ISWC OpenIR  > 水保所知识产出(1956---)
DSSAT 模型 在黄土丘陵区 的适用性评价及水肥 管理应用
邹 龙
Subtype硕士
Thesis Advisor冯 浩
2014-03
Degree Grantor中国科学院研究生院
Place of Conferral北京
KeywordDssat 黄土高原 春玉米
Abstract

农业技术转化决策系统( DSSAT,Decision Support System for Agrotechnology
Transfer)是一个以模拟作物生长为基础的机理模型系统,可逐日模拟作物生长和发育
过程,可响应许多因素,包括作物遗传特性、管理措施、环境、氮素和水分胁迫、病
虫害等,主要用于农业试验分析、农业产量预报、农业生产风险评估、气候对农业的
影响评价等,也可用于作物光温、降水生产潜力的模拟、作物栽培方案的优化等,可
以为农业技术选择提供合理有效的决策和预测。CERES(Crop-Environment Resource
Synthesis System)系列模型是密西根州立大学Ruthie教授等在20世纪80年代初建立的
谷类作物模拟模型,它不仅能模拟作物生长发育的主要过程,还能模拟土壤养分平衡
(矿化、硝化、反硝化、固氮、淋溶、吸收和利用等)与水分平衡(有效降雨、径流、蒸
发、蒸腾、土壤水分的垂直流动与渗漏等),其作为DSSAT模型中的主要作物生长模
型之一在本世纪初在我国开始得到应用。
本研究以陕西安塞农田生态系统国家野外科学观测研究站 2002、2003、2005 年
春玉米田间试验结果为基础,利用 DSSAT4.5 中的 CERES-Maize 模型对春玉米生育
期、产量、生物量进行模拟,并结合田间实测数据进行对比分析,对 CERES-Maize
模型在该区春玉米种植中的适用性进行评价,并且在完成模型参数调试的基础上,设
定一系列的施肥和灌溉情景,利用 DSSAT-CERES 模型在平水年份 2005 年的光照、
气温、降雨等条件下模拟不同情景下春玉米生长的情况,结合模拟结果对该区春玉米
种植中的水肥效应进行分析探讨,得到以下一些结论:
1)在正常降雨条件的情况下,CERES-Maize 模型能够较好的模拟黄土高原丘陵
沟壑区春玉米的生长过程、地上部分生物量以及收获产量—2002 年和 2005 年对春玉
米生育时期的模拟误差均在 3 天以内,产量模拟的归一化的均方根误差(n-RMSE)
和平均相对误差(MRE)均小于 10%,生物量模拟的误差稍大,但均在可接受范围
内。
2)在降雨较少的年份,模型对作物生产实际情况的再现能力相对变弱—2003 年模拟的生育期时间和春玉米实测生育期时间基本吻合,唯独出苗日期相差较大为 9
天;产量模拟的 n-RMSE 和 MRE 分别为 6.47%和 5.03%;地上生物量模拟误差稍大,
模拟结果的 n-RMSE 和 MRE 分别为 31.81%和 30.21%,说明 CERES-Maize 模型还不
能很好的模拟干旱年份下作物的生长过程。
进一步对模型的模拟过程分析发现,模型在春玉米模拟过程中,对作物系数 Kc
的确定与实际研究中有所差异,据相关研究表明,Kc 在作物生长过程中是介于 0-1
之间变化的,然而模型在模拟过程中默认 Kc 为定值 1,对 Kc 的高估使得模型对作物
的干旱胁迫的模拟加重,从而导致模拟结果低于实测的结果。值得说明的是 2003 年
春玉米生育期间降水分布与其余两年差别较大,表现为前期降水很少,后期降水较多,
在春玉米生长后期的降水补充在一定程度弥补了前期干旱造成的生物量和产量损失,
灌浆期间增加的干物质首先用来满足籽粒的需要,剩余的才会转移到茎,所以模型可
以很好的对籽粒产量进行模拟,而对生物量的模拟误差偏大。要使模型更好的运用于
干旱地区,以上模拟过程需要进一步完善和改进。
3)在平水年分(2005 年为例)的气象条件下的一系列预测模拟结果表明:雨养
条件下,对不同有机肥用量和氮肥用量处理下的春玉米生长进行模拟,春玉米产量和
地上生物量的结果均表明,在控制氮肥用量的同时可以通过增施有机肥料来弥补由于
少使用氮肥而带来的产量损失;就肥料的投入量对春玉米产量的增加效果而言,7500
kg/hm 2 的有机肥料和 400 kg/hm 2 氮用量处理(M1N4)与 15000 kg/hm 2 有机肥料和 300
kg/hm 2 氮用量处理(M2N3)下春玉米的产量、生物量以及氮的偏生产率均相差不大,
且都维持在较高水平,两个处理的效果基本相同并且都优于其他处理方式。
4)对灌一水条件下各个施肥处理进行模拟,结果表现为:在有机肥和 N 肥用量
较低的情况下,生育期内灌一水(无论是哪个生育期)对春玉米产量影响可能不大,
此时,春玉米产量的主要限制因素可能是肥料用量;然而在有机肥和 N 肥使用量均
较大的条件下,灌水对产量的增加有明显的促进效果。此外,同一施肥水平下不同时
期的灌水处理产量均表现为在灌浆期灌水的处理(Ig)的产量均高于其他处理,这说
明避免春玉米在灌浆期间受到严重水分胁迫对保证春玉米产量有重要意义。
关键词 :DSSAT,黄土高原,春玉米

Other Abstract

The Decision Support System for Agrotechnology Transfer is a model which based on
modeling the growth of crop. It can simulate the daily growth of crop and respond many
factors, such as the Genetic characteristics of crop, management measures, environment,
the duress of water and nitrogen, pests, and so on. It was used to analysis the agricultural
experiment, predict yield, and evaluate the risk of produce and the influence of climate. It
also can work on the simulation of productive capacity based on photosynthesis,
temperature and rain, the optimal schedule for cultivation, and so on. The model also was
used to provide reasonable and effective policy decision and prediction. CERES
(Crop-Environment Resource Synthesis System) is a model about corn which was found
by Ruthie who a professor of Michigan State University in 1980s. It wasn’t only used to
model the main process of crop growth, also to simulate the balance of soil nutrients
(mineralization, nitration, denitrification, nitrogen fixation, leaching, absorbing and
utilizing, and so on.) and water (effective rain, run off, evaporation, transpiration, and the
move of soil water in plumb and level). It was one of the main crop models of DSSAT and
was used in China in the starting of 21th.
This study was based on the field experiment of spring maize in 2002, 2003 and 2005,
at Ansai Farm Ecosystem National Field Research Station Shaanxi. The CERES-Maize
which one of the main crop models of DSSAT4.5 was used to simulate the growth period,
yield and aboveground biomass of spring maize. To evaluate the applicability of the model
in spring maize planting, the simulated results were compared with the measured result. After the debugged of parameters, the CERES was used to model the growth situation
among the different treatments of spring maize under the conditions of 2005. The
discussion of the effect of water and fertilizer in the planting of spring maize was based on
the result of simulation. Some conclusions were received.
1) The CERES-Maize model could do well in simulating the growth period,
aboveground biomass and yield of spring maize at the Hill and Gully Area of the Loess
Plateau with normal rainfall. The error on growth period was only 3-days, both n-RMSE
and MRE of simulated yield were less than 10%, although there’s no good work on
aboveground biomass, the result could be accepted.
2) CERES could not reappear the situation of spring maize in field accurately in the
condition of less rainfall. It had a good job on growth period simulation except emergence
date which the error was 9days. This model was also good at the yield simulating, with the
n-RMSE and MRE of simulated yield were 6.47% and 5.03%, respectively. The model
could not be good at aboveground biomass modelling with little rainfall, with both
n-RMSE and MRE of simulated biomass were more than 30%.
There’s some work be done to study the process of the model, some consequences
were displayed, the Kc of model which was evaluated with 1, was different from the one in
our research which was a value ranged from 0 to 1. Because of over evaluating of Kc, the
duress of drought was over estimated and the simulated values were less than the measured
one, during the process of growth of spring maize. In the other hand, the distribution of
rainfall which little at the beginning of the period and more and more in the later in 2003
was most different from the other two years. The rainfall in the later reduced the loss of
yield and above ground biomass because of the little rain in the start of period. During the
grain-filling period the increased dry material was assigned to seed first, others were
moved to stem, so the yield of spring maize could be simulated well, but there’s a big error
in simulating of above ground biomass. In order to do some good work in dry areas with
this model, it must be optimized deeply.
3) Under the normal rainfall meteorological conditions (for example 2005), the output
of model was that the result of growth simulation of spring maize with the treatment of
different manure and nitrogen showed us it can reduce the loss of yield which because of
the less nitrogen were used in the production of spring maize, by added more manure in the  condition of rainfed. Both the two treatment which treatment M1N4 with manure: 7500
kg/hm 2 and nitrogen: 400 kg/hm 2 , treatment M2N3 with manure: 15000 kg/hm 2 and
nitrogen: 300 kg/hm 2 could bring us the equal effect on yield and aboveground biomass
increasing, productivity of nitrogen. Both of them were the better treatments rather than
others.
4) With one irrigation during the process of spring maize growing, the result told us
maybe it have no useful to increase the yield that during the growth stage irrigated water
once with both of manure and nitrogen were little used, this time the limitative factor of
yield may be fertilizer; on the other hand it had apparent effect on the yield increasing that
there’s one irrigation during the growing stage under the more dosage of manure and
nitrogen. In addition, the treatment Ig was the best one among the treatments in the equal
dosage of fertilizer. This indicated that it is important to ensure the yield that assured
there’s no duress during the growing stage of spring maize especially in the stage of
grain-filling.
Key word: DSSAT, The Loess Plateau, spring maize

Language中文
Document Type学位论文
Identifierhttp://ir.iswc.ac.cn/handle/361005/8985
Collection水保所知识产出(1956---)
Recommended Citation
GB/T 7714
邹 龙. DSSAT 模型 在黄土丘陵区 的适用性评价及水肥 管理应用[D]. 北京. 中国科学院研究生院,2014.
Files in This Item:
File Name/Size DocType Version Access License
DSSAT 模型 在黄土丘陵区 的适用性(1754KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[邹 龙]'s Articles
Baidu academic
Similar articles in Baidu academic
[邹 龙]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[邹 龙]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.