Faliagka et al. (2026) Adaptation of the VegSyst model to predict crop nutrient uptake and water needs for precise soilless crop fertigation in greenhouses

Identification

• Journal: Irrigation Science
• Year: 2026
• Date: 2026-03-26
• Authors: Sofia Faliagka, Ioannis Naounoulis, Marisa Gallardo, Nikolaos Katsoulas
• DOI: 10.1007/s00271-026-01104-6

Research Groups

• Department of Agriculture Crop Production and Rural Environment, Laboratory of Agricultural Constructions and Environmental Control, University of Thessaly, Volos, Magnesia, Greece
• Department of Agronomy, University of Almeria, Almería, Spain

Short Summary

This study adapted and validated the VegSyst model for precise soilless crop fertigation in greenhouses, integrating climate forecasts to predict water and macronutrient needs for cucumber and tomato, demonstrating reduced nutrient leaching and increased agronomic efficiency without compromising yield.

Objective

• To adapt the existing VegSyst model for soilless crop fertigation by incorporating climate forecasts and a greenhouse climate model to predict dry matter production, crop irrigation requirements, and tailored nutrient requirements.
• To validate the adapted model for dry matter production (DMP) and uptake of macronutrients (nitrogen (N), phosphorus (P), potassium (K), and calcium (Ca)) in cucumber and tomato crops under soilless greenhouse conditions.
• To support the integration of the adjusted model into a decision support system (DSS) to provide greenhouse farmers with weekly nutrient schedules, thereby mitigating water and nutrient waste.

Study Configuration

• Spatial Scale: Greenhouse facilities of the University of Thessaly (39°22’ N, 22°44’ E, 85 m altitude), Velestino, Greece. Experiments were conducted in separate greenhouse compartments, each covering a cultivated area of 183 m², with plants grown on perlite slabs.
• Temporal Scale:
  • Cucumber (Cucumis sativus cv. Columbia): Autumn-winter (1 October 2021 to 17 December 2021, 77 days) and spring-summer (4 February 2022 to 31 May 2022, 116 days).
  • Tomato (Lycopersicum esculentum cv. Protector × Ekstasis): Winter-spring (4 February 2022 to 4 July 2022, 151 days).
  • Climate forecasts were used for 7-day periods.

Methodology and Data

• Models used:
  • VegSyst model: Adapted to simulate daily water and macronutrient demands, dry matter production (DMP), and critical nutrient content.
  • Penman-Monteith (P-M) equation: A simplified version was used to estimate future crop water requirements (transpiration, ETc), incorporating an automatic calibration mechanism.
  • Greenhouse climate model: Used to estimate internal greenhouse climate conditions based on external weather forecasts and predefined control settings.
• Data sources:
  • Climate data: Short-term (7-day) external weather forecasts obtained from a web-based greenhouse Decision Support System (DSS); internal greenhouse air temperature (°C) and solar radiation (W m⁻²) measured by sensors at 1.5 m and 4.0 m above ground level, respectively.
  • Crop measurements: Destructive measurements of above-ground dry matter production (DMP) and N, P, K, Ca content in leaves, stems, and fruits (9-10 measurements per trial, n=4 per treatment). Weekly measurements of plant height (cm), number of leaves, number of fruits, and final yield (kg m⁻²). Weekly measurements of crop net photosynthesis (µmol CO₂ m⁻² s⁻¹), stomatal conductance (mmol H₂O m⁻² s⁻¹), and transpiration (mmol H₂O m⁻² s⁻¹) using an LCpro T instrument.
  • Water and nutrient solution analysis: Daily irrigation and drainage volumes (L m⁻²). Weekly samples of irrigation and drainage analyzed for NO₃⁻, NH₄⁺, K⁺, Na⁺, Ca²⁺, PO₄³⁻ concentrations (mM, µM), pH, and electrical conductivity (EC, dS m⁻¹).
  • Literature-derived coefficients: For VegSyst model (e.g., critical nutrient dilution curve coefficients, radiation use efficiency (RUE), initial and maximum fraction of intercepted photosynthetically active radiation (PAR)).

Main Results

• Model Performance:
  • Good correlation between simulated and measured cumulative DMP for tomato (R² = 0.94) and CUC-2 (R² = 0.95). Marginal performance for CUC-1 (R² = 0.89). The model generally underestimated DMP.
  • Strong correlation for simulated and measured N uptake across all scenarios (R² = 0.88 to 0.98). N simulation for tomato was underestimated but within the acceptance range; for cucumber, it was more accurate (acceptable range).
  • Adequate simulation of K uptake in CUC-1 (R² = 0.94) and CUC-2 (R² = 0.96), with acceptable statistical indices.
  • Poor performance for Ca uptake simulation across all trials (Willmott index below criteria, relative error > 25%).
  • Acceptable performance for P simulation only for tomato and CUC-2. Marginal performance for CUC-1.
• Nutrient Leaching:
  • Significant reductions in nutrient leaching were achieved with the adaptive treatment (T2) compared to the conventional treatment (T1).
  • Nitrate (NO₃⁻) leaching reduced by 22.0% (CUC-1) and 15.5% (CUC-2).
  • Phosphate (PO₄³⁻) leaching reduced by 16.8% (CUC-1), 5.7% (CUC-2), and 39.5% (tomato).
  • Potassium (K⁺) leaching reduced by 25.2% (CUC-1) and 16.6% (CUC-2).
  • Calcium (Ca²⁺) leaching reduced by 14.0% (CUC-1) and 12.1% (CUC-2).
  • For tomato, NO₃⁻, K⁺, and Ca²⁺ concentrations in T2 drainage were higher than T1, likely due to marginal/poor model performance for these nutrients in tomato.
• Crop Productivity, Water Productivity (WP), and Agronomic Efficiency (AE):
  • Total fruit yield did not differ significantly between conventional (T1) and adaptive (T2) treatments across all cropping scenarios.
  • Water productivity (WP) did not differ significantly between treatments.
  • Agronomic efficiency (AE) significantly increased in the T2 treatment:
    • CUC-1: AEN (7.0%), AEP (4.4%), AE_K (10.5%) increased.
    • CUC-2: AEN (10.1%), AEP (20.2%), AEK (10.5%), AECa (7.1%) increased.
    • Tomato: AEP (27.0%), AECa (10.1%) increased.
• Crop Morphological and Physiological Characteristics:
  • Plant height, number of leaves, and total marketable fruits per square meter showed no significant differences between treatments for both cucumber and tomato.
  • Net photosynthesis, ratio of photosynthesis to radiation, transpiration, and stomatal conductance showed no statistically significant differences between treatments for CUC-2 and tomato.

Contributions

• First adaptation and validation of the VegSyst model for soilless crop fertigation, specifically incorporating climate forecasts to predict future daily nutrient uptake and water needs.
• Demonstrated the potential of integrating climate forecasting with nutrient uptake models to provide weekly nutrient schedules, thereby reducing water and nutrient waste in soilless greenhouse systems.
• Showed significant reductions in nitrate and phosphate leaching (up to 22% and 40% respectively in cucumber trials) and increased agronomic efficiency without compromising overall productivity.
• Highlighted the model's nutrient-specific predictive reliability, identifying limitations for calcium uptake simulation under the tested conditions.

Funding

• PRECIMED project (Project Acronym/Code: “PRECIMEDPrima2018-09”, Project application number: 155331/I4/19.09.18).
• Funded by the General Secretariat for Research and Technology of the Ministry of Development and Investments of Greece under the PRIMA Programme.
• PRIMA is an Art.185 initiative supported and co-funded under Horizon 2020, the European Union’s Programme for Research and Innovation.

Citation

@article{Faliagka2026Adaptation,
  author = {Faliagka, Sofia and Naounoulis, Ioannis and Gallardo, Marisa and Katsoulas, Nikolaos},
  title = {Adaptation of the VegSyst model to predict crop nutrient uptake and water needs for precise soilless crop fertigation in greenhouses},
  journal = {Irrigation Science},
  year = {2026},
  doi = {10.1007/s00271-026-01104-6},
  url = {https://doi.org/10.1007/s00271-026-01104-6}
}