The objective of this study was to investigate the effect of soil tillage and crop type on soil respiration (Rs) in a typical Central European agricultural site characterized by crop rotation. The weekly Rs and supporting environmental variables were measured under different crop types (including winter and summer crops) for 5 consecutive years under mouldboard ploughing (MP) and no-tillage (NT) treatments. The long-term mean Rs was 0.093 mg CO₂ m⁻² s⁻¹ in NT versus 0.086 mg CO₂ m⁻² s⁻¹ in MP. Soil respiration was significantly higher in NT treatment regardless of crop type and weather conditions in most of the study period. The difference between the treatments was larger for summer crops than for winter crops. The observed differences were more pronounced during the growing season, which cannot be explained solely by the observed plant production related data. Soil temperature did not differ in the two contrasting treatments, but soil water content was significantly higher in the NT treatment which might contribute to the observed differences in Rs. Six models were tested to simulate Rs in MP and NT treatment based on the observed environmental variables (soil temperature and soil water content). Normalized Difference Vegetation Index (NDVI) was also included as a predictor in four models to serve as a proxy for root activity. Inclusion of NDVI clearly improved the performance of the Rs models when the entire dataset was simulated including vegetated and non-growing season data. Performance and structure of the proposed Rs models varied between crop types and also between treatments (MP and NT). The preferred models explained 42% and 44% of the observed Rs variance in MP and NT, respectively. We provided explicit Rs model equations for the entire time series and also for sunflower, maize and the non-growing season period. The results suggest that there is added value in the construction of crop-specific Rs models, and also treatment specific models. Methodology related uncertainty of the Rs observations calls for longer datasets and improved modeling approaches including Bayesian and probabilistic methods.

本研究的目的是调查土壤耕作和作物类型对以轮作为特征的典型中欧农业地点土壤呼吸 (Rs) 的影响。在犁板犁 (MP) 和免耕 (NT) 处理下，连续 5 年在不同作物类型（包括冬季和夏季作物）下测量每周 Rs 和支持环境变量。NT 中的长期平均 Rs 为 0.093 mg CO ₂ m ^-2 s ^-1与 0.086 mg CO ₂ m ^-2 s ^-1在议员。在大部分研究期间，无论作物类型和天气条件如何，NT 处理的土壤呼吸都显着更高。夏季作物处理之间的差异大于冬季作物。观察到的差异在生长季节更为明显，这不能仅用观察到的植物生产相关数据来解释。两种对比处理的土壤温度没有差异，但 NT 处理的土壤含水量显着更高，这可能导致观察到的 Rs 差异。基于观察到的环境变量（土壤温度和土壤含水量），测试了六个模型以模拟 MP 和 NT 处理中的 Rs。归一化差异植被指数 (NDVI) 也被包括在四个模型中作为预测因子，作为根活动的代理。当模拟整个数据集（包括植被和非生长季节数据）时，包含 NDVI 明显提高了 Rs 模型的性能。所提出的 Rs 模型的性能和结构因作物类型和处理（MP 和 NT）而异。首选模型分别解释了 MP 和 NT 中观察到的 Rs 方差的 42% 和 44%。我们为整个时间序列以及向日葵、玉米和非生长季节提供了明确的 Rs 模型方程。结果表明，构建特定作物的 Rs 模型以及处理特定模型具有附加值。Rs 观测的方法学相关不确定性需要更长的数据集和改进的建模方法，包括贝叶斯和概率方法。