G. Alagarswamy

East African food security as influenced by future climate change and land use change at local to regional scales

Climate change impacts food production systems, particularly in locations with large, vulnerable populations. Elevated greenhouse gases (GHG), as well as land cover/land use change (LCLUC), can influence regional climate dynamics. Biophysical factors such as topography, soil type, and seasonal rainfall can strongly affect crop yields. We used a regional climate model derived from the Regional Atmospheric Modeling System (RAMS) to compare the effects of projected future GHG and future LCLUC on spatial variability of crop yields in East Africa. Crop yields were estimated with a process-based simulation model. The results suggest that: (1) GHG-influenced and LCLUC-influenced yield changes are highly heterogeneous across this region; (2) LCLUC effects are significant drivers of yield change; and (3) high spatial variability in yield is indicated for several key agricultural sub-regions of East Africa. Food production risk when considered at the household scale is largely dependent on the occurrence of extremes, so mean yield in some cases may be an incomplete predictor of risk. The broad range of projected crop yields reflects enormous variability in key parameters that underlie regional food security; hence, donor institutions’ strategies and investments might benefit from considering the spatial distribution around mean impacts for a given region. Ultimately, global assessments of food security risk would benefit from including regional and local assessments of climate impacts on food production. This may be less of a consideration in other regions. This study supports the concept that LCLUC is a first-order factor in assessing food production risk.

Durations of the photoperiod-sensitive and -insensitive phases of time to panicle initiation in sorghum

Abstract The development of sorghum [Sorghum bicolor (L.) Moench] is influenced by genes that control sensitivity to photoperiod, and their interaction with photoperiod and temperature. While temperature influences development throughout the life cycle of plants, photoperiod influences the vegetative stage (from seedling emergence to panicle initiation). In order to simulate plant development, it is essential to know when sorghum plants first become sensitive to photoperiod, and how long that photoperiod sensitivity persists. Ten cultivars with different levels of photoperiod sensitivity were grown in pots under natural climatic conditions both in short days (SD: 8 h day−1) and long days (LD: 17 h d−1). Plants were transferred at different times after seedling emergence from SD to LD and vice versa. The time to panicle initiation (PI) for each transfer treatment was determined. In cultivars that remained continuously in SD, the time to PI varied from 16 to 27 d, whereas, in continuous LD it varied from 22 to 37 d. The cultivars started reacting to photoperiod 4–9 d after seedling emergence. After sensing photoperiod stimuli, inductive effects among cultivars persisted for 4–14 d in SD, and for 15–33 d in LD depending on their intrinsic photoperiod sensitivity. The sensitivity ended 2–5 d before panicle initiation. This interval, between completion of the photoperiod-inductive phase and the actual observation of PI under the microscope, represents the time required for the photoperiod-inductive stimulus to promote sufficient cell division and growth at the shoot apex for the morphological change to become visible as a shiny globular structure. We conclude that photoperiod sensitivity in these sorghum cultivars ends shortly before or at the PI stage. Our results support the assumptions followed in several crop simulation models that sorghum remains photoperiod-sensitive until the completion of the vegetative stage.

Evaluation and application of the CROPGRO-Soybean simulation model in a Vertic Inceptisol

Crop simulation models are valuable research tools in agricultural decision making. In order to increase its general applicability, models need to be evaluated in diverse conditions. To achieve this, CROPGRO-Soybean model was evaluated on Vertic Inceptisols in a climatically variable semi-arid tropical condition. The model predicted reasonably the temporal changes in leaf area index, biomass and grain yield. The model was used to develop yield–evapotranspiration (ET) relationship, and to assess the influence of soil water-storage capacity on yield. Yield was linearly related to ET and was reduced non-linearly as soil depth decreased. The yield reduction was minimal when depth decreased from 90 to 67 cm but severe reduction occurred when depth decreased below 45 cm. There exists a threshold soil depth (37 cm), below which crop productivity in Vertic Inceptisols cannot be sustained, even in good rainfall years. There is an urgent need to develop sustainable natural resource management technology to prevent further degradation of Vertic Inceptisol. CROPGRO-Soybean model can be successfully used as a research tool to evaluate the risks associated in adapting such technologies.

Soybean–chickpea rotation on Vertic Inceptisols: I. Effect of soil depth and landform on light interception, water balance and crop yields

During 1995-1997 a field study was conducted at the ICRISAT Centre, Patancheru, Andhra Pradesh, India, on a Vertic Inceptisol watershed to study the effect of two soil depths, shallow (<50 cm soil depth) and medium-deep (=50 cm soil depth), and two landform treatments, flat and broadbed-and-furrow (BBF) systems, on productivity and resource-use efficiency of a soyabeans-chickpeas rotation. Soyabeans grown on flat landform on medium-deep soil had a higher leaf area index and more light interception compared with the soyabeans grown on the BBF landform. This resulted in an increase in mean seed yield for the flat landform (2.12 t/ha) compared with the BBF landform (1.87 t/ha). However, the landform treatments on shallow soil did not affect soyabean yields. The soyabean yield was higher on the medium-deep soil (1.76 t/ha) than on the shallow soil (1.55 t/ha) during 1995-1996, but were not different during 1996-1997. In both years chickpea yields and total system productivity (soyabean + chickpea yields) were greater on medium-deep soil than on the shallow soil. Total run-off was higher on the flat landform (25% of seasonal rainfall) than on the BBF landform (20% of seasonal rainfall). This concomitantly increased profile water content (10-30 mm) of both soils in BBF compared with the flat landform treatment during 1995-1996, but not during 1996-1997. Deep drainage was higher in the BBF landform than in flat, especially for the shallow soil. Across landforms and soil depths, water use (evapotranspiration) by soyabeans-chickpeas rotation during 1996-1997 ranged from 496 to 563 mm, which accounted for 54-61% of the rainfall. These results indicate that while the BBF system is useful in decreasing run-off and increasing infiltration of rainfall on Vertic Inceptisols, there is a need to increase light use by soyabeans on BBF during the rainy season to increase its productivity. A watershed-based farming system needs to be adopted to capture significant amount of rain water lost as run-off and deep drainage. The stored water can be used for supplemental irrigation to increase productivity of soyabean-based systems leading to overall increases in resource-use efficiency, crop productivity and sustainability.

Soybean–chickpea rotation on Vertic Inceptisols II. Long-term simulation of water balance and crop yields

A field study was conducted on a Vertic Inceptisol during 1995-1997 seasons at the ICRISAT Centre, Patancheru, India, to study the effect of two landforms (broadbed-and-furrow (BBF) and flat) and two soil depths (shallow and medium-deep) on crop yield and water balance of a soyabean-chickpea rotation. Using two seasons experimental data, a soyabean-chickpea sequencing model was evaluated and used to extrapolate the results over 22 years of historical weather records. The simulation results showed that in 70% of years total runoff for BBF was >35 mm (range 35-190 mm) compared with >60 mm (range 60-260 mm) for flat on the shallow soil. In contrast on the medium-deep soil it was >70 mm (range 70-280 mm) for BBF compared with >80 mm (range 80-320 mm) for the flat landform. The decrease in runoff on BBF resulted in a concomitant increase in deep drainage for both soils. In 70% of years, deep drainage was >60 mm (range 60-390 mm) for the shallow soil and ranged from 10 to 280 mm for the medium-deep soil. In 70% of years, the simulated soyabean yields were >2.2 t/ha (range 2.2-3.0 t/ha) and were not influenced by landform or soil depth. In the low rainfall years, yields were marginally higher for the BBF than for the flat landform, especially on the shallow soil. Simulated chickpea yields were higher for the medium-deep soil than for the shallow soil. In most years, marginally higher chickpea yields were simulated for the BBF than for the flat landform on both soil types. In 70% of years, the chickpea yields were >0.5 t/ha (range 0.5-1.5 t/ha) for the shallow soil, and >0.8 t/ha (range 0.8-1.96 t/ha) for the medium-deep soil. Total productivity of a soyabean-chickpea rotation was >3.0 t/ha (range 3.0-4.15 t/ha) for the shallow soil and >3.45 t/ha (range 3.45-4.7 t/ha) for the medium-deep soil in 70% of years. These results showed that in most years BBF increased rainfall infiltration into the soil and had marginal effect on yields of soyabeans and chickpeas. Crop yields on Vertic Inceptisols can be further increased and sustained by adopting appropriate rain water management practices for exploiting surface runoff and deep drainage water as supplemental irrigation to crops in a watershed setting.

The influence of extended vegetative development and d2 dwarfing gene in increasing grain number per panicle and grain yield in pearl millet

As in other cereals, grain yield in pearl millet is directly related to grain number m−2. Hence, an attempt was made to increase grain numbers by (1) increasing the length of vegetative period (to increase potential grain number per panicle and increase leaf area and light interception before flowering) by using a non-inductive long photoperiod during the early stages of crop growth, and (2) using a dwarfing gene to vary assimilate partitioning between panicle and stem prior to flowering. Extended day length delayed panicle initiation (PI) and flowering and increased leaf area index and assimilate production. Time to flowering was directly related to assimilate allocation to individual panicles, and to grain number per panicle. Delayed PI, however, reduced panicle numbers. Dwarf hybrids partitioned more assimilates to panicles and less to stems, which was also associated with more grains per panicle. The increase in grain number per panicle was offset by decrease in panicles per plant so that neither longer vegetative stage nor dwarfing gene caused in increase in grain yield. With increase in grain numbers per panicle in the dwarfs, grain density (grains cm−2 surface area of panicle) also increased resulting in the dwarf hybrids producing smaller grains and failing to benefit from the increased grain numbers per ear.

Overview of Crop Models for Assessment of Crop Production

Assessment of the impact of climate change on food production systems and the policies needed to adjust to these changes, begins with the biophysical evaluation of crop production as influenced by soil properties, weather variability with climate change superimposed upon it, and potential changes in agrotechnology. The biophysical outcomes must be combined with economic and trade models for final decisions regarding optimal policies to deal with climate change. The outcomes should be appropriately responsive to the most likely climate changes of increased air temperature and higher atmospheric carbon dioxide (CO2)and to the changing technology required for adjusting to the projected climate changes. The biophysical outcomes will be no better than the general circulation model projections of the actual climate change.