高粱产量预测模型研究——以屯留区为例

doi:10.11924/j.issn.1000-6850.casb20200100030

摘要/Abstract

摘要：

研究屯留区高粱产量对气候环境条件的响应,探讨气候因子与高粱产量的关系,为屯留区高粱高效生产提供理论和技术参考。首先,通过粒子群优化算法对指数平滑系数进行优化,进而运用三次指数平滑法将高粱产量分解成趋势产量与气象产量2部分。其次,采用灰色绝对关联法得出对高粱气象产量影响较大的气候因子。最后,对屯留区2004—2017年高粱实际产量进行多元线性回归模型拟合,并利用2018年高粱实际产量对其模型进行验证。结果表明：影响高粱产量的主要气候因子有5月降雨量、5月光照时长、6月平均相对湿度、6月降雨量、7月降雨量、7月光照时长、8月降水量、9月光照时长。2018年高粱产量验证多元线性回归模型的精确度为81.06%,表明该模型能相对准确地拟合屯留区的高粱产量,可为高粱产量的准确预测提供了方法借鉴。

关键词: 气候, 产量, 三次指数平滑法, 粒子群优化算法, 灰色绝对关联, 多元线性回归

Abstract:

To study the response of sorghum production to climate and environmental conditions in Tunliu District we explore the relationship between climatic factors and sorghum yield to provide theoretical and technical references for efficient production of sorghum. Firstly, the exponential smooth factor coefficient was optimized by the particle swarm optimization, and then the sorghum production was decomposed into two parts: trend production and meteorological production by the exponential smoothing for three times. Secondly, the absolute grey relational analysis was used to obtain the climate factors which had great influence on the meteorological production of sorghum, Finally, the multivariable linear regression model was fitted to the actual sorghum production in Tunliu District from 2004 to 2017. The model was verified by using the actual sorghum production in 2018. The results showed that the main climatic factors that affecting sorghum production were rainfall and duration of light in May, average relative humidity and rainfall in June, rainfall and duration of light in July, precipitation in August, and duration of light in September. In 2018, the accuracy of the multivariable linear regression model for sorghum production verification was verified to be 81.06%, which indicated that the model could fit the sorghum production in Tunliu District relatively accurately, and provided a method for precise prediction of sorghum production.

Key words: climate, production, exponential smoothing for three times, particle swarm optimization, absolute grey relational analysis, multiple linear regress

中图分类号:

S162.5

赵汝男, 杨华, 徐民孜, 杨怀卿. 高粱产量预测模型研究——以屯留区为例[J]. 中国农学通报, 2020, 36(14): 28-33.

Zhao Runan, Yang Hua, Xu Minzi, Yang Huaiqing. Prediction Model of Sorghum Production: An Example of Tunliu District[J]. Chinese Agricultural Science Bulletin, 2020, 36(14): 28-33.

图/表 4

年份	t	Y_t/(kg/hm²)	$S t 1$	$S t 2$	$S t 3$	Y_bt/(kg/hm²)	Y_ct/(kg/hm²)
	0		13900.00	13900.00	13900.00
2004	1	13680	13760.61	13811.68	13844.04	13481.82	198.18
2005	2	13950	13880.61	13855.35	13851.21	14106.78	-156.78
2006	3	14070	14000.61	13947.39	13912.15	14279.80	-209.80
2007	4	14190	14120.61	14057.14	14004.01	14370.87	-180.87
2008	5	14220	14183.58	14137.25	14088.43	14291.48	-71.48
2009	6	14280	14244.67	14205.31	14162.49	14326.27	-46.27
2010	7	14400	14343.09	14292.61	14244.93	14501.76	-101.76
2011	8	14520	14455.18	14395.61	14340.40	14650.20	-130.20
2012	9	14565	14524.76	14477.44	14427.23	14632.38	-67.38
2013	10	14550	14540.75	14517.56	14484.46	14530.34	19.66
2014	11	14700	14641.65	14596.18	14555.25	14799.52	-99.52
2015	12	15000	14868.70	14768.85	14690.59	15302.02	-302.02
2016	13	15750	15427.09	15185.91	15004.42	16529.97	-779.97
2017	14	16200	15916.81	15649.01	15412.83	16883.27	-683.27

年份	t	Y_t/(kg/hm²)	$S t 1$	$S t 2$	$S t 3$	Y_bt/(kg/hm²)	Y_ct/(kg/hm²)
	0		13900.00	13900.00	13900.00
2004	1	13680	13760.61	13811.68	13844.04	13481.82	198.18
2005	2	13950	13880.61	13855.35	13851.21	14106.78	-156.78
2006	3	14070	14000.61	13947.39	13912.15	14279.80	-209.80
2007	4	14190	14120.61	14057.14	14004.01	14370.87	-180.87
2008	5	14220	14183.58	14137.25	14088.43	14291.48	-71.48
2009	6	14280	14244.67	14205.31	14162.49	14326.27	-46.27
2010	7	14400	14343.09	14292.61	14244.93	14501.76	-101.76
2011	8	14520	14455.18	14395.61	14340.40	14650.20	-130.20
2012	9	14565	14524.76	14477.44	14427.23	14632.38	-67.38
2013	10	14550	14540.75	14517.56	14484.46	14530.34	19.66
2014	11	14700	14641.65	14596.18	14555.25	14799.52	-99.52
2015	12	15000	14868.70	14768.85	14690.59	15302.02	-302.02
2016	13	15750	15427.09	15185.91	15004.42	16529.97	-779.97
2017	14	16200	15916.81	15649.01	15412.83	16883.27	-683.27

表1. 趋势产量与气象产量分解结果

月份	气候因子	相关系数
5月	降水量	0.9779
5月	光照时长	0.7217
6月	平均相对湿度	0.7491
6月	降水量	0.8597
7月	降水量	0.8940
7月	光照时长	0.7797
8月	降水量	0.7058
9月	光照时长	0.7346

表2. 气候因子与高粱气象产量灰色绝对关联分析结果

图1. 实际产量拟合结果

图2. 实际产量拟合模型残差结果

参考文献 25

[1]	刘茂柯, 唐玉明, 任道群 , 等. 酿酒高粱籽粒酿造性能的比较[J]. 中国酿造, 2012,31(11):111-114.
[2]	Tooth S . Research resource review: Olav Slaymaker, Tom Spencer, and Christine Embleton-Hamann (eds), Geomorphology and Global Environmental Change. Cambridge: Cambridge University Press[J]. Progress in Physical Geography, 2010,34(6):865-868.
[3]	刘彦随, 刘玉, 郭丽英 . 气候变化对中国农业生产的影响及应对策略[J]. 中国生态农业学报, 2010,18(4):905-910.
[4]	杜连仲, 李卫东, 王韧 . 豫中高粱产量与气象因子关系的统计分析[J]. 河南农业大学学报, 1985(1):52-57.
[5]	侯玉虹, 陈传永, 郭志强 , 等. 春玉米不同产量群体叶面积指数动态特征与生态因子资源量的分配特点[J]. 应用生态学报, 2009,20(1):135-142.
[6]	Niu X Z, William E, Cynthia J H , et al. Reliability and input-data induced uncertainty of the EPIC model to estimate climate change impact on sorghum yields in the U.S. Great Plains[J]. Agriculture, Ecosystems & Environment, 2009,129(1-3):268-276.
[7]	陈娟, 罗宇翔, 穆彪 , 等. 茅台酒用高粱产量品质与气象因子研究[J]. 高原山地气象研究, 2012,32(1):73-76.
[8]	李慧明, 李霞, 平俊爱 , 等. 我国高粱机械化发展前景及配套栽培技术研究[J]. 现代农业科技, 2015(14):45-46.
[9]	Lukas W, Markus O, Barbara A , et al. Sorghum, a sustainable feedstock for biogas production? Impact of climate, variety and harvesting time on maturity and biomass yield[J]. Biomass and Bioenergy, 2017,106:137-145.
[10]	Boubacar D, Li M, Tang C C , et al. Biomass yield, chemical composition and theoretical ethanol yield for different genotypes of energy sorghum cultivated on marginal land in China[J]. Industrial Crops and Products, 2019,137:221-230.
[11]	Kokou A A, John P A, Lamers J B N , et al. Climate change impact on water- and nitrogen-use efficiencies and yields of maize and sorghum in the northern Benin dry savanna, West Africa[J]. Field Crops Research, 2019,235:104-117.
[12]	赵建武, 白文斌, 刘贵锋 , 等. 不同播期、积温、降水量对高粱农艺性状形成及产量的影响[J]. 农学学报, 2014,4(4):1-4.
[13]	J Wolf, K Ouattara, I Supit . Sowing rules for estimating rainfed yield potential of sorghum and maize in Burkina Faso[J]. Agricultural and Forest Meteorology, 2015, 214-215:208-218.
[14]	Thierry K T, Daniel F, Ibrahima N , et al. The sowing date and post-flowering water status affect the sugar and grain production of photoperiodic, sweet sorghum through the regulation of sink size and leaf area dynamics[J]. Field Crops Research, 2016,192:67-77.
[15]	安盼盼, 明博, 董朋飞 , 等. 黄淮南部玉米产量对气候生态条件的响应[J]. 作物学报, 2018,44(3):442-453.
[16]	张镇涛, 杨晓光, 高继卿 , 等. 气候变化背景下华北平原夏玉米适宜播期分析[J]. 中国农业科学, 2018,51(17):3258-3274.
[17]	张福耀, 平俊爱, 赵威军 . 中国酿造高粱品质遗传改良研究进展[J]. 农学学报, 2019,9(03):21-25.
[18]	Zhou B, Yue Y, Sun X, Ding Z et al. Maize kernel weight responses to sowing date-associated variation in weather conditions[J]. The Crop Journal, 2017,5(1):43-51.
[19]	赵东妮, 王艳华, 任传友 , 等. 3种水稻趋势产量拟合方法的比较分析[J]. 中国生态农业学报, 2017,25(3):345-355.
[20]	魏庆伟, 石俊峰, 张亿博 . 冬小麦趋势产量与气象产量分离方法对比分析[J]. 山东农业科学, 2019,51(8):127-132.
[21]	Kennedy J, Ebeehaet R . Particle swarm optimization [C].IEEE International Conference on Neural Networks (Perth,Australia), Piscataway:IEEE Serv-ice Center,1942-1948, 1995.
[22]	刘小妹, 柯林, 于俊杰 . 灰色绝对关联度的改进模型[J]. 数学的实践与认识, 2018,48(10):16-22.
[23]	刘小妹, 柯林, 于俊杰 . 三维灰色绝对关联度的改进模型[J]. 统计与决策, 2018,34(22):20-24.
[24]	王占林, 张海春 . 基于多元回归的高寒地区油菜产量预测模型[J]. 中国农学通报, 2019,35(14):32-35.
[25]	Faried A H A, Khairy H A, Safwat A A . Comparative study and multiple linear regression analysis for assessment of chromatographic behavior of structurally related β-blockers on different stationary phases.[J]. Journal of separation science, 2019,42(24).