32 The underground portion of a plant that serves as support draws food and water from the surrounding soil, and stores
BIO-PHYSICAL MODEL and CROP GROWTH MODEL H. OZAN ERUYGUR Ph. D candidate, Graduate school of economics Middle East Technical University, Turkey.
1. INTRODUCTION Bio-economic models are economic models that integrate in formation coming from biophysical models. Commonly, they are mathematical programming models (MPM). The basic idea behind this approach is the fact that the way of representing technology in an economic model can be ameliorated in a very substantial way by the use of simulated data obtained from a biophysical model results(17). In addition, we can easily integrate environmental parameters associated with different agricultural techniques, such as soil erosion(18) or water pollution of different sources (nitrates(19) or chemical pesticides4) in bio-economic models. This gives us the opportunity to examine the possible results of externalities like erosion, pollution and salinity in the agricultural production processes when constructing agricultural economic models. It is obvious that this kind of economic modeling will represents the structure of real economic activity in a more accurate way. For example, while dealing with the economic dimension of agricultural water use, it is necessary to obtain detailed representation of the available set of techniques. In other words, to have a good knowledge of the production functions; these functions are commonly estimated by econometric procedures. However, usual econometric methods are based on statistical inference as in formation source. This approach presents
17
Flichman, G. (1995), Bio-Economic Models, Integrating Agronomic, Environmental
and Economic Issues with Agricultural Use of Water, CIHEAM, Montpellier. 18
The wearing away of the land surface by running water, wind, ice, or other geological
agents, including such processes as gravitational creep. In other words, the removal of material from the surface of the land by weathering, running water, moving ice, wind and mass movement. 19
Any compound containing the N03- radical.
145
different types of problems20). Firstly, lack of observations may not give the opportunity to the estimation of the function's parameters (the number of observations has to be higher than the number of parameters). Secondly, sometimes it is difficult to guarantee the random character of samples used (Boussemart et al, 1 994). Thirdly, in the case of using cross section data, the individual techniques that proportionate the basic data for the estimations of the production functions are influenced by the present structure of relative prices. Generally, the variation among different farms in respect to the proportions of factors that are used for a certain crop at a certain moment is quite low. Lastly, in the case of using time series data, the problem of technological progress measurement appears. Even the most advanced procedures trying to deal with this problem are not very convincing (Boussard 198821) In other terms, in the case of estimating agricultural production functions using statistical data from the past (time series), or from the present, through cross-data analysis, it will be very difficult to represent prosperity and accurately the technical universe. That is why, as well, we consider that the engineering production function approach, is a more appropriate procedure of the agricultural water use. In order to construct engineering production functions. we should obtain technical coefficients of production from results obtained by agronomic experimentation and survey data and not from statistical data adjusted to "a priori" defined mathematical functions. Here, a very practical problem arises: it is almost impossible to obtain all the necessary data using these sources. Generally, agronomic research is not done in order to obtain appropriate input data for economic models. Most of time it is very difficult, for example, to examine separately the influence of fertilization, irrigation, weather, soil quality, previous crop in the rotation and other factors on agricultural production. The same practical problems arise even in the
20
Flichman. G. ( 1 995). Bio-Economic Models, Integrating Agronomic. Environmental
and Economic Issues with Agricultural Use of Water. CIHEAM, Montpellier 21
Flichman, G. ( 1995). Bio-Economic Models. Integrating Agronomic. Environmental and Economic
Issues with Agricultural Use of Water. CIHEAM. Montpellier.
146
developed countries where very old experimental agricultural stations exist. In order to partially overcome these problems the agronomic simulation modeling or bio-physical modeling approach is used. This kind of procedure allows new ways of implementing the engineering production function approach. By using the bio-physical modeling, it becomes possible to simulate production outcomes concerning almost any type of agricultural technical schedule. Bio-physical models give us the chance to represent the effects on yields, coming from changes in the quantities of fertilizer22, the levels of irrigation, the types of equipment, the characteristics of alternative cultivars , and rotation schemes etc. In addition, these models allow the analysis of weather variability on production as well as on environmental parameters (pollution, soil erosion), related with different type of agricultural techniques. We can thereby take into account the interactions between irrigation and weather. Most of time it is very difficult, for example, to examine separately the influence of fertilization, irrigation, weather, soil quality, previous crop in the rotation(23) and other factors on agricultural production. The same practical problems arise even in the developed countries where very old experimental agricultural stations exist. In order to partially overcome these problems the agronomic simulation modeling or bio-physical modeling approach is used. This kind of procedure allows new ways of implementing the engineering production function approach. By using the bio-physical modeling, it becomes possible to simulate production outcomes concerning almost any type of agricultural technical schedule. Bio-physical models give us the chance to represent the effects on yields coming from changes in the quantities of fertilizer (24), the levels of irrigation, the types of equipment, the characteristics of alternative cultivars (25), and rotation schemes etc. In addition these models allow the analysis of weather variability on production as well as on
22
A material that is added to the soil to supply one or more p]ant nutrients in available form.
23
The practice of purposefully alternating crop species grown on the same plot of land
24
A material that is added to the soil to supply one or more plant nutrients in a readily
available form. 25
A specially developed agricultural plant variety. 147
environmental parameters (pollution, soil erosion), related with different type of agricultural techniques. We can thereby take into account the interactions between irrigation(26), weather.
fertilization(27), type of varieties and tillage systems in the
process of agricultural production. Moreover, in the case of economic analysis of agricultural water use, the possibility of simulating water response functions considering the .interactions of irrigation with the rest of agronomic management is very advantageous. However, the information and data gained from biophysical simulation models is useful but only to the extent that the results are accurate. Therefore, the predictive use of this kind of model implies the need of getting this model calibrated and validated. In other words, the model should be able to reproduce the behavior of a real system in order to allow us to change some parameters of that system (usually policy parameters as prices taxes or tariffs) and make forecasting analysis about the impact of these changes on the system. With the very large technological progress in the computer opportunities of our time, application of these kinds of sophisticated bio-physical models becomes easy. Even the complex processes of calibration and validation of these kinds of models becomes easily applicable in a relatively very short of time comparing to pass years. Therefore, there is an increasing trend in the use of bio-physical simulations and integrating them to the economic models and in the end constructing bio-economic models when trying to analyze the possible effects of some policy changes in agriculture. In addition, these models not only allow us to examine agricultural production results but also to examine environmental results of some agricultural policy changes and applications. Soil Erosion, soil and water pollution resulting from fertilization, soil salinity resulting from over irrigation are some hazardous examples of environmental results of agricultural productions. These concepts are called externalities in economics language and causes to some large social costs to the total welfare of the society. If we are only interested in the private costs ignoring the social costs of externalities, we will do bi_~: mistakes while trying to increase the social welfare of a society.
26
The artificial application of water to the soil for the benefit of growing crops.
27
Applying fertilizers.
148
Bio-physical models give us the opportunity to take into account these kinds of externalities and hence by using the bio-economic models to decrease them. By using simulation models we can examine the possible effects of agricultural policies on environment and agricultural production and in the end on social welfare without applying them. Agricultural production is a natural production meaning that it depends largely on natural conditions like weather, soil, location and this condition can be represented in cyber space with the help of positive science disciplines like soil science, biology and meteorology. By means of this positive science disciplines, the bio-physical models are created in cyber space, that is, in computer environment. Therefore, bio-physical models are very sophisticated interdisciplinary models and imply the need of interdisciplinary collective studies. Economists should use these models in the agricultural economics. Natural resource economics and environmental economics in order to construct more realistic models which can easily be used in the economic analysis. It can be said that bio-physical models provide economists with a very useful and necessary natural laboratory which can not be generally obtained in economics. However, one should keep in mind that these models are interdisciplinary models and therefore economists should be very careful in using them since the simulation models must be calibrated and validated before to use. Economists should work with scientists from other disciplines in order to get some accurate simulation models. In order to work in such a study economist should know some basic concepts of bio-physical models. Here, it is very important to notice that all the bio-physical models should be calibrated and validated with the help scientists from other disciplines. The main aim of our study is to show how CropSyst, as a bio-physical model, works. Calibration and validation procedures are explained but these are only some mathematical procedures, every time they should be confirmed by experts from other disciplines or by scientists who created these models as our model was done. One of the widely used bio-physical models is EPIC. It is a mechanistic simulation model used to examine long-term effects of various components of soil erosion on crop production (Williams et al., 1983). EPIC is a public domain model that has been used to examine the effects of soil erosion on crop production in over 60 different countries in Asia. South America and Europe. In addition, EPIC has an economic component in order to facilitate some economic analysis. Another widely used bio-physical model is CropSyst. This is a user-friendly,
149
conceptually simple multi-year multi-crop daily time step simulation model. It was developed to study the effect of cropping systems management on productivity and the environment. The model simulates the soil water budget, soil-plant nitrogen budget, crop canopy and root growth, dry matter production, yield, residue production and decomposition, and erosion. In the first chapter, we will look at some simulation models, especially EPIC and CropSyst, which are the most widely used models in these kinds of bio-physical simulation models. In the second chapter, we will start to learn CropSyst, and in doing that we will use a calibrated and validated model as an example. We are calling our model as YIM Model. The model is created for Konya-Karapinar region and our crops are wheat (winter), sugar beet and barley. In this chapter we will give some parameter ranges for Turkey and we will explain to which data we need. In the third chapter we will explain the calibration processes of CropSyst which is the most difficult part of CropSyst modeling. In the fourth chapter we will give our simulation model results and them with observed wheat, sugar beet and barley production value, Like all other models like it. uses a very different language for Economists. Therefore, we give some explanations as foot notes and to the end of the study we added a glossary for some terms that CropSyst uses.
150
2. BIOPHYSICAL MODELS Table 1. shows a list of some bio-physical models. Generally, these models have tended to be complicated, elaborate, and needed an enormous amount of data to be validated and used. However, EPIC and CropSyst are the most famous and practical ones. It was also found that complex models didn't always give more accurate results.
In the next section we will look at EPIC and CropSyst.
2.1. EPIC The Erosion-Productivity Impact Calculator (EPIC) model was developed to assess the effect of soil erosion on soil productivity. It was used for that purpose as part of the 1985 RCA (1977 Soil and Water .Resources Conservation Act) analysis. Since the RCA application, the model has been expanded and refined to allow simulation of many processes important in agricultural management. The model requires input from GRASS GIS layers. These include soil series and weather data, although the model can generate the necessary weather parameters. The model also requires management information that can be input from a text file. Currently, there are may management files that exist for EPIC and an effort is underway to catalogue these files and provide them to users. The model provides output on crop yields, economics of fertilizer use and crop values. EPIC is a continuous simulation model that can be used to determine the effect of
151
management strategies on agricultural production and soil and water resources. The drainage area considered by EPIC is generally a field-sized area, up to 100 ha. The major components in EPIC are weather simulation, hydrology, erosion-sedimentation, nutrient(*28) cycling, pesticide fate, plant growth, soil temperature, tillage, economics, and plant environment control. Recently, most of the EPIC model development has been focused on problems involving water quality and global climate/C02chang:e. Example applications include: (1) 1985 RCA analysis; (2) 1988 drought assessment; (3) soil loss tolerance tool; (29) Australian sugarcane model (AUSCANE); (5) pine tree growth simulator; (6) global climate change analysis (effect of C02, temperature, and precipitation change on runoff and crop yield); (7) farm level planning; (8) five-nation EEC assessment of environmental agricultural policy alternatives; (9) Argentine assessment of erosion/ productivity; (10) USDA-Water Quality Demonstration Project
Evaluation; (11) N leaching index national analysis.
2.1.1. ECONOMIC COMPONENT OF EPIC The economic component of EPIC is more accurately represented as a crop budget and accounting subsystem. Costs (and income) are divided into two groups: those costs which do not vary with yield and those that do. These groups will be addressed in turn. .All cost registers are cleared at harvest. All operations after harvest are charged to the next crop in the cropping sequence. Tillage and (pre-harvest) machine operation costs are assumed to be independent of yield. These operation costs must be calculated outside of EPIC and are inputted as one variable into the tillage file. This cost cell contains all costs associated with the single operation or activity (e.g., a chiseling activity includes fuel, labor. depreciation, repair, interest. etc.. for both the tractor and the chisel). A budget generator program like the Micro Budget Management System (MBMS) is convenient for making these calculations. This is an updated interaction program developed from the Enterprise Budget Generator. The MBMS is more compatible with EPIC in that it has output capabilities
28
The elements C, H, O, P, K, N, S, Ca, Mg, K, B, Mn, Cu, Zn, Mo, Cl, Co, Si and F which are required
for plant growth. 29
Any substance or mixture intended for preventing, destroying, repelling, killing, or
mitigating problems caused by any insects, rodents, weeds, nematodes, fungi, or other pests; and any other substance or mixture intended for use as a plant growth
152
to itemize cost by machine operation. This information (when converted to metric units) can be inputted directly into the equipment file in EPIC. Farm overhead, land rent, and other fixed costs can be charged to the crop by first creating null operations in the equipment file with machine number and cost information only and then triggering the cost in EPIC with a null activity. Government payments can be credited by using negative cost entries in the same way. Costs which are yield and management dependent are entered into EPIC in two regions of the input data. Seed costs, seeding rates, and crop prices are entered in the crop parameter file for each crop code. Seed costs are calculated as the produce of seeding rate and cost per kilogram. Amendment costs are calculated similarly. The amendments include elemental N and P, irrigation water, and lime. Total cost per hectare is based on the sum of costs for machinery operations, seed, and amendments. Market value per hectare is based on the product of crop yield and net crop price. Net crop price is the market price minus the harvest, hauling, and other processing costs which are yield dependent. The net price must be determined outside EPIC. When valid cost figures are entered into these EPIC input
cells. the model will return annual cost and returns by
crop. EPIC budget information is valuable not only for profit analyses but also risk analyses, since the annual distributions of profits and costs can be captured. Risk analyses capability greatly enhances the analytical value of EPIC for economic studies. The greatest value of EPIC to economic analyses is not its internal economic accounting, but the stream of physical outputs on daily, monthly, annual, or multi-year periods that can be input into economic models, budget generators, and risk analysis systems. EPIC estimates crop yields, movement of nutrients and pesticides, and water and sediment yields. Changes in inputs necessary to respond to changes in management, soil quantity and quality, climate (i.e. global warming), droughts, etc. are also estimated. These outputs become inputs into economic and natural resource models facilitating comprehensive analyses of alternative policies and programs. However EPIC has no GUI (Graphical User Interface), that is. it does not run over Windows operating system instead it does run over DOS operating system which is a very old computer operating system and therefore, it is a bit difficult to work with EPIC. In contrast, CropSyst is a computer program Windows 95/98/NT compatible and therefore it is used widely. In the next section we will look at CropSyst.
153
154
2.2 CropSySt CropSyst (Cropping Systems Simulation Model) is a multi-year, daily time step crop growth simulation model, developed with emphasis on a friendly user interface and GUl, and with a link to GIS(30) software and a weather generator (Stockle, 1996). Link to economic and risk analysis models is under development. The model’s objective is to provide an analytical tool to study the effect of cropping systems management on crop productivity and the environment. For this purpose, CropSyst simulates the soil water budget, soil plant nitrogen budget crop phonology(31), crop canopy and the root(32) growth biomass(33) production crop yield residue production and decomposition(34), soil erosion by water, and pesticide(35) fate. These are affected by weather, soil characteristics, crop characteristics, and cropping system management options including crop rotation, cultivator(36) selection, irrigation, nitrogen fertilization, pesticide applications, soil and irrigation water salinity, tillage operations, and residue management(37). The model code is written in Pascal (DOS version) and C++ (Windows and Windows 95 versions). An advanced user-friendly interface allows users to easily manipulate input files, verify input parameters for range errors and cross compatibility, create simulations, execute single and batch run simulations, customize outputs, produce text
30
(Geographic Information System) A set of tools for collecting, storing, retrieving,
transforming and displaying spatial referenced data. 31
Phonology is the study of the response of living organisms to seasonal and climatic changes to the
environment in which they live 32
The underground portion of a plant that serves as support draws food and water from
the surrounding soil, and stores food. 33
The weight of living organisms (plants and animals) in an ecosystem at a given point in time,
expressed either as fresh or dry weight. 34
The process by which materials are broken down into simpler compounds by decomposers.
35
Chemicals that kill organisms that are injurious to man or to the crops and animals upon which he
depends for food, fiber, and shelter. These organisms include insects, mites, microorganisms, weeds, and rodents. Pesticides include insecticides, fungicides, herbicides, and others. 36
Farm implement or machine designed to stir the soil around a crop as it matures to promote growth
and destroy weeds 37
Management of rice straw and stubble after harvest.
155
and graphical reports, link to spreadsheet programs, and even select a preferred language for the interface text. Simulations can be customized to invoke only those modules of interest for a particular application (e.g., erosion and nitrogen simulation can be disabled if not desired), producing more efficient runs and simplifying model parameterization. CropSyst executable program manual and tutorials can be retrieved directly over Internet (38). . The model is intended for crop growth simulation over a unit field area (m2). Growth is described at the level of whole plant and organs. The water budget in the model includes, precipitation, irrigation, run of(39), interception, water infiltration(40) water redistribution in the soil profile (41), crop transpiration(42) , and evaporation. Users may select different methods to calculate water redistribution in the soil profile and reference (43), evapotranspiration. Water redistribution in the soil is handled by a simple cascading approach or by a finite difference approach to determine soil water fluxes. The latter allows accounting for upward flow (and chemical transport) from a water table(44) whose depth from the soil surface needs to be specified over time). CropSyst offers three options to calculate grass reference ET (evapotranspiration). In decreasing order of required weather data input, these options are the Penman-Monteith model, the Priestley-Taylor model, and a simpler implementation of the Priestley-Taylor model, which only requires air temperature. Crop ET is determined from a crop coefficient at
38
http://www.cahe.wsu.edu/～bsyse/faculty/stockle/CropSyst/cropsyst.htrnl
39
The portion of the total precipitation on an area that flows away through stream channels. Surface
runoff does not enter the soil. Groundwater runoff or seepage flow from groundwater enters the soil before reaching the stream. Runoff occurs when the rainfall rate exceeds the soil's infiltration capacity. On sloping areas, runoff is a concern since it can carry soil particles, nutrients, and other chemicals with it. 40
The process whereby water enters the soil through the surface.
41
A vertical cross-section of a given soil, showing the different layers or horizons, if present.
42
The evaporation of water vapor from plants, mostly through stomata.
43
The loss of soil moisture due to evaporation from the soil surface and transpiration by plants.
44
The upper limit in the soil or underlying material permanently saturated with water.
156
full canopy and ground coverage determined by canopy(45) leaf area index(46). The nitrogen budget in CropSyst includes N transformation, ammonium sorption, symbiotic fixation(47), crop N demand and crop N uptake. Nitrogen transformations of net mineralization(48) , nitrification(49) and denitrification(50) are simulated. The water and nitrogen budgets interact to produce a simulation of N transport within the soil. Chemical budgets (pesticides, salinity(51)), including pesticide decay and absorption, are also kept and interact with the water balance. A11 balances within the model are reported in case of departures within set threshold values. Crop development is simulated based on thermal time (GDD(52)) required to reach specific growth stages. The accumulation of thermal time may be accelerated by water
45
The leafy parts of vines or trees.
46
The ratio between the total leaf surface area of a plant and the surface area of ground that is
covered
by the plant. Leaf area index (LAI) is widely used to describe the photosynthetic and transpirational surface of plant canopies. LAI can also be simply defined as the amount of leaf surface area per unit ground area. 47
The process or processes in a soil by which certain chemical elements essential for
plant growth are converted from a soluble or exchangeable form to a much less soluble or non exchangeable form, for example, phosphate fixation. 48
The conversion of an element from an organic form to an inorganic state as a result of microbial
decomposition. The process by which organic residues in the soil are broken down to release mineral nutrients that can be utilized by plants. 49
The biochemical oxidation of ammonium to nitrate.
50
The gaseous loss of nitrogen by either biological or chemical mechanisms, but exclusive of ammonia
volatilization. 51
The amount of soluble salts in a soil, expressed in terms of percentage, parts per million, or other
convenient ratios. 52
Growing degree days (GDD) are used to estimate the growth and development of plants and insects
during the growing season. The basic concept is that development will only occur if the temperature exceeds some minimum developmental threshold, or base temperature (Tb). The base temperatures are determined experimentally and are different for each organism.
157
stress. Thermal time may be also modulated by photoperiod(53) and vernalization (54) requirements whenever pertinent. Daily crop growth is expressed as biomass increase per unit ground area. The model accounts for four limiting factors to crop growth: water, nitrogen, light, and temperature. Four input data files are required to run CropSyst. These are; Location, Soil, Crop and Management files. Separation of files allows for an easier link of CropSyst simulations with GIS software. A Simulation Control file combine the input files as desired to produce specific simulation runs. In addition, the Control file determines the start and ending day for the simulation, defines the crop rotations to be simulated, and set parameters requiring initialization. The location file includes information such as latitude, weather file code name and directories, rainfall intensity parameters (for erosion prediction), freezing climate parameters (for locations where soil might freeze), and local parameters to generate daily solar radiation(55) and vapor pressure deficit values. The soil file includes surface soil Cation Exchange Capacity（56） and pH(57), required for ammonia volatilization( 58 ), parameters for the curve number approach (runoff
53
Period of a plant's daily exposure to light. In other words, the total number of hours of daylight.
54
The process in which a seed is subjected to a period of cold, causing changes that allow germination to
occur. In other words, treatment of germinating seeds or seedlings with low(or high) temperatures to induce flowering at maturity. 55
Measure of the intensity of the sun's radiation reaching the earth's surface.
56
The total amount of exchange able cations that a soil can adsorb.
57
A value used to express relative acidity or alkalinity. The logarithm of the reciprocal of the
concentration of hydrogen ions in a medium, Like water or soil (log10{ l/[H+] })･ pH values range from O to 14, giving the relative acidity or alkalinity of a medium, with a pH of 7 being neutral, and lower values being acidic, higher values, alkaline. pH values are the negative logarithm of the hydrogen ion concentration of a soil solution. The degree of acidity or alkalinity of a soil expressed in terms of the pH scale, from 2 to 10. 58
Gaseous loss of a substance to the atmosphere.
158
calculation), surface soil texture( 59 ) (for erosion calculation), and five parameters specified by soil layer: Layer thickness, Filed capacity. Permanent Wilting point(60), Bulk density(61), and Bypass coefficient. The latter is an empirical parameter to add dispersion to solute transport. particularly when using the cascading approach for soil water redistribution. The Management file includes automatic and scheduled management events. Automatic events (irrigation and nitrogen fertilization) are generally specified to provide optimum management for maximum growth, although irrigation can also be set for deficit irrigation. Management events can be scheduled using actual date, relative date (relative to year of planting), or using synchronization with phonological events (e.g., number of days after flowering). Scheduled events include irrigation (application date, amount, chemical or salinity content), nitrogen fertilization (application date, amount, source-organic and inorganic, and application mode-broadcast, incorporated, injected), tillage operations (primary and secondary tillage operations, which are basically related to residue fate), and residue management (grazing, burning, chopping, etc.) The crop file allows users to select parameters to represent different crops and crop cultivars using a common set of parameters. This file is structured in the following sections; Phonology (Thermal time requirements to reach specific growth stages, modulated by photoperiod and vernalization requirements if needed), Morphology (Maximum LAI, root depth, specific leaf area and other parameters defining canopy and root characteristics), Growth (transpiration-use efficiency normalized by VPD, light-use efficiency, stress response parameters, etc.). Residue (decomposition and shading parameters for crop residues). Nitrogen Parameters (defining crop N demand and root uptake).
Harvest Index (unstressed harvest index and stress sensitivity parameters),
and Salinity Tolerance. CropSyst includes the simulation of Crop development and growth, Water and nitrogen
59
The relative proportions of the various soil separates in a soil. In other words, Soil texture is a
description of the relative amounts of clay, sand and silt present in the soil. Soil texture determines the amount of water a soil can make available for plant growth. 60
The moisture content of a soil at which plants (specifically sunflower plants) wilt and fail to recover
their turgidity when placed in a dark, humid atmosphere. The wilting point is commonly estimated by measuring the 1 5-bar percentage of a soil. 61
The mass of dry soil per unit bulk volume. The bulk volume is determined before the soil is dried to
constant weight at I 05' C. It has been called apparent density.
159
balance, Salinity, Residue fate, Soil erosion by water, CO2 effects and Pesticide fate. We can perform all of these simulations by using CropSyst. In addition, CropSyst give us the opportunity for some management practices. These management options are crop rotation (including fallow and multi-year forage crops), irrigation (fresh and saline water), nitrogen fertilization (mineral and organic sources), residue management and tillage. Model validation is a necessary requirement for model application. In order to do a reliable validation, several steps must be taken, and each of them may be a source of errors which will influence the final result.
160
Table 2. Necessary Inputs for CropSyst Necessary Inputs
Outputs
Meteorology
Growth
Precipitation
yield, nitrate
Temperature min.
pollution, erosion. etc.
Temperature max. Solar Radiation Management & Production dates & types of : Rotation Planting Tillage Irrigation Yield Etc. Soil Composition of soil physiques
and
hydrologic
properties Crop Phenologic et biologic data Location Data of location Latitude etc.
161
stages
of
crop,
Experimental data Model(theory)
Parameter
Model(program)
Inputs
Simulations
Experimental data
Validation Figure 1. Validation in CropSyst As a general rule, if there are discrepancies between observed and simulated data, the technical structure of a model should be the last factor to suspect. Validation errors means; 1. Model inadequate,
2. Lack of calibration, 3. Errors in code, 4. Errors in input,
5. Error in the use (the model does not simulate a key process for the environment under study), 6. Error in the experimental data.
Simulation File Soil Crop File Location File Management File Weather File Figure 2. File structure of CropSyst Calibration is an important part of CropSyst simulation models. Before starting the calibration of crop parameters we should check the quality of out weather data to which the model CropSyst is very sensitive. The calibration of crop phonology is a concept related to GDD. The first step to take is the correct definition of the crop cycle. Accumulated CDD must allow matching emergence and physiological maturity dates of crop.
These are; a. Planting date, b. Emergence date, c. Flowering
date, d. Max LAI date, e. Maturity date, f. Harvest date.
162
It is very important for most species to exactlv match flowering dates. In addition, accumulated GDD must be calculated using meteorological data of 3 to 4 versions at least. The calibration of crop morphology is another part of calibration of simulation models. The key parameters to characterize the morphology of the canopy in CropSyst are the maximum root depth and the light extinction coefficients (it relates to orientation of leaves).
163
3. Calibratin of CroSyst 3.1
CALIBRATION : CROP PHENOLOGY
Before starting phenology calibration of the simulation of our crops, we will make an example for this calibration process by using the crop of maize Suppose that we have the real date if growth stages for maize cultivation as follows (MM.DD): Planting date (05.15), Emergence date(05.23), Max LAI(08.10), Begin Flowering(08.04), Grain Filling(08.10), Maturity(09.28)
We are running our simulation dates and we get the following 5imulation dates Planting date(05.01), Emergence date(06.02), Max LAI(07.2 l), Begin Flowering(07.15), Grain Filling(07.2 l), Maturity(08.13) The comparison of simulation data and actual data is given in Table 19. Table 19. Observed and simulated data (1) Data Fitting
Actual Data
Simulated Data
Fitting
(Month. Day)
(Month. Day)
Planting date
05.15
05.01
No
Emergence date
05.23
06.02
No
Max LAI
08.10
07.21
No
Begin Flowering
08.04
07.15
No
Grain Filling
08.10
07.21
No
Maturity
09.28
08.13
No
None of the dates are fitting each other. We want to calibrate our model so that simulated model date fits to the actual one. Firstly we should change the planting date. We are changing “Planting date” for our crop by using the “Rotation" menu of CropSyst in the simulation file. For the calibration of with stages dates of a crop we need to use the concept of GDD (62). Tx is minimum temperature, Tn is base temperature for our crop (Tb for maize is of 8) We should calculate the following values: emergence 62
Growing Degree Days
164
Σ GDDi → I=planting flowering Σ GDDi → I=planting grain filling Σ GDDi → I=planting Max LAI Σ GDDi → I=planting maturity Σ GDDi → I=planting
Accumulated GDD up to emergence data Accumulated GDD up to flowering data Accumulated GDD up to grain filling data Accumulated GDD up to Max LAI data Accumulated GDD up to maturity data
In order to make the calculations, we are creating an Excel file in order to calculate accumulated GDD values for the corresponding growth stages dates by using the Weather file in which there are meteorological data as maximum and minimum temperature values. That is_ we are opening our weather file for the year that we want to make the calibration by using Excel In the following table we see an illustration for this weather file. We are writing the following formula in F2 cell in order to compute the GDD: =IF(((D2+E2)/2 - 876) >0; (D2+E2)/2 – 8;0) By using this formula we are calculating GDD for all the days of the year. To the G138 cell we are writing the following for formula to calculate the accumulated GDD for all days of year =F138+G137 We arc looking the GDD values for the dates that we actually have for maize At emergence date (05 23),
accumulated GDD is 46 (cell G144)
At Max LAI date {08.10),
accumulated GDD is 883 (cell G208)
At Begin Flowering date(08.04),
accumulated GDD is 1006 (cell G2 1 7)
At grain filling date (08.10),
accumulated GDD is lOS6 (cell G222)
At maturity date (09 2S).
accumulated GDD is 1753 (cell G27 1 )
We arc noting these values as above. Then we are opening crop file, and we select "Phenology" option
Figure 44. CropSyst Crop file phenology form for maize example.
165
That is, we are changing the default values with the values that we get above and then we are saving the file. We run the simulation again, we get the harvesting results given in Table 20 Table 20. Observed and simulated data (2). Dates
Actual Data
Simulated Model 1
Sim.Modcl 2
(Month.Day）
(Month.Day)
(Month.Day)
Planting date
05.15
05.01
05.15
Max LAI
05.23
06.02
05.23
Begin Flowering
08.04
07.15
08.10
Grain Filling
08.10
07.21
08.10
Maturity
09.28
08.13
09.28
As we can see all the growth stages dates of maize are the same with the real ones. Thus the model is calibrated in terms of phenology. These are the basic steps in order to perform the phenology calibration Now we will make the phenologv calibration for our simulation crops which are wheat sugar beet and barley in the following section
3.1.1. Sugar Beet Note that Tb is of 1 for 5ugar beet- You can obtain that value by loading default values for sugar beet in the crop file of CropSyst. We have the real data for sugar beet cultivation as follows (M~1 DD): Planting date
04.01
Emergence date 04.10 Maturity date
10.01
We are running our 5imulation file by using the crop file for sugar beet and we get the following dates Planting date
04.01
Emergence date 05.02 Maturitv date
09.13
The comparison of simulation data and actual data is given in Table 21.
166
Table 21. Observed and simulated data (3). Dates
Actual Data
Simulated Model 1
Sim.Modcl 2
(Month.Day）
(Month.Day)
(Month.Day)
Planting date
04.01
04.01
Yes
Emergence date
04.10
05.02
No
Maturity
10.01
09.13
No
We see that the 2ctual emergence date and maturity dates do not fit to the real (observed) ones. The model should be calibrated in terms of these dates so that simulated model dates fit to the real (observed) ones. As we have done in the example, we should calculate the accumulated GDD values of sugar beet for the year that we want to make the calibration. Having calculated the accumulated GDD values, we are noting corresponding GDD values for the emergence and maturity dates At emergence date,
accumulated GDD is 70
(cell G92)
At maturity date,
accumulated GDD is 3000 (cell G272)
We are opening crop file for sugar beet that we prepared before, and we select "Phenology" option. Figure 45. CropSyst Crop file phenology form for sugar beet. We are changing the default values with the values that we calculated above and then we are saving the crop file of CropSyst for sugar beet with these new values. We are running the simulation again by using the crop file for sugar beet that we saved with new GDD values. and we get the simulated harvesting results given in Table 22.
Table 22. Simulation results (1).
Table 23. Observed and simulated Dates
data(4).
Actual Data
Sim.Model 1
Sim. Model 2
(Month.Day)
(Month.Day)
(Month.Day)
planting date
04.01
04.01
04.01
Yes
Emergence date
04.10
05.02
04.10
Yes
Maturity
10.01
09.13
10.01
Yes
167
Fitting
All the dates are fitting each other. thus the model is calibrated in terms of phenology.
3.1. 2. Wheat Note that Tb is of 0 for wheat (winter). You can obtain that value by loading default values for wheat (winter) in the crop file of CropSyst. We have the real (observed) data for wheat cultivation as follows (MM.DD). Planting date (10.07), Emergence date (10.30), Begin Flowering (05.20), Max LAI(06.01), Grain Filling(06.01), Maturity (07.08) We are running our simulation file by using the crop file for wheat and we get the following dates: Planting date (10.07), Emergence date(10.15), Begin Flowering(06.06), Max LAI(06.16), Grain Filling(06.16), Maturity(07.03). The comparison of simulation data and actual data is given in Table 24 Table 24. Observed and simulated data (5) Dates
Actual Data
Simulated Data
Fitting
(Month. Day)
(Month. Day)
Planting dale
10.07
10.07
Yes
Emergence data
10.30
10.15
No
Max LAI
06.01
06.16
No
Begin Flowering
05.20
06.06
No
Grain Filling
06.01
06.16
No
Maturity
07.08
07.03
No
We see that the observed emergence, maturity, Max. LAI, flowering, grain filling and maturity date do not fit to the real (observed) ones. The model should be calibrated in terms of these dates so that simulated model dates fit to the real (observed) ones. As we have done in the example, we should calculate the accumulated GDD values of wheat for the year that we want to make the calibration. Having calculated the accumulated GDD values., we are noting corresponding GDD values for emergence. Max LAI, flowering grain filling and maturity dates.
168
At emergence date,
accumulated GDD is 240
At Max LAI date
accumulated GDD is 1565
At Begin Flowering date,
accumulated GDD is 1365
At grain filling date accumulated GDD is 1565 Al maturity date accumulated GDD i5 2300 We are opening crop file for wheat that we prepared before and we select "Phenology" option. Figure 46. CropSyst Crop file phenology form for wheat. We are changing the default values with the values that we calculated above and then we are 5aving the crop file of wheat with these new values. We are running the simulation again by using the crop file for wheat that we saved with new GDD values, and we get the simulated
harvesting results given in Table 25.
Table 25. Simulation results (2]. Planting
Emergence
Flowering
Grain
filing
Dates
Date
Date
Date
1996/10/07
1996/10/30
1997/05/20
1997/06/01
Peak LAI
Maturity Date
1997/06/01
1007/07/06
The comparison of simulation data and actual data is given in Table 26. Table 26. Observed and simulated data (6) Dates
Actual Data
Simulated Model 1
Sim. Model 2
(Month. Day)
(Month. Day)
(Month. Day)
Planting date
10.07
10.07
10.07
Emergence date
10.30
10.15
10.30
Max LAI
06.01
06.16
06.01
Begin Flowering
05.20
06.06
05.20
Grain Filling
06.01
06.16
06.01
Maturity
07.08
07.03
07.08
All the dates are fitting each other. thus the model is calibrated in terms of phenology.
169
3.1.3
Barley
Note that Tb is of 0 for wheat barley, You can obtain that value by loading default values for barley in the crop file of CropSyst. We have the real (observed) data for barley cultivation as follows (MM.DD) Planting date(10.07), Emergence date(10.28), Begin Flowering(04.30) Max LAI(05.12), Grain Filling(05.12), Maturity(06.28) We arc running our simulation file and we get the following dates Planting date (10.07), Emergence date (10.15), Begin Flowering (06.06) Max LAI (06.16), Grain Filling (07.03), Maturity (07.03) The comparison of simulation data and actual data is given in Table 27. Table 21. Observed and simulated data (7). Dates
Actual Data
Simulated Model 1
Fitting
(Month. Day)
(Month. Day)
Planting date
10.07
10.07
Yes
Emergence date
10.28
10.15
No
Max LAI
05.12
06.16
No
Begin Flowering
04.30
06.06
No
Grain Filling
05.12
06.16
No
Maturity
06.28
07.03
No
We see that the observed emergence, maturity, Max LAI, flowering, grain filling and maturity date do not fit to the real (observed) ones. The model should be calibrated in terms of these dates so that simulated model dates fit to the real (observed) ones. As we have done in the example we should calculate the accumulated GDD values of barley for the year that we want to make the calibration. Having calculated the accumulated GDD values_ we are noting corresponding GDD values for emergence. Max LAI, flowering. grain filling and maturity dates. These values are as following, At emergence date,
accumulated GDD is 227;
At Max LAI date,
accumulated GDD is 530;
At Begin Flowering date, accumulated GDD is 440. At grain filling date,
accumulated GDD is 530;
At maturity date,
accumulated GDD is 1385.
We are opening crop file for barley that we prepared before. and we select “Phenology" option. Figure 47. CropSyst Crop file phenology form for barley.
170
We are changing the default values with the values that we calculated above and then we are saving the crop file of barley with these new values We are running the simulation again by using the crop file for barley that we saved with new GDD value5 and we obtain the simulated harvesting results given in Table 28
Table 28. Simulation results (3). Planting
Emergence
Flowering
Grain filling
date
Date
Data
Date
1996/10/07
1996/10/28
1997/04/30
1997/05/12
Peak LAI
Maturity Date
1997/05/12
1997/06/28
The comparison of simulation data and actual data is given in Table 29. Table 29. Observed and simulated data (8). Dates
Actual Data
Simulated Model 1
Simulated Model 2
(Month. Day)
(Month. Day)
(Month. Day)
Planting date
10.07
10.07
10.07
Emergence date
10.28
10.15
10.28
Max LAI
05.12
06.16
05.12
Begin Flowering
04.30
06.06
04.30
Grain Filling
05.12
06.16
05.12
Maturity
06.28
07.03
06.28
All the dates are fitting each other. thus the model i5 calibrated in terms of phenology.
171
3. 2. Calibration: Crop Yield 3. 2. 1. Water Use In order to calibrate water use, we used the concept of water stress index that CropSyst give us in the daily output report after the crop simulation. The irrigation dates and amounts for our crops are given in Table 30 Table 30. Irrigation dates and amounts for Sugar best simulation. Sugar beet Irrigation
dates
Amount mm
(Montb. Day) 4.15
60.000 mm
5.15
100.180 mm
6.19
86.180 mm
7.3
86.180 mm
7.17
86.180 mm
81
86.180 mm
8.9
86.180 mm
8 19
86.180 mm
8.28
86.180 mm
9.10
86.180 mm
Table 31. Irrigation dates and amounts for Wheat (irrigation) simulation Wheat (irrigated) Irrigation dates
Amount (mm)
(Month. Day) 4.15
110.870 mm
5 13
127.530 mm
10.16
112.720 mm
172
Table 32. Irrigation dates and amounts for Barley (irrigated) simulation. Barley(irrigated) Irrigation dates
Amount (mm)
(Month. Day) 4.08
114.790 mm
4.24
106.400 mm
10.14
82.680 mm
3.2.2. Nitrogen Use For nitrogen use we used the automatic nitrogen option of CropSyst for optimum fertilization for each crop. If we do not use this option we can make the calibration of nitrogen use by using the nitrogen, stress value given in the daily output report We should determine the nitrogen application dates and amounts by means of thi5 nitrogen stress index
3. 2. 3. CROP YIELDS We run our model and we get the simulation results for crop yields. In order to 5ay that our model represents reality we should look at the simulated and observed .yields of our crops. Simulated and observed actual yields are given in Table 33 Table 35. Observed and simulated yields (9). Crop
Observed Yield (kg/ha)
simulation Yield (kg/ha)
Sugar Beet
673 5.0
6790.0
Wheat (irrigated)
4200.0
41 20.6
Barley (irrigated)
3 500
3493.7
Then our crop simuiation5 are calibrated both in terms of phenology and Yields. If we look at our model all the necessary growth stage dates for our crops, water use5, nitrogen uses and crop yields are fitting to the observed ones. This was the our aim in order to get a calibrated bio-physical model representing reality
173
4. Conclusion We have constructed and calibrated a bio-physical model By using this model, we can obtain all the neces5ary technical coefficients that we need in order to construct a bio-economical model We can also use this model to obtain engineering production functions for our crops. In many cases economical models needs technical coefficients and production functions. As an example, we can give an agricultural irrigation water pricing model. Suppose that we want to prevent the waste in the use of water because of irrigation in agricultural production. We want to look at the possible effects of some water pricing methods in the total welfare of society, productivity, and yields etc. In each water pricing methods, because of changing cost restriction functions of farmers maximizing their profits, farmers will use different amount of water as irrigation. However, changes in the amounts of irrigation water use will cause changes in yields of crops. Therefore, for a linear programming model constructed in order to look at the possible effects of different water pricing policies, the needed technical coefficients for the crop productions can be obtained by our bio-physical model YIM. Our model representing the real production conditions and situations will
give us the different
yield levels for different water use amount because of different water pricing policies. If farmers are using less water resulting from, for example, an increasing block water pricing policy, the yields will diminish. By using our bio-physical model we can obtain these technical coefficients and take into account the less water use effect on agricultural production Both EPIC and CropSyst give us the opportunity to generate these needed coefficients. The economic models, which use the bio-physical model data, are called bio-economical models. These models are generally descriptive models. By means of these kind of models we can look at the policy changes effects on the agricultural productions. Our YIM model is generated for Konya-Karapinar region of Turkey for wheat (winter), barley and sugar beet. This bio-physical model can be a part of water pricing policy bio-economical model. We constructed a bio-physical model what we called YTM and we did the calibration of this model. These kinds of bio-physical models are very u5efuL in agricultural economical analysis for the reasons stated before. The use of bio-physical models have an increasing trend in agricultural economics and give us the opportunity to represent the reality and the opportunity of looking at effects of possible agricultural policy changes.
174
BIBLIOGRAPHY Blanco. M, Sumpsi, J.M., Ortega, C., Garrido, A., Iglesias, E ( 199S). "Water Pricing Policies, Public Decision Making and Farmer’s Response Implications for Water Policy”, Agricultural Economics, 19 ( 1-2), pp. 193- 202. Blanco, M. (1999). La Economia Del Agua: Analysis De Politicas De Modernization Y Mejora De Regadio. En Espana, Thesi s Doctoral. Universidad Politechnica De Madrid. Madrid. B]anco, M (1995). Analysis des impacts socio-economiques et des effects sur l’environment des politiques agricoles : modelisation de l’utilisation agricole des resources en eam dans la region espagnole de Castille-Leon, These Master of Science, Institute Agronomique Mediterraneen de Montpellier, Montpellier. Boussard. J_M. (1988). Economie de l'Agriculture, Paris, Economica. Boussemart, J.P.,
Flichman, G., Jacquet, F., Lefer, H. B.(1994). “L’evaluation
micro-economiques de la reform de la PAC par un modele bio-economique: une illustration sur la Beauce et la region de Toulouse”, Communication presentee dans les IIemes Journee de Microeconomie Appliquee, Marseille. Bryant, K. J., Benson, V. W., Kiniry, J. R., Williams, J. R., and Lacewell, R. D. (1992), “Simulating Corn Yield Response to irrigation Timing Validation of the Epic Model”, J. Prod. Agr., 5 (April-June 1992), pp. 237-42. Bryant. K. J., Mjelde, J. W. (1993). "An Intraseasonal Dynamic Optimization Model to allocate irrigation Water between Crops”, American Journal of Agricultural Economics, 75, Issue 4, p. 1021. Chapra, S C . Boyer, J.M and Runkel. R_L. (1993), “Advanced decision support systems for environmental simulation modeling,” The 20th Anniversary Conference on Water Management in the 90’s, Seattle, WA. Chenery, H.B (1949). “Engineering Production Function”, Quaterly Journal of Economics, 63, pp. 507-531.
175
Stockle, C. O., Martin, S. A. and Campbell, G. S. (1994) “CropSyst, a cropping system simulation model: Water/nitrogen budgets and crop yield”, Agricultural Systems, 46, pp. 335-359. Cole, G.W.. Lyles, L., and Hagen. L., J. (1982). A simulation model of daily wind erosion soil loss". 1982
ASdE Winter Meeting, Paper No. 82-2575.
Cummings. R. G.. Nercissiantz, V. (1992), “The Use of Water Pricing as a Means for Enhancing Water-Use Efficiency in Irrigation- Case Studies in Mexico and United States". Natural Resources.Journal Vol 32. Iss 4, pp 73l- 755. Cakmak. E (1997), “Water Resources in Turkey Availability , Use, and Management”, in D. D. Parker, Y. Tsur, Decentralization and Coordination of Water Resource Management, Kluwer Academics Publishers, Massachusetts. pp. 33-44. Doorenbos, J. and Kassam, A.H. ( 1986). “Yield response to water". Irrigation and Drainage Papers 33, Food Agri. Org. United Nations. Rome. Droesbeke, F., Hallin, M., Lefevre, CL (1986). Programmation Lineaire Par l’Exemple, Paris, Eliipses Ertas, R. (1980). Konya Ovasinda Bugdayin ve Seker Pancarinin Lizimetrede Saptanan Su tuketimleri, Konya Topraksu Arastirma Enstitusu Mudurlugu Yayinlari, Konya. Eruygur, O. (1998). Comparison de Differentes Politiques de Tarification de I’eam: Cas de Eskisehir, Projet de These Master of Science, Institute Agronomique Mediterraneen de Montpellier(IAMM), Montpellier. Easter. K_ W.. Feder. G_ (1994). Water Institutions, Incentives and Markets, Department of Agricultural and Applied Economics. University of Minnesota. Minnesota Faucheux, S., Noel, J F. Economie des resources naturelles et de l’environment, Armand Collin, Paris. Flichman, G.(1995). Bi0-Economic Models, Integrating Agronomic, Environmental and
176
Flichman, G.(1990), International Comparisons of Efficincy in Agritulture, OECD Development Center, No. 21, Paris. *.. Flichman, G., Garrido. A., Varela Ortega, C (1994), "Agricultural policy and technical choice: a regional analysis of income variation, soil use and environmental effects under uncertainty and market imperfections", in Environmental and land use issues: an economic perspective. Seminaire de l'EAAE (European Association of Agricultural Economists), Zaragoza. Fuentes, M. A. (1998). Leseffets de la reform de la politique agricole commune (PAC) sur l’agriculture espagne:
analyse du cas de la Catalogne, Institut Agronomique
Mediterraneen de Montpellier, Montpellier. Gokcora, H_ (1969), Bilki Yatistirme ve Islahi, Ankara Universitesi Ziraat fakultesi Ya Yinlari, Yayin No: 366,Ankara. Hansen. S., Jensen. H.E., Nielsen, N.E., & Svendsen. H (1991), "Simulation of nitrogne dynamics and biomass production in winter wheat using the Danish simulation model DAISY”, Fertilizer Research, 27, pp 245-259. Hazell, P. B. R., & Norton, R.D. (1986), Mathematical Programming for Economic Analysis in Agriculture, MacMillan Publishing Company, New York. Kasnakoglu, H., Ok. E. (1993). Principles of Linear Programming, Middle East Technical University,
Faculty of Economic and Administrative Sciences, Ankara.
Kaufmann, R.K., Snell, S.E. (1996) "A Biophysical Model of Corn Yield Integrating Climatic and Social Determinants", American Journal of Agricultural Economics, 78, pp. 1082-1091. Koy Hizmetleri Genel Mudurlugu (1988). Tiirkiye'de Uretile'l Tarim Urunlerinin Uretim Girdileri 58, Rehber No. 10,
ve Maliyetleri Rehberi, Koy Hizmetleri Genel Mtidurlugu Yayin No. Ankara.
Koy Hizmetleri Genel Mudurlugu (1992). Konya Ili Arazi Varligi, Koy Hizmetleri Genel
177
Mudurlugu Yayinlari, Ii Rapor No.42, Ankara. Koy Hizmetleri Genel Mudurlugu( 1 991 ). Konya ve Karaman Illeri Verimlilik Envanteri, TOVEP Yayin No. 60, Ankara. Koy Hizmetleri Genel Mudurlugu(1999), Toprak ve Su Kaynaklari Arastirma Yilligi 1998, Yarin No. 108, Ankara. Law, A. M., and Kelton, W.D (1991). Simulation Modeling and Analysis, McGraw-Hill Inc., New York. Louhichi, K. (1997), Utilisation d’un modele bio-economique pour analyzer I’impact des politiques agri-environmentales: cas des techniques de conservation des eaux et du sol en Tunisie. Application a la ferme Sawaf, Institut Agronomique Mediterraneen de Montpellier, Montpellier. Ministere de l’Agriculture du Canada, Direction de la Recherche (1976), Glossaire des Termes de la Science des Sols, Publication 1459. Canada. Plevne G. (1999), Une Comparaison des Politiques Environnemetales pour le Probleme d’Erosion: Cas de Beypazari (Turquie), These Master of Science, IAMM, Montpellier. Putman J., Dyke P. (1987), The erosion productivity calculator as formula for resource conservation act appraisal, US Department of Agricultural & Natural Resource Division (ERS), Texas. Sharpley, A. N. and Williams, J. R. (1990), EPIC (Erosion Productivity Impact Calculator) 1. Model Documentation, U. S. Dept. Agric. Tech. Bull. No. 1768. Soganci, A., Demiryurek M. (1996), Konya ille Karapinar ilcasi Tarim isletmelerimin Ekonmik Analizi ile Optimal Isletme Organizasyonlari ve Yeterli Gelirli Isletme Buyuklugu, Koy Hizmetleri Genel Mudurlugu Konya Arastirma Enstitusu Mudurlugu Yayinlari, Konya.
178
Steduto. P., Pocuca, V., Caliandro, A., Debaeke, P. (1985), “An evaluation of the crop-growth simulation submodel of EPIC for wheat grown in a Mediterranean climate with variable soil-water regimes". to be published in the European .Journal / of Agronomy, 1985. Talpaz, H., and Mjelde, I. W.
(19S8), “Crop irrigation Scheduling via Simulation-Based
Experimentation", Western, Journal of Agricultural Economics, 13, pp. 184-92. Tietenbcrg, T. (1988). Environmental and Resource Economics. Second edition, Boston Scott: Foresman Company. Tlili, M. (1995). Essai de simulation, d’impact des ,nuvelles mesures economiques (PASA 2) sur le sectour cerealier en Tunise. Etude de cas: Behaya, Institut Agronomique Mediterraneen de Montpellier, Montpellier. Tsur, Y., & Dinar, A. (May 1995), “Efficiency and Equity Considerations in Pricing and Allocating
Irrigation Water”, Policy Research Working Paper 1460, The World Bank.
Agricultural and Natural Resources Department, Agricultural Policies Division Tsur, Y., & Dinar, A. (1996), “The rekative efficiency and implementation costs of alternative methods for pricing irrigation water”, The World Bank Economic Review, 11(2), pp. 243-262. Van Evert, F. K., and Campbell, G. S. (1994), “CropSyst: A collection of object-oriented simulation models of agricultural systems”, Agron. Journal, 86, pp. 325-3317. Vicien C.(1989), Les modeles de simulation comme outil pour la construction de fonctions de production: une application a la mesure de l’efficacite de la production agricole, Montpellier. Institut Agronomique Mediterraneed de Montpellier, Motpellier. Williams, J. R., Jones, C. A. and Dyke, P. T. (1984), “The EPIC model and its application”, In Proc. ICRISAT-IBSNAT-SYSS Symp. On Minimum Data Sets for Agrotechnology Transfer, March 1983, Hyderabad, India. Pp. 111-121.
179
