registered impressive increases in China, from 4.5 to 5.75 t/ha (+0.9 % per annum). The Corn Belt of China, stretching from the northeast to south-west of the country, cuts across 11 provinces (Heilongjiang, Jilin, Hebei, Henan, Shandong, Inner Mongolia, Liaoning, Shanxi, Yunnan, Sichuan, and Shaanxi) accounts for 81 % of area under maize and nearly 83 % of total maize produced (Hu and Zimmer 2013).

Although developed countries, particularly the USA, contribute predominantly to maize production, demand for maize in developing countries is expected to surpass the demands for both wheat and rice by the year 2020 (Pingali and Pandey 2001). The growth in demand for human consumption of maize in the developing world is predicted to be 1.3 % per annum until 2020. Moreover, rising incomes are expected to result in a doubling of consumption of meat across the developing world (Naylor et al. 2005), leading to a predicted growth in demand for feed maize of 2.9 % per annum. Hence, there is need for at least a 2 % per annum increase in maize production to meet this growth in global human population and shift in dietary preferences (Ortiz et al. 2010). Maize demand is projected to see 87 % rise by 2020 as compared to its demand in 1995 (IFPRI 2003). An array of factors are contributing to this sharp increase in maize demand, including increase in per capita income, changing diets, and a rapidly growing poultry sector (Shiferaw et al. 2011). For instance, India’s maize demand has been forecast to grow by 36 per cent in the next four years touching 30 million tons in 2017 and double within the next nine years to touch about 44 million tons by 2022.

3.2 The Challenge of Improving Maize Productivity in the Tropics

Both production and productivity have to be significantly improved if the developing world has to successfully meet the rapidly growing demand for maize. The average maize yields in several of the African countries are still below 1 t/ha, while many countries have only 1–2 t/ha, due mainly to poor soils and farmers’ limited access to fertilizer or improved maize seed. Similarly, maize yields in many of Asian countries remain low, with India, Nepal, and the Philippines achieving ≈2 t/ha, Indonesia and Vietnam ≈3.5 t/ha, Thailand almost 4 t/ha, and China 5 t/ha, compared to the world average of 4.7 t/ha in 2005 and current USA average of 9.4 t/ha (Prasanna et al. 2010).

Several factors, including overdependence on rainfall, frequent droughts, yield losses due to pre- and post-harvest pathogens and insect–pests, weeds, poor agronomic management, and lack of access to quality seed, continue to affect maize production and productivity in the developing world, particularly in SSA , Asia , and Latin America . It is notable that eight major maize-producing countries—China, India, Indonesia, Nepal, Pakistan, Philippines, Thailand, and Vietnam—taken together, produce 98 % of Asia’s maize and 28 % of global maize. In most of these countries, maize is predominantly grown under rainfed conditions by the smallholder, resource-poor farmers. Increasing maize yield by even 1 t/ha in the low-yielding countries in Africa and Asia could deliver a much higher relative impact than does the same increase in the high-yielding environments.

3.2.1 Drought Stress

With most maize production dependent on rainfall, especially in the developing world, maize is particularly vulnerable to drought and its yields fluctuate more widely from year to year than is the case for rice and wheat, which are more commonly irrigated. Thus, drought is recognized as the most important constraint across the rainfed lowland and upland environments in the developing world. For instance, over 80 % of maize grown in South and South-east Asia is rainfed, with an average yield that is less than half of the irrigated maize. There is further increase in rainfed maize area at 1.8 % per year, which is six times more than the irrigated area (Edmeades 2007). The decline in the irrigated area is mostly due to the diminishing groundwater table that puts the irrigated area under threat. This situation is likely to exacerbate in the coming decades due to climate change , often leading to inadequate and/or uneven incidence of rainfall in the crop season alongside temperature changes (IPCC 2007; ADB 2009). Alleviating the effects of drought alone could increase average maize yields by 35 % across Asia-7 (excluding China) and by 28 % in south-west China (Gerpacio and Pingali 2007).

3.2.2 Waterlogging Stress

Waterlogging is a major problem for maize production in several maize agro-ecologies where rainfall is erratic and intense, and the soil drainage capacity is poor. Over 18 % of the total maize production area in South and South-east Asia is frequently affected by floods and waterlogging problems, causing production losses of 25–30 % annually (Cairns et al. 2012). The problem of waterlogging during the crop cycle is exacerbated due to climate change in some maize-growing regions in the developing world; for example, the distribution patterns of rainfall rather than total annual rainfall are predicted to change in South Asia and in many areas in SSA (IPCC 2007). Flood and waterlogging frequently affect more than 18 % of the total maize production area in South and South-east Asia causing production losses of 25–30 % annually (Zaidi et al. 2010; Cairns et al. 2012).

Climate change effects are expected to further complicate the already difficult situation of uneven/poor rainfall distribution pattern. Countries in the Greater Himalayan region—Bangladesh, Bhutan, northern India, and Nepal—are facing increased frequency and magnitude of extreme weather events, resulting in flooding, landslides, and devastation of agricultural crops, besides negative impacts on ecological health. The coastal areas of Bangladesh, India, the Maldives, and Sri Lanka are at high risk from projected sea level rise that may cause displacement of human settlements, saltwater intrusion, loss of agricultural land and wetlands, and a negative impact on tourism and fisheries (Ahmad and Suphachalasai 2014). Flash floods occur not only during the seedling stage but also at the flowering and grain filling stages, often forcing the farmers (especially in Bangladesh and Eastern India) to harvest maize ears before physiological maturity.

3.2.3 Heat Stress

Maize is particularly vulnerable to the reproductive stage heat stress. Climate projections also suggest that elevated temperatures, especially in the drought -prone areas of SSA and rainfed areas in South Asia , are likely to result in significant crop yield losses (Cairns et al. 2013a). From analysis of over 20,000 historical maize trial yields in Africa, Lobell et al. (2011) reported a yield reduction of 1 and 1.7 % for every degree-day above 30 °C under optimal rainfed and drought conditions, respectively. Temperatures are expected to increase in SSA by an average of 2.1 °C by 2050 (Cairns et al. 2012). The most important effects of elevated temperatures on maize yield reduction include shortened life cycle, reduced light interception, and increased sterility (Cairns et al. 2012).

3.2.4 Poor Soil Fertility

Declining soil fertility and expanding soil acidity, low phosphorus availability, and aluminium toxicity affect maize yields on about 4 million hectares of land worldwide (Shiferaw et al. 2011). The problem of poor soil fertility is particularly severe in SSA where all the maize mega environments are affected (Pingali and Pandey 2001). Use of fertilizer and restorative crop management practices remains relatively low and inefficient in many developing countries, particularly in SSA (Smale et al. 2011).

3.3 Developing Climate-Resilient Maize Varieties: Some Major Initiatives

The future of maize production and, consequently, the livelihoods of several million smallholder farmers worldwide are based to a great extent on breeding for high-yielding and stress-resilient varieties. The technological opportunities for maize improvement have increased tremendously in recent years. Significant strides have been made particularly with regard to understanding the phenotypic and molecular diversity in the maize germplasm, identification of genes/QTLs influencing diverse traits, especially tolerance to important biotic and abiotic stresses , developing precision phenotyping protocols, and utilizing marker-assisted or genomics-assisted breeding strategies for improving stress resilience in maize. Some of the major initiatives on developing climate-resilient maize for the tropics are highlighted below.

3.3.1 Drought Tolerance

Understanding the environmental conditions that contribute to drought and effectively unravelling genetic variability for drought tolerance in appropriate environments are two critical factors for the success of breeding for drought tolerance. CIMMYT’s work since 1970s on characterization of drought-prone environments in the tropics, identification of suitable secondary traits and trait donors in breeding for drought tolerance, optimizing procedures for undertaking managed drought stress phenotyping trials, developing drought-tolerant (DT) maize germplasm through extensive multi-location and multi-year experiments, and disseminating the stress-tolerant cultivars in partnerships with various public and private organizations, holds considerable significance for improving the livelihoods of the resource-poor farmers in the developing world.

CIMMYT is presently implementing an array of projects in SSA , Asia, and Latin America for developing and deploying climate-resilient varieties. Under the Drought Tolerant Maize for Africa (DTMA) project, jointly implemented by CIMMYT and IITA, in close collaboration with NARS and private sector institutions in 13 countries in Africa, nearly 180 drought-tolerant maize varieties have been released during 2007–2014, with close to 60 % of these being hybrids. These varieties perform as well as or better than the commercial varieties currently available on the market under optimum (no water deficit stress) conditions and outperform the best commercial checks by at least 25–30 % under drought stress and low-input conditions. DTMA has also facilitated production and delivery of about 52,000 tons of DT maize seed in 2014 in partnerships with about 90 seed companies, benefiting an estimated 5 million African households.

The DT varieties developed by CIMMYT typically have a combination of traits that confer them tolerance to drought conditions; these include reduced barrenness under drought stress, short anthesis-silking interval, reduced leaf senescence (as compared to susceptible germplasm), and longer leaf area duration during grain filling (Edmeades 2008; Bruce et al. 2002). Some of the DT varieties developed by CIMMYT and released in SSA have wide adaptation. For example, one of the most popular DT varieties (ZM521), developed at CIMMYT-Zimbabwe, is currently grown in several countries in eastern and southern Africa, including Angola, Burundi, Ethiopia, Kenya, Malawi, Mozambique, South Africa, Tanzania, Zambia, and Zimbabwe.

The Water Efficient Maize for Africa (WEMA) Project is another important public–private partnership , that is intensively engaged in developing and deploying drought-tolerant and insect-resistant white maize varieties in five target countries in SSA (Kenya, Tanzania, Uganda, Mozambique, and South Africa), through a combination of conventional breeding, marker-assisted breeding, and transgenes.

3.3.2 Heat Stress Tolerance

Compared to other abiotic stresses associated with climate change , especially drought stress, work on developing and deploying heat stress-tolerant tropical/subtropical maize is still in its infancy. Studies undertaken by the CIMMYT team to identify heat stress-tolerant tropical maize lines among the elite, DT maize germplasm developed in Mexico, Asia , and Africa revealed high vulnerability of most of the tropical maize germplasm, including commercial cultivars in South Asia , SSA, and Latin America , to reproductive stage heat stress. Several of the DT parents developed by CIMMYT and widely used in hybrid maize breeding in eastern and southern Africa were found to be highly susceptible to drought stress under elevated temperatures; a notable example is CML442 × CML444 that is used as the female parent in several commercial hybrids. Therefore, intensive efforts are required to ensure that the most widely used DT inbred lines and hybrids also possess tolerance to combined drought and heat stresses, especially for deployment in drought-prone areas where temperatures are predicted to increase.

Despite the above-mentioned limitation, a few CIMMYT inbred lines with high levels of tolerance to drought as well as combined drought and heat stress, most notably La Posta Sequia C7-F64-2-6-2-2 and DTPYC9-F46-1-2-1-2, have been identified. Such lines are presently being utilized in developing elite germplasm with tolerance to combined drought and heat stress (Cairns et al. 2012, 2013b). CIMMYT is presently implementing two major research projects, supported by USAID under the Feed the Future initiative, for developing and deploying heat-resilient maize for SSA and Asia. The Heat-Tolerant Maize for Asia (HTMA) Project, initiated in 2012, brings together public and private institutions based in South Asia (Bangladesh, India, Nepal, and Pakistan), besides Purdue University, USA, for accelerated development and deployment of heat stress-resilient maize germplasm.

3.3.3 Waterlogging Tolerance

Considerable variation was observed among maize inbreds in tolerance to waterlogging /flooding of older seedlings (Mano et al. 2006; Zaidi et al. 2010). At EMBRAPA-Brazil, recurrent selection over 12 cycles resulted in the development and subsequent release of the waterlogging-tolerant BRS4154 maize line, with 20 % yield advantage under waterlogging compared to the original source (Ferreira et al. 2007). Both additive and non-additive gene actions contribute to the expression of waterlogging tolerance (Zaidi et al. 2010). QTL mapping undertaken by CIMMYT-Asia team, using single-nucleotide polymorphisms (SNPs), revealed five QTLs on chromosomes 1, 3, 5, 7, and 10 conferring waterlogging tolerance; these QTLs together explained approximately 30 % of phenotypic variance for grain yield under waterlogging stress, with effects ranging from 520 to 640 kg/ha for individual genomic regions (Zaidi et al. 2015).

3.4 Improved Germplasm with Package(s) of Adaptive Traits

For developing climate-resilient maize germplasm, breeding programmes also need to incorporate, more efficiently, packages of traits such as abiotic and biotic stress tolerance. Climate change projections suggest more frequent weather extremes, and occasionally, more than one within one crop season (e.g. drought and waterlogging) could also happen, thereby increasing on the short-run the likelihood of crop failures and on the long-run major production declines (Zaidi and Cairns 2011).

In southern Africa and South Asia, maize farmers are likely to require varieties with tolerance to drought stress at elevated temperatures. Similarly, tolerance to both drought and waterlogging is becoming increasingly important for some areas in South and South-east Asia and may be required by farmers in small areas in SSA (Cairns et al. 2013a). Under the BMZ-funded Abiotic stress-Tolerant Maize for Asia (ATMA) project led by CIMMYT , and with partners in India, Bangladesh, the Philippines, Indonesia, and Vietnam, significant progress was made towards development of improved maize germplasm adapted to South and South-east Asia, and with enhanced levels of tolerance to drought, waterlogging, or combined stress tolerance. New hybrid combinations were developed by crossing promising stress-tolerant lines (DT and/or waterlogging tolerance) and evaluated across moisture regimes, including managed drought and waterlogging stresses, and optimal conditions. A set of approximately 50 promising hybrids are at advanced stage and ready for large-scale adaptive trials.

For effectively developing climate-resilient maize varieties with packages of adaptive traits, breeding programmes need to be reoriented for simultaneous selection under combinations of stresses. This requires establishment of a strong network of managed-stress phenotyping /screening sites and use of standardized protocols for specific combinations of stresses predicted in the target environments. Future maize genotypes should also be equipped with a more efficient rooting system to improve water-use efficiency. From the breeding perspective, this will require developing high-throughput root screening systems (rhizotronics) under both field and controlled conditions and identifying key root traits associated with improved water and nutrient capture in the field. Such efforts could lead to identification of suitable trait donors for improving water and nutrient use efficiency of maize varieties.

3.5 Approaches for Enhancing Genetic Gains in Stress-Prone Environments

The ability to develop, in a cost- and time-efficient manner, elite maize hybrids with high yield potential and necessary adaptive traits (abiotic and biotic stress resilience) will be critical for the improved productivity and diversification of cropping systems. While conventional breeding has been successful in developing an array of elite maize hybrids, rapid advances in breeding tools and techniques, especially doubled haploidy (DH) , high-throughput phenotyping for traits of interest, mechanization of breeding operations (to the extent possible), molecular marker-assisted breeding, and decision support tools/systems (Fig. 3.1) offer excellent opportunities for improving genetic gains and enhancing breeding efficiency . A few of these important components will be briefly discussed here.

Fig. 3.1
figure 1

Components for enhancing genetic gains in maize

Full size image

3.5.1 Doubled Haploid (DH) Technology in Maize Breeding

Greater access to low-cost hybrid seed and more rapid development of improved hybrids are vital to increase maize productivity and enhance income opportunities to maize growers, while meeting the demands of food, feed, and nutrition security. Development of stable and productive inbred lines to produce hybrid seed is the cornerstone of successful and affordable hybrid maize technology. The DH technology is now a powerful tool to accelerate development, identification, and use of elite breeding lines. DH not only significantly reduces the time and resources required for generating homozygous lines, but also enhances “forward breeding” (Geiger and Gordillo 2009; Prasanna et al. 2012).

The in vivo haploid induction using temperate haploid inducers (genetic stocks with high haploid induction capacity) has been adapted by an array of commercial maize breeding programmes in Europe, North America, and more recently in Asia (especially China), but the lack of tropically adapted haploid inducer lines impeded the application of DH technology in the tropical maize breeding programmes (Prigge et al. 2012). Tropically, adapted haploid inducers with a haploid induction rate (HIR) of up to 10 % have been recently developed through collaboration between CIMMYT and the University of Hohenheim (Prigge et al. 2011; Prasanna et al. 2012). These tropicalized haploid inducers have already been shared with a large number of interested institutions in Africa, Latin America , and Asia, for research or commercial use under specific terms and conditions. The availability of tropicalized haploid inducers is expected to significantly enhance the efficiency of DH line production, increase seed set and rates of induction, and reduce the costs of inducer line maintenance and seed production. CIMMYT is also in the process of further optimizing the DH protocol and developing second-generation haploid inducer lines adapted to specific tropical environments, especially SSA , Asia, and Latin America, to further widen the application of DH technology in maize breeding programmes.

While DH technology is the primary mode of deriving new inbred lines by several large private sector breeding programmes, National Agricultural Research Systems (NARS) and small and medium enterprise (SME) seed companies in several Asian countries have so far not derived benefits out of maize DH technology for various reasons. CIMMYT, in partnership with Kenya Agricultural and Livestock Research Organization (KALRO), has also established a centralized maize DH facility at Kiboko (Kenya) for strengthening maize breeding programmes, including those of NARS and SME seed companies in SSA. A similar facility is being planned for Asia, in partnership with Asian institutions.

3.5.2 High-Throughput and Reasonably Precise Phenotyping

Field-based phenotyping still remains a major bottleneck for future breeding progress. Phenotyping capacity of several institutions in Asia is lagging far behind the capacity to generate genomic information. Phenotyping capacity is constrained in many countries, limiting our ability to breed better cultivars with higher grain yield and stress resilience (Prasanna et al. 2013b; Araus and Cairns 2014). Field-based phenotyping of appropriately selected traits, using low-cost, easy-to-handle tools, is now possible and should become an integral and key component in the maize breeding programmes. There is also a distinct need for the public and private institutions to come together and establish “phenotyping networks” for comprehensive and efficient characterization of genetic resources and breeding materials for important target traits.

Molecular breeding strategies, such as genome-wide association studies (GWAS), marker-assisted recurrent selection (MARS), and genome-wide selection (GS), are being implemented by several institutions worldwide. However, genotypic predictions for both MARS- and GS-based strategies depend heavily on a single phenotyping cycle, thus increasing the need for reliable phenotyping methodologies (Cobb et al. 2013).

Combining advances in aeronautics and high-performance computing is paving the way for the development of field-based phenotyping platforms (Araus and Cairns 2014); such platforms could range from ground-based to the aerial. Recently, under the MAIZE CGIAR Research Program, the Crop Breeding Institute (Zimbabwe), University of Barcelona (Spain), AirElectronics (Spain), Consejo Superior de Investigaciones Científicas (Spain), and CIMMYT developed an UAV (unmanned aerial vehicle), called “Skywalker”. The “Skywalker” is able to carry a payload of up to 1 kg and can carry several sensors including thermal, multispectral, and digital cameras. This UAV is currently being used to identify genotypic variability in plant water status under drought stress and biomass production and senescence under drought, heat, and low N stresses (Cairns et al. 2012b) at CIMMYT-Harare, with promising results.

Beyond such technological advances, there is also an immense need for measuring and reducing the effects of field variability, thereby increasing the genetic signal-to-noise ratio to detect real differences between genotypes. CIMMYT is making intensive efforts for characterizing field variability at the key phenotyping sites worldwide and for improving field-based phenotyping through various approaches, such as monitoring soil moisture using neutron probes/time-domain reflectometer (TDR), non-destructive estimation of biomass (using NDVI or Normalized Differential Vegetation Index), and analysing canopy behaviour using Infrared thermography and functional aspects of roots using Rhizotronics (Prasanna et al. 2013b).

Breeding programmes of majority of the NARS and SME seed companies in the developing countries have limited capacity for undertaking precision phenotyping, particularly under repeatable and representative levels of abiotic stresses in the field. Intensive efforts are therefore required to build the capacity of the institutions on methods to characterize and control field site variation (for improving repeatability), adopting appropriate experimental designs, selection of “right” traits for phenotyping, proper integration, analysis, and application of heterogeneous data sets, in addition to generating better awareness of technological advances with regard to phenotyping.

3.5.3 Genomics-Assisted Breeding

Molecular marker-assisted or genomics-assisted breeding is the way forward in effectively meeting the greater challenge of developing cultivars with combinations of relevant adaptive traits, including biotic and abiotic stress tolerance, besides nutritional quality. With the rapid reduction in genotyping costs, new genomic selection technologies (Bernardo and Yu 2007, Heffner et al. 2009) have become available that allow the maize breeding cycle to be greatly reduced, facilitating inclusion of information on genetic effects for multiple stresses in selection decisions.

High-density genotyping using platforms such as genotyping by sequencing (GBS) is now an integral part of CIMMYT’s maize molecular breeding strategies. GWAS is being implemented by CIMMYT-GMP for identification of genomic regions associated with an array of important traits, especially abiotic stress tolerance and disease resistance, coupled with validation of the significant genomic leads in an array of tropical/subtropical biparental populations (Prasanna et al. 2014). MARS and GS are being implemented by CIMMYT and partners through several projects in the tropics, especially for the improvement of complex traits. A recent comparative study of pedigree selection, MARS, and GS undertaken across 8–10 biparental populations demonstrated the superior performance of hybrids derived from Cycle 3 of both MARS and GS schemes over pedigree selection in most populations compared with the Cycle 0. The overall gain per year for MARS and GS under managed drought and well-watered environments was two-to threefold higher than the gain achieved via pedigree selection (Beyene et al. 2014).

CIMMYT is also employing joint GWAS and linkage mapping approach for identifying breeder-ready markers for resistance to major diseases affecting tropical maize such as Turcicum leaf blight (TLB), grey leaf spot (GLS), maize lethal necrosis (MLN), common rust, ear rots, and corn stunt complex. A recent example is with regard to the maize streak virus (MSV), a major disease that affects maize productivity in several countries in SSA . CIMMYT Maize Program has fine-mapped and identified SNP markers for a major QTL for MSV resistance (msv1) and validated these markers on a set of DH lines that have been phenotyped for responses to MSV in different locations in SSA (Sudha Nair et al., manuscript in preparation). Forward breeding and MABC are ongoing using a three-marker haplotype for msv1 selection. Simultaneous with the marker discovery and validation, strategies for incorporating validated markers in breeding pipelines, through both conventional and DH-based breeding schemes, are also being developed and implemented (Prasanna et al. 2014).

3.5.4 Breeding Informatics and Decision Support Tools

A careful balance of many diverse elements is required to design and implement an appropriate decision support system that provides an optimal combination of time, cost, and genetic gain (Xu et al. 2012). Such a system would need to include the following: (a) managing and analysing large amounts of genotype, pedigree, phenotype, and environment data; (b) selecting desirable recombinants through an optimum combination (in time and space) of phenotypic and genotypic information; and (c) developing breeding systems that minimize population sizes, number of generations, and overall costs while maximizing genetic gain for traditional and novel target traits (Prasanna et al. 2014).

Effective management of product and trait pipelines in breeding programmes requires effective management of pedigree, phenotypic, and genotypic databases, accurate forecasting of genotyping and phenotyping services, as well as optimized decision-making tools/system. Standardized software tools for forecasting, project management, and conventional and molecular breeding data review can streamline the process from initial discovery to final deployment of products through coordinated workflows.

3.5.5 Tapping the Vast Genetic Diversity in Maize

Although maize hybrids represent the most economically important portion of the species, breeding populations, open pollinated varieties (OPVs), and landraces contain the majority of the allelic diversity, much of which has never been incorporated into improved maize cultivars. The CIMMYT Gene Bank holds ~27,000 maize entries, of which ~24,000 are landraces/OPVs collected from diverse regions in Latin America , Africa, and Asia , and held in trust since several decades (Ortiz et al. 2010; Prasanna 2012). Many favourable alleles for an array of useful traits, including tolerance to biotic and abiotic stresses and nutritional quality, are available in these invaluable genetic resources, following natural and farmer’s selections over the decades/centuries.

Maize has enormous genetic diversity that offers incredible opportunities for genetic enhancement. There is no lack of favourable alleles in the global maize germplasm that contribute to higher yield, abiotic stress tolerance, disease resistance, or nutritional quality improvement. However, these desirable alleles are often scattered over a wide array of landraces or populations. Our ability to broaden the genetic base of maize and to breed climate-resilient and high-yielding cultivars adaptable to diverse agro-ecologies where maize is grown will undoubtedly depend on the efficient and rapid discovery and introgression of novel/favourable alleles and haplotypes (Prasanna 2012).

A well-characterized and well-evaluated germplasm collection would have greater chances of contributing to the development of novel and improved varieties and, consequently, greater realization of benefits for the resource-poor farmers. Simultaneously with the wider adoption of high-throughput molecular tools, there is a distinct need to establish global phenotyping network for comprehensive and efficient characterization of genetic resources and breeding materials for an array of target traits, particularly for biotic and abiotic stress tolerance and nutritional quality. This would significantly accelerate genomics-assisted breeding, diversification of the genetic base of elite breeding materials, creation of novel varieties, and countering the effects of global climate changes .

Seeds of Discovery (SeeD) is a novel project, funded by the Mexican Government, which aims to discover the extent of allelic variation in the genetic resources of maize and wheat, through high-density genotyping/resequencing, multi-location phenotyping for prioritized traits, and novel bioinformatics tools for discovery and use of favourable alleles and haplotypes associated with important traits (Kevin Pixley, personal communication).

3.6 Delivering Climate-Resilient Maize Varieties to the Smallholder Farmers

Developing and releasing climate-resilient varieties is by itself not adequate to lift communities from climate vulnerability. Affordability and access of smallholder farmers to quality seed of stress-resilient maize varieties in the vulnerable agro-ecologies is highly important. This certainly warrants innovative approaches and partnerships to reach the unreached and to make a difference to their livelihoods.

In recognition that a common constraint for SME seed companies and of rapid scale-up of new varieties is parental line maintenance and foundation (basic) seed production, CIMMYT has been providing appropriate technical support for these activities, at least in the initial phases of variety commercialization, on a case-by-case basis in SSA , Latin America , and Asia . The basis for determining this support is the “seed road maps” that are developed with partner institutions for effective scale-up, promotion, and delivery of improved varieties to the smallholders in the target geographies.

Experiences of CIMMYT strongly indicate that besides strengthening the seed sector (especially the SME seed companies), appropriate government policies and adoption of progressive seed laws and regulations are vital for improving smallholder farmers’ access to improved seed and for overcoming key bottlenecks affecting maize seed value chain. This is particularly important in the areas of policy, credit availability, seed production, germplasm, and marketing. A proactive approach that combines promising technological, institutional, and policy solutions to manage the risks within vulnerable communities implemented by institutions operating at different levels (community, subnational, and national) is considered to be the way forward for managing climate variability and extremes (Shiferaw et al. 2014). Geographic information system (GIS) could play an integral role in targeting breeding programmes by predicting regions of vulnerability, targeting germplasm movement, and identifying future climates for agricultural production environments (Cairns et al. 2013a).

CIMMYT’s recent initiative of establishing the International Maize Improvement Consortium (IMIC) in Asia and in Latin America , in partnership with nearly 80 SME seed companies, is a huge step forward. The underlying principles of this partnership include research prioritization that is client-determined, a more focused, demand-driven approach for product development, while drawing synergies through a collaborative testing network for targeted impacts.

3.7 Conclusions

Intensive multidisciplinary and multi-institutional efforts are required to develop and deliver climate-resilient tropical/subtropical maize germplasm for the benefit of smallholder farmers. Genetics and breeding alone cannot solve the complex challenge of enhancing maize productivity at the smallholder farm level, especially in the face of depleting/degrading natural resources and changing climates. There is a distinct need for effective complementation of improved maize cultivars by suitable conservation agriculture practices as well as institutional and policy innovations that support maize growth and development. This includes understanding the smallholder farmers’ affordability and access to quality seed, measures to overcome constraints in adoption of high-yielding, stress-resilient, and nutritionally enriched maize varieties, and partnerships and policies to significantly enhance maize production and utilization.

The use of higher spatial resolution modelling is essential for the identification of high-priority geographic areas for development and deployment of improved germplasm suited to the future climates. Temperature thresholds for current cereal varieties and the interaction of heat stress with other components of climate change (especially drought and biotic stresses ) must also be considered. The application of biophysical and economic models in maize improvement, decision support, and foresight requires implementing harmonized procedures for data acquisition, incorporating diverse and actual data sets (cultivar-specific data, climatic data, soil data, important macro- and micro- nutrients, pests/pathogens data, crop management practices, and socio-economic data) in a meaningful way for reliable predictions and practical utility (Prasanna et al. 2013a).

Recent research has led to the development of a suite of soil and crop management practices for increasing resource-use efficiency while maintaining soil health, and mitigating greenhouse gas emissions (Govaerts et al. 2009). Increased use of natural resource management is required to reduce agricultural impacts and to increase efficient water use. However, one must recognize the fact that, in general, most of the modern high-yielding varieties were developed using conventional tillage and crop establishment practices in high-input environments. There could be significant variety x management interaction, with variability in the response of current maize germplasm to resource conservation technologies. A new generation of maize cultivars, suitable for conservation agriculture-based practices, need to be bred. In addition, achieving increased adaptation action will necessitate integration of climate change-related issues with other risk factors, such as climate variability and market risk, and with other policy domains, such as sustainable development (Howden et al. 2007).