check
Cotton2K Model | Plant Sciences and Genetics in Agriculture
Avishalom Marani

Prof. Avishalom Marani

In Memoriam
Plant Science and Genetics

IN MEMORY OF

Prof. Avishalom Marani

1924 - 2020

Cotton physiology and breeding; Moisture stress in field crops; Quantitative genetics in plant breeding; Heterosis in cotton.

Contact Us

 

Mailing Address:
The Robert H. Smith Institute of
Plant Sciences and Genetics
in Agriculture
Herzl 229, Rehovot 7610001, Israel

Administrator: 
Neomi Maimon 
Tel: 972-8-948-9251,
Fax: 972-8-948-9899,
E-mail: neomim@savion.huji.ac.il

Secretary of teaching program:
Ms. Iris Izenshtadt
Tel: 972-8-9489333
E-mail: Iris.Izenshtadt@mail.huji.ac.il

Director: 
Prof. Naomi Ori
Tel: 972-8-948-9605
E-mail: naomi.ori@mail.huji.ac.il

 

Cotton2K Model

Cotton2K Model version 4.0

A. Marani

Version history

The COTTON2K cotton simulation model is descended from GOSSYM-COMAX. Its main purpose was to make the model more useful for conditions of cotton production under irrigation in the arid regions of the Western USA. Since many changes have been made in the model, it has been given a new name: CALGOS (for CALifornia GOSsym). The present version, which is completely revised has been renamed COTTON2K.

The 1997 version of CALGOS had options both for Windows 3.1 and for Windows 95. The user interface had been compiled by Microsoft Visual C++ version 4, and the model itself by Microsoft PowerStation FORTRAN version 4.

The 1998 and 1999 versions could be run by Windows 95, 98 and Windows NT4 only. The user interface had been compiled by Microsoft Visual C++ version 5 and the model itself was compiled by Digital Visual Fortran version 5.

The COTTON2K user interface has been compiled by Microsoft Visual C++ version 6, and it makes more use of Windows 9x/NT conventions for editing, opening and saving files, etc. The model itself was compiled by Compaq Visual Fortran version 6.5 for the Windows environment.

In Versions 3.0 and later, the model itself has been translated from Fortran to C++, and both the user interface and the model itself have been compiled by Microsoft Visual C++ .NET 2003.

Version 4.0 has been optimized for Windows XP, but it can also be run by Windows 98, Windows NT4 and later versions of Windows.

Main characteristics of the Cotton2K model

This is a process-level model. It simulates the processes occurring in the soil, plant, and in the microenvironment, and the interactions between these processes and the management inputs applied to the field.

The main characteristics of COTTON2K, with emphasize on the differences between it and GOSSYM, will be summarized here. Most of these modifications, which have made CALGOS and COTTON2K more suitable than GOSSYM, or other previous cotton models, for use in the irrigated arid regions, may be summarized as follows:

 

1. Water Relationships

Potential evapotranspiration is computed on an hourly basis, using equations derived from those adopted by CIMIS (California Irrigation Management Information Service). In order to enhance the accuracy of the computation of potential evapotranspiration, a procedure for estimating hourly values of weather parameters from the daily values has been implemented. Note that in addition to the daily weather input parameters used by GOSSYM (global radiation, maximum and minimum temperatures, rainfall, wind) daily average dew-point temperatures are now also required as input.

The root sub model has been modified, especially concerning the responses of root growth and activity to differential soil moisture conditions. Average soil water potential is computed as an average for the whole root zone, weighted by root activity in each soil cell. This soil water potential is used for modifying the potential evapotranspiration and computing the actual transpiration by the plants.

Water movement in the soil is computed as a combination of implicit and explicit numerical procedures. This is done at an hourly time step.

 

Leaf water potential is computed on the basis of the average soil water potential, plant resistance to water transport, and potential transpiration. The leaf water potential is then used to compute empirical water-stress factors. These water-stress factors affect the growth rates of plant parts, aging rates of leaves and bolls, photosynthesis, and abscission rates of leaves, squares and bolls.

2. Irrigation

In addition to surface methods of irrigation (sprinklers, furrows), the option of drip irrigation has been implemented as input to the model.

The model can also be used to predict the irrigation requirements of the crop, under given weather scenarios and soil conditions, for drip as well as for other methods of irrigation.

 

3. Nitrogen Relationships.

The processes of nitrogen mineralization (from decomposing organic matter) and nitrification in the soil have been modified. Modules for denitrification, N immobilization under high C/N ratios, urea hydrolysis, and transport of nitrate and urea in the soil have been added.

Uptake of N by the plants is assumed to be affected by the growth requirements of each plant part, and it is simulated as a Michaelis-Menten procedure. New procedures have been devised to simulate the allocation and reallocation of nitrogen to all plant parts. Nitrogen stress factors are computed, and their effects on plant growth rates, aging of leaves and bolls, and abscission of squares and bolls are simulated.

4. Plant Growth and Phenology.

Leaf growth is simulated separately for blades and petioles, using the monomolecular growth function. The parameters of the growth function are dependant on the node position of each leaf. The routines for leaf aging and abscission have also been completely revised.

Boll growth is simulated separately for seed-cotton and for burrs, using improved growth functions. The logistic function is used to simulate the growth of seedcotton, whereas burr growth is assumed to be linear for the first three weeks after flowering. The routines for boll aging, square and boll abscission, and boll opening have also been revised.

5. Abscission of Squares and Bolls.

The rate of abscission of squares and bolls is assumed to be affected by carbon stress, water stress, and nitrogen stress. There is a time lag (usually 5 to 6 days, depending on temperature) between the occurrence of the physiological stress and the actual abscission caused by it. The susceptibility of each square or boll to shedding is simulated as a function of its physiological age and the severity of stress.

6. Soil and Canopy Temperatures.

The temperature of the soil surface is computed by solving the energy balance equation at the soil surface: heat conductance in the soil, incoming short wave radiation, incoming long wave radiation (from sky and from canopy), outgoing long wave radiation, sensible heat transfer, and latent heat of evaporation.

The temperature of the plant canopy is similarly computed by solving the energy balance equation at the canopy interface with the air: incoming short wave radiation, incoming long wave radiation (from sky and from soil), outgoing long wave radiation, sensible heat transfer, and latent heat of transpiration.

Heat flux in the soil is computed as a combination of implicit and explicit numerical procedures. All these procedures are done at an hourly time step. The incorporation of these processes enables the model to simulate the effect of a plastic mulch covering the soil surface.

7. Time Steps used in the Model.

Most of the procedures in the model, as in many other models, are computed in a daily time step. However, in order to increase the accuracy of the simulation, we can now utilize the enormous computing power of today's personal computers, and compute some procedures in an hourly time step. Although weather input data are on a daily basis, the model can estimate the hourly values of these data. The "heat units" (or "physiological age") concept, used to express the effects of temperature on growth rates and phenology, is now computed at an hourly time step.

The following procedures are now computed at an hourly time step: transpiration and evaporation from the soil surface, water and nitrogen movement in the soil, heat flux in the soil, energy exchanges at the soil-plant-air interfaces, soil and canopy temperatures, prediction of plant germination and emergence.

Scientific principles on which the model is based

Cotton Plant physiology

Growth rates are related to temperature, using the concept of ìheat unitsî also referred to as ìdegree daysî. This is, however, modified as follows: calculations are based on hourly heat unit accumulation, using computed hourly temperature values. The threshold value is assumed to be 12 C. One ìphysiological dayî is equivalent to a day with an average temperature of 26 C, and is therefore equal to the sum of heat units per day divided by 14.

A linear relationship is assumed between temperature and heat unit accumulation in the range of 12 C to 33 C. The effect of temperatures higher than 33 C is assumed to be equivalent to that of 33 C.

In addition to temperature, carbon stress, water stress, and nitrogen stress have a strong effect on all simulated rates of growth and development.

Carbon stress: Potential growth of each organ is driven by its age and position, as well as by temperature. These ìpotentialî growth rates are computed for roots, stems, each leaf blade and petiole, each square, and for seedcotton and burs in each boll. The sum of these potential growth rates is the ìcarbon sinkî.

Gross photosynthesis is computed from radiation, plant cover (radiation interception), temperature, CO2 content in the air, water stress, and nitrogen content in the leaf blade. Subtracting photorespiration, maintenance respiration and growth respiration results in net photosynthesis. This, together with the supply from mobilized starch stored in the leaves and in the taproots, is the ìcarbon sourceî.

When the carbon source is less than the carbon sink, the potential growth cannot be realized, resulting in a condition called ìcarbon stressî. This is numerically expressed as the ratio between source and sink (1 = no stress, and 0 = a most severe stress). There are usually two main periods of carbon stress in cotton:

  • (1) For a period of three to four weeks after germination, carbon stress is usually caused because there is not enough leaf area to sustain growth;
  • (2) Beginning 2 to 4 weeks after the start of flowering, carbon stress is usually caused by the strong sink caused by boll growth. This condition continues until most of the bolls reach maturity.

When carbon stress occurs, growth is reallocated according to the priorities of the different plant organs. Highest priority is for square and boll growth, lowest priority for stem and root growth. Carbon stress also reduces the rates of appearance of new nodes, rate of stem growth in height, and is considered to be the main cause of square and boll shedding.

Water stress: Potential transpiration is computed by the modified Penman equation (CIMIS version). Actual transpiration is modified by the light interception factor of the plant canopy, and by the average soil water potential (which is computed for soil cells containing active roots only, an average weighted by the amount of active roots in each soil cell).

Early morning (maximum) leaf water potential is derived directly from the average soil water potential. The minimum leaf water potential occurs when transpiration rate is maximal (usually in the early afternoon). The product of the maximum transpiration rate and the total plant resistance to water transport decreases the minimum leaf water potential. These leaf water potential values are the basis for computing several empirical water stress factors (where 1 = no stress, 0 = most severe stress).

The water stress factors affect rates of photosynthesis, leaf aging, growth in height, shedding of bolls, rate of boll maturation, allocation of photosynthates, and growth rates of all organs.

Nitrogen stress: Cotton2k computes the rates of urea hydrolysis in the soil, mineralization of organic N, nitrification of ammonium N, denitrification of nitrate N, and movement of soluble N (nitrate and urea) in the soil. It also computes the uptake of N by plant roots.

The nitrogen in plant organs is computed in the following way. The model first computes the N requirements for growth. Then it calculates the supply of N from uptake and from reserves. The model then computes the allocation of N to the plant organs, and the concentrations of N in plant dry matter. If supply of N does not cover the requirements for growth, the model computes nitrogen stress factors. There is a feedback of computed N requirements to the N uptake routine. Nitrogen stress affects growth, new node production, leaf aging and abscission.

Cotton Plant Phenology

The model simulates the development of vegetative branches, fruiting branches, their nodes and the associated appearance of leaves and squares in each node. This involves a number of processes and rates: production of new pre-fruiting nodes and leaves; appearance of first square; production of new fruiting branches and new nodes on existing fruiting branches. These rates are a function of temperature, stresses, and in some cases also of population density.

Soil processes

The model simulates the capillary flow and gravity flow of water and nitrate and urea in the soil, redistribution of water and nitrate and urea after irrigation (surface or drip), and evaporation of water from the soil surface.

Agrometeorology

Using daily weather data in the input files, hourly values of temperature, global radiation, etc., are estimated, and used to compute evapotranspiration, soil temperatures and plant temperatures, as well as rates of growth and development.

Input data needed

Although the model works in metric units, there are options for input and output in English units.

Climate data

For each day during the cotton season, the following weather data are needed: Radiation, Maximum air temperature, Minimum air temperature, Rainfall, Daily wind run (if not available ñ input of seasonal average is needed), and Dew point temperature (if not available ñ it will be estimated by the model).

There are two types of weather files: (1) actual weather ñ can be used when the simulation is run after the climate data were measured; (2) predicted weather ñ based on previously recorded weather scenarios.

Agricultural input data

The Agricultural Input File contains the following information about the agricultural input for a simulation run.

Irrigation application. For each irrigation that has been applied, or is planned to be applied, the following data are required: Date, Effective amount of water applied (in inches or mm), Method of irrigation (sprinkler, furrow, or drip), and Location of the drip tubes (if it is a drip system) - the horizontal distance is measured from the mid-point between two plant rows, and the vertical distance is measured from the soil surface (inches or cm may be used).

Irrigation prediction. The model can be used to predict the optimal irrigation regime. The following data are required for using this option: Dates of starting and stopping the predicted irrigation, Minimum number of days between successive irrigations, Maximum amount of water to apply in each irrigation, Method of irrigation (sprinkler, furrow, or drip), Location of the drip tubes (if it is a drip system), Required depth of soil to be wetted (if it is a furrow or a sprinkler system). The recommended wetting depth for cotton is usually 90 cm (or 36 inches).

Nitrogen fertilizer application. For each application that has been applied, or is planned to be applied, the following data are required: Date of application, Effective amount of nitrogen applied as NH4, NO3 or urea (in lbs. per acre or kg per hectare), Method of application (broadcast, side dressing, foliar, or drip), Location of the application (if the method is side dressing or drip).

Cultivation. For each cultivation the following data are required: Date of cultivation, Depth cultivation

Defoliation. For each application of defoliation the following data are required: Date of application, Method of application (broadcast, sprinkler, or banded), Band width (in cm or inches), if it is a banded application, Rate of application, Units of application rate (lbs. per acre, gal per acre, oz per acre, acre per lb., or acre per gal).

Defoliation prediction. The model can be used to predict the optimal defoliation regime. The following data are required for using this option: Percentage of boll opening at the first defoliation (usually between 65% and 90%), The date to defoliate even if boll opening has not reached the defined level.

Pix application. For each application of Pix the following data are required: Date of application, Method of application (broadcast, sprinkler, or banded), Band width, if it is a banded application, Rate of application, Units of application rate (lbs. per acre, gal per acre, oz per acre, acre per lb., or acre per gal).

Water table and salinity data. For cases of shallow water table conditions (less than 2 m), or saline soils, the following data are required: Date (of start of this water table or salinity condition), Depth of water table (in cm below soil surface), Soil salinity, from saturated soil extract, measures in milimhos per cm (dS/m in SI units), averaged for the soil layers with active roots

Soil characteristic data

General Soil Properties (common to all soil layers):Soil Water potential at field capacity - The default value for this property is -0.3 bars, but it may vary in extreme sandy or clay soils. Use bar units.Soil Water potential for free drainage - The default value for this property is -0.15 bars, but it may vary in extreme sandy or clay soils. Use bar units.

Properties specific for each soil profile layer: Up to nine layers can be defined. If there is no detailed information about this soil, at least one layer should be defined.Depth to the end of this soil layer.Parameters for the 'van Genuchten' equation - This equation relates the soil water potential to the soil water content. There are four parameters: Residual water content (by volume); Saturated water content (by volume); Alpha coefficient; Beta coefficient.Hydraulic conductivity - Input at least either the saturated conductivity, or the conductivity at field capacity. This will be the basis for computing the relationship of actual hydraulic conductivity with soil water content. Use cm per day units. Soil Bulk density.Clay and Sand content - percent by weight of dry soil.

Soil initial data

Soil data at the start of a simulation run. The data are for eleven successive 15 cm (or 6 inches) deep layers of the soil, and a 12th layer which extends down to the bottom of the soil slab (total of 200 cm, or 80 inches):

Water content, as percent of field capacity.Soil nitrate and soil ammonium content, as kg per ha, or lbs. per acre of N for each layer.Soil organic matter, as percent of soil dry weight.

Simulation Profile data

The profile file defines a single simulation run. It points to the other input files used for this simulation. Other data needed for this file:

Site location: Latitude and longitude (degrees), elevation above sea level.

Dates: Start and end of simulation, Planting or Emergence.

Crop Data: The cultivar used (a number of cultivars, for which the model has been calibrated, are available now. If another variety is actually used, choose the calibrated variety that is most similar in its phenology and morphology).

Site: Choose one of several availables sites. (The climate functions have been calibrated for the California San Joaquin Valley, the Arizona Phoenix-Tucson area, and the Israel coastal plain and upper Galil. Choose the site closest to your actual climatic conditions).

Field data: Row spacing, Number of Plants per Row Unit, Skip-rows (Yes or No; if Yes - input skip row width). Skip row width is the smaller distance between two adjacent rows. When skip rows are defined, "row spacing" will mean the average distance between rows.

The profile file is also used to indicate the required optional outputs (some basic output will always be produced), and the units of the output (English or metric units). It also indicates if site numbers and weights of plant parts will be output on a per plant or per unit area basis.

The output data

There are three types of output files:

1. Text files : Summary of results; summary of input; detailed daily output; plant maps; plant vigor data.

2. Charts : graphs of 18 output variables.

3. Soil maps: two-dimensional graphs of the soil slab.

Problems of calibration for new cultivars or new areas

The model has been validated using extensive data sets from California, Arizona and Israel. It has presently been calibrated for the following cultivars: Acala SJ-2, GC-510, Maxxa, Deltapine 61, Deltapine 77, and Sivon.

Cotton cultivars differ in many of their properties. This is expressed in the model by the values of parameters used in equations describing the following:Rates of leaf growth, stem growth, square and boll growth.Rates of appearance of new nodes (prefruiting, main stem, fruiting branch).Time to square, to flower, and to open boll.Susceptibility of abscission to stresses.

The weather related procedures have been tested and calibrated for the following regions: California San Joaquin Valley, Arizona (Phoenix - Tucson area), Israel Coastal Plain, and Israel Upper Galil (Hula valley area).

Differences between sites (geographical areas) are expressed in the model by the values of parameters used in equations describing the following:Estimates of hourly wind speeds and hourly dew point temperatures.Estimates of cloud type correction for computing hourly long wave radiation emitted from the sky.Difference between time of daily maximum temperature and solar noon.Estimates of daily deep soil temperatures (at 2 m depth).Estimate of hourly relative humidities.

The latest version of Cotton2K enables advanced users to create their own calibration files for a new cultivar or for a new site.

Possible uses of the model by extension staff and growers

Education: Use output files and charts to show the response of cotton to different irrigation or nitrogen fertilizer regimes, planting dates, plant density, or to different weather scenarios.

Management: Use the ëirrigation predictioní option for planning irrigation regime for different soil types and weather scenarios. Try to eliminate Nitrogen stress by modifying the fertilizer regime (Hint: disregard N stress that coincides with C stress). Use the ëdefoliation predictioní option to get an idea when to defoliate.

Using plant mapping results : As a result of pests or diseases, or deficiencies in P or K or other nutrients, or incorrect input data ñ actual plant development in the field may be significantly different from model predictions. In this case, create a ë*.MAPí file from plant mapping results, and rerun the model any time during the growing season, using this file as input. Also, actual weather data will replace the predicted weather data for any rerun during the growing season.