Soil-crop models are widely used as valuable tools to assess the combined effects of cropping practices, soil management and climate on the agro-environmental indicators. They provide a wide range of predictive information that are useful to design and evaluate innovative cropping systems. However, intercropping modeling is still under development, especially for grain legumes-based intercropping system. We performed here the first calibration of the STICS (v 9.2) model on chickpea grown under contrasting nitrogen (N) levels during two copping seasons (2018/2019 and 2019/2020). This calibration allowed us to simulate a wide range of agronomic scenarios (climate, N-fertilization and cropping system) to optimize intercrops (durum wheat-chickpea) management. 37 parameters were estimated by using a sequential optimization method. Our results showed that STICS performs well in predicting Leaf Area Index (LAI), above ground biomass (AGB) and N uptake (AGPN) for both intercropped and sole cropped species, with satisfactory model efficiency (EF ranged from 0.62 to 0.93). In addition, grain yield was correctly predicted by the model with small error (NRMSE≤13%) especially for wheat crop (EF≥0.50), while it was less correctly predicted for chickpea crop (EF≤0.24 and NRMSE≤21%). STICS predicted well root depth under the conditions of our field study (EF ≥ 0.65 and NRMSE ≤ 37%). For soil outputs variables, the model simulated adequately soil water content with a satisfactory model efficiency (EF ≥ 0.65) and low relative error (NRMSE ≤. 8.8%) especially for sole cropped and intercropped chickpea. The soil N stocks were less adequately predicted (EF ≤ 0.28) with high relative error (NRMSE ≥ 56%) in sole cropping system, while it was moderately adequately predicted (EF ≤ 0.44) in intercropping. Under the two contrasted years and N-application conditions of this study, the temporal dynamic was well reproduced by the model for both plant and soil outputs with low simulation errors. RMSE values were lesser than 0.6 m²m^-2 (9%), 0.2 t ha^-1 (14%) and 30 kg ha^-1 (12%), respectively for LAI, grain yield and AGPN of sole cropped chickpea. The dynamic of soil water content was also well reproduced among all N-application rate and during the two cropping year, with RMSE equal to 27 mm (<10%). The present work provides the first calibration for chickpea sole crop and an evaluation for durum wheat-chickpea intercrops, which will allow to use the STICS model to simulate scenarios of innovative cropping practices based on crop diversification (i.e. grain legumes and cereals) and N-fertilization management.

土壤作物模型被广泛用作评估种植实践、土壤管理和气候对农业环境指标的综合影响的宝贵工具。它们提供了广泛的预测信息，可用于设计和评估创新的种植系统。然而，间作建模仍在开发中，尤其是基于谷物豆类的间作系统。我们在此对两个种植季节（2018/2019 和 2019/2020）在对比氮 (N) 水平下生长的鹰嘴豆进行 STICS (v 9.2) 模型的首次校准。这种校准使我们能够模拟各种农艺情景（气候、施氮肥和种植系统）以优化间作（硬粒小麦-鹰嘴豆）管理。使用顺序优化方法估计了 37 个参数。我们的结果表明，STICS 在预测间作和单作物种的叶面积指数 (LAI)、地上生物量 (AGB) 和氮吸收 (AGPN) 方面表现良好，模型效率令人满意（EF 范围为 0.62 至 0.93）。此外，该模型对小麦作物（EF≥0.50）的谷物产量预测正确，误差较小（NRMSE≤13%），而对鹰嘴豆作物（EF≤0.24，NRMSE≤21%）的预测精度较低。STICS 在我们的实地研究条件下（EF ≥ 0.65 和 NRMSE ≤ 37%）预测了井根深度。对于土壤输出变量，该模型以令人满意的模型效率 (EF ≥ 0.65) 和低相对误差 (NRMSE ≤. 8.8%) 充分模拟了土壤含水量，特别是对于单作和间作鹰嘴豆。土壤氮库预测不够充分（EF ≤ 0. 28) 在单作系统中具有较高的相对误差 (NRMSE ≥ 56%)，而在间作中它被适度充分预测 (EF ≤ 0.44)。在本研究的两个对比年份和施氮条件下，模型很好地再现了植物和土壤输出的时间动态，模拟误差很小。RMSE 值小于 0.6 m² m ^-2 (9%)、0.2 t ha ^-1 (14%) 和 30 kg ha ^-1 (12%) 分别用于单季鹰嘴豆的 LAI、谷物产量和 AGPN。土壤含水量的动态在所有施氮量和两季期间也得到了很好的再现，RMSE 等于 27 mm (<10%)。目前的工作提供了鹰嘴豆单一作物的首次校准和硬质小麦-鹰嘴豆间作的评估，这将允许使用 STICS 模型模拟基于作物多样化（即谷物豆类和谷物）和 N-施肥管理。