RiceSM水稻模型参数敏感性分析与适应性研究

doi:10.19788/j.issn.2096-6369.230215

摘要/Abstract

摘要：

作物模型可定量描述作物生长发育过程及其与环境因子的关系，在农业生产管理决策等方面具有重要的应用价值。模型参数调试是作物生长模拟模型进行应用前的重要步骤，且往往需要大量时间和精力进行调试，敏感性分析可以以较高的效率筛选出敏感参数，是模型本地化的重要环节，对模型的应用有重要意义。文章研究基于Morris法和EFAST法对RiceSM模型的作物参数进行了敏感性分析，筛选出输出变量中成熟期、叶面积指数、地上部生物量、产量的敏感参数，并比较分析两种方法结果的异同。结果表明，移栽至拔节阶段的基本发育系数K3、出苗到移栽阶段的基本发育系数K2、移栽到拔节阶段叶干物质的分配系数CLV1是影响RiceSM模型主要输出结果的最敏感参数，两种方法得到的敏感参数结果基本一致，但各敏感参数的重要程度略有差异。以筛选出的敏感参数为基础，基于长沙、常德两站的农气观测资料对RiceSM模型进行调参与验证。验证结果表明，早稻和晚稻叶面积指数模拟值与实测值的归一化均方根误差(NRMSE)在21.63%~47%之间，早稻和晚稻茎、叶、穗、地上部生物量和产量模拟值与实测值的NRMSE分别为4.77%~39.51%、5.46%~6.64%、3.78%~4.15%和2.78%~3.52%和9.29%~12.12%之间。调参后的模型能够较好地模拟早稻和晚稻生物量、叶面积指数的动态变化和产量形成过程。研究结果可为RiceSM模型的本地化、参数优化和推广应用提供支持。

关键词: 水稻, RiceSM模型, 敏感性分析

Abstract:

Crop model can quantitatively describe crop growth and development processes and their relationships with environmental factors, and have important applications in agricultural production management decisions and other areas. Model parameter debugging is an important step before crop growth simulation models are applied, and often requires a lot of time and effort for debugging. Sensitivity analysis can screen out sensitive parameters with high efficiency, and is an important part of model localization, which is of great significance for model application. Sensitivity analysis was conducted on the crop parameters of the RiceSM model based on the Morris method and EFAST method to screen out the sensitive parameters of maturity, leaf area index, total biomass and yield among the output variables, and to compare and analyze the similarities and differences between the results of the two methods. The results showed that basic development factor from transplanting to jointing stage K3, basic development factor from seeding to transplanting stage K2 and dry matter distribution coefficients of leaf from transplanting to jointing stage CLV1 were the most sensitive parameters affecting the main output results of RiceSM model, and the results of the sensitive parameters obtained by the two methods were generally consistent, but the importance of each sensitive parameter differed slightly. The validation results showed that the normalized root mean square error (NRMSE) of simulated and measured values of early and late rice leaf area index ranged from 21.63% to 47%, and the NRMSEs of simulated and measured values of stem, leaf, spike, aboveground biomass and yield of early and late rice ranged from 4.77% to 39.51%, 5.46% to 6.64%, 3.78% to 4. 15% and 2.78% to 3.52% and between 9.29% and 12.12% respectively. The model was able to better simulate the dynamics of bio-mass, leaf area index and yield formation in early and late rice. The results of the study provide a reference for the localization of the model.

Key words: rice, RiceSM model, sensitivity analysis

王雪莹, 陈先冠, 汤顺杰, 冯利平. RiceSM水稻模型参数敏感性分析与适应性研究[J]. 农业大数据学报, 2023, 5(2): 97-108.

WANG XueYing, CHEN XianGuan, TANG ShunJie, FENG LiPing. Sensitivity Analysis and Adaptability Evaluation of RiceSM Model[J]. Journal of Agricultural Big Data, 2023, 5(2): 97-108.

图/表 9

表1

图1

图2

表2

表3

图3

图4

图5

表4

参考文献 23

[1]	汪欢欢. 未来气候条件下水稻生产力模拟研究[D]. 南京: 南京农业大学, 2014.
	Wang H H. Research on rice productivity simulation under future climate conditions[D]. Nanjing: Nanjing Agricultural University, 2014. (in Chinese)
[2]	殷新佑, 戚昌瀚. 水稻生长日历模拟模型及其应用研究[J]. 作物学报, 1994(3): 339.
	Yin X Y, Qi C H. Rice growth calendar model and its application study[J]. Acta Agronomica Sinica, 1994(3): 339. (in Chinese)
[3]	高亮之, 金之庆. RCSODS-水稻栽培计算机模拟优化决策系统[J]. 计算机农业应用, 1993(3): 14-20.
	Gao L Z, Jin Z Q. RCSODS-Computer simulation and optimization of rice cultivation decision system[J]. Computer Applications in Agriculture, 1993(3): 14-20. (in Chinese)
[4]	曹宏鑫, 金之庆, 石春林, 等. 中国作物模型系列的研究与应用[J]. 农业网络信息, 2006(5): 45-48, 51.
	Cao H X, Jin Z Q, Shi C L, et al. Researches and application of crop model series in China[J]. Agriculture Network Information, 2006(5): 45-48, 51. (in Chinese)
[5]	王亚莉, 贺立源. 作物生长模拟模型研究和应用综述[J]. 华中农业大学学报, 2005, 24(5): 529-535.
	Wang Y L, He L Y. A review on the research and application of crop simulation model[J]. Journal of Huazhong Agricultural University, 2005, 24(5): 529-535. (in Chinese)
[6]	叶芳毅, 李忠武, 李裕元, 等. 水稻生长模型发展及应用研究综述[J]. 安徽农业科学, 2009, 37(1): 85-89.
	Ye F Y, Li Z W, Li Y Y, et al. Development and application of rice productivity models[J]. Journal of Anhui Agricultural Sciences, 2009, 37(1): 85-89. (in Chinese)
[7]	谭君位. 作物模型参数敏感性和不确定性分析方法研究[D]. 武汉: 武汉大学, 2017.
	Tan J W. Study on parameter sensitivity and model uncertainty analysis of crop model[D]. Wuhan: Wuhan University, 2017. (in Chinese)
[8]	王连喜, 张阳, 李琪, 等. 作物模型参数敏感性分析现状与展望[J]. 气象科技, 2018, 46(2): 382-389.
	Wang L X, Zang Y, Li Q, et al. Current status and prospects of sensitivity analysis of crop model parameter[J]. Meteorological Science and Technology, 2018, 46(2): 382-389. (in Chinese)
[9]	蔡毅, 邢岩, 胡丹. 敏感性分析综述[J]. 北京师范大学学报(自然科学版), 2008, 44(1): 9-16.
	Cai Y, Xing Y, Hu D. On sensitivity analysis[J]. Journal of Beijing Normal University(Natural Science), 2008, 44(1): 9-16. (in Chinese)
[10]	邢会敏, 相诗尧, 徐新刚, 等. 基于EFAST方法的AquaCrop作物模型参数全局敏感性分析[J]. 中国农业科学, 2017, 50(1): 64-76. doi: 10.3864/j.issn.0578-1752.2017.01.006
	Xing H M, Xiang S Y, Xu X G, et al. Global sensitivity analysis of AquaCrop crop model parameters based on EFAST method[J]. Scientia Agricultura Sinica, 2017, 50(1): 64-76. (in Chinese) doi: 10.3864/j.issn.0578-1752.2017.01.006
[11]	谢松涯, 张宝忠. 基于全局敏感性分析的WOFOST模型参数优化[J]. 中国农村水利水电, 2018, 434(12): 29-34.
	Xie S Y, Zhang B Z. Optimization of WOFOST model parameters based on global sensitivity analysis[J]. China Rural Water and Hydropower, 2018, 434(12): 29-34. (in Chinese)
[12]	宋明丹, 冯浩, 李正鹏, 等. 基于Morris和EFAST的CE-RES- Wheat模型敏感性分析[J]. 农业机械学报, 2014, 45(10): 124-131, 166.
	Song M D, Feng H, Li Z P, et al. Global sensitivity analyses of DSSAT-CERES-Wheat model using Morris and EFAST methods[J]. Journal of Agricultural Machinery, 2014, 45(10): 124-131, 166. (in Chinese)
[13]	胡钧铭, 江立庚, 吕永成. 水稻模拟模型研究与发展趋势[J]. 农业网络信息, 2007,(4): 7-10, 17.
	Hu J M, Jiang L G, Lv Y C. Research and development tendency of simulation and model in rice[J]. Agriculture Network Information, 2007,(4): 7-10, 17. (in Chinese)
[14]	高蓓, 胡凝, 高茂盛. 水稻ORYZA2000模型在陕西省的适应性评价[J]. 西南师范大学学报(自然科学版), 2016, 41(5): 74-80.
	Gao B, Hu N, Gao M S. On adaptability evaluation of ORY -ZA2000 in Shanxi province[J]. Journal of Southwest China Normal University(Natural Science), 2016, 41(5): 74-80. (in Chinese)
[15]	姚凤梅, 许吟隆, 冯强, 等. CERES-Rice模型在中国主要水稻生态区的模拟及其检验[J]. 作物学报, 2005, 31(5): 545-550.
	Yao F M, Xu Y L, Feng Q, et al. Simulation and validation of CERES-Rice model in main rice ecological zones in China[J]. Acta Agronomica Sinica, 2005, 31(5): 545-550. (in Chinese)
[16]	郭建茂, 王星宇, 刘慎彬, 等. 基于稻田实测温度的水稻模型ORYZA2000应用[J]. 中国农业气象, 2020, 41(4): 211-221.
	Guo J M, Wang X Y, Liu S B, et a1. Application of rice model ORYZA2000 based on measured temperature in rice fields[J]. Chinese Journal of Agrometeorology, 2020, 41(4): 211-221. (in Chinese)
[17]	汤顺杰. 水稻生长模拟模型系统(RiceSM)软件设计[D]. 北京: 中国农业大学, 2022.
	Tang S J. Rice growth simulation model system (RiceSM) software design[D]. Beijing: China Agricultural University, 2022. (in Chinese)
[18]	Morris M. Factorial Sampling plans for preliminary computational experiments[J]. Technometrics, 1991, 33(2): 161-174. doi: 10.1080/00401706.1991.10484804
[19]	Campolongo F, Cariboni J, Saltelli A. An effective screening design for sensitivity analysis of large models[J]. Environmental Modelling & Software, 2007, 22(10): 1509-1518.
[20]	史鑫蕊, 梁浩, 周丰, 等. 稻田土壤—作物系统模型参数敏感性分析与模型验证[J]. 农业机械学报, 2020, 51(5): 252-262, 271.
	Shi X H, Liang H, Zhou F, et al. Sensitivity analysis and parameter estimation for soil-rice system model[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(5): 252-262, 271. (in Chinese)
[21]	Saltelli A, Tarantola S, Campolongo F, Ratto M. Sensitivity analysis in practice: A guide to assessing scientific models[M]. John Wiley and Sons, 2004: 20-78.
[22]	孟怡君, 邱小雷, 刘蕾蕾, 等. RiceGrow 水稻模型品种参数敏感性分析[J]. 农业大数据学报, 2021, 3(3): 23-32.
	Meng Y J, Qiu X L, Liu L L, et al. Sensitivity analysis of genetic parameters of RiceGrow model[J]. Journal of Agricultural Big Data, 2021, 3(3): 23-32. (in Chinese)
[23]	曹秀霞, 安开忠, 蔡伟, 等. CERES-Rice模型在江汉平原的验证与适应性评价[J]. 中国农业气象, 2013, 34(4): 447-454.
	Cao X X, An K Z, Cai W, et al. Validation and adaptability evaluation of CERES-Rice model in the Jianghan plain[J]. Chinese Journal of Agrometeorology, 2013, 34(4): 447-454. (in Chinese)

参数 Parameters	定义 Definition	初始值 Initial values	变化范围 Range of variation
K1	播种到出苗阶段基本发育系数 Basic development factor from sowing to seedling stage	-1.5	-1.95~- 1.05
K2	出苗到移栽阶段基本发育系数 Basic development factor from seedling to transplanting stage	-3.15	-4.095~2.205
K3	移栽到拔节阶段基本发育系数 Basic development factor from transplanting to jointing stage	-3.0	-3.9~-2.1
K4	拔节到孕穗阶段基本发育系数 Basic development coefficient from jointing to booting stage	-1.5	-1.95~-1.05
K5	孕穗到抽穗阶段基本发育系数 Basic development coefficient from booting to heading stage	-1.25	-1.625~-0.875
K6	抽穗到黄熟阶段基本发育系数 Basic development coefficient from heading to yellow ripening stage	-3.0	-3.9~-2.1
K7	黄熟到成熟阶段基本发育系数 Basic development coefficient from yellow ripening to ripening stage	-1.98	-2.574~-1.386
Pmax	最大光合速率 Maximum photosynthetic rate	4.0	2.8~5.2
P1	播种到出苗阶段温度反应特性遗传系数 Temperature genetic coefficient from sowing to seedling stage	0.9	0.63~1.17
Q1	播种到出苗阶段光周期反应特性遗传系数 Light genetic coefficient from sowing to seedling stage	1.157	0.8099~1.5041
CLG1	移栽到拔节阶段根干物质的分配系数 Dry matter distribution coefficients of root from transplanting to jointing stage	0.58	0.406~0.754
CLT1	移栽到拔节阶段茎干物质的分配系数 Dry matter distribution coefficients of stem from transplanting to jointing stage	0.02	0.014~0.026
CLV1	移栽到拔节阶段叶干物质的分配系数 Dry matter distribution coefficients of leaf from transplanting to jointing stage	0.4	0.28~0.52

输出变量 Output variables	敏感性分析方法 Sensitivity analysis methods	敏感参数 Sensitive parameters
成熟期 Ripening stage	Morris	K2、K3、K6、Q2、K7
成熟期 Ripening stage	EFAST	K3、K2、P2、Q2、K7、K6
叶面积指数 Leaf area index	Morris	K3、CLV1、K2、K6、Pmax、Q3
叶面积指数 Leaf area index	EFAST	K3、CLV1、K2、K6、Pmax、Q3
地上部生物量 Above-ground biomass	Morris	K3、K6、CLV1、K2、Pmax、Q3、K4
地上部生物量 Above-ground biomass	EFAST	K3、CLV1、K2、K6、Pmax、Q3、K4
产量 Yield	Morris	K3、K6、CLV1、K2、Pmax、Q3、K4、K7
产量 Yield	EFAST	K3、K2、CLV1、K6、Q2、K7、Pmax、Q3、K4

站点 Site	类型 Type	K2	K3	K4	K6	K7	Q2	Q3	Pmax	CLV1
长沙 Changsha	早稻 Early rice	-2.5	-3.05	-1.5	-3.0	-1.98	0.50	0.52	4.0	0.42
长沙 Changsha	晚稻 Late rice	-3.0	-3.00	-1.5	-3.0	-1.50	0.49	0.51	4.0	0.43
常德 Changde	早稻 Early rice	-2.5	-3.01	-1.5	-3.0	-1.50	0.49	0.50	4.0	0.42
常德 Changde	晚稻 Late rice	-3.0	-3.00	-1.5	-3.0	-2.15	0.52	0.50	4.0	0.40

	项目 Project	N	Xobs±SD	Xsim±SD	P(t * )	α	β	R²	RMSE	NRMSE (%)
早稻 Early rice	叶面积指数 Leaf area index	82	2.5± 1.9	2.3± 1.6	0.43	0.83	0.21	0.94	0.54	21.63
	茎生物量Stem biomass (kg/hm²)	61	1 910.6± 1 193.4	1 536.2± 1 294.6	0.1	0.94	-250.24	0.74	754.88	39.51
	叶生物量Leaf biomass (kg/hm²)	61	1 210.4±733.2	994.9± 812. 1	0. 13	0.78	44.74	0.50	80.34	6.64
	穗生物量 Panicle biomass (kg/hm²)	61	2 546.3±2 836.0	2 276.5±2 674.7	0.59	0.91	-31.27	0.92	105.62	4. 15
	地上部生物量 Above-ground biomass (kg/hm²)	61	5 663.0±4 064.9	4 801.1±4 364.0	0.26	1.02	-1 001.39	0.91	199.29	3.52
	产量 Yield (kg/hm²)	12	4 979.7± 1 029.0	5 196.2±999.0	0.61	0.81	1 173.45	0.69	603.50	12. 12
晚稻 Late rice	叶面积指数 Leaf area index	82	3.4±2.3	2.8±2.0	0.10	0.66	0.59	0.60	1.59	47.00
	茎生物量Stem biomass (kg/hm²)	60	2 413.8± 1 460.6	2 048.7± 1 504.6	0.18	0.87	-58.01	0.72	115. 14	4.77
	叶生物量Leaf biomass (kg/hm²)	60	1 534.0± 860.7	1 396.6±952.4	0.41	0.84	113.75	0.57	83.75	5.46
	穗生物量 Panicle biomass (kg/hm²)	60	3 080.0±3 320.2	3 336. 1±3 605.0	0.67	1.05	86.92	0.94	116.50	3.78
	地上部生物量 Above-ground biomass (kg/hm²)	60	7 237.8±4 937.2	6 753. 1±5 449.2	0.61	1.06	-944.36	0.93	200.87	2.78
	产量 Yield (kg/hm²)	12	6139.9±342.0	6 432.6±582.4	0. 15	0.83	1 322.21	0.24	570.43	9.29
注：N为样本数，Xobs为实测平均值、Xsim为模拟平均值、SD为标准差、P(t)为t检验显著性、α、β、R²分别为模拟值与实测值线性相关的斜率、截距、线性相关系数、RMSE 为均方根误差、NRMSE为归一化均方根误差。P(t)中，*表示模拟值和实测值无显著性差异的可信度为95%。