AQUACROP model-based assessment of deficit irrigation strategies for cotton production

Introduction

Water allocation for agriculture is shrinking across the former Soviet cotton belt, where river diversions that once supported 2.5 million hectares of irrigated cotton are now contested by upstream hydropower demands [1,2] . Farmers need tools that help them decide how much water to cut and when, without gambling on an entire season's income. Crop simulation models offer a way to test irrigation scenarios virtually before committing real water to real fields [3] . AQUACROP, developed by the FAO, is a water-driven model that simulates yield response to water with relatively few input parameters compared with models like DSSAT or APSIM [4,5] . It tracks canopy cover, transpiration, biomass, and harvest index on a daily step, making it well suited to deficit irrigation analysis. The model has been validated for cotton in semi-arid zones of Turkey, Pakistan, and Uzbekistan, but not yet for the continental climate conditions of central Russia [6,7,8] . This research calibrated AQUACROP for cotton using three seasons of field data from Moscow State University, validated it against independent observations, and then ran four deficit irrigation scenarios to identify the strategy that maximises water productivity without unacceptable yield loss [9,10,11] .

Simulation parameters and model configuration AQUACROP version 7.0 was configured with local soil hydraulic properties derived from laboratory analysis of the Moscow research station's chernozem soil: saturated hydraulic conductivity 14.2 mm h⁻¹, field capacity 34.8%, wilting point 17.3%, and saturation 48.1%. Crop parameters were initialised from the FAO's default cotton file and adjusted during calibration: canopy growth coefficient (CGC = 0.0123 d⁻¹), canopy decline coefficient (CDC = 0.0071 d⁻¹), reference harvest index (HI₀ = 0.42), and maximum rooting depth (1.2 m). Climate inputs included daily Tmax, Tmin, rainfall, ETo, and solar radiation from the university's meteorological station [4] .

Climatic conditions

The research covered three growing seasons (May-October 2020-2022) at the Lomonosov Moscow State University agronomy station (55.70° N, 37.53° E). Mean growing-season temperature ranged from 18.4 °C (2020) to 19.7 °C (2022). Cumulative ETo averaged 612 mm. Seasonal rainfall varied from 187 mm (2021) to 263 mm (2020). The 2021 season was the driest and served as the primary calibration year; 2020 and 2022 were used for validation.

Material and Methods

Material

Cotton variety Bukhara-6 was sown in mid-May each year at 80,000 plants ha⁻¹ on a deep chernozem soil at the university's field station. Plots (10 m × 8 m) received four irrigation treatments via surface drip: full (100% ETc), mild deficit (80% ETc), moderate deficit (65% ETc), and severe deficit (50% ETc), in a randomised complete block design with four replications [9] . Soil moisture was monitored with TDR probes at 15, 30, and 60 cm. Seed cotton yield was hand-harvested from 6 m² per plot.

Methods

Calibration followed a stepwise approach: canopy cover was fitted first against weekly LAI measurements, then biomass against destructive harvests at 30-day intervals, and finally yield against end-of-season data [4] . Goodness-of-fit was assessed by R², RMSE, normalised RMSE (NRMSE), and the index of agreement (d). After validation, the model was run for all four deficit scenarios using 2020-2022 weather data. WUE was calculated as simulated yield divided by total water input (irrigation + effective rainfall) [11,12] .

Experimental design

The randomised complete block design with four irrigation levels and four blocks yielded 16 plots per season. For modelling purposes, plot-level data were averaged within each treatment to provide treatment-mean time series for calibration and validation. Model simulations were run as single-point, single-season runs without spatial variability. Moderate deficit irrigation saved 31% of water while reducing yield by 19.1%. WUE increased from 0.73 under full irrigation to 0.86 kg m⁻³ under both moderate and severe deficit, but severe deficit sacrificed 37.6% of yield an unacceptable loss for most commercial growers.

Results

Comprehensive interpretation

The model consistently showed that moderate deficit irrigation occupies a sweet spot: WUE plateaus at 0.86 kg m⁻³ for both moderate and severe strategies, but the yield cost of moving from moderate to severe is an additional 18.5 percentage points far steeper than the 11.5-point jump from mild to moderate. This non-linear response reflects cotton's sensitivity to water stress during boll development, which falls under severe deficit but is largely maintained under moderate scheduling.

Discussion

The NRMSE of 7.4% places AQUACROP's performance for Moscow-region cotton within the acceptable range reported for semi-arid cotton in Uzbekistan (6.8%) and Turkey (8.2%) [6,7] . The slightly higher error in 2020 likely reflects that season's above-average rainfall, which introduced drainage uncertainty not fully captured by the model's simple water balance [4,12] . The moderate deficit strategy's 31% water saving aligns with the 25-35% range reported by AQUACROP simulations for cotton in the Aral Sea basin [8] . For Russian cotton growers, who face both water scarcity and a short growing season, this strategy offers a pragmatic compromise. The 19.1% yield penalty is partially offset by lower pumping costs and reduced waterlogging risk on heavy soils [13,14] . One limitation is that AQUACROP doesn't simulate pest and disease dynamics, which can intensify under water stress. Cotton bollworm pressure, for example, tends to increase in deficitirrigated fields because stressed plants produce fewer defensive compounds [15] . Coupling AQUACROP with a pest risk module would give a more complete picture of deficit irrigation's net economic effect. Multi-decadal climate scenarios should also be tested to assess strategy robustness under projected warming [16] .

Limitations

AQUACROP's simplified soil water balance doesn't account for preferential flow in cracking soils, and its lack of a nitrogen module means nutrient interactions with water stress are ignored [5, 17] . The three-season calibration dataset, while adequate, captures limited inter-annual climate variability. Extending the analysis with 30-year synthetic weather series would strengthen the robustness of strategy recommendations [16,18] .

Conclusion

References

1. Bekchanov M, Lamers JP, Martius C. Pros and cons of adopting water-wise approaches in the lower reaches of the Amu Darya. Water. 2010;2(2):200-216.
2. Dukhovny VA, de Schutter JL. Water in Central Asia: past, present, future. CRC Press. 2011;1:1-408.
3. Jones JW, Hoogenboom G, Porter CH, Boote KJ, Batchelor WD, Hunt LA. The DSSAT cropping system model. European Journal of Agronomy. 2003;18(3-4):235-265.
4. Steduto P, Hsiao TC, Raes D, Fereres E. AquaCrop the FAO crop model to simulate yield response to water. Agronomy Journal. 2009;101(3):426-437.
5. Raes D, Steduto P, Hsiao TC, Fereres E. AquaCrop: the FAO crop model to simulate yield response to water. Reference Manual. 2018;1:1-302.
6. Farahani HJ, Izzi G, Oweis TY. Parameterization and evaluation of the AquaCrop model for full and deficit irrigated cotton. Agronomy Journal. 2009;101(3):469-476.
7. Hussein F, Janat M, Yakoub A. Simulating cotton yield response to deficit irrigation with the FAO AquaCrop model. Spanish Journal of Agricultural Research. 2011;9(4):1319-1330.
8. Garcia-Vila M, Fereres E. Combining the simulation crop model AquaCrop with an economic model for the optimization of irrigation management at farm level. European Journal of Agronomy. 2012;36(1):21-31.
9. Doorenbos J, Kassam AH. Yield response to water. FAO Irrigation and Drainage Paper 33. 1979;1:1-193.
10. Allen RG, Pereira LS, Raes D, Smith M. Crop evapotranspiration: guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. 1998;1:1-300.
11. Geerts S, Raes D. Deficit irrigation as an on-farm strategy to maximize crop water productivity in dry areas. Agricultural Water Management. 2009;96(9):1275-1284.
12. Hsiao TC, Heng L, Steduto P, Rojas-Lara B, Raes D, Fereres E. AquaCrop the FAO crop model to simulate yield response to water: parameterization and testing for maize. Agronomy Journal. 2009;101(3):448-459.
13. Fereres E, Soriano MA. Deficit irrigation for reducing agricultural water use. Journal of Experimental Botany. 2007;58(2):147-159.
14. Pereira LS, Oweis T, Zairi A. Irrigation management under water scarcity. Agricultural Water Management. 2002;57(3):175-206.
15. Showler AT. Effects of water deficit stress, shade, weed competition and kaolin particle film on selected foliar free amino acid accumulations in cotton. Journal of Chemical Ecology. 2002;28(3):631-651.
16. Rosenzweig C, Elliott J, Deryng D, Ruane AC, Müller C, Arneth A, Boote KJ. Assessing agricultural risks of climate change in the 21st century. Proceedings of the National Academy of Sciences. 2014;111(9):3268-3273.
17. Vanuytrecht E, Raes D, Steduto P, Hsiao TC, Fereres E, Heng LK, Garcia Vila M. AquaCrop: FAO's crop water productivity and yield response model. Environmental Modelling and Software. 2014;62:351-360.
18. DeTar WR. Yield and growth characteristics for cotton under various irrigation regimes on sandy soil. Agricultural Water Management. 2008;95(1):69-76.
19. Dagdelen N, Yilmaz E, Sezgin F, Gurbuz T. Water-yield relation and water use efficiency of cotton in western Turkey. Agricultural Water Management. 2006;82(1-2):6385.
20. Oweis T, Hachum A. Water harvesting and supplemental irrigation for improved water productivity of dry farming systems. Agricultural Water Management. 2006;80(1-3):5773.
21. Kirda C. Deficit irrigation scheduling based on plant growth stages showing water stress tolerance. FAO Water Reports. 2002;22:3-10.