Published in Proceedings of International Symposium of Environments and Biometeorology (1995.7). Beijing: Chinese Agricultural ScienTech Press. 1996, p.567-573
Dong Zhanshan, Han Xiangliang
Beijing Agricultural University
ABSTRACT
A Cotton Management System suitable for the Huang-Huai-Hai region in China has been established by application of techniques of crop simulation model and knowledge engineering. It is composed of cotton simulation model (GOSSYM), cotton management expert system (CMES), graphical user interface (GUI) and database. It can suggest the applications of nitrogen fertilizer, irrigation and plant growth regulator.
Finally, the computer simulation experiments to select the optimal row spacing and plant population have been done through running COTMAS. The results showed that suggesting optimal values by COTMAS are similar to practical ones. Fertilizing is a basic action in cotton management, so an example suggesting application of nitrogen fertilizer was provided in this paper.
Key words cotton, simulation model, management system, expert system
China is one of the biggest cotton producing country in the world. Its cotton-planting history is long; its regions suitable for planting cotton is wide. So far, total lint yield of the country has been the most in the world, but its lint yield per unit is lower than other countries that have advanced cotton-planting techniques. A great progress in study of cotton management system has been made, such as GOSSYM-COMAX, which has extended in 14 states in USA cotton belt. For elevating the technological level of cotton cultivation in our country and accelerating the application of new techniques in cotton production, it is necessary to do the research of cotton simulation model and management system.
Huang-Huai-Hai cotton-planting region locates in Huang-Huai-Hai plain in North China. It has flat land, deep soil layer, soft soil texture, convenient drainage, plenty sunshine, proper rainfall, moderate temperature, fast elevation of temperature in spring, and many fine days in fall. These properties are suitable for early growth in cotton seedling, stable growth in middle stage, and boll opening in the later season. That the region is selected as a research zone has a very practical meaning.
Cotton crop management system suitable for Huang-Huai-Hai region is programmed by using object-oriented programming method with Borland Pascal 7.0. Source lists of program have 18000 lines. The compiled executable program COTMAS.EXE can run in MS WINDOWS 3.1, and it can use Chinese interface if Chinese Star 2.0 is running.
1 Structure of COTMAS
COTMAS is composed of graphical user interface (GUI), cotton management expert system (CMES), cotton simulation model (GOSSYM) and database. Its basic structure is shown in Fig. 1, which also expresses the relations of modules of COTMAS.
Figure 1 Basic Structure of COTMAS
2 Modules of COTMAS
2.1 Graphical User Interface (GUI)
Users can input the data into database of COTMAS, and make decisions by using functions of COTMAS, and output the information of its database. GUI is programmed by application of ObjectWindows library of Borland International, and source lists have 5000 lines.
The GUI is a dialog window in which there are 12 icons that stand for 12 functions. Users can click one icon to perform a system function. Every function has a individual dialog window that includes controls and buttons. Users can use the system easily if he/her can use Windows.
2.2 Cotton Management Expert System (CMES)
Since the emergence of cotton seed, farmers start to dynamically manage cotton growth in the field. Management focuses on whether a measure is applied or not, and its application date and amount. For managing large cotton fields, experts’ directions are very important to gain a good economic, social and ecological efficiency.
In the decision of cotton management, a short management schedule, such as fertilizing, irrigation and using plant growth regulator, is suggested by checking the growing potential of cotton plant.
By application of object-oriented programming, an abstract object is established to express the unit of management of cotton field. The management knowledge is ruled, and programmed in the object. When decision-making, reference is performed by information transferring, first, determine the application date and base amount, then run GOSSYM to see the result of lint yield for some times, finally CMES to suggest a schedule of management to user.
Cotton management expert system (CMES) is an expert system to determine application date and amount of agronomy measures. It includes 3 expert sub-systems: nitrogen fertilizer expert sub-system, irrigation expert sub-system and PGR (plant growth regulator) expert sub-system. In this system, knowledge and program are linked. CMES source lists have 2000 lines.
2.3 Cotton Simulation Model (GOSSYM)
Gossym is a dynamic model, and can simulate the growth and development of cotton plant and lint yield in physiological processing level. In essence, the model is a material balance model of water and nitrogen in root rhizosphere and of carbon and nitrogen in plant; it includes sub-model of water balance, nitrogen balance, carbon balance, forming and partitioning of photosynthesis products, morphological formation.
Gossym simulates the photosythesis, growing, morphological formation in a day, but simulate moving of the soil water in 10 times a day.
Gossym can simulate the effect of environmental variables on cotton plant. These variables include meteorological condition, such as daily solar radiation, maximum and minimum temperature, wind speed, rainfall, and agronomic measure, such as plant population, row spacing, cultural measure, nitrogen fertilizing, and irrigation.
By using the dataset obtained in Cotton Research Institute, CAAS, located in Anyang, Henan, in 1990, and the modified variety parameter file and soil hydrological file, calibration of GOSSYM show that this model can generally simulate the growth and development of cotton plant and lint yield of cotton in Huang-Huai-Hai Region in China. So GOSSYM can be introduced to our country, used in study of cotton science, and integrated in cotton management system suitable for our country.
GOSSYM was programmed in FORTRAN. For simply integrating the system, it is programmed in BORLAND PASCAL again, and includes 11000 lines in source lists.
2.4 Database
Database is a data-gathering location of COTMAS, and includes various files that are made by GOSSYM, CMES and GUI.
In Fig.1, identifying climate pattern model (IDCLIM) and simulator of weather (SIMWTH) are included, but both models are not linked with COTMAS in which has interfaces of two models.
3 Functions of COTMAS
3.1 Application of COTMAS in cotton management
? Decision for irrigation date and amount
Irrigation timing and amounts are determined from water stress output, from effect on yield and from nearness to crop maturity. GOSSYM practically predicts when the crop will enter physiological stress due to lack of soil water. Irrigation to alleviate that future physiological stress is the decision supported by CMES.
? Decision for nitrogen date and amount
Nitrogen timing and amounts are also determined from nitrogen stress output, from effect on final yield and from effect on vegetative growth. GOSSYM practically predicts when the crop will enter physiological stress caused by nitrogen shortage. CMES estimates the required nitrogen amount from the hypothetical lint yield, then runs GOSSYM some times, and suggests an application schedule of nitrogen to alleviate the future physiological stress.
? Decision for plant growth regulator
Plant growth regulation can decrease the detrimental effects of rapid vegetative growth by shortening the plant and causing photosynthate to be used for reproductive growth. DPC is a currently popular plant growth regulation, and its application timing and amount are determined from status of growth and development of cotton. DPC timing is divided into 3 stage: square, first bloom and boll. In normal climate pattern, DPC amount is 0~15 g/ha in square stage, 30~60 g/ha in first bloom, 45~60 g/ha in boll stage. GOSSYM can simulate the effect on yield of DPC, then CMES can decide the application date and amount.
? Estimation of lint yield
GOSSYM can simulate the lint yield of simulated field and estimate lint yield before maturity of cotton plant, but it is not able to estimate lint yield of cotton in large region.
3.2 Application of COTMAS in research of cotton science
? Simulation of cotton experiment design
Crop simulation model can act as a tool to assist field experiment design after its validation proved in several climate patterns and several places. It can be used to simulate the plant growth and development of cotton under various environments, and then you can select the effective plan of experiment so that the traditional regression or orthogonal design is not selected for field experiment. Before planting, you can simulate the growth and development by using hypothetical future weather file, then select optimal row spacing, plant population, and planting date.
? Relationship research between cotton and environment
Originally, crop simulation model was used to simulate relationship between crop and environment, so it is a powerful tool to study effect on crop of various environmental factors. For example, Reddy (1989) studies effect of carbon dioxide (CO2), ozone (O3) on cotton crop.
? Application in cotton breeding
In cotton breeding, a new strain will be planted at many environment sites when it is selected for 2~3 year. If it grows well and gains high lint yield and has good quality, it would be extended in selected region. Crop simulation model can simulate its growth and lint yield at various ecological environments that stand for various geographical sites. Those sites that have high yield and good quality are best planting regions of the strain.
? Application in research of cotton potential productivity
We can use crop simulation model to study the climate productivity and soil productivity in any location to determine the prospective about planting cotton.
4 Application Example of COTMAS
Now, COTMAS is used to determine the best row spacing and plant population, and suggest a nitrogen schedule.
4.1 The proper plant population and row spacing
Crop simulation model can simulate rapidly growth and development of crop under various ecological environments including climate, soil and agronomic culture inputs by computer, and this is called computerized crop experiments. By analyzing the results of computerized experiment, some regular things can be found and served cotton production.
Optimum row spacing and plant population on cotton cultivar CRI12 and CRI 17 have been simulated by GOSSYM (Table 1, Table 2, respectively). CRI12, CRI17 are planted in April 13, May 3, respectively.
Table 1 Simulated lint yield of CRI12 under some row spacing and plant population (kg/ha)
Plant population (kilo-plants/ha.)
15
30
45
60
75
60
1377
1399
1391
1382
1369
1384
80
1580
1713
1707
1686
1669
1671
100
1569
1917
1953
1938
1902
1861
120
1466
1851
1946
1959
1781
1801
average
1498
1720
1749
1741
1680
1678
Table 1 showed that proper plant population of CRI12 is 30~60 kilo-plant/ha, and that proper row spacing is 1 meter. The results are similar to practical measurements. CRI12 is a middle-maturity variety of cotton, and has taller height, and need larger nutritional spaces. Generally, the optimum plant population under high-yield condition is 45 kilo-plant/ha. (Zhu Mingzhe, 1988; Jiang Guozhu, 1990).
Table 2 Simulated lint yield of CRI17 under some row spacing and plant population (kg/ha.)
Row spacing (cm)
Plant population (kilo-plants/ha.)
Average lint yield
45
60
75
90
105
60
1249
1252
1240
1226
1220
1237
80
1477
1511
1513
1505
1487
1499
100
1446
1513
1426
1376
1336
1419
average
1391
1425
1393
1369
1348
Table 2 showed that proper plant population of CRI17 is 45~60 kilo-plant/ha, and its proper row spacing is 0.8~1 m, simulated lint yield is 1426~1513 kg/ha. The planting schedule of CRI17 made by its breeders is that sowing date is April 28 to May 8, and that plant population is 60~75 kilo-plant/ha with 0.9 m row spacing.
4.2 An example of nitrogen schedule made by COTMAS
A nitrogen decision-making example by COTMAS has been given under following conditions: CRI12, sowing on April 12, emergence on April 23, 45 kilo-plant/ha of plant population, 0.8 m of row spacing, 120 kg/ha of nitrogen content in soil before sowing. How many amounts of nitrogen fertilizer will be needed to be sidedressed and when will be applied if 2000 kg/ha lint yield will be wanted to be gotten?
According to needing 0.12~0.18 kg pure nitrogen by producing 1 kg lint yield, producing 2000 kg/ha needs 240~360 kg N /ha. This amount subtracting 120 kg/ha of soil nitrogen content is equal to 120~240 kg/ha that is rough range sidedressing nitrogen amount. Then CMES runs GOSSYM several times to determine how many amounts of N and when will be applied. Finally, CMES suggests that application date is July 27 and pure nitrogen of application is 140 kg/ha.
Reference
Baker, D. N., J. R. Lambert and J. M. McKinion (1983) GOSSYM: A simulator of cotton growth and yield. S.C. Agri. Exp. Stn. Tech. Bull. 1089
Chinese Cotton Research Institute (Ed.) (1983) Chinese Culture of Cotton. Shanghai Science and Technique Press (in Chinese)
Deng Shaohua et al. (1992) Studies on mathematical model on the relationship between chemical regulation and lint yield. Acta Gossypii Sinica. 4(Supplement): 42~52 (in Chinese)
Dong Zhanshan (1994) Development of computer system for cotton crop management. Chinese Agronomy Journal. 10(4):41~43 (in Chinese)
Dong Zhanshan, Pan Xuebiao et al. (1992) CPMSS/CGSM: A decision-making system for cotton crop management. Proceedings of 1st Youth Agronomy Symposium. Chinese Science and Technique Press. pp.427~432 (in Chinese)
Dong Zhanshan, Pan Xuebiao et al. (1993) Design and implementation of decision-making support system for cotton crop management CPMSS. Agricultural Application of Computer. (1):16~19 (in Chinese)
Han Xiangling, Qu Manli (1987) The exploitation and utilization of agroclimatic resources in Huang-Huai-Hai region. Beijing Agricultural University Press. pp.66~75 (in Chinese)
Jiang Guozhu et al. (1990) Studies on optimal decision model of agronomic measures for high yield and good quality in cotton. Acta Gossypii Sinica. 2(1):51~57 ( in Chinese)
Lemmon, H. E. (1986) COMAX: An expert system for cotton crop management. Science, 233:29~33
Liu Wen, Wang Enli, Han Xiangling (1992) Study of cotton simulation model. Chinese Agrometeorology. 13(6):10~16 (in Chinese)
McKinion, J. M., D. N. Baker et al. (1989) Application of the COSSYM/COMAX system to cotton crop management. Agricultural System. 31:55~65
McKinion, J. M., T. L. Wagner (1994) GOSSYM/COMAX: A decision support system for precision application of nitrogen and water. In: Shi Yuanchun & Cheng Xu (Ed.) Integrated Resources Management for Sustainable Agriculture. Beijing Agricultural University Press. pp.32~38
Pan Xiaokang, Dong Zhanshan et al. (1991) Study on control measures and diagnosis indices of good yield of cotton. Agricultural Application of Computer. supp. pp.80~85 (in Chinese)
Reddy, V. R., D. N. Baker et al. (1989) Analysis of effects of atmosphere carbon dioxide and ozone on cotton yield trends. J. Environ. Qual. 18:427~432
Tom, S. (1993) Borland Pascal 7.0 programming for Windows. Random House, Inc.
Wang Suyun (1993) Properties and development of special belt of cotton production in China. Information Research of Agriculture. (4):12~18 (in Chinese)
Zhou Jichu (1991) Some problems in culture of good yield of CRI12. Chinese Cotton. 18(3):32 (in Chinese)
Zhu Mingzhe et al. (1988) Study on mathematical model of agronomic measures for high yield, good quality and high profit of cotton. Agronomy Journal of North China. 3(1):13~20 (in Chinese)
