品种参数调试是利用作物生长模型进行模拟前的重要步骤，其调试往往花费大量时间和精力，敏感性分析可以帮助识别敏感参数，提高调试效率。本研究针对水稻生长模型RiceGrow，运用SimLab和MATLAB软件，采用EFAST法对水稻品种参数进行敏感性分析，得出不同地区和不同气候情景下（1981-2015年的历史气象数据和全球未来增温2.0℃气候情景）该模型的参数敏感性，并通过TDCC（Top-Down-Coefficient of Concordance）系数计算敏感性排序一致性。结果表明，影响开花期和总干物质量的最敏感参数为最适温度（OT，Optimum Temperature），其次为温度敏感性（TS，Temperature Sensitivity）、光周期敏感性（PS，Photoperiod Sensitivity）、基本早熟性（IE，Intrinsic Earliness），对成熟期和全生育期的最敏感参数为OT，TS、IE、PS、基本灌浆因子（BFF，Basic Filling Factor）也是敏感参数，影响产量的敏感参数主要为最大光合速率（AMX，Maximum CO₂ assimilation rate）、比叶面积（SLA，Specific Leaf Area）、收获指数（HI，Harvest Index），其次包括IE、TS、BFF、OT、PS；各个地区和不同气候情景下敏感参数较为一致但敏感性排序差异较大，增温气候情景下的多数参数敏感指数略有增加，少数略有减小；不同气候情景下的参数敏感性变化较小，不同地区之间的变化较大。在对生育期和总干物质量输出变量进行调参时，需要重点调试OT；在低温高纬度的地区需重点调试和温度、光周期及光合有关的参数；在对产量进行调参时，需要重点关注AMX、HI、SLA。LAI相对生长速率和消光系数不敏感，可在参数调试中忽略，也可在模型中剔除进行模型简化。研究结果将为作物模型的本地化、提高参数估计效率提供支持。

Genetic parameter calibration is an important step before applying the crop growth model, which often calls for a lot of time and effort. Sensitivity analysis can help to identify sensitive parameters, improve calibration efficiency, and simplify the model. Using Simlab and Matlab software, this study analyzed the sensitivity of rice genetic parameters of RiceGrow model by EFAST method and obtained the parameter sensitivity of the model in different regions and under different climate scenarios (historical meteorological data from 1981 to 2015 and global future warming 2.0℃ climate scenarios). The TDCC (Top-Downward-Coefficient of Concordance) coefficient was used to calculate the sensitivity ranking consistency. The results showed that Optimum Temperature (OT) was the most sensitive parameter affecting flowering period and total dry matter, followed by Temperature Sensitivity (TS), Photoperiod Sensitivity (PS) and Intrinsic Earliness (IE). OT was the most sensitive parameter affecting maturity period and the whole growth period. TS, IE, PS and Basic Filling Factor (BFF) were also sensitive parameters. The sensitive parameters affecting yield are mainly maximum CO₂ assimilation rate (AMX), Specific Leaf Area (SLA) and Harvest Index (HI), followed by IE, TS, BFF, OT and PS. The sensitivity parameters in all regions and under different climate scenarios are relatively consistent, but the sensitivity ordering varies greatly. The sensitivity indexes of most parameters under warming climate scenarios slightly increase, while a few slightly decrease. The variation of parameter sensitivity under different climate scenarios is small, while which among different regions is large. When calibrating the model for phenology and dry matter, OT is the most sensitivity. In areas with low temperature and high latitude, the parameters related to temperature, photoperiod and photosynthesis should be focused. When calibrating the parameters of the yield, we need to focus on AMX, HI, SLA. Relative growth rate of LAI is not sensitive, so it can be ignored in parameter calibration, and can also be eliminated from the model to simplify the model. The results would be used to localize crop model and provide a way to improve the efficiency of parameter calibration.