Few Successfully Used Models In Agrometeorology

Large scale evolution of computers since 1960 allowed to synthesize detailed knowledge on plant physiological processes in order to explain the functioning of crops as a whole. Insights into various processes were expressed using mathematical equations and integrated in simulation models. In the beginning, models were meant to increase the understanding of crop behaviour by explaining crop growth and development in terms of the understanding physiological mechanisms. Over the years new insights and different research questions motivated the further development of simulation models. In addition to their explanatory function, the applicability of well-tested models for extrapolation and prediction was quickly recognized and more application-oriented models were developed. For instance demands for advisory systems for farmers and scenario studies for policy makers resulted in the evolution of models, geared towards tactical and strategic decision support respectively.

Now, crop growth modeling and simulation have become accepted tools for agricultural research.A few models used in agrometeorological studies are:

1. The de Wit school of models

In the sixties, the first attempt to model photosynthetic rates of crop canopies was made (de Wit, 1965). The results obtained from this model were used among others, to estimate potential food production for some areas of the world and to provide indications for crop management and breeding (de Wit, 1967; Linneman et al., 1979). This was followed by the construction of an Elementary CROp growth Simulator (ELCROS) by de Wit et al. (1970). This model included the static photosynthesis model and crop respiration was taken as a fixed fraction per day of the biomass, plus an amount proportional to the growth rate. In addition, a functional equilibrium between root and shoot growth was added (Penning de Vries et al., 1974). The introduction of micrometeorology in the models (Goudriaan, 1977) and quantification of canopy resistance to gas exchanges allowed the models to improve the simulation of transpiration and evolve into the BAsic CROp growth Simulator (BACROS) (de Wit and Goudriaan, 1978).

2. IBSNAT and DSSAT Models

In many countries of the world, agriculture is the primary economic activity. Great numbers of the people depend on agriculture for their livelihood or to meet their daily needs, such as food. There is a continuous pressure to improve agricultural production due to staggering increase in human population. Agriculture is very much influenced by the prevailing weather and climate. The population increase is 2.1 per cent in India. This demands a systematic appraisal of climatic and soil resources to recast an effective land use plan. More than ever farmers across the globe want access to options such as the management options or new commercial crops. Often, the goal is to obtain higher yields from the crops that they have been growing for a long time. Also, while sustaining the yield levels they want to :

1 . Substantially improve the income.

2. Reduce soil degradation.

3. Reduce dependence on off-farm inputs.

4. Exploit local market opportunities.

5. Farmers also need a facilitating environment in which;

a. Affordable credit is available.

b. Policies are conducive to judicious management of natural resources.

c. Costs and prices of production are stable.

6 Another key ingredient of a facilitating environment is information, such as :

a. An understanding of which options are available.

b. How these operate at farm level.

c. The impact on issues of their priority.

To meet the above requirements of resource poor farmers in the tropics and sub tropics IBSNAT (International Benchmark Sites Network for Agro-technology Transfer) began in 1982. This was under a contract from the U.S. Agency for International Development to the University of Hawaii at Manoa, USA. IBSNAT was an attempt to demonstrate the effectiveness of understanding options through systems analysis and simulation for ultimate benefit of farm households across the globe. The purposes defined for the IBSNAT project by its technical advisory committee were to :

1. Understand ecosystem processes and mechanisms.

2. Synthesize from an understanding of processes and mechanisms, a capacity to predict outcomes.

3. Enable IBSNAT clientele to apply the predictive capability to control outcomes.

The models developed by IBSNAT were simply the means by which the knowledge scientists have and could be placed in the hands of users. In this regard, IBSNAT was a project on systems analysis and simulation as a way to provide users with options for change. In this project many research institutions, universities, and researchers across the globe spent enormous amount of time and resources and focused on:

1 Production of a "decision support system" capable of simulating the risks and consequences of alternative choices, through multi-institute and multi-disciplinary approaches.

2. Definition of minimum amount of data required for running simulations and assessing outcomes.

3. Testing and application of the product on global agricultural problems requiring site-specific yield simulations.

The major product of IBSNAT was a Decision Support System for Agro-Technology Transfer (DSSAT). The network members lead by J.W. Jones, Gainesville, USA developed this. The DSSAT is being used as a research and teaching tool. As a research tool its role to derive recommendations concerning crop management and to investigate environmental and sustainability issues is unparalleled. The DSSAT products enable users to match the biological requirements of crops to the physical characteristics of land to provide them with management options for improved land use planning. The DSSAT is being used as a business tool to enhance profitability and to improve input marketing.

The traditional experimentation is time consuming and costly. So, systems analysis and simulation have an important role to play in fostering this understanding of options. The information science is rapidly changing. The computer technology is blossoming. So, DSSAT has the potential to reduce substantially the time and cost of field experimentation necessary for adequate evaluation of new cultivars and new management systems. Several crop growth and yield models built on a framework similar in structure were developed as part of DSSAT package. The package consists of : 1) data base management system for soil, weather, genetic coefficients, and management inputs, 2) Crop-simulation models, 3) series of utility programs, 4) series of weather generation programs, 5) strategy evaluation program to evaluate options including choice of variety, planting date, plant population density, row spacing, soil type, irrigation, fertilizer application, initial conditions on yields, water stress in the vegetative or reproductive stages of development, and net returns.

Other Important Models

Crop models can be developed at various levels of complexity. The level of complexity required depends on the objective of the modeling exercise. The top-down approach to model design (Hammer et al., 1989; Shorter et al., 1991) is appropriate for models aimed at yield prediction. In this approach, complexity is kept to a minimum by commencing with a simple framework and only incorporating additional phenomena or processes if they improve the predictive ability of the model. Sinclair (1986), Muchow et al. (1990) and Hammer and Muchow (1991) have adopted this method in developing models of soybean, maize and sorghum respectively.

The EPIC, ALAMANC, CROPSYST, WOFOST, ADEL models are being successfully used to simulate maize crop growth and yield. The SORKAM, SorModel, SORGF as also ALMANAC models are being used to address specific tasks of sorghum crop management. CERES - pearl millet model, CROPSYST, PmModels are being used to study the suitability and yield simulation of pearl millet genotypes across the globe. Similarly, the two most common growth models used in application for cotton are the GOSSYM and COTONS models. On the same analogy the PNUTGRO for groundnut, CHIKPGRO for chick pea, WTGROWS for wheat, SOYGRO for soybean, QSUN for sunflower are in use to meet the requirements of farmers, scientists, decision makers, etc., at present. The APSIM, GROWIT added with several modules are being used in crop rotation, crop sequence and simulation studies involving perennial crops.

Under Indian sub-continent conditions, crop yield forecasting based on meteorological data is very important from several points of view. Using crop yield as dependent variable and weather factors as independent variables, empirical statistical models for predicting crop yield have been reported by Mall et al. (1996). Also, Mall and Gupta (2000) reported a successful empirical statistical - yield weather model for wheat in the Varanasi district of Uttar Pradesh, India. Rajesh Kumar et al. (1999) developed a stepwise regression model which states that the pigeonpea yield variation by weather variables is upto 94 per cent in Varanasi District, U.P., India. Using a thermal time and phasic development model. Patel et al. (1999) found that pigeonpea phenophases depended upon the available heat units and accounted for 98 per cent of total variation in Anand, Gujarat State, India.

Elsewhere in the world there were studies (Robertson and Foong, 1977; Foong, 1981), which predicted yields based on climatic factors using mathematical modeling. Factors such as water deficit, solar radiation, maximum and minimum temperatures which play a vital role at floral initiation were taken into consideration to construct these models. In Southern Malaysia Ong (1982) found a high correlation between yield of oil palm and rainfall and dry spells as also temperature and sunshine using a step-wise regression approach. Chow (1991) constructed a statistical model for predicting crude palm oil production with trend, season, rainfall, etc. Amissah-Arthur and Jagtap

(1995) successfully assessed nitrogen requirements by maize across agro ecological zones in Nigeria using CERES-maize model. Hammer et al. (1995) using local weather and soil information correlated peanut yields with estimates from PEANUTGRO, a model in the CERES family and gave a regression with high coefficient (r2 = 0.93) of variation. The construction of contemporary crop models entails the combination of many algorithms for physiological processes and impact of environmental factors on process rates (Monteith, 2000). This clearly indicates that in the development of models and their application for solving problems at field level on agrometeorological aspects are given due weightage.

Was this article helpful?

0 0

Post a comment