Modeling of Climate Change Impacts on Sorghum Yield under Different Inputs and Management Practices in Mali

Abstract

Sorghum (Sorghum bicolor (L.) Moench) is one of the most important crops for human food and animal feeding worldwide. In Mali, sorghum is a vital cereal crop for food security. Despite its importance in the country’s food security, average farmer yield of sorghum remains low (less than 1 t ha-1). Climate variability and change, poor soil fertility, limited improved seeds adoption, and management practices are major factors of this low productivity. The objective of this thesis was to evaluate through experimentation and modelling approaches, the effects of management practices (sowing dates and fertilizer types) on improved sorghum cultivars to enhance the adaptive capacity of farmers on sorghum production in Mali Sudano-Sahelian zone. Experiments were conducted at Samanko, N’Tarla and Cinzana during the 2021 and 2022 growing seasons. In each site, a split-split plot design was used as experimental design with four fertilization types [F1 (the control, no fertilization); F2 (the mineral recommended fertilization rate for sorghum and millet production in Mali, 100 kg ha-1 DAP + 50 kg ha-1 urea); F3 (the organic amendment rate in Mali, 5000 kg ha-1 cow dung); and F4 (the combined option made of half doses of F2 and F3, ie., 50 kg ha-1 DAP + 25 kg ha-1 urea + 2500 kg ha-1 cow dung)] as main factor, two sowing dates (SWD1 and SWD2, medium-early and late sowing dates, respectively) as subfactor and three sorghum varieties (Soubatimi, Jakunbe and Pablo) as sub-sub factor, replicated three times. Days to 50% flowering, plant leaves number at 50% flowering, plant height at harvest, panicle length at harvest, mean weight of 100 grains, aboveground biomasses at 50% flowering and at harvest, grain yield and water use efficiency of grain yield (WUE_GY) were evaluated. From the two cropping seasons trials, DSSAT/CERES-Sorghum model was calibrated with the F2 data and validated with the three other fertilizer types (F1, F3 and F4) data. After validation, seasonal simulation analyses on 31 years (1990-2020) were performed to evaluate the effect of management practices [different fertilization strategies (F1, F2, F3 and F4), and varying sowing windows (early sowing window, SW1 from 1st to 20th of June; medium sowing window, SW2 from 21st of June to 10th of July; and late sowing window, SW3 from 11th to 30th July)] on grain yield, and WUE of the three calibrated cultivars across agroecological zones (Samanko for the Southern Sudanian zone, N’Tarla for the Northern Sudanian zone, and Cinzana for the Southern Sahelian zone). Results concerning the thesis agronomic aspects revealed that experiments sites soil nitrogen and phosphorus contents were low, and the rainfall was also variable during the growing seasons in the study agroecological zones. These environmental factors had different effects on the growth and development of the tested varieties. Moreover, analyses of variance (ANOVA) of trials data showed that days to 50% flowering for the three varieties was mostly shortened under the organo-mineral fertilization F4 compared to the three other fertilizer types at Samanko and N’Tarla. First sowing date (SWD1) and organo-mineral fertilization have considerably increased growth parameters like plant leaves number at 50% flowering and plant height at harvest at Samanko. Likewise, results revealed that SWD1 produced higher biomass and grain yield, as well as higher WUE, compared to SWD2 (the second sowing date) at Samanko during both seasons. SWD1 outperformed SWD2 for biomass and grain yield, with increases of 65% and 30%, respectively. The organo-mineral fertilization F4 recorded the highest biomass and grain yield and the highest WUE compared to other fertilization types. During both cropping seasons, the hybrid Pablo produced the highest biomass (4083 and 6795 kg ha-1) and grain yield (1765 and 2740 kg ha-1), as well as the highest water WUE (3.83 and 6.15 kg ha-1 mm-1) at Samanko while the variety Soubatimi produced the highest biomass (6171 and 7085 kg ha-1) at N’Tarla. Regarding the study modelling aspects, the calibration and validation processes gave satisfying results with good agreements between simulated and observed values for phenology (anthesis and maturity dates), leaves number, aboveground biomass and grain yield. Long-term management practices’ simulations (fertilization strategies, and sowing windows) indicated that optimized nitrogen fertilizer application (mineral or organo-mineral) and early sowing window, applied alone or in combination, could substantially increase sorghum yields under current climate conditions in Mali Soudano-Sahelian region. In conclusion, this study highlights the value of crop simulation models for identifying and promoting effective management strategies, such as optimized organo-mineral fertilization and adjusted sowing dates to enhance sorghum production and bolster food security in the Soudano-Sahelian region, providing actionable insights for farmers and policymakers.